Weekly mean reversion with Dow 30 stocks performed well in 2018 despite the stock market correction. The strategy, which is based on a formula from a text on probability theory, returned 13.5%, while its more conservative version that offers additional protection returned 9.3%, both after commissions.
Although mean-reversion in daily timeframe suffered from whipsaw, in the weekly timeframe the results were excellent given the market woes last year. Some traders did not see the whipsaw coming and continued to trade mean-reversion algos in daily timeframe only to see losses accumulating. Instead we decided early on to employ mean-reversion in weekly timeframe for avoiding daily fluctuations induced mainly by algo trading. This was a good move judging from the results.
The mean-reversion strategy we employ for Dow 30 stocks in weekly timeframe was analyzed in this article. Here we only present backtesting results for the last three years.
Market: Stocks in current composition of DJIA
Time-frame: Weekly (adjusted data)
Strategy type: Mean-reversion, long-only
Score: Based on PSI5 (a proprietary algo)
Backtest period: 01/08/2016 – 12/28/2018
Maximum open positions: 20
Commission per share: $0.01
Position size per stock: Available equity/20
Trade entry: Open of next weekly bar after entry signal (no look-ahead bias)
Trade exit: Open of next weekly bar after exit signal (no look-ahead bias)
Below are the returns for the strategy without (left) and with the whipsaw filter.
Last year the strategy gained 13.5% without the filter and 9.3% with the filter. In our weekly signals reports we employ the strategy with the filter.
Below are the two equity curves, left without filter and right with the filter.
Underwater equity curves are shown below, without and with the filter.
The table below summarizes the result for 2018 and compares to SPY total return.
|Without filter||With filter||SPY TR|
If you have any questions or comments, happy to connect on Twitter: @priceactionlab
Charting and backtesting program: Amibroker