Wondering which rating was most accurate?

A guy posting on Reddit is tracking Net, Kenpom, BPI and Sagarin.

So far Net with a decent lead:

Current standings:

NET: 520 (95.8%)
KenPom: 510 (92.5 %)
BPI: 510 (92.5%) BartTorvik: 510 (92.5%)
Sagarin: 500 (88.1%)


I used none of those and currently sit at 510 points. So I’m as smart at all those computers. :joy::joy:

1 Like

After seeing the talking heads ragging on how horrible the NET is because it had a bunch of smaller schools rated too high I filled out a bracket where I picked the highest rated NET team in each game. I have 520 points. Looks like the NET is smarter than the talking heads.


Final results are:

KenPom, BPI, and Bartorvik tied for 1st with 1270 (.975)
NET: 1240 (.965)
Sagarin: 940 (.875)
Chalk: 920 (.863)

Personally, I did better than all of them. However, I had Virginia as my champion so that gave me lots of point from just the championship game.

and even with being that far off, he will have major adjustments for his preseason ratings that have nothing to do with how teams finished the season and 100% to do with moving all the Big6 teams up in his rankings and the rest of the pack down to keep the media happy.

1 Like

I thought AP was very accurate for us. We hung around 11-14 most of the year and made the sweet 16.


Just curious. When you did these brackets for the ratings, did you do the scores based off their pre tourney ratings or post tourney ratings? Because KenPom kept updating his ratings moving teams like Texas Tech up from 9th to 5th making it look like it was a good predictor of games like Texas Tech against Michigan. But not so much.

The guy that did it filled them out w pre tourney ratings.
Keep in mind that since the committee uses these ratings to seed they can have a self fulfilling piece to them. Like how many more rounds would teams like Houston and Wofford make it if they used NET for seeding and none of the others who causes teams like those to get tougher seeds than their net deserved.

1 Like