Defense

[pnkranger]

So...I was the primary detractor of our defense early in the season, and it was BAD. The last four games we have really shown a toughness on defense. Our doubles are better, quicker, our traps are creating offense, and we are rotating in the halfcourt.

Still a ways to go...but we are improving.

In the last four games, our defense efficiency has improved from 190th in NCAA to 120th. This late in the season that is a HUGE jump.

 

[jerry]

Still not where it needs to be though. Especially on the perimeter.

 

[NickZepp]

OSU is going to score some points because of who they are. But their shooting percentage and shot selection last night was poor. A lot of that had to do with our defense.

 

[garyeb2]

So...I was the primary detractor of our defense early in the season, and it was BAD. The last four games we have really shown a toughness on defense. Our doubles are better, quicker, our traps are creating offense, and we are rotating in the halfcourt.

Still a ways to go...but we are improving.

In the last four games, our defense efficiency has improved from 190th in NCAA to 120th. This late in the season that is a HUGE jump.

Doesn't that huge jump mean that the 1st number wasn't any good. The defense hasn't got much better. Woodard still can't guard anyone, Buddy won't, Cam is still busting weak side help rotations, Booker is still getting picked off by every screener, we still can't protect the rim, and Spangler will still let some of them go instead of fouling.

Even if those numbers had any validity, the shift would more likely represent a regression toward the mean due to a larger sample size. We have never had a "putrid or "terrible" defense. Our defense from the get go was decent enough considering our built in limitations.

While I would agree that the defense is improving, it was never as bad as the 190 number and is much better than the 120.

 

[ChongoChingi]

Doesn't that huge jump mean that the 1st number wasn't any good. The defense hasn't got much better. Woodard still can't guard anyone, Buddy won't, Cam is still busting weak side help rotations, Booker is still getting picked off by every screener, we still can't protect the rim, and Spangler will still let some of them go instead of fouling.

Even if those numbers had any validity, the shift would more likely represent a regression toward the mean due to a larger sample size. We have never had a "putrid or "terrible" defense. Our defense from the get go was decent enough considering our built in limitations.

While I would agree that the defense is improving, it was never as bad as the 190 number and is much better than the 120.

Couldn't we argue that the competition has gotten better and that going from 190 to 120 against those teams is proof that the defense has improved significantly?

 

[coolm]

those are good points gary. we probably were never 190, you're right. But I do think we're improving. It just seems to be a really slooooooow progression and that's part of being so young across the board and not necessarily being a team of recruited defenders. I happen to like the skills coach gathered this group to promote. It's fun ball again.

 

[pnkranger]

Gary, it doesn't mean the data is bad because the data is a snapshot based on how we performed to this point and not a predictive tool to say how well we'll be playing defense in March. Comparatively we were playing defense at the 190 level, regardless of whether we could have been playing at the 60 level with more effort and focus.

To this point, across the entire season, our defense is probably in the 100 to 120 range in overall performance. The last four games our defense was a top 60 defense, which accounts for the improvement in our efficiency numbers.

To say that a defense who allowed UTA to score 89 points at home or 102 points to La Tech wasn't pretty terrible is being too nice.

To your point though, it would be silly to look at the number (120) and assume that we are the 120th best defense right now. We're much better than that. If you really want to use the data effectively, you can break our season down by segments and see the current state (last four games, say) vs each four game segment of our season to see the improvement. Or compare our defense against top 100 opponents vs 100+ opponents to predict how well we'll play vs quality of opponents.

Or, you can see how game by game temp impacts our defensive efficiency. Are we more efficient when the game is up tempo or when possessions are limited.

As a flat number 120 doesn't mean much...and I didn't use it to signify anything more than the fact that we've improved tremendously over the last four games.

 

[pnkranger]

Couldn't we argue that the competition has gotten better and that going from 190 to 120 against those teams is proof that the defense has improved significantly?

Sort of...but the numbers are efficiency as filtered by quality of opponent. So strength of competition is built in.

 

[garyeb2]

Couldn't we argue that the competition has gotten better and that going from 190 to 120 against those teams is proof that the defense has improved significantly?

We are agreeing that the defense has gotten better result lately. What I'm saying that the numbers being used to support one position or another are not valid for that purpose.

 

[SoonerBounce13]

getting better.

 

[pnkranger]

We are agreeing that the defense has gotten better result lately. What I'm saying that the numbers being used to support one position or another are not valid for that purpose.

But they are valid because the only thing I'm using the numbers to say is that we have improved tremendously. Not arguing in any way that we are currently the 120th best defense.

Also, you continually dismiss these numbers, which are widely regarded in college basketball circles, but don't often give any reasoning ... I would be interested to know why ken Pom, sagarin, espn's BPI are not appropriate tools for discussing basketball?

 

[AdaSooner]

those are good points gary. we probably were never 190, you're right. But I do think we're improving. It just seems to be a really slooooooow progression and that's part of being so young across the board and not necessarily being a team of recruited defenders. I happen to like the skills coach gathered this group to promote. It's fun ball again.

^
This.

Our defense is definitely better than it was several weeks ago when we had a lengthy discussion about it. This team is capable of playing lockdown D in spurts when they needs stops. Doesn't mean they always succeed, just that we have seen it more than once in the last few games. I'm encouraged by the improvements we have seen.

 

[stormspencer]

But they are valid because the only thing I'm using the numbers to say is that we have improved tremendously. Not arguing in any way that we are currently the 120th best defense.

Also, you continually dismiss these numbers, which are widely regarded in college basketball circles, but don't often give any reasoning ... I would be interested to know why ken Pom, sagarin, espn's BPI are not appropriate tools for discussing basketball?

It's not only the numbers that signify our improvement. Our eyes tell us that the defense is improving. As Gary pointed out above, we still have deficiencies based upon our limitations. A month ago, we couldn't defend a ball screen to save our lives. We've improved in that aspect a little bit. We are much improved at switching off defenders and our rotations have gotten better as well. And the defensive efficiency numbers bear this out. And as PNK said, these metrics (Kenpom, BPI etc.) are held in high regard amongst the basketball community. IMO, Bilas is the best CBB analyst around. And he continually quotes these metrics as a measure of a team's ability (or inability) in a particular facet of the game.

To dismiss these metrics (that actually semi-prove OU's improvement) is silly and short-sighted. We have improved significantly on defense since the conference season started....both the numbers and our eyes tell us that.

 

[garyeb2]

But they are valid because the only thing I'm using the numbers to say is that we have improved tremendously. Not arguing in any way that we are currently the 120th best defense.

Also, you continually dismiss these numbers, which are widely regarded in college basketball circles, but don't often give any reasoning ... I would be interested to know why ken Pom, sagarin, espn's BPI are not appropriate tools for discussing basketball?

They are fine for discussing basketball. But, it is a mistake to make definitive judgments using them. I like the Sagarin numbers. I have enjoyed them for years. Incorporating strength of schedule gives a better rendering of comparative strength than pure w/l records. RPI is pretty good too. It weights a team's w/l record, the opponents w/l record, and the opponents opponents w/l record. Ken Pom has something to sell and his customers have something to talk about. All good stuff.

Good stuff that is marginaly better than nothing. I'm having trouble following your position shifts. I think you are saying now that the 120 number doesn't really mean that OU is the 120th best defense. And that the 190 number really didn't mean that OU had the 190th best defense (even though that was your earlier assertion) but, some way or another the difference between 120 and 190 has some meaning. If the gap between two meaningless numbers has some meaning to you, we probably can not agree.

What got this started was your use of Ken Pom number's to discount winning in a disproportionate way. Winning is the dominate statistic. OU is 17-4 after playing a competitive schedule. We have built in limitations and areas that need continuous improvement. We are not bad at anything. Any stat or rating that indicates that this team is bad at anything is flawed.

I'm a numbers guy. I like stats that have some value. Does a stat have any predictive value? If it does, I'm in. How much predictive value? The more the better. The stuff that Pom, Sagarin, and the others generate have some but very limited predictive value. There is much better available.

Betting markets are now, have always been, and always will be the best, most efficient, and reliable predictor of outcomes. It doesn't matter if it is horse races, elections, or sporting events. Betting markets far out pace anything else in prediction out comes. A knowledge and understanding of betting markets renders a greater understanding of the relative strengths of opponents that the type of stats you are relying on.

 

[garyeb2]

It's not only the numbers that signify our improvement. Our eyes tell us that the defense is improving. As Gary pointed out above, we still have deficiencies based upon our limitations. A month ago, we couldn't defend a ball screen to save our lives. We've improved in that aspect a little bit. We are much improved at switching off defenders and our rotations have gotten better as well. And the defensive efficiency numbers bear this out. And as PNK said, these metrics (Kenpom, BPI etc.) are held in high regard amongst the basketball community. IMO, Bilas is the best CBB analyst around. And he continually quotes these metrics as a measure of a team's ability (or inability) in a particular facet of the game.

To dismiss these metrics (that actually semi-prove OU's improvement) is silly and short-sighted. We have improved significantly on defense since the conference season started....both the numbers and our eyes tell us that.

Ah, you are closing in on the argument. Ken Pom's numbers do not measure what you eyes are telling you. If you believe that you have seen improved switches, rotations, and busting ball screens, mayby you have.

Ken Pom's number measure outcomes. Results, if you prefer that word. Ken Pom's numbers could be measuring what you eyes are telling you or not. Like any other small sample example the fluctuations in Ken Pom numbers are just as likely to be the result of randomness.

 

[Max Power]

One thing to look at is that our defense has gotten better at the same time as Hornbeak getting healthy

 

[garyeb2]

One thing to look at is that our defense has gotten better at the same time as Hornbeak getting healthy

That is what my eyes are telling me.

 

[stormspencer]

They are fine for discussing basketball. But, it is a mistake to make definitive judgments using them. I like the Sagarin numbers. I have enjoyed them for years. Incorporating strength of schedule gives a better rendering of comparative strength than pure w/l records. RPI is pretty good too. It weights a team's w/l record, the opponents w/l record, and the opponents opponents w/l record. Ken Pom has something to sell and his customers have something to talk about. All good stuff.

Good stuff that is marginaly better than nothing. I'm having trouble following your position shifts. I think you are saying now that the 120 number doesn't really mean that OU is the 120th best defense. And that the 190 number really didn't mean that OU had the 190th best defense (even though that was your earlier assertion) but, some way or another the difference between 120 and 190 has some meaning. If the gap between two meaningless numbers has some meaning to you, we probably can not agree.

What got this started was your use of Ken Pom number's to discount winning in a disproportionate way. Winning is the dominate statistic. OU is 17-4 after playing a competitive schedule. We have built in limitations and areas that need continuous improvement. We are not bad at anything. Any stat or rating that indicates that this team is bad at anything is flawed.

I'm a numbers guy. I like stats that have some value. Does a stat have any predictive value? If it does, I'm in. How much predictive value? The more the better. The stuff that Pom, Sagarin, and the others generate have some but very limited predictive value. There is much better available.

Betting markets are now, have always been, and always will be the best, most efficient, and reliable predictor of outcomes. It doesn't matter if it is horse races, elections, or sporting events. Betting markets far out pace anything else in prediction out comes. A knowledge and understanding of betting markets renders a greater understanding of the relative strengths of opponents that the type of stats you are relying on.

When we had a defensive efficiency rating of 190, we were not a very good defensive team. We have significantly improved that number over the last several weeks (to 120). In order to improve our rating that significantly, we have probably been playing at defensive efficiency rating of around "60-70" during the last several weeks. Thus, by playing at a much higher level (per the raw numbers)...we have improved our overall defensive numbers.

Regarding prediction value, I wouldn't be quick to use Kenpom. Kenpom has some built -in discrepencies that he has been trying to account for over the last several years.http://www.hagrin.com/346/kenpom-wisconsin-problem

However, Kenpom does provide about the best "up to date" snapshot of how effecient a team is playing on offense and defense (with pace/luck adjustments) at a given point in time. But again, I wouldn't use it as the most effective tool in predicting the outcome of a game. We have been an underdog the last 7 games within most metric simulations, but we've won 6 of them. Regarding Sagarin, it now has two predictor measurements, but again, it's not perfect and has its shortcomings.

 

[ArdmoreSooner]

I have been hard on this teams defense myself. I watched a couple games in person and on TV. There is no question that the defense has gotten better. This team is still improving which is what you really want to see. Doesn't mean they are the best defensive team, but this team is so good offensively they don't have to be to be really good. If they continue to work to get better defensively, then they will be very competitive in March and have a chance at a run with the right match ups.

 

[j2d2]

Doesn't that huge jump mean that the 1st number wasn't any good. The defense hasn't got much better. Woodard still can't guard anyone, Buddy won't, Cam is still busting weak side help rotations, Booker is still getting picked off by every screener, we still can't protect the rim, and Spangler will still let some of them go instead of fouling.

Even if those numbers had any validity, the shift would more likely represent a regression toward the mean due to a larger sample size. We have never had a "putrid or "terrible" defense. Our defense from the get go was decent enough considering our built in limitations.

While I would agree that the defense is improving, it was never as bad as the 190 number and is much better than the 120.

I think you are right.