I was wondering if anyone is currently doing statistical arbitrage trading at the retail level, since this type of trading is widely used by the hedge funds firms.
I am interested to find out, share and discuss more about this subject as follows:
…data processing and manipulation (i.e. gaps), instruments, timeframes, backtesting methods (i.e. in-sample/out-of-sample), number of observations, cointegration testing methods (i.e. Dickey-Fuller, Johansen), results interpretations & trade signals, strategies (i.e. mean reversion), automated & semi-automated solutions (API’s and third party software) etc.

You'll never be able to do it, there's a reason why it's limited to hedge funds. Unless you got millions hidden under your mattress the edge is too small, the speed is too fast and the money needed is too much. You are competing against HFT's and special situation funds full of PHD's.

PeakGrowth 

You'll never be able to do it, there's a reason why it's limited to hedge funds. Unless you got millions hidden under your mattress the edge is too small, the speed is too fast and the money needed is too much. You are competing against HFT's and special situation funds full of PHD's.

Statistical arbitrage is not limited to hedge funds. Equities HFT does require institutional grade data, infrastructure, and modest capital ($100k+) and monthly volume requirements, but those are certainly within reach of motivated individuals and small trading groups.

I do know a fellow trader who makes his living liquidity providing stat arb on the GLD ETF vs futures. All he is doing is collecting the liquidity premium. None of his alpha comes from knowing which direction either instrument will go, he just gets rebate checks each month from the exchanges. Lime brokerage provides the datafeed/server co-loc. I believe he uses a third-party FIX library from onixs.

This is the NYSE ARCA program he participates in if anyone is interested: (warning PDF link) Futures Incentive Program.

I also did basic stat arb co-integration trading back in 2007-2008 when the market was mad volatile. The strategy made a lot but stopped working not long after the markets calmed down, but truth be told, equities trading sucks. Its like the wild west vs futures. I enjoy trading on CENTRAL regulated futures exchanges, and there are plenty of opportunities in derivatives. Not to mention more preferential tax treatment and reporting.

baywolf 

Statistical arbitrage is not limited to hedge funds. Equities HFT does require institutional grade data, infrastructure, and modest capital ($100k+) and monthly volume requirements, but those are certainly within reach of motivated individuals and small trading groups. I do know a fellow trader who makes his living liquidity providing stat arb on the GLD ETF vs futures. All he is doing is collecting the liquidity premium. None of his alpha comes from knowing which direction either instrument will go, he just gets rebate checks each month from the exchanges. Lime brokerage provides the datafeed/server co-loc. I believe he uses a third-party FIX library from onixs. This is the NYSE ARCA program he participates in if anyone is interested: (warning PDF link) Futures Incentive Program. I also did basic stat arb co-integration trading back in 2007-2008 when the market was mad volatile. The strategy made a lot but stopped working not long after the markets calmed down, but truth be told, equities trading sucks. Its like the wild west vs futures. I enjoy trading on CENTRAL regulated futures exchanges, and there are plenty of opportunities in derivatives. Not to mention more preferential tax treatment and reporting.

"2007-2008" and "stopped working" pretty much summed it up.

You can give it a try now but good luck.

PeakGrowth 

You'll never be able to do it, there's a reason why it's limited to hedge funds. Unless you got millions hidden under your mattress the edge is too small, the speed is too fast and the money needed is too much. You are competing against HFT's and special situation funds full of PHD's.

You can say this for any form of trade you put on really. Tools are tools, trading ideas are trading ideas.

I have heard the newer way to do stat arb is to find your pairs and then bet on that relationship breaking at some point in the future.

You can also just put on a directional trade by only taking one leg and just use stat arb techniques to think about where that instrument is in relation to other instruments.

animalSpiritArb1 

You can say this for any form of trade you put on really. Tools are tools, trading ideas are trading ideas. I have heard the newer way to do stat arb is to find your pairs and then bet on that relationship breaking at some point in the future. You can also just put on a directional trade by only taking one leg and just use stat arb techniques to think about where that instrument is in relation to other instruments.

I would disagree, not all trading ideas are equal.

If you trade like most retails do, which is directional, there are always opportunities because markets will always trend and it will always range in some shape or form. You can find edges by following big money trying to get in and out. i.e. it's always possible from now to the end of time to have a fundamental/technical bias (such as the market going up because of x reason), and then test your hypothesis by buying or selling the market.

However, stat arb is a whole different ball game. Stat arb is dominated by institutional money - with stat arb, you aren't trying to coat tail them, you are trying to beat them in two ways.
1. Find an inefficiency that they haven't found yet. If they haven't found it and you have, I can bet you they won't be far behind and they'll remove your edge soon enough. You are also telling yourself you're smarter and faster than 1,000 PHD's with unlimited computing power, money and speed.
2. If you try to arb something obvious like a futures contract and its corresponding cash market, you are trying to beat them at their bread and butter game - you won't win. You'll be too slow, comms will be too high, and capital requirements will be too high. Stat arb is two way, if you go long cash and short futures you need to pony up for both sides.

Edges come and go, once they're gone, they probably won't come back. Even if you find one, you better start looking for another because you don't know when it'll be gone.

Take options for example, 20 years ago when nobody knew what was going on, options would be mispriced and you can arb the bananas out of it. However, every inefficiency and free lunch is gone because MM's have machines that arb everything and squeeze out a few dollars whenever someone makes a stupid mistake.

You might be able to arb some dual listed stuff, or arb some of the smaller less liquid Asian markets. The less popular the market, the bigger the chance you'll find an edge.

PeakGrowth 

However, stat arb is a whole different ball game. Stat arb is dominated by institutional money - with stat arb, you aren't trying to coat tail them, you are trying to beat them in two ways.

Institutional stat arb has been dead for a decade at least. This is not new stuff we are talking about. It is an old misnomer sales pitch that predated any real computing power. Maybe there was something close to an arbitrage for Morgan Stanley but it is still twisting the meaning of arbitrage. Doing any kind of multi instrument statistical analysis and then trading on that analysis can in some way be considered stat arb. To say you can't find any trades using multi instrument statistical analysis though is absurd unless you believe all trading is a waste of time.

The 1000 smart guys in a room argument applies to all markets and all instruments. We very well may all be fools that should just be indexed and doing something else with our time.

I guess that this is maybe not the right forum to discuss statistical arbitrage
As a matter of fact, I am currently running market neutral strategies...and it is working ;-)
My intention was to discuss various aspects of the cointegration concepts and related trading practices

in my experience, (which is energy derivatives and not equities)...

PeakGrowth 

The less popular the market, the bigger the chance you'll find an edge.

and

animalSpiritArb1 

To say you can't find any trades using multi instrument statistical analysis though is absurd unless you believe all trading is a waste of time.