am i missing something or wasnāt ndongo already at GT?
Sometimes the best gets are the ones you keep
1 Like
Maliq Brown will be interesting to watch.
6 Likes
Which is why team recruiting rankings are less informative than they used to be. A program could have year after year of top 10 classes, but if none of them stay more than a year ā¦
3 Likes
Yeah - there should be a 2 year delay on all team recruiting rankings
1 Like
TL:DR - 247 ranks us 8 in the ACC in their latest ACC Power Rankingsā¦
UNC doesnāt have a #2 roster IMO. Cadeau canāt be a 3rd option. Theyāre relying on Davis and Tyson?
Edit: never mind, Ian Jackson and Drake Powell should be really good out the gate.
forget about the fact that he picks us 8th and has never heard of Blissā¦Louisville as third? yeah they got some interesting pieces but cāmon.
1 Like
If we set out to predict a teamās performance next season, you could envision two ends of a spectrum of how to do this:
- 100% team-based: All we know is last seasonās performance (or maybe the last five seasons) and so we expect the good teams to be good again and the bad teams to be bad again.
- 100% player-based: All we know is how each player on the roster performed last season and maybe a projection based on recruit rank for players with no college experience. We have no idea how their current team did last season.
You can see the ways in which both models would be wrong. Model #1 wonāt accurately predict teams that have had a lot of turnover, or coaching changes, etc. Model #2 might have a hard time accounting for anything related to the team context, like coaching effects, adjustment for players changing levels of play.
Good preseason models have elements of both team-level performance and player-level performance incorporated, though I think the current transfer environment might push at their ability to adjust for how the same player might perform differently on a different team. Louisville would look terrible under a Model #1 approach (so terrible the last couple years) but pretty good under a Model #2 approach (lots of players who were good at their previous school).
5 Likes
Probably hard for any of us to get our heads around Louisville being well coachedā¦ but they will be
Yep, I would describe Torvikās rostercast as elements of both. But looking at a few teams you can see the ways in which itās limited ā¦
- it has to project player minutes and usage with mass attrition. Not sure how assumptions are generated, but some are pretty funky. For example, it assumed something like 15 mpg for Dunn this past season lol.
- It really doesnāt know how to treat freshman and redshirts. It has Bliss and ARob both as total non-contributors next year.
Tbh, it would be great to have some structured fan or beat reporter contributions to the model for each team as a way to adjust assumptions. Last season we could have told Bart that Dunn was going to get 25 mpg. This season we could tell him Bliss and ARob are probably going to get minutes. Hard to do at scale across 350+ teams, but thatād be as close as you can get to something really meaningful.
Sports Illustrated used to do this, Luke Winn (works for the Raptors now I think) and Dan Hanner (economist for the government in his day job) put together a player-based model and they did some legwork to source info on rotations. I think Bart has made some adjustments like that before, but more likely closer to the season start as thereās more information.
2 Likes
Oh nice, Iād seen their work on the best offensive coaches that had a really thoughtful and nuanced model, accounting for recruiting strength, attrition and development. Hadnāt seen the pre-season predictive stuff. RIP Sports Illustrated as we knew it.
Anyone know/remember our Torvik/Kenpom preseason O/D predictions last year? I donāt.
Looks like Torvik was 103/39 for 62 overall.
As noted above, Torvikās model seemed to be unaware of RD defensive monster potential.
2 Likes
I think weāre just going to have to accept that weāve lost the benefit of the doubt. Gotta earn back the reputation as an elite program.
6 Likes
Will say the current Torvik cast has Sharma playing but Bliss out of the rotation which is quite funny
4 Likes
I think the system just doesnāt know how to compute redshirts, which ā¦ fair enough.
4 Likes
Oh I mean totally reasonable, but just something to think about when considering the output
2 Likes
2 Likes
I remember this dude. I think HGN at the time was like āoh yeah thatās a no goā. Then he got kicked off of ODU lol.
Dude is a bucket but seems to be a complete clown.
3 Likes