Man of the Match vs. NC State

[4decadedukie]

Scheyer played a great game, and he is a WONDERFUL young man; not just a stellar athlete, but the kind of undergraduate I am proud -- without constraints -- to have at Duke. I was surprised and somewhat taken aback a few weeks ago when a DBR poll did not identify Jon as a future NBA presence; I disagree. He probably has as fine an understanding of basketball details and nuances as any of our terrific players have had in many years, and I will be unsurprised to see him in a critical "front office" NBA role (including coaching).

With all this said, I voted for our Captain (but wish I could have cast two ballots). When we needed it most, DeMarcus came through (including at the charity stripe).

[Jumbo]

No argument from me that Scheyer played a great game. In fact, as I said, I voted for him. But it just seems to me that you say that plus/minus stats are misleading only when they don't verify what you felt from the game. But when they support your concepts, they are revealing?

In my opinion, the stat is just too sensitive to random variation to be of any use in a single game situation. Maybe Nelson's plus/minus was due to bad play. Or, maybe Nelson happened to be on the floor when some of his teammates had their worst moments. Or maybe he happened to be in during a stretch in which State caught fire. There are just so many possibilities.

I happen to think both players played very well. I agree that Scheyer played better. But he didn't play 21 net points better. Nelson had a terrific game as well.

CDu, I'm not sure why you seem so aggressive in attacking my opinion. Let me take a step back and try to explain a little better.

Single-game plus/minus numbers can be both revealing and misleading. Much of that has to do with minutes played. A guy who plays 35 minutes will have a plus/minus rating that is less subject to random fluctuations than someone who plays 10 minutes. That much is obvious. So, in this case, Nelson and Scheyer both played substantial minutes. It's a lot different than the people who were advocating Taylor King earlier in the season based on limited playing time.

I'm not saying Scheyer was 21 points better than Nelson. But his plus/minus advantage wasn't the result of his simply being on the floor for one pivotal run. There were several points during the game where Duke gained points with Nelson on the bench and several points where Duke lost points when Scheyer went out. For instance:
-Duke was down four when Scheyer replaced Nelson. Duke was down one when Nelson came back in.
-Duke was down three when Nelson went back out again. When he came back, Duke was down one. (Scheyer was in this whole time.)
-Duke was up one when Scheyer finally came out again. By the time he came back in, Duke was down four.
-In the second half, Duke went on a quick 5-0 run to cut the deficit from 10 to 5 when Henderson replaced Nelson.
There were a few other examples. But the point was that, for whatever reason, Duke functioned better on multiple occasions with Scheyer in the game and with Nelson out. Obviously, Nelson was awesome at the end of the game and equally obviously, I have no desire for Nelson's minutes to get cut or anything like that. But Markie was not a consistent force against State (the plus/minus numbers support that hypothesis) and Scheyer was (again, the numbers back that up). That's all.

[Jumbo]

I was surprised and somewhat taken aback a few weeks ago when a DBR poll did not identify Jon as a future NBA presence...

What poll are you talking about. I've been singing this tune for a while, but everything I've heard tells me that Jon will indeed play in the NBA.

[CDu]

CDu, I'm not sure why you seem so aggressive in attacking my opinion. Let me take a step back and try to explain a little better.

Single-game plus/minus numbers can be both revealing and misleading. Much of that has to do with minutes played. A guy who plays 35 minutes will have a plus/minus rating that is less subject to random fluctuations than someone who plays 10 minutes. That much is obvious. So, in this case, Nelson and Scheyer both played substantial minutes. It's a lot different than the people who were advocating Taylor King earlier in the season based on limited playing time.

I'm not saying Scheyer was 21 points better than Nelson. But his plus/minus advantage wasn't the result of his simply being on the floor for one pivotal run. There were several points during the game where Duke gained points with Nelson on the bench and several points where Duke lost points when Scheyer went out. For instance:
-Duke was down four when Scheyer replaced Nelson. Duke was down one when Nelson came back in.
-Duke was down three when Nelson went back out again. When he came back, Duke was down one. (Scheyer was in this whole time.)
-Duke was up one when Scheyer finally came out again. By the time he came back in, Duke was down four.
-In the second half, Duke went on a quick 5-0 run to cut the deficit from 10 to 5 when Henderson replaced Nelson.
There were a few other examples. But the point was that, for whatever reason, Duke functioned better on multiple occasions with Scheyer in the game and with Nelson out. Obviously, Nelson was awesome at the end of the game and equally obviously, I have no desire for Nelson's minutes to get cut or anything like that. But Markie was not a consistent force against State (the plus/minus numbers support that hypothesis) and Scheyer was (again, the numbers back that up). That's all.

My disagreement isn't with your opinion that Scheyer was the player of the game. We share that opinion. My disagreement is with your use of plus/minus as evidence of him being better. Sorry for sounding aggressive about it. It's just that the use of single-game plus/minus is a pet peeve of mine.

I think you're understating the problems of single-game plus/minus by stating that more minutes means less risk of random fluctuation. Sure, there's less risk, but there's still a LOT of risk - too much to take the numbers too seriously, in my opinion. There just aren't enough minutes in a single game to control for all the variables that go into plus/minus. So much can happen in a single game while a player is on the floor (or not on the floor) that is out of that player's control. Hence, I don't think single-game plus/minus is useful for any analytical purposes. It's just too difficult (even for the very educated basketball mind) to weed through all of the variables and accurately control for them.

As such, I think the post comparing the two players' plus/minus values was irrelevant to the comparison. That's all. Again, I think plus/minus has the potential for being a very interesting and useful tool. But I think it's useful over a 25+ game season - not over a single game. Over a longer stretch, it's reasonable to assume that the random uncontrollable events that happen from game to game sort of even out, allowing us to get a picture of the player's true worth to the team. But a single game plus/minus doesn't provide that opportunity. Is it possible for a single-game plus/minus to exactly match up with a player's true value in that game? Sure, but I'd say it's VERY unlikely.

Sorry again for sounding antagonistic. It's just a hot-button issue with me. We agree that Scheyer was the player of the game. We just disagree on the value of single-game plus/minus.

[Jumbo]

My disagreement isn't with your opinion that Scheyer was the player of the game. We share that opinion. My disagreement is with your use of plus/minus as evidence of him being better. Sorry for sounding aggressive about it. It's just that the use of single-game plus/minus is a pet peeve of mine.

I think you're understating the problems of single-game plus/minus by stating that more minutes means less risk of random fluctuation. Sure, there's less risk, but there's still a LOT of risk - too much to take the numbers too seriously, in my opinion. There just aren't enough minutes in a single game to control for all the variables that go into plus/minus. So much can happen in a single game while a player is on the floor (or not on the floor) that is out of that player's control. Hence, I don't think single-game plus/minus is useful for any analytical purposes. It's just too difficult (even for the very educated basketball mind) to weed through all of the variables and accurately control for them.

As such, I think the post comparing the two players' plus/minus values was irrelevant to the comparison. That's all. Again, I think plus/minus has the potential for being a very interesting and useful tool. But I think it's useful over a 25+ game season - not over a single game. Over a longer stretch, it's reasonable to assume that the random uncontrollable events that happen from game to game sort of even out, allowing us to get a picture of the player's true worth to the team. But a single game plus/minus doesn't provide that opportunity. Is it possible for a single-game plus/minus to exactly match up with a player's true value in that game? Sure, but I'd say it's VERY unlikely.

Sorry again for sounding antagonistic. It's just a hot-button issue with me. We agree that Scheyer was the player of the game. We just disagree on the value of single-game plus/minus.

Right, I was focusing on the plus/minus issue as well. I think what you aren't recognizing is that, if this were research, I wouldn't be using plus/minus as a primary source. It's purely supporting evidence to back up a point established in other ways. In this case, I observed that Duke played better with Scheyer in the game from the first second he stepped on the court, that he played consistently well throughout the game, and that DeMarcus Nelson had a rough first half, including some silly fouls that put him on the bench and State at the line. After noticing those trends, and already deciding that Scheyer was my MOTM, I added up the plus/minus numbers. Sure enough, they matched my hypothesis.

All I'm saying is that while it's silly to use plus/minus stats for a single game as the primary tool for evaluation, it's also silly to dismiss them out of hand because a single game is a small sample size. In some cases, those stats are still telling.

[CDu]

Right, I was focusing on the plus/minus issue as well. I think what you aren't recognizing is that, if this were research, I wouldn't be using plus/minus as a primary source. It's purely supporting evidence to back up a point established in other ways. In this case, I observed that Duke played better with Scheyer in the game from the first second he stepped on the court, that he played consistently well throughout the game, and that DeMarcus Nelson had a rough first half, including some silly fouls that put him on the bench and State at the line. After noticing those trends, and already deciding that Scheyer was my MOTM, I added up the plus/minus numbers. Sure enough, they matched my hypothesis.

All I'm saying is that while it's silly to use plus/minus stats for a single game as the primary tool for evaluation, it's also silly to dismiss them out of hand because a single game is a small sample size. In some cases, those stats are still telling.

And my point is that it seems like people (not just you, so don't please don't take it personally) assume plus/minus is misleading when it doesn't support their hypothesis and telling when it does support their hypothesis. That's a completely natural stance, but without analyzing every single possession it lends itself to being a faulty conclusion.

Your argument that Scheyer was man of the match stood strong without the plus/minus argument in, my opinion. It is also my opinion that the plus/minus in this case may very well be exaggerating the difference between the two players for this game.

[Chitowndevil]

I give it to Demarcus, because he went 8-9 from the line, where he hasn't exactly been automatic, and more importantly scored 7 of Duke's points during an 11-3 run to close out the game, including 4-4 from the stripe. As K said after the game he just has an amazing will to win. As nice as his season stat line looks I don't think it even comes close to capturing his value to the team this season.

But really there's no wrong answer, as without Scheyer that 11-3 run doesn't matter in the outcome.

[kidslow]

DeMarcus stepped up and put the team on his shoulders to close that game out. Without his leadership example, there's no way Duke would have won. I believe he scored most of his points during the final 10 minute crunch, when they mattered most. After Lance missed those free throws and the lead had been stretched back to near double-digits, the announcers were faulting Duke for sitting back and taking 3s instead of taking the ball at the defense. Nelson then took the ball right at the hole as if on queue and drew the foul. I can easily recall at least 10 of his points coming in the final 8 minutes of the game. That 3-pointer when he stepped back over the line was clutch. I don't know how anyone can lobby for another MotM here, because Markie took charge and willed the win yesterday. It was evident to me as an observer that he made all the difference between a W and a L yesterday. Stats (+/- etc) never accurately reflect small but meaningful contributions nor the intangibles like leadership and such.