Reply to thread

Message

[QUOTE="Fly Rodder, post: 4912694, member: 2133"]
That's because there is no perfect algorithm. These are all best fit. But KenPom is pretty freaking good, since the "luck" component is so small as to be nearly inconsequential (e.g. for Duke it's like less than 0.4% of their KenPom adjEM).

It's just a measure of departure from the expected record/adjusted efficiency margin. He didn't name it luck because it is a MEASURE of luck, he just didn't have a better name for it.  But it captures some of the "what the hell is happening" in Team's records. It's not sustainable or predictive in any way (coaches/teams aren't luckier than others year over year). It's a measure of error. Basically if you think a team is better (or worse) than their record indicates, take a look at the luck factor to see they're underperforming or overperforming their statistical metrics. Lots of weird things can happen when you have enough data points and we've all seen how basketball games can change on a community college transfer drilling a low-percentage jumper as time expires or a ball bouncing off the rim the wrong way.
[/QUOTE]

Verification

What is a Syracuse fan's favorite color?