I have a fixed sample set to which I apply set of strategy settings and derive series of profit factors. I want to filter out the setting which gives consistently high pf and avoiding some big swings at the same time. So what I did was sum(pfs) / stddeviation(pfs) simile sharpe but using pf instead of returns. It doesn't seem to work. Here's an example: sum(pfs_s)/std(pfs_s) 15.7184805417 sum(pfs)/std(pfs) 37.9844408612 pfs [0.77037037037034339, 1.2999999999999545, 0.74285714285713855, 0.8242424242424049, 0.82432432432427116, 0.90666666666664542, 0.77037037037034339, 0.68539325842692356] pfs_s [0.79999999999996541, 2.166666666666591, 1.2999999999999545, 0.57777777777775752, 1.2999999999999545, 0.79999999999996541, 0.96296296296292927, 2.5999999999999091] pfs series outputs 37.98 but clearly pfs_s looks much better to me with most of pf above 1. What's the best filter I can use in this scenario? Maybe I can use some weighted avg. I would appreciate any suggestions.