October 24, 2022

Mean Reversion Check Up 2022

A common question I get is whether mean reversion is still working. My response is I am still trading a mean reversion strategy but the edges seem to get smaller. Over the year I have investigated this. I was asked again recently and wanted to investigate again. Here are the results of my 2022 investigation.

The Rules

Test date range 1/1/2000 to 9/30/2022. I wanted to keep the rules simple. I tested various ways with the 200-day moving average. The reason for this is that some people only trade stocks above the 200, while I like to trade without. In general, this will have more volatility but better returns.

Buy Rules

  • Stock is a member of the Russell 3000
  • 2-period RSI is less than 1
  • Close is greater than $3
  • 21-day moving average of Close*Volume is greater than $1 million
  • [Stock is above 200-day moving average, the stock is below 200-day-moving average, NA]
  • Enter on next open

Sell

  • 2-period RSI is greater than 70

The spreadsheet you can get also has results for the S&P500 stocks.

Results for Above the 200-day Moving Average

Average %p/l

Two things I found interesting. In the current bear market and the one in 2020, the returns are negative. Returns for 2000, 2002, 2008 and 2011 bear markets were all positive. As we can see the edge is getting smaller.

Number of Trades

Here we can see the number of trades fluctuated but basically stayed flat. As expected, bear markets have fewer trades.

% Winners

This trend is also flat. What is interesting this year is the worst for % winners.

Results for Using No 200-day Moving Average

For my mean reversion strategy, I do not use a 200-day moving average.

Average %p/l

This agrees with my personal trading. Edges getting smaller. The 2020 bear market sucked and but this current one has been OK.

Number of Trades

Trade count is staying flat per year.

% Winners

And % winners are also staying most flat too.

Spreadsheet

Fill in the form below to get the spreadsheet with lots more information. Included are results for those below the 200-day moving average and on the S&P500 stocks.

Final Thoughts

These tests confirmed what I see in my trading, that the edges are continuing to get smaller but are still there. The % winners and # of trades are staying about flat.

Will we get to where the edges make it not worth the work? I don’t know.

 

Backtesting platform used: AmiBroker. Data provider: Norgate Data (referral link)

Good quant trading,

Fill in for free spreadsheet:

spreadsheeticon

 
Privacy Policy
Visited 148 times, 3 visit(s) today

Click Here to Leave a Comment Below

Shawn Crichley - October 24, 2022 Reply

I’ve noticed the same thing. Seems to work better in smaller countries where it’s been less arb’d away. Also still seems to work on oversold broad ETFs. Wonder where you think it’s still worth pursuing. SP500 seems very efficient.

    Cesar Alvarez - October 26, 2022 Reply

    From talking to others, I agree that the mean reversion works better in other/smaller countries.The edges are there in the smaller stocks but just not as big as they used to be.

Bing - October 25, 2022 Reply

2020 would have been a great year for most MR systems. I am surprised the sample system you chose had such a big loss. 2022 on the other hand has been a disaster for me even though there has been plenty of volatility

    Cesar Alvarez - October 26, 2022 Reply

    Remember that the strategy I used was very very basic. My own strategy did well in 2020 and is flat this year.

Thomas Musselman - October 26, 2022 Reply

https://www.etfstream.com/features/short-term-momentum-in-stocks-commodities-and-crypto/

short-term momentum dominated the US,
Japanese, and German stock
markets between 1926 and the 1980s. Furthermore, the trends were often similar across
markets, highlighting synchronous global equities trading.

However, short-term momentum stopped working in the US
and Japan in the late 1970s, in
Germany in the 2000s, and in the UK around 2005.

Since then
markets were dominated by mean-reversion, e.g. if the market fell last week, then go long this
week.

Mean reversion started working in Hong Kong in 2000

Sunil S - June 11, 2023 Reply

Captcha not visible

    Cesar Alvarez - June 12, 2023 Reply

    You need to start typing into the boxes for it to show up. It is annoying that they changed how these work. Sorry.

Ami - May 4, 2024 Reply

This might be a spurious relationship, but assuming stocks are sensitive to interest rates makes the following argument plausible. If you overlay your Figure 1 with e.g. ^TNX (yahoo) they appear to move in similar directions. So if this is trully the case, if you use TNX (rather than SPX) to separate different market regimes, you may get a better filter.

Ami - May 4, 2024 Reply

Correction, I meant: If you use MA(TNX, 200) rather than MA(C, 200) (or both) to separate different market regimes, you may get a better filter.

Leave a Reply: