Hello! I am testing trading strategies and I was wondering if anyone could tell me what is pretty much the minimum workable expectancy? I imagine even if it's positive, if it's too small it's kind of waste to trade it in light of commissions etc. So anyone else that does testing, what is the minimum expectancy you are usually shooting for before you consider a strategy viable? Thanks for any input!
the statistical notion of expectancy might not be exactly what you need. Take it this way: %Win x Avg win > %Loss x Avg Loss where %= percentage of and Avg = average win or average loss, the sum of all your gains/losses divided by their number This general formula is both simple and effective as a starter.
I would also add to the right hand side > other costs (such as data feeds) > opportunity cost of time spent trading (eg. wage income)
Thoughts: 1) it is unusual for models/backtest to yield anything of significant expectancy. Many backtest, few make profits. 2) your worst drawdown during testing may have little to do with real world - it is common for real world DD to be several times that of your testing periods. 3) any expectancy must subtract all trading costs: slippage, commission, data/other fixed costs, reasonable risk, errors in your judgement/transmission/broker/etc., large outliers that may blow you out, risk of bankruptcy, your time, cost of not having your money at a broker not collecting normal returns/interest, etc. 4) In addition, paper trading/demoing does not test real world trading, so your real world performance is likely to be hindered by your emotional/psychological makeup 5) Do not mistake winning% for expectancy. Many newbies so do, and it means little. I would say, after ALL potential costs are subtracted, you want (profit factor) at least 1.3 with reasonable metrics. But everyone will give you a different number.