I don't actually know. I can guess - I've seen requests for data. All brokers and traders have to keep accurate records of every order, cancel, etc. They may be required to explain why certain events occurred as they did. Mistakes do happen, but if a company did something that the SRO though was manipulative, they would at least be asked to stop. Having not stopped it would be very hard to claim it was an accident. We err on the safe side of compliance, as did my prior employer.
Why 4 hours? You think HFT is a leech, but you don't understand the system. Take market making. HFT market makers keep the spreads at 1c, meaning when you buy (or sell) you get the stock within 1c of its actual value according to the invisible hand of the market. You could immediately sell (or buy) that stock for a loss of only 1c. ALL THAT MONEY (1c per share!!!) goes to one or more market makers.
Now understand this: market makers did the same thing before HFT. But back then they couldn't do it efficiently, and thus the spreads were AT LEAST 25c, meaning you immediately lose, on average, 12.5c per stock.
If you knew anything, you'd understand that HFT market making makes the markets more efficient, and cheaper to trade on.
Your broker charges you more than any HFT company profits from your trades.
ES is the symbol for SPY futures. ES trades in New Jersey and SPY trades in Chicago. The price of the future is directly tied to the price of the underlying. Hence when the price changes (which happens all the time) in one venue you have to get the information to the other venue and adjust your orders there, or risk being swept and losing money.
You realize you contradicted yourself? If I take a fraction of 1c out of your investment per share, how much did you lose in order to starve to death?
Note that only 1 market maker could possibly be involved or the spreads would be wider... you buy (invest in, if you prefer) 100 shares. The seller sold them to you at the going rate. Exactly how do you think you got that "going rate"? You realize that the best ask would only be 1c higher than the price you got?
No, of course not. You know nothing, but have an opinion. And you're willing to let someone die over it.
In terms of black letter regulations, I admit it's been a while since I did the series 7...
Manipulation is definitely covered in black letter law, and flickering would come under that blanket regulation.
FYI We used "flicker monitors" to detect bad behavior in our algorithms so we could shut them down before there were any worse consequences. I guess I might have assumed there was a particular regulation involved. But any attempt to manipulate the price is definitely illegal.
If some HFT firms are using flickering, then I agree it should be outlawed.
You have to be careful, though to ensure that legitimate activity isn't affected: As a market maker placing orders the the top of book leaves you exposed to sweeps which is a significant risk. They have to be able to move their orders around, regardless of how long the order has been sitting. Otherwise market makers will be unable to profit from the narrow spreads we have today, and wider spreads means higher cost to traders - even institutional and typical investors.
You mean rebates. It works like this: You want to start a new "exchange" (ECN). No one wants to trade there - why would they, there isn't anyone buying or selling there: no liquidity. You come up with an incentive fee schedule that will encourage market makers, and liquidity providers: In every trade there is a passive and an aggressive side. Charge the aggressive side a fee (almost all exchanges do), but then rebate some of it to the passive side (almost always a market maker). Hence, companies can make money by providing the market making service to the new exchange. Traders are encouraged by plenty of liquidity and low fees (compared to the existing exchanges). The liquidity is there because of the incentives.
Note that market making is very risky: leaving passive orders around the top of book is dangerous - when stocks change in value aggressors "sweep" the book, which is expensive for a market maker. The make a very small amount from most trades, but can lose it all on a single sweep.
They have to be very low-latency to make it profitable.
And yes, it's a service. Good luck running an exchange without market makers. Why would anyone submit orders to your empty books? What quotes would you publish?
2009 was a high point in HFT in equities - I know what I'm talking about here. Trading took a huge hit due to the economy. Lower trading means less money for HFT. HFT makes money from busy markets, high liquidity and moderate volatility.
Oh, and by the way, the ECNs got started AFTER NASDAQ. The get a lot of their volume because they attracted market makers, so traders are confident that there is good liquidity there with good pricing. Guess what? The market makers are HFT. Of course, NASDAQ bought Island, and NYSE bought Arca... I guess they were too successful.
Except that its untrue: they charge for TRADEs. High message rates are a cost. They actually charge a fee by volume traded to the aggressor, and rebate at a slightly lower rate to the passive participant, thus making a profit.
This is bullshit. 1. Many exchanges have message rate limits. 2. Send ing then cancelling repeatedly is illegal, and is monitored for by the exchange 3. ECNs (basically exchanges) charge for TRADEs NOT for orders, quotes or other messages. Infrastructure for high messaging rates costs them, so they have an incentive to keep rates DOWN. In fact, they have a minimum message per trade ratio, to control this. 4. There is no way to be a mitn. Each participant sends orders and cancels to the exchange. The exchange matches orders and creates trades, in what's called a matching engine. Participants have no way to access this, and no way to "get in the middle" of other participants orders.
You have no idea what you are talking about at all.
I've worked in HFT for 7 years, at 2 companies, and I can tell you from this experience that you are wrong. Entering and order and cancelling immediately repeatedly goes by many names, e.g. flashing, and is illegal. Companies that do it will at a minimum get fined (eliminating possibly profit from it), and can be expelled from the exchange - meaning no future profit. ALL KINDS OF MANIPULATION ARE ILLEGAL. Being HFT doesn't change that.
BTW I've seen the kinds of fines that the SROs can hand out (this was from a mistake, not even manipulation), and they are enough to make you blanch.
The SEC has been investigating HFT for years, learning whatever they can, and believe me, any company that can singlehandedly push the markets around is taken very seriously. A working stock market is the SEC's #1 concern.
HFT uses that same trades that people have used for years, such as arbitrage, but using technology to make it more efficient.
Trades are usually broken due to execution outside the NBBO. I.e. they were outside regulations. It doesn't often happen, but the exchanges and ECNs can get out of whack, when there is too much volume for their systems.
Only a few things, though: 1. They aren't fitting a curve to actual sea levels, and then extrapolating. 2. There is evidence that changing the chemical composition of the atmosphere can have an effect - just look at CFCs. 3. The most accurate way to test climate models would be to make a series of predictions, and then compare to actual data. Unfortunately that would take a couple of decades. Considering the accelerating rate of gas release, some of us think it's unwise to wait for such an experiment to complete before taking action. 4. The sea level rise is estimated from the extra volume of liquid water from ice melt. We know ice is melting. 5. A sea rise of a few inches won't necessarily affect your highest tide for many years.
Many think, well let's just wait and hope; or the evidence is unclear, so we should do nothing. I think that's an awful mistake. By the time we realize we were wrong it will be too late to react. We should do the best science we can, and make the best decision we can taking in all the available evidence, and cost vs benefit.
I'm pretty sure they mean more accurate. Many people incorrectly use "precise" and "accurate" interchangeably. The article mentions using faster computers. Anyone who's done modelling knows that you can do more steps in the same amount of time, resulting in increased accuracy. They also mention better modelling.
Just because YOU are ignorant of the methods and the available accuracy doesn't mean everyone is. What's your preference, ignore the possibility that we could be destroying our world because predicting the future is difficult? Yeah, good plan.
Correlation is an important tool. It might be the most important tool.
But, climate scientists use more than correlation. They build ever more accurate models, and test them for their ability to make predictions. Like a lot of science actually.
Er, no. Pronunciation and spelling have little to do with each other, at least in English. English, eight, light, naught, subpoena, psychology, knife... the list goes on and on. Missile is spelled "missile" regardless of how you say it.
Slashdot Mirror
I don't actually know.
I can guess - I've seen requests for data. All brokers and traders have to keep accurate records of every order, cancel, etc. They may be required to explain why certain events occurred as they did.
Mistakes do happen, but if a company did something that the SRO though was manipulative, they would at least be asked to stop. Having not stopped it would be very hard to claim it was an accident.
We err on the safe side of compliance, as did my prior employer.
Why 4 hours?
You think HFT is a leech, but you don't understand the system. Take market making. HFT market makers keep the spreads at 1c, meaning when you buy (or sell) you get the stock within 1c of its actual value according to the invisible hand of the market. You could immediately sell (or buy) that stock for a loss of only 1c. ALL THAT MONEY (1c per share!!!) goes to one or more market makers.
Now understand this: market makers did the same thing before HFT. But back then they couldn't do it efficiently, and thus the spreads were AT LEAST 25c, meaning you immediately lose, on average, 12.5c per stock.
If you knew anything, you'd understand that HFT market making makes the markets more efficient, and cheaper to trade on.
Your broker charges you more than any HFT company profits from your trades.
Why is ignorance so loud?
I can give you an example.
ES is the symbol for SPY futures. ES trades in New Jersey and SPY trades in Chicago. The price of the future is directly tied to the price of the underlying. Hence when the price changes (which happens all the time) in one venue you have to get the information to the other venue and adjust your orders there, or risk being swept and losing money.
No manipulation.
Because I actually know that's not true.
You realize you contradicted yourself? If I take a fraction of 1c out of your investment per share, how much did you lose in order to starve to death?
Note that only 1 market maker could possibly be involved or the spreads would be wider... you buy (invest in, if you prefer) 100 shares. The seller sold them to you at the going rate. Exactly how do you think you got that "going rate"? You realize that the best ask would only be 1c higher than the price you got?
No, of course not. You know nothing, but have an opinion. And you're willing to let someone die over it.
Based on what?
In terms of black letter regulations, I admit it's been a while since I did the series 7...
Manipulation is definitely covered in black letter law, and flickering would come under that blanket regulation.
FYI We used "flicker monitors" to detect bad behavior in our algorithms so we could shut them down before there were any worse consequences. I guess I might have assumed there was a particular regulation involved. But any attempt to manipulate the price is definitely illegal.
If some HFT firms are using flickering, then I agree it should be outlawed.
You have to be careful, though to ensure that legitimate activity isn't affected: As a market maker placing orders the the top of book leaves you exposed to sweeps which is a significant risk. They have to be able to move their orders around, regardless of how long the order has been sitting. Otherwise market makers will be unable to profit from the narrow spreads we have today, and wider spreads means higher cost to traders - even institutional and typical investors.
"The major trading indicies are OK with this, because they are paid on a per-transaction basis..." is completely untrue.
You mean rebates. It works like this:
You want to start a new "exchange" (ECN). No one wants to trade there - why would they, there isn't anyone buying or selling there: no liquidity.
You come up with an incentive fee schedule that will encourage market makers, and liquidity providers:
In every trade there is a passive and an aggressive side. Charge the aggressive side a fee (almost all exchanges do), but then rebate some of it to the passive side (almost always a market maker).
Hence, companies can make money by providing the market making service to the new exchange. Traders are encouraged by plenty of liquidity and low fees (compared to the existing exchanges). The liquidity is there because of the incentives.
Note that market making is very risky: leaving passive orders around the top of book is dangerous - when stocks change in value aggressors "sweep" the book, which is expensive for a market maker. The make a very small amount from most trades, but can lose it all on a single sweep.
They have to be very low-latency to make it profitable.
And yes, it's a service. Good luck running an exchange without market makers. Why would anyone submit orders to your empty books? What quotes would you publish?
2009 was a high point in HFT in equities - I know what I'm talking about here. Trading took a huge hit due to the economy. Lower trading means less money for HFT. HFT makes money from busy markets, high liquidity and moderate volatility.
Oh, and by the way, the ECNs got started AFTER NASDAQ. The get a lot of their volume because they attracted market makers, so traders are confident that there is good liquidity there with good pricing. Guess what? The market makers are HFT.
Of course, NASDAQ bought Island, and NYSE bought Arca... I guess they were too successful.
Except that its untrue: they charge for TRADEs.
High message rates are a cost.
They actually charge a fee by volume traded to the aggressor, and rebate at a slightly lower rate to the passive participant, thus making a profit.
This is bullshit.
1. Many exchanges have message rate limits.
2. Send ing then cancelling repeatedly is illegal, and is monitored for by the exchange
3. ECNs (basically exchanges) charge for TRADEs NOT for orders, quotes or other messages. Infrastructure for high messaging rates costs them, so they have an incentive to keep rates DOWN. In fact, they have a minimum message per trade ratio, to control this.
4. There is no way to be a mitn. Each participant sends orders and cancels to the exchange. The exchange matches orders and creates trades, in what's called a matching engine. Participants have no way to access this, and no way to "get in the middle" of other participants orders.
You have no idea what you are talking about at all.
I've worked in HFT for 7 years, at 2 companies, and I can tell you from this experience that you are wrong.
Entering and order and cancelling immediately repeatedly goes by many names, e.g. flashing, and is illegal. Companies that do it will at a minimum get fined (eliminating possibly profit from it), and can be expelled from the exchange - meaning no future profit.
ALL KINDS OF MANIPULATION ARE ILLEGAL.
Being HFT doesn't change that.
BTW I've seen the kinds of fines that the SROs can hand out (this was from a mistake, not even manipulation), and they are enough to make you blanch.
The SEC has been investigating HFT for years, learning whatever they can, and believe me, any company that can singlehandedly push the markets around is taken very seriously. A working stock market is the SEC's #1 concern.
HFT uses that same trades that people have used for years, such as arbitrage, but using technology to make it more efficient.
Trades are usually broken due to execution outside the NBBO.
I.e. they were outside regulations. It doesn't often happen, but the exchanges and ECNs can get out of whack, when there is too much volume for their systems.
More like they are trying to maintain the illusion that they are working on our behalf.
I agree with most of what you said.
Only a few things, though:
1. They aren't fitting a curve to actual sea levels, and then extrapolating.
2. There is evidence that changing the chemical composition of the atmosphere can have an effect - just look at CFCs.
3. The most accurate way to test climate models would be to make a series of predictions, and then compare to actual data. Unfortunately that would take a couple of decades. Considering the accelerating rate of gas release, some of us think it's unwise to wait for such an experiment to complete before taking action.
4. The sea level rise is estimated from the extra volume of liquid water from ice melt. We know ice is melting.
5. A sea rise of a few inches won't necessarily affect your highest tide for many years.
Many think, well let's just wait and hope; or the evidence is unclear, so we should do nothing.
I think that's an awful mistake. By the time we realize we were wrong it will be too late to react.
We should do the best science we can, and make the best decision we can taking in all the available evidence, and cost vs benefit.
To test the models appropriately takes at least a decade. So, yeah, they're working on it.
In the meantime, they have a number of models written by different groups with different assumptions, etc., and they give broadly similar results.
But I'm sure you have a better approach?
I'm pretty sure they mean more accurate. Many people incorrectly use "precise" and "accurate" interchangeably.
The article mentions using faster computers. Anyone who's done modelling knows that you can do more steps in the same amount of time, resulting in increased accuracy. They also mention better modelling.
Just because YOU are ignorant of the methods and the available accuracy doesn't mean everyone is.
What's your preference, ignore the possibility that we could be destroying our world because predicting the future is difficult?
Yeah, good plan.
Correlation is an important tool. It might be the most important tool.
But, climate scientists use more than correlation. They build ever more accurate models, and test them for their ability to make predictions. Like a lot of science actually.
Global warming doesn't care whether you believe in it.
I'd bet they'll find some excuse to blacklist the key in the not too distant future, and make it a bit more painful to use a non-windows OS.
We can define "better" any way we like.
Going by the responses, I guess I *was* trolling.
Oops! :)
Er, no.
Pronunciation and spelling have little to do with each other, at least in English.
English, eight, light, naught, subpoena, psychology, knife... the list goes on and on.
Missile is spelled "missile" regardless of how you say it.
This my favorite answer so far. :)