Duke basketball 2015-2016 discussion thread

[SeanMayTriedToEatMe]

If Plumlee is putting up a well-below-average plus-minus per minute while putting up amazing efficiency stats otherwise, I have to wonder whether our lineups can be improved.

Surprise - freshmen aren't as good as upperclassmen in the early season... except Kennard, who is secretly an all-around stud. Just wait til his 3pt% gets into the 40s, which I'd expect to happen soon.

 

[deeyoukayeee]

K focused on turnovers in his press conference. We gave up 17 pts on turnovers. We take care of the ball and our defense looks spectacular. Our offense hurt our defense.

That's pretty much what the eye test said too. You're gonna have a shitty fucking defensive efficiency when you give the other team that many fast breaks. I'd wager our half court defense was good to elite. But unfortunately it seems like the turnovers are going to be a problem for a while longer

I don't think TOs will be nearly as problematic vs. teams who start fewer than three really quick, excellent defensive guards. Which is every team in the country other than UK. Add in the fact that 3-5 of those horrific DT/Grayson floaters right in the teeth of two great shotblockers were de facto TOs, and that K will have those guys throttle it down going forward, and I think the abhorrent fastbreak baskets against numbers are unlikely to be duplicated.

However, ball movement and getting it to guys like BI and Luke in the right position for them to score will be an ongoing concern.

 

[childress22]

I think it's foolish to make any conclusions based on three games. However, if someone were to propose a starting lineup that had Kennard in there at the point, Grayson and Matt on the wings, and Amile and Marshall down low, I wouldn't hate it or anything.

 

[Slap the Floor]

Am I the only one pissed off Sulaimon is averaging 6 fucking assists per game?

 

[rome8180]

Low sample size obviously. But it's going to be really annoying if he puts it all together finally. I mean, I don't think any sane Duke fan would have denied his potential. In every season at Duke he had games where he played like an absolute legend.

 

[childress22]

I'm going to stick the overall team plus-minus stuff in this thread. We don't have statsheet.com to even give us partial plus-minus data anymore, and I don't know of any other site that keeps track of the season totals, so Duke-Forum.com will be tasked with keeping track of it this season for Duke and possibly beyond.

Data gathered from scacchoops.com box scores.

Note [A lot of my Siena numbers were off by 2 earlier; everything in this thread now is fixed and acurate].

Okay, so I've computed all those plus-minuses, and SCACC hoops is dreadfully low. You can get a sense for the fact that it has our regulars having an "average" plus-minus of 20. In reality, our regulars have an average plus-minus of 46.6 per position, or 29.125 per player. (Vrankovic and Pags cancel each other out, and Obi's -8 ends up bringing us to the 45 points by which we have outscored opponents).

Every single regular player has a more positive +/- than reported by SCACC. The general pattern is more or less similar (everyone is between 6 and 17 points lower than their true value) The real numbers are below:

Allen +48
Ingram +11
Jefferson +49
Jeter +13
Jones + 33
Kennard +45
MP3 +15
Thornton +19

The other guys have less impressive values, and obviously small sample sizes.
Vrankovic +2
Pagliuca -2
Obi - 8

 

[childress22]

SMTTEM, if I include all the values for the individual games, would you mind re-doing the table?

Here they are, for Siena, Bryant, and Kentucky

Allen: +28, +27, - 7
Ingram: +3, +22, -14
Jefferson: +26, +25, -2
Jeter: -3, +12, +4
Jones: +24, +21, -12
Kennard: +15, +28, +2
MP3: +10, +20, -15
Thornton: -5, +34, -10
--
Obi, -8, NA, NA
Pagliuca N/A, -2, N/A
Vrankovic -3, +5, NA

 

[SeanMayTriedToEatMe]

This brings up a big problem with plus-minus in general - it's going to be measured differently by different sources, especially when automated. When Duke commits a foul and sends an opponent to the line for free throws at 8:03 of the 1st half, and Jones is subbed out for Kennard after the first of two free throws, at 8:03 of the 1st half, which player is being accounted for as being on the court when each of the free throws goes in? If Kennard then plays until 4:04 left in the 1st half, how many minutes did he play? It was exactly 3:59, but how is the automated source rounding this segment for Kennard? It's doubtful that it's using seconds in each segment. There's significant margin for error in simply adding up minutes for players in each game.

Note that if you simply reduce the total minutes by 8 (1 minute per player), the average plus-minus comes out to 30.3 per player, which is apparently what it should be. Is this your typical rounding error in individual segments that adds up to make the total look off? More likely, it's a combination of this problem and the source's procedure for determining who should get credit for certain points, like in the free throw example.

I'd prefer to just use the automated source, because (1) the +/- goodness of each Duke regular relative to each other doesn't change much at all using your numbers instead (e.g., Plumlee and the freshmen minus Kennard are still below average, while the upperclassmen minus Plumlee are above average), and (2) if the purpose of all of this, besides educating ourselves, is to make a point to someone who doesn't want to believe what numbers tell them (e.g., on TDD), then they're not going to be persuaded by "some guy on Duke-Forum.com got these numbers manually by reading play-by-plays so we should trust these numbers over a somewhat decent website (scacchoops)."

If it's wrong, it's going to be consistently wrong for each player using consistent automated procedures, and all the issues should balance out for the players relative to each other over the season. If someone can get per possession stats, then that's worth going through the trouble, since there is no site that does that for college basketball.

 

[childress22]

I'm happy to do it manually if it gets us the correct numbers. In terms of the FT problem, the shots should always be credited to the lineup that was on the floor when the foul occurred; GoDuke gets this right by always listing the subs after the FTs have occurred (even when there's a miss and a rebound by a player that just subbed in).

I didn't pull minutes data, but I could, if we wanted it to be more accurate. At this point, I've got it all in there for every line-up segment, it would just be a matter of compiling it for the individual players.

The biggest problem with SCACC hoops is not the minute rounding, it's the +/- values. They're all low -- In Derryck's case, they have him as a +4 instead of +19 over three games. That's a big difference that would just accumulate over time, probably.

The SCACC hoops stuff is off for weird reasons. For some games, they arbitrarily have the same stretch of time listed twice, and added twice. For others, I think they must be missing some stretches of time.

 

[SeanMayTriedToEatMe]

The good thing about playing Kentucky, win or lose, is that we should be close to indifferent to how Kentucky performs for the rest of the season. We don't want them to be too good, to the point where they're clearly on a different level from Duke, but we also don't mind if they crush everyone, since it helps our computer numbers and the perception of how good Duke is. Kentucky games are just going to be meh to me this season, and that's a good thing.

 

[SeanMayTriedToEatMe]

Zack set up a program for per possession stats for all players last year. He can probably do that again, and I'd just wait for that. We know per minute is not good, so we're just biding time with garbage.

 

[ZackM]

Zack set up a program for per possession stats for all players last year. He can probably do that again, and I'd just wait for that. We know per minute is not good, so we're just biding time with garbage.

I can certainly do that again, but it's not going to happen for at least a week. I'm on vacation until after Thanksgiving and all of that data is on a different computer.

 

[childress22]

(1) the +/- goodness of each Duke regular relative to each other doesn't change much at all using your numbers instead (e.g., Plumlee and the freshmen minus Kennard are still below average, while the upperclassmen minus Plumlee are above average).

Mostly true, although the shittiness of the SCACC numbers on per-minute values is magnified where there are fewest minutes. In this case, that's Jeter, who in real life is not a +0.10 ppm guy, but a +.042 ppm guy.

 

[physicsfactor]

Zack set up a program for per possession stats for all players last year. He can probably do that again, and I'd just wait for that. We know per minute is not good, so we're just biding time with garbage.

I can certainly do that again, but it's not going to happen for at least a week. I'm on vacation until after Thanksgiving and all of that data is on a different computer.

Oh come on. This is bullshit.

 

[bdotling]

Zack set up a program for per possession stats for all players last year. He can probably do that again, and I'd just wait for that. We know per minute is not good, so we're just biding time with garbage.

I can certainly do that again, but it's not going to happen for at least a week. I'm on vacation until after Thanksgiving and all of that data is on a different computer.

Oh come on. This is bullshit.

Seriously dude, Duke basketball never sleeps

 

[rome8180]

I'm more impatient for this blog.

 

[aiw]

I don't know why this tweet didn't show up before, but now it looks like Keeley deleted it.

She responded to a tweet that said "Why doesn't this asshole play Jeter and Obi?" with "Obi probably won't see the floor this year."

 

[rome8180]

Gonna be a long one. Maybe not without some good moments though (i.e. every time Ingram is on the bench).

 

[Pantone287]

So, it's settled. Thornton is the PG, Amile is the C.

Only question remaining is can Ingram overtake Kennard for the PF spot. Right?

 

[rhfarmer]

Can Matt Jones pull his head out of his ass for a full season is a remaining question.