The State of Short-Term Mean-Reversion: August, 2009

01Sep09

*** CLICK TO ZOOM ***

This is our monthly health check of short-term mean-reversion in the US market.

Why a health check? Because short-term MR (by “short-term”, think for example RSI(2)) is so important to what we swing-traders are doing right now, because at this moment in history, it’s the most effective directional trade.

See the July report for details re: how this report is calculated, but in a nutshell, we’re using daily mean-reversion as a simple proxy for all other short-term strategies. Put another way, we assume a trader went long the S&P 500 at the close if the market closed down, and short if it closed up, and look at the results each month in terms of (a) average daily return, (b) average return relative to volatility (daily SD), and (c) win percentage.

I would of course never recommend anyone actually trade this strategy, but I think it makes a good proxy, because this tendency for the market to retrace very recent gains is exactly why all of these short-term indicators work the way they do.

A New Addition: RSI(2) Stretch

Note that I’ve also added a fourth metric: RSI(2) Stretch.

The problem with looking at short-term MR in a purely binary sense (up days and down days) is that it doesn’t capture the fact that at this moment in history the more the market becomes stretched, the more likely it is to retrace (like a rubber band).

For example, consider four fictional daily returns: -0.3%, -0.3%, -0.3%, +0.5%.

Binary daily mean-reversion would see that sequence as a failure because it would have been equally long on days 2, 3, and 4, resulting in a negative total return. But an indicator based on the stretch of the market (like scaling in/out of RSI(2)) might not have ratcheted up exposure until day 4 (if at all), meaning that that sequence might have been profitable.

The RSI(2) Stretch graph tries to capture this by looking at the return vs. vol. of a strategy that goes long or short based on how close RSI(2) is to its extremes. An RSI(2) reading of 0 would equal 100% long exposure, a reading of 25 = 50% long exposure, 50 = no exposure, 75 = 50% short exposure, 100 = 100% short exposure, etc (and all points in between).

So what is the current state of short-term mean-reversion?

Depending on how you play it, it’s been under the weather the last two months.

July performed well in a binary sense, but very poorly in terms of stretch. Interestingly, I saw this same disparity in my own strategies. YK approaches daily MR in a more binary sense and performed very well last month (+13.0%), while Scotty looks at daily MR in a more stretched sense and performed poorly (-1.4%).

This month, it was binary MR that performed poorly, while stretched MR was neutral. And more importantly, 1-year averages for the metrics that matter most have fallen very close to 5-year averages.

Short-term MR has clearly lost some of its steam relatively to its peak about a year ago. I’m not going to get too excited about these results just yet, but it’ll be very interesting to see if this trend continues over the coming months.

Happy Trading,
ms

 

To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed, or following us on Twitter.

Filed under: State of Short-term MR   |  11 Comments