Intraday Pair Trading/Stat Arb

Quote from Shanb:

Also I read through some of the responses and couldn't even understand what some of the questions were. Talking about the ultra-specific quant and exceuction stuff. Although I realize this is important, Ive seen many people make a killing with relatively simple ideas and execution. You would think this would be arbed away, but experience has taught me this is not the case...specifics do matter though.

I asked some of the pair guys in my office what they were using for some of their work and almost LOL. This guys are pulling in 6-figures a month and they wouldn't know if some of you guys were speaking English!
More...

Haven't really read through the whole thread, but scanning this post caught my eye.

Very often, quants and MBAs from Wharton etc. think that they need a lot of sophisticated analysis to make money in the markets and all simple ideas must have been arbed away by gazillion hedge funds out there.

What these ivy tower types don't realize is that they need to come out of their efficient markets shell and think for a moment that markets might be really inefficient! These kids are smart and if they can only broaden their perspective, they have enough brains to build firms like Rentec and Citadel.

E.g. A book by an Indian on pairs trading talk about applying co-integration to do proper pairs trading. Absolutely stupid idea for an intra-day pairs trade. First, cointegration will work only on longer periods, and there also its not certain whether it will work or not.

Just my 2 cents.
 
Quote from gmst:

Haven't really read through the whole thread, but scanning this post caught my eye.

Very often, quants and MBAs from Wharton etc. think that they need a lot of sophisticated analysis to make money in the markets and all simple ideas must have been arbed away by gazillion hedge funds out there.

What these ivy tower types don't realize is that they need to come out of their efficient markets shell and think for a moment that markets might be really inefficient! These kids are smart and if they can only broaden their perspective, they have enough brains to build firms like Rentec and Citadel.

E.g. A book by an Indian on pairs trading talk about applying co-integration to do proper pairs trading. Absolutely stupid idea for an intra-day pairs trade. First, cointegration will work only on longer periods, and there also its not certain whether it will work or not.

Just my 2 cents.
More...


Six figures a month from equity pairs trading? Holy cow. That's pretty impressive, given that these guys are not using algos.

And yes, co-integration for intraday trading seems almost useless.
 
Pardon my ignorance, but why would you not wish to test cointegration for intraday pairs trading? What would you recommend as a better signal? I read a paper last night that was testing for cointegration on intraday trading intervals and there was significance. Very open and curious to hear about better signals.

Warm regards and happy holidays.
 
Quote from gmst:

Haven't really read through the whole thread, but scanning this post caught my eye.

Very often, quants and MBAs from Wharton etc. think that they need a lot of sophisticated analysis to make money in the markets and all simple ideas must have been arbed away by gazillion hedge funds out there.

What these ivy tower types don't realize is that they need to come out of their efficient markets shell and think for a moment that markets might be really inefficient! These kids are smart and if they can only broaden their perspective, they have enough brains to build firms like Rentec and Citadel.

E.g. A book by an Indian on pairs trading talk about applying co-integration to do proper pairs trading. Absolutely stupid idea for an intra-day pairs trade. First, cointegration will work only on longer periods, and there also its not certain whether it will work or not.

Just my 2 cents.
More...

I agree with much of what you have said. But, co integration is a statistical quality of time series. If a pair is co integrated and stays co integrated it will likely be a profitable pair to trade a mean reversion approach. This is a function of the behavior of the spread.

The flack that this type of approach gets sometimes is that there is a high exposure to tail risk. Cointegration may exist in a spread one day, but that pair that has some news or something else blow out the spread. If you are using something like legs or time stops etc. this could potentially blow you out. Cointegration does not exist into perpetuity, you have to be diligent with filtering pairs and knowing when not to trade. You have to be able to interpret news properly to hedge against this.


The guys that trade pairs and do well in the office understand the products they are trading. The nuance of the news and know when to play the convergence and when to stand aside!
 
Quote from Shanb:

I agree with much of what you have said. But, co integration is a statistical quality of time series. If a pair is co integrated and stays co integrated it will likely be a profitable pair to trade a mean reversion approach. This is a function of the behavior of the spread.

The flack that this type of approach gets sometimes is that there is a high exposure to tail risk. Cointegration may exist in a spread one day, but that pair that has some news or something else blow out the spread. If you are using something like legs or time stops etc. this could potentially blow you out. Cointegration does not exist into perpetuity, you have to be diligent with filtering pairs and knowing when not to trade. You have to be able to interpret news properly to hedge against this.


The guys that trade pairs and do well in the office understand the products they are trading. The nuance of the news and know when to play the convergence and when to stand aside!
More...


Yes, interpretation of news and events and other "soft" factors is pretty critical. At my desk we use algos, but it's basically a grey-box system. I do a lot of adjustments and other discretionary decisions based on what's going on in the market. This includes filtering, not trading certain pairs on a given day, flattening early, etc. Inevitably, for a mean reverting intraday strategy, things get nasty when one stock makes a large move due to a news item.
 
My partner and I have used and tested a variety of statistical methods for automated pairs and would like to exchange ideas with others who may have gone down this road already. If anyone is interested and would like to have an offline discussion, please send me a note. Regards
 
Hi RK. I have nothing to sell. It's just an invitation to have an exchange of ideas that goes beyond what can be handled in a chat room format.
 
You must log in or register to reply here.
Back
Top