2024-2025 MBB computer projections thread

[Sigmapolis]

I believe they don't even have to be in the same region to both play in Milwaukee, so it's possible they wouldn't see each other until the final.

This reason is why I called them our "roommates" and not our opponents.

If we're at least a #4 seed you avoid playing any of the other Top 16 teams until the second weekend.

Anything under that level is a scrum of the committee trying to make the bracket work.

 

[Cytasticlone]

The main downer of WI being in Milwaukee will be ticket availability right? I don't remember exactly how sessions work but them being in the same session, or set of games, would be brutal. Would the 2/3 seeds get separate sessions or do seeds not matter at that point and it's all however they want to schedule the games?

 

[madguy30]

The main downer of WI being in Milwaukee will be ticket availability right? I don't remember exactly how sessions work but them being in the same session, or set of games, would be brutal. Would the 2/3 seeds get separate sessions or do seeds not matter at that point and it's all however they want to schedule the games?

Yeah that would depend on sessions.

 

[KennyPratt42]

I think I would rather be a #3 seed in a region with Bama/Florida/Tenn/Houston as the #1 seed than be a #2 seed with Auburn/Duke as the #1 seed.

IMO Auburn and Duke are the clear cut best two teams. Teams 3-10 are pretty comparable.

The goal is the Final 4. Avoid Auburn/Duke until you get to San Antonio.

Auburn has separated themselves. Duke has helped their metrics (and perception) a lot by killing teams in quad 2b and quad 3 games. In their 10 quad 1 and quad 2a games they have been good, but I'm not sure they have clearly separated themselves from Houston, Florida, Tennessee, Alabama, and Michigan St.

 

[Sigmapolis]

The main downer of WI being in Milwaukee will be ticket availability right? I don't remember exactly how sessions work but them being in the same session, or set of games, would be brutal. Would the 2/3 seeds get separate sessions or do seeds not matter at that point and it's all however they want to schedule the games?

Having Wisconsin for a roommate in Milwaukee has several disadvantages --

-- if we're playing in the same session... intense competition for tickets and higher prices

-- packed arena... with Wisconsin fans likely rooting against us during our first two games because upsets are fun and they'll probably remember when we upset them a few years ago

-- the smartest Wisconsin fans will also know TJ is stealing good players from them all the time

 

[ForeverIowan]

Auburn has separated themselves. Duke has helped their metrics (and perception) a lot by killing teams in quad 2b and quad 3 games. In their 10 quad 1 and quad 2a games they have been good, but I'm not sure they have clearly separated themselves from Houston, Florida, Tennessee, Alabama, and Michigan St.

Im not so sure we have a decent matchup for Cooper Flagg. It would be interesting to see how we would defend him. Id assume some combination of Jefferson/Heise.

 

[not-the-manager]

interesting data point here, Nate Silver's model appears to currently have ISU outside the top 16. I'm guessing this is mostly due to his weighting of recent games

 

[NiceMarmot]

interesting data point here, Nate Silver's model appears to currently have ISU outside the top 16. I'm guessing this is mostly due to his weighting of recent games

Saw UConn at 13 and thought what the hell? Then saw "Ratings carry over from season to season". Kenpom has UConn at 36, Torvik has them at 37, EvanMiya at 25. How do you have UConn at 13 and not wonder if your model might need to be recalibrated a bit before going live with it Nate?

 

[Cyclonepride]

interesting data point here, Nate Silver's model appears to currently have ISU outside the top 16. I'm guessing this is mostly due to his weighting of recent games

Back to the drawing board lol

 

[not-the-manager]

Saw UConn at 13 and thought what the hell? Then saw "Ratings carry over from season to season". Kenpom has UConn at 36, Torvik has them at 37, EvanMiya at 25. How do you have UConn at 13 and not wonder if your model might need to be recalibrated a bit before going live with it Nate?

He did say it’s a beta version. I saw that ratings factor in past performance, but it was interesting because the top ~11 are about what you’d expect—with a major omission, ISU. It has to be Oklahoma State and Houston weighing them down though, ISU’s performance last season was very good

 

[ClonesFTW]

He did say it’s a beta version. I saw that ratings factor in past performance, but it was interesting because the top ~11 are about what you’d expect—with a major omission, ISU. It has to be Oklahoma State and Houston weighing them down though, ISU’s performance last season was very good

TJ wants no association with anything "beta".

 

[Sigmapolis]

I love me some Nate Silver but Elo Ratings are a stupid way to measure college basketball teams.

The teams change so much between seasons and heck even within seasons.

College basketball isn't like chess where you're going to receive a much higher volume of data over years where the only stat that matters is W/L (unlike in college basketball where we know exactly how close games were unlike in chess where there's no objective standard about how close a match was... maybe an advanced chess engine can make an estimate, but there's no single way like looking at the score).

Using game-level data instead of possession-level data makes your sample size from "reasonable to have a sense of how a team is doing" into "this is a guess from a very small sample size."

I never liked FiveThirtyEight for its sports ratings for the same reason outside of making some broad historical comparisons between "good eras" and "bad eras" for certain teams in various leagues.

Torvik and KenPom and the like are much better looks at this matter.

 

[CloniesForLife]

I love me some Nate Silver but Elo Ratings are a stupid way to measure college basketball teams.

The teams change so much between seasons and heck even within seasons.

College basketball isn't like chess where you're going to receive a much higher volume of data over years where the only stat that matters is W/L (unlike in college basketball where we know exactly how close games were unlike in chess where there's no objective standard about how close a match was... maybe an advanced chess engine can make an estimate, but there's no single way like looking at the score).

Using game-level data instead of possession-level data makes your sample size from "reasonable to have a sense of how a team is doing" into "this is a guess from a very small sample size."

I never liked FiveThirtyEight for its sports ratings for the same reason outside of making some broad historical comparisons between "good eras" and "bad eras" for certain teams in various leagues.

Torvik and KenPom and the like are much better looks at this matter.

Especially in the NIL era.

 

[werdnamanhill]

I don't want to play Wisconsin. They're a whole different beast than the last time we played them in the tourney. It would be a great game though.

 

[cyfan92]

So we are likely going to play BYU again in KC on Thursday? Since both teams are locked into the 4/5 seed lines

 

[ClubCy]

The main downer of WI being in Milwaukee will be ticket availability right? I don't remember exactly how sessions work but them being in the same session, or set of games, would be brutal. Would the 2/3 seeds get separate sessions or do seeds not matter at that point and it's all however they want to schedule the games?

Committee picks the seeds and locations. Tv picks the sessions and what time you play.

 

[Sigmapolis]

Especially in the NIL era.

Yeah. They're basically new teams every year.

I can see why Torvik puts data from last year into his preseason projections, but by the middle of the season those should be removed from the sample and teams have to stand on their own.

 

[not-the-manager]

I love me some Nate Silver but Elo Ratings are a stupid way to measure college basketball teams.

The teams change so much between seasons and heck even within seasons.

College basketball isn't like chess where you're going to receive a much higher volume of data over years where the only stat that matters is W/L (unlike in college basketball where we know exactly how close games were unlike in chess where there's no objective standard about how close a match was... maybe an advanced chess engine can make an estimate, but there's no single way like looking at the score).

Using game-level data instead of possession-level data makes your sample size from "reasonable to have a sense of how a team is doing" into "this is a guess from a very small sample size."

I never liked FiveThirtyEight for its sports ratings for the same reason outside of making some broad historical comparisons between "good eras" and "bad eras" for certain teams in various leagues.

Torvik and KenPom and the like are much better looks at this matter.

I was hoping someone knowledgeable would chime in. I don't even think I knew Silver did sports ratings until today, and figured there's at least one major factor behind his ratings seeming to be an outlier (ISU, UConn, Michigan State and St. John's in top 5). But it shocked me so thought I'd see if anyone knew why

 

[RonBurgundy]

So we are likely going to play BYU again in KC on Thursday? Since both teams are locked into the 4/5 seed lines

There is a scenario for us to win the tiebreaker against Arizona and take #3 seed in Big 12. If we win out and KU can beat Arizona in Phog Allen the last game of the year, we would have same record as Arizona, split our games, but we beat Tech to claim the tiebreaker (I think).

 

[CySmurf]

Yes, it's too much to ask.
ASU isn't beating TTech.

...watch ASU beat Tech now.