Wanted to give a short heads-up in this forum post.
I used this factor function for a while in short and put option systems, as was also found in the ranking system I refered to:
CurQEPSStdDev/abs(CurQEPSMean)
As described here (IMPORTANT: Fix for StdDev of split-sensitive estimates has been released) this previously gave non realistic returns. I still think it is a good factor to use, but the returns that were previously obtained were wrong.
Best,
Victor