Does your system have a quality knob? I've been following along and I don't recall seeing any thresholding info other than occasional "we have a large edge here" comments. How does performance look if you filter it down to only the "best" (relative to given book odds?) odds?
I use human judgement to decide which bets to make and haven't been marking them by quality/edge explicitly. When I do training runs I choose a required edge, since one must. I do list my fair value for every game, so one could go back and look at which ones have bigger/smaller edge, and certainly one could do a 'trim off anything with tiny edge especially totals' or something, but I wanted to avoid multiple hypothesis testing.
Even though I ended up losing 62.6% of my tiny stake, I enjoyed doing it and that's what gambling's about I guess. Thanks!
Does your system have a quality knob? I've been following along and I don't recall seeing any thresholding info other than occasional "we have a large edge here" comments. How does performance look if you filter it down to only the "best" (relative to given book odds?) odds?
I use human judgement to decide which bets to make and haven't been marking them by quality/edge explicitly. When I do training runs I choose a required edge, since one must. I do list my fair value for every game, so one could go back and look at which ones have bigger/smaller edge, and certainly one could do a 'trim off anything with tiny edge especially totals' or something, but I wanted to avoid multiple hypothesis testing.