Mean reverting strategies and volatility

September 27, 2010

(This article was first published on Quantitative thoughts » EN, and kindly contributed to R-bloggers)

Mean reverting strategies are beating on mean reversion of the prices. There are various flavors of mean reverting strategies, but as a proxy I chose RSI(2). You can find many entries on blogosphere about this strategy, but nowadays its popularity dried up.

What made me wondering is that there was an idea about correlation between return of such strategy and market volatility. That means, that during high volatility periods this strategy produces higher return and during low volatility -lower.

To test this idea, I built two strategies. RSI Simple goes long, if RSI(2) indicator is below 10 or it goes short, if RSi(2) is above 90. It closes the open position, then indicator is above/below 70/30 respectively.
RSI Garch follows the same rules as RSI Simple, except one additional rule. By using Garch model I forecast volatility for next day. If  volatility value is greater that 65 % of values during the year (252 days), then order is executed. By adding this filter I can catch most volatile days of the last year.

RSI simple RSI Garch
Profit factor 2.14 2.22
Winning % 0.69 0.7
Profit avg. 197 250
Loss avg. -207 -272
Sharpe 0.93 1.0278

The result of Garch filter are slightly better, but the question remains – is it worth of adding it?

Next thing I tried to look at correlation between volatility of the market and return. Actually, it was not so trivial to implement that. The problem is the duration of the investment, which is not fix in days (it can take between 1 to 19 days to close a position). So, I had two approaches- either measure volatility at the beginning of the investment or to fix max investment time (for example 5 days) and then measure volatility on the last day of investment. Despite the differences the results are similar and I will present former.
As you can see from the graph below there’s correlation between return of RSI(2) strategy and volatility of the last 20 days. However, I would attribute this correlation to the fact, that return (profit or loss) tend to be higher, then volatility is higher. This thought is supported by R^2, which was only 0.1.


You can see return ranges against volatility ranges on the next graph. As in the first example above, I can’t see hard evidence of the correlation.


Volatility filter can slightly improve return of RSI(2) strategy, but it is not significant. In the future I will run the same test on pairs trading strategy (another flavor of mean reverting strategies).

To leave a comment for the author, please follow the link and comment on their blog: Quantitative thoughts » EN. offers daily e-mail updates about R news and tutorials on topics such as: Data science, Big Data, R jobs, visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...

If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Tags: ,

Comments are closed.


Mango solutions

RStudio homepage

Zero Inflated Models and Generalized Linear Mixed Models with R

Quantide: statistical consulting and training




CRC R books series

Six Sigma Online Training

Contact us if you wish to help support R-bloggers, and place your banner here.

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)