donaldkagan

Hi Nate, I'm curious as to what you mean by this statement. Are you playing with semantics? What would you call an exponentially increasing error rate in a deterministic system? Are you just critiquing the lay view of "chaos". I'm quite familiar with Bayesian probability, etc. And although I do have a few colleagues who are antagonistic academics within the real of physics, I have never heard someone claim that chaos is bogus. Have you any links to elucidatory papers on this topic? Perhaps you are a PhD and have been published re: "Chaos is bogus"? I have searched Applied Physics Letters and although there are many papers dealing with chaos, I have not been able to find anything claiming it to be bogus. If I sound glib I do apologize, it's just that I am fascinated by bold propositions, which if accurate would change how I understand the world. Thanks for your time, DK

I recently participated in an online session where an "expert" took a group of would be traders through what he called pair trading. It was a very peculiar experience. The method taught by this expert, was to create a pair of some sort (no correlations/co-integration calculations) and then trade the pair as though it were a single stock, by which I mean to say he would take RSI readings, Stochastic's, and various candle formations and interpret them as buy/sell signals. The most shocking thing about this for me was how eagerly the participants ate it up. I hadn't realized how insulated I was from the general population of retail traders, and how badly my estimation of their awareness of very basic (non quant) principles of arbitrage was. Mean reversion was not mentioned once, I actually have trouble fully understanding how one can trade in that fashion. Needless to say, my list of questions regarding stat arb with futures went unasked.

Thanks for the welcome Urma, I tend to agree with you regarding the paucity of traders who have a solid grasp of stat arb. I have had many a frustrating conversation with professional traders who like to ask "why don't you just trade one side of the pair?" I found your chart to be very interesting, although I'm sure it's a gross simplification and perhaps fundamentally wrong, that indicator seems as though it's a highly granular realtime record of open interest. Have you had any experience with a macro arbitrage strategy with futures? Are there any instrument specific challenges that wouldn't be found in equities?

Certainly when pair trading it is critical to make sure that the instruments are linked in some way beyond mere price action. A pair that I have traded very profitably for a long time is PBR / PBR.A. It's quite clear why these two stocks would show correlation overtime. After I create a basket of stocks from similar industries I further filter for obvious fundamental issues, mergers etc, I then use a combination of correlation and covariance to finalize the selection. I'm not sure why the same process wouldn't work with futures. I need to think about how contract expiration would impact the overall system.

Between different markets. Gold and platinum for instance. Anything that has clear historical correlation and exhibits mean reversion.

I am curious as to whether any members of this forum have used pair trading methodology in the futures market. I have used statistical arbitrage to great success with equities, and my limited backtesting seems to suggest that futures offer an attractive opportunity. Thanks for your time, D.K.