Criteria for portfolio composition
1. Choose strategies with good performance which you like. Remember, a spoonful of tar can spoil a whole barrel of honey. The more strategies you use, the higher the diversification, and the lower the profitability, max. drawdown, and volatility of your portfolio.
The more strategies you use, the less you are at risk in the long run, and the better you can trade in different market regimes.
2. Allocate a certain trading lot size for each strategy (not for a set-file!). For example, 0.04lots on every $1000. Divide this lot size between set-files based on the same strategy because they might have a very high correlation, especially if they use the same pair. For example, if you’re going to use 4 presets of ‘Daily_levels‘ strategy, then the lot size for each preset should be 0.01lots on every $1000 (0.04/4 = 0.01).
3. Don’t give a significant trading advantage (for example, a larger lot size) to any strategy because absolutely any strategy will sooner or later fail. If the failed strategy has a larger lot size it will be difficult for others to cover losses.
4. The ratio of expected income of M/MR strategies should be at least 70%/30%. For example, the expected profit of M in a portfolio is 100, MR is 50. The ratio = 67%/33%.
5. Your portfolio should withstand a long losing streak. No doubt such a streak will occur in the future. So multiply your estimated MDD by 2-3 times to cover the survivor bias.
Reliable backtest results
Good backtests require accurate historical data, but most brokers don’t provide their own historical data. They usually provide poor quality data from Metaquotes:
I know two ways to test with high quality.
1) The first way is to use the TDS2 software with the Dukascopy tick data and real variable spreads (Unfortunately, TDS2 is not for free).
2) The second is to use the Alpari MT4 terminal (free). Alpari provides some of the best historical data. They close daily charts at 5 pm New York time. Just open a demo account through MT4 and download their data via ‘History Center‘.