Australia May 'Pause' Trades To Tackle High-Frequency Trading

angry tapir (1463043) writes "The Australian Securities and Investment Commission (ASIC), a government financial watchdog, is reportedly contemplating the idea of implementing a 500 millisecond delay on trades in an effort to put the brakes on high-frequency trading. ASIC last year knocked back the idea and stated that fears about HFT were overblown. However, in a government inquiry today representatives of the organization said the idea of a 'pause' is still on the table."

Comments

Won't work

[EmagGeek]

If you simply change everyone's temporal frame of reference by the exact same amount, you have done nothing, really. Everyone will simply account for the 500ms delay, and trades will still execute in the same order.

Re:Won't work

[captainpanic]

The way I understand it is that traders (computers) have to hold on to shares for a minimum of 500 ms, which means that whatever the market does in those milliseconds cannot be acted upon. However, others can act in the meantime.

Personally, I think that it should be law that if you buy shares in any company (or fund or whatever), you have to hold on to them for a minimum of a week or a month. Shares represent actual physical companies which own factories and employ real people. Those things don't change in 500 ms. They change over a much larger amount of time. And I believe that the stock market would be healthier if this was reflected in its trading. Obviously, when new information comes out (press release: "The factory of company X has just gone up in flames"), everybody's counter should be set to zero, but shares sold in such a case cannot be bought back a fraction of a second later (because whoever just bought them has to hold on to them for a week/month).

I don't pretend that this plan is waterproof. I'm sure someone will shoot a big hole in it in the replies below... I just wish that the stock market would represent what it's supposed to represent: a place where people can invest in our real economy.

Re:Won't work

[Anonymous Coward]

The problem isn't the order per se, it's disparate knowledge. High frequency traders can "jump the queue" because they can react faster to the same information. The point of a delay isn't to disable their execution advantage, it's to disable their knowledge advantage. When you can't get an order in faster than 500ms, you can not know the market movement in the 500ms before the order is in. They can still jump the queue, but they would have to do so "blindly", which prevents most of the shenanigans that HFT is used for.

Re:Won't work

[operagost]

Obviously, when new information comes out (press release: "The factory of company X has just gone up in flames"), everybody's counter should be set to zero

This is enough to show why your idea won't work... unless you plan is really to collapse the economy. What information is major enough to allow immediate sales of stock, and who gets to choose?

Re:Won't work

[JoeyRox]

It will work. The majority of HFT's illicit profits accrue from speed arbitrage *between* the exchanges, not from a speed advantage at any particular exchange. A co-located HFT server at an exchange sees an order, and, in anticipation of that order representing a larger order that can't be filled in full at that same inside "best" price at that exchange, trades ahead of the order by sending a buy/sell order to other exchanges faster than the original buyer/seller can, resulting in a riskless vig for the HFT trader. By delaying orders on all exchanges by 500ms, the benefit of early-access to incoming orders on any particular exchange is eliminated because all the exchanges will have 500ms of order price discovery incorporated into their SIP, the consolidated price representing the aggregate of the best prices for all the exchanges.

Re:Won't work

[OzPeter]

I just wish that the stock market would represent what it's supposed to represent: a place where people can invest in our real economy.

I purpose the the stock market should really go back to its roots, and that every share should be attached to a genuine item of stock - be that cow, pig or chicken. And that you are responsible for housing and feeding all the stock that you own.

This would also have the interesting effect of changing our perception of Bull and Bear markets.

Re:Won't work

[gurps_npc]

Incorrect. One of the major issues is that HFT look at multiple markets. They see a trade go down in market a, than instantly - in less than 100 ms, change their own order in market b.

By putting in a delay of 500 ms, you prevent this kind of behavior.

Why is that behavior bad? Because for high volume traders they have to split a single order over multiple markets - mainly because others are ALSO splitting among multiple markets.

That is, say I have 500,000 shares to sell. Currently 5 different markets all show a price of 35.2, at 100,000 shares being offered

Moreover, all 5 markets offers are from you, as you are the main guy buying right now.

It is NOT fair for you to take 100,000 of my order at 35.2, then instantly cancel your four other 100,000 orders and replace them at 35.4

You offered to buy all 500,000 at 35.2, not just that 100,000 and should not be allowed to cheat me by raising your price for the remaining 400,000.

A delay of 500 ms means you can't see that your first order is executed until after all your other orders are ALSO executed.

This is one simple example of how HFT try to unethically game the system.

Re:Won't work

[frinsore]

The problem with requiring someone to hold onto shares for a specific amount of time is that it doesn't prevent HFT but instead adds a barrier to entry. If I buy 5 shares of IBM but then must wait a week before I can sell those 5 shares what prevents me from selling these other 5 shares I bought last week? If there was a company that held several stocks that were readily trade able other people could contract selling the shares & replacing them for a nominal fee. This would be similar to shorting stocks.

One possible solution would be to trade shares on a fixed schedule, ex: every minute but not between minutes. This would provide everyone the same amount of time to react to the information. The difficult part would be deciding between who wins the bids, since time is an unfair determining factor that leaves some other metric like price bid or bidder's history or a random element.

Re:Won't work

[KingOfBLASH]

Well let's say you want to buy a share, who do you buy it from? Or let's say you want to sell a share, who do you sell it to?

It used to be you'd actually have to find someone to step in and take the contra side of your transaction. That's a pain in the ass, will cost you time and money, and in the event you need to sell and everyone else wants to sell you're screwed. All of this would mean that unless you had lots of money to invest, the stock market was not for you.

Fast forward to today. We have people willing to take a position, any position. They provide "liquidity" for the market by buying the share you wanted to sell, in the hopes that they can turn around and sell it for a fraction of a cent more when someone comes along with a buy order. They actively manage their inventory of shares (yes that's a thing), and adjust prices in the event information comes out causing a large price change in the shares.

This is a service that needs to continue if you want modern markets to maintain their efficiency.

Now here's the problem. Back when the "marketmakers" were actual human beings buying and yelling at each other in trading pits one would not be substantially faster than another. But, using computers, there's an arms race for speed. If you can get a few miliseconds (or even nanoseconds) faster than your competition, you can take all of the profitable orders. This means if you plough enough money into speed, you can just own the market. In addition, because computers are so fast, your computer can make many millions of silly trades before a human trader can push the big red stop button.

Now a solution needs to come about. But, because of the need for market makers speed can't really be limited to holding onto shares for months. (Sorry). 500 ms basically breaks the arms race since it's a very easy speed to obtain. So, you can't just plough money into being the fastest kid on the block.

Re:Won't work

[Immerman]

I agree this would seem like a good solution. Perhaps lowering the limit to a day or an hour though to keep things fluid. As long as the time period is considerably longer than the communication delays you've removed a lot of the ways in which HFT can game the system. A short enough time period would also pretty much remove any major advantages to "resetting the clock" - Yeah, it sucks if you bought stock in a company ten minutes before the factory catches fire, and will have to wait for another 50m before you can sell, but there's always a certain amount of unavoidable risk in the market - I doubt increasing it by such a small margin will have a huge effect.

Re:Won't work

[captainpanic]

There's basically 3 types of information for traders:
1. Official public information from the company itself (press releases, annual report, etc).
2. Analysis by external parties
3. Unofficial (non-public) information from within the company - trading using this is called insider trading, and this is already against the law.

So, when companies release information (category 1), everybody can trade. I do realize now that large companies have press releases on a daily basis. So, maybe the timespan of a week/month was a little long. But at least something that reflects the timespan between press releases (hours? at least not milliseconds) would still be a good idea.

Re:Won't work

[heypete]

Personally, I think that it should be law that if you buy shares in any company (or fund or whatever), you have to hold on to them for a minimum of a week or a month. Shares represent actual physical companies which own factories and employ real people. Those things don't change in 500 ms. They change over a much larger amount of time. And I believe that the stock market would be healthier if this was reflected in its trading. Obviously, when new information comes out (press release: "The factory of company X has just gone up in flames"), everybody's counter should be set to zero, but shares sold in such a case cannot be bought back a fraction of a second later (because whoever just bought them has to hold on to them for a week/month).

A week or a month might be a bit too long, but something along the order of 1-5 minutes might be reasonable.

Alternatively, one might also have the exchange do batch orders: traders submit their orders to the exchange, the exchange groups them all together, and then processes them all periodically (say, every 30 seconds or something), then displays the results. Since the results are not released until after the batch is fully processed there's no advantage to submitting an order at 29.999 seconds compared to any other time within that window. This way trades can be executed reasonably quickly on a human scale and HFT doesn't have any particular advantage.

Re:Won't work

[ysth]

There's no need to set a minimum time; what is needed is a minimum tax or fee. It could be .01% and still completely put a stop to abusive trading.

Re:Won't work

[L4t3r4lu5]

It's front-running by machine. If a person-trader did this, they'd be in jail.

"... the illegal practice of a stockbroker executing orders on a security for its own account while taking advantage of advance knowledge of pending orders from its customers."

Re:Won't work

[Anonymous Coward]

The 500 ms delay will NOT solve the problem. I will by 100 shares every millisecond and keep it for 500 ms and sell it. So I will always have 50000 shares at any moment. This way I will not gain or loose money since at every ms I am buying and selling 100 shares at (almost) the same price. If I really want to "trade" in the old sense, then I just delay the selling or buying of those 100 shares, and I am in business.

Re:Won't work

[Gr8Apes]

It would be better to have random delays introduced from 0-30s, which causes out of order sequencing on trades, making HFT relatively unreliable and unusable, since the high speed links currently used to facilitate those ms advantages will be entirely negated.

Re:Won't work

[Anonymous Coward]

If you simply change everyone's temporal frame of reference by the exact same amount, you have done nothing, really. Everyone will simply account for the 500ms delay, and trades will still execute in the same order.

Then you have no idea how HFT works.

Re:Won't work

[Anonymous Coward]

It would not be practical to use your metrics for resetting trade restrictions as there is no system in place for notifying the exchange of every official statement made by a company (as an example running a new TV commercial, or updating their web sight or Facebook page would qualify).

Re:Won't work

[Anonymous Coward]

It would be better to have random delays introduced from 0-30s, which causes out of order sequencing on trades, making HFT relatively unreliable and unusable, since the high speed links currently used to facilitate those ms advantages will be entirely negated.

This does seem like the most solid solution.

You can probably even set the delay a lot lower. A delay between 0 and .5s if randomly assigned to every trade would completely wreck the speed advantage HFT system rely on.

Re:Won't work

[Anonymous Coward]

My thoughts exactly, they will still game the system, only time is quantized at a different interval now.
I think they should delay the trades variably and inversely with the volume of shares traded.

So the folks that execute massive HFT trades will have the longest to wait, and less ability to game the system over the small traders that aren't moving around as much money "influencing" the system to their benefit as the HFT traders do.

Re:Won't work

[squiggleslash]

Nope. That's not relevant to preventing HFT.

Quick explanation of HFT, at least, the form that hit the headlines recently. Suppose you have two exchanges, let's call them VAMPIRES and ANGELS (to pick names at random, I mean, if I've accidentally used names reminiscent of a real exchange I apologize as that's not my intention...)

A legitimate trader wants to buy 1000 shares of X at $10. It sees 500 on VAMPIRES, and 500 on ANGELS. So it immediately, simultaneously, places both orders.

Well, it turns out VAMPIRES is kinda rigged. They've made sure their connections to every trader are the fastest possible, whereas ANGELS just uses regular telco connections. They advertise this as a feature, and everyone believes them, but it turns out the founders of VAMPIRES have a hidden agenda. They've made sure they have a fast connection to VAMPIRES, and arranged with the phone company to have an equally fast connection to ANGELS. After "resigning" from VAMPIRES to give the appearance of being uninvolved, they monitor all transactions on VAMPIRES. As soon as they see all shares for X have been told at $10, they immediately place an order for all shares of X at $10 on ANGELS, correctly deducing that the only reason someone would buy ALL the shares of X on VAMPIRES is because they're buying ALL the shares on ALL exchanges for X for $10.

Because they have a fast connection to both exchanges, the HFT traders can see the trade that just happened on VAMPIRES and successfully transmit their trade to ANGELS in less time than it takes the legitimate trader's trade to be transmitted to ANGELS.

So what does the 500ms delay do? Answer: it makes it impossible to see the trade that occurred on VAMPIRES before the accompanying trade has been received by ANGELS too. The founders of VAMPIRES sees the trade 500ms+latency after it was sent by the legitimate trader. They can place the order on ANGELS anyway, but their trade will arrive 500ms+theirlatency-legittrader'slatency on ANGELS so it'll arrive afterwards and the legitimate trader will get their shares unmolested.

An alternative, but it's not terribly reliable, is for the legitimate trader to determine the latencies to each exchange and then send each order with an appropriate delay to make sure they arrive at about the same time at each exchange. It's not 100% fool proof, but RBC was able to get around HFT traders using the technique.

Re:Won't work

[cryptolemur]

I gather the best way to 'encourage' investors to aim for long term profits, would be to simply make the tax be absurdly high (like 99.9999%) for HTC and then converge it to normal according to the time one has held a particular stock before sale. This way you can always make profit (if there's profit to be made), but even the gambler would be interested in the long term health of the general economy, and of the business in particular they have invested in.

Overnight, we'd have a stable, healthy, growing economy.

Re:Won't work

[rahvin112]

And then we go back to the grand old days of the market where a broker (called market maker) skims 3-5% off every trade because he's the buyer and seller of last resort.

Personally I'll take the HFT.

Re:Won't work

[romco]

>By delaying orders on all exchanges by 500ms, the benefit of early-access to incoming orders on any particular exchange is eliminated because all the exchanges will have 500ms of order price discovery incorporated into their SIP, the consolidated price representing the aggregate of the best prices for all the exchanges.

If you delay everyone 1/2 second some servers will still have an advantage. I think the 1 to 5 second delay would work better.

Re:Won't work

[PRMan]

I think a 30 minute or 1 hour limit would take care of your concerns.

Re:Won't work

[PRMan]

They get around it by looking at OTHER PEOPLE'S orders (as if that's better).

Re:Won't work

[PRMan]

This was the example given on the 60 Minutes piece.

Re:Won't work

[Anonymous Coward]

1) Put in a limit order at price X. You won't get worse than price X, unless you were hoping to get a better price from the generous market?

2) When you're sweeping the market at a particular price, algos are going to recognize a sweep is happening and are going to pull quotes on other exchanges, which will drive up the price. Market sweeps usually infer that the buyer knows something the market doesn't, and lots of algorithms will play defensively against that sweep.

Re:Won't work

[Anonymous Coward]

Now here's the problem. Back when the "marketmakers" were actual human beings buying and yelling at each other in trading pits one would not be substantially faster than another.

Back when market makers were actual human beings, they were far more corrupt and screwed the investors (legally and illegally) to a staggering degree. I am bemused that a tech-savvy crowd like slashdot believes computers -- as used in HFT -- could come anywhere close to the insidious cleverness of humans in manipulating the markets.

HFT is a godsend to everybody but the NYC "specialists".

Re:Won't work

[Anonymous Coward]

It is not cheating you - I'm looking for someone to buy 100,000 shares at 35.2. If my inventory shrinks, I'm not as desperate to sell and my price rises.

Re:Won't work

[TheRaven64]

The important issue is the ratio between investors and speculators. You need speculators in the market to provide liquidity, but you don't want too many because liquidity is the positive spin on volatility. If you have too high a ratio of speculation to investment then the market becomes completely decoupled from the thing it's trying to represent and it becomes a dangerous place for investors (and companies) because they can lose all of their money as a result of something completely unrelated to the actual profitability of the company. If you have too few speculators, then it becomes difficult to buy and sell shares.

The problem with HFT is not really HFT itself, it's that it magnifies the effects of speculators on the market, meaning that you need far fewer speculators with far less capital to have a disproportionate effect on the functioning of the market.

Re:Won't work

[Dare nMc]

I think a better option would be a fee that is triggered off a pattern of undesirable trading patterns over a extended time period (perhaps a year). IE brokerages that has a high percentage of short term trades gets a extra fee, along with customers and funds that do the same. Not preventing a single (or a few) actions.

> you have to hold on to them for a minimum of a week or a month

In general I think that would make momentum stocks much worse, more bubbles and crashes. IE if a stock takes off going up, you will create a bubble anytime you lock out very many sellers. Also locking out sellers in one often locks out buyers in another stock, thus creating more market volatility, as well.
For example as a fundamentals trader I look at a industry that looks promising, and find individual stocks and place a value on it, then look for stocks to give me a chance at a desired return, but expect half of that. For example I might decide I like ag-equipment, and put a values on Cat for $90 and Deere for $85. I buy the one at a better value, if I buy cat at 8am for $85, it is at $90 later in the day, and deere drops to $80 that same day, despite planning to hold cat stock for 2 years to get 10% return, I would take the 5% today in Cat (preventing a price bubble) and instead buy the new better value (Deere in this case.) The point of the market is to keep valuations in check (not too high or too low), so preventing me from selling the Cat stock, and buying the Deere stock hurts the market arbitrage. For example I have had an entire year not selling a single stock, and I have had days where I bought and sold several stocks the same day multiple times. Encouraging people to hold a stock (or houses, businesses...) despite it being overvalued, causes more problems than HFT.

Re:Won't work

[Anonymous Coward]

  As soon as they see all shares for X have been told at $10, they immediately place an order for all shares of X at $10 on ANGELS, correctly deducing that the only reason someone would buy ALL the shares of X on VAMPIRES is because they're buying ALL the shares on ALL exchanges for X for $10.

This is not really a common case. More commonly each HFT has sell orders out on every exchange, but really only wants to trade on at most one or two of those orders. As soon as they get a message that their standing order at ANGELS executed, they cancel their other outstanding orders, including at VAMPIRES. If all the standing orders at VAMPIRES were from HFTs then there is no more quantity at $10 and the price hops up to the next level, $10.01.

To the buyer, it looks like they've been front-run (somebody took out all those $10 orders!). But really it's just the HFTs wanting to limit their exposure and canceling standing orders. Katsuyama likely understands this but explained it misleadingly.

If HFT were eliminated from the market, the phenomenon migh twell disappear. But it's pretty likely the market would widen from bids at 9.99 and offers at 10.00 to bids at 9.97 and offers at 10.02, with those standing orders coming from "slow" traders. The HFTs would no longer be making money, but the slow market makers would be collecting even more money than HFT ever did.

Personally I think the most insidious issues pointed out by Lewis are not really HFT related, and it's a pity they aren't being discussed with the hysteria aimed at HFT:

(1) A large and increasing proportion of market activity is in "dark" pools that are not required to report their orders, limiting the ammount of information available to regulators and the public. Brad Katsuyama's IEX "exchange" is actually a dark pool, and he's being rather disingenuous about that.

(2) Brokers like Charles Schwab, whose customers are the little guys, are sending all their customer orders to hedge funds like Citadel for execution, where those orders end up mostly in dark pools and other internal crossing networks. Citadel and its type, in turn, use HFT techniques to handle the overflow. Arguably the best information discovery comes from individual investor flow, so why is Schwab allowed to hide it like that in return for a payment of a hundred million dollars?

Re:Won't work

[bobbied]

If you simply change everyone's temporal frame of reference by the exact same amount, you have done nothing, really. Everyone will simply account for the 500ms delay, and trades will still execute in the same order.

The problem with your idea is that "trading" can be done outside of the exchanges too. You can stand on the street corner and trade with a passersby or do the digital equivalent if you wanted. In fact, many traders do just this. They have back channel trading connections that they use to shave fractions of pennies off of trades on one exchange by using another. There is no way you can force trading to be done only at specific times in these situations and if you apply this rule only to the exchanges, then the high frequency traders will only have another advantage.

One could put some uncertainly in the trading process, by randomly choosing orders which are executable in the queues or introducing random delays, but I still think that there is no real way to fix this nor is there an easy answer for this problem.

About the ONLY thing I can figure *might* work is to make programed trading totally transparent to all. Make everybody who is using some kind of automated trading platform publish their trading logic for all to see before they can use it. Require a minimum amount of time between publishing of the algorithms and when they can be used. They should also outlaw the use of anything beyond level 2 quotes in automated trades on an exchange and require the publishing of all trades which are automated within a short time frame (say a few seconds) of their execution. Automated trading is defined as a computer program filling out all or part of an order, setting price, limits, stops etc. If a human enters the order, it's NOT automated trading, even if they are copying the data from an automated system, but they must TYPE in each part of the order to enter it, no Cut and Paste, or just selecting "GO" from a list.

Even that idea has problems...

Re:Won't work

[jalopezp]

Batch orders are seemingly a good idea. Several markets operate in an auction style batches, where each participant presents their schedule (a list of possible prices coupled with an amount to buy or sell for each price) and at set intervals a computer determines the price and quantity that clear the market. We have algorithms that do this in a utility-maximising way, but it is definitely more effort on the part of the traders. Electricity exchanges are the only ones I can think that use this method. APX, for example, does this for day-ahead UK power trading.

But you don't actually need auctions for this to work. You can use a similar public-outcry method (just like in the NYSE) but separated by breaks. The London Metal Exchange does this in its trading sessions. They are split up into 5 minute intervals, each of which can only be used to trade one particular metal. It's pretty cool to watch.

Both these markets are very liquid, so I doubt splitting up the trading day of even the busyest exchanges (NYSE, NASDAQ, LSE) would cause any perceivable damage to the market's fluidity.

Re:Won't work

[Anonymous Coward]

You really think that a small tax would put things back to the days of the 1970's with respect to spread? Especially with the rest of the computerization and price pressure that the market has brought to bear? Not to mention that, back then, quotes were only given to the hoi poloi in 1/8 dollar increments? You really think we'd move off the 1 cent increments as a result of a minor tax?

You, my friend, are either a shill or someone with a vested interest or hallucinating. If it's the latter, please let us know what you're on and where we can get some - the paranoia effects alone are quite fascinating.

Re:Won't work

[Anonymous Coward]

I think you screwed up your example. But setting that aside, if you don't want to sell your shares at 35.2 instead of 35.4, then don't.

Re:Won't work

[suutar]

This is essentially what IEX does (described in this article); putting in the delay does have an effect because it increases the minimum turnaround time which reduces the "fast" players' ability to see what the "slow" players are doing and react. Or, put another way, a 1ms round trip can do 100 times as many transactions as a 100ms round trip, but a 501ms round trip is not that much better than a 600ms round trip.

Re:Won't work

[Anonymous Coward]

No, you don't understand the problem.

The problem is that HFT produces orders at one price, and then cancels them when someone tries to buy it, call it arbitrage, or call it cheating. So when you try to buy (without a limit order) you will always be ripped off, as you won't get it for the advertised price, but instead some larger arbitrage gap.

In NY the figured out the solution was to add 350ms on their own exchange so that HFT can't do that. This AFAIK has not been implemented on any exchange but one, but it is the cause of current investigations into HFT in the US.

The delay makes it impossible to jump-ahead. BATS exchange was jumping ahead in the US because it was the closest to NYC Stock exchange.

Re:Won't work

[Gauchito]

Why is it not fair?

I'll tell you what will happen if you add that rule: There will be no market for you to hit at 35.2 in the first place. The extra liquidity provided by the HFT let's you get part of the trade at that price. Why are they obligated to trade with you? If you really want to buy at 35.2, leave your order there. If someone else wants to trade with you, great, you'll get your fill. If not, then I don't see why anyone should.

In your scenario, you got part of your trade at 35.2. That's great. If the liquidity provider hadn't been there, the market would have been wider to begin with, and unless another big player like you wants to take the other side at 35.2 (lower probabililty than a market maker being there), you'll be paying more for both legs of your trade anyway.

Look at it this way: the guy you just bought from is now short, and knows there is someone out there willing to buy more than he got hit for. He's in a tougher spot than you (that's market making), he needs to buy back as well, and you can probably afford to pay an extra tick and still make money. If he pays an extra tick to close out of his position, he's out $1000. If he makes money, great for him, but it's no difference to you. You're taking advantage of the liquidity he's providing. If he wants to cancel to reduce his risk, fine. You still got a cheaper price on part of your trade than you would have without him being there. How do I know? Because he's the only one at that price in your example.

What you want is the same liquidity, but without the chance of that liquidity going away when you take an action in the market. In essence, you want public exchanges to act as dark pools, you want your buying interest to only be reflected in the market after all your buying is done. If you do that consistenly, there will be no point to market making, and watch the markets widen again, and you having to call brokers and pay through the nose for them to handle large volume for you. The current system is better.

Re:Won't work

[Anonymous Coward]

The way I see it it's not much different from insider trading. It's acting on information that is not yet available to other parties.

Re:Won't work

[mederbil]

HFTs are not front-running. Read this and educate yourself: http://clarkgaebel.com/you-are....

If you don't route your orders correctly, other traders can take advantage of that. Your trade becomes public knowledge when it is filled and other traders can take advantage of that information at other exchanges. There is no knowledge of the rest of your trade being sent to other exchanges, thus it isn't front running. It's basically just a good guess.

Re:Won't work

[Actually, I do RTFA]

To defend the GP's point:

What information is major enough to allow immediate sales of stock, and who gets to choose?

Well, on one side we have information about disasters involving the company (factories exploding, recalls, CEO death). That clearly is major enough.

On the other side we have information about how the rest of the market values that company, taken to its extreme with HFT. That clearly is not major enough and what is trying to be blocked.

I think we can all agree that there is a lot of grey in between the two extremes. But the difficulty of arriving at a solid point does not abrogate the concept, any more than sorites paradox prevents a heap of sand from existing.

But, I put forth, that although this is a way of reviving the GP's point, it is not the only solution.

We could easily, and perhaps profitably, allow no information to reset the counter. What would the advantage be?

Re:Won't work

[Actually, I do RTFA]

You seem to be the first person who is actually trying to explain something instead of just opaquely insisting "HFT adds liquidity'. So, since I don't understand that position at all, allow me to ask a series of questions that hopefully help me out.

You're saying when there are 500k being offered at 35.2 (split into 100k across 5 exchanges), in reality there is only 100k being offered on any of 5 exchanges, because as soon as 100k are snapped up, they'll withdraw their offer on the other 4 exchanges an up them to 35.4. And that this is expected and ethical, because otherwise they would have started at 35.4?

But even HFT seek to make money, so why wouldn't they start at 35.4 if they thought they could get it?

Do we have any evidence that this actually lowers the spreads? I mean, I understand there is a correlation, but electronic trading in general lowers the spread, plus it seems timed to the switch to decimal. Has anyone controlled for those factors?

You mention market making. But isn't that factor only for thinly traded stocks? Wouldn't widely traded stocks already have a market.

How does this "extra liquidity" occur? I thought HFT was arbitrage. Arbitrage, pretty much by definition, can only occur when you already have both parties lined up for you to transact with.

Why do you say that he's the only one at that price. Presentably, there could be infinite people offering at that price, all of whom withdraw their offer and up it. I mean, that may be a different example, but it seems like otherwise you're picking on a poorly constructed example instead of a proper point.

Re:Won't work

[Gr8Apes]

My solution was intended to not only nix the advantages of those with higher speed links over other high speed links, but also remove the advantage over those without high speed links. 30s may actually be too short. Perhaps a 5 minute window would be more appropriate.

Re:Won't work

[pnutjam]

The difficult part would be deciding between who wins the bids, since time is an unfair determining factor that leaves some other metric like price bid or bidder's history or a random element.

That's easy... of course my brother-in-law/fraternity brother/ex-coworker wins. Same as now.

Re:Won't work

[MatthiasF]

What you just described is a huge farce.

The unnaturally high liquidity being created by HFT is actually being supported by most of the exchanges (through kick-back fees) not because it makes it easier for investors to buy and sell shares, but because the HFT is hiding a huge market correction that has been happening since 2009.

While billions of dollars in shares are being passed along at high speed, the rate of decline on shares during a correction slows to a crawl because it is hiding investor sentiment beneath the high speed veil. Since you can make more money on a stock gaining (sky is the limit) over declining (zero is the bottom), almost all of the HFT is aimed at gains and not losses.

So, over the last five years, the HFTs have been pumping stock prices up while actual volume of trades HAS DECLINED.

Look at any graph of stock trade volume over the last 10 years. You will see a very steady downward trend on volume and an unnatural upward trend on stock prices.

http://finance.yahoo.com/echar...^DJI+Interactive#symbol=^DJI;range=5y (Hit Max)

You might also notice two huge days of volume on December 6th and 19th of 2013. If you can figure out what happened on those two days, you'll discover where your parent's retirement went.

Re:Won't work

[Anonymous Coward]

You simply lack basic understanding of the feedback loop made possible by one exchange, and front running made possible two or more exchanges.

This is not a network problem. The sum are more than its parts. Etc, etc.

Captcha: analytic

Re:Won't work

[Anonymous Coward]

Can you explain why this would be the case? Why can't a seller and buyer meet up without a middle man?

Re:Won't work

[towermac]

World of Warcraft (and I guess other mmos) do this already - the one second global cooldown. It's the great equalizer. Pretty much everybody (unless you're in Iraq) has the same latency. Of course, in WoW, most people want things to be fair; even the ones that live across the street from Blizzard.

There's no real incentive on Wall Street to be "fair" though.

Re:Won't work

[BasilBrush]

And you think you know better than The Australian Securities and Investment Commission because...?

If you don't know how it will help, that's more likely to mean you don't understand than ASIC don't.

Re:Won't work

[KingOfBLASH]

Utter rubbish.

I used to work for a high frequency firm, and I can tell you the downwards strategy is as important as the upward strategy. While in theory it is true that potential gains on the upside are unlimited and potential gains on the downside are limited to the share price this presupposes that a stock price can go to infinity. Realistically, this is not true, and realistically barring a catastrophe a stock does not go to zero. Even then a stock does not go to zero immediately when becoming worthless (see Dead Cat Bounce)

HFT is focused in on short term moves. On a short period of time, say the order of seconds, it will dominate. On a longer period of time, say hours, days, or weeks, investor sentiment will dominate. Even when HFT goes completely wonky and accidentally manipulates the market, over a period of days investor sentiment will still dominate.

Additionally, stock ownership represents ownership in an underlying company, that on average will experience growth. As profits grow, the value of the shares grow. Therefore growth in the present value of stocks is the natural state. Additionally, while you might take the counter argument that not all stocks should grow, most indices will cut out underperforming stocks. Therefore, an index is a bad indicator for the stock market at large because you are artificially selecting for winners and losers.

Re:Won't work

[davester666]

Except they game the situation. For example, you see an buy offer for $X for Y shares, you go 'sell $X for Y shares'. They get to see your sell offer, remove their buy offer and put in a new buy offer of $X-fraction for Y shares. Lather, rinse, repeat.

Re:Won't work

[Kaenneth]

They would just be selling contracts to buy/sell in advance of the trading window; actual ownership wouldn't change, but if would be bought an sold on a futures market

http://en.wikipedia.org/wiki/F...

Re:Won't work

[gurps_npc]

Because you are trying to force the guy doing a market order to instead take a limit order, and allowing the guy who is claiming to be do a limit order to actually be doing a market order. There are advantages to doing a market order - instant sale (liquidity). When I do market orders that is what I get.. There are advantages to doing a limit order - guaranteed price - at the cost of liquidity.

Your method is designed to let an unethical criminal who happens to spend a lot of money to create a HFT trading platform get to create a new kind of order that has the advantages of both a market order AND a limit order. He is NOT entitled to that liquidity. He traded it away in order to get the better price.

And that advantage is ONLY available to wealthy unethical criminals that have access to HFT. (note I am not saying that all HFT are unethical criminals, just that some of them are.).

What I am saying is that if you want the liquidity fine, you have to pay the price of liquidity - a lack of control over price. If you want control over price, you have

Markets are by law AND by theory supposed to be fair. It is not ethical, nor right to let the wealthy people and ONLY the wealthy people get a certain kind of special order type that grants both liquidity and also guaranteed price.

In my example if the guy is not willing to offer 500,000 shares total at that price, then he doesn't have to. It was perfectly legal for him to offer 20,000 shares at 35.2 on 5 different markets.

I am not taking advantage of the HFT - he is taking advantage of me. He SOLD his liquidity to get the slightly better price, he is NOT allowed to sneakily steal that liquidity back after he sold it.

I bought liquidity that he offered for sale. that is MY profit, not his to steal back.

Re:Won't work

[AK Marc]

Why not just batch trades? Come up with a "fair" way to pick whose trades get executed, and only trade every [timeperiod] (about as long as we'd expect 10,000 shares traded, or some other value). So high-traded stocks could go every 0.01s, but lower trades stock may hold for 10s between trades. The trades are executed simultaneously, so any "benefit" within that trade period will not help at all, and it would eliminate flodding fake offers with the intent of manipulating the market.

Re:Won't work

[Quirkz]

Personally, I think that it should be law that if you buy shares in any company (or fund or whatever), you have to hold on to them for a minimum of a week or a month. Shares represent actual physical companies which own factories and employ real people.

I get scolded if I make changes to my retirement account which result in moving in and out of a stock any time within a three-month window. I rebalanced once and then adjusted some holdings a few weeks later and got a letter about it. Of course I'm sure the fund manager isn't restricted like that, they just didn't want me to do it.

Re:Won't work

[Gauchito]

Something is fairly priced if it's at the highest price a buyer is willing to buy at and the lowest price a seller is willing to sell at. In your example, how high are you willing to buy? From his point of view, he's willing to sell at 35.2 given current market conditions. You buying that much changes those conditions. I think it's fair for him to pull. When you see the market, you see a system before you affect it. Once you affect, I think it's fair for the system to change. You buying should raise the price, and it does. This is no different than what usually happens in the market, only at a more granular and faster scale.

You have no inherent right to that liquidity, just like he has no inherent right to your flow. If you sent a market order to take out 35.2, don't be surprised if you pay more for the liquidity. If you're sending a market order, you don't care about the price, is what you're telling the market, so in an efficient market you should get filled at the lowest price someone is willing to sell to you, and you should get the worst price the market has to offer. I think that's fair. If you care about the price and send a limit order, then yes, someone has to be willing to sell to you, with the caveat that now there's information in the market about a new buyer, so people should be willing to sell to you only with that information in mind.

What you want is the benefit of a dark pool in a public market, you want to be able to buy across exchanges atomically without any information being available to any seller before all your buying is done. That's not what a public market should do, in my opinion. It should show any new information available as quickly and as broadly as possible, and that's what HFTs are essentially doing. They are making money by removing the information inefficiency between markets.

Re:Won't work

[sjames]

None. You bought it knowing the minimum hold time. Besides that, who do you think is going to be the buyer when you try to sell the burning factory?

Re:Won't work

[AK Marc]

One possible solution would be to trade shares on a fixed schedule, ex: every minute but not between minutes. This would provide everyone the same amount of time to react to the information. The difficult part would be deciding between who wins the bids, since time is an unfair determining factor that leaves some other metric like price bid or bidder's history or a random element.

Highest buys and lowest sells first, then the earliest submission is executed, then random numbers would make a good tie breaker after that.

Re:Won't work

[ultranova]

When you see the market, you see a system before you affect it. Once you affect, I think it's fair for the system to change. You buying should raise the price, and it does.

This sound suspiciously like insider trading: the guy with a supercomputer at the exchange's basement gets price-affecting information before anyone else does (because speed of light) and can react to it before other supercomputers, much less mere mortals, have had a chance to.

Re:Won't work

[Gauchito]

It's not insider trading, because the information is public. He's not the only one getting it, dozens are getting that information at the same time. HFT is a very competitive space with hundreds of firms doing it and competing against each other. They all have access to that information at the same time.

Re:Won't work

[Darinbob]

What they need is a RANDOMIZED delay of 250-500ms. That stops anyone from poaching trades (ie, noticing a high volume trade and then getting your own order in first). Ie, all incoming trades are given a randomization and place in a queue. Even a longer delay would be better (slow down the high frequency traders and not just high speed traders, which aren't the same thing).

The problem is that exchanges won't go for it. They make a lot of money by being the middle-men, and these high speed traders are increasing the volume which means the exchanges make more money.

Re:Won't work

[Darinbob]

This "liquidity" is a vague term used by high speed traders to justify their poaching of legitimate trades. What they really mean is that they increase VOLUME of trades. What they are doing is intercepting a trade and getting in a buy or sell first. So they double the number of transactions. However the individual investor holding onto a stock that wants to be able to sell it does not gain any extra ability to sell (liquidity) because of these extra traders, and is very likely to be gaining less money because of them.

This is not a service that improves the system, it is more like a parasite that feeds off of the system.

Re:Won't work

[Anonymous Coward]

All these posts like this forget the purpose of the specialist on the stock exchange floor. Brokers and institutions come to him with an order and he places the trades at specific price levels that he knows are going to happen (because HE is in charge of the price of that security). HFT is just one small facet of the overall system.

There is a reason most volume is at the open and close of the market. There is a reason the biggest spikes of intraday volume occur at the intraday highs and lows. And many transactions between specialists and favorable institutions don't even get printed on the intraday volume numbers.

Most of you posting to this story are slightly out of touch.

And don't forget about the Forex market. Most of you don't seem to care about HFT there (where it is much more prominent).

Re:Won't work

[Anonymous Coward]

Check your charts with your protractor and drafting triangles. The specialists are still quite in control of things. The computers just help them handle the volume these days.

Re:Won't work

[AK Marc]

My suggestion should even out some of these little spikes. The reason there is corruption is because corruption is allowed and rewarded.

Re:Won't work

[Citizen of Earth]

The way I see it, you can eliminate the advantages of HFT while keeping the markets highly responsive by imposing a "clocking" scheme on exchanges. When an order is received by an exchange, it is not executed immediately but stored in a queue to wait for the next clock tick. When that comes, the order queue is shuffled into random order and then executed sequentially. Make the clock ticks wait a random period between 40ms and 50ms and any timing advantage of HFT or geography is nullified. The exchanges are still highly responsive; they just do randomized batch processing. All of the requests they receive in the previous clock period ought to be processed within the new clock period (with perhaps some occasional spill-over, in which case the new clock tick is stretched).

Re:Won't work

[ultranova]

It's not insider trading, because the information is public. He's not the only one getting it, dozens are getting that information at the same time.

No, dozens aren't getting the information at the same time. They get the information when it enters their past light cone, which for some is later than for others.

HFT is a very competitive space with hundreds of firms doing it and competing against each other. They all have access to that information at the same time.

Even if every HFT firm gets the information at once, that doesn't mean that the market (everyone) does and has had time to react before they send their orders. Which is what needs to be shown to prove it's not insider trading.

Re:Won't work

[Gauchito]

The information being acted on is published by the exchange to everyone, notifying everyone of a trade executed in its market. For it to be insider trading, it would have to be non-public information, which a trade published by the exchange definitely isn't.

Re:Won't work

[rahvin112]

HFT, through the very nature of what it is and that there are dozens of competing HFT firms, has been driving buy/sell spreads down to micro percent spreads. You can execute a market order these days and actually get the price on the ticker. 10 years ago that wasn't the case and that is because of HFT. The firms buy and sell so rapidly that the spread has decreased from the 3-5% normal to fractional percents that often don't even add a penny to the price.

All the opposition to HFT I see is people who've never traded stocks in their life. How many trades do you make per year?

Re:Won't work

[rahvin112]

Without HFT we go back to the bad old days of Market makers. It's a world I'd prefer not to see again. This is a fact, HFT is providing the volume and market, without it there isn't the market available in anything but blue chips to drive buy and sell during the day. You might not understand this but if you buy or sell any volume that isn't a whole lot (100 shares) you went through a market maker. HFT has eliminated that, even with small partial lots the HFT moves in and buys and sells on the spread between markets. It's creating markets in almost every stock (even pink sheets) that didn't exist 10 year ago.

Only idiots who've never traded in their lives and think they are experts on this think HFT is always bad. Those of us that have been investing for long enough know that what existed before HFT was significantly worse.

Re:Won't work

[rahvin112]

Because the market operates on lots. A Lot is 100 shares. Say you want to buy 28 shares of a stock, you put in your order and it shows up on the trading boards. Joe blow investors wants to sell 8 shares, Jane wants sell 15 and Loki wants sell 10. The market maker (just a broker that trades in the stock), buys all three for sale lots then sells you the 28 shares, he plays a spread on each and takes roughly 3-5% of the transaction buy executing the trades at higher or lower prices to net his spread. The remainder shares he holds until the next transaction.

If you trade is clean even lots and do so nice even numbers you can probably net a transaction with another investor that eliminates the middle man. If you don't you get fucked by the market maker. Who doesn't trade in even lots? Small investors.

HFT has eliminated the market maker. The systems are buying and selling across dozens of different exchanges and driving the spread down to fractional percents. 10 year ago if you made a market order for stocks it would execute 3% less than the market was when you put the order in. If you put a limit on it the stock would execute at the price you set but miraculously there would still be 1-3% premium lost. I've had this happen to me a dozen times. It doesn't happen since HFT has taken over the market. Billions of shares are being traded daily now as the HFTs each try to beat each other for that 0.00001% margin between stock exchanges and markets.

Re:Won't work

[Anonymous Coward]

HFT is skimming, just legal...for now.

Re:Won't work

[epiccollision]

it is front running, whether you admit it or not. There is no trader acting upon information, its just a computer and they get there ahead of the trade by virtue of physics or the NYSE/NASDAQ sells you premium access...aka front-running.

Re:Won't work

[Anonymous Coward]

Suppose you have two exchanges, let's call them VAMPIRES and ANGELS (to pick names at random, I mean, if I've accidentally used names reminiscent of a real exchange I apologize as that's not my intention...)

...

it turns out the founders of VAMPIRES have a hidden agenda. They've made sure they have a fast connection to VAMPIRES, and arranged with the phone company to have an equally fast connection to ANGELS. After "resigning" from VAMPIRES to give the appearance of being uninvolved, they

Vampires... Bats...

Cummings said in his resignation "As BATS prepares to become an exchange, my ownership of a broker-dealer precludes me from serving in management." Cummings returned to his position as CEO at Tradebot

Hmm.

Re:Won't work

[Anonymous Coward]

It's not really. The market is open. The machines all have equal opportunity to buy something when it hits the market cheap.

Yes, colo access costs a lot etc. but once all the big players are in the same datacenter as the exchange they have no advantage other than the software they use.

Shameless Plug: http://www.objecttrading.com/

Re:Won't work

[Anonymous Coward]

It is NOT fair for you to take 100,000 of my order at 35.2, then instantly cancel your four other 100,000 orders and replace them at 35.4

It's just business. The biggest companies in the world do exactly that, thousands of times a day. Sometimes it works, sometimes someone beats them to it, sometimes it turns out to be a bad position to have taken.

Re:Won't work

[Anonymous Coward]

That's simply not true - putting a 500ms delay in the processing of orders will cause a major disruption to HFT algorithms.

HFT works by posting a bunch of small orders (e.g. buy 100 shares at 9.00, sell 100 shares at 9.01). When someone comes in with an order to e.g. "buy 5000 shares at market", the HFT side will sell the 100 shares at 9.01, then immediately enter a new order to sell at e.g. 9.02. Simultaneously, it will go out and try to buy 4900 shares at 9.02 on another exchange, secure in the knowledge that it already has a buyer for those shares at 9.02 lined up. Because the HFT machine is closer (cable-length wise) to the second exchange than the 5000-buyer's broker, the HFT order will get to the 2nd exchange ahead of the 5000-buyer's broker's order, and hence will get the better price on that 2nd exchange. The 5000-share-buyer won't be able to see what's happened behind the scenes, but the HFT algo has made a tiny, almost risk-free profit. All the 5000-buyer would see is that his average fill price would be e.g. 9.01585; he'd wonder why he only got 100 shares at 9.01, but with no visibility as to what the HFT algo has done behind the scenes, it'll remain a mystery.

Adding in a 500ms pause will completely disrupt the HFT algorithm. It won't see the "buy 5000 at market" order come in until 500ms later, which is well beyond the time it would take for the 5000-buyer's broker's order to get the remaining 4900 shares at the 2nd exchange. That means the HFT algo won't be able to outrace the order, and the 5000-buyer's fill price will be that bit closer to 9.01.

This is exactly the logic used in building the IEX (Investor's Exchange), which is often regarded as one of very few "HFT-free" exchanges. IEX implemented the "pause" by forcing all incoming and outgoing traffic to go through a giant loop of fibre before it gets to the exchange - something like 40km of fibre in a box that does nothing except slow down the rate at which orders can get to the exchange, and that fill information gets out of the exchange back to the HFT servers. That giant loop of fibre makes things just slow enough to f*ck up the HFTs...

All of this isn't to say that HFT algos won't come up with a new way of gaming the system, but this particular approach won't work any more

Re:Won't work

[Gr8Apes]

That only nullifies the ordering issue (first in, maximum profit) You still get in on the next tick at 'x' price, which should be close to what you were looking at. Your solution only stops the ever faster links and programs from improving results by merely being first in. Mine also addresses the issue of attempting to ride the small changes and skimming profit. It would reduce market volume, and likely volatility, as profit would have to go looking elsewhere other than the small fluctuations. Maybe even enhancing that aspect of the stock market that is its raison d'etre: actually investing in an activity.

Re:Won't work

[KingOfBLASH]

Citation please?

While you are correct that trades are typically allocated on a first come first served basis (creating an incentive to those who are speedy) most of what you wrote is bollucks

Re:Won't work

[400_guru]

If you simply change everyone's temporal frame of reference by the exact same amount, you have done nothing, really. Everyone will simply account for the 500ms delay, and trades will still execute in the same order.

Exactly. So look up IEX. Their solution is to 'position' themselves electronically to be *EXACTLY the same number of ms away from all the major exchanges in the U.S. So when they execute trades across 5 exchanges they all happen at EXACTLY the same time on all the exchanges. Thus those HFT guys react too late, all the needed buys for the specific transaction have already occurred. HFT Guys loose. Bummer.

It's not all about speed it's about timing.

Re:Won't work

[tolkienfan]

A stock's price is affected by many things. A few simple examples: Commodities - many (most?) companies either use them or are affected in some way. Similar or related stocks - suppose something happens to AMD - the price of NVIDIA will be affected. Likewise if the cost of silicon changes. Derivatives - options market makers have to constantly change their positions in underlying stocks. So trading in options affects stock prices. Currencies - many companies have assets priced in other currencies. They may have competition in other countries. They may have suppliers in other countries. Weather - some stocks are affected by the weather either directly or indirectly (e.g. via commodity prices) Etc... So the value of stocks changes on much smaller tinescales than you imagine. Forcing people to hold a stock that may decline in value makes them riskier. Additional risk translates into a higher cost to trading, which results ultimately to lower volumes, less liquidity and a damaged economy. This kind of idea is a bad solution to a nonexistent problem.

Re:Won't work

[tolkienfan]

What problem are you trying to solve? An HFT company can make an extra fraction of a cent on a trade because they pay for bettet infrastructure? This was true when phones were first introduced. What do you think us worse about using modern computers and modern networks?

Re:Won't work

[TechnoJoe]

Mod parent up. Something a lot simpler is a Financial Transaction Tax. For example, one cent per share, for shares held less than 30 days. When calculating age, you need to start with the newest shares first, to prevent people from avoiding the tax just by having a large pool of shares.

In cases like "factory X has just gone up in flames" it would be worth paying the tax. In other cases, probably not.

Re:Won't work

[Meski]

We do that now. CGT has a factor for how long you've held the share. ( > 12 months)

Re:Won't work

[tolkienfan]

HFT trades on exchanges, where that kind of preferential treatment cannot occur. More BS from the ignorant.

Re:Won't work

[tolkienfan]

This is absolutely correct.

Re:Won't work

[tolkienfan]

This just isn't true. HFT isn't speculation. And the current markets are highly efficient, meaning all information about an instument is very quickly included in it's price, largely because of HFT. If this weren't true arbitrage would work, and largely it doesn't anymore.

Re:Won't work

[tolkienfan]

A rare moment of upvoted clarity. Slashdot used to be a place of rationality. Now it's dominated by loud ignorance.

Re:Won't work

[tolkienfan]

That's not how it works. Liquidity is the amount of shares available to buy or sell at any moment in time. They are provided by market makers who keep "quotes" active at the best prices at the exchanges. The exchange matches orders to quotes, and nobody can get in between the way you describe. High liquidity means that someone wanting to buy or sell quickly can - any market order will immediately match at the exchange. Low liquidity is the opposite - you may have difficulty finding a buyer or seller for you to trade with.

Re:Won't work

[AK Marc]

It "can't" occur, but does. Reality trumps your desire.

Re:Won't work

[Darinbob]

But some firms doing high speed trading are defending their actions as "good" for the market because it adds "liquidity", which it really doesn't.

This is in the news last week because of publication of Flash Boys.
http://www.barnesandnoble.com/...

The author basically disputed that the HFT was adding liquidity.

Re:Won't work

[tolkienfan]

I'm sorry, the facts are on my side. Equity exchanges match in price time priority, regardless.

Re:Won't work

[tolkienfan]

I'm afraid he's wrong. Many studies have demonstrated that HFT market makers help stabilize markets during high volatility due to the liquidity they provide. I've seen first hand the volume of passive trades market makers provide. Its very real. Plus it's how they make money - so it wouldn't be very profitable if it weren't real, would it?

Re:Won't work

[AK Marc]

When you own the clock, that statement is meaningless.

Re:Won't work

[Anonymous Coward]

Incorrect, there is no knowledge that it is part of pending orders. Just pure guess work sent to another exchange to beat the 'possibility' that there might be.

Re:Won't work

[tolkienfan]

Nope. It isn't meaningless, it completely refutes your original statement. You seem to think that a speed advantage is everything, which is nonsense. If it were so, HFT would dominate in trading profits, but they don't. Not by a long way.

Re:Won't work

[Darinbob]

No, they make money by getting in trades first before someone else. Ie, what the book is talking about are traders who sit and wait until they see a big trade and then they get in the trade first because they're faster. This doesn't add ability to sell more easily. The traders in the middle make money though, with no risk as long as they're always first, but that's not a good thing for the market.

If there are studies saying this is good, then the studies were funded by HFT trading houses.

Re:Won't work

[tolkienfan]

I'm talking about independent studies. Even some where the researchers were expecting the reverse result. Regardless, what you're describing isn't correct. Orders are executed in time order. There isn't a way for am HFT company to "get in front" once they "see a big trade". I've been working in HFT for nearly 8 years. I actually know how this stuff really works. We don't manipulate markets, or jump the queue, or any of the usual accusations. We'd be in jail. We use strategies that are basically the same as those that existed before computers. It's simply that competition has driven down latency, and margins, but increased efficiency and liquidity.

Re:Won't work

[AK Marc]

Speed is an advantage when you can jump in line. See an order placed in LA, then get to NYC before the order does, and trade in a manner expecting the order to arrive, and defraud the person who placed the order in LA. I've been told that's illegal now, but it wasn't always, and it's how the HFT houses made their initiall billions to keep screwing everyone (and there's nothing in place to keep it from happening). Did you not read the story about the trade in Chicago that executed before the announcement it was based on was made?

Re:Won't work

[tolkienfan]

There isn't any way for an HFT company to see an order "on its way" to an exchange. HFT companies CAN see the trades and react, but that's public info, and anyone can react. That story you are referring to is nonsense, the times quoted were quite within reasonable transmission times. Even if they weren't it would mean somebody on the inside had leaked information to a trader. You can't blame HFT for that.

Re:Won't work

[Guru2Newbie]

There isn't any way for an HFT company to see an order "on its way" to an exchange.

Leaked info? No, try sold your information, and it's not insider trading. It's called Payment for order flow where brokers sell your intent to buy or sell to HFT companies. Before the trade takes place.

Quote from WSJ posting on 4/6/2014: "Shares of E*Trade Financial Corp. ETFC +0.05% , Charles Schwab Corp. SCHW +0.27% and TD Ameritrade Holding Corp. AMTD +0.26% tumbled last week amid concerns that regulators would ban a practice that allows brokerages to collect hundreds of millions of dollars a year in revenue by selling orders to middlemen who use high-frequency strategies to trade with the brokers' customers. The practice, called payment for order flow, has gained more attention since the release of "Flash Boys," a book by Michael Lewis that argues the markets are "rigged" to benefit high-frequency traders, allegations that are stirring up long-running questions about the fairness of markets."

Re:Won't work

[Actually, I do RTFA]

The issue isn't a HFT company making a fraction of a cent on a trade. It's a HFT making a fraction of a cent on every trade.

In your example, a better infrastructure allowed companies to make a profit proportional to the number (and, to some degree, skill) of their employees. In the current state, it depends solely on the infrastructure present.

Or, to put it a different way, the combination of a winner-take-all system with a solved-problem combined with a hardware-race yields an undesirable result; at least to me.

Re:Won't work

[Imsdal]

You completely miss the very obvious point that information from other sources may have a bearing on the value of a company. If I spot a trend in the subway I may draw the conclusion that some companies stand to gain more from that trend than other companies. Your static view of the world is way too simplified.

Re:Won't work

[tolkienfan]

I know all about payment for order flow (although we don't do it). Brokers have been allowed to match orders without sending them to an exchange forever. This is a simple extension of that. Why is it a problem? Regulations mean that the order must trade at the NBBO whether it goes to the exchange or not. The customer gets the best price. All trades and their times have to be submitted to regulators to ensure the rules are being followed. This is more haymaking. Unless, of course, you have evidence that some company is not following these rules... then I suggest you contact the authorities. But you don't, do you?

Re:Won't work

[tolkienfan]

So it's simply the efficiency of scale that you object to? Even knowing that before HFT, the market makers used to collude to keep the spreads at a quarter, and now (since HFTs compete) spreads are mostly a cent? And you still haven't mentioned any harm. Btw all orders would go thru a specialist even before HFT, so it's not "a trade" vs "all trades". Last thing: if you don't want to pay the market maker, simply join the best ask or best bid, and don't pay the spread. It's entirely your choice when you submit your order. Of course, your broker will still charge fees, but you can't blame that on HFT.

Re:Won't work

[Actually, I do RTFA]

So it's simply the efficiency of scale that you object to?

Well, it changes the entire situation. It's similar to how many people are okay with an officer tailing a suspect, but object to a dragnet search.

I do object to it on a small scale. But not enough to really care about.

Even knowing that before HFT, the market makers used to collude to keep the spreads at a quarter, and now (since HFTs compete) spreads are mostly a cent?

Mere correlation. HFT occurred at the same time (and as a result of) other efficiencies in the market. I have yet to find any evidence that HFT lowers costs to the retail investor.

Also, I find the spread to be a fairly meaningless comparison. HFT increases the number of trades. So, the spread isn't as unabigous a marker as it seems.

Although, I may be wrong. Please let me know if I seem off.

you still haven't mentioned any harm.

HFT's are making money, making billions. That money isn't from any value created. So it must be coming from somewhere else. That makes them a deadweight loss - siphoning cash from the market.

Re:Won't work

[Actually, I do RTFA]

but object to a dragnet search.

Sorry, stupidly explained. Object to a GPS device being attached to everyone's car. Same end result for one person, but totally different in aggregate.

Re:Won't work

[tolkienfan]

If you submit market orders, you are paying the bid ask spread. It's per share, so it's completely unambiguous. Put it this way: suppose the theoretical value of XYZ is $98.7633 then the best bid will be $98.76 and the best offer will be $98.77. If you submit a market order to buy a 100 lot, you'll be matched at $98.77, paying $0.0067 per share to the market maker, totalling $0.67 for the lot. But, you don't have to do that. You could instead submit a limit order to buy 100 at $98.76. And you'd likely get filled after a fairly short time. So: 1 people CHOOSE to pay the spread. So they must see a benefit. 2 your broker likely charges more in fees. The benefit is quality of fill... you'll get filled instantly at the nbbo in the first case. In the latter case you may not get filled at all... the price may change. These small fluctuations are next to meaningless to a long term investor, but they keep the markets efficient and risk down. There's much less risk to owning a stock when you know you can sell at within a penny of its value at a moment's notice. As for "other efficiencies" occurring at the same time, it's simply false. HFTs shrunk spreads by competing with the specialists, who used to keep the spreads at 25c in many cases. The specialist was unable to compete with 1c spreads -their profits dried up. This isn't theoretical, it's fact. HFTs are also supremely good at arbitrage, to a degree that prices are always very very close to ideal. HFTs are better for your trading than your broker. You seem genuine, unlike many here, so I hope you continue to discuss this.

Re:Won't work

[tolkienfan]

Suppose a screw manufacturer cut their price by halving their profit, but as a result sells 10 times as many screws. You wouldn't call that an ambiguous comparison, would you? It's exactly the same with shares. Market makers cut their per share profit, but increased their volume. As a result shares are much cheaper to trade, and more trading takes place. The big difference with shares is you're not buying in the first market (from the manufacturer), you're buying aftermarket, so the buyer needs a way to find the seller and they need to agree on a price. The most efficient means yet found for this is the market maker. Any more efficient way woukd naturally supplant the market maker buy undercutting their profit.

500 ms

[rossdee]

Thats half a second in laymans terms

If you need to sell some stock or commodity within a second of buying it, then something is wrong

Re:500 ms

[StripedCow]

True.

However, I still think it is wiser to slowly increase the delay from 0ms to 500ms over several months, because that would prevent any shock waves going through the markets.

Re:500 ms

[rvw]

Thats half a second in laymans terms

If you need to sell some stock or commodity within a second of buying it, then something is wrong

They need a 20 minute sell ban, plus a 1ct price tag for each buy.

Re:500 ms

[thaylin]

Why? if there is nothing wrong with HFT then why would there be shock waves?

Re:500 ms

[Anonymous Coward]

I've thought of three ways to address this:

1: A small cost per transaction. That way, the HFT traders can do their trades... but with even a tiny tax per transaction, it will slow this down.

2: A random delay on each transaction of 0-100 milliseconds on top of the 500ms delay. Cryptographically random and done in a way that is fair to all. The delay will not make the playing field perfectly even, but it will give the guys who don't have the billions in server farms a chance at some success.

3: In US prisons, prisoners can only move between cells from the start of the hour to the tenth minute (i.e. from 2:00 to 2:10.) A similar concept should be in trading, where trades can only be executed at certain time intervals, such as the first part of each minute or allowed for five seconds, disallowed for five, etc.

Even a random delay on transactions of 0-25 milliseconds will make things fair... of course, unless the RNG is checked out to be fair, that can be gamed to a player's advantage very easily.

Re:500 ms

[Immerman]

If there's nothing wrong with HFT then this would be a pointless change. I have to agree that a slow increase might be wise - we've already established that the majority of trading by volume is by HFT algorithms, which by their nature are incapable of exercising common sense, and have in fact caused massive swings in the market simply through their emergent reactions to random noise and each other.

Re:500 ms

[Anonymous Coward]

Thats half a second in laymans terms

If you need to sell some stock or commodity within a second of buying it, then something is wrong

So you reference the layman and totally miss the point. When the layman wants to sell, he can either sell in 100 ms via high-speed arbitrage. Or he can go to the end of the queue after all of Wall Street or London or Hong Kong has acted on the information.

The "cost" of you being able to buy or sell in a 100 ms timeframe (or less) is that the other end of the trade can flip that in X ms.

Keep in mind when you place an order you can limit the price in either direction.

The old system is to place a trade order and wait. And wait. And wait. It made it hard to put in a limit order because if you guess wrong, you're trade is not going through that day or that week. As is, I can look at the spread - gauge my urgency - and typically pick a number that executes a trade often before my screen can refresh.

Of course you could always place orders without limits in which case your trade could execute at 100 for a share that was only worth 10. Sucks to be you and sure as fuck you will wish there was HFT to make your purchase at 10.001 or 9.999.

What price/value do I place on that? Don't know. But this HFT are like the bitcoin miners except the game is latency and market info. It will be more and more people chasing smaller spreads. Likely, it is a business that has to colocate at or near the exchange. But so does the McDonalds/Burger King/Wendy's if they want to make money off the exchange as well.

Arguably the exchange could perform this service but they would have to switch from focusing on what they are good at, to what they may not be (and there could be bias that prevents that).

Sorry if its not fair to traders away from the exchange. The same is true if Sonics/In-and-Out Burder/DairyQueen wants to sell burgers to the exchange workers without the capital investment to locate just off site.

Either you compete or STFU!

How does this simply not move the goalposts?

[Junta]

If the whole point is to be x microseconds ahead of the other guys wouldn't a 500 ms delay simply mean the exact same game would become 'after 500 ms, still be a few microseconds ahead of the other guys'.

I would imagine a more effective approach would be to process trades 4 times per second. A request for a trade always gets processed in the slot after the next slot (meaning no less than a 250 ms delay, but no more than 500 ms delay). Within a given slot of trading activity, randomly shuffle the requests so that someone beating someone else by less than 250 ms doesn't actually affect things.

Re:How does this simply not move the goalposts?

[Anonymous Coward]

If the whole point is to be x microseconds ahead of the other guys wouldn't a 500 ms delay simply mean the exact same game would become 'after 500 ms, still be a few microseconds ahead of the other guys'.

I would imagine a more effective approach would be to process trades 4 times per second. A request for a trade always gets processed in the slot after the next slot (meaning no less than a 250 ms delay, but no more than 500 ms delay). Within a given slot of trading activity, randomly shuffle the requests so that someone beating someone else by less than 250 ms doesn't actually affect things.

HFT is making money off of network noise.

Re:How does this simply not move the goalposts?

[homb]

One of the major problem is when an HFT sees your making a trade in exchange A, it assumes you're going to be hitting the other exchanges for similar trades and beats you to them. I don't see how putting a delay in a trade at a single exchange would help.

Re:How does this simply not move the goalposts?

[Githaron]

They could also use cooldowns rather than a delays. HFTs will then put themselves at a disadvantage if they are constantly making trades because they will have to wait the whole cooldown before they can react. Meanwhile, everyone else will have no delay since their last trade was long before (at scale).

Re:How does this simply not move the goalposts?

[Junta]

Well my thought would be that multiple exchanges would implement the same scheme. In that case, someone coming in as late as 249 milliseconds after you has a 50/50 shot of being ahead of your trade anyway. Yes, one exchange wouldn't be enough, but the more exchanges that did the scheme, the less this would help.

Re:How does this simply not move the goalposts?

[StripedCow]

How about: incoming trades are delayed by a random amount of time (within reasonable limits).

Re:How does this simply not move the goalposts?

[DarkOx]

This solution does more or less solve that problem as well. Nobody but the exchange will have a view of the orders for at least a 250ms but possibly as long as a half second under the parents scheme. That will anyone selling or buying in enough volume to be a market mover on anything but the smallest of micro caps likely has the technology to submit orders to each exchange they mean to trade on inside that window where the HFT guys will not have the opportunity to do any new price discovery.

That should effectively thwart the front running ( if its really happening on the scales many claim ). It will help the retail level investor who owns funds, that may have been themselves victims of front running. It will also minimally impact the operation of the market and price discovery behaviors for not HFT trading. This may help prevent a flash crash type situation as well because the HFT machines will have to wait to see if an exchange is really dropping before the open up their sales and short sales on other exchanges. I think its a good solution in those terms; but I remain unconvinced that HFT is really the problem many say it is; the evidence just isn't there.

Re:How does this simply not move the goalposts?

[Junta]

In this case, the question is 'what's the downside?' If HFT isn't really a problem, then what harm would it be to level the playing field to 250 ms or whatever quantums? If HFT is a big deal, then this would fix it. If it is not, then it wouldn't change things much.

Certainly some financial institutions are heavily investing in HFT relevant schemes, so they at least believe that HFT impact can be significant.

Re:How does this simply not move the goalposts?

[StripedCow]

This could be more difficult to implement, because you'd need to keep track of all the trades.

Re:How does this simply not move the goalposts?

[GameboyRMH]

Agree, if it's a *delay* it does nothing. If it's an *interval* it would help. And I think a 5sec interval would be good. That would eliminate all technological competition, putting the servers in the exchange basements with hot-swappable, consumable CPUs and bleeding-edge network gear on par with a RasPi on the other side of the planet.

Re:How does this simply not move the goalposts?

[Pro923]

i agree with this 100%. Even make it a minute, for example... So trades only execute every minute on the minute - everything just buffers up until that second hand reaches zero again. I can't see any negative to this, and it would certainly seem to eliminate the shenanigans of using connection latencies to skim off of the inefficiencies of the market.

Re:How does this simply not move the goalposts?

[KingOfBLASH]

Won't work. How do you suppose trades actually go through and prices get discovered? Trading and price discovery sort of works like an auction. An auction is not effective if you randomly scramble the order the bids come in.

I say like an auction because there's not actually an auctioneer, it's more complicated than that. Some people do what's called passive execution. They put a quote out there that says "I'll buy XYZ at $x and sell at $y." Other people do what's called aggressive execution. They see your order to buy for $x and they "hit" it, and take the other side of the order.

Algorithms ultimately move the price they're willing to buy or sell at up or down, which is what creates price discovery. So if I'm willing to buy at a higher price than you, my quotes will get "hit" and you'll either need to match my offer price, or stay out.

That's where the speed advantage comes in. If I publish a quote to say I'll buy at $x + $0.05, and it takes my competitors 500 ms to react, and you can react in 100ms, you can buy shares from my competitors and sell to me before the market stabilizes.

That's the point of the delay. Right now there's a speed arms race where the fastest trader basically takes all the chips. If you can react to price discovery in 100ms, and it takes me 500ms, I effectively can't trade anymore.

Re:How does this simply not move the goalposts?

[Impy the Impiuos Imp]

Before outlawing it in a pique of jealousy at some perceived cosmic injustice, is this kind of trafing actually a problem?

If the real danger is a scenario where computers issues millions of trades in a few seconds in some feedback loop watching each other and making predictions, a 500ms slowdown could be just what the doctor ordered, slowing things down by 3-4 orders of magniude.

Re:How does this simply not move the goalposts?

[Tom]

If the whole point is to be x microseconds ahead of the other guys wouldn't a 500 ms delay simply mean the exact same game would become 'after 500 ms, still be a few microseconds ahead of the other guys'.

No, because HFT works by exploiting the tiniest of price differences and they are likely to vanish in those 500 ms.

I would imagine a more effective approach would be to process trades 4 times per second. A request for a trade always gets processed in the slot after the next slot (meaning no less than a 250 ms delay, but no more than 500 ms delay). Within a given slot of trading activity, randomly shuffle the requests so that someone beating someone else by less than 250 ms doesn't actually affect things.

That would work as well, but is more complicated and you could run into trouble when your slots reach capacity.

Re:How does this simply not move the goalposts?

[Junta]

No, because HFT works by exploiting the tiniest of price differences and they are likely to vanish in those 500 ms.

Perhaps then much of the problems the mass media tags with 'HFT' might hint at more general artifacts of low latency trading, or I could be full of it. A lot of the observed 'weirdness' would be mitigated through larger quantums of trading. The so-called 'flash crashes' where algorithmic trading of some sort goes nuts faster than people can correct for would be significantly slowed if trades could not react to each other as frequently.

That would work as well, but is more complicated and you could run into trouble when your slots reach capacity.

Considering the volume of data inherent, the 'capacity' of a slot can be pretty damn high as to be inexhaustible from a practical perspective. Sure, the code should have a contingency for the condition and that should be tested, but it is unlikely to be a frequently hit contingency.

Re:How does this simply not move the goalposts?

[Junta]

Won't work. How do you suppose trades actually go through and prices get discovered? Trading and price discovery sort of works like an auction. An auction is not effective if you randomly scramble the order the bids come in.

I'll confess to not being very well versed in this field and anything I say on the matter should be scrutinized.

That said, with a more coarse grained timeline, wouldn't the same processes hold, but just at a larger timescale? E.g. a 'bidding war' would span multiple trading quantums. If two bids go in within 250 ms of each other, then a randomization of the order shouldn't fundamentally change things. At an auction in person, for example, if two hands go up within 250 ms of each other, the auctioneer has no idea who went first. In effect, the auctioneer considers two such bidders in 'random' order, yet discovery in that case is not seen as unfairly random.

I guess I don't understand how the scheme would break things. I think I'm coming to understand now why a static delay can have an effect specifically against HFT, but was thinking that other algorithmic low-latency trading schemes are in play/likely for the future (i.e. I wasn't sure if non-HFT algorithmic trading also causes problems like flash crashes).

Re:How does this simply not move the goalposts?

[Githaron]

Not all the trades. Just the timestamp of the last trade for each user.

Re:How does this simply not move the goalposts?

[KingOfBLASH]

Sort of like an auction but not like an auction. There is no centralized keeper of bids, people make offers, and people take offers. HFT relies on speed because if I am faster than you I will take all the good offers and leave none of the bad offers.

A randomization of the bidding process would require change in the underlying mechanism as well

Re:How does this simply not move the goalposts?

[Tom]

Yes. Flash crashes are basically positive feedback loops at 10,000 rpm.

Considering the volume of data inherent, the 'capacity' of a slot can be pretty damn high as to be inexhaustible from a practical perspective. Sure, the code should have a contingency for the condition and that should be tested, but it is unlikely to be a frequently hit contingency.

One thing I can guarantee is that if exhausting the capacity would give a trader any kind of advantage in any way, then it will be done.

The computer systems these people have put up are incredibly. They can easily bring your trading system to a halt if they are determined to do that.

Re:How does this simply not move the goalposts?

[DarkOx]

Honestly I don't think it would do any harm to implement something like this. I don't think it will do much good personally, but I would not be opposed to it.

Re:How does this simply not move the goalposts?

[Anonymous Coward]

One of the major problems is when two algorithms malfunction and bounce between each other forcing the price of an instrument up or down.

Slowing down feedbacks like that, as well as the obvious levelling of the field (500ms vs 550ms is better than 200us vs 50ms), could be a good thing.

Re:How does this simply not move the goalposts?

[StripedCow]

Anyway, if trades are load balanced over various machines using various constraints (for instance, not just the user's ip address), those timestamps must also be distributed.

Re:How does this simply not move the goalposts?

[Junta]

There must already exist *some* scenarios in which two trades are practically 'simultaneous' and therefore the ordering within that quantum is ambiguous. Chaotic factors like network jitter already cause bids to jump in front of each other in a manner that does not necessarily reflect the precise order in which they were fired off. You already do not have a system where bids do not have sequence preserved. In fact, that's what the whole HFT business seems to take advantage of, that you can exert enough resources to jump in front of a offer you *know* is coming thanks to latencies, which clearly shows that ordering is not preserved.

So I guess my question is given that the current state of affairs where order is not preserved, but a door is open wide enough so that a big enough player can spend enough money to unfairly game the jitter, why would lifting the floor of that jitter to the level where all parties are equally impacted be a big game changer to the underlying mechanism (aside from the obvious of eliminating the ability for one party to jump in front of another reliably).

Re:How does this simply not move the goalposts?

[KingOfBLASH]

Two people hitting an offer at exactly the same trade is an EXTREMLY rare event because literally these guys are polling based on a difference of micro, nano, or picoseconds.

Only delay sell orders?

[crow]

I was first thinking that there should be a minimum holding period of five seconds, but I'm not sure that's technically feasible. How about only delaying sell orders? That would effectively create a minimum holding period.

Re:Only delay sell orders?

[Anonymous Coward]

If only sell ordered were delayed, one would only have to short sale a stock and immediately buy it back. This does not guaranty a minimum holding period.

Or a tax.

[Anonymous Coward]

A pause just creates an arms race. Taxation is a good antidote to accumulation of money without creation of wealth.

And don't give me any of the liquidity bullshit - investment, to be rational, must be a long term exercise. And there's no reason why market makers can't charge less without the HF bullshit - hell, public or private sectors could create a non-profit market maker.

Re:Or a tax.

[Whammy666]

I'd support a transaction tax so long as the funds were used for better enforcement and prosecution of Wall Street crooks.

Re:Or a tax.

[jythie]

Or an even simpler solution, extend the laws that make a human doing this illegal to cover automated systems too. HFT is only 'legal' because they can scream 'but it is being done by technology!', but as various court cases with file sharing have shown, assuming the judge feels like it, novel technological implementations are not a panacea against legal repercussions. Unless of course you have good lawyers, weak regulation, and an industry that is profiting from the transgression, in which case a judge might indeed magically find that the technicality is enough to get them off.

If you really want to shake things up

[StatureOfLiberty]

Just make a rule that you cannot own shares less than some arbitrary time. Say 10 minutes or maybe an hour.

Better article

[homb]

There's a gripping article over at the NY Times (adapted from a just released book) that explains very well the pitfalls of HFT, where the problems are mostly due to the haves and have-nots, just like in most things. The article is at http://www.nytimes.com/2014/04...

Not having a level playing deck in an exchange is a major problem for the correct functioning of said exchange.

Re:Better article

[Anonymous Coward]

haves? have-nots? did you per chance write the article about the "windows xp black market"?

Re:Better article

[jythie]

yeah, but the exchanges can really profit from inertia here. They only have to be fair enough for investors to not leave for exchanges with less usage on them. Kinda like PayPal, no one likes them, sellers and buyers know the company is corrupt, but it is where the buyers and sellers are so if you want access to other people using it you have to use it to. PayPal profits off this as long as they keep their corruption low enough to not cause a mass exodus. Same with the exchanges, investors know they are corrupt, but that is where the other investors are, so as long as they do not get TOO corrupt, investors will use them. People have tried building alternative exchanges that avoid the various conflicts of interest, but not enough people are using them for people to, well, use them.

Re:Better article

[RandomStr]

The story of how the IEX came to be is a compelling one, and considering the inertia behind it, it would seem that most exchanges would be well advised to consider implementing an analogous system; even if it only ends up being a short term solution...

Article is not very clear.

[140Mandak262Jamuna]

The Australian Financial Review today reported that ASIC had told the inquiry into Australia's financial system chaired by David Murray the regulator would consider introducing a half-second clamp on trades to remove HFT's speed advantage.

HFT work by seeing the order in one exchange at one price, and the same thing is available in another exchange for a slightly different price, simultaneously buy in one, sell in another and pocket the difference. Plain arbitrage, something Commander Vanderbuilt apparently did back in the days when news traveled on horseback during the day time. And he traveled at night in his sailing ship and raced ahead of the news, dumping bonds from bankrupt New York corporations.

These exchanges communicate the prices between themselves and take slightly longer than 350 milliseconds for the news to travel between exchanges. These big trading companies have faster access to both exchanges and are able to act on them. Would it be enough to delay all orders by 0.5 sec? Even if one trading firm sees the price difference, before it could act on it, the news would have traveled and it could no play micro second arbitrage.

This is my understanding. It might mean any trader must hold the instruments for 0.5 sec before trading it again. Not really sure what the article means by clamp.

Re:Article is not very clear.

[rvw]

Make it 20 minutes so people can understand what's going on. A sell ban for 20 minutes. And preferably a 1ct tax for each buy.

Re:Article is not very clear.

[locofungus]

Not really. That used to be possible but there are now so many people doing it that there aren't sufficient arbitrage opportunities to make enough money to cover your costs.

Instead it's more like this:

Consider a hypothetical stock that is worth exactly $1.

What used to happen was that the market makers would have a bid/ask of 0.95/1.05. (For a modern market that still trades with those sorts of spreads, look at gold metal. For small investors, a 5-10% bid/ask spread is fairly typical)

Someone came along and said, hey, I can make money even if I under cut the banks on both sides. So they started offering 0.96/1.04, another competitor, 0.97/1.03 until eventually you end up at 1.00/1.01.

Of course, in real life it's not quite that simple, the price actually moves. While the banks were offering 0.95/1.05, excepting exceptional circumstances, they don't need to update their prices all that often in order to avoid losing money. Of course, they do want to update their prices but it's a fairly leisurely process.

HFT traders can keep their spreads so low because they update their bid/ask prices constantly.

They make money, not by having a big margin, but by having tiny margins but capturing a lot of trade by having the best price. HFT has taken money from the banks and given it back to the investor.. The banks hate it and would love to see it stopped.

That said, there are things that (some) HFT firms are allegedly doing that aren't ideal. One of the things is to enter a new bid/ask into the order book and then cancel it again so quickly that nobody else can take advantage of it.

For example, current market 0.99/1.01. HFT puts in a 1.00 bid and then instantly cancels it again. People see that 1.00 order and try to move towards it - for example the person with the 1.01 ask might cancel it and enter a 1.00 ask instead to try and match the 1.00 only to find it's been cancelled in the 50us it took to react to it. The 1.00 ask now gets cancelled again but the HFT has now entered a 1.01 ask and is now front of the order book.

I have no idea how prevalent this is but if it is, it's easily preventable (orders have to be good in the market for a minimum (short) length of time before they can be cancelled) without having to lose all that is good in HFT.

Re:Article is not very clear.

[Splab]

Considering the book "Flash Boys" is about how fucked up HFT is and how bad it is for investors, one does wonder what kind of agenda you are pushing, claiming the exact opposite.

Re:Article is not very clear.

[squiggleslash]

What you're describing is not HFT, at least, not the definition used by people concerned about HFT.

Re:Article is not very clear.

[wiredlogic]

From High Frequency Trading Acceleration using FPGAs Leber, Geib, and Litz:

High frequency traders utilize a number of different strategies, including liquidity-providing strategies, statistical arbitrage strategies and liquidity detection strategies. In liquidity-providing strategies, high frequency traders try to earn the bid-ask spread which represents the difference of what buyers are willing to pay and sellers are willing to accept for trading stock. High volatility and large bid-ask spreads can be turned into profits for the high frequency trader while in return he provides liquidity to the market and lowers the bid-ask spread for other participants, adopting the role of a market maker. Liquidity and low ask-bid spreads are desirable as they reduce trading costs and improve the informational efficiency of asset price. Traders that employ arbitration strategies, on the other hand, try to correlate pricing information between related stocks or derivates and their underlying prices. Liquidity detection comprises strategies that seek to discover large orders by sending out small orders which can be leveraged by the traders. All strategies have in common that they require absolute lowest round-trip latencies as only the fastest HFT firm will be able to benefit from an existing opportunity.

Re:Article is not very clear.

[Anonymous Coward]

What you're describing is not HFT, at least, not the definition used by people concerned about HFT.

That's because, due in part to misleading descriptions by Michael Lewis, almost no one concerned about HFT really understands HFT, while the grandparent poster clearly does.

    --Someone in HFT

Re:Article is not very clear.

[squiggleslash]

The grandparent clearly understands $ATHING but we're not talking about $ATHING and the Australian initiative is not aimed at stopping $ATHING. So his point is irrelevent. Given nobody seems to be using the term HFT the way he and you are, it's safe to say you're simply wrong at this point.

Re:Article is not very clear.

[Anonymous Coward]

HFT doesn't just work on that, in fact a lot of it doesn't work like that at all.

They are just seeing bargains arrive on the market before anyone else.

I have a wise idea.

[Lieutenant Halfabeef]

Captchas on all trades.

Re:I have a wise idea.

[StripedCow]

Doesn't work: http://www.danielmiessler.com/...

Re:I have a wise idea.

[Immerman]

Ah, but we're not necessarily trying to lock out computer trading, just remove its massive advantage over human traders. I would bet that a mechanical turk-based captcha bypass would take considerably longer than a human trader typing in the captcha personally. Of course that might just mean the HFTers employ rooms full of captcha-typers, but you've still removed the ability to react on millisecond timescales.

Re:I have a wise idea.

[StripedCow]

In that case a short delay added to the order should do the job just fine. No need to mess around with captchas.

Re:I have a wise idea.

[Immerman]

But what fun would that be?

Also - a captcha based system would actually have two effects - adding semi-random delays into the system, and increasing the transaction costs. After all you need to pay that room full of captcha-typing monkeys, and bananas aren't cheap in infinite quantities.

Install random delay

[Whammy666]

A better system is to install a random delay of between 1 and 5 seconds. This would level the playing field completely and kill off the HFT parasites.

Re:Install random delay

[Anonymous Coward]

I think that would make the problem worse.

The article from a few days ago; showed how a firm was able to beat the HFT guys. They did this by having their trades all show up at the same time at the different exchanges. Your way would make it worse.

Front running is about seeing the 'random' in the network. That is why they want shorter and faster networks. 2-5ms is enough to front run.

Re:Install random delay

[dargaud]

Much better system: tax it in reverse to the time it is held. Something like 10 years 0%, 1 year 5%, 1 hour 50%, 1us 90%.

Re:Install random delay

[RogueWarrior65]

Until somebody gets a hold of the random number generator algorithm. You'd almost have to use some sort of biological process to generate the seed (no pun intended).

Re:Install random delay

[Anonymous Coward]

What's wrong with good old radiation-based RNGs?

Re:Install random delay

[Anonymous Coward]

Combine that with a required 6 second hold on stock before reselling... In the event that the HFT parasites figure out how to manipulate the random 1-5 second delay in their favor (which they will), they will still not be able to front run a multi-exchange trade in progress. I chose 6 seconds because it could take up to 5 sec random delay + 1 sec latency to complete multi-exchange trade.

Re:Install random delay

[Tom]

This.

Anything predictable will just be gamed. These people have made a multi-billion dollar business out of gaming the system. You'd have to be insane to assume they'll just pack up and leave on the first sign of resistance.

Introduce randomness into the system and put a tiny tax (one professor of the field recommended 0.01%) on trades even if they don't complete. Because HFT also does a lot of trade-ohwaitIdidntmeanit-cancel abuse. Sometimes thousands of orders per second that they never had any intention to actually completing. Their margins on each trade are tiny, so even the smallest transaction tax will drive them out of business. Meanwhile, honest traders won't care. If you buy stock for 5k and pay 50cents additional tax on that, you're highly unlikely to be the slightest bit bothered.

Re:Install random delay

[Anonymous Coward]

They all ready do by having to turn ever trade the right way up!

Re:Install random delay

[mederbil]

This is in fact a terrible idea if your goal is to prevent HFTs.

HFTs will just assume a 3.0s delay since they're making many trades very frequently.

But infrequent traders will have to assume a 5.0s delay because they don't trade enough to invoke the law of large numbers.

So in the end you give HFTs a 2.0s edge for no reason.

ASIC (disambiguation)

[tepples]

I wonder how hard this proposal by the Australian Securities and Investment Commission (ASIC) will hit makers of Application Specific Integrated Circuits (ASIC) designed to evaluate quotes and request trades.

End the Accounting tricks

[worker17]

Instead of just playing the numbers, why don't governments stop the manipulation entirely? You buy a stock, you hold it for 3 DAYS. The market adjusts for the sales and purchases instead of being artificially stimulated. The microsecond barons have to do some REAL work instead.

Re:End the Accounting tricks

[mwvdlee]

This.
All I keep seeing is proposed delays in seconds or minutes at best.
Trading shares is effectively gaining and selling ownership of a company.
There is no valid case for wanting to own part of a company for mere seconds.
There is no benefit for most companies either, so why do they allow an exchange to permit these risks to their business?
Are there no exchanges that enforce a "minimum ownership duration" rule for the companies they list?

Re:End the Accounting tricks

[DarkOx]

3 Days is a looong loooong time. I don't think minimum holding periods are fair to anyone, at least not above intervals of more than a few seconds. At three days imagine this situation.

On the 1st Joe buys 100 shares of $OIL_COMPANY at $10 a share. On the 3rd Jim buys 100 shares of $OIL_COMPANY for $11 a share. On the night of the third Joe and Jim are sitting at the bar watching the news, and discover $OIL_COMPANY just had a tanker run aground and its probably going to destroy a major fishery, the damages are almost certain to bankrupt $OIL_COMPANY.

Joe should be allowed sell his shares at the open on the fourth leaving Jim and everyone like him holding the bag? Joe has owned the company longer, possibly collected more dividends, paid less for the shares, he should be allowed to keep his profits at the expense of Jim and others like him? even though he and Jim both have the same quality information?

I don't see any justice or stabilizing effect to be gained with such long hold times, it will just discourage investing over all because anyone who makes a new purchase has to take the risk of sitting in a burning building while everyone else heads for the exists, until their hold period expires.

I can see doing trades in baskets of orders entered in say 60 second intervals and processed in random order inside the interval. That prevents a class system where guys with access to HFT have a shot at the getting in or out first all the time and moves it to where anyone with an Internet connection, a E-trade account, and a willingness to sit with three or four of the major news networks on a few tvs has a fair chance of transacting on whatever securities they are playing as quickly as anyone else. That seems okay.

Re:End the Accounting tricks

[operagost]

So when the CEO is indicted on fraud charges, or a factory burns down, you get to hold your stock while it tanks? This takes far less than a day to happen.

Re:End the Accounting tricks

[Immerman]

I agree, but:

> why do they allow an exchange to permit these risks to their business?

How exactly does frequent trading present risks to a business? The only impact stock value has on a company is with regard to its credit worthiness and resistance to hostile takeovers, and short-duration trades are unlikely to have a long-term impact on stock value. Perhaps there could be a trader just sitting around waiting to buy up controlling stock in Company X, but they'd have to be pretty sneaky to avoid tipping their hand - that's a pretty high volume there. And I really doubt bankers pay much attention to the HFT spikes when determining credit worthiness.

Re:End the Accounting tricks

[Anonymous Coward]

Instead of just playing the numbers, why don't governments stop the manipulation entirely? You buy a stock, you hold it for 3 DAYS. The market adjusts for the sales and purchases instead of being artificially stimulated. The microsecond barons have to do some REAL work instead.

Perfect is the enemy of good enough.
I happy if we see a delay introduced at all. Moving from a 500ms delay to a 3 day delay is can wait if it means that the 500ms delay actually is enforced.
If it gives a positive effect we can start thinking about the 3 day version.
I would hate to see this not getting implemented just because of a disagreement on the duration of the delay.

Re:End the Accounting tricks

[Anonymous Coward]

Yes! Shit happens!

Re:End the Accounting tricks

[u38cg]

I would like to sell you this contract to buy my share in three days. Price: that of one unit of stock.

Also, most of the solutions here forget that you can sell stock anywhere, including under the shade of a buttonwood tree. Make the exchanges illiquid and people will just fuck off elsewhere.

Re:End the Accounting tricks

[Anonymous Coward]

Because the sort of firms that can afford to buy data centers in the same building as the exchange to execute trades milliseconds faster, can also afford to buy a pile of shares every day and then do todays trading with the shares they bought 3 days ago. Thereby bypassing any impediment to their trading speed.

Re:End the Accounting tricks

[PRMan]

Would you be against a 1 minute delay? Or 30 seconds?

Re:End the Accounting tricks

[fostware]

"I can see doing trades in baskets of orders entered in say 60 second intervals and processed in random order inside the interval. That prevents a class system where guys with access to HFT have a shot at the getting in or out first all the time and moves it to where anyone with an Internet connection, a E-trade account, and a willingness to sit with three or four of the major news networks on a few tvs has a fair chance of transacting on whatever securities they are playing as quickly as anyone else. That seems okay."

Except anyone (human) won't submit 1000's of small trades to even the odds. You put a trade for 100 shares in as well as 99 other people buying ~100 - it's fair. My (previously used for HFT) computer submits 1000 lots of 25. Suddenly I might not get the full 100, I need to start cancelling when I get my quota, but I'm more likely to get my 100. If the computer plays it right, it could adjust the purchase quantity against projected movements. Add 1000 previously employed HF traders doing the same. Humans have a nigh-on impossible chance of getting a single trade, infact we'll end up in the same situation we're in now, where a large chunk of small trades are machines competing against machines.

Personally, I'm all for a 60sec hold before being able to sell. Give people a chance.

Re:End the Accounting tricks

[hunnybunny]

Yep. Tough. This acts as an incentive to *know* the company you're investing in. Not just to hit the button if X>Y.
Caveat Emptor.
Shares may go down as well as up.
All that jazz.

Re:End the Accounting tricks

[DarkOx]

This is a good point, but if they try to game the system stuffing a large batch of orders for very small numbers of shares to in order to as you describe stuff the box increasing their odds of acquiring or selling at least some shares over other peoples complete orders, any advantage they gain should be lost in increased commission fees; (hopefully)

Re:End the Accounting tricks

[mwvdlee]

HFT is a risk to the traded businesses because they deal with relatively large transactions and high speed.
Have a few HFT algorithms misread the market and compete (unless their code is 100% bug-free), and a stock can very quickly spiral out of control through feedback loops.

Re:End the Accounting tricks

[Immerman]

Yes, and that can be horrible (or wonderful) for the stockholders, but how does that affect the business itself? Stock price has no impact on their cash reserves, employee roster, institutional assets, etc. Unless they need credit, or wish to sell fresh stock, they are perfectly capable of functioning in exactly the same fashion whether their stock is worth $10,000 or $0.01. It's certainly possible that HFT could collapse (or wildly inflate) their stock price, but so long as their are rational actors in the market as well that stock should quickly recover to reflect something closer to the actual value of the company. Even if for some reason it takes a few weeks or months to do so, the only risk I see is that the company gets stuck needing to raise money while it's stock is artificially deflated.

Increased risk/volatility and tax

[Anonymous Coward]

So, every trade is now delayed by 500 milliseconds. If it's like that for everyone, across the board, how is that exactly deterring HFT? In fact, I'd argue that broker houses would take more risk in their programming of HFT to tackle this delay. I'm betting if this is implemented, we see a HFT runaway and market drop event within 6 months.

In the US, a micro-tax per trade would likely cut down on this absurdity. Not that my opinion matters, since I'm paying $6 a trade with the irrelevant sum of money I play with in the market.

Why not just a small transaction fee?

[Wycliffe]

This might work or just have the delay a random amount between 1 and 5 seconds but I
think the better solution would be to just increase the transaction cost as presumably this
is putting a fair amount of load on the system as well.
A simple transaction cost of maybe 1cent per share wouldn't affect a normal buyer at all,
would bring in money to the exchange but would put a huge damper on buying and selling
thousands of shares per second.
High Frequency Trading is kindof like email spam. The only do it because it is profitable.
A transaction cost should make it unprofitable unless they are scalping. If they are
scalping then the best solution is to maybe both increase the transaction cost and
add a random delay of 1-5 seconds. The increased transaction cost could also help
offset any loss that might come from the reduced volume of trading as presumably they
already do get a little something per transaction.

Re:Why not just a small transaction fee?

[Immerman]

Well, it would at least put a damper on HFT schemes where the profit is less than 1 cent. How common is that though? A HFT jumping in the middle of a 5-cent transaction discrepency still stands to make 4 cents.

Re:Why not just a small transaction fee?

[L4t3r4lu5]

1 - 5 cents is the bread and butter of HFT. The fact that it's only
The trouble is that it's actually billions of dollars annually because it's on every transaction.

Re:Why not just a small transaction fee?

[L4t3r4lu5]

Ah fuck me not closing my tags properly.

1 - 5 cents is the bread and butter of HFT. The fact that it's only < 10 cents that nobody thinks it's that big a deal.

The trouble is that it's actually billions of dollars annually because it's on every transaction.

Re:Why not just a small transaction fee?

[Wycliffe]

1 - 5 cents is the bread and butter of HFT. The fact that it's only < 10 cents that nobody thinks it's that big a deal.
 

So it sounds like the transaction fee might need to be closer to 10cents per share.
HFT exists because the current (transaction fee + transaction cost) is too low. Just like with spam, trying to
artificially increasing the transaction cost for an electronic transaction is bound to fail but unlike spam
increasing the transaction fee is easy to implement.
I don't think you want to eliminate all arbitrage as some can be good for the market. You just want to
eliminate the fractional cent arbitrage that doesn't benefit the market but taxes it (both physically and financially)
and a small fee would be a perfect solution to this.

Re:Why not just a small transaction fee?

[Archangel Michael]

The problem isn't that it cost billions, the problem is that "average" people cannot accrue the benefit of that profit, directly. And this seems unfair to the "fairness police".

The issue is, that this liquidity has already saved billions of dollars because it is liquid. Nobody is complaining about the increased efficiency of the market (liquidity is efficiency). If you want to play this game, buy the equipment to play it. There are more fundamental issues with Stock Trading IMHO, that HFT is not one of them.

Re:Why not just a small transaction fee?

[L4t3r4lu5]

There's no increased liquidity; Buyer and seller(s) were already engaged in a transaction (had reached Exchange A, not Exchange B), and all orders would have been fulfilled by the Exchange B once the order arrived there (Gross simplification, I know). Someone with a faster link between A and B buying all of $Stock at B and selling for slightly higher before the order from A arrives isn't adding liquidity. If a person did it, it would be called front running and it's illegal.

Re:Why not just a small transaction fee?

[Anonymous Coward]

Of course it costs billions. Money comes out of the system, but nothing is produced. Now either they are creating money out of nothing, or they are making it unavailable to people who are actually doing something productive (i.e. long term investments).

Re:Why not just a small transaction fee?

[Guru2Newbie]

A simple transaction cost of maybe 1cent per share wouldn't affect a normal buyer at all,

Oh, yes it would. Say I'm a small trader with a small account of $10,000. No margin. I'm trading stocks under $2.00, and I buy 1000-2000 shares at a time (that's $2000-$4000 per position). I get charged $8 to buy or sell, so that's $16 in fees for one round-trip (buy and sell) trade.

Adding a $0.01 fee PER SHARE (not per transaction) adds in a $10 (1000x0.01) to $20 (2000x0.01) fee, so that's $20-$40 in extra fees per trade, plus the $16 per trade = $36-$56 just for ONE round-trip trade.

With fees like that, I might as well telephone in each trade to a broker and hand them the fees, and give up on individual trading altogether. I'd have to make that up in higher and higher profit for every trade that I make. Sometimes I'm only making $50 gross on a trade, and $34 net profit after the broker fees. That's not a lot.

Re:Why not just a small transaction fee?

[Wycliffe]

You're already paying this fee. The HFT are already scalping you probably more than this per trade.
But you're also apparently a day-trader which is basically a similiar type position as a HFT though
not quite as fast. Stocks are shares of a company's current assets and future profits. It makes no
sense to have the amount of churn we do as companies assets and profits don't change that fast.
In alot of ways we would be better off if there was a higher transaction fee and/or a minimum holding
period of 1 hour/day/week. Day trading and HFT is calculated gambling and is a zero sum game.
It should be discouraged.

Banks deflecting attention from themselves

[FreeUser]

High frequency trading isn't the issue. The banks are the real "insiders", and are pointing fingers at small, high frequency prop shops to deflect attention from themselves, and to get back to the bad old days when they could really gouge their customers with wide spreads.

High frequency traders make their money by having better pricing models, narrowing spreads in the market, and being able to execute and then get out of a position quickly to lock in their profits and eliminate risk. The banks like to be the middleman, with wide spreads, so that they can pocket the difference.

The net result of high frequency traders is that the rest of us can get a stock much closer to their actual value (due to narrow spreads). Yes, the high freqency traders make good money by selling the stock $0.005 off the "real" value to me and then immediately getting out of the position by reselling it a millisecond later and locking in that $0.005 profit, but I have only paid a premium of $0.005 instad of the $0.35 or worse the banks would love to gouge me for (and used to, a few short years ago).

We get rid of high freqency trading and we'll be back to the bad old days, when the real insiders really did gouge us, and we all paid far too much for our investments, and were able to sell at far too little, with the likes of Goldman Sachs pocketing the enormous difference.

As for the front-running nonsense on 60 Minutes, that's always been illegal (contrary to what we're being told), and it is not at all how high frequency trading works. If someone was in fact doing that, then they're in a whole world of hurt with the SEC (and rightly so), but this entire exercise appears much more like a distraction: blame small outsider firms who've made the marketplace more effecient and tightened spreads for problems created by corruption within the big banks, and hope no one notices...at least until the next bank-induced crash.

Re:Banks deflecting attention from themselves

[homb]

As for the front-running nonsense on 60 Minutes, that's always been illegal (contrary to what we're being told), and it is not at all how high frequency trading works. If someone was in fact doing that, then they're in a whole world of hurt with the SEC (and rightly so), but this entire exercise appears much more like a distraction: blame small outsider firms who've made the marketplace more effecient and tightened spreads for problems created by corruption within the big banks, and hope no one notices...at least until the next bank-induced crash.

This is absolutely not illegal. Here's how HFT gets one of its profit lines:
Large trades often spread across multiple exchanges. Buy 30,000 shares here, 15,000 there, etc... The regular broker submits one purchase and it gets distributed across exchanges. As soon as it hits the first exchange, say after 30ms, an HFT algo picks up on the trade and assumes that it'll happen as well on the other exchanges. So it races ahead and front-runs in the other exchanges before the regular distributed trade has a chance to arrive there.
There is nothing illegal whatsoever, since the trades are public. It's just that the HFT optimized their routes.

Re:Banks deflecting attention from themselves

[OzPeter]

There is nothing illegal whatsoever, since the trades are public. It's just that the HFT optimized their routes.

Sure not illegal per se, but only a finite number of people can get that sort of access, so now the playing field isn't level.

Re:Banks deflecting attention from themselves

[homb]

There is nothing illegal whatsoever, since the trades are public. It's just that the HFT optimized their routes.

Sure not illegal per se, but only a finite number of people can get that sort of access, so now the playing field isn't level.

Exactly. That's one of the major complaints regarding HFT, and why the IEX exchange why created.

Re:Banks deflecting attention from themselves

[locofungus]

And why is this a problem?

I'm looking to sell say 50K on two exchanges - and I need to sell so I'm going to accept the best bid.

markets A and B are both offering 1.00 bid with a 1.02 ask. 20K on market A and 30K on market B are available.

I put in my 20K to market A. and I get my $20K. HFT trader with wow whiz algorithm spots my 20K order on A and races ahead to get a 30K 1.01 ask on B before I get there.

I get $30.3K on market B. Your greedy evil HFT has just made me keep more of my money and give less of it to those greedy bank market makers with their excessively wide spreads.

Re:Banks deflecting attention from themselves

[homb]

Wrong. HFT trader will bid 1.00 and sell, then as your trade comes in it won't be executed and you'll be forced to sell lower, say 0.98, which he'll gladly buy back from you. He got a completely unnecessary spread out of your pocket.

Re:Banks deflecting attention from themselves

[Immerman]

>being able to execute and then get out of a position quickly to lock in their profits and eliminate risk.

If they're not carrying any risk, then by what right to they lock in profits? The entire point of a stock exchange is to allow people to carry risk in exchange for profits. HFTers are simply gaming the system.

Re:Banks deflecting attention from themselves

[Anonymous Coward]

That's not front running. Front running would be: You (an investor) ask me (a broker) to buy 100 stocks, I buy 100 stocks and then "sell" them to you at a higher price, pocketing the difference. I already knew you were going to buy (through me) 100 stocks. This is: You (an investor) buy 100 stocks, I (another investor) assume that means you're going to buy another 100, so I buy 100 and hope to sell them to you at a higher price. At no point did I know for a fact that you were going to buy the stock back off me until you actually did.

Re:Banks deflecting attention from themselves

[jythie]

Well, technically HFT is illegal, it is just a matter of if a judge will agree hold them to it or not. Generally courts have found, well, at least for small fries, novel technical solutions do not make something legal when doing it by hand would not be. HFT is automated front running, nothing more.

Re:Banks deflecting attention from themselves

[jythie]

Not really, the spread is still there, the buyer just pays on the higher end of it.

Re:Banks deflecting attention from themselves

[j-beda]

Well, technically HFT is illegal, it is just a matter of if a judge will agree hold them to it or not. Generally courts have found, well, at least for small fries, novel technical solutions do not make something legal when doing it by hand would not be. HFT is automated front running, nothing more.

It is clearly illegal for your broker to purchase or sell securities on their own account before making the trades that you have ordered - that's the "classic" front running.

But I don't think it is illegal for a third party to use public information to make a trade. If you loudly tell someone on the bus that you are going to order a million shares of XYZ Corp as soon as you get to the office (maybe because E.F. Hutton told you to, a la classic 19870s commercial: https://www.youtube.com/watch?... ) anyone on that bus is free to whip out their smartphone and "front-run" you. I don't know that I would even call this type of thing unethical.

Figuring out technical solutions to this type of "abuse" by introducing delays or taxes is probably better than trying to prohibit them within a legal framework.

Re:Banks deflecting attention from themselves

[sjames]

The problem is the HFTs don't actually narrow the spreads, they just pocket them. They were never actually in the market to sell off or acquire stock, they're just the middle man.

They just crapflood the market with insincere offers to feel you out (or up) and then jump in line to make sure they grab the good deal you nearly got and then offer you a slightly less good deal. They do the same to the other side of the transaction.

If you don't get a good deal from the banks, just don't trade with them.

Re:Banks deflecting attention from themselves

[Anonymous Coward]

This is absolutely not illegal. Here's how HFT gets one of its profit lines:
...So it races ahead and front-runs in the other exchanges...

And front-running is illegal. So how is this not illegal?

Re:Banks deflecting attention from themselves

[Anonymous Coward]

I hope I'm not giving them any ideas, but since this is effectively "front running" could I exploit this behavior to "pump and dump" against the HFT?

Re:Banks deflecting attention from themselves

[jythie]

Though in this case, the first case is essentially what is happening. The exchange is giving your trade information to HFT traders in such a way that it can be utilized before it arrives at the other exchanges, even though all they are supposed to be doing is forwarding your trades to the other exchanges.

Re:Banks deflecting attention from themselves

[j-beda]

Though in this case, the first case is essentially what is happening. The exchange is giving your trade information to HFT traders in such a way that it can be utilized before it arrives at the other exchanges, even though all they are supposed to be doing is forwarding your trades to the other exchanges.

Well, the exchange IS supposed to publish information on what trades everyone solicits, that's pretty much a description of their function. Publishing that information in such a way so as to not create situations where a preferred group gets it ahead of everyone else does seem like a worthwhile goal.

I'd increase that to ten seconds at a minimum

[Karmashock]

There is no earthly reason for these commodities and stocks to trade hands faster then that. What are you doing?

The primary issue here is that human beings can't keep up with it. And that's extremely dangerous. If the computer gets confused then it can smash the market before anyone can do anything about it. But if its doing its thing in ten second pulses then you can likely stop it.

The secondary issue is that the market is very unfair with high frequency trading because it gives people with a better connection a huge advantage over everyone else. Its like having a time machine. Its the insider trading of knowing what the price is going to be in .2 seconds.

Pulse the system and most of that advantage goes away. Sure, your might get your order in faster if your system placed it faster but there's less information to react to... fewer iterations of the price to buy or sell against. You buy and sell on the pulse.

The problem after this will be the dark markets... the in house trading and between house trading of stocks, bonds, futures, etc. And putting any rules on the market tends to encourage the houses to use the dark markets more and more.

Which is fine. You control that by putting laws on the houses that they can't accept certain types of money if they're doing a lot of in house trading. The money you don't let them have is the pension money. The mortgage money. The big safe pots of money that the people give to the market makers largely to keep safe and grow at some reasonable rate.

The big houses need that money or they can't make the big buys. They can't leverage it to bend markets. And that means they have to choose... do they want to go big into the dark market or have access to the pension money? Because you make it a choice and they'll mostly choose the pensions. Which means the ones that will go after the dark market will be the smaller guys... the hustlers. And whatever they might or might not do, without the liquidity of the safe money... they won't really matter.

Re:I'd increase that to ten seconds at a minimum

[Anonymous Coward]

With a "pulse" it would have the opposite problem, people would be fighting over milliseconds to get their orders in fast right after each one because they would be processed first.

Re:I'd increase that to ten seconds at a minimum

[Karmashock]

True but the market would be slowed down a good deal. The ability of the computers to feedback on each other would be radically slowed down... ideally enough that a human being might actually follow the ebb and flow.

Which is ultimately my point here.

Might some market makers be able to enter orders faster or slower? Maybe. Though if you really wanted to take things to the next level you could introduce a system of uniform latency.

That is... you keep moment to moment stats on everyone in the network to figure out their latency. Or possibly you just have statistics on given IP ranges to get a rough idea of their geography. And then you delay order by those fractions of a second so that there is no geographic advantage when placing orders. All systems have the same ping be it next door or around the world.

The delay would be on the order of a half second which is well below the threshold of a human trader. Yes, a human twitch video game player would easily notice. But we're talking about trades with a 10 second pulse. The idea would be not to time the pulses but just trade normally and not worry about them.

Obviously the same people that created the high frequency trading scheme will try to exploit the pulse. That's fine. The instant they figure something out, counter it. And keep countering it. Make it expensive and unprofitable to try and cheat the system. The people just trading without trying to be sneaky about it will be unaffected.

random delay not enough...

[Junta]

Again, you have an 'average' 3 second baseline to compete against. What you really want to do is accumulate trades into a queue, have said queue stop taking new trades for some period of time, then process that queue in random order. Then there truly is no difference whatsoever between trades getting in within a quantum of the trade processing slice.

Why are trades (pre-purchase) public anyway?

[phillk6751]

Wouldn't it just make more sense that the trades be kept secret, using encryption until the transaction is completed? Transactions like this should prevent someone else stepping into front of line. Because this is basically what's happening, is that these machines are taking advantage of a security flaw that allows them to see a transaction before it's complete, It would be similar to going to a grocery store where you're ready to pay and someone behind you hands the cash to the cashier before you can get your money out and takes your stuff. That kind of behavior certainly wouldn't be tolerated at a grocery store (they'd definitely get punched in the face!), and shouldn't be tolerated on the stock market. The delay really wouldn't work....it would only take 5 seconds longer for you to find out you've been had, which means they're now stealing your time too. Any moderately useful encryption where public/private keys are used, like PGP, for any transaction prior to being completed would fix this.

Re:Why are trades (pre-purchase) public anyway?

[homb]

Because this is basically what's happening, is that these machines are taking advantage of a security flaw that allows them to see a transaction before it's complete

No. They see the completed transaction at one exchange for X shares, and assume you're doing the same thing at the other exchanges. They just race there faster and preempt your transactions that are on the way.

And they also consistently post fake offers that they retract in order to analyze the market appetite.

Re:Why are trades (pre-purchase) public anyway?

[Anonymous Coward]

If you cannot see the trades, how will you act to buy some stocks someone else put a sell order for?

Re:Why are trades (pre-purchase) public anyway?

[SydShamino]

The dark pools run by some of the big brokerage houses are doing what the GP describes though, and while that's not HFT it is another problem with the markets.

Just tax every trade

[Squidlips]

Capital Gains or some such tax. That will stop this crap....

TAX THEM!

[Anonymous Coward]

Add a 1% tax to all stock SELL orders where the seller has held the security less than a day.
Lower the tax to 1/2% for SELL orders where the seller has held the security for less than a week.
Lower to 1/4% for securities held less than a year.

This scheme would:
a) Raise a large amount of revenue
b) Constitute a 'use tax', kind of like a road toll.
c) Only affect people engaged in short term trading (e.g. wall street manipulators)
d) Act as a brake to prevent market volatility (e.g. the flash crash)
e) Be immediately shot down by Teapublicans asshats, so it won't happen.

Re:TAX THEM!

[Boronx]

Capital gains tax will ruin 'murka!

Re:TAX THEM!

[Khashishi]

I don't agree with basing the tax on time held. That's just an extra pain to audit. A simple low flat tax on each trade will already stop HFT. Even with a tiny tax, trading multiple times per microsecond will add up very quickly.

Re:TAX THEM!

[Khashishi]

You mean the same Teapublicans who push for elimination of income tax in favor for sales tax? Except that sales tax only applies to poor people. Rich people don't buy things; they invest in things.

Re:TAX THEM!

[esperto]

I'd go further, every time I see some HFT news I think that the best way to discorage would be a hyperbolic cotangent tax, and adjust the weights on the fly. So if you trade in miliseconds you pay like 30% of the total value of the shares, if you do in 1 second it would be like 10%, 1 minute 1%, and so on. if people start abusing again, adjust the weights further.

Re:TAX THEM!

[Anonymous Coward]

Problem is, you didn't consider the effect on 401k's and other retirement accounts. I can change my investments anyway I want as much as I want for my 401k. That could potentially cause a pretty hefty tax increase on the middle class simply because they are trying to save for retirement. It would also complicate their taxes considerably.

Now if you only tax the company that administers the 401k/IRA/etc. then eventually they're going to start capping trades and imposing delays on people who just want to shift their portfolio around.

Anyway, to sum up my feelings: first two points - good. Last one should be gotten rid of.

Re:TAX THEM!

[Anonymous Coward]

It would only tax middle class people engaged in day trading or very short term trading. The trigger points can certainly be optimized, in any case. There can even be a minimum threshold--if you profit less than $X in a year from ultra short term trading, you don't pay the tax.

As far as complicating taxes, it would all be spelled out in your year end brokerage statement, just like dividends and capital gains are now.

I should also add that the tax should be eliminated or reduced for trades that occur within 401K and other retirement account, just like you don't pay taxes now on those accounts until you cash out or retire.

Yikes

[Primate Pete]

A week?! A month?!!

How do you propose to compensate me and others for the loss of value and liquidity created by your arbitrary market rules and centrally controlled economy? Will you or the government either put up part of the purchase price to compensate for your partial control, or allow me to write off losses caused by the proposed rules?

What's wrong with me immediately changing my mind after a trade?

Re:Yikes

[lorinc]

Stop the bullshit. You're not changing your mind, you're trying to gain a lot very quickly by gambling.

If you don't understand the implications of what you're doing, please go to the casino instead of messing up our global economy.

Re:Yikes

[ysth]

The concept is that the market is supposed to be for investing. Investing implies certain loss of liquidity (no idea what you mean by loss of value). That said, see my response.

Re:Yikes

[Primate Pete]

No. Your concept of the market is incorrect.

The market organized to facilitate transactions, to buy and sell things easily, not to save money. Banks are for saving money--that's what they were invented for. However, whenever an investment is made, there's a risk, even in a bank. You might not get some or all of your money back. Banks can guarantee interest rates on small deposits because they pay you less than your risk is worth, that leaves them extra money to cover the times when their reinvestments of your money fail. (p.s. your money is not sitting in the vault, and they don't guarantee the safety of large deposits).

Loss of the ability to trade at will reduces the price that buyers are willing to pay for an asset. Nobody want to buy an asset that they won't be able to sell--it is not worth as much to buyers. So, slowing the market down reduces the values of the goods, services, and securities being traded.

High frequency traders are a serious problem, but not as serious a problem as a imposing a 1-week waiting period on market transactions.

Re:Yikes

[i kan reed]

The traders are the only people who gain tangible benefit from that, though. It's only their insistence that makes the spread so small, and the duration so large.

The rest of us are interested in laws that facilitate investment, you're interested in laws that let your manipulate people with less immediate knowledge of the market than you.

Re:Yikes

[fuzznutz]

A week?! A month?!! How do you propose to compensate me and others for the loss of value and liquidity created by your arbitrary market rules and centrally controlled economy? Will you or the government either put up part of the purchase price to compensate for your partial control, or allow me to write off losses caused by the proposed rules? What's wrong with me immediately changing my mind after a trade?

How about instead a Ultra-Short Capital Gains tax rate at 100%. Trade as fast as you want but unless you hold it for a week or month, you pay out all your profits. You're as liquid as you need to be. You don't have to ban it, just de-incentivise it.

Re:Yikes

[guises]

There are certainly problems with that plan, but the idea that you deserve some kind of compensation for basic changes to trading or how the market works is staggeringly egocentric.

Re:Yikes

[Primate Pete]

Yes, it is only the traders that gain benefit--they are the ones participating in the transactions. Fast traders (HFTs), slow traders (e.g., retirement investors), and traders with all time frames in between assume the risks and accumulate the profits and losses.

No matter how slowly the transactions move, the participants with more current information have an advantage. This is true whether you're talking about making a fast profit using stocks or investing (speculating?) in real estate because you think it will be valuable decades in the future. Information always matters, and you can't eliminate the advantage of information and still have a free (or free-ish) market.

Of course, if you're interested in making use of the money owned by market participants to achieve social/economic/political goals, you may have a legitimate point. But at that point, we're not talking about investing, we're talking about taxation. I may support that, but let's be clear about what we're doing. Until then, the people who own the money get to use it for their own profit.

Re:Yikes

[AK Marc]

How do you propose to compensate me and others for the loss of value and liquidity created by your arbitrary market rules and centrally controlled economy?

So you are ok with losses caused by arbitrary market rules and a centrally controlled economy, so long as you have liquidity? Because that's what you have now.

Will you or the government either put up part of the purchase price to compensate for your partial control, or allow me to write off losses caused by the proposed rules?

You have the options you have now. Accept the risk or don't play.

Re:Yikes

[AK Marc]

No matter how slowly the transactions move, the participants with more current information have an advantage.

The problem is that's false. A slow investor with more information will still "lose".

Re:Yikes

[sjames]

That's a nice sense of entitlement you've got there.

How much should we compensate muggers for outlawing mugging?

How much do you propose to pay in compensation for the damage caused by HFT?

Re:Yikes

[geekoid]

"The market organized to facilitate transactions, to buy and sell things easily"
and cheaper.

"High frequency traders are a serious problem, but not as serious a problem as a imposing a 1-week waiting period on market transactions.
nothing you said supports that statement.
Don't like the rules, don't play the game. The fact is, people churning stocks for a quick buck hurt everyone else.

Re:Yikes

[geekoid]

Do you understand the HFT ride the bubble they create by buying, or selling?
buy 100,000 share, price goes up slightly* then as sell happens immediately. less then a second is many cases. Rinse repeat * a million.

*as you would escpet, but then you expect an 'immediate' dip.

Re:Yikes

[geekoid]

yes, that would work. By work I mean it will solve the 'issue' that use as an excuse to prevent regulation. The real reason isn't changing ones mind, it's riding a bubble they create.

IT's like when people say, you can' ban guns I need then to hunt. And then you say, fine ban all guns that hold more then 3 rounds.
Then they stammer and come up with as different excuse.

People just making shit up so the don't have to challenge their own narrative.

Re:Yikes

[i kan reed]

Incorrect. This is not what HFT does. I'm not going to defend those jerks, but I do suggest making a fact-based argument, because it makes my, similar, argument, look better.

Re:Yikes

[Anonymous Coward]

Go look up arbitrage. Short time frames and a necessary part of the market. Please go check it out instead of messing up our global economy.

Re:Yikes

[tolkienfan]

In what way do you think HFT messed up the global economy? I bet you can't back that claim up.

Re:Yikes

[tolkienfan]

This is just plain false. Suppose you know that AMD is going to rise 5c, and place an order at an exchange for the current best bid. The HFT market maker has less info, and trades with you at yiur given price. He takes at most 1c and you take 4c. This is the best you can do because of the spread, and the market maker may well end up losing in the deal, since he sold at the bottom. What magic do you think the HFT company has?

Re:Yikes

[Meski]

And this is called 'rent-seeking'

Re:Yikes

[tolkienfan]

That's a complete fabrication. What you describe is called manipulation and is specifically illegal. The SEC and exchanges monitor for this kind of activity. Nobody would stay in business long doing that.

Re:Yikes

[tolkienfan]

What damage has HFT caused? Be specific.

Re:Yikes

[AK Marc]

They make their 1c many more times than I make my 4c, so they still end up ahead.

Re:Yikes

[tolkienfan]

You claimed that speed is better than information, and my example shows how ridiculous that is. Speed can allow you to lose money quickly too, you know? Your new argument is lost on me. It sounds simply like jealousy.

Re:Yikes

[AK Marc]

Jealous that I don't have a billion dollars to milk innocent investors out of millions more? HFT is criminally stealing money from "honest" investors. HFT make fraudulent offers that are designed to manipulate the market, then are canceled if they aren't profitable.

Re:Yikes

[tolkienfan]

HFTs don't do those things. That's libelous. Manipulation and fraud are illegal, and constantly monitored for by the exchanges, sros and the SEC. You're making a lot of accusations, but you have no evidence, do you?

Re:Yikes

[sjames]

I think it should be fairly clear if you look through the thread, and I don't have the days it might require to make a full list, but lets start with tipping the entire market in favor of the big players, de-stabilizing the market, creating irrational highs and flash crashes, completely de-coupling ROI in the market from anything resembling investment in economically valuable business, and leaching money from the economy.

HFTs crapflooding the market with insincere offers are no better than trolls in a forum doing much the same.

Re:Yikes

[tolkienfan]

The market was in favor of the big players long before HFT. In fact HFTs tend to be small shops. The two I've worked for started with 2 people. So basically the reverse of what you claim is true. Studies show that in the whole, HFT improves stability. The occasional mistake is less likely to have a big impact since circuit breakers were introduced. I doubt you can find evidence that HFT reduces stability. As for HFT creating irrational highs, it's utter crap. HFT makes money from small fluctuations. Large price swings are caused by big events. Investment still works as it ever did - ask your broker. As for leaching, how does any trading make money? We buy and sell and profit. The only difference is competition has driven down margins and latency. Are you from Salem?

Re:Yikes

[AK Marc]

If doesn't matter what I have for proof. You'd ignore it anyway. You believe in the religion of greed. HFT is a leach, causing damage to the economy, and rife with fraud. Yes, I have proof. More than one HFT has lost in court. They may be better at hiding it now, but they don't do it any less.

Re:Yikes

[tolkienfan]

Everything you've said so far is demonstrably false, and now you claim you have proof. Where is it? Which company? Put up or shut up. Even if some HFTs had entered into some bad behaviors (and you've presented no evidence), that doesn't show that HFT is itself necessarily bad. What is it about HFT that you think is wrong? You can't answer because you don't actually understand it. You don't know what we do or how we interact with marketplaces. For some reason your ignorance doesn't stop you from making wild unsupported claims.

Re:Yikes

[sjames]

Are you from Salem?

No, I just took my blinders off and passed on the cool aid.

Of course, I don't have to wear the blinders to be able to look myself in the mirror.

Re:Yikes

[AK Marc]

http://slashdot.org/comments.p...

Others have said it. There are books about it. There have been public investigations about it. That you choose not to see doesn't make it not happen.

Re:Yikes

[tolkienfan]

Others have said it, so it must be true? Lewis book is the only one I'm aware of, and I haven't read it yet. Public investigations? Show me. I responded to the comment you linked to, btw. More haymaking. You really don't understand any of this stuff, yet feel entitled to spout other people's unsubstantiated nonsense. You have yet to come up with a single solid piece of support for your accusations.

Re:Yikes

[AK Marc]

There's nothing I could say to change your mind, no evidence I could present. So why do you keep asking when it doesn't matter to you? Why do you think HFT is a good thing? The billions they extract from the system is worth it for the tiny increase in liquidity they assert they add (the only "identified" benefit I've ever seen mentioned). If they are extracting billions, and adding to value, then, by definition, they are thieves. I don't feel it's my problem that I'm unable to prove their level of thievery to your satisfaction. You don't seem interested in the truth, but instead in supporting thievery, regardless of the facts.

Re:Yikes

[tolkienfan]

That was a completely valueless comment. It did nothing to furthur the discussion. It didn't present any viewpoint or evidence. You merely call me a thief and state that I wouldn't accept any proof if you did present it. I'm asking for any shred of evidence for other's benefit... because I know you have none - as evidenced by your lack of ability to present it. I'm the expert in this discussion, so it should be obvious why I keep asking. Some people have had genuine questions for me about HFT, and want to learn. You clearly want to remain ignorant. So be it.

You realize a market is already...

[Anonymous Coward]

'centrally controlled', don't you? Be it by the market owners, or whatever your local equivalent of the SEC is, the market is already stacked for the paper pushers to win steady, and everybody else to claw over each other for their share of the scraps.

neo-liberal rubbish anyway

[Anonymous Coward]

HFT does nothing for real economics: no real jobs are created, no real infrastructure, ideas or concepts developed.
In essence its a non-transitive sink hole where real 'entropy'/work flows from the 'real' the economy represented by fiat currency.

And it does so using this real money: a waste of computing power and brain power that could be used elsewhere.
Eventually laws of physis will catch up with a system designed like this: if money is tied to real 'work' at some point and this system is left to its own devices laws of entropy say its going to crash.

Ge rid of the system altogether it already destabilises the economy.

Markets should be synchronous

[Strange Attractor]

Perhaps it would be better to accumulate orders without making them public for intervals of say a second or a minute. Then at the end of each interval execute buy and sell orders with overlapping prices.

HFT = a cost to society

[advid.net]

What really annoy me with HFT, besides not being "fair", it that it as a cost and that the society doesn't benefit from it.

Building a stock exchange with top-notch computers if fine, since there is a need fulfilled here for our society.

But building new warehouses as close as possible to stock exchange computers to house top speed fiber connected computers, just to lower the delays from 600ms down to 10ms or so, to allow HFT, is a waste of resources.
No one needs that, it's just a smart way to build a sucking vampire over information systems. And this cost is always somehow reflected to society.

One big bank of my country paid a lot to move all its crucial infrastructure abroad, in such new buildings, to be able to compete in HFT.
Who's paying for those efforts? The company, the bank, instead of doing something more useful to society (investments to improve their services, etc).

Re:HFT = a cost to society

[tobe]

It really is since they're potentially inflating the prices paid for equities.. not only by the individuals but their pensions funds. HFT shops open and close with 0 positions.. they do not hold stock past the close of business. They are simply skimming cents from transactions and that's costing people real money over the lifetimes of their investments. The stock exchange has lost all sight of it's original function to raise capital for growth investment.

Re:HFT = a cost to society

[zoefff]

To be the devils advocate:
The beneficiars are the manufacturers of said equipment and warehouses, which pay their employees and their suppliers for the work done, who in turn will pay their employees as well. And by the way these employees will buy houses, cars and food from other companies, which .... Need to go on? And every part will pay some tax as well.

The investments are probably paying themselves off. One can debate if the service delivered is a usefull one, but in the end it translates to 'real' economic value in terms of production of other goods. Because economy is about pumping money around. The faster, the better. And if the investment pays off really good, then there is money for investing in improved banking services as well. A win-win situation here.... Ok, I'm a bit sarcastic here.

So, my conclusion, yes, there is a cost, but yes, there is also a benefit to society in terms of economic activity (real). And I think it is better to focus on the unfairness part of it, which is on the part of the big, fast, but not so fast traders. But it is a rat race and I'm confident it will be sorted out. The end result is no one will make money this way and that again there will be a fee asked for trading. Problem solved.

Re:HFT = a cost to society

[advid.net]

You are not the devils advocate.

Your argument boils down to :
" Any pointless activity that consume goods is good for the economy. "

If you think this kind of behaviour drives us in the right direction for the material well-being, you must be quite an economist ! :)

Re:HFT = a cost to society

[Anonymous Coward]

High-Frequency Traders arent actually investors, they're just middle-men getting rich by siphoning money out of the system through automated front-running. This takes money out of the hands of real investors and companies, which is a detriment to everyone (except the ones doing it, but obviously they dont give a fuck about society or anyone else).

Outlaw the whole practice ...

[Anonymous Coward]

HFT is just wholesale theft by from the rest of us perpetrated by the big financial institutions which have direct access to the system.

They're not 'earning' money, they're skimming money.

bin trades and add randomness

[RichMan]

a) resolution in 5 second trade blocks
            - don't resolve trades with bid/ask when they arrive. Accumulate 5 seconds of incoming bid/ask then resolve. Pass the remaining unresolved to the next interval.
b) Arrival time +10 to 20 seconds random interval -> fuzzy bin boundaries and fixed "lag". Not guaranteed to get simultaneous orders into the same trade bin. Keeps trading honest and less able to be gamed.
c) fixed lag and binning means no insta-cancelling orders.
d) no cancelling once posted into the (b) queue. Yes you trade in the lag. So does everyone else. The traders can factor that into their margins and strategies.

Re:bin trades and add randomness

[tobe]

Yep.. this is basically what i'd propose but with a shorter clock, 100ms, and without the artificial jitter. Shouldn't be required with low enough clock frequency. I'd also probably allow cancels but only after multi period persistence.

It won't work.. second entity can short

[mrops]

Even the 500ms won't work. HFT have a lot of infrastructure and resources. You can simple model your HFT in a way that there are two entities trading, one buys it, if the HFT system decides to sell it and is restricted to do so for 500ms or a week, the other entity shorts the same trade, in effect achieving the same results. IMO, you should be allowed to short trades along with this 500ms block for this to properly work.

Obviously, you can have multiple entities trading on behalf of the HF Traders.

Re:It won't work.. second entity can short

[Anonymous Coward]

What you propose is illegal already, at least in the USA.

Re:It won't work.. second entity can short

[Meski]

And this article is about Australia, *not* the USA.

Would adding a CAPTCHA be feasible

[Anonymous Coward]

Should automated systems be trading at all?

Re:Would adding a CAPTCHA be feasible

[Khashishi]

Exchanges could easily make HFT unviable if they wanted to. But why would a private exchange ever want to do that?

HFT algorithms already manage this

[peter303]

According to last years ACM issue on HFT algorithms, trades already presented in time-tic goups. Just currently much shorter than you propose. The ACM articles discussed tricks to jump to the head of a group/queue.

HFT also manages the randomness in order presentation. At millisecond resolution TCp/IP collisions cancelations become a serious issue. The smae ACM issues discusses how to game this.

Evey constraint you popose algorthmically will probably be beaten by another clever computer executed algorithm.

TAX

[s122604]

I'm an active trader, on the order of about a dozen trades a month, not a "day trader" or a high frequency trader. I'd gladly pay a nickle a trade in tax, with the proceeds to go to better regulation. It would also have the effect of seriously hampering the high frequency traders business model, which is a great side effect IMHO.

Re:TAX

[geekoid]

.07% would be better.

Add Delay

[randallman]

They could switch the trade system to .NET. As London discovered, delay functionality is already built in.

serious problems with networking equipment in HFT

[lkcl]

http://www.nytimes.com/2014/04...

this article explains in depth what the problem is. the SEC has now been alerted to the problem, and is investigating. the people who found the issue believed originally that this was deliberate, but it actually just turned out to be a systemic problem of the speed differentials between different routes that high-frequency trades come in at.

what they originally discovered was that they could see a price on a screen, but the moment that they put in the bid to a number of brokers, the price would DISAPPEAR. they thought that this was deliberate, that someone was scamming them: it turned out that this wasn't true, but it took a couple of years of investigation to find out. what they did was they put in *individual* bids *directly*, and found that they were accepted. they then investigated various combinations, introducing delays into the bids, and found, amazingly, that it was down to the *time of arrival at the exchange* of their bids as they were sent via numerous brokers.

so it was only when they invented a tool (which they called "Troy") that *deliberately* introduced networking delays, such that the bids would (as best they could manage) arrive within milliseconds of each other at the exchange, that they managed to trade successfully.

if however any one of those bids happened to go via a different ISP, or a different router, or any other random combination, then the bids would *FAIL*.

the problem it turns out is that these delay effects are well-known. most of the money in high-frequency trading is therefore made by seeing a slightly slower broker's prices, then putting in an undercutting bid *knowing full well* that the other broker has a slower network. and this aspect of high-frequency trading is what is currently under investigation by the SEC.

*this is why the introduction of networking delays is so absolutely important*.

the people who discovered this phenomenon basically had to set up their own independent exchange in order to solve the problem. they needed to introduce a delay of 350ms as a way to make things fair for everyone. they did this by basically putting in 38 miles of fibre-optic cable in a shoe-box in the basement of the server farm that they leased.

it turns out that once investors discovered this, they began *specifically demanding* that their trades *exclusively* be brokered through this new exchange that had this 350ms shoe-box delay. it actually caused a lot of embarrassment for a number of brokers and trading houses because the brokers were explicitly disobeying their client's instructions, because the brokers didn't understand how important this really is.

anyway: you really have to read that article (or the book) fully because it's quite complex, and it's basically an inherent flaw down to the fact that the internet (TCP/IP) is routed randomly, thus introducing gross unfairness that has become the subject of intense investigation, very recently.

so yes, *all* trading should be done with at least a 350ms delay.

Gamed? Like how YOU "played yourself"?

[Anonymous Coward]

APK blew you away for your libel http://slashdot.org/comments.p... and in programming on multiple levels (which you tried to hide by minus mods earlier there no less - ashamed big talker? You ought to be) http://slashdot.org/comments.p...

Re:Gamed? Like how YOU "played yourself"?

[Anonymous Coward]

Funny Tom always shuts up and disappears after that post (not). He must be busy "eating his words" and at least Tom's somewhat polite (now that apk humbled the hell out of him for Tom's numerous mistakes) and doesn't talk with his mouth full (of his own words he had to eat)

Tom = know nothing libeler

[Anonymous Coward]

APK blew you away for your libel http://slashdot.org/comments.p... and in programming on multiple levels (which you tried to hide by minus mods earlier there no less - ashamed big talker? You ought to be) http://slashdot.org/comments.p...

Re:Tom = know nothing libeler

[Anonymous Coward]

Funny Tom always shuts up and disappears after that post (not). He must be busy "eating his words" and at least Tom's somewhat polite (now that apk humbled the hell out of him for Tom's numerous mistakes) and doesn't talk with his mouth full (of his own words he had to eat).

Re:Won't work: Nope the only way it will work.

[Anonymous Coward]

I read through most of the replies. Most of them don't consider the 500ms is both for placing an order and canceling an order, which is why it will work. Read further below. This post goes the extra mile and suggests we don't need any intraday trades. I will counter this post and clarify the others.

This post completely misses the point of HFT. "HFT" is a "good" thing in certain instances. It produces liquidity, within the range of prices. Now remove computers from HFT; it used to be real humans who did the HFT stuff instead of algorithms. So you could hit the exchange with a million dollar order and not see the price move from 100 to 150. Some other exchanges handle this with market makers or what have you in the pink slip world. So, every single market participant including the guy who holds stuff for years, needs the liquidity at least twice in his lifecycle or if you are a hold till your grave guy, at least once.

Before anyone accuses me of being in favor of HFT. Let me quickly point out why the 500ms delay will work. There is an exchange in the US which is privately run which uses a 350ms delay, by using a 30 mile box of fiber. Many of the answers below talk about arbitrage between exchanges. All of them or most of them missed why it will work even within the exchange!

If the current state of the market order book is:

10000 shares at $10
20000 shares at $10.1
30000 shares at $10.2

Remember the 350ms delay? Orders which are on the exchange stay for at least 350ms. The computers cannot have orders that stay for 20ms and then leave. HFT's can absorb or are designed to absorb a hit on the market book. So the first 5000 or so orders or whatever the market book is at a price may result in a loss due to some investor pumping in cash at a price point. It is usually the big fish which have bigger orders. When such big orders hit the market the HFT's are turned off. Why? The HFT's are just trying the squeeze out 0.1 or 0.4 or whatever minimal profits they are aimed to get out. So a big move in a particular direction due to big funds operating is not the right time for the HFT's!

Now all of that is tangential. Let's get back to the 500ms delay.

Someone wants to buy 60,000 shares at 10.2 The order arrives at the exchange and is on the market books and is executing at 10.1. Everyone can see this. The HFT's in the above case will see this @10.1 the orders at $10.1 and $10.2 will simply vanish within the milliseconds it takes for the program to realize it is time to vacate the floor to the GoldMans of the world. So HFT's "cheat" the market off liquidity. The other arbitrage context applies here, but I won't go into it as other posts explain why. Since there is a 350ms delay, the order book can't change for 350ms even after the 10000 shares have been bought at $10. So the playing field is leveled and the HFT's will have to provide a "useful" market function as well. Provide liquidity even within the exchange and can't just turn off when there is a huge order. Please note that 10,000 shares may mean nothing to these HFT's. Scale accordingly.

So the key is it takes 350ms to send and order to the exchange and 350ms to cancel the order! There in lies the leveling of the playing field. It is actually the best way to prevent the HFT's from being what they are today. Playing on the markets without contributing liquidity! Works brilliantly!

So it is not a game of 350ms + few seconds out. The computers are stuck for 350ms! even when they have further information. They cannot influence the markets by pulling out trades! Practically, the orders will remain in place for 350ms + latency time of the the fiber outside the market for a cancellation request placed as soon as the other order hit the market. So you induce a delay of almost a second back and forth. 350ms for a order to go in. 350ms for an order to be cancelled. 700ms will kill the HFT's which don't contribute to liquidity. But will keep the good human controlled HFT's in play.

I am sure, some of you will point out

Put the exchanges on discrete clocks.

[tobe]

1 Hz should be about right, maybe 10. Bids and Offers arrive and are visible but aren't acted on until the next clock edge at which point the settlements take place and a new trade queue is started.

Should do it.. as long as the frequency is reasonably low this scheme seems to enable real-time price information to be able for all players in the market.

And ban dark pools.

Re:Put the exchanges on discrete clocks.

[geekoid]

SO who gets first option to buy something?
Lets say you have 10,000 shares of a company available.
In the 1 second, there are 10 bids to each buy 2000 shares.
Who gets to buy it first?
Who gets to see the share disappear without buying?

Re:Put the exchanges on discrete clocks.

[Anonymous Coward]

The 10 bidders each get 1000 shares fulfilled.

Tax == Arbitrage

[mangu]

Imposing a tax only means the profit threshold is raised. That creates the market distortion called "arbitrage", where the relative costs between different transactions are not symmetric.

A .01% tax per transaction would mean that for me, a small trader, there would be a net loss unless my own profit per trade is lower than .01%. For a bigger trader, the cost per trade is lower, therefore they would gain and advantage over us, the smaller guys.

The true solution? Let it be, do not change anything.

Apart from some guys who get a lot of profit selling books claiming HFT is bad, no one actually makes very much on HFT. The margins are very low, extremely low, so you need to invest a lot of capital to get any profit from it.

Getting a small profit from economy of scale is something that hurts no one, it happens in every sector of the economy. As a small investor, I have an indirect gain from the higher liquidity when the big investors go into HFT.

The economy is not a zero-sum game, there are situations where everyone profits and situations where everyone loses. With HFT everyone gains, with taxes everyone loses.

Re:Tax == Arbitrage

[Anonymous Coward]

Only if you tax the profits, not the total value of the transaction.
If you tax the total value of the transaction, then HFT is impaired because it can't make rapid large transactions without losing a lot to tax. If HFT were trading small amounts very fast then it wouldn't have such an impact on the market. However, it doesn't only trade small amounts very fast, it trades large amounts very fast also. Thus, a transactions tax makes HFT less viable (if viable at all) and provides an incentive for stability.

Re:Tax == Arbitrage

[AK Marc]

Apart from some guys who get a lot of profit selling books claiming HFT is bad, no one actually makes very much on HFT.

Then why do they have so much spare cash they are spending millions of dollars on faster networks?

The economy is not a zero-sum game,

The economy isn't a zero sum game, but a single trade is, in that the more someone else makes from it, the less you do, and if you know the sell price when you buy, you can define it in terms that it is a zero sum game. I am not trying to hurt the other guy to make my position better, but HFTs do do that.

Tax them

[PeterJFraser]

Tax each one penny

Delaying everyone won't help

[Anonymous Coward]

Delaying trades by 500ms won't achieve a thing. They'll just be fighting to execute a trade after 500.008ms instead of 500.009ms.

But bundling up trades into batches, then resolving and matching the buys and sells in every 500ms window (or rather, a random 250ms to 750ms window) might do, though. It would mean that they wouldn't place an order unless they actually wanted to buy or sell the shares in question, which is tbh a pretty bloody sensible idea.

ANY delay destabilizes the system

[mangu]

If you need to sell some stock or commodity within a second of buying it, then something is wrong

Oh, yeah? Then, please, tell me in your infinite wisdom how long I should wait? Ten years? Twenty?

The fact is that ANY delay in a feedback system tends to destabilize it. In mechanical systems this is called "backlash" and there is extensive research on how to eliminate it and cope with the problems it causes. Anyone who proposes to artificially introduce backlash in a feedback system know nothing about what he is talking about.

In a market it would be trivially easy to manipulate prices if an artificial delay were involved, especially for the bigger traders. Put a buy order for a million shares and watch the prices rise, then sell at the higher price that would result a half second later. The same principle would work no matter how long the delay is.

Markets work so well because there is negative feedback in many different loops all over the economy. Some of these loops have shorter response times, other are slower to respond. If you invent an artificial delay that overlaps everything, this creates a well defined eignevalue that anyone with the proper technical knowledge could exploit.

Re:ANY delay destabilizes the system

[Anonymous Coward]

"Put a buy order for a million shares and watch"...as you're forced to execute on that order, instead of being able to pull it nanoseconds later, before other people's buys go through, but not before you got to see their bids.

Even better, tax it

[dutchwhizzman]

Put a tax on fast trading, but don't outlaw it. If you want to trade within a second, you pay a 50% profit tax. Trade within a day, 40%, trade within a week, 30%. If you lose at all, not matter how long you're holding on to stock, you don't get to subtract that from your profits. If something happens, you're free to trade, but you have to pay taxes regardless if you make a profit on the deal. If you want to cut your losses, you're tax free since you're losing on the deal. See it as a gambling tax, since either you have insider knowledge, you are manipulating the market, or you were gambling if you can consistently make money on HFT. Because providing legally accepted proof of any of these is hard, just tax it as gambling. Any proof that it's not gambling will automatically be proof of one of the others, which are illegal.

People say this won't work unless it's introduced globally, but I think that's not going to be the case. Sure, HF traders will hate it, but the companies that are actually putting up their shares will be happy to oblige and still register their stock at exchanges that will have this sort of taxes implemented. People that are actually interested in investing in a company will not mind either, since they have faith in the company and won't have to deal with market whims and price manipulations as much. Finally, the really good traders will find a way to still make money out of this, but they will have to actually look at the economy and what companies are doing, instead of whatever else they are doing now.

Don't pause them

[geekoid]

tax them at .07%.

This is so stupid.

[Anonymous Coward]

The problem with metatrading has nothing to do with timing. It has everything to do with the fact that it is FREE. There is literally no downside to buying stocks for a profit margin that basically amounts to a rounding error. Add a 0.5% tax to stock sales. Problem fucking solved.

What works

[no-body]

is a fee - best a federal tax - per trade. That will put brakes on that whole nonsene but never fly with those bribed polititians, regrettably.
All that delay stuff is baloney!

Egocentrism

[Primate Pete]

Possibly, but I'm not sure I'd choose the word 'egocentric.' I do think that you can count on the fact that *everyone* in the market is working strictly for his/her own benefit. Even the 401(k) investors and the people who want to slow the transactions down.

Yes, screw them.

[ze_jua]

I had this idea years ago. But a more perverted versions : Random delay btw 0-1500ms.

+ Mandatory random latency on financial information streams

+ Tobin tax

The benefit

[Anonymous Coward]

Is liquidity.
http://en.wikipedia.org/wiki/High-frequency_trading#Effects

Answer a question Tom

[Anonymous Coward]

How'd "eating your words" taste? http://slashdot.org/comments.p... were they flavorful (lol) seasoned with "the bitter taste of SELF-defeat" + YOUR FOOT IN YOUR MOUTH you bigmouth libelous Open SORES bullshitter?

Re:Answer a question Tom

[Anonymous Coward]

Funny Tom stfu disappearing after that post (not). Tom's busy "eating his words". Tom's polite (now that apk humbled him http://slashdot.org/comments.p... after that libel of Tom's for Tom's numerous mistakes). Tom doesn't talk with his mouth full (of his own words he had to eat).

I'm from the government and I'm here to help

[Anonymous Coward]

it turns out that once investors discovered this, they began *specifically demanding* that their trades *exclusively* be brokered through this new exchange that had this 350ms shoe-box delay. it actually caused a lot of embarrassment for a number of brokers and trading houses because the brokers were explicitly disobeying their client's instructions, because the brokers didn't understand how important this really is.

It sounds like 350ms works really well today, and the market is starting to wise up and demand this on its own. Why make it harder to switch to a lower delay some time in the future by attaching regulation to it?

Re:serious problems with networking equipment in H

[BitcoinBenny]

Your entire post is factually incorrect.

" and it's basically an inherent flaw down to the fact that the internet (TCP/IP) is routed randomly"

The Internet and TCP/IP are not routed randomly, your basic failure to understand that leads me to believe that you shouldn't be advocating for adding latency to HFT infrastructure you don't understand.

Quantum trading

[sjames]

I would propose quantum trading instead. Essentially, every trade that is entered in the same quantum happens simultaneously. For example, make it 10 minutes. It doesn't matter when within that 10 minutes you made your offer, it will all happen at the same time. And, btw, nobody gets to see what others have entered until the quantum is complete.

Sales tax

[darkwing_bmf]

Just charge a tax on the total value of whatever is bought/sold. Like 0.1%. This would eliminate any incentive for the micro-game of HFT.

Re:Sales tax

[dkf]

Just charge a tax on the total value of whatever is bought/sold. Like 0.1%. This would eliminate any incentive for the micro-game of HFT.

It shouldn't be a sales tax. It should be a tax that is required to be paid at the point where the request to trade goes in, even if it is subsequently cancelled. That stops anyone from doing tricks like trying to probe what's going on using phantom trades. It wouldn't even need to be much of a tax to discourage the bad behaviour; like that it wouldn't deter anyone who's actually interested in holding the instrument for any substantial period of time.

Big Red Button

[DarthVain]

This is currently the problem. Zero liability currently. There have been a number of LARGE examples of this, where things have gone awry, and the company loses like 500 Million. The response has been to halt trading, and reverse all the trades. To me this is cheating. They may have lost, but that just means that someone else was the winner.

If people want to use these methods, then they take the risks. They don't get to call a "redo" because things didn't work out in the way they thought it should.

After a couple of big losses like this, people might think twice about using such a service, or at least account for it within their threshold of risk. They do not own a licence to make money.

Re:Big Red Button

[tolkienfan]

No, the broken trades did not eliminate the half billion dollars that Knight lost. Quite the reverse. The broken trades were mostly "clearly erroneous" trades because they executed at wildly wrong prices. Lots of traders did make money at Knight's expense. Such is the risk of trading.

Re:Big Red Button

[DarthVain]

So if I clearly screw up, and lose a bunch of money, can I request a do over? Didn't think so.

Perhaps if trades didn't happen in milliseconds, and actually had human QA, mistakes like that would not happen (or at least not of that magnitude).

From what I read about the Knight situation, was that a development instance was put live rather than production which was untested, contained a flaw, which caused the problem. Then again that could just be management throwing the technical staff under the bus, we'll never know for sure.

Re:Big Red Button

[tolkienfan]

Reread my post. KNIGHT DID NOT GET A DO OVER. Some trades were busted, and that's only in clear cut cases. But at the end of the day Knight lost half a billion dollars. Those trades settled.

Re:Big Red Button

[tolkienfan]

Oh, and it's not whether YOU screw up, it's when the exchange screws up. E.g. if a match occurs outside the NBBO. And yes, in that case you CAN request a do over.

No need to go this far ...

[BrianPRabbit]

Placing a sales tax of, say, 1% would go far at discouraging any excessive amount of trading, for some definition of "excessive", while helping to shift the tax burden off Those less able to afford it.

Good start. Needs more delay.

[LaughingRadish]

Ever since high speed trading became a thing, I've been thinking that a delay of around a minute would be ideal to prevent the shinanigans associated with the same.

Tom = multiple account using troll

[Anonymous Coward]

And libeler: How'd "eating your words" taste? See here http://slashdot.org/comments.p... were they flavorful (lol) seasoned with "the bitter taste of SELF-defeat" + YOUR FOOT IN YOUR MOUTH you bigmouth libelous Open SORES bullshitter? As to the rest of my subject, let's let TOM speak shall we:

"I'm having great conversations on this site with one of my alias accounts" - by Tom (822) on Monday April 07, 2014 @02:29PM (#46686259) Homepage

FROM -> http://slashdot.org/comments.p...

Glorify those that game the system?

[See Attached]

Why should we glorify the process of bilking the system, and manufacturing ^H^H^H^H^H^H^H^H^H^H sucking money out of the stream? The process adds nothing, not to the company or the people or the products. We have been glorifying the gamers for.. a long long time. I recall a visit to a museum in my home town when I was 10 where they had a wampum mill on display that drilled Clam shells to make fake Lene Lenape Indian currency. It struck me wrong then too.

Tom = multiple account using trash

[Anonymous Coward]

And libeler: How'd "eating your words" taste? See here http://slashdot.org/comments.p... were they flavorful (lol) seasoned with "the bitter taste of SELF-defeat" + YOUR FOOT IN YOUR MOUTH you bigmouth libelous Open SORES bullshitter?

As to the rest of my subject, let's let TOM speak shall we:

"I'm having great conversations on this site with one of my alias accounts" - by Tom (822) on Monday April 07, 2014 @02:29PM (#46686259) Homepage

FROM -> http://slashdot.org/comments.p...

Tom = multiple /. account using scum

[Anonymous Coward]

And libeler: How'd "eating your words" taste? See here http://slashdot.org/comments.p... were they flavorful (lol) seasoned with "the bitter taste of SELF-defeat" + YOUR FOOT IN YOUR MOUTH you bigmouth libelous Open SORES bullshitter?

As to the rest of my subject, let's let TOM speak shall we:

"I'm having great conversations on this site with one of my alias accounts" - by Tom (822) on Monday April 07, 2014 @02:29PM (#46686259) Homepage

FROM -> http://slashdot.org/comments.p...

Re:serious problems with networking equipment in H

[swillden]

they needed to introduce a delay of 350ms as a way to make things fair for everyone.

Slight (well, three orders of magnitude :)) correction: That's 350 us, not 350 ms (microseconds, not milliseconds). So the proposed delay here is 1429X what the IEX guys found they needed to add.

Wouldn't work

[Anonymous Coward]

The HFT'ers will just parallel process transactions from DIFFERENT accounts all linked in the background thereby bypassing any effect at the counter.

How about, instead....

[nobuddy]

We simply hold everyone responsible for their trades. No bailouts, not rescues, no help. You fuck up, YOU burn.

Restore regulation to help keep the inventors of new scams from their short term gains before the general surge crashes that particular market segment. I think a lot fewer major scams would happen if the downside of trading was left intact.

Even better idea

[jonwil]

Just introduce a 0.001% tax on all transactions
(not just shares but other traded instruments like bonds and commodities).

Anyone buying shares or bonds or whatever to keep long term will see almost no impact from the tax. Even on a million dollar transaction, the tax would only be $1000 (so even big funds or corporate buy-outs or whatever wouldn't be affected by the tax). It would make high frequency trading (and day traders etc) unviable though.

Latency

[Anonymous Coward]

Latency is the killer of throughput. 500ms is a life time if you are trying to do millions of trades.

how about

[publiclurker]

you just grow up and face the fact that the world does not owe you a way to make money by sitting on your ass. No matter how self-important you think you are.

shares are investment in a company

[bigtreeman]

This is the wrong approach. Shares are there to provide capital for a company to do something. Make the investor leave the investment in the company for a minimum of one month. This will cause people to be more careful in their investments and change the focus to what the company is going to do in the longer term, not immediate speculative gain.

I have a much simpler solution

[Anonymous Coward]

Change $1/trade tax

Delay to human speed

[AndyCanfield]

I suggest that stock trades be delayed for a length of time comparable to a human being. A human investor can buy or sell in what time frame - 10 seconds? It would seem to me, then, that the only way human traders can hope to compete with computerized trading programs is to force the computers to wait ten seconds from submitting the trade to the trade being actuated. Otherwise, the mess we have today will solidify into the stockmarket being a robot playground with humans not allowed on the tarmac.

Ah, Memories

[CmdrTamale]

Reminds me of a story. About 40 years ago some engineers I knew got excited by the day trading of that time - in commodities. The fun went out of it when one of them couldn't sell in time and had to take delivery of 40 tons of cabbages.

As I understand it, the root tool of HFC is being able to cancel orders quickly. Let them buy and sell as fast as they like, but make them stick to their contracts.

There did I finally get something a little bit right?
--
Intelligence is realizing that nobody knows what they're talking about. Wisdom is realizing that you don't, either.

Make time discrete

[Anonymous Coward]

Drop first-come-first-served policy. Every N seconds, all the trades received in the previous N seconds get executed in a proportionally fair(*) manner. Time becomes discrete N second events, removing the millisecond bickering.

(*) service all requests where supply exceeds demand and an equal proportion of request where demand exceed supply.