Welcome!

By registering with us, you'll be able to discuss, share and private message with other members of our community.

SignUp Now!

Pomeroy

I added some stuff to the 2pt data, including Kentucky's data for this season and free throw data. Most importantly, I included a new stat in the rightmost column that incorporates free throws, since I think it's silly to analyze shot selection defense without considering fouls.

The new stat is the proportion of points the opponent gets from free throws and layups/dunks. The goal of all of this is to be able to analyze the "true" performance of a defense, so that we can predict how it will perform in the future. With opponent layups/dunks and free throws being the worst results for a defensive possession and more skill-based than 3pt performance, which Pomeroy and others have shown is more random/luck-based, I believe the layup+dunk+free throw proportion of total points is a useful predictive stat. This is obviously not perfect, since some free throws result from bad shots and 3-pointers, but I believe the vast majority of free throws result from good shots (which are not counted as shots if missed) and a disproportionately high amount of free throws result from layups/dunk attempts. Regardless of whether any of this correct, it is inarguable that opponent free throws are generally the worst outcome of any defensive possession, so they should not be ignored.

A few notable observations are Duke's lack of fouling and Kentucky's high likelihood of overall defensive regression. Their opponent 3pt, midrange and FT accuracy have all been awful. Their opponent 3pt accuracy is lower than Duke's, which everyone cited as the reason Duke's defensive efficiency would regress. I think their opponent midrange accuracy is not based too much on luck, but I do think it will rise - 22.5% is absurd. Overall, Duke's defense looks to be in good shape, and it's clear that last year's defense was simply atrocious (as opposed to being weighed down by bad luck)

nH9ZgF0.jpg
 
Wiscy result bumped our D down 0.1, and our O up 0.3, so net positive there.

Inching closer and closer to Kentucky
 
Wisconsin D skyrocketed too. I thought they were 20th after the Duke game.
 
Lol at the gap between our offense and the #2 offense.

Our defense is fine, especially given our offensive dominance, but I can't help but fantasize about what it could be if K showed the same forward-thinking and willingness to adapt on that end of the floor as he does on offense.
 
If all you did to Kentucky's numbers was make their opponent FT accuracy equal to Duke's opponent FT accuracy, Kentucky's adjusted defensive efficiency would rise by 3.1 points on Kenpom, to 86.2. They would drop to 4th in defense and significantly below Duke overall.

So yes, opponent FT accuracy has mattered so far.
 
SeanMayTriedToEatMe said:
If all you did to Kentucky's numbers was make their opponent FT accuracy equal to Duke's opponent FT accuracy, Kentucky's adjusted defensive efficiency would rise by 3.1 points on Kenpom, to 86.2. They would drop to 4th in defense and significantly below Duke overall.

So yes, opponent FT accuracy has mattered so far.
Click to expand...

The mere prospect of knowing you have to fight their trees for an offensive rebound if you miss has rattled their opponents into shooting 10% worse from the line. Another genius move by Calipari.


Where would Duke's be if our FT numbers were the same as UK?
 
DurhamSon said:
SeanMayTriedToEatMe said:
If all you did to Kentucky's numbers was make their opponent FT accuracy equal to Duke's opponent FT accuracy, Kentucky's adjusted defensive efficiency would rise by 3.1 points on Kenpom, to 86.2. They would drop to 4th in defense and significantly below Duke overall.

So yes, opponent FT accuracy has mattered so far.
Click to expand...

The mere prospect of knowing you have to fight their trees for an offensive rebound if you miss has rattled their opponents into shooting 10% worse from the line. Another genius move by Calipari.


Where would Duke's be if our FT numbers were the same as UK?
Click to expand...

88.2ish

Good for the #6 defense.
 
Damnit, 75% allowed could potentially not regress to the mean. We could just be the unlucky ones who get fucked at the end of the bell-curve this year: Notre Dame last year finished the season with 75% allowed


Average in D1 was 70%
 
But that 75% has 0 predictive value. So it doesn't really matter.
 
So our defense is probably like #11 or so if our opponents shoot the average 70%?

I am going to go ahead and assume that's our true defensive rating, since we obviously can't control how well opponents shoot FTs, and they're just as likely to have a subpar performance from the line in a critical moment as they are an above average performance.
 
rome8180 said:
So our defense is probably like #11 or so if our opponents shoot the average 70%?

I am going to go ahead and assume that's our true defensive rating, since we obviously can't control how well opponents shoot FTs, and they're just as likely to have a subpar performance from the line in a critical moment as they are an above average performance.
Click to expand...

D1 average is about 68% this season. Which would still result in a #6 defense.
 
I didn't realize the average was that low. Sweet. From now on, we are the #6 defense in my mind. Fuck Pomeroy.

#1 offense by a huge margin and #6 defense seems like a title team. Just need UK to get beat by some plucky mid-major who plays a lot of zone and bombs threes.
 
Honestly i'm not sure why Pomeroy doesn't adjust his algorithm to account for FT fluctuation when FT defense literally has nothing to do with how well a team defends from a skill/technical perspective. He should just at the end of games use (team's season FT average*attempts) to get the score numbers to calculate efficiency, instead of using actual game numbers. So if Crap-U shoots 17-20 from the line, but 65% on the season, subtract 4 points from the final efficiency calculation.
 
But, we shouldn't adjust based on the D1 average, we should adjust based on the actual players that shot the FTs.
 
Some jackass will argue with him that the Grateful Red or Izzone is actually significantly influential, though
 
If I have the time, I'll work on the Z-rating again during Christmas break. I will only credit teams for expected FT points instead of actual FT points. This will effect game to game Oeff, but won't change the total. Defense on the other hand will change from game to game and total.
 
That would be an insane amount of work.

Duke doesn't foul a lot, but I wonder if they are fouling more guards (i.e. - players who tend to shoot a higher percentage from the line)? I don't really know why that would be the case, but it could affect the number.
 
Someone should email Ken about this. I doubt he overhauls anything but it'd be interesting to hear his thoughts on the matter.


pomeroy = pumice
 
You must log in or register to reply here.

Chat users

  • No one is chatting at the moment.

Chat rooms

  • General chit-chat 0

Forum statistics

Threads
1,067
Messages
424,890
Members
624
Latest member
Bluegrass Blue Devil
Back
Top Bottom