Short-Term Mean-Reversion Becoming Stronger: Wood’s Light Bulb

Recently I’ve shown that daily stock market mean-reversion is becoming stronger at an accelerating pace and attempted to explain why. Understanding this shift is key to trading short-term swing strategies like RSI(2) or daily follow-through.

Woodshedder (from my blogroll) made a very interesting observation in response to my post. He said that short-term mean-reversion isn’t necessarily accelerating in terms of the percentage of days that close in a given direction (i.e. win %), only in terms of the magnitude of the change when they do (i.e. average return), which could just be a function of increasing volatility in the market.

Here is my response:

click to enlarge

The graph above is calculated in a similar fashion as in my previous posts, a six-year rolling average following either an up close (blue) or down close (red), de-trended to remove the influence of bull/bear markets, on the S&P 500 from 1950.

Rather than looking at next-day average returns, I’m looking at the probability that the next day closed up. I’ve de-trended the result by subtracting the probability that the next day closed up for all days over that same six-year period.

I know it's a little complicated to explain in a blog post, so let me give an example. If over a given six-year period the market closed up 55% of the time, but following a close down it closed up 60% of the time, the red line would read 5%. Got it?

For comparison, below is my original graph using the same methodology, but looking at average returns.

click to enlarge

Response to Wood’s Light Bulb

Daily mean-reversion is becoming stronger both in terms of average return (which we’ve previously shown) AND the probability of the next-day’s closing direction (which the first graph above shows).

HOWEVER, the probability of the next-day’s closing direction (win %) is not accelerating at as aggressive a pace as average return (i.e. the slope of the line is shallower). As Wood alluded to, part of the acceleration in daily follow-through as I showed it previously was a result of increasing volatility in the market in and of itself.

Well done Wood.

Comments

[Fred Voetsch:]

This data is like peak earnings; dangerous in the hands of a fool but useful in the hands of a wise investor.

If you are using this data to convince yourself that: dropping economic indicators; dropping earnings; negative earnings on the S&P 500; a housing market that is accelerating downward; rapidly rising unemployment; continued massive deleveraging; and a historically and currently overpriced stock market do not matter, then you are a fool and deserve to lose even more in 2009.

www.generationaldynami...

online.wsj.com/mdc/pub...

www.businesscycle.com/.../

Feb 16 12:51 PM

[MWRjr:]

What this article implies is that momentum strategies such as "the trend is your friend" will not work any more. Many strategies like Valueline's are based on momentum.

Feb 16 04:25 PM

[Prudent Man...:]

Is the "mean" up or down? Like econometrics, this attempt to understand Human Nature by statistical measurement is pure sophistry and makes little logical sense.

How many "quads" have any funds left to invest?

Feb 16 06:25 PM

[Michael S:]

RE to freddyv and PrudentMan:

My standard cut-n-paste to these type of comments:

Your anecdotal shot from the hip is appreciated for what it is...an anecdotal shot from the hip. With all due respect - click through to the blog - checkout the independently-audited real-time returns of our programs (which are in large part built off of these fundamental concepts) and then tell me (preferably empirically or at least thoughtfully) why you've come to the conclusion you have. ms

Feb 17 11:28 AM

[Woodshedder:]

Uh, you guys really do not get it.

I find that encouraging though because I was really worried this edge was going to be eroded quickly.

Guess I was wrong about that!

Feb 23 09:50 PM

[Michael S:]

Wood on SA? What is the world coming to?

Good to see you sir.

Feb 24 02:43 AM