Big 12 Four Factors - FINAL 2014 Rankings

[kso_FAN]

But we've continued to have one of the lowest in the league, and it's stayed low way longer than it should have if it was random.

I'm not sure this is correct.

Which part?

We do have one of the best 3PT% defenses in the country (28.7%, #9) and its still really good in Big 12 games only (31.8%, #2).

I'm still most surprised by our 2PT% defense. Our 2PT% defense after the OOC was 47.9% compared to 43.7% in Big 12 games. That's a significant improvement against much better competition.

[Stevesie60]

I want to make fun of michigancat for being wrong, but those conversations haven't gone well for me in the past.

[michigancat]

But we've continued to have one of the lowest in the league, and it's stayed low way longer than it should have if it was random.

I'm not sure this is correct.

To expand, the Big 12 average 3FG is 35.4%, and we're at 31.8% on defense. Our conference opponents have attempted 173 3fg's in conference and made 55. If our opponents shot the average, they would have made 61. I think you can point to 6 makes/misses over 11 games as being random.

But we've continued to have one of the lowest in the league, and it's stayed low way longer than it should have if it was random.

I'm not sure this is correct.

Which part?

We do have one of the best 3PT% defenses in the country (28.7%, #9) and its still really good in Big 12 games only (31.8%, #2).

Conference games, especially, (as shown above). With regards to our crazy overall number, I think you can look at oscar's 10 year history and clearly see that number is random.

[sys]

Okay, so you and kenpom have both said that you can have some control over forcing opposing team to take low quality shots. How can you both still say it's random? Of course there's "luck" involved to some extent, but there's plenty of luck involved in 2pt% defense as well. And I bet if you were able to break it down, though I doubt you (or anyone) would be able to, the teams that continually force rushed 3's just happen to be luckier throughout the year.

don't think of it as either random or nonrandom.  think of it as the % of overall variation that is explained by random (stochastic) factors and by nonrandom factors.  for example (numbers are purely hypothetical), 80% of the variation may be random, 20% related to defensive quality, defensive style, opponent quality, etc.

[kso_FAN]

Conference games, especially, (as shown above). With regards to our crazy overall number, I think you can look at oscar's 10 year history and clearly see that number is random.

Fair enough, but I think you could argue that oscar's defense had a really good run from 06-11, including one season under 30%. While this season is still low, this does point to him coaching "good 3PT defense" to an extent. The real abnormal season (just as abnormal as this one) was giving up nearly 38% in his last (terrible) season at Illinois, the only season from 2006 through this year that he had a team allow >33% from 3PT.

[Stevesie60]

Okay, so you and kenpom have both said that you can have some control over forcing opposing team to take low quality shots. How can you both still say it's random? Of course there's "luck" involved to some extent, but there's plenty of luck involved in 2pt% defense as well. And I bet if you were able to break it down, though I doubt you (or anyone) would be able to, the teams that continually force rushed 3's just happen to be luckier throughout the year.

don't think of it as either random or nonrandom.  think of it as the % of overall variation that is explained by random (stochastic) factors and by nonrandom factors.  for example (numbers are purely hypothetical), 80% of the variation may be random, 20% related to defensive quality, defensive style, opponent quality, etc.

But if 80% is random for every team, wouldn't the 20% be observable? Or are seasons too short to have enough 3pts attempted in order to see a significant difference?

[michigancat]

Conference games, especially, (as shown above). With regards to our crazy overall number, I think you can look at oscar's 10 year history and clearly see that number is random.

Fair enough, but I think you could argue that oscar's defense had a really good run from 06-11, including one season under 30%. While this season is still low, this does point to him coaching "good 3PT defense" to an extent. The real abnormal season (just as abnormal as this one) was giving up nearly 38% in his last (terrible) season at Illinois, the only season from 2006 through this year that he had a team allow >33% from 3PT.

Yeah, he definitely coached some nice defenses over the years - I'm not denying that. I'm just pointing out that the 3FG defense this year is largely random. (sys's post explained it much more eloquently).

[sys]

Okay, so you and kenpom have both said that you can have some control over forcing opposing team to take low quality shots. How can you both still say it's random? Of course there's "luck" involved to some extent, but there's plenty of luck involved in 2pt% defense as well. And I bet if you were able to break it down, though I doubt you (or anyone) would be able to, the teams that continually force rushed 3's just happen to be luckier throughout the year.

don't think of it as either random or nonrandom.  think of it as the % of overall variation that is explained by random (stochastic) factors and by nonrandom factors.  for example (numbers are purely hypothetical), 80% of the variation may be random, 20% related to defensive quality, defensive style, opponent quality, etc.

But if 80% is random for every team, wouldn't the 20% be observable? Or are seasons too short to have enough 3pts attempted in order to see a significant difference?

i'm not going to say no, but without thinking about it too deeply - it would be hard to detect a small effect on a very noisy variable.  the game to game variance is really large in relation to the possible signal you'd be looking for.  not to say it couldn't be done.  one could easily model how large an effect would have to be in order to be able to detect it with a given variance and sample size - but i haven't done that and am not going to do it.

[kso_FAN]

I was curious to see what it looked like:

[michigancat]

From Kenpom's "Boeheim exception" post:

While I believe the Boeheim 2-3 zone influences opponents’ three-point percentage, I don’t want to go overboard. They have a couple of systematic advantages that help their numbers. First, their non-conference schedule is largely home games and second, the Big East does not contain a lot of good perimeter-shooting teams. We can remove these two effects by looking at how the Orange have ranked in Big East play in conference games only.

Since 2004, here’s how those rankings looked: 1, 7, 15, 8, 4, 2, 4, 9, and 4. The Orange may have the league’s best 3P% defense, as much as such a thing exists, but it’s really difficult to see that consistently displayed over an 18-game stretch. Compare that to the Big East’s best two-point percentage defense over that time.

Here’s how UConn has ranked in the Big East in that category over the same nine-year stretch: 1, 1, 1, 2, 1, 1, 1, 6, and 3. You just can’t dominate three-point percentage defense like that because there is so much variance involved from game-to-game, which I would say is due to the defense’s lack of influence on three-point percentage. (It’s true that there are fewer three-point attempts than two-point attempts which contributes to increased variance, but even accounting for this, two-point defense is more consistent than three-point defense from year-to-year.)

[Powercat Posse]

Kstate in Big 12 games is only allowing 28.1% of shots to be 3pt.   Tcu (31.3) and WV (32.6) are 2nd/3rd.   However TCU is 9th (37.7) and WV 6th (36.0) in 3pt% defense, while we are 2nd at 31.8%

 So no surprise that Ksu leads the league allowing only 21.9% of our opp points to be from the 3pt line.  We do a good job of not letting a high% of shots to be from 3pt and for various reasons, teams do not shoot a good % against us.   Taking an ave of the other 9 teams in the conf, they allow 27.8% of pts to be from 3 pt makes

[Stevesie60]

I can't be reading that right. He's saying UCONN finished ranked 1, 1, 1, 2, 1, 1, 1, 6, and 3 in 3pt% defense but still claims it's random?

[kso_FAN]

I can't be reading that right. He's saying UCONN finished ranked 1, 1, 1, 2, 1, 1, 1, 6, and 3 in 3pt% defense but still claims it's random?

No, that's their 2PT% defense.

[Stevesie60]

I can't be reading that right. He's saying UCONN finished ranked 1, 1, 1, 2, 1, 1, 1, 6, and 3 in 3pt% defense but still claims it's random?

No, that's their 2PT% defense.

Ah, gotcha. I guess my thought is that it's such a fine line, that it would be really hard to maintain from season to season when you have different personal and are facing different personal, so I could see why it would vary like that. But over the course of a season it seems to be a trend. Obviously I say that only having tracked our 3pt% defense this year. But late in our non-con it was obvious that it was low, and we were regularly holding teams below their 3pt% average. And we have continued to do that up till now (outside of a couple of games, of course), over halfway through our conference schedule. So it's hard for me to believe that we've continued on that trend even though it's random. I feel like I kind of share the view of sys' in that there is obviously going to be a large variance, but to pretend we don't have any control (I'd even say it's more than 20%, which is what sys suggested) just seems silly at this point.

[michigancat]

I can't be reading that right. He's saying UCONN finished ranked 1, 1, 1, 2, 1, 1, 1, 6, and 3 in 3pt% defense but still claims it's random?

No, that's their 2PT% defense.

Ah, gotcha. I guess my thought is that it's such a fine line, that it would be really hard to maintain from season to season when you have different personal and are facing different personal, so I could see why it would vary like that. But over the course of a season it seems to be a trend. Obviously I say that only having tracked our 3pt% defense this year. But late in our non-con it was obvious that it was low, and we were regularly holding teams below their 3pt% average. And we have continued to do that up till now (outside of a couple of games, of course), over halfway through our conference schedule. So it's hard for me to believe that we've continued on that trend even though it's random.

How can you say we "followed the trend" when we were 26.2% in OOC and 31.8% in conference games? That's a big jump.

[Stevesie60]

The trend being holding teams below their average, often to a season low. You know our conference schedule is tougher than our non-con, yes?

[sys]

I feel like I kind of share the view of sys' in that there is obviously going to be a large variance, but to pretend we don't have any control (I'd even say it's more than 20%, which is what sys suggested) just seems silly at this point.

that's not my view.  i think it is overwhelmingly random.  i threw out 20% as a purely hypothetical example.  whatever the percentage of variation attributable to non-stochastic factors is, i think it is quite low.  perhaps undetectable.

[The Tonya Harding of Twitter Users Creep]

wish we could be that good at 2 pt defense. that's like, really good.

[michigancat]

The trend being holding teams below their average, often to a season low. You know our conference schedule is tougher than our non-con, yes?

In conference, our opponents have shot over their average from three in 5 of 11 games.

[Stevesie60]

The trend being holding teams below their average, often to a season low. You know our conference schedule is tougher than our non-con, yes?

In conference, our opponents have shot over their average from three in 5 of 11 games.

Yeah, I make things up a lot to try to prove my point. I guess you didn't get the memo to not call me out on them.

[Powercat Posse]

I just hope we keep doing what we have been doing.   This season is not over, but in the last 6 seasons of Big 12, our 21.9% of pts allowed from 3pt makes is the 3rd best and the 31.8% on 3pt Def is the 6th best (overall in those 6 seasons).  And I would guess the 28.1% of shots being 3pt is a pretty good mark over those 6 yrs too

What I also like is  only 2 of 11 games have we let a Big 12 team to shoot 30%(or more) of their shots from 3pt and make 35.0% or higher.
Last year our offense was pretty efficient and the 3pt shot was a big factor.  13 of 18 games, 30% or more of our shots were from 3pt, while hitting 35.0% or higher

[cDubya]

This is how sports boards are supposed to look! My god, the stats. 

[kso_FAN]

The middle of this league continues to be crazy. There has been lots of discussion about road losses around here, so here is a league breakdown of that as well:



While the road has been awful, at least the Cats have avoided a home loss. At minimum, hopefully the Cats win out at home and get 1 road win. 11-7 would be nice, but I'd accept 10-8. At best the Cats might be able to slide to a 7, but the 8/9 line seems likely and the 6 seems a long way away.

[Powercat Posse]

Poor TCU.  Maybe they could win a game if they got to play at home. 0-12 on the road.....LOL.  J/K, I know it was a typo.

If we 10-8, I will deal with it, but I wont accept it.  Not at this point.   10-8 meens we either have a let down and lose at home or we were only able to finish 1-8 on road.  11-7 for me is fine.  I think that would keep us out of the 8 vs 9 seed game, and its hard to be disappointed with a 7 seed in the Big Dance

Posted this in another thread, but the main problem with our offensive woes on the road tracks back to 3PTA and 3PT%.   
6 road games = 29.5% and only 27.4% of shots were 3pt (26.3 3PT% and 26.1% 3 pt shots in the 5 losses)
6 home games = 40.4% and 33.5% of shots were 3pt

[kso_FAN]

Oops on TCU. They are 0-6 home and away.

I don't mind the lower 3PTA% on the road with that 3PT%. Seems to be a real effort to take it to the paint. We had bad moments, but our offense did some really great things against Baylor's zone with a ridiculous assist%. That said, at some point down the stretch you need players to make plays that don't require assists and we didn't get that.