re: How the Playoff Committee uses Data to Place Teams

It’s weird when a metric that tends to correlate positively with a result has as much weight as the actual result itself.

I.e. having a good turnover margin is just as important to indicate how good a team is as whether the team has actually won the game.

I’d be a little nervous letting the average moron play around with these %s

Also the more metrics you use, the less weight one individual metric has. So you really need to start cranking down the %s. Some of these probably have 1:1 correlations too and are likely redundant. It would be like having winning percentage as a metric and then number of wins as another metric. I’d like to see a CART for these metrics