As you might be aware, I simulate every game using my Power Rankings every time I run the power rankings so I can generate a predicted season result.
I’ve decided that I might do this more often as an experiment. Tentatively, I will be running this post on Mondays (projecting the games thru Friday) and Friday’s (projecting the Friday, Saturday and Sunday games). I’ll also be keeping track of some of the ongoing results.
Projection #1 (1/02 thru 1/04) went up friday.
The results so far look like:
|Model Straight Win Loss||60.0%|
|Just pick the Home team Win Loss||52.0%|
|Home team average Error||3.3|
|Road Team average Error||-3.3|
Small sample explains the high standard deviation. Home teams outperformed the projection, again small sample, but something to watch for. On a straight win/loss eval, the model outperformed just picking the home team.
Remember this is for science and recreational purposes only of course. I take no responsibility for any possible losses. For wins, I simply ask that you buy a beer if you see me at Sloan.
Here’s the week projected point margins:
A few notes:
- The Hawks and Cavs are facing a rough stretch. I feel more trust in the Hawks.
- Pistons (who are 5-0 and plus 18.2 since releasing Josh Smith) are favored over Atlanta on Friday. I think the result of their game tomorrow at San Antonio will reveal whether the Model’s current love for the Motor City (currently top 10) is warranted.
- Sixers have a significant chance at winning tonight versus the ailing Cavs.