Avg win 339 Avg loss 469 Wins 78% Losses 22% # Wins 66 # Losses 19 The system has a positive expectancy: (339 * 78% + -469 * 22%)/469 = 0.34 Profit Factor = (66 * 339)/(19 * 469) = 2.5 P/L Ratio = 339/469 = 0.7 Profit Factor * P/L Ratio = 1.75 Profit Factor * P/L Ratio * Win % = 1.37 One of the MM books I've read says not to trade a system with a expectancy below 0.6 while another says that Profit Factor * P/L Ratio * Win % is the most important and it should be 1.2+. I'm thinking this may be a marginal system but would like to hear other traders' thoughts. Thanks (This is a monthly system which is why the trade counts aren't higher)
1) When will you feel confident enough to trade "it" with real money? 2) Systems characterized by average profits being less than the average loss generally require more effort because you are in a race against time to avoid a large loss that WILL erase a lot of profit, provided that you adhere to your exit strategy. 3) You have to keep in mind that you may have tested your system during an "optimal interval". By the time you begin trading, the market environment can change and produce different statistics, to your detriment.
I'd say that PF=2.4 and %Win=78% are encouraging. the biggest missing piece of data, imo, is Drawdown and its relation to profit. I'd like to see Profit/DD > 3 but preferably higher and (avg annual profit)/DD > 1.
The biggest concern I have is with the number of trades. If it is an intraday system - and I think it is not - then it is not acceptable, way too low significance. If it is a position system then it is possibly not statistically significant. I do not think it is a trend following system because of hit rate.
Have you included transaction costs/friction in the numbers? What time frame is this on? 66 trades seems like a very narrow sample size.
I cant even begin to look at the numbers with that small of sample size of trades. That means absolutely nothing. How did you arrive at these results? Any sort of optimization and I guarantee you will lose money. Get more markets to get more data. Period.
How true! If your strategy's long-term "% winners" really is 78%, then you may not see say 5 losers in a row for on average 200 - 300 trades or so ... plenty of time to build up a nice profit before the first big drawdown comes along ... However, if your stats resulted from testing during an "optimal interval", and your strategy now reverts to a %-winners closer to 50%, you'll more likely get 5 losers (wiping out 8 winners) in a row every 50 trades or so, and lots more 2, 3, and 4 losers in a row in between Perhaps a key to trading a strategy like this to have also a "strategy management methodology" that allows you to start now (in case it does persist at generating 78% winners), but alerts you quickly to stop trading if the real stats deviate significantly from what you measured in testing ...? Or is this heresy?