3/10/24 Model Update

[Aike]

Cats improved very slightly with the big win yesterday.

What’s crazy is that we won at Tennessee without really playing at our best. We went 12/16 assist/turnover which is atrocious, especially for this team. Gives us some stuff to clean up going into postseason.

We also allowed UT to go 17/7 in the same category. This is a bit more concerning, imo. I’ve been calling all year for us to get more disruptive on the defensive end.

We did block 9 shots, which is great. We’ve developed an identity around blocking shots. I’d like to see us generate more steals, and generally do a better job of preventing teams from running their stuff.

Here’s the latest update:

1. UCONN 3.23
2. Houston 2.98
3. Arizona 2.92
4. Auburn 2.82
5. Tennessee 2.68
6. Purdue 2.47
7. Iowa St 2.23
8. Kansas 2.18
9. Alabama 2.14
10. Marquette 2.11
11. Kentucky 1.95
12. BYU 1.83
13. Duke 1.773
14. Michigan St 1.769
15. UNC 1.68
16. Creighton 1.63
17. Texas 1.61
18. Colorado 1.60
19. Gonzaga 1.58
20. TCU 1.52
21. Baylor 1.43
22. Virginia 1.42
23. St Mary’s 1.39
24. FAU 1.34
25. Iowa 1.33


Rest of SEC
26. Mississippi St
29. Florida
53. Mississippi
68. LSU
69. South Carolina
80. ATM
96. Arkansas
98. Missouri
104. Georgia
198. Vanderbilt

 

[Cowtown Cat]

That’s wild that Tennessee only turned it over 7 times and we were in control basically from the jump. I’m telling you, if we play with that type of intensity the rest of the way, we are gonna be a tough out! Our guards are unreal.

 

[podgejeff_]

So a question about the model (and apologies if you've answered this before): does it take recency into consideration or does it weigh the season as a whole?

 

[BBUK]

That’s wild that Tennessee only turned it over 7 times and we were in control basically from the jump. I’m telling you, if we play with that type of intensity the rest of the way, we are gonna be a tough out! Our guards are unreal.

Double Huh?

 

[phunterd]

It would be interesting to see your model only taking into account the last 10 games or so.

 

[LineSkiCat14]

Count me I as someone who would like to see more recent numbers. And id like to see it done this morning. TIA.

 

[nickhorvathsuxazz]

Count me I as someone who would like to see more recent numbers. And id like to see it done this morning. TIA.

Get on it Aike! Sunday services can wait.

 

[UKBB4Ever]

Our biggest problem is defensive rebounding.

That really screws kenpompom type numbers.

I think UT had 15 offensive rebounds yesterday.

That’s 15 possessions that KenPomPom doesn’t count.

Add 15 possessions to UT yesterday and the defensive numbers change completely.

But the bottom line is is doesn’t matter. The game was still played the way it was played.

The nerd numbers have nothing to do with it.

Their star player scored 40 points but the nerd numbers have him as a -8.

Guess Barnes should not have played him? They win the game without his -8, right?

 

[Montana81]

I think it will be hard for the computers to predict this team because we’ve had so many guys in and out because of injury and 2, we’ve let off the gas a few times when the game was already in hand and that’s effecting our metrics.

We do need to get better at closing games out

 

[podgejeff_]

Everything has Kansas and Kentucky as similarly rated teams, and they even have the scoreboard this year.

But I don't think there's a team that exists that would rather face us than Kansas right now.

 

[Aike]

So a question about the model (and apologies if you've answered this before): does it take recency into consideration or does it weigh the season as a whole?

I’ve built them with recency and found that they didn’t predict any better than those without.

I wouldn’t ignore recency completely. Teams who make large moves late are definitely worth tracking, imo.

I just ran a model for the past month - going to post that in a separate thread.

 

[Aike]

I think it will be hard for the computers to predict this team because we’ve had so many guys in and out because of injury and 2, we’ve let off the gas a few times when the game was already in hand and that’s effecting our metrics.

We do need to get better we closing games out

It’s a fair point. Not a great sample size with our full roster.

 

[Aike]

Everything has Kansas and Kentucky as similarly rated teams, and they even have the scoreboard this year.

But I don't think there's a team that exists that would rather face us than Kansas right now.

Their healthy starting 5 is a top 5 team, imo. However, they aren’t healthy.

 

[CatPhight]

@Aike

How far did Florida drop after that Q3 loss?

 

[Aike]

Our biggest problem is defensive rebounding.

That really screws kenpompom type numbers.

I think UT had 15 offensive rebounds yesterday.

That’s 15 possessions that KenPomPom doesn’t count.

Add 15 possessions to UT yesterday and the defensive numbers change completely.

But the bottom line is is doesn’t matter. The game was still played the way it was played.

The nerd numbers have nothing to do with it.

Their star player scored 40 points but the nerd numbers have him as a -8.

Guess Barnes should not have played him? They win the game without his -8, right?

No offense, but this really isn’t the conversation for you. I get it, math is for nerds and only you really understand the game. Cool. I acknowledge your genius.

 

[Aike]

@Aike

How far did Florida drop after that Q3 loss?

Their score dropped a little and they dropped one spot.

 

[Cowtown Cat]

I think it will be hard for the computers to predict this team because we’ve had so many guys in and out because of injury and 2, we’ve let off the gas a few times when the game was already in hand and that’s effecting our metrics.

We do need to get better at closing games out

Yep! Like many are saying, this UK team is kinda the sleeping giant in a sense. Damn the metrics, we are a top 8 team. Easily. Probably top 5. When we were locked in, that is.

 

[Cowtown Cat]

Their healthy starting 5 is a top 5 team, imo. However, they aren’t healthy.

Their lack of depth is killing them. Like I said to begin the season, they lack depth and they lack multiple shooters, playmakers. I said if they take even one injury to a starter, they’re done. Well, here we are! Good!

 

[podgejeff_]

I’ve built them with recency and found that they didn’t predict any better than those without.

I wouldn’t ignore recency completely. Teams who make large moves late are definitely worth tracking, imo.

I just ran a model for the past month - going to post that in a separate thread.

I really asked that question in regards to both us and Kansas, as evidenced by my previous post.

The last month Kansas has lost to Tech, Baylor, BYU, and that beatdown by Houston.

Meanwhile we all know what we've done in the last month.

I can't imagine that your model run on the last month keeps Kansas at 8. I say that without any idea of how it works, of course. But things are shaping up to be a Bill Self special this March for them.

 

[Aike]

I really asked that question in regards to both us and Kansas, as evidenced by my previous post.

The last month Kansas has lost to Tech, Baylor, BYU, and that beatdown by Houston.

Meanwhile we all know what we've done in the last month.

I can't imagine that your model run on the last month keeps Kansas at 8. I say that without any idea of how it works, of course. But things are shaping up to be a Bill Self special this March for them.

Here’s that update for past month only:

Past Month Model

 

[BBUK]

Cats improved very slightly with the big win yesterday.

What’s crazy is that we won at Tennessee without really playing at our best. We went 12/16 assist/turnover which is atrocious, especially for this team. Gives us some stuff to clean up going into postseason.

We also allowed UT to go 17/7 in the same category. This is a bit more concerning, imo. I’ve been calling all year for us to get more disruptive on the defensive end.

We did block 9 shots, which is great. We’ve developed an identity around blocking shots. I’d like to see us generate more steals, and generally do a better job of preventing teams from running their stuff.

Here’s the latest update:

1. UCONN 3.23
2. Houston 2.98
3. Arizona 2.92
4. Auburn 2.82
5. Tennessee 2.68
6. Purdue 2.47
7. Iowa St 2.23
8. Kansas 2.18
9. Alabama 2.14
10. Marquette 2.11
11. Kentucky 1.95
12. BYU 1.83
13. Duke 1.773
14. Michigan St 1.769
15. UNC 1.68
16. Creighton 1.63
17. Texas 1.61
18. Colorado 1.60
19. Gonzaga 1.58
20. TCU 1.52
21. Baylor 1.43
22. Virginia 1.42
23. St Mary’s 1.39
24. FAU 1.34
25. Iowa 1.33


Rest of SEC
26. Mississippi St
29. Florida
53. Mississippi
68. LSU
69. South Carolina
80. ATM
96. Arkansas
98. Missouri
104. Georgia
198. Vanderbilt

I know it's late in this model year but what stat would have caused these Cats to have a higher efficiency number? (If I stated that correctly?) (Must be some glaring deficiency... (I know defensive numbers may have been the cause...)) Thanks

 

[crawfords corner]

Healthy Kentucky model: 2.61
Not fully healthy for most of season Kentucky model: 1.95(not sure how to account for the overlap maybe you can Aike)

 

[Aike]

I know it's late in this model year but what stat would have caused these Cats to have a higher efficiency number? (If I stated that correctly?) (Must be some glaring deficiency... (I know defensive numbers may have been the cause...)) Thanks

2 main things. Need to rebound better and force more turnovers.

 

[Aike]

Healthy Kentucky model: 2.61
Not fully healthy for most of season Kentucky model: 1.95(not sure how to account for the overlap maybe you can Aike)

That’s probably close enough. Remember that we were all the way up to 2nd overall after we played Miami, which was before we got our bigs back.

I don’t think it has as much to do with overall health, as just general tinkering.

To be honest, I still don’t think we’ve played our best basketball.