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

05Oct09

20091005.01
*** 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 index swing-traders are doing right now because at this moment in history, it’s the most effective directional trade.

See the July and August reports for details re: how this report is calculated. In a nutshell, we’re using both daily mean-reversion (the likelihood that down days follow up days, and vice-versa) and RSI(2) as simple proxies for all other short-term strategies. I would never recommend anyone actually trade either as I’ve defined them here, but I think they make 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.

What is the state of short-term mean-reversion?

Very much under the weather.

3-month averages for the two metrics I think best capture short-term MR, “Return vs Vol” and “RSI(2) Stretch” (upper right and lower right graphs), are hovering near their lowest levels since short-term mean-reversion became the play du jour around the turn of the century. Both metrics have reached much lower points in individual months, but have rarely failed this consistently for this long.

Back in the real-world, I’ve seen a number of alt.-investment programs, that I know are trading short-term MR in some form, go off the rails the last few months. We’ve been fortunate that, despite leaning heavily on short-term MR in our own programs, our combined performance has been more or less flat.

Do I think short-term MR will stage a comeback? Yes. I think the breakdown over the last few months is tied to the strong protracted rally, but that this slow grind up will come to an end and that short-term MR will again be the play du jour once we get to the other side. Do I know that with any certainty? Absolutely not.

More to follow.

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.



7 Responses to “The State of Short-Term Mean-Reversion: September, 2009”

  1. 1 CarlosR

    Michael,

    I’ve been looking forward to this report, although I have to say it shows pretty much what I expected.

    I think these results raise some interesting questions, such as should the YK strategy rely less on MR? Or, knowing that YK is adaptive, should the adaptive parameters have a shorter time constant, so that they are responding more to what happened in the last month, and less to what happened 1, 5, or even 20 years ago?

    As the strategy designer, I’m sure you wrestled fully with all this in the past — I’m just wondering if it might be worth a re-visit, that’s all. I think most of us would agree that no strategy design should be carved in stone, and I think you feel that way also, based on your previous changes to YK (adding the AMF, for example).

    • 2 MarketSci

      RE to Carlos: all very good comments. My $0.02…

      There have been limited opportunities to observe big fundamental changes in the market (intermediate-term MR in the 1970′s, short-term MR around 2000, and a few other instances in very distant history). Based on those limited observations, when these type of things flip, they tend to do so over a longer period of time, not overnight. YK’s adaptive lookback was based on those limited previous observations. Will the next shift be more sudden? There’s no way of knowing.

      I can say pretty confidently that shortening the adaptive lookback to focus on say the last month or even quarter has been a bad idea historically. Even very strong periods of the market working one way like we’ve seen this decade have been broken up by months and quarters where the observation completely broke down. Taking a bit of a longer lookback has kept us out of all that noise.

      Last thought – the addition of other inputs beyond short-term MR are going to mitigate some of the pain of periods when MR breaks down. YK and Scotty are good examples over the last few months – despite short-term MR being in the toilet, the strategies have managed to scratch close to even.

      So long story short, yes, I’m always very open to changing direction on our strategies. But I think in this particular instance, it’s too early to start shifting away from short-term MR any faster than we’re already doing.

      michael

      • 3 CarlosR

        “I can say pretty confidently that shortening the adaptive lookback to focus on say the last month or even quarter has been a bad idea historically. Even very strong periods of the market working one way like we’ve seen this decade have been broken up by months and quarters where the observation completely broke down. Taking a bit of a longer lookback has kept us out of all that noise.”

        Perfect, that’s what I wanted to hear. Thanks!

  2. 4 jh

    good update .. .I’m sure it will come back

    can you give a monthly update? thanks

  3. Hi Michael,

    Is there any chance you could do a post on what RH “would have done” if it had been using the abnormal market filter?

    I recall that you did something like that for YK after it had a big loss, and I think it would be very interesting as well as instructive to see it for RH.

    Thanks,

    -FM

  4. 6 mike

    The markets basically have 2 states- trending and reversion. So when short term MR is getting killed, its probably a good time to switch the MR system on. When short term MR is doing very well, its time to turn on the trending system. I’d be interested in seeing a test of a “timing the system” idea like this.

    • I have did a test of Mean Reversion to the 150 /200 DMA after the index hits peaks of ~ +40% above 200 DMA for INDIA`s NIFTY index. ( corresponding S&P 500 peak 200 DMA spread is +20%). Also since most global markets are at highest levels above their 200 DMAs , based on that back test, most global markets indicate high probability to revert to their 200 DMA in next 2/3 months. details at my blog


Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s