There seems to be two very different opinions about stat arb. Some people (usually the practicioners of it) die by it and use it religiously. Others, however, believe that it is an outdated methodology and it cannot realize the returns it use to ten years ago mostly because the arena is so competitive. So what is the future of stat arb?

It is not uncommon to find two stat arb traders to be on opposite sides of a trade, because they are trading different time horizons. Also, there are many other players in the market who play by completely different rules (fundementalists, techies, etc). As long as this is the case, stat arb shouldn't die off completely. It is true that competition is becoming really fierce, and there are too many institutions now who share profits. The natural effect would be less stellar returns in the future. IMHO, stat arb will become very similar to classical arb in the future: only the biggest and the best institutions will be able to extract reasonable profits from it due to the competition.

Stat arb is just a name given to any statistically profitable trading strategy - i.e. all profitable trading strategies. Therefore if stat arb is dead then trading is dead. I think we are a long way from this yet.

That is completely wrong Plus. Statistical arbitrage is in fact a very specific trading strategy. There are many many different types of strategies that can be "statistically" profitable that are not statistical arbitrage.

of course I am aware of the most common definition.

QuoteOriginally posted by: plusStat arb is just a name given to any statistically profitable trading strategy - i.e. all profitable trading strategies. Therefore if stat arb is dead then trading is dead. I think we are a long way from this yet.stat arb refers to shorter term equity strategies of holding periods 2-20 days, all market neutral. the usual strategy commonly referred to as stat arb by hedge fund & investment management community is "very specifically" (as rmb623 mentioned) stock vs its replicating basket (and occasionally basket vs a replicating basket). a few stat arb firms may add a some other strategies on the fringes like index arb, earnings announcement, etc but by and large all strategies conform to a stock vs another equivalent and over a short time-frame. so to that extent i completely agree with rmb623 that you don't know anything on stat arb (otherwise you wouldn't have called it as "any statistically profitable strategy").

AGAIN I am aware of the standard definition.

stat arb is just a fancy name for hocus pocus. they used to call it chartism; then technical analysis; when tech analysis got debunked, they changed the name to stat arb. what will they call it next?

QuoteOriginally posted by: KackToodlesstat arb is just a fancy name for hocus pocus. they used to call it chartism; then technical analysis; when tech analysis got debunked, they changed the name to stat arb. what will they call it next? I disagree. I think if you have a good statarb strategy there's no way you can actually see with your own eyes a buy/sell signal. So that's totally different than technical BS. However you're right in the sense that a few years ago you would just chart a ratio or spread with its rolling standard deviation that would give you a buy/sell signal if it was crossed... I don't think anybody that makes money right now does that anymore.