There are obviously better ways to demonstrate how a mean-reversion strategy works during bull and bear markets. It’s probable that a blogger somewhere has covered this issue a lot more scientifically than I’m about to right now. If I had to bet who it was, I’d say Michael Stokes over at MarketSci. Poke around over there and see what you can find.
From what I can tell about mean-reversion strategies, they love volatility. Love it, love it. They thrive on it.
It makes sense then, at least to me, that MR strategies would outperform in bear markets.
My research seems to confirm this.
Below are equity curves from two different mean-reversion systems.
Exhibit A:
Exhibit B:
Each strategy risked 10K every trade without compounding gains or using leverage. Exhibit A used $50.00 on each side of the trade to account for commissions and slippage, and Exhibit B used .01/share for commissions with no allowance for slippage. Not ideal for the sake of comparison, but this was never really intended to be overly scientific in the first place.
Not only does short-term mean reversion vary in bull or bear markets, but it also varies by decade.
Selling rips and buying dips worked well in the last couple of bear markets, and buying dips was great great in the last bull (while arguably it was better to be neutral rather than short the bull rips).
Go back a decade and you get different results.
Go back ANOTHER decade and you get still different results.
So I’m not as much of a fan of mechanical short-term reversion systems over the long haul simply because they don’t show the same robustness that other methods show over the decades.
Actually MarketSci did a four part series on Mean Revision:
Compilation of Articles
and he references you as well…
Yeah, very familiar with his work. Great stuff.
Woodshedder,
If you’re near Orlando there is a tribute to Sedacca tomorrow, its posted at Minyanville
I a dum E.. my hed hertz