HFT robots seeing orders before the exchange ?
- Thread starter softdown
- Start date
For whatever its worth, I bought tick data and went back and quantified this for one of my models (the study covered something like 5000 trades over the course of 1.5 years). It was a considerable undertaking to go back and compare my fills using IB SMART vs. NBBO...however, at the time I was convinced that if I did it, I would come to some conclusions that would greatly improve my execution. In the end, the fills I got from SMART dark pools we were way better than anything I could devise using directed limit orders (tried combinations of both passive and marketable). Of course, no replication is perfect since I couldn't model hidden orders at the venue I would have directed orders to, but the results were so overwhelming that I doubt it would have mattered. I was surprised (and disappointed) that there wasn't an obvious improvement for my execution over marketable SMART routed orders. Again, this was only for one unique model and I'm sure results differ based on strategy/underlyings, but I thought I'd share the findings.
I'm giving DeeDeeTwo full credit for being highly experienced and fully competent (although you never know...) ... and I'm sure he/she can fully justify the quoted statement from the perspective of the DeeDeeTwo trading program (I can tell my trading program is very different because my average spreads are still around $0.10 ... DeeDeeTwo seems to be taking a lot of liquidity) ... but for all who are following this thread and interested in liquidity provision strategies for illiquids, it behooves you to know the facts:
The "enemies" (HFTs, predatory algos, dark pool volume-suckers, order internalizers of all stripes, and everyone else I'm too ignorant to name) have DECIMATED the little-guy retail liquidity provision game since they started ramping up in 2007, and the situation gets more grim every day ... they take almost ALL of the good, low-risk liquidity I used to feast on and leave guys like me with only the bad, high risk trades ... yes I still make a decent living but ROC was 50%+ from 2002 - 2009, fell to about 30% in 2010 and has been less than 20% for 2011 - 2012, and still falling like a rock ... and that lousy 20% is earned with FAR more risk than I used to take to get 50%. Proceed at your own risk ... if the SEC doesn't do something about cancerous order internalization and/or preferential access to better/faster data feeds then things will only get worse at a rapid rate.
That *should* be happening. The existence of "little-guy retail liquidity" is an historical aberration made possible by occasional step functions in technology. Once the transformation is done, it naturally goes away.
No argument with Random.Capital ... I've been prepared for 2 years to die off and be resurrected with a new trading program that doesn't compete with the "enemies" ... just trying to save time and pain for some ET readers.
Also, forgot to mention in response to DeeDeeTwo's first post: The "enemies" didn't care about tiny-volume issues until the last couple years ... now they LOVE them and are taking all the $$$. This makes perfect sense since once you've got the code written for the higher-volume issues it's just a switch-flip to apply it to everything, even the lowest-volume issues.
Also, forgot to mention in response to DeeDeeTwo's first post: The "enemies" didn't care about tiny-volume issues until the last couple years ... now they LOVE them and are taking all the $$$. This makes perfect sense since once you've got the code written for the higher-volume issues it's just a switch-flip to apply it to everything, even the lowest-volume issues.
here is another example of what US stock market have became.
pathetic 600$ total order..look at the execution of it.
last 100+ shares are sold at current bid at that time. bid i think was 200 x 3.77
this whole thing is not even reported in T&S(cause they are all odd chunks). good luck for all of "tape readers" or volume examiners . good luck to find some transparency here.
then there is couple executions @ 3.85-3.84(last trade before 4:00 included)
but reported closing and "last" trade is 3.75. go f** figure all this mess. just one little tiny example..i'm not saying that this is fair or nor fair,,hft or not hft..i'm just saying that this whole thing stinks and it shouldn't be like this. and this is precisely why all humans abandon US stock market. cause anyone can smell that something is not right(or maybe everything) .keep it up SEC! great job!
PS: hey..even madoff is saying it's not right..
pathetic 600$ total order..look at the execution of it.
last 100+ shares are sold at current bid at that time. bid i think was 200 x 3.77
this whole thing is not even reported in T&S(cause they are all odd chunks). good luck for all of "tape readers" or volume examiners . good luck to find some transparency here.
then there is couple executions @ 3.85-3.84(last trade before 4:00 included)
but reported closing and "last" trade is 3.75. go f** figure all this mess. just one little tiny example..i'm not saying that this is fair or nor fair,,hft or not hft..i'm just saying that this whole thing stinks and it shouldn't be like this. and this is precisely why all humans abandon US stock market. cause anyone can smell that something is not right(or maybe everything) .keep it up SEC! great job!
PS: hey..even madoff is saying it's not right..
This post was muddied from the start...
Because the ZH example was micro-volume.
Low volume would be something exotic trading 100,000 shares/day...
That eliminates ETFs, ETNs, or tracking vehicles.
Exotic means it requires real human EXPERTISE to avoid pitfalls...
Which "HFT robots" totally lack...
Being about as "intelligent" as other forms of spam.
Also, the real issue here...
Is the imperviousness of Limit Order flooding to "HFT robots"...
Go ahead and spoof me, penny jump me, whatever, etc...
I only wanna fill one of out or every 5-10 Limit Orders...
And I'm completely ignoring 90% of the noise out there.
Saying that you have trouble with "HFT robots"...
Is like saying spam prevents you from sending email.
I'm still retaining a solid respect for DeeDeeTwo's knowledge, competence and experience ... but it's sounding more and more like DeeDeeTwo is fundamentally a taker of liquidity ... even in the explicit DeeDeeTwo example given above where one limit order was marketable and one was not, the non-marketable order was "quasi-liquidity-taking" because of its price-aggressiveness relative to the recent price action of the symbol.
If indeed DeeDeeTwo is basically a taker of liquidity then I have had no reason to post to this thread ...
However, if DeeDeeTwo were to post that he/she had been doing a considerable volume of trading using truly passive liquidity provision strategies in illiquid issues for several years covering the period from at least 2006 - 2011, and had been measuring the ROC, average realized spreads, average holding periods, and risk/reward profile using even the simplest and crudest metrics, and if DeeDeeTwo were to still make essentially the same assertions as have been made already in this thread, then I would have to respectfully conclude that DeeDeeTwo is either suffering from a severe mental impairment or is deliberately misrepresenting the facts.
The strictly mathematical, logical, measurable facts are simply too overwhelming for any rational observer to refute or deny... yes, it remains possible to pull money out of the market by providing liquidity ... but the amount of money that can be pulled out, especially relative to the risk taken, has plummeted drastically since 2007 and the situation continues to get worse at a steady, methodical rate ... and the primary culprits are HFTs, runners of predatory algos, and predatory, super-aggressive order internalizers (ok, the only real culprit is the SEC which has allowed and encouraged the present situation where asymmetric market information and "legal front-running" are the SEC-blessed norms).
If indeed DeeDeeTwo is basically a taker of liquidity then I have had no reason to post to this thread ...
However, if DeeDeeTwo were to post that he/she had been doing a considerable volume of trading using truly passive liquidity provision strategies in illiquid issues for several years covering the period from at least 2006 - 2011, and had been measuring the ROC, average realized spreads, average holding periods, and risk/reward profile using even the simplest and crudest metrics, and if DeeDeeTwo were to still make essentially the same assertions as have been made already in this thread, then I would have to respectfully conclude that DeeDeeTwo is either suffering from a severe mental impairment or is deliberately misrepresenting the facts.
The strictly mathematical, logical, measurable facts are simply too overwhelming for any rational observer to refute or deny... yes, it remains possible to pull money out of the market by providing liquidity ... but the amount of money that can be pulled out, especially relative to the risk taken, has plummeted drastically since 2007 and the situation continues to get worse at a steady, methodical rate ... and the primary culprits are HFTs, runners of predatory algos, and predatory, super-aggressive order internalizers (ok, the only real culprit is the SEC which has allowed and encouraged the present situation where asymmetric market information and "legal front-running" are the SEC-blessed norms).
The primary culprit is retail pulling nearly half a trillion dollars out of these markets.