Phase IV - Wake 2 through Unc 1

[devilsince1977]

After Justise moved up to PF, our starting perimeter was Tyus Jones (3.0 def reb per game), Quinn Cook (3.0 drpg) and Matt Jones (1.5 drpg), for a total of 7.5 drpg. Our starting perimeter now is (as you pointed out) Gary Trent (3.2 drpg), Grayson Allen (2.9 drpg), and Trevon Duval (1.4 drpg), for a total of 7.5 drpg.

Also, while height and jumping ability may have some bearing on the rare occasions when a guard goes "in among the trees" to snare an interior rebound, the vast majority of perimeter player rebounds happen, not surprisingly, out on the perimeter, on long rebounds. And in those cases, height and jumping ability have a lot less bearing than positioning and boxing out (plus, to a large extent, luck).

In any event, it's hard to fathom why 7.5 drpg from the perimeter might have been acceptable in 2015 but the exact same number is not acceptable now.

(FWIW, the entire team is slightly better at defensive rebounding this season then we were in 2015, but it's pretty close.)

I never compared 2018 to 2015. But since you brought it up; Justise didn't move to starting PF until the Clemson game in on Feb. 21st, Plus he still played the 3 for several minutes a game after the change. Justise averaged 5.2 drpg for the season. Your 7.5 per game number is cherry picking stats since you totally eliminate Justise's time playing at the 3. Which was at least half of his minutes for the season. This years starters have not changed so the stats are less distorted than 2015.

Size and leaping ability may not be as important for the perimeter players as the bigs, but is sure doesn't hurt. The fact that Tyus and Quinn were getting 6 per game; while Trevon and Grayson are only getting 4.3 could be evidence that Trevon and Grayson need to do better. Luck tends to come to those who are in position to receive it.

[Kedsy]

Because in 2015 we played small ball, and purposely exchanged the ability to rebound in return for quickness on the perimeter. In 2015 we were among to top teams in the country in opponents 3PT% and forcing Turnovers. This year we're playing big and are terrible are defending the 3 and forcing turnovers, so we need to do much better on rebounding to make up for it.
Further more, your numbers are off because in 15 we played a much slower tempo of 66 possession per game versus 71 ppg in 2018, add that to the fact this year's trio actually averages 6 more total minutes played per game means this year's trio is not rebounding nearly as well as the perimeter did in 2015.

Well, first of all the 2015 was tied for 220th in opposing turnover percentage, so not really among the top teams in the country at forcing turnovers (we were 37th in total turnovers forced, but that's mostly because we played more games than most teams; on a percentage basis we were much closer to the bottom third). Second, while it's true the 2015 team held its opponents to a lower 3-point percentage, this year's team holds its opponents to a lower 2-point percentage. The result being that the opposing eFG% are not very different (46.5% in 2015 and 47.1% this season). So there wasn't as much justification for lower rebounding in 2015 as you suggest.

And if you'd rather go with DR% to counteract the discrepancies in possessions and minutes, there's still not much difference:

Tyus: 9.9%
Quinn: 9.2%
Matt: 7.4%

Gary: 9.7%
Grayson: 8.6%
Trevon: 4.9%

Some difference, but not all that much, especially when you factor in that on a percentage basis, this season's perimeter guys are competing with a lot more big guys for the boards.

Again, if the 2015 perimeter players were satisfactory at defensive rebounding, then this year's perimeter guys' defensive rebounding is not what's causing the problem (with the caveat that our overall defensive rebounding was still unacceptable in the UNC and St. John's games).

[Kedsy]

I was talking about defensive rebounds. Grayson averages 2.9 per game, Gary 3.2 and Trevon 1.4. That is 7.5 per game. With the heavy minutes they play it is not good. Also Tyus and Quinn were about 6'1 while Trevon and Grayson are taller and elite jumpers.

I never compared 2018 to 2015.

OK, if you say so.

[Kedsy]

But since you brought it up; Justise didn't move to starting PF until the Clemson game in on Feb. 21st,

Also, since our 2015 defense was not all that good before 2/21/15 and got a lot better after, I don't think the comparison is cherry-picking at all.

[devilsince1977]

OK, if you say so.

In my original post listing problems with the defensive rebounding; I didn't mention 2015. I did later compare the backcourt size and athleticism in response to jv100. So I did compare players from both teams. I stand corrected.

Here is my "Cherry Picking" - In the six NCAA tournaments games in 2015 Justise played 178 total minutes. I added up the minutes played by Jah, Marshall and Amile, subtracted that from 80 minutes(the center and PF minutes) which left 109 total PF minutes for Justise. That leaves Justise playing 69 minutes or 38.8% of his minutes as a 3 in the tournament. He had a total of 52 DRB. 38.8% of the 52 rebounds is 20.2 rebounds. That is 3.4 perimeter drpg. Add that to the 7.5 drpg and you end up with 10.9 drpg for the perimeter players in the 2015 NCAAT; when Duke was a good to very good defensive team. 2015. That is a significant difference and one that the 2018 perimeter players should strive for. It they had 11 drb instead of 6 drb verses UNC, that would equal 5 less shots for UNC and 5 more shots for Duke. With 10 more shots the shot disparity is much less and Duke almost certainly wins.

We are not a good defensive team so far this year and defensive rebounding in our losses has not been very good. The starting perimeter players are getting 5.8 drpg in those losses.

In this phase, DRB by the perimeter was not a strength and in the losses it was a weakness.

[Kedsy]

In my original post listing problems with the defensive rebounding; I didn't mention 2015. I did later compare the backcourt size and athleticism in response to jv100. So I did compare players from both teams. I stand corrected.

Here is my "Cherry Picking" - In the six NCAA tournaments games in 2015 Justise played 178 total minutes. I added up the minutes played by Jah, Marshall and Amile, subtracted that from 80 minutes(the center and PF minutes) which left 109 total PF minutes for Justise. That leaves Justise playing 69 minutes or 38.8% of his minutes as a 3 in the tournament. He had a total of 52 DRB. 38.8% of the 52 rebounds is 20.2 rebounds. That is 3.4 perimeter drpg. Add that to the 7.5 drpg and you end up with 10.9 drpg for the perimeter players in the 2015 NCAAT; when Duke was a good to very good defensive team. 2015. That is a significant difference and one that the 2018 perimeter players should strive for. It they had 11 drb instead of 6 drb verses UNC, that would equal 5 less shots for UNC and 5 more shots for Duke. With 10 more shots the shot disparity is much less and Duke almost certainly wins.

We are not a good defensive team so far this year and defensive rebounding in our losses has not been very good. The starting perimeter players are getting 5.8 drpg in those losses.

In this phase, DRB by the perimeter was not a strength and in the losses it was a weakness.

We are not a bad defensive rebounding team this year (for Duke, we're the 2nd best ever), but in 3 of our 5 losses I agree it was a weakness.

Since PFs get more rebounding opportunities than SFs, I'm not sure you can simply apply a straight percentage like you did, but I don't really have a better alternative for an estimate. Though if you're going to add a non-starter's rebounds to the 2015 group then you'd have to add, e.g, Alex O'Connell's and Javin DeLaurier's and Marvin Bagley's (and lately Jack White's) wing minutes as well (there are an average of 24 perimeter minutes which were not taken up by GA/GT/TD this season), which would add to the 2018 wing totals.

Using our imperfect estimate, I've compared this year's perimeter rebounding in our losses to our 2015 championship team's perimeter rebounding during the NCAA tournament. Using your straight line formula for Justise Winslow's perimeter minutes in 2015 and Marvin Bagley's perimeter minutes in 2018, it looks like during the 2015 tournament run our perimeter players grabbed approximately 29.4% of available defensive rebounds while during this year's losses our perimeter players grabbed 18.3% of available defensive rebounds. So if our perimeter players had rebounded like national champions during the championship run, they would have grabbed a bit less than 4 additional defensive rebounds per loss. Which probably would have helped, since all five losses were relatively close.

Except they were losses -- can you really expect a team to rebound like champions during losses? For example, in our 2015 loss to NC State, our perimeter players only rebounded 13.6% of available defensive rebounds, worse than our current team in its average loss.

I conclude our perimeter players could have rebounded better in our losses (though of course our interior players could have rebounded better in our losses too), but I still disagree it was the primary reason for the Ls.

[Kedsy]

We are not a bad defensive rebounding team this year (for Duke, we're the 2nd best ever), but in 3 of our 5 losses I agree it was a weakness.

Since PFs get more rebounding opportunities than SFs, I'm not sure you can simply apply a straight percentage like you did, but I don't really have a better alternative for an estimate. Though if you're going to add a non-starter's rebounds to the 2015 group then you'd have to add, e.g, Alex O'Connell's and Javin DeLaurier's and Marvin Bagley's (and lately Jack White's) wing minutes as well (there are an average of 24 perimeter minutes which were not taken up by GA/GT/TD this season), which would add to the 2018 wing totals.

Using our imperfect estimate, I've compared this year's perimeter rebounding in our losses to our 2015 championship team's perimeter rebounding during the NCAA tournament. Using your straight line formula for Justise Winslow's perimeter minutes in 2015 and Marvin Bagley's perimeter minutes in 2018, it looks like during the 2015 tournament run our perimeter players grabbed approximately 29.4% of available defensive rebounds while during this year's losses our perimeter players grabbed 18.3% of available defensive rebounds. So if our perimeter players had rebounded like national champions during the championship run, they would have grabbed a bit less than 4 additional defensive rebounds per loss. Which probably would have helped, since all five losses were relatively close.

Except they were losses -- can you really expect a team to rebound like champions during losses? For example, in our 2015 loss to NC State, our perimeter players only rebounded 13.6% of available defensive rebounds, worse than our current team in its average loss.

I conclude our perimeter players could have rebounded better in our losses (though of course our interior players could have rebounded better in our losses too), but I still disagree it was the primary reason for the Ls.

Actually, the more I think about it, my analysis isn't entirely correct. The perimeter players from 2015, playing in full championship form, might have obtained almost 4 additional defensive rebounds per 2018 loss, but more than 70% of those rebounds went to Duke anyway (grabbed by the 2018 Duke bigs). So really, the difference comes out to a little more than one (1) defensive rebound per game that the perimeter players might have taken away from our opponents.

Would an additional defensive rebound per game have helped us, in five games decided by an average of 5 points per game? Sure. Would it have turned any of those losses into wins? I don't think there's any way to tell. Was that one extra rebound the primary reason for our losses. I've upped my opinion to "no way."

[SilkyJ]

But Marvin's defense down the stretch was fantastic. Maybe he was gassed. Maybe getting him just a little rest could have helped. But you are correct, defensive rebounding killed us. GoDuke!

He had one cool block. Come on.

Rebounding was absolutely the issue in the second half. No disagreement there. But it wasn't Bagley's fault. More below.

Yeah, no.

Bagley had an 18% defensive rebound rebound rate in the second half. That is a solid number. It is 1-2 rebounds different than in the first half. And a 19% versus a 13% offensive rebound rate from first half is literally 1 rebound difference.

Bagley was otherworldy rebounding in the first half. He was just very good rebounding in the second half. The difference was a whopping 2-3 rebounds. Likely statistical noise. Note: he also had 2 blocks and a steal in the last 7 minutes of the second half. He was not the problem, and fatigue was not the problem for him. The problem was that UNC changed their approach in the second half, and made life more difficult for Bagley on both ends. And we also switched defenses multiple times, which didn't help.

But the real problem is that nobody else was getting rebounds. We had at least one and often two other bigs on the floor in the second half. Those guys were not getting rebounds. Wendell Carter and Marques Bolden combined for 7 rebounds in 45 minutes. They were the rebounding problem for us in the second half. Because no player should be expected to be getting 31% of the defensive rebounds.

I am with CDu here. If you are arguing that Bagley was gassed, why even include the first part of the second half in your analysis, when he was more rested than the late first half? Would a tired player be better in the first five minutes of the half rather than the last, when Bagley excelled?

He is a 19 year old kid with great conditioning who had five days between games. I am not saying it is impossible for Bagley to get gassed (though he does seem superhuman) just that I see no evidence of it. At least my eye test watching last night and your numbers don't convince me as such.

Edit: also, a lot of those open three UNC got were very early in the game. I might buy fatigue for lazy defense later on, but not the first twelve minutes when the only thing keeping them close was Williams' hot hand.

I'll concede that we are dealing with the law of small numbers, but I am factoring in my "eye test" here and I think Marvin, and others on the team, looked tired. Maybe they were just being lazy, but we got beat in transition several times and didn't seem to have pop on defense/the boards. They were just way more active than we were. You can ignore the numbers if you want, but the facts are that we got beat in transition several times and his reb% was lower in the 2nd half.

Now, our defense was also just bad.

the 1:52 mark here...I mean Marvin is still in the paint when Williams catches it 30 feet from the hoop on the wing. That's just unacceptable. (The at the 2:07 mark he puts his head down as he runs back...that's usually a sign of being tired. This doesn't have all the UNC transition buckets, but my guess is there are a couple that look like that.)

UNC is also a touch matchup for us b/c their bigs take ours out of the paint they shoot 3s well (and we defend 3s poorly). So there's a lot going on here and I don't think there's a silver bullet, but I do think that having our best players be a bit fresher would help)

Also, since our 2015 defense was not all that good before 2/21/15 and got a lot better after, I don't think the comparison is cherry-picking at all.

You are smart to remind us of this. Our Kenpom D-rating was in the 50s IIRC heading into march and finished at 11? I always forget this and have to hope that our defense gels with experience...its just tough to rely on our defense getting hot in March.