均值回归策略 (mean reversion strategy)

mean reversion system

hi gurus, this is a strategy I have made last year and would like to discuss with you guys about possible improvements.

The logic behind the system is simply mean reversion, say, if price becomes overbought it’s likely to fall back.

I trade on hourly timeframe and use SMA. The system calculates average swing length in past several months. Swing length here refers to the price difference between upcross MA and downcross MA. When ongoing swing length is greater than 5 times average swing length, I consider the price enter overbought area and sell short.

If price continues go up, I sell short every time a new bullish candle formed, so position get stacked all the way to the peak of this swing and every new trade always higher than previous one. When price finally close below MA, I close out all trades in the “basket” no matter in profit or not.

I backtested 20+ pairs and found some pairs got okay results while others don’t. the best pair is EU, below is backtest from 2010 to present using tickdata suite.



A small live account forward test as below.



Yes it got blown in November last year because I was too optimistic and increased lots size unwisely. quite stupid. It’s a good lesson I have learned.. after that I deposited some small money and restarted it again.

My trading experience is barely more than 2 years, and I’m not quite aware of many good techniques about risk control and money management. I use moving average as natural stoploss and exit, and increase lots size based on equity growth. If you think this system has some potential please give 2 cents about how to make it more efficient?
Your opinions are more than welcome.