Charting Duke v. Princeton

[pfrduke]

Individuals
Miles Plumlee 54-22 (+32)
Kyrie Irving 65-39 (+26)
Kyle Singler 63-38 (+25)
Seth Curry 58-34 (+24)
Nolan Smith 71-48 (+23)
Andre Dawkins 49-30 (+19)
Mason Plumlee 48-34 (+14)
Josh Hairston 31-20 (+11)
Ryan Kelly 26-18 (+8)
Tyler Thornton 20-17 (+3)

Per 40 Minutes
Miles Plumlee +75.3
Kyrie Irving +41.6
Seth Curry +40.0
Kyle Singler +38.5
Andre Dawkins +34.5
Josh Hairston +33.8
Nolan Smith +30.7
Mason Plumlee +28.0
Ryan Kelly +22.9
Tyler Thornton +13.3

Lineups (Score, times used, margin)
Irving-Smith-Singler-Mason-Miles (23-11, 3x, +12)
Irving-Curry-Dawkins-Mason-Miles (8-2, 1x, +6)
Irving-Smith-Curry-Singler-Miles (8-2, 1x, +6)
Smith-Curry-Dawkins-Hairston-Kelly (7-2, 1x, +5)
Thornton-Curry-Dawkins-Hairston-Kelly (13-10, 1x, +3)
Irving-Smith-Dawkins-Hairston-Kelly (4-1, 1x, +3)
Irving-Curry-Dawkins-Singler-Miles (3-0, 1x, +3)
Thornton-Smith-Curry-Singler-Miles (5-3, 1x, +2)
Irving-Smith-Singler-Hairston-Mason (3-1, 1x, +2)
Smith-Curry-Dawkins-Singler-Kelly (2-0, 1x, +2)
Irving-Curry-Singler-Mason-Miles (2-0, 1x, +2)
Irving-Smith-Dawkins-Singler-Mason (5-4, 1x, +1)
Smith-Curry-Dawkins-Singler-Miles (3-2, 1x, +1)
Irving-Smith-Dawkins-Hairston-Mason (2-2, 1x, 0)
Irving-Smith-Dawkins-Singler-Kelly (0-0, 1x, +0)
Irving-Smith-Dawkins-Singler-Miles (2-2, 1x, 0)
Thornton-Smith-Curry-Singler-Hairston (2-4, 1x, -2)
Irving-Smith-Curry-Singler-Mason (5-9, 1x, -4)
Irving-Smith-Dawkins-Mason-Kelly (0-5, 1x, -5)

[Bob Green]

Thanks for charting the numbers pfrduke. The +/- numbers can be deceiving, especially when looking at only one game, however, it is encouraging to see our starting line-up be the most effective line-up and Miles Plumlee at +32. I thought Miles looked good in the low post yesterday and was a bit surprised to read so many negative comments in the post game thread. It'll be interesting to watch the +/- numbers for the first dozen games to see if any trends develop.

[superdave]

Thanks for charting the numbers pfrduke. The +/- numbers can be deceiving, especially when looking at only one game, however, it is encouraging to see our starting line-up be the most effective line-up and Miles Plumlee at +32. I thought Miles looked good in the low post yesterday and was a bit surprised to read so many negative comments in the post game thread. It'll be interesting to watch the +/- numbers for the first dozen games to see if any trends develop.

Thank you for calling out Miles for good play. I think his defense will be a big factor for us this season. The light bulb has come on for him in a lot of ways - he is going after his duties and trusting his teammates more to take care of theirs.

[airowe]

Thanks for doing these pfrduke. I know we have more talented players on the team this year, but I was a bit taken aback that only 2 lineups were used more than once. With so many different combinations possible, it will be interesting to see how much these lineups settle down into more regularity as the season progresses.

[COYS]

Thanks for doing these pfrduke. I know we have more talented players on the team this year, but I was a bit taken aback that only 2 lineups were used more than once. With so many different combinations possible, it will be interesting to see how much these lineups settle down into more regularity as the season progresses.

I second the thanks, pfrduke! I was also surprised that certain lineups were not used more than once, especially the Irving, Smith, Curry, Singler, Mason lineup that was actually at -4 for the day. I would expect that lineup to be used a bit more along with the Irving, Smith, Curry, Singler, Miles lineup in the future . . . but maybe not.

I thought Hairston played excellently in minimal minutes, even if his plus/minus stats aren't overwhelming compared to some of our other guys. He looked really active on defense and surprisingly smooth on his steal/breakaway dunk.

[-jk]

Thanks for posting this, and especially for doing the crunching.

As Bob noted, a single game's +/- doesn't tell the whole story. PFR, will you be able to do a cumulative, season long +/- that we can sticky?

Also, there were a couple things I'd asked Jumbo if he could do last season, but he wasn't able to with his data. You seem to have more automated data, so perhaps you can to this crunching.

First, as the season progresses, I would be very interested in seeing a running subset of the last (say) 4 games - enough games to see movements within a season as a player begins to assert himself, is sick or playing through an injury, or what-have-you, but something beyond a one game anomaly.

Also, in a cumulative total, if the margin of victory (or, alas, defeat) could be worked in, that also would be illuminating. Comparing a +32 and +38.5/40min performance with a +2 and +2/40min performance makes them seem radically different. Comparing Kyle's +32 and +38.5/40min and Duke's +37 for the Princeton game's margin to his +2 and +2/40min and Duke's +2 Butler margin - well, they become remarkably consistent. (Of course playing almost 40 minutes a game last season, Kyle's +/- tracked game margins closely!)

Within any one game, since we know the margin, I don't know that it really matters that it be in the chart - we can just put the final score in the post. (Glancing through last season's charts, I find I've forgotten some of the margins - gettin' old). For a season or a subset, though, it could be very compelling.

Anyway, I offer them as two ideas to make the +/- data more useful, knowing full well I'm not up to the task.

-jk

[Listen to Quants]

Thanks for posting this, and especially for doing the crunching.

<snip>

Also, in a cumulative total, if the margin of victory (or, alas, defeat) could be worked in, that also would be illuminating. Comparing a +32 and +38.5/40min performance with a +2 and +2/40min performance makes them seem radically different. Comparing Kyle's +32 and +38.5/40min and Duke's +37 for the Princeton game's margin to his +2 and +2/40min and Duke's +2 Butler margin - well, they become remarkably consistent. (Of course playing almost 40 minutes a game last season, Kyle's +/- tracked game margins closely!)

Within any one game, since we know the margin, I don't know that it really matters that it be in the chart - we can just put the final score in the post. (Glancing through last season's charts, I find I've forgotten some of the margins - gettin' old). For a season or a subset, though, it could be very compelling.

Anyway, I offer them as two ideas to make the +/- data more useful, knowing full well I'm not up to the task.

-jk


Indeed, I would very much like to second this last suggestion. An 'opponent correction' puts +/- numbers into context nicely. And thanks for the work on this.

[COYS]

Thanks for doing these pfrduke. I know we have more talented players on the team this year, but I was a bit taken aback that only 2 lineups were used more than once. With so many different combinations possible, it will be interesting to see how much these lineups settle down into more regularity as the season progresses.

One other interesting fact. Bob Green and others have consistently encouraged us to think of Dawkins as a 3 when projecting future lineups. It's only one game, but so far the Princeton game seems to prove Bob's stance. Dawkins was used exclusively as a 3 in every single lineup he was in, most often matching up with the taller and athletic Kareem Maddox. I had thought that Hairston might see some time at the three as a defender/screener a la Thomas last season and possibly competing with Dawkins for minutes there, but a two bigs plus Hairston lineup wasn't used for this game. As Coach K said in his post game presser, Dawkins really has improved. If he continues to play well as the 3, then Duke becomes a deeper team in the post with Hairston able to be a backup to Kelly, Mason, and Miles and in the perimeter, with Dawkins as a back up for Singler and a strong option as the 3 for a smaller lineup when Kyle slides to the 4. Singler is still the teams most unique player in terms of skill set, but if Dawkins is able to continue his promising start, Duke becomes even deeper and more versatile.

[Rich]

One other interesting fact. Bob Green and others have consistently encouraged us to think of Dawkins as a 3 when projecting future lineups. It's only one game, but so far the Princeton game seems to prove Bob's stance. Dawkins was used exclusively as a 3 in every single lineup he was in, most often matching up with the taller and athletic Kareem Maddox. I had thought that Hairston might see some time at the three as a defender/screener a la Thomas last season and possibly competing with Dawkins for minutes there, but a two bigs plus Hairston lineup wasn't used for this game. As Coach K said in his post game presser, Dawkins really has improved. If he continues to play well as the 3, then Duke becomes a deeper team in the post with Hairston able to be a backup to Kelly, Mason, and Miles and in the perimeter, with Dawkins as a back up for Singler and a strong option as the 3 for a smaller lineup when Kyle slides to the 4. Singler is still the teams most unique player in terms of skill set, but if Dawkins is able to continue his promising start, Duke becomes even deeper and more versatile.

I believe this was mentioned in another thread, but I think that Dre and Kyle go up against each other in practice so it seems that coach K is committed to using Dre in the 3 spot. I agree that adding depth to the 3 only makes this team better since we have a number of guys that can play at 1 and 2.

[MChambers]

From these great numbers (thanks, pfrduke), it looks to me that the lineups where Kyle was at the 3 outscored Princeton 28-12. Lineups where he was at the 4 (or maybe even the 5, when he was paired with Josh) did less well, 27-20.

Interesting to note that Kelly and Singler were rarely on the court at the same time, and never with Kyle at the 3.

In case anyone wants to check my work, I've put a "B" next to lineups with Kyle at the 3 and an "S" next to lineups with Kyle at the 4.

Lineups (Score, times used, margin)
Irving-Smith-Singler-Mason-Miles (23-11, 3x, +12) B
Irving-Curry-Dawkins-Mason-Miles (8-2, 1x, +6)
Irving-Smith-Curry-Singler-Miles (8-2, 1x, +6) S
Smith-Curry-Dawkins-Hairston-Kelly (7-2, 1x, +5)
Thornton-Curry-Dawkins-Hairston-Kelly (13-10, 1x, +3)
Irving-Smith-Dawkins-Hairston-Kelly (4-1, 1x, +3)
Irving-Curry-Dawkins-Singler-Miles (3-0, 1x, +3) S
Thornton-Smith-Curry-Singler-Miles (5-3, 1x, +2) S
Irving-Smith-Singler-Hairston-Mason (3-1, 1x, +2) B
Smith-Curry-Dawkins-Singler-Kelly (2-0, 1x, +2) S
Irving-Curry-Singler-Mason-Miles (2-0, 1x, +2) B
Irving-Smith-Dawkins-Singler-Mason (5-4, 1x, +1) S
Smith-Curry-Dawkins-Singler-Miles (3-2, 1x, +1) S
Irving-Smith-Dawkins-Hairston-Mason (2-2, 1x, 0)
Irving-Smith-Dawkins-Singler-Kelly (0-0, 1x, +0) S
Irving-Smith-Dawkins-Singler-Miles (2-2, 1x, 0) S
Thornton-Smith-Curry-Singler-Hairston (2-4, 1x, -2) S
Irving-Smith-Curry-Singler-Mason (5-9, 1x, -4) S
Irving-Smith-Dawkins-Mason-Kelly (0-5, 1x, -5)

[johnb]

If I were Andre, I'd definitely want to be a 3, where there'll be minutes when Singler slides to the 4 and when he goes to the bench. There are downsides to guarding guys who are bigger, but that may be offset by the reality that those bigger and presumably slower guys may not be able to effectively cover him on the outside. Plus, playing the 3 sets him up to compete with Michael Gbinje for a starting job next year, as opposed to looking for time amongst Curry, Rivers, Thornton, and Cook (+/- Kyrie).

[Kedsy]

From these great numbers (thanks, pfrduke), it looks to me that the lineups where Kyle was at the 3 outscored Princeton 28-12. Lineups where he was at the 4 (or maybe even the 5, when he was paired with Josh) did less well, 27-20.

Interesting to note that Kelly and Singler were rarely on the court at the same time, and never with Kyle at the 3.

In case anyone wants to check my work, I've put a "B" next to lineups with Kyle at the 3 and an "S" next to lineups with Kyle at the 4.

Lineups (Score, times used, margin)
Irving-Smith-Singler-Mason-Miles (23-11, 3x, +12) B
Irving-Curry-Dawkins-Mason-Miles (8-2, 1x, +6)
Irving-Smith-Curry-Singler-Miles (8-2, 1x, +6) S
Smith-Curry-Dawkins-Hairston-Kelly (7-2, 1x, +5)
Thornton-Curry-Dawkins-Hairston-Kelly (13-10, 1x, +3)
Irving-Smith-Dawkins-Hairston-Kelly (4-1, 1x, +3)
Irving-Curry-Dawkins-Singler-Miles (3-0, 1x, +3) S
Thornton-Smith-Curry-Singler-Miles (5-3, 1x, +2) S
Irving-Smith-Singler-Hairston-Mason (3-1, 1x, +2) B
Smith-Curry-Dawkins-Singler-Kelly (2-0, 1x, +2) S
Irving-Curry-Singler-Mason-Miles (2-0, 1x, +2) B
Irving-Smith-Dawkins-Singler-Mason (5-4, 1x, +1) S
Smith-Curry-Dawkins-Singler-Miles (3-2, 1x, +1) S
Irving-Smith-Dawkins-Hairston-Mason (2-2, 1x, 0)
Irving-Smith-Dawkins-Singler-Kelly (0-0, 1x, +0) S
Irving-Smith-Dawkins-Singler-Miles (2-2, 1x, 0) S
Thornton-Smith-Curry-Singler-Hairston (2-4, 1x, -2) S
Irving-Smith-Curry-Singler-Mason (5-9, 1x, -4) S
Irving-Smith-Dawkins-Mason-Kelly (0-5, 1x, -5)

I maybe missing something, but when I add up the "S" lineups I get 35-26. Still less well than when Kyle was at the 3, though.

[MChambers]

I maybe missing something, but when I add up the "S" lineups I get 35-26. Still less well than when Kyle was at the 3, though.
I missed a couple of lineups when I added it up. (Consistent with my mathematics classes at Duke: understood the concepts but inevitably made one or two careless errors.)

[stickdog]

It's interesting to me that while Curry & Dawkins were both in the game together, Duke outscored Princeton 36-16.

Those dudes can really stretch a defense.

[superdave]

It's interesting to me that while Curry & Dawkins were both in the game together, Duke outscored Princeton 36-16.

Those dudes can really stretch a defense.

When they play together it should be know as the ZoneBuster lineup. It sounds like a diet pill but is actually a cure for unimaginative defenses.

[Orange&BlackSheep]

When they play together it should be know as the ZoneBuster lineup. It sounds like a diet pill but is actually a cure for unimaginative defenses.

I just watched that game with no hope in my heart. Can anyone name a team that had more great three-point shooters on it? I think that Singler, Smith, Irving, Dawkins, Curry, and Kelly could all conceivably finish the year above 40% from three. Personally, I feel that the shot is going in whenever any one of those guys gets an open look. There are going to be games this year where the team gets on a roll and just buries teams with a flurry of threes.

[airowe]

One other interesting fact. Bob Green and others have consistently encouraged us to think of Dawkins as a 3 when projecting future lineups. It's only one game, but so far the Princeton game seems to prove Bob's stance. Dawkins was used exclusively as a 3 in every single lineup he was in, most often matching up with the taller and athletic Kareem Maddox. I had thought that Hairston might see some time at the three as a defender/screener a la Thomas last season and possibly competing with Dawkins for minutes there, but a two bigs plus Hairston lineup wasn't used for this game. As Coach K said in his post game presser, Dawkins really has improved. If he continues to play well as the 3, then Duke becomes a deeper team in the post with Hairston able to be a backup to Kelly, Mason, and Miles and in the perimeter, with Dawkins as a back up for Singler and a strong option as the 3 for a smaller lineup when Kyle slides to the 4. Singler is still the teams most unique player in terms of skill set, but if Dawkins is able to continue his promising start, Duke becomes even deeper and more versatile.


I believe this was mentioned in another thread, but I think that Dre and Kyle go up against each other in practice so it seems that coach K is committed to using Dre in the 3 spot. I agree that adding depth to the 3 only makes this team better since we have a number of guys that can play at 1 and 2.

Dawkins matched up against Singler in all 4 practices I saw before the season. He did a very impressive job pestering Kyle too, and you have to imagine he is gaining confidence with each practice. If he's able to even slow down the perceived NPOY (unless you're Gary Parrish :rolleyes:), mentally that has to make him realize that he can do the same, if not better, to other guys with his size and nearly his skill set. Barring anything unforeseen, expect Andre to be Duke's starting small forward in the 2012 season, unless he transfers to Stanford. :eek: :p