Where does the team go from here? A look into recent results for clues...

[Kedsy]

I would suggest there is a real risk of small sample size effect here. Young has shot 67% over the past 3 games, which is basically right in line with his norms. He had two subpar shooting games the two games prior to that. So I am not sure how much they figured out vs him just having two bad games and course correcting since.

As for Lively, that awesome rebound rate is mainly driven by one awesome game against Miami.

That said, if Lively has more games like Miami moving forward, his minutes will go up.

Maybe. Ryan has only been taking four shots a game, recently. Other than BC, he hasn’t passed much, and he hasn’t been going to the free throw line. Even if his efficiency hasn’t suffered recently (which I think it has, but I’m not sure to what extent), that’s not so much of an offensive impact. Dereck has taken more shots in recent games and made just one fewer basket, in 65% of the minutes. Since we all agree Lively has a much bigger defensive impact, at Young’s recent usage rate, perhaps his offensive efficiency doesn’t justify playing more minutes than Lively. Just a thought, anyway.

[CDu]

Our teams historically have performed great in November and December and struggled in January, even in years where there are no major roster changes. I typically chalk it up to elite freshmen having an advantage out of the gate when games are more about raw talent, opponents adjusting to our play and us failing to counter those adjustments, more true road games (Wake was our first this season), and the grind of conference play wearing down younger players. I think our numbers were naturally going to come down in January no matter what, irrespective of Lively's minutes. I also don't think his minutes increased THAT much across the two windows you provided. I'm not a stats guy so I don't have proof, but again I think our numbers are more about the team as a whole and less about Lively playing 15 minutes instead of 10 or whatever.

ETA: Also Roach missed three games during the "bad" stretch and was clearly not 100% leading up to his missed games. That's clearly going to impact our numbers.

Yeah, agreed (obviously). As I mentioned in a previous post, OE and DE are literally just translations of points per possession. So efficiency margin (OE-DE) is literally plus/minus translated to a per 100 possessions rate. AdjOE and AdjDE are just scaling factors that take into account quality of opponent and location (in some formulations). So if a player has a better plus/minus, that means the team had a better efficiency margin (OE-DE) when he was in, and thus a better Adjusted efficiency margin.

It is true that the team has played worse over the games beginning with Wake. But that is almost certainly a function of the guard play cratering over most of this stretch, and totally independent of the bigs. Roach being hurt and then out is a big reason.

So most likely the team adjusted efficiency numbers have dropped for both guys over this stretch. But regardless, the plus/minus numbers have favored Young, which means the team’s efficiency margin and adjusted efficiency margin has been better with Young on the floor than when Lively has been on the floor.

Hopefully Lively can continue to improve though. If he does, he could well overtake Young and see his minutes increase accordingly. It is hard to imagine Young can play any better than he has, whereas Lively has plenty of room for improvement. If he can start being a more consistent impact player, he will play more.

[CDu]

Maybe. Ryan has only been taking four shots a game, recently. Other than BC, he hasn’t passed much, and he hasn’t been going to the free throw line. Even if his efficiency hasn’t suffered recently (which I think it has, but I’m not sure to what extent), that’s not so much of an offensive impact. Dereck has taken more shots in recent games and made just one fewer basket, in 65% of the minutes. Since we all agree Lively has a much bigger defensive impact, at Young’s recent usage rate, perhaps his offensive efficiency doesn’t justify playing more minutes than Lively. Just a thought, anyway.

His 4 shots per game recently isn’t really any different than his season average (4.35 per game). And he has had similar 2-3 game stretches of low volume and/or low offensive impact this year. He shot just 50% on 4 shots per game against Xavier and Purdue. He shot “just” 60% on 2.5 attempts per game against Iowa and UMES. Conversely, in this 3-game stretch, in two games he shot 70% on 5 attempts per game. So I am not sure I buy the “teams have figured him out” theory so much as he just isn’t a high usage player and one mediocre game will thus stand out. In this case, offensively that is the Miami game, and rebounding it is the Clemson game.

But again, simply looking at plus/minus and the individual efficiency metrics, it is hard to argue that Lively should be playing as much as Young. Hopefully Lively will step up though and change that story. Because I agree that our best hope of being a really good team probably relies on Lively figuring it out.

[DukieInBrasil]

They could star in an upbeat youth about town show: Young & Lively
They could really help the team out if they could tap into that spry energy of being Young and Lively
They could wear the opponents down playing 40 minutes per game with some Young 'n Lively energy

just trying to Lively up the energy in this thread...

[trinity79]

They could star in an upbeat youth about town show: Young & Lively
They could really help the team out if they could tap into that spry energy of being Young and Lively
They could wear the opponents down playing 40 minutes per game with some Young 'n Lively energy

just trying to Lively up the energy in this thread...

Perhaps this could become the team's theme song for this season:

"You're gonna lively up yourself (Lively up yourself)
And don't say: "No"
You lively up yourself (Lively up yourself)
Big daddy said so, y'all
You lively up yourself (Lively up yourself)
And don't be no drag
You lively up yourself (Lively up yourself)
'Cause reggae is another bag"

--- Bob Marley and the Wailers, track 1 from African Herbsman

[Billy Dat]

They could star in an upbeat youth about town show: Young & Lively
They could really help the team out if they could tap into that spry energy of being Young and Lively
They could wear the opponents down playing 40 minutes per game with some Young 'n Lively energy

just trying to Lively up the energy in this thread...

Perhaps this could become the team's theme song for this season:

"You're gonna lively up yourself (Lively up yourself)
And don't say: "No"
You lively up yourself (Lively up yourself)
Big daddy said so, y'all
You lively up yourself (Lively up yourself)
And don't be no drag
You lively up yourself (Lively up yourself)
'Cause reggae is another bag"

--- Bob Marley and the Wailers, track 1 from African Herbsman

First off, I want to praise the discourse on this thread. Kudos to CDu for starting it with such a thoughtful examination and for the rest of you for pushing it along.

Second, I think this proposed TV show, which is a great idea, was actually already produced as a motion picture, and the casting of Young and Lively was pretty good. But, they got the sport wrong and didn't realize that the prior coach had already retired. They were close, the offensive coordinator could've played Coach Scheyer.

Any Given Sunday.jpg

[Kedsy]

But regardless, the plus/minus numbers have favored Young, which means the team’s efficiency margin and adjusted efficiency margin has been better with Young on the floor than when Lively has been on the floor.

But again, simply looking at plus/minus and the individual efficiency metrics, it is hard to argue that Lively should be playing as much as Young.

Per the official box scores, in our last six games Dereck Lively's aggregate plus/minus has been +9 while Ryan Young's aggregate plus/minus has been -23.

I know how flawed plus/minus is, but for some reason you keep citing it. Since my theory is teams have lately figured Ryan Young out, we can't go back too far, but at least in our "n" most recent games (whether that's three games, four games, five games, or six games), the plus/minus numbers have not favored Young.

[CDu]

Per the official box scores, in our last six games Dereck Lively's aggregate plus/minus has been +9 while Ryan Young's aggregate plus/minus has been -23.

I know how flawed plus/minus is, but for some reason you keep citing it. Since my theory is teams have lately figured Ryan Young out, we can't go back too far, but at least in our "n" most recent games (whether that's three games, four games, five games, or six games), the plus/minus numbers have not favored Young.

I was citing season-long numbers, as that's the only thing that remotely comes close to a reasonable sample size (even then it's woefully insufficient, but it's way better than a game-level or 5-10 game-level sample). That difference over 6 games basically comes down to two games: NC State (difference of -17) and most importantly Pitt (difference of -27). In the other 4 games, Young had a better +/- in 3: +8 difference vs Miami, +5 vs VT, and +17 vs BC. He had a -18 difference against Clemson, which basically offset the BC game. And worth noting that you stopped short of including the FSU game, which was a +35 for Young. So over the last 7 games, Young has the edge in +/- by a margin of +3 (+17 for Young, +14 for Lively).

So I don't think the data supports the "they've figured out Young" theory either, but rather just that individual game +/- are just incredibly noisy. And as such, looking at +/- over any small segment of games is going to be meaningless. For example, one of Young's worst differentials came against NC State, but that game had very little to do with the bigs as it was our guard play that buried us (I did a breakdown of the possessions in that game in an earlier thread). Conversely, one of Lively's worst was ironically against Miami, despite that being his best individual performance and one of Young's weaker ones. I'm not ready to say that Young has been figured out on the back of essentially a single game (the Pitt game accounts for almost the entire +/- difference over the past 6 games), especially as we had better numbers with Young on the floor in each of the last two games.

That's why I have used the aggregate +/- for the season: because looking at small samples of games renders +/- way too noisy.

[Kedsy]

I was citing season-long numbers, as that's the only thing that remotely comes close to a reasonable sample size (even then it's woefully insufficient, but it's way better than a game-level or 5-10 game-level sample). That difference over 6 games basically comes down to two games: NC State (difference of -17) and most importantly Pitt (difference of -27). In the other 4 games, Young had a better +/- in 3: +8 difference vs Miami, +5 vs VT, and +17 vs BC. He had a -18 difference against Clemson, which basically offset the BC game. And worth noting that you stopped short of including the FSU game, which was a +35 for Young. So over the last 7 games, Young has the edge in +/- by a margin of +3.

So I don't think the data supports the "they've figured out Young" theory either, but rather just that individual game +/- are just incredibly noisy. And as such, looking at +/- over any small segment of games is going to be meaningless. For example, one of Young's worst differentials came against NC State, but that game had very little to do with the bigs as it was our guard play that buried us (I did a breakdown of the possessions in that game in an earlier thread). Conversely, one of Lively's worst was ironically against Miami, despite that being his best individual performance and one of Young's weaker ones. I'm not ready to say that Young has been figured out on the back of essentially a single game (the Pitt game accounts for almost the entire +/- difference over the past 6 games), especially as we had better numbers with Young on the floor in two of the last three games.

That's why I have used the aggregate +/- for the season: because looking at small samples of games renders +/- way too noisy.

How can you use a season-long number to counter a theory that teams have figured Young out lately? FSU didn't double Young every time he touched the ball. Our last several opponents did. Have you not noticed that? He's obviously going to be less effective against a double-team unless he makes great passes out of it (which he did against BC but not really any other game).

In any event, my point is that when a low-usage, high efficiency player becomes less efficient (which Young has, compared to earlier in the season), his offensive advantage becomes much less valuable. If you combine that with (a) Young's defense is miles worse than Lively's; and (b) Lively's been contributing more on offense in the last four games (at least more than he used to), I think a very good case can be made that Lively's minutes should increase at the expense of Young. Citing season-long plus/minus numbers isn't much of a counter-argument.

[CDu]

How can you use a season-long number to counter a theory that teams have figured Young out lately? FSU didn't double Young every time he touched the ball. Our last several opponents did. Have you not noticed that? He's obviously going to be less effective against a double-team unless he makes great passes out of it (which he did against BC but not really any other game).

I didn't use a season-long number to counter that theory. I said that the fact that Young had the better +/- in the last 2 games and in 3 of the last 6 is what counters the theory. I don't think one game 4 games ago is sufficient to support that theory, especially since +/- is so noisy.

So my point was that we simply don't have any sufficient evidence to suggest that the team is playing better with Lively in the game than with Young. Certainly not at the season-long level, and not really at the shorter-term level.

In any event, my point is that when a low-usage, high efficiency player becomes less efficient (which Young has, compared to earlier in the season), his offensive advantage becomes much less valuable. If you combine that with (a) Young's defense is miles worse than Lively's; and (b) Lively's been contributing more on offense in the last four games (at least more than he used to), I think a very good case can be made that Lively's minutes should increase at the expense of Young. Citing season-long plus/minus numbers isn't much of a counter-argument.

But he hasn't become less efficient. His ORtg for the season is 137.1. His ORtg in conference play is 144.3. And over his past 3 games, he's only one make off his average, so it isn't like he has been in this massive efficiency drought recently.

And again, I'm only citing season-long numbers because +/- is meaningless over small sample size. I'm not using that to counter your theory. I'm just saying that your theory isn't really supported by any evidence, and that the only evidence that is even remotely worthwhile says Young has been better. They are two complementary thoughts, not both being used to address the same point.

[Kedsy]

I didn't use a season-long number to counter that theory.

Is that right?

I think opponents might be on to Ryan Young.

Young has a better plus/minus than Lively. Which means that the team has had better results when he has been on the floor.

But regardless, the plus/minus numbers have favored Young, which means the team’s efficiency margin and adjusted efficiency margin has been better with Young on the floor than when Lively has been on the floor.

So I am not sure I buy the “teams have figured him out” theory so much as he just isn’t a high usage player and one mediocre game will thus stand out. In this case, offensively that is the Miami game, and rebounding it is the Clemson game. But again, simply looking at plus/minus and the individual efficiency metrics, it is hard to argue that Lively should be playing as much as Young.

Per the official box scores, in our last six games Dereck Lively's aggregate plus/minus has been +9 while Ryan Young's aggregate plus/minus has been -23.

I was citing season-long numbers

[Kedsy]

And again, I'm only citing season-long numbers because +/- is meaningless over small sample size.

I have seen countless posts by you saying plus/minus is meaningless, period. Even over a full season. With that in mind, I'm flabbergasted to see you using plus/minus to support your position now.

I'm not using that to counter your theory. I'm just saying that your theory isn't really supported by any evidence, and that the only evidence that is even remotely worthwhile says Young has been better. They are two complementary thoughts, not both being used to address the same point.

I think this is a situation where there's no reliable statistical evidence, either way. I think that if you're debating a theory about the past few games, there is no possibility of a decent sample size. But I also think it's clear that opponents are doubling Young a lot more than they used to (though of course, there's no available stat for that). I think it's obvious that a guy who's double-teamed is going to be less efficient and less effective than the same guy when he isn't double-teamed. I think it follows that Young's offensive value to the team has decreased when he's double-teamed. Which logically makes Lively's defensive superiority more of a factor and Young's offensive superiority less of a factor.

I'm not sure why you can't admit to all that, nor do I understand why you're hyping season-long plus/minus as "the only evidence that is even remotely worthwhile," when in the past you've said that sort of evidence is not remotely worthwhile.

[CDu]

I have seen countless posts by you saying plus/minus is meaningless, period. Even over a full season. With that in mind, I'm flabbergasted to see you using plus/minus to support your position now.

I definitely think +/- is meaningless, even over a full season. But azzefkram wasn't interested in discussing BPM (which clearly favors Young), so I was trying to be polite and use a framework that azzefkram might acknowledge. But I would absolutely prefer to be talking about things like BPM, and in fact that's where I started in this discussion a few steps before you jumped in.

I think this is a situation where there's no reliable statistical evidence, either way. I think that if you're debating a theory about the past few games, there is no possibility of a decent sample size.

I completely agree. I'm not saying your theory is wrong. I'm just saying there aren't data to support it. There are two separate discussions going on: (1) has the team been more successful on the floor with Young vs Lively? And (2) has that story flipped recently? The overwhelming consistent answer to (1) is "yes." I was using BPM to say that, and then switched to +/- begrudgingly to show that even the +/- (which are meaningless) don't support the idea that the team plays better with Lively in. That's what azzefkram and I were discussing.

You changed the discussion to talk about the second question. I don't think we have enough information to assess the second question.

I'm not sure why you can't admit to all that, nor do I understand why you're hyping season-long plus/minus as "the only evidence that is even remotely worthwhile," when in the past you've said that sort of evidence is not remotely worthwhile.

I'm saying that if one is going to use +/- at all, then season-long is the only +/- stat that is even remotely close to worthwhile. As I said above, I'd much rather be talking about this using BPM. I definitely believe that even season-long +/- is pretty worthless, but way less worthless than a single-game +/-. But the participants in the discussion refused to consider BPM, so I was taking the only other track possible.

And I am not saying your theory is wrong. I'm saying that we don't have any sufficient evidence to assess it at this point. It's certainly possible that your theory is right. I just wouldn't say the +/- is a strong argument for it, nor would I say his efficiency is a strong argument for it (at least not yet; that could eventually prove the case).

[CDu]

Is that right?

What a garbage, disingenuous post, Kedsy. Come on. Not even all of those were responses to you, and some came before you even joined the discussion.

I was using season-long plus/minus as an alternative explanation in my discussion with azzefkram.

You then jumped in well into the discussion with your 6-game sample, saying that it isn't true that +/- favors Young and introducing your "recent double-team" theory. I clarified that I was using season-long, where it is absolutely true that +/- favors Young (and noted that even over a 7-game sample, +/- favors Young). Not that I think +/- is a good measure, but azzefkram wasn't interested in individual stats so that's the best I could do.

I'm not sure why you even included a few of those comments, as they are totally irrelevant to the discussion of your theory (and as mentioned, weren't even responses to you). There was only comment in there that was a response to your theory. And it doesn't use season-long +/- as a counter to your theory.

So, yes, in summary, I am not using season-long +/- to counter your theory. I don't have a strong opinion one way or another about your theory. I just don't think +/- is a good way to test it (and I agree that there isn't a good way to test it), nor do I think the +/- evidence over the last 6 games even strongly supports the theory anyway. Totally separate thought from my use of the season-long +/-, which was for a conversation with another poster.

[Kedsy]

What a garbage, disingenuous post, Kedsy. Come on. Not even all of those were responses to you, and some came before you even joined the discussion.

Sorry. It seemed pretty clear to me you were using season-long plus/minus to counter my theory.

[CDu]

Sorry. It seemed pretty clear to me you were using season-long plus/minus to counter my theory.

Apparently looks can be deceiving.

I was using season-long +/- in this thread. But I was not using them to counter your theory.

[Kedsy]

I'd much rather be talking about this using BPM.

How possible is it to calculate BPM over the past four games? (That's when I noticed teams doubling Ryan more.) I've never tried to calculate that stat myself.

Per Torvik, I see Ryan's offensive BPM in the last four games were: 4.1, 7.3, -10.3, and 2.1; Not sure if averaging makes any sense, but that's an average of 0.8 (obviously skewed by the bad performance in the Miami game).

Ryan's defensive BPM in those games were: -3.0, 6.4, 5.8, 1.7; an average of 2.7. Does that suggest his defensive contributions on average outweighed his offensive contributions? (honest question) If so, I may have my doubts about the value of this stat (at least on the defensive side).

Dereck's oBPM in the last 4 games: -4.5, 1.8, 1.3, -0.1; an average of -.4.

Dereck's dBPM in the last 4 games: 6.6, 11.1, 3.2, -2.4; an average of 4.6.


Overall BPM:

Ryan: 1.1, 13.6, -4.5, 3.8; average 3.5

Dereck: 2.0, 12.8, 4.5, -2.6; average 4.2


Obviously a small sample, but any theory about recent games would necessarily involve a small sample.

[azzefkram]

I definitely think +/- is meaningless, even over a full season. But azzefkram wasn't interested in discussing BPM (which clearly favors Young), so I was trying to be polite and use a framework that azzefkram might acknowledge. But I would absolutely prefer to be talking about things like BPM, and in fact that's where I started in this discussion a few steps before you jumped in.



I completely agree. I'm not saying your theory is wrong. I'm just saying there aren't data to support it. There are two separate discussions going on: (1) has the team been more successful on the floor with Young vs Lively? And (2) has that story flipped recently? The overwhelming consistent answer to (1) is "yes." I was using BPM to say that, and then switched to +/- begrudgingly to show that even the +/- (which are meaningless) don't support the idea that the team plays better with Lively in. That's what azzefkram and I were discussing.

Here's a quote for you from Basket-reference.com.

"Box Plus/Minus is good at measuring offense and solid overall, but the defensive numbers in particular should not be considered definitive. Look at the defensive values as a guide, but don't hesitate to discount them when a player is well known as a good or bad defender."

Using BPM to compare Ryan and Dereck is putting your thumb on the scale. Plus BPM doesn't always favor Ryan. BPM versus top 100 teams favors Dereck. Given that 7 of Duke's remaining 11 games will be against teams in the top 100, maybe that stat should carry more weight.

I don't think you have any good evidence that the team performs better when Ryan is on the floor. That being said, I guess I don't have any good evidence that the team performs worse when Ryan is on the floor (though AdjOE and AdjDE as team stats would seem to be of more relevance than BPM (an individual stat)).

I appreciate your trying to educate me on this and most times you are spot on. I just don't think this is one of those times.

[CDu]

How possible is it to calculate BPM over the past four games? I've never tried to calculate that stat myself.

Per Torvik, I see Ryan's offensive BPM in the last four games were: 4.1, 7.3, -10.3, and 2.1; Not sure if averaging makes any sense, but that's an average of 0.8 (obviously skewed by the bad performance in the Miami game).

Ryan's defensive BPM in those games were: -3.0, 6.4, 5.8, 1.7; an average of 2.7. Does that suggest his defensive contributions on average outweighed his offensive contributions? (honest question) If so, I may have my doubts about the value of this stat (at least on the defensive side).

Dereck's oBPM in the last 4 games: -4.5, 1.8, 1.3, -0.1; an average of -.4.

Dereck's dBPM in the last 4 games: 6.6, 11.1, 3.2, -2.4; an average of 4.6.


Overall BPM:

Ryan: 1.1, 13.6, -4.5, 3.8; average 3.5

Dereck: 2.0, 12.8, 4.5, -2.6; average 4.2


Obviously a small sample, but any theory about recent games would necessarily involve a small sample.

I don’t have a way to calculate BPM. I just have the data BBRef provides at the season level (wasn’t even aware that Torvik provided it at all). I would think any calc should be minutes-adjusted (BPM is a rate stat, but more more minutes with one rate stat should contribute more to the overall estimate than less minutes with another rate stat (hope that makes sense).

Basically, those stats from Torvik show that the two have taken turns being the clearly better player over the past two games, and we’re about equal the two games prior to that. And worth noting that in the game in which Young was bad (Miami) he played his fewest minutes in weeks and less than Lively (who would have surely played more were it not for some foul trouble; eventually they went to zone and that allowed him to stay in). Whereas in the 6-game sample you mentioned before, Young had the better BPM thanks to a clear edge in the State and BC games.

If Lively can play more like the Miami game and avoid foul trouble, his minutes will go up. If he isn’t able to distinguish himself from Young (or plays worse, as the BPM numbers suggest in the VT game), inertia will work against his minutes.

Hopefully the former happens. I think we can agree that we know Young’s ceiling and that there isn’t room for more upwards mobility from him. It is now up to Lively to make the case to return to the lineup by consistently playing as well or better than Young.

[CDu]

Here's a quote for you from Basket-reference.com.

"Box Plus/Minus is good at measuring offense and solid overall, but the defensive numbers in particular should not be considered definitive. Look at the defensive values as a guide, but don't hesitate to discount them when a player is well known as a good or bad defender."

Using BPM to compare Ryan and Dereck is putting your thumb on the scale. Plus BPM doesn't always favor Ryan. BPM versus top 100 teams favors Dereck. Given that 7 of Duke's remaining 11 games will be against teams in the top 100, maybe that stat should carry more weight.

I don't think you have any good evidence that the team performs better when Ryan is on the floor. That being said, I guess I don't have any good evidence that the team performs worse when Ryan is on the floor (though AdjOE and AdjDE as team stats would seem to be of more relevance than BPM (an individual stat)).

I appreciate your trying to educate me on this and most times you are spot on. I just don't think this is one of those times.

Again though, AdjOE and AdjDE are just scaled plus/minus numbers, and those favor Young. As noted in my discussion with Kedsy, I don’t think those stats are very useful, and would strongly favor BPM - even with its limitations - to any of the OE/DE, +/- stats. But even looking at the +/- stats, they don’t suggest the team has played better with Lively in than with Young in this season.

I can certainly agree that we don’t have the evidence to say definitively one way or the other. But the stats that we do have (BPM as well as all the iterations of +/-) favor Young so far this season. That doesn’t inherently mean Young has been the better player; just that there is clearly more evidence suggesting he has been better than there is that Lively has been better, and not much evidence to the contrary.

Hopefully Lively changes that though.

I would also note that while Lively has the higher BPM against the top-100, Young has the higher BPM against the top-50 (in each case, the difference is small though, so take it with a grain of salt either way). So one could make the argument that Young should play more in the games against the top-50 and the teams outside the top-100 (so all but the Wake game) moving forward . That was said tongue-in-cheek, just to be clear.