PDA

View Full Version : Charting Duke vs. Maryland



Jumbo
01-27-2008, 11:47 PM
Thanks to Georgia Tech, Virginia and the good folks at FSN, my I missed the first few minutes of charting. Thankfully, the gamebook shows all subs, so all is well. The numbers are here. And they're interesting.

Again, I'll start with the usual caveat that single-game plus/minus stats can be very misleading. That said, here's some stuff that backed up what seemed obvious while watching the game:
-Kyle Singler didn't seem to be on his game tonight, and he was a -1. I'm sure that I don't need to go back to the archives to conclude that other than Pitt (when he was -6), that was his only negative finish of the year.
-Jon Scheyer entered the game as Duke's plus/minus leader. He seemed to struggle and he was +3.
-DeMarcus Nelson played 38 minutes, so it's pretty obvious that his numbers would mesh with Duke's overall total. In fact, there were a grand total of four points scored with Marie out of the game -- two for Duke, two for Maryland.

But there was some counterintuitive stuff as well:
-Dave McClure didn't seem to do much tonight. But in only eight minutes, he was +8. That might be a case of being in the game at the right time for a run. But it's interesting.
-Here's the biggie, and I think it's really instructive for a certain debate raging in another thread. Nolan Smith was fabulous tonight. People are wondering about Paulus. And yet, Paulus was a +10 tonight. In other words, Duke was outscored by a point when he was out of the game. Smith, but contrast, was a +3 (Duke outscored Maryland 6-3 when they were in together, and Scheyer was technically the PG during an offense/defense sub in the final minute when Maryland scored a point).

Those are some juicy numbers to ponder. Here are more:

Individuals
Greg Paulus 53-43 (+10)
DeMarcus Nelson 91-82 (+9)
David McClure 20-12 (+8)
Gerald Henderson 81-74 (+7)
Lance Thomas 41-34 (+7)
Jon Scheyer 60-57 (+3)
Nolan Smith 46-43 (+3)
Kyle Singler 67-68 (-1)
Taylor King 6-7 (-1)

Per 40 Minutes
David McClure +40
Greg Paulus +21.1
Lance Thomas +15.6
DeMarcus Nelson +9.5
Gerald Henderson +8.5
Nolan Smith +6
Jon Scheyer +4.3
Kyle Singler -1.3
Taylor King -8

Lineups
Smith-Scheyer-Nelson-Henderson-Thomas (x2) 14-8 (+6)
Paulus-Scheyer-Nelson-McClure-Singler 6-1 (+5)
Paulus-Scheyer-Nelson-Henderson-Thomas (x2) 4-0 (+4)
Paulus-Nelson-Henderson-Singler-Thomas (x4) 23-21 (+2)
Paulus-Nelson-Henderson-McClure-Singler 5-3 (+2)
Paulus-Smith-Nelson-Henderson-Singler 3-1 (+2)
Smith-Scheyer-Henderson-McClure-King 2-0 (+2)
Paulus-Smith-Scheyer-Nelson-Henderson (x2) 3-2 (+1)
Smith-Nelson-Henderson-McClure-Singler 2-2 (0)
Smith-Scheyer-Nelson-McClure-Singler 2-2 (0)
Smith-Nelson-Henderson-McClure-King 0-0 (0)
Smith-Nelson-Henderson-Singler-Thomas 0-0 (0)
Scheyer-Nelson-Henderson-Singler-Thomas (x2) 0-1 (-1)
Smith-Scheyer-Nelson-King-Singler 4-5 (-1)
Paulus-Scheyer-Nelson-Henderson-McClure 3-4 (-1)
Smith-Scheyer-Nelson-Singler-Thomas 0-2 (-2)
Smith-Scheyer-Henderson-King-Thomas 0-2 (-2)
Smith-Scheyer-Nelson-Henderson-Singler (x4) 16-19 (-3)
Paulus-Scheyer-Nelson-Henderson-Singler (x3) 6-11 (-5)

shadowfax336
01-27-2008, 11:50 PM
Why are we playing Smith at all!!
Paulus should be playing 40 minutes a game!
And anyone who thinks Smith should play any minutes clearly doesn't have the best interests of the team in mind...

Jumbo
01-27-2008, 11:52 PM
Why are we playing Smith at all!!
Paulus should be playing 40 minutes a game!
And anyone who thinks Smith should play any minutes clearly doesn't have the best interests of the team in mind...

LOL.

devildeac
01-27-2008, 11:53 PM
Yea, GP really sucked tonight and deserves to lose his starter role(insert internet sarcasm alert here and see other thread). Good stuff once again, Jumbo. What great progress by NS and how nice it is to have a healthy GP this year.

dukie8
01-27-2008, 11:58 PM
i'll point out that the numbers are skewed slightly by the last minute of the game when we were up, they were fouling, we were making our fts and they were missing their shots. part of the problem with +/- stat tracking is that not all points are the same.

Troublemaker
01-28-2008, 12:04 AM
I'm watching the FSN replay of the game right now (because my TV provider was on the fritz during the live game), and I can say that Greg played really well in the first 8 minutes of the game. The part of the game that EVERYONE missed.

Jumbo
01-28-2008, 12:08 AM
i'll point out that the numbers are skewed slightly by the last minute of the game when we were up, they were fouling, we were making our fts and they were missing their shots. part of the problem with +/- stat tracking is that not all points are the same.

Well, yeah. That's why I included my caveat. We were up 4 when they started fouling -- Lance hit those two FTs. Here's how my log looked during that sequence:
86-82
Paulus-Scheyer-Nelson-Henderson-Thomas
87-82
88-82
Paulus-Smith-Scheyer-Nelson-Henderson
89-82
89-83
89-84
Paulus-Scheyer-Nelson-Henderson-Thomas
90-84
91-84
Smith-Scheyer-Nelson-Henderson-Thomas
Paulus-Smith-Scheyer-Nelson-Henderson
92-84
93-84
FINAL

So, once the fouling started, if we're doing the Smith/Paulus thing, it looks like Paulus was +5 and Smith was +1.

Jumbo
01-28-2008, 12:09 AM
I'm watching the FSN replay of the game right now (because my TV provider was on the fritz during the live game), and I can say that Greg played really well in the first 8 minutes of the game. The part of the game that EVERYONE missed.

I'd love to know if I have my lineups right for those eight minutes. I've already caught one mistake in the final 20 seconds in the official gamebook. (They have Paulus coming out of the game, then somehow hitting 2 FTs...)

Devilsfan
01-28-2008, 12:18 AM
If you don't have the first eight minutes these numbers are meaningless. So I ask why bother.

Jumbo
01-28-2008, 12:24 AM
If you don't have the first eight minutes these numbers are meaningless. So I ask why bother.

Read. I posted everything in English:

Thanks to Georgia Tech, Virginia and the good folks at FSN, my I missed the first few minutes of charting. Thankfully, the gamebook shows all subs, so all is well. The numbers are here. And they're interesting.

I got the subs from the gamebook. So all the numbers are valid. Thanks for the contribution, though.

Taco
01-28-2008, 12:55 AM
Why isn't K playing David McClure at all 5 positions for 40 minutes a game; I'm incensed.

Troublemaker
01-28-2008, 01:12 AM
I'd love to know if I have my lineups right for those eight minutes. I've already caught one mistake in the final 20 seconds in the official gamebook. (They have Paulus coming out of the game, then somehow hitting 2 FTs...)

Here's what I have:
20:00 - starting lineup = -5 (MD 13, Duke 8 when they sub)
16:16 - JS in for GH, DM in for LT = +5 (MD 14, Duke 14)
14:42 - NS in for GP = 0 (MD 16, Duke 16)
13:22 - TK in for KS, GH in for DN = +2 (Duke 18, MD 16)
12:54 - LT in for DM = -2 (MD 18, Duke 18)
11:22 - starting lineup

Jumbo
01-28-2008, 02:01 AM
Here's what I have:
20:00 - starting lineup = -5 (MD 13, Duke 8 when they sub)
16:16 - JS in for GH, DM in for LT = +5 (MD 14, Duke 14)
14:42 - NS in for GP = 0 (MD 16, Duke 16)
13:22 - TK in for KS, GH in for DN = +2 (Duke 18, MD 16)
12:54 - LT in for DM = -2 (MD 18, Duke 18)
11:22 - starting lineup

The gamebook had the same thing, so we're all set. Thanks for the help, TM.

jzp5079
01-28-2008, 08:13 AM
Call Mr. Hollinger, tell him to drop his PER rating and just keep +/- 's instead.

I hope you realize that this rating system is SEVERELY flawed when McClure has an 8 point lead on Singler.

CDu
01-28-2008, 08:15 AM
Sounds like most of it is captured in the conversation.

- McClure's plus/minus is almost certainly a function of being on the floor at the right time, as he did next to nothing in the game and played very few minutes.
- Paulus benefits slightly from being on the floor at the end when they're fouling. And apparently he had a good first few minutes. He was also on the floor early in the second half when Nelson starting going nutty and Duke made up a large portion of the deficit. So again, plus/minus for a single game can be misleading (as you said).

In this case, it appears that the plus/minus for Paulus is buoyed by both good play and good timing. The stat line should make it obvious that both Smith and Paulus had solid games. Look at the points-per-shot, Smith's rebounds, and Paulus' steals. I'd say they were both really effective, and that in this case the plus/minus is very misleading to suggest that Paulus was substantively more effective than Smith. So I don't think that the plus/minus stat last night is particularly instructive in the Paulus/Smith debate. Fruit for discussion? Certainly (which is why we're discussing). Instructive? Probably not so much.

Surfsideron
01-28-2008, 08:18 AM
"Why are we playing Smith at all!!
Paulus should be playing 40 minutes a game!
And anyone who thinks Smith should play any minutes clearly doesn't have the best interests of the team in mind..."


What! I hope you are kidding. The more minutes Smith can get the better. He's our best chance aginst Lawson. I love Paulus, but Smith needs to be on the floor as much as possible.

Exiled_Devil
01-28-2008, 10:12 AM
Correct me if I 'm wrong (Jumbo - you are the one who would know on this) but I thought that the +/- could give us some insight into players that 'make the team better' in otherwise intangible ways. People are dismissing Greg's performance as being in for free throws at the end, or Dave's for being luck in his timing.
However, another way to look at this that Greg didn't turn over in the final free throw-intensive period, and provided a sound outlet to his teammates. Dave provided a presence that could pull the defense out and screen, and also played sound defense - isn't the point of the +/- to overcome the fixation on individual performance and allow us to see things we don't pay attention to as much?

Exiled

Jumbo
01-28-2008, 11:36 AM
Call Mr. Hollinger, tell him to drop his PER rating and just keep +/- 's instead.

I hope you realize that this rating system is SEVERELY flawed when McClure has an 8 point lead on Singler.

Um, yeah. Like, maybe when I wrote (http://www.dukebasketballreport.com/forums/showpost.php?p=91247&postcount=1), "Again, I'll start with the usual caveat that single-game plus/minus stats can be very misleading."

Or how how about when I wrote (http://www.dukebasketballreport.com/forums/showpost.php?p=91247&postcount=1) "But there was some counterintuitive stuff as well:
-Dave McClure didn't seem to do much tonight. But in only eight minutes, he was +8. That might be a case of being in the game at the right time for a run. But it's interesting."

BTW, Hollinger uses plus/minus. So do the more advanced NBA sites, like 82games.com. And every team uses it as one of many, many statistical measures for evaluating players.

Jumbo
01-28-2008, 11:41 AM
Correct me if I 'm wrong (Jumbo - you are the one who would know on this) but I thought that the +/- could give us some insight into players that 'make the team better' in otherwise intangible ways. People are dismissing Greg's performance as being in for free throws at the end, or Dave's for being luck in his timing.
However, another way to look at this that Greg didn't turn over in the final free throw-intensive period, and provided a sound outlet to his teammates. Dave provided a presence that could pull the defense out and screen, and also played sound defense - isn't the point of the +/- to overcome the fixation on individual performance and allow us to see things we don't pay attention to as much?

Exiled

I'd say the best way to describe plus/minus in a single-game setting is "a piece of the puzzle." You're right -- it should give us a window into intangible contributions. At the same time, it is flawed in some ways. Five guys' plus/minus increased by two points when Nelson stole the inbounds pass early in the second half and scored. That was a solo effort. Everyone's plus/minus benefits when Maryland misses the front end of a one-and-one. And so forth.

You're right to assert that if Duke outscored Maryland by eight with McClure in the game, he couldn't have been hurting the team that much, and very well might have been doing some subtle things to help the other guys. But plus/minus (unless there are glaring one-game blips) is meant to be more of a long-term stat. That said, I still enjoy seeing the numbers each game and, in some cases, I do think they mirror players' contributions. Lance Thomas is the best example I can think of -- his plus/minus has been excellent in each of the last three games, and we can all agree that they've probably been his best three games of the season.

Jumbo
01-28-2008, 11:42 AM
"Why are we playing Smith at all!!
Paulus should be playing 40 minutes a game!
And anyone who thinks Smith should play any minutes clearly doesn't have the best interests of the team in mind..."


What! I hope you are kidding. The more minutes Smith can get the better. He's our best chance aginst Lawson. I love Paulus, but Smith needs to be on the floor as much as possible.

The poster was kidding/making fun of the whole Paulus/Smith debate.

detule
01-28-2008, 12:13 PM
Thanks Jumbo for taking the time.

I always find the plus/minus stats interesting.

CDu
01-28-2008, 01:32 PM
I'd say the best way to describe plus/minus in a single-game setting is "a piece of the puzzle." You're right -- it should give us a window into intangible contributions. At the same time, it is flawed in some ways. Five guys' plus/minus increased by two points when Nelson stole the inbounds pass early in the second half and scored. That was a solo effort. Everyone's plus/minus benefits when Maryland misses the front end of a one-and-one. And so forth.

You're right to assert that if Duke outscored Maryland by eight with McClure in the game, he couldn't have been hurting the team that much, and very well might have been doing some subtle things to help the other guys. But plus/minus (unless there are glaring one-game blips) is meant to be more of a long-term stat. That said, I still enjoy seeing the numbers each game and, in some cases, I do think they mirror players' contributions. Lance Thomas is the best example I can think of -- his plus/minus has been excellent in each of the last three games, and we can all agree that they've probably been his best three games of the season.

Well said. Plus/minus stats are still interesting and worth discussion, but making conclusions based on them is debatable. The key is to be VERY careful not to over-analyze single-game plus/minus.

One thing I'm concerned with with regard to plus/minus is the fact that it is still largely driven by the success of the team. As such, a player who plays a lot of minutes on a bad team is likely to have a lower (more negative) plus/minus than a teammate who plays fewer minutes. It's possible that this can be accounted for in the plus/minus per 40 stat, but really all that is going to do is create a larger standard error for the lesser player (due to fewer minutes). It could very well be that the lesser player, by nature of playing fewer minutes on the bad team, may have a better (less negative) plus/minus per 40 minutes. I'd certainly not argue that such a player is the better or more valuable player, but the plus/minus might well suggest such a thing.

And of course plus/minus is useless in comparing players across teams, as a star player who plays for a sub-.500 team is almost certain to have a negative plus/minus.

Saratoga2
01-28-2008, 01:36 PM
I'd say the best way to describe plus/minus in a single-game setting is "a piece of the puzzle." You're right -- it should give us a window into intangible contributions. At the same time, it is flawed in some ways. Five guys' plus/minus increased by two points when Nelson stole the inbounds pass early in the second half and scored. That was a solo effort. Everyone's plus/minus benefits when Maryland misses the front end of a one-and-one. And so forth.

You're right to assert that if Duke outscored Maryland by eight with McClure in the game, he couldn't have been hurting the team that much, and very well might have been doing some subtle things to help the other guys. But plus/minus (unless there are glaring one-game blips) is meant to be more of a long-term stat. That said, I still enjoy seeing the numbers each game and, in some cases, I do think they mirror players' contributions. Lance Thomas is the best example I can think of -- his plus/minus has been excellent in each of the last three games, and we can all agree that they've probably been his best three games of the season.

I think that what a player contributes has also got to be adjusted for the talent that is playing against him. Singlar was in a lot when Osby and Gist were in for instance. Osby seemed to tire out and Singlar got a rest then as well. Clearly if a player is in when the opponents best are out, that player is likely to have a better (+/-). I still appreciate the work that Jumbo is doing to compile this list and think it would be very difficult to adjust the numbers for opponent quality.

Singlar also fouled out before we had the final few minutes when Duke piled on the points with free throws at the end of the game. Just pointing out that these numbers are very tricky to apply but probably have far more meaning when looked at over the course of many games.

CDu
01-28-2008, 01:40 PM
I think that what a player contributes has also got to be adjusted for the talent that is playing against him. Singlar was in a lot when Osby and Gist were in for instance. Osby seemed to tire out and Singlar got a rest then as well. Clearly if a player is in when the opponents best are out, that player is likely to have a better (+/-). I still appreciate the work that Jumbo is doing to compile this list and think it would be very difficult to adjust the numbers for opponent quality.

Singlar also fouled out before we had the final few minutes when Duke piled on the points with free throws at the end of the game. Just pointing out that these numbers are very tricky to apply but probably have far more meaning when looked at over the course of many games.

Very true. A reserve on a team with 7-8 guys who could all start (like this Duke team) is likely to have a slight advantage over the starters, who will generally (a) face better players or (b) face the best players when they are more fresh. And in the case of a guy who's in for a specific matchup, that can't be captured either. Plus/minus useful and interesting, but clearly not infallible.

jzp5079
01-28-2008, 01:43 PM
Um, yeah. Like, maybe when I wrote (http://www.dukebasketballreport.com/forums/showpost.php?p=91247&postcount=1), "Again, I'll start with the usual caveat that single-game plus/minus stats can be very misleading."

Or how how about when I wrote (http://www.dukebasketballreport.com/forums/showpost.php?p=91247&postcount=1) "But there was some counterintuitive stuff as well:
-Dave McClure didn't seem to do much tonight. But in only eight minutes, he was +8. That might be a case of being in the game at the right time for a run. But it's interesting."

BTW, Hollinger uses plus/minus. So do the more advanced NBA sites, like 82games.com. And every team uses it as one of many, many statistical measures for evaluating players.


With all due respect I meant my comment to be read with sarcasm in mind...

Jumbo
01-28-2008, 02:15 PM
I think that what a player contributes has also got to be adjusted for the talent that is playing against him. Singlar was in a lot when Osby and Gist were in for instance. Osby seemed to tire out and Singlar got a rest then as well. Clearly if a player is in when the opponents best are out, that player is likely to have a better (+/-). I still appreciate the work that Jumbo is doing to compile this list and think it would be very difficult to adjust the numbers for opponent quality.

Singlar also fouled out before we had the final few minutes when Duke piled on the points with free throws at the end of the game. Just pointing out that these numbers are very tricky to apply but probably have far more meaning when looked at over the course of many games.

SinglEr!!!

dukestheheat
01-28-2008, 06:39 PM
Jumbo,

I love these statistical posts and they tell me one thing, at least: a number on a page doesn't tell the full, complete story of what (went down) on the landscape. I see this in my line of work, constantly. 'This account is down, this account is up....make sense of these numbers, why this way, why not that way..'. Stuff like that.

Though Nolan only garners a +3, clearly Duke flowed very well when he was in the game; we truly benefitted in more ways than just what that stat, +3, could tell anyone who looked at it.

Again, I do appreciate this analysis and I look forward to it, but there are many times where we need to read between the lines and I think the point with Smith applies here.

dth.

duketaylor
01-30-2008, 09:53 AM
Thanks, Jumbo. I love this statistical breakdown, even if a one-game analysis reveals less than a larger body of work. It's still apparent Duke plays it's best ball with Greg in the game. I'd like to see more of Nolan and Greg in together and see what these numbers would reflect.