General question on how traders approach the optimization of mechanical trading systems using the System testers. Working through various trading systems using system tester, I have been balancing optimizations between the number of positive to loss trade ratio versus the profit per trade to loss ratio. Usually the balance comes down to dropping the number of trades and the profit per trade in order to increase the number of profitable trades to loss trade ratio. For example can optimize a system around:

case A: Profit/Loss trades = 3/2 at a Profit/Loss per trade ratio of 1.0 or

Case B: Profit/Loss trades = 6/9 at a Profit/Loss per trade ratio of 2.27

In either case the net final ratio comes out to 1.5 (1.5*1.3 or .66*2.27) with more trades occuring in case B.

To me it you would want to case A to minimize trade expenses and have to participate in fewer trades rather then having to hit a number of ones in case B trying to get the bigger gains on those trades. the other argument is that useing Case B, you can further improve your good trades by brining in other items that may not be considered, such as chart inspection for divergance, head shoulder patterns etc.

So - Am I missing something really basic in optimizing using the above metrics? Or is this the same quandry that individuals work through. If so what are the typical ratios that make a "good" system. FYI Big trend TK out of the can seem do to ~8/10 at 2.0 ratios.

Normally you want to maximize the net profit (after brokerage expenses) but there are some additional factors you must also keep an eye on while developing a trading system. For instance, watch the drawdown. Will you stay with a system that doesn't allow you to sleep at night? See more at my blog post, Develop a Trading System That You Won’t Second Guess or Abandon.