Regarding tonight’s game: The road to the NCAA Championship, brought to you by ESPN. For the record, I have Kansas winning 71.07 to 70.74. I suppose this is just a statistical way of saying that it’s a toss-up, but I’m sticking to it. Enjoy the game!

### Like this:

Like Loading...

*Related*

Briefly … could you please explain how your numbers, re: Kansas winning 71.07 to 70.74, were determined?

My projections were actually pretty basic: For each team, each statistic, I interpreted each game they played as a sample giving information about each team’s “true” value for that statistic. Then, I added to this sample each team’s opponent’s opponents’ values for that statistic, which I interpreted as giving information about the true value each team allowed. So, for my projection of Kansas’ three-point percent, I used all of Kansas’ games, plus the three point percentages for each of Memphis’ opponents in their games against Memphis. Thus I combined estimates of what Kansas was able to do, and what Memphis has allowed. For the final score projection, I took each team’s offensive and defensive efficiency projection, and multiplied them by my total possessions projection. I’d like to point out that my projection predicted a tie, and we had a tie going into overtime. Also the total sum of points, compared to my projected sum of points, was off by 0.83%. Not bad… in a way, although I did

notcall overtime.Thanks.

From where did you get the opponent’s opponent’s stats?

Opponents stats, and college stats in general are a hassle to find. We really need a better source, with all players and all teams, for and against, for all time in one giant CSV file. But, for this project, I found stats at statsheet.com, which has a lot of info, but is not formatted very well for this kind of thing.