Yes, I believe this is true for all computer models. All seasons are small sample sizes in the big scheme of things. But the outcry would be too great to leave in preseason weights.
I believe an exception to this is in baseball. Rest of season player projections (and therefore the resultant team projections) are still considering like the previous three seasons.
KenPom explains the much-discussed coaching effect rankings:
tldr:
And kind of implicitly dependent on the validity of whatever talent metric Ken is using
Yeah, it could basically be described as an error term for his talent measure. Heâs choosing to attach that to coaching, but it could just as easily be unrelated. Does the article talk about any persistence in the effect YoY?
Not directly.
Fun chart:
Didnât know 7 would be so much more popular than 6!
Interesting little read by KenPom about scoring going up, up, up.
Ken Pomeroy (the singer, not the stat guy) has a small recurring part in The Lowdown on FX/Hulu. Very funny modern noir for people interested in that kind of stuff.
EvanMiya with a look at how international pros have fared so far this season compared to traditional freshmen and expectations.
Of note for us, De Ridder and GrĂźnloh have been massive successes, while the most of the Euro imports have been mixed bag. Left unsaid is whether that was due to better evaluation, better coaching, or the opportunity we offered with essentially a blank roster. Itâs probably a combo of all three.
Here a chart from the post that has De Ridder and GrĂźnloh as the third and fourth best European player versus preseason expectations if you just want to look at pictures.
Seems like a pretty good success rate (for euros overall) when you consider itâs considered a pretty good frosh class overall
I just did a rough count but seems like 30% expected to be top 100 and 20% actually have been
Secretary General Odom
He still believes in NATO
Gonna go with âskill diffâ in evals for why our Euros have been so good. Should probably do a formal comparison, but I think Nick Kalinowskiâs International player rankings have faired pretty well: International Basketball Recruits Dashboard
Saint-Supery, Gruenlooh, De Ridder, Steinbach, Fru all highly rated in Nickâs model and have done well. Demir, Zougiris, Mantzoukas all misses where Nick thought they would translate and they didnât. Dominguez and Domon did better than Nick expected. Dorn and Zrno were guys Nick correctly didnât rate highly.
Interesting mix of teams here jacking up threes.
https://twitter.com/hoopvision68/status/2014875386539033007?s=46&t=vkjgQUekzGC7z44tIfnIRQ
Definitely makes more sense to weight the best teams in each conference more. Will be interesting to see if the other numbers guys follow suit.
https://twitter.com/evanmiya/status/2017306693844267167?s=46&t=vkjgQUekzGC7z44tIfnIRQ
Yeah, Iâve never been a fan of KenPomâs ranking of the conferences by the median team. Itâs a different meaning of top or best than I think the average fan would consider.
E.g., in 2019, the ACC had an all-timer year. We had 5 teams in the top 15 of KenPem, and three 1-seeds, yet we were the 3rd best league per that way of ranking leagues.
Found a helpful website for tracking WAB: WAB Watch
Uses the T-Rank version, so it is different from the NCAA version (but not crazily different). Described here:
Super interesting. It loves our win at Louisville (no surprise), but a pretty big hit for losing to UNC, which surprised me.
-0.52 for the UNC means that the hypothetical bubble team would be expected to win that game at home 52% of the time. That seems fair-ish.
And just for reference, the interpretation for what positive values for each win means is âprobability of the bubble team losing that game.â
I like the WAB by quadrant. Gives insight into where youâre getting bang for your buck.
The site presentation makes my eyes bleed a little though.
