Gator Country

tebowharvin · 03-17-2013, 04:40 PM

UPDATED FOR REAL BRACKET AND SUNDAY'S KENPOM

ASSUMPTIONS:
-Kenpom's Pythagorean rankings are very close to accurate
-Every site is neutral

Running log5 on the tournament I found that:

Florida has a 24% chance of winning it all, Louisville 20%, Indiana 16%, remaining field's teams each <10%

Florida has a 51% chance of reaching Atlanta, Louisville 48%, Indiana 49%, Gonzaga 28%, Ohio State 23%, Kansas 17%, Duke 19%, remaining field's teams each <15%

tebowharvin · 03-17-2013, 04:41 PM

I'll post a follow-up tonight after the bracket is announced.

UFish · 03-17-2013, 04:43 PM

Did KenPom predict us to win all the games we lost?

**mdgator05** · 03-17-2013, 04:45 PM

Not familiar with the model. Does it have a state dependence/momentum factor built into the model? If so, I would be shocked the numbers are that high. If not, that does seem to be a reasonably large omission.

tebowharvin · 03-17-2013, 04:47 PM

Quote:

Originally Posted by UFish

Did KenPom predict us to win all the games we lost?

You mean his mathematical system? For each individual game yes, but the real question is how easy is it to understand that a team having a 90% probability to individually beat teams A, B, C, D, and E also has only a 60% chance of winning all 5 of those games...?

xenythx · 03-17-2013, 04:47 PM

Yeah I'm pretty sure KenPom's metrics are skewed by us beating the bad teams on our schedule so badly. I'd bet we were predicted to win every game we lost this season by a really high percentage.

tebowharvin · 03-17-2013, 04:49 PM

Quote:

Originally Posted by mdgator05

Not familiar with the model. Does it have a state dependence/momentum factor built into the model? If so, I would be shocked the numbers are that high. If not, that does seem to be a reasonably large omission.

Show me a model that does, first of all.

Second of all, one would think it is reasonable to assume teams have probabilities of getting "hot" or gaining "momentum" relative to their strengths, therefore making a variable based on such "heat" or "momentum" moot.

**gator1986** · 03-17-2013, 04:50 PM

So what's the outlook on our seeding? What is our number? 2 or 3? West or Midwest?

tebowharvin · 03-17-2013, 04:51 PM

Quote:

Originally Posted by xenythx

Yeah I'm pretty sure KenPom's metrics are skewed by us beating the bad teams on our schedule so badly. I'd bet we were predicted to win every game we lost this season by a really high percentage.

He does weight more recent games. Also, if one can beat the crap out of a team, it will. I think it says a lot that few teams have been as CLINICAL as us at some points.

Jonas · 03-17-2013, 04:55 PM

Apparently even KenPom thinks his analysis overrates the Gators. I think we are a good/great team, but our efficiency is skewing some of these statistical models I think. The issue is that we don't play as efficient when games get close I think partially due to our lack of an inside game.

tebowharvin · 03-17-2013, 04:56 PM

Quote:

Originally Posted by gator1986

So what's the outlook on our seeding? What is our number? 2 or 3? West or Midwest?

The bracket I used had us as the 2 in Gonzaga's out west. I would say us losing to an unranked team on a neutral floor with many watching SLIGHTLY hurts us, but these probabilities will not shift very much. There is only so much ADDITIONAL info that you can learn in, what, the 33rd(?) game of the season.

Additionally, while seed number matters marginally it is getting a decent pod or region that counts for more. If we end up in a Duke, Indiana/Louisville, bracket, we will be up to probably 5% points lower.

What to take from this is that we have a great chance at making a run at Atlanta: nearly a coin flip.

tebowharvin · 03-17-2013, 04:59 PM

Quote:

Originally Posted by Jonas

Apparently even KenPom thinks his analysis overrates the Gators. I think we are a good/great team, but our efficiency is skewing some of these statistical models I think. The issue is that we don't play as efficient when games get close I think partially due to our lack of an inside game.

How often do you see a team claw back from the 5-point range because of inside action, and how often do you see them claw back because they just make decently open outside shots?

tebowharvin · 03-17-2013, 05:00 PM

Okay, Joe Lunardi updated the bracket again, but it is not much different, so I am just going to update when the real one is announced. (It takes about 10 minutes to update it, and I would rather do other things in that time.)

Jonas · 03-17-2013, 05:14 PM

Quote:

Originally Posted by tebowharvin

How often do you see a team claw back from the 5-point range because of inside action, and how often do you see them claw back because they just make decently open outside shots?

I don't disagree with what you saying, and I'm far from a basketball expert, but I sometime we just need to get some easy high percentage 2s to stop the other teams runs. Too many times we amass a big lead, but then we got on long scoring droughts. Some of this is because of the lack of a dependable inside game I think.

UFish · 03-17-2013, 05:18 PM

Quote:

Originally Posted by tebowharvin

You mean his mathematical system? For each individual game yes, but the real question is how easy is it to understand that a team having a 90% probability to individually beat teams A, B, C, D, and E also has only a 60% chance of winning all 5 of those games...?

I'm a fan of advanced metrics. What I'm saying though is that in the upcoming tournament everything is one game. So while those odds are great (obviously I'd rather have a high chance of winning than a low chance), KenPom will likely predict us to win every game we play in the tourney unless we run into an Indiana or a Louisville. It doesn't mean anything.

tebowharvin · 03-17-2013, 05:21 PM

Quote:

Originally Posted by Jonas

I don't disagree with what you saying, and I'm far from a basketball expert, but I sometime we just need to get some easy high percentage 2s to stop the other teams runs. Too many times we amass a big lead, but then we got on long scoring droughts. Some of this is because of the lack of a dependable inside game I think.

True. But on the flip side, scoring outside taking good shots and running our normal offense can get hot FAST as we have seen in different games. One must take EVERYTHING into account. Teams only occasionally come back from a medium deficit late, and just because we haven't yet doesn't make us choke artists we have only had 5 chances too.

**fairfaxgator** · 03-17-2013, 05:25 PM

Did these models say that we need a 20 point lead with 4 min left to achieve these percentages?

Cuz...a 10 point lead with 4 min will not cut it...

tebowharvin · 03-17-2013, 05:27 PM

Quote:

Originally Posted by fairfaxgator

Did these models say that we need a 20 point lead with 4 min left to achieve these percentages?

Cuz...a 10 point lead with 4 min will not cut it...

I wonder what percentage of this board either is not black or white or understands probability. We have a really small sample size of being up said amount late, fairfax.

**mdgator05** · 03-17-2013, 05:31 PM

Quote:

Originally Posted by tebowharvin

Show me a model that does, first of all.

Second of all, one would think it is reasonable to assume teams have probabilities of getting "hot" or gaining "momentum" relative to their strengths, therefore making a variable based on such "heat" or "momentum" moot.

Unless I saw some actual stats disproving it, I would be very surprised if there wasn't state dependence in this data.

Think about it like a random walk. So maybe every team is subject to the same normal distribution of playing better or worse on a given day, with a mean at where they played the last game and some standard error set evenly for all teams.

Even in this year, I would be shocked if there wasn't state dependence in this data. It is a pretty crucial component. Basically, if you took your predicted model and found significant bias, in which, to use Florida as an example, they consistently outperformed the projection early in the year versus underperforming the projection later in the year, you would have a pretty good argument that model accuracy could be improved by the inclusion of some measure of state dependence.

Not meaning this as an attack as I appreciate the use of stats to analyze teams above fairly biased observations as given by the talking heads. Just thinking that might help model accuracy if it isn't there. And if it is, that is a pretty surprising result.

Jonas · 03-17-2013, 05:35 PM

Quote:

Originally Posted by tebowharvin

True. But on the flip side, scoring outside taking good shots and running our normal offense can get hot FAST as we have seen in different games. One must take EVERYTHING into account. Teams only occasionally come back from a medium deficit late, and just because we haven't yet doesn't make us choke artists we have only had 5 chances too.

I generally agree with you, but aren't 5 games enough to expect it to average out? I know we came back against Bama, which is great, but I'm slightly worried that when the game gets close at the end, our players might get tight. Maybe it's just variance. Hard to say without really knowing what the players are thinking.