The Problem with Algos

[Jrinne]

I use algos (e.g. VWAP). I do not intend to stop because I am not smart enough (or motivated enough) to develop an advanced program for trading (for now anyway).

But I also accept that I may be exploited if my trades are a big percentage of the daily volume. From someone who sounds like he has exploited (and attempted to prevent exploitation):

Every market participant leaves a characteristic footprint in the trading records, and with enough patience, you will find a way to anticipate a competitor’s next move. For example, TWAP algorithms leave a very particular footprint that is used by predatory algorithms to front-run their end-of-day trading (usually hedging) activity (Easley, López de Prado, and O’Hara [2011]).

de Prado, Marcos López. Advances in Financial Machine Learning (Kindle Locations 1343-1345). Wiley. Kindle Edition.

FWIW.

-Jim

[Barn]

There is no way to avoid making a market “footprint”. If you make a trade then you had an influence (large or small). The smaller the time frame you plan on being in that position increases the risk factor of being in that position. That’s why backtested trading statistics are good to look at and see what the avg holding time is for winners / losers and avg gain and exposure time. Those are important risk statistics IMO. Personally I’m using limit GTC orders and I may miss the odd trade but the majority I do get. If it’s at the expense of a bit of drift then I accept that rather than chase a stock and accept slippage. We’ve all seen the effect of slippage on sims.

[yuvaltaylor]

There is no way to avoid making a market “footprint”. If you make a trade then you had an influence (large or small). The smaller the time frame you plan on being in that position increases the risk factor of being in that position. That’s why backtested trading statistics are good to look at and see what the avg holding time is for winners / losers and avg gain and exposure time. Those are important risk statistics IMO. Personally I’m using limit GTC orders and I may miss the odd trade but the majority I do get. If it’s at the expense of a bit of drift then I accept that rather than chase a stock and accept slippage. We’ve all seen the effect of slippage on sims.

I’m curious about two things. Do you set your limit orders before or after the open? And do you set them at the current price or below (for buys) and above (for sells)?

[Jrinne]

Personally I’m using limit GTC orders and I may miss the odd trade but the majority I do get. If it’s at the expense of a bit of drift then I accept that rather than chase a stock and accept slippage. We’ve all seen the effect of slippage on sims.

Cool.

-Jim

[Barn]

I’m curious about two things. Do you set your limit orders before or after the open? And do you set them at the current price or below (for buys) and above (for sells)?

I set them before the open at previous close. I’m sure you’ll agree that most SIM’s have a better outcome IF you can fill at that price. If the stock gaps then I’m fine with changing to the daily open price but I’m pretty cautious about chasing stocks beyond the open price. If I had a lot of turnover on my port then I would be much more aggressive about getting filled. Also I may use a %vol if I’m trading very liquid stocks or possibly VWAP if I’m using very illiquid stocks.

Are saying a large limit order on the books cannot be exploited?

What’s to exploit? I’ve agreed to trade at that price. Machines have no idea what my holding period will be (I don’t even know that). If an algo sees an order on the book then it has the option to out bid, but so does any human. I’m not going to lose sleep wondering if I missed a fill because it may have been an aglo that out bid me. If I wanted the trade then I should have increased my bid. That’s the market and there are just somethings that we need to accept.

With the caveat that LARGE orders are relative to liquidity and in those cases where you are trading >10% of the avg daily liquidity then you need to make a choice. Patience vs slippage …

[Jrinne]

Are saying a large limit order on the books cannot be exploited?

What’s to exploit? I’ve agreed to trade at that price. Machines have no idea what my holding period will be (I don’t even know that). If an algo sees an order on the book then it has the option to out bid, but so does any human. I’m not going to lose sleep wondering if I missed a fill because it may have been an aglo that out bid me. If I wanted the trade then I should have increased my bid. That’s the market and there are just somethings that we need to accept.

With the caveat that LARGE orders are relative to liquidity and in those cases where you are trading >10% of the avg daily liquidity then you need to make a choice. Patience vs slippage …

Looks like you got this!

Best,

-Jim