2024-2025 MBB computer projections thread

[Letterkenny]

I agree. Never understood the love, no pun intended, for him. He is okay, but certainly not the alpha guy. Keshon is a much better player than him.

He's a terribly inefficient chucker. He's shot over 40% one season in his 5 year career. Puts up decent scoring number just because he takes a ton of shots. You could give his shot volume to about any players and they'd put up the same PPG.

Career 37.5% FG. 31.9% 3PT. That is terrible. I can't believe none of his coaches ever reigned him back a bit.

 

[Letterkenny]

Computers still love Houston...

View attachment 139746

I'm going to redo my analysis of schedule strengths next week, but it wouldn't surprise me if Houston dodged more landmines than did Iowa State and Kansas when it comes to the conference race.

Are they still living off last years numbers, or are the computers going strictly off this season by now?

 

[VeloClone]

Are they still living off last years numbers, or are the computers going strictly off this season by now?

Exactly what I was thinking. I think most metrics still have some legacy component influencing the numbers...

 

[Sigmapolis]

Are they still living off last years numbers, or are the computers going strictly off this season by now?

Exactly what I was thinking. I think most metrics still have some legacy component influencing the numbers...

I think I remember Torvik's rule of thumb being by about January 1 all the data from last season and the preseason projections will have worked itself out of the system and teams "stand on their own."

Therefore, there might be some legacy data in the system but... not much now.

I'd guess Houston is 90%+ "standing on its own" by this point.

 

[cycloneman003]

He's a terribly inefficient chucker. He's shot over 40% one season in his 5 year career. Puts up decent scoring number just because he takes a ton of shots. You could give his shot volume to about any players and they'd put up the same PPG.

Career 37.5% FG. 31.9% 3PT. That is terrible. I can't believe none of his coaches ever reigned him back a bit.

The absolute worst type of player. His highs are as high as any player in the country, but the lows are brutal and far more frequent. Doesn't stop him from chucking it though!

 

[VeloClone]

I think I remember Torvik's rule of thumb being by about January 1 all the data from last season and the preseason projections will have worked itself out of the system and teams "stand on their own."

Therefore, there might be some legacy data in the system but... not much now.

I'd guess Houston is 90%+ "standing on its own" by this point.

Houston has played 10 and has 2 to go before the end of the year. If that rule of thumb stands in their case then it is likely 10/12 based on this season or about 83.3%.

 

[rosshm16]

Something to remember about (most) "analytics metrics" is they do not explicitly factor in winning and losing, they look at how well you were expected to do. That's why we try to beat all these buy-game teams by 40 points vs. get up by 25 and let the third-stringers have a go.

Houston has blown the doors off all the bad teams they've played except for Troy (only beat them by 20), and their losses are all neutral-floor losses to kenpom #1 by 5, kenpom #9 by 5, and kenpom #42 by 3. Probably only one of those is actually a "loss" from kenpom/torvik perspective.

I'm still not sure why they're ranked above many of the teams they're ranked above right now on kp and torvik though.

 

[VeloClone]

Something to remember about (most) "analytics metrics" is they do not explicitly factor in winning and losing, they look at how well you were expected to do. That's why we try to beat all these buy-game teams by 40 points vs. get up by 25 and let the third-stringers have a go.

Houston has blown the doors off all the bad teams they've played except for Troy (only beat them by 20), and their losses are all neutral-floor losses to kenpom #1 by 5, kenpom #9 by 5, and kenpom #42 by 3. Probably only one of those is actually a "loss" from kenpom/torvik perspective.

I'm still not sure why they're ranked above many of the teams they're ranked above right now on kp and torvik though.

Regarding Ken Pom:

The first thing you should know about this system is that it is designed to be purely predictive. If you’re looking for a system that rates teams on how “good” their season has been, you’ve come to the wrong place.
---Straight from the horse's mouth

Ratings Explanation | The kenpom.com blog

 

[cyfan92]

Computers still love Houston...

View attachment 139746

I'm going to redo my analysis of schedule strengths next week, but it wouldn't surprise me if Houston dodged more landmines than did Iowa State and Kansas when it comes to the conference race.

No trip to Ames helps them A LOT while we have to go to their place again..

 

[Sigmapolis]

No trip to Ames helps them A LOT while we have to go to their place again..

Meanwhile Iowa State and Kansas have to home/home each other.

That's not too far from spotting Houston a one-game lead even before we begin.

Houston and Kansas do have a home/home series. The only "missing" game is UH at Hilton.

That's a small disadvantage for KU and a huge one for ISU.

 

[Sigmapolis]

Houston has played 10 and has 2 to go before the end of the year. If that rule of thumb stands in their case then it is likely 10/12 based on this season or about 83.3%.

I seem to remember somebody (might be Torvik, might be KenPom) uses the last 13 games of the previous season and rolls off the oldest one as the newer ones become available.

So if Houston has ten games under its belt now, then three games from last year still matter.

 

[Sigmapolis]

Something to remember about (most) "analytics metrics" is they do not explicitly factor in winning and losing, they look at how well you were expected to do. That's why we try to beat all these buy-game teams by 40 points vs. get up by 25 and let the third-stringers have a go.

Houston has blown the doors off all the bad teams they've played except for Troy (only beat them by 20), and their losses are all neutral-floor losses to kenpom #1 by 5, kenpom #9 by 5, and kenpom #42 by 3. Probably only one of those is actually a "loss" from kenpom/torvik perspective.

I'm still not sure why they're ranked above many of the teams they're ranked above right now on kp and torvik though.

Change "Houston" to "Iowa State" and roll the clock back twelve months and...

That sounds a lot like where we were at this time last year.

 

[VeloClone]

I seem to remember somebody (might be Torvik, might be KenPom) uses the last 13 games of the previous season and rolls off the oldest one as the newer ones become available.

So if Houston has ten games under its belt now, then three games from last year still matter.

Thanks for the info.

So the number would be more like 76.9%. Even further from the estimated 90%+. So, if the 13 games is correct almost one quarter of Houston's evaluation at this point is based on end of last year's performance.

 

[CychiatricWard]

Computers still love Houston...

View attachment 139746

I'm going to redo my analysis of schedule strengths next week, but it wouldn't surprise me if Houston dodged more landmines than did Iowa State and Kansas when it comes to the conference race.

Not dumb to you but dumb to that model. At this point in the season it really doesn’t make any sense. Not to say they can’t or won’t, I just don’t see it.

 

[Sigmapolis]

Not dumb to you but dumb to that model. At this point in the season it really doesn’t make any sense. Not to say they can’t or won’t, I just don’t see it.

No offense taken.

We’ll see how Houston does in the B12.

 

[alarson]

Not dumb to you but dumb to that model. At this point in the season it really doesn’t make any sense. Not to say they can’t or won’t, I just don’t see it.

I mean, their losses include a couple close ones to top teams. I see why the metric still likes them. Maybe not quite that high but still

 

[CychiatricWard]

I mean, their losses include a couple close ones to top teams. I see why the metric still likes them. Maybe not quite that high but still

True but they still have 3 losses. Has to count for something. They lost to the 56 net/43 kenpom SDSU. They are right next to Iowa on that list. Not a terrible loss by any stretch but again still a loss.

 

[alarson]

True but they still have 3 losses. Has to count for something. They lost to the 56 net/43 kenpom SDSU. They are right next to Iowa on that list. Not a terrible loss by any stretch but again still a loss.

It doesn't though. This is a predictive metric, not a resume based one

 

[SeventhSon]

Stop the count. No need to look any further.

We have found the One True Metric: https://fivestarbracketology.com/2024/12/19/cbb-contenders-pretenders/

 

[CySmurf]

Is that Margaret? Life hasn’t been kind!

Looks a little like 80s Meatloaf.