Wagering...I Mean Clairvoyance Update

As I think I have said before, I am working on a system that predicts NBA games against the spread. I am still waiting for a big enough sample size to be completely confident in my system, but I have had some limited success recently.

In its simplest form, my system tracks every possession of every team and measures how the team performs per possession against what the average NBA team does (on both offense and defense). I was unhappy with how the initial system worked, so I tweaked the system to have four different variations -- my original version, a recent performace (last 10 games) version, a weighted version (that weighs more heavily recent performance), and a version that is weighted and incorporates some regression to the mean.

Once I have reached my threshold sample size, I will begin posting picks for recording purposes. For right now, my best performing system is the weighted one with regression to the mean, which has gone 30-18-1 since I have started tracking it.

I have two concerns that I have yet to effectively address. I am trying to incorporate something that accounts for back-to-backs since it seems that teams play worse on the second half of them. Also, when I expect a team to win in a blowout, I have not found an effective way to account for garbage time.

I expect to begin posting picks early in the new year (once all teams have played 30 games).