So you don't trade the correction in big long term trend? And you end-up having a double strengthened trend-following strategy, that's not a mean-reverting strategy, and it may not be better than a standard-alone trend-following strategy even.
Interesting, could you please elaborate on "I prefer to play these as soon as price stalls and with a tight stop; others will average as price continues to move against them, looking for the eventual reversion to mean. "? Thanks!
+2 This technique can be very powerful when applied to day trading as well, look for mean-reversion setups on 3 and 5 minute charts during pullbacks within trends on the 1 hour and daily charts. This takes a lot of discipline and patience though, and I think that's where a lot of people (myslef included) tend to screw it up. The key to a good mean reversion setup is to, as Mr. Rogers would say, wait long enough for the money to just be sitting there in the corner and all you have to do is go pick it up. It's too easy to be casual and lazy and start building a position after the market's made a couple legs, without considering the move into the higher-frame context. If you try to mean-revert at the moment the higher-frame trend is getting taken behind the woodshed (in other words, ceasing to be a trend), you've got a good chance of getting chopped up, seeing your max loss, or a blow-up, depending on your money management strategy.
Because mean reversion strategies require leverage and scale, and then they ultimately turn out to be martingale strategies, and martingale strategies are doomed to failure. See LTCM, Niederhoffer, Amaranth, etc.
I dont think a experienced trader would bad mouth a Mean reversion strategy. Look up Lescor thread at this site as an example. EF
Nice thread. Three things have been put on the table: trends, retraces and reversals. As we review their definitions, differentiating between each becomes the signal generator for the timing to make money all the time each is in play. The best aspect of this differentiation is that each is telegraphed before the event ensues. Some people have mentioned using substitutes for signals; These solutions (substitutes) are less effective and efficient than using leading tells of the signals.
I can't imagine any knowledgeable person badmouthing mean reversion strategies. But I do believe they typically have to be traded in pairs of correlated instruments rather than on individual instruments in order to work well. Otherwise you will continually be beat to death by macroeconomic effects you can't predict. If those pairs have a sound fundamental reason they must remain correlated, then what you have is a long term arbitrage. That's how a bond yield curve arbitrage works, for example. If the pair is correlated without any fundamental reason (it's just an observed fact) then obviously you need to be a little more careful. For example the TED spread (which represents the Eurodollar vs. T-bill pair) usually trades between 25 and 150 basis points. There's a lot of money to be made trading that the spread will converge on 30. But every once in a while it blows up to 300+ basis points, and if you don't have a way to protect yourself in the event that that happens you WILL go broke before it gets around to converging.
This sort of question is why "reversion" strategies on a single instrument are fraught with peril. After all, what is it supposedly reverting to? Itself? Doesn't it always have the same value as itself? Not a very strong philosophical starting point for a trading strategy IMO.