Slashdot Mirror
NASDAQ Trading Halted Due To "Technical Issue"
barlevg writes "The Wall Street Journal reports that trading was halted midday Thursday due to an as-yet unnamed technical issue. Says SEC spokesperson John Nester, 'We are monitoring the situation and in are close contact with the exchanges.'"
Have you tried turning it off and on again?
Does this make anyone else happy for some reason?
"When information is power, privacy is freedom" - Jah-Wren Ryel
Some HFTs went haywire, and the little-guy traders started to get the half cent per transaction that they actually deserve. We can't have that, can we?
nuff said. MUAHAHAHAHAH ! [ disclaimer: this post is neither anonymous nor cowardly ]
Religous speak to God. Insane are spoken to by God. When all shut up, one can finally hear Shostakovich in peace
Their stock has been slipping, and right when it dipped below 500 per share, NASDAQ shits the bed.
Coincidence? I think... yeah, probably, actually.
...which is why I keep all my money in bitcoins.
I'm an old school (1930s era) value investor - Ben Graham type of investor.
I don't give shit. I don't care what the know nothings on CNBC have to say (I don't think Becky is all that, BTW) nor do I care what Warren Buffet has to say - publicly.
I got an alpha of 20% right now and it's freaking me out because I'm thinking there's something wrong with my calcs. No, overall I'm up like 50+% year to date but I'm freaking out because i KNOW - I KNOW - there's luck involved and I WANT to weed it out so I can plan better.
If I were a Hedge fund or mutual fund manager, I'd be interviewed in the press and folks would be patting me on the back for my "brilliance" - fucking morons- all of them.
I got LUCKY and I'm too stupid to figure out where!!
- Flop sweating their asses off
- Furiously searching their email for that ass-covering memo to their boss about the pricey "redundant this" or "redundant that" that the boss was too cheap to get
- Wondering if there is enough alcohol on earth for what they will need later tonight
---- "Logoff! That cookie shit makes me nervous!" - A. Soprano
Quick, everybody spread rumors about shutdowns and buyouts, those Wall Street fatcats will shit their pants! >:D
"When information is power, privacy is freedom" - Jah-Wren Ryel
Technically, someone screwed up a bid and lost a billion dollars. Or, at least, that's the most common cause of stoppages I hear about these days.
I've got a dollar that says there will be a roll-back.
a.k.a Market Crash
Same thing with the crooked banks, when they have a Holiday now, and you can't get your money out, it is a "Hacker" or a "Terrorist" or some other garbage.
-Hack
Got Geometrodynamics? Awe, too hard to figure out? Too bad.
They forgot to restock their supply of Greed Oil.
Seriously. Is there any real need (beyond that for connected players to be able skim money off the top) for anyone to be able to sell and buy stock (or commodities) in a tiny fraction of a second, instead of say, once every fifteen minutes or even longer?
Must be Windows Updates forcing a reboot. That's usually the problem.
In case you didn't know Goldman is never supposed to lose money on a trade, which is why they can report 100+ days of consecutive trading profits, which is a mathematical impossibility in a non-rigged market.
This happened right around the time that the new Neverwinter module was launched. Coincidence? I think not. Perhaps the people in charge installed the Neverwinter client and were playing the new area instead of monitoring the exchange.
"Technical Issue" = Apple below 500.
Well known error condition.
Slow news day?
Does it really add any value to society and therefore worthy of generating money from actually getting nothing done?
Seems its a game that we are all financing..... a sad sad game.
I don't understand the problem. Being first in line is not always advantageous. It just makes you first in line. First to make a good deal. But first to make a bad deal too. Nothing "magical" about it that generates you magical profits.
Seven puppies were harmed during the making of this post.
Towards the end of the article it says that this is the latest in a string of technology-related mishaps affecting exchanges and brokers as markets over the past two decades have migrated to electronic systems. That's an absurd statement considering two decades is a very long time. Besides, computers were first used in trading in the 1960's. What's happened in recent years that's made market systems more vulnerable?
I do not block ads. I do block third party scripts.
Sure.
In the old days (80s, 90s), when it was seconds, the middle men grabbed 12.5 to 25 cents per share. Before then, when it was minutes, they would grab 50 cents. Costs for the average small investor have fallen by over 90%. If you invest in index funds your costs have fallen by over 95%. (But wait you say – I don’t trade my index funds. Look at your funds expense ratio, pull out the supplementary prospectus information on what portion of that is trading costs for the past 20 years, and gap.)
Or, to put it another way, would you rather have dozens of HFT fighting for your business or an oligopoly of clique, cozy partnerships. Not saying it is perfect but that it is an improvement.
Seriously. Is there any real need (beyond that for connected players to be able skim money off the top) for anyone to be able to sell and buy stock (or commodities) in a tiny fraction of a second, instead of say, once every fifteen minutes or even longer?
Time is money. Time passes as new information is acquired or transmitted. Shorter time intervals will therefore always be desirable for making a market more efficient and to accurately reflect reality. The more time it takes to buy and sell, the more risk there is in buying or holding stock, because the stock is less liquid and its value can change dramatically in very little time.
Yes. Just like you can't make a very good car if you only have one cylinder fire ever 15 seconds (though you could make something with wheels that lurched about). I've explained the technical details about a dozen times on /. before, so at this point I'll just say: you know that guy who believes "I don't understand it so it must be easy"? Don't be that guy. Education, then opinion.
Socialism: a lie told by totalitarians and believed by fools.
NASDAQ down today. All of Google down for a few minutes recently. Many of Apple's services down today. Some of Microsoft's services down last week. I wonder what's up with that?
There is $1B+/year in HFT profits that indicate otherwise. It's not just that they trade first but that much of their trading is risk-free by getting in the middle of other slower orders on both sides of the market.
beyond that for connected players to be able skim money off the top
Let me stop you right there.
It's better to vote for what you want and not get it than to vote for what you don't want and get it.
- E. Debs
Because they can. If you don't like it, try coming up with a system that restricts trading to only at 15 minute intervals without tripping over yourself trying to find out whether all transactions are valid or not. This isn't a "We're so powerful, we can push rules that make us win," it's "This is the way computer systems work, deal with it."
... the group of MBAs ...
- Flop sweating their asses of
- Furiously searching their email for that ass-covering memo where the IT guy said "Yeah, this should work"
- Wondering if there is enough coke on earth to get them through the rest of the day
For these guys, there are only two universal truths:
1) This is absolutely, positively, 100% the IT guy's fault
2) He can not fix this without the IT guy.
The impotent rage would be palpable.
Yep, because they know how to trade. HFT won't guarantee a winning trade however, when the market moves the other way. And believe me it happens often. There's more to it than just having a fast connection. For an example I can point to KCG (Knight Capital Group), whose HFT programs lost them what was it, 400+ million in a few minutes when FB was launched? It broke the company, anyway. HFT is not always "good". You have to know how to use it, too.
Seven puppies were harmed during the making of this post.
Seriously. Is there any real need (beyond that for connected players to be able skim money off the top) for anyone to be able to sell and buy stock (or commodities) in a tiny fraction of a second, instead of say, once every fifteen minutes or even longer?
Time is money. Time passes as new information is acquired or transmitted. Shorter time intervals will therefore always be desirable for making a market more efficient and to accurately reflect reality. The more time it takes to buy and sell, the more risk there is in buying or holding stock, because the stock is less liquid and its value can change dramatically in very little time.
Wait, wait, wait. I'm grokking some of the other points in favor of high-frequency trading, but are you actually claiming that the stock market reflects reality?
"There is $1B+/year in HFT profits"
Which means there is also $1B+/year in losses to someone, caused by HFT. Guess what? If it's the HFT traders making a profit, it's the investors taking the hit.
"National Security is the chief cause of national insecurity." - Celine's First Law
adds money into the system from the people willing to "invest" into the funds that do this
most of the trading every day is a tiny percentage of shares traded by hedge funds
I think we need more often than every fifteen minutes, but I'm not sure we need more often than every 15 seconds.
Every trade is an information exchange. It puts information into the market. If you have trades every 15 minutes, and two actors have conflicting information, it will take a large amount of time before they agree - ie, the price stabilizes. This provides a bandwidth limit to the amount of information that can be transmitted into the market - in other words, it makes the information less accurate. In practice, this will probably manifest as much larger spreads, which makes any exchange more expensive.
Now, at some point information pushing stops being the case, and it turns into somebody scalping off the information that's already in the market. That's what HFT is about - somebody analyses the price movement, and extracts the information content from that - bringing that information to the market a millisecond earlier, but taking the value that was really generated by whoever was doing the trade that started the small price movement. In practice, as far as I can tell, this is usually front-running, and effectively takes away value from longer term investors. Example: pension fund PF is going to sell $100M of MSFT. Nobody is going to buy all of that at once, so they put it out in $1M lots at market value, as quickly as their pipe can do it. HFT trader HF sees that there are $1M lots coming in, and shorts a bunch of $1M lots of MSFT to drive the price down (taking the value of each of these), and then covers their shorts by buying the MSFT shares from PF at the now depressed prices. In practice, how they do this is a tiny bit more complicated than this, but it morally end up with the same thing.
Seriously. Is there any real need (beyond that for connected players to be able skim money off the top) for anyone to be able to sell and buy stock (or commodities) in a tiny fraction of a second, instead of say, once every fifteen minutes or even longer?
The market operates for the near-exclusive benefit of those HFT players. The SEC ensures that a competing market cannot replace it("them"), while the revolving door between Wall Street, Treasury, and The Fed, continues unabated.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
A billion does not seem to be too much to pay for liquidity in a $36 trillion dollar market.
You don't make money by "getting in the middle of slower orders" in any risk-free way. Markets don't work that way. You do make money by being the first to trade on "news", but better that than insider trading (i.e., better to trade 1 ms after than 1 week before). You do make money by taking a little risk as a market maker, but HFT has squeezed profits there very thin.
Do you have any idea how tiny $1 B/year is compared to the amount of stock that trades each year? I'm sure profits are much higher, just on first-after-news trading, and that doesn't mean there's a problem.
When I trade (being a little guy who doesn't follow the markets constantly) I get a better price thanks to HFT: the bid-ask gap is tiny these days, often 1 cent, and my broker makes $10 on the same trade the HFT guy makes $0.01 or so.
Socialism: a lie told by totalitarians and believed by fools.
I told the manager we should test it thoroughly but he just wouldn't listen ;)
Assuming Direct Control.
Latest news: They are going to start trading at 2:45 EDT.
My prediction is a brief panic sell-off, followed by a return to normalcy, considering putting in a few limit orders.
Runs on kittens, puppies and unicorn blood now.
I just wish the market operated on a tick.
As things are now, they're supposed to be first come, first served. That's why all the HFTs pay mega bucks to be collocated in the same datacenter as the exchanges. This means any hickups affect the order that things are processed. It's non-deterministic by design.
It makes much more since to run on a tick. All transactions would processed once every second. You still need some sort of ordering, but it would allow for many things to happen. Including realtime offsite backups, and moving the servers without everyone screaming bloody murder.
Think about it, the stock exchanges can't be in more than one data center because otherwise the HFTs would freak. It's just too non deterministic to even have a failover location.
They could even have a bidding war for transaction priority. Instead of HFTs paying for microwave links, they could just straight up bid to have their transactions processed first in the queue.
So lets pretend that we've just completed writing this code, as opposed to having just completed sabotaging it -Altera
... Prior to re-opening the market
I think what you want to say is “only a billion” – as in it is small and falling number in historical terms.
Then let’s view trading costs as analogous to waste friction in a mechanical system – the lower the better. How should we measure this waste? Read up on “Implementation shortfall”. It’s the gold standard in the industry. (Rarely implemented because it’s complex, but still, the best theoretical method of measuring waste.)
Now, anybody claiming much about how much made what when or the effect of which market maker or HFT needs to be treated with a large grain of salt. Those numbers are highly propitiatory. However, we can, and have, measured the cost of trading at mutual funds – which have to publish a lot of public information.
Rewind the clock to the 90s and you can see that the costs have fallen by 80%. Why is that? In the old days, the 30 odd market makers on the NYSE would make billions each year – mostly risk fee. We can probably trace about ½ the gain to HFT.
This software is all written in Java these days, isn't it?
Probably just paused for a few hours for garbage collection.
The obvious flaw in your argument is that trading costs have gone down due to electronic trading which is different from high-frequency trading. With electronic trading, instead of brokers having to call other brokers to make the trades, buyers and sellers can be found almost instantaneously to fill orders. That is a good thing and no one is arguing against that. What people hate is how high-frequency trading is used to underhandedly find a buyer's maximum buy price and the seller's minimum sell price so that the traders can keep the difference in addition to charging their commission fees. HFT is not required for a company to participate in electronic trading and yet it is prevalent throughout the market because Wall Street owns Capitol Hill. And I wouldn't be surprised if this shutdown was caused by a HFT algorithm that went haywire, although there doesn't seem to be any evidence of that at the moment.
I think this is a terrible problem with education in America. People are afraid of the market, don't understand it, don't want to understand it, but that's due to simple lack of education. And it's important to know the basics, since it will likely affect your standard of living in retirement.
Just like there's a certain minimum amount you need to know about how cars work before you can drive safely - not all that much, but there are a several hours about it in most drivers ed classes - there's a certain minimum amount you need to know about how markets and investments work. Where's the public education for that? Are we so intent on class warfare that we'll cut off our nose to spite our face here?
I've already made more from my stock and mutual fund investments this year than I'll make in salary. Now I'm going to post a response to the guy who talked up how scary market investing is so I can scare more people away from the market.
Just kidding about scaring people away. Making money in the stock market isn't a zero sum game. The more players there are and the more money invested, the better it is for everyone except the people who sit on the side lines becaused they're scared/don't understand/don't want to learn. If you have any "spare money", you are absolutely stoooopid if you aren't investing it. If you don't want to expend too much effort; low cost, broad based index mutual funds (e.g., S&P 500) are an easy way to at least track the market.
Cheers,
Dave
They that can give up essential liberty to obtain a little temporary safety deserve neither safety nor liberty.
Ben
I was searching for Archer quotes, but I accidentally found this:
http://finance.yahoo.com/blogs/breakout/hindenburg-omen-very-ominous-high-technical-warning-sign-163004190.html
I wonder if there is a relation?
When I trade (being a little guy who doesn't follow the markets constantly) I get a better price thanks to HFT: the bid-ask gap is tiny these days, often 1 cent, and my broker makes $10 on the same trade the HFT guy makes $0.01 or so.
Only if the broker trades on the open market, which is his last resort. If he trades internally, he takes the whole $10.01. If he trades on a dark market, he takes $10.005.
... and cockiness you will do no better than throwing random darts at a dart board.
Hint: You cannot predict a system as complex as the stock market, no matter how many successes at prediction (or was it luck?) you think you've had in the past.
Read up on fragility and perhaps more importantly Antifragility.
I was responding specifically to the OP ½ cent mark. In the 90s, under fully electronic trading, the bid/ask spread was between 12.5 to 25 cents . Today it is under a penny. Why the compression? Computers are not like fairy dust – sprinkling them around does not automaticity solve problems. It is the fact that we have gone from 4 market markers to dozens of HFT acting as de facto market makers.
(Which then brings up a tricky question of what a HFT is. The big boring index funds trade like a HFT in an electronic exchange even though their intentions are radically different.)
As for the tick system – The problem is that you are assuming that at any given second at a given price that supply and demand are equal. When there is an imbalance how do you pick which trades gets execute and which ones are delayed. There are markets that work on “order book” where a exchange agent manages the flows. In practice, these markets have higher trading costs once you factor in indirect costs.
(And I am not so sure about stock exchanges not being in more than one place. NASDAQ is heavily decentralized. Speculating, I would guess the outage was due to a software not hardward.)
You KNOW the NSA has a server (or two) in the middle of all this.. Makes ya wonder eh?
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
HFT won't guarantee a winning trade however,
Yes, it does. At least the way it has been set up. HFT lets some traders outbid others by submitting a whole series of offers and then canceling the one that won't make them any money. Its an abuse of a feature demanded by HF traders to unwind 'system errors' created by such rapid transactions. In reality, its like being able to put a bet down at a roulette table and then pick it back up quickly when you see where the ball lands.
Its fraud and HF traders should be in prison. Except that the brokerages like the commission business.
Have gnu, will travel.
The stock market was complete shit until HFT came along. Got it.
HFT does improve market efficiency. In other words, the price you pay at Exchange A for Equity Z is going to be almost exactly the price you'd pay at Exchange B for Equity Z. This applies to multi-listed instruments only, of course. Not sure how important that is to the average Joe, probably not much.
All the regulator has to do is introduce a very small charge for every share traded.
Let's say something like USD $0.0001 per share.
Feed the funds collected back to the exchanges to pay for the networking and the compute consumed by the very high-speed traders.
Problem solved.
This really only affected the gamblers in the market using NASDAQ options, not people buying and selling actual shares.
It didn't impact limit orders up/down, or any buy or sell orders, just the gamblers that exaggerate noise in the market signal to bet against other computer systems that exaggerate more noise.
No actual investors were harmed, only gamblers. You could see it in the calm reaction of the shares - instead of bouncing up and down as gamblers tried to tweak it up and down for profits, it remained stable, since the reality was that the share pricing does not actually trade that much without gamblers distorting the signal.
-- Tigger warning: This post may contain tiggers! --
The rise of HFT is similar to bitcoin. To encourage liquidity in the market, exchanges started to offer a slight rebate to organizations that put standing side orders to trade a stock at the current best price (buy or sell) which would be executed if there was a temporary liquidity problem. Analgously, Bitcoin offered a little bit of payback to those that did the transaction hash work to help assure short transaction validation times.
Initially, large trading platforms that were acting as Supplemental Liquidity Providers (SLP) were dabbling in HFT as a way of indirectly chasing these rebates (by gaming the price and acting as a SLP, you had a net price advantage in that stock due to the rebate). With bitcoin, the miners dabbled with high performance hardware to chase the early large bitcoin rewards, but really didn't try to game it (e.g., flooding other node with bogus transactions, whilst you worked on the most promising hashchain) since the economic didn't work that way.
Of course once you have this high-tech high-speed trading platform, you can attempt to use it to game the price even in non-liquidity crunch situations. With bitcoin there is this potential fear that with enough computing resources you can force your own hash chain, but it probably isn't worth it. With stocks... well, that's where we are with HFT...
Could they be proactively stopping another feedback loop?
I swear to God...I swear to God! That is NOT how you treat your human!
Average daily volume on the NASDAQ is a little over 400,000,000 shares for the last three months. Even with the outage, today's volume was 924,433,630 shares. Do you really think slowing that down to only trading on some specific interval would work? Plus, you have to remember that people decide to buy or sell on something other than your schedule and the whole point of a stock exchange is to price the security and execute the trade. All putting in some sort of fixed interval would do is force traders to manually price each trade and then somehow fix that price up to the next trade interval. It simply wouldn't work.
Cheers,
Dave
They that can give up essential liberty to obtain a little temporary safety deserve neither safety nor liberty.
Ben
I understand. I just chose a post that seemed to be related to write my little rant.
I'm not assuming supply and demand are equal. What I am saying is that the current method of choosing which get executed, and which get delayed is stupid and non deterministic. It's first come, first served, which results in companies paying millions of dollars to shave off 1/10 millisecond worth of latency. That number is for example use only, If you want a real number, Google it.
Let me give an example for a single tick.
Start of Tick.
1. The servers take all new buy and sell orders and process them.
*Each buy and sell order is compared against both the new ones, and the old ones in the database.
*Whenever a buy/sell match is detected a transaction takes place.
**When this happens the relevant buy/sell orders are removed from the database and the list of new orders.
***********See bottom for when there are multiple possible matches.
2. The new orders (minus any removed) are entered into the database.
3. The new orders (minus any removed) are sent to all parties.
End Of tick.
When there are multiple possible transactions there are a few things to keep in mind.
1. The seller wants the most money per item.
2. The buyer wants to pay the least money per item.
3. Buyers and sellers are working off knowledge of what was in the database of the last tick. They might not even know about this better deal that happened this tick.
4. Some orders are large, and some are small.
There are multiple ways of resolving this, including my earlier notion of bidding for a higher priority. I'm not going to say what's better. What I am going to say is fist come first served has resulted in some really stupid things.
One more thing:
Currently, buyers and sellers set a publicly available buy/sell amount. They also set a private amount with a minimum/maximum buy/sell price. The trick is HFTs constantly create and then cancel transactions to find these minimum and maximum amounts. I wouldn't be surprised if most brokerage services do this as well. To get rid of all this nonsense, I am against allowing transactions to be canceled. You'll see my proposed model doesn't even allow it. I am also for one publicly available number. This half-assed hiding does nothing but add complexity, and invite people to cheat the system.
Hope you enjoyed the dump. I'm not fully informed of how the system fully works, but on the other hand there are probably only a handful of people who really do. I wonder how old the codebase really is.
So lets pretend that we've just completed writing this code, as opposed to having just completed sabotaging it -Altera
I suspect it's very much more than the billion $ figure I was responding to, given the investments being made in new fiber to shave microseconds, to hire $1M programmers and quants, etc.
You claim of causality based on correlation rings hollow. There have been other significant changes which increase liquidity, which you're ignoring. Common investors being able to make near-immediate trades via the Internet rather than with a phone call to a broker. Changes from fractional dollar to penny (or finer) valuations. Ready access to real time quotes for common investors. $5 or $10 flat priced trades, rather than $xx + % (+ more for odd lots).
HF traders are leeches. They're not altruists, there for the good of the market, they're there to suck blood.
"National Security is the chief cause of national insecurity." - Celine's First Law
Investors don't dump a stock milliseconds after an earnings report is released.
Real investors buy something and hold it for months or YEARS. When a stock drops 40% in value in seconds due to earnings being less than expected, thats probably a good time to BUY!
Do you know anything about “order driven market”(OD)?–Because your method sounds like that type of system (If I am wrong, let me know). NASDAQ, and most of the world, uses quote driven (QD) system.
OD systems do work. The German Bourse and Paris Bourse were 2 good examples. However, most OD stock exchanges have moved to (QD) in the past 20 years because they are more efficient. The same stock could be trading both on the London and Paris exchange and the London exchange would offer the lower cost.
Now let me be specific about cost – I am talking about total cost, both explicit, implicate, and opportunity costs. You may refer to it as quality but the industry calls it a cost. QD systems are more certain. Your order is more likely to be executed at the price you have specified. With OD systems your order tends to go through a couple of “ticks” and the price wanders from your decision point so the price is different. A lot of trades are canceled.
There are decades worth of data to back this up and I am approaching this form a purely pragmatic viewpoint, I am not claiming that QD systems are inherently or theoretically better than OD systems – and I am sure we can find a few cases where OD systems are better.
I'm not sure we need it more often than every 24 hours.
Take down the exchange!
Less causation then you think.
Of the 80% reduction in costs, 20% is attributed to compressed bid/ask spread and 60% For both of those figures, most of the gain in both of figures can be attributed to high speed algorithmic trading systems.
Who uses speed algorithmic trading systems? Pension funds, trading desks, mutual funds, hedge funds, market markets, etc. Which of these are HFT? That’s a subjective judgment – kind of boils down to the person’s intent.
And another reply - I just remember some OD markets – “Dark Pools”, like Instinet are OD, not QD.
(And I dislike Dark Pools, not because they are OD but because they are dark.)
What I am saying is that the current method of choosing which get executed, and which get delayed is stupid and non deterministic. It's first come, first served
"First come, first served"? That's quite deterministic.
What people hate is how high-frequency trading is used to underhandedly find a buyer's maximum buy price and the seller's minimum sell price so that the traders can keep the difference in addition to charging their commission fees.
And who on Slashdot trades in such a way that this would matter? You're basically speaking of HFT gaming of computer traders. And there are ways to keep that from being an issue (such as slowing down your rate and volume of purchase at those less favorable prices). If those people don't like giving free money to HFT, then debug their damn trading programs.
I see no reason to fix yet another non-problem in the trading of securities.
"claiming that the stock market reflects reality?" More accurately it reflects people's perception of reality, more like a instantaneously conducted poll. HFT just allows you to say if a then I feel good or if b I feel bad or if... profit! People can be wrong that is why mr invisible market will always have bubbles, bursts etc because we can't take in all info and process it quickly enough it is just a better planning method than any other because people who are consistently wrong either are giving their money away or drummed out but there will always be people to replace them.
Money; derived through interception of stock trades and pick from analysts.
NSA copied "SETI@Home" and turned its parts into TERROR@Home.
In terms of stock picks and intercepted bets, NSA wants to beat the House (Global Stock Markets).
That's all! It was never about fighting 'Terrorists' if Terrorists does not include the US NSA, DoJ and White House.
Very simple story of greed. That's all. Nothing more.
Unless you mean they realized the money they were stealing didn't TECHNICALLY exist.
They increase liquidity. Suppose I wanted to sell stock, and it took a week to find a buyer. I might be happy to find somebody who'd cut that down to a day for a small slice of the sale value. This is what stock exchanges were about; they'd charge some money, and allow your broker to find somebody to buy your stock fast. There's times when an investor won't want to wait fifteen minutes, and you can make several decisions in that time.
Now, I'm not sure what liquidity is worth, and I don't know why it really helps to have transactions take less time than a human can think about them, and I don't even know, really, how much they skim, so I can't say if it's worth it. There's no actual need for HFT, since we had stock markets long before it, but it may be worth what it costs.
"When you have eliminated the unacceptable, whatever is left, however improbable, must be the truthiness" - Holmes
It is best to be first in line if you are scalping movie concert tickets.
And the HFT is like a person who knows what the people at the end of the line want. If you see lots of people in line to buy a pen, then you buy out the stock of all of them at the front, then start your own line selling pens. Then, when nobody wants your pens, you return the unsold ones for a refund. HFT works for trade snipers, who look at trade requests, jump to the front of the line, buy it, then turn around and sell it to the guy next in line. You don't have to make much per trade if you make 1,000,000 trades per second. $0.01 per trade is enough to make you more than a million dollars a year, doing nothing but being a leach.
Learn to love Alaska
It's like you see the end of the line, and start walking towards it. Someone sees you going towards the line and sprints and gets there first. They did so *because* they saw you headed for the line, but you couldn't see them until they were in line. Is that still deterministic? Does it matter from who's point of view?
Learn to love Alaska
It seems you are rejecting the high-level vision of his change for a minor implementation detail. Why couldn't you set up a QD system that was clocked? I think the only "fair" system is to do it that way. Eliminate the gains from better access and co-location and such. I don't care about the QD/OD difference, I care about the arbitrage others can use to abuse me, but I can't employ the same tactics against them. The system is inherently unfair.
Learn to love Alaska
Minor implementation differences and fairness. Hmmm.
What you see as a minor implementing difference is the key difference between QD and OD.
I should point out that OD markets can be abused. The obvious one is if you can figure out the orders in the orderbook / database. If you can figure this out you will have the edge on your completion. (This is particularly true when you are trying to sell a large block of stock that can’t be cleared on a single trade.)
As for fairness, do you want to be theoretically fair or practically fair. Would you be happier if you had to pay more for a system that was “fair”. Look up Shortfall Implementation – I think wiki has a very short article on it. It is the gold standard on determining order cost and quality. For the past 50 years QD systems offer better costs then OD systems. It has narrower margins, deeper liquidity, and more robust during dynamic swings.
The issue that I have is that the two systems have been tried out side by side and QD offers the better value which, I would argue, is more fair.
I get it, you will shoot down anything. Makes you feel smart to tell how others can't fix it. But you are obviously not smart enough to think of a single way of implementing a clocked market that wasn't worse than today in apparently every way. And, in typical slasdot style, anything that you can't think of in 10 seconds is impossible. That, and looking for obvious edge cases that would be fixed in implementation, even if not fully specified in a "would this work" discussion and shooting them down just shows you want to prove others wrong, not discuss the facts. Trade size would be limited. If you wanted to trade a block of 4x max trade size, you'd have to submit 4 (or more) trades. If there was a max trade size, and no minimum (other than the unity of a stock, but that's not even a strict limitation today - but partial shares paid as part of dividend reinvestment may not be traded on the open market, only with the dividend payer, I can't be arsed to look, as it's a detail irrelevant to the overall questions here), then the person with 4x max trade had a decision to make of whether to submit it as 4 trades (minimum number) or the maximum number of trades allowed.
And paying more for a system that's "fair" would be "fair" only if the extra payment was fairly distributed. With HFT, the payment isn't distributed, it's concentrated in a few leaches hands.
It would be hard to create a market from scratch that was as unfair as the one we have now, while appearing to be fair to the uneducated. But yes, it would be easy to create a worse market - it's just that the unfairness in the simplistic markets that can be described with less than a doctoral thesis will be obvious, even if the limitations wouldn't survive into implementation, the simple minded are required to point them out in the least productive way possible.
Learn to love Alaska