HFT Myths

[JamesVU2000]

It's hard to believe that Getco spends 73 million on co
Location. I guess it's obvious why the exchanges like them

 

[hft]

Quote from Cdntrader:

What are the typical HFT percentage returns on capital invested like?

Interesting thread. Thanks for the effort.

ROC as it's typically measured isn't all that pertinent to much of HFT. You keep your working capital as low as possible as you would in any other business and can achieve pretty high numbers (500%+), but bigger concerns are blowout risk and long-term expenditures, and that raw ROC gets cut down drastically as you pay out more to employees as profit sharing percentages keep increasing.

 

[hft]

Quote from blah12345678:

Just wait - HFT will implode because the rewards will shrivel while the costs skyrocket...

It won't implode; it will reach an equilibrium where only the strongest firms will survive. The industry surrounding internet companies didn't implode, but bad firms went down and growth slowed to a reasonable pace (which is all already happening in recent years in HFT).

 

[stock777]

there are still plenty of trade set ups that work, but i don't like being frontrun on every entry,exit.

if you are trading for small gains, it takes too big a bite out of the p&l.

if you like to swing trade, not so much.

 

[Cdntrader]

Quote from hft:

ROC as it's typically measured isn't all that pertinent to much of HFT.

Ya but what is it?

Are you basing 500% returns on leveraged capital?


I was under the impression that hft was very small % return on very large leveraged capital invested with low risk.

 

[hft]

I was under the impression that hft was very small % return on very large leveraged capital invested with low risk.

500% return on all of the working capital you have in the firm is realistic ($100M net annual profit on $20M in working capital let's say).

Take this simple example to illustrate why that number is misleading. You trade ES@CME only. You hold a maximum intraday position of 100, trade about 5% of average daily volume, and net roughly 1/10 of a tick per round-turn. That comes out to somewhere in the neighborhood $10M/yr in PNL on roughly $200K in margin requirements. Take off X amount of money for colocation expenses and what not and your ROC is still through the roof (~5000%). Use leveraged capital in the equation and you get a little closer to an ROC within an order of magnitude of the metric as typically measured (~100%), but that's just fudgery with numbers.

The more valid metrics are things like blowout/sweep risk (analyzed in a number of non-standard ways), return as a multiple of max drawdown with certain likelihoods over certain time intervals, decay of strategy vs. capital invested in getting up and running, etc.

From a pure ROC standpoint most (good)firms will effectively pay out employees more significantly, in effect diminishing the owners'/shareholders' bottom-line ROC to more reasonable levels.

 

[jb514]

How much, if any of your trading is single product/spreading? Are you mainly trading say the front month of a product in isolation? Or are you mainly spreading a product vs different expirations or different products all together?

 

[Cdntrader]

Quote from hft:



500% return on all of the working capital you have in the firm is realistic ($100M net annual profit on $20M in working capital let's say).

Take this simple example to illustrate why that number is misleading. You trade ES@CME only. You hold a maximum intraday position of 100, trade about 5% of average daily volume, and net roughly 1/10 of a tick per round-turn. That comes out to somewhere in the neighborhood $10M/yr in PNL on roughly $200K in margin requirements. Take off X amount of money for colocation expenses and what not and your ROC is still through the roof (~5000%). Use leveraged capital in the equation and you get a little closer to an ROC within an order of magnitude of the metric as typically measured (~100%), but that's just fudgery with numbers.

The more valid metrics are things like blowout/sweep risk (analyzed in a number of non-standard ways), return as a multiple of max drawdown with certain likelihoods over certain time intervals, decay of strategy vs. capital invested in getting up and running, etc.

From a pure ROC standpoint most (good)firms will effectively pay out employees more significantly, in effect diminishing the owners'/shareholders' bottom-line ROC to more reasonable levels.

Fascinating. Thanks.

I suppose the early hft'ers must have had a real free for all. Akin to early SOES bandits.

 

[hft]

Quote from jb514:

How much, if any of your trading is single product/spreading? Are you mainly trading say the front month of a product in isolation? Or are you mainly spreading a product vs different expirations or different products all together?

It's largely either single product in isolation, single product across multiple exchanges, or closely related products. Other HFT's do more of the multiple product loose spreading, but then you're delving more into less HF and more T.

 

[jb514]

I'm not a futures trader so bear with me.

Would you say your, and other hft market making strategies apply more towards products that trade like barcodes like the 30 year, as opposed to a more whippy product like crude oil.

How often is the position you have on important to you? If you are long are you immediately adjusting the strategy to try and flatten out? Or does the strategy behave the same whether you have plus or minus 25 contracts and only starts to get biased towards flattening out when you have say +- 50 contracts?