APBRmetrics

The 5 most used lineups against Pacers in Finals were +29, 170% of the margin in the series. The other 69 lineups were almost all dinks (less than 1 minutes game) and -12. The dink game was lost. Title was won with what concentration there was at the top. But slightly more time was given to the unsuccessful dink game.


Carlisle won with his 5 most used lineup too but not by as much and lost dink game as well but by slightly less.

Series are generally won overhelmingly with most used lineups as I have shown with past titlewinners.

Pacers starters used 40% more than Thunder's and were moderate positive to moderate negative. The next 4 lineups won it for Thunder. In fact, starters with Wallace over Hartenstein was enough by itself on paper.

I don't expect near trades but here are a few ideas that came to mind anyways:

Dieng (and pissibly other) to Raptors in a sign n trade for Boucher.

Hartenstein to Bulls for Vucevic.

Hartenstein and perhaps other to Blazers for R Williams and perhaps other.

I Joe for Konchar and stuff.

Some combo of players (Joe, Dieng, Topic, others) for Giddey in sign n trade at "right" price.

Of the probable strongest 4 teams in west:

Thunder were best ranled on both offensive and defensive efficiency and pace. Rockets 2nd on defense, Nuggets second on offense. Timberwolves with a better balanced pair and slowest pace.

Thunder were 1st on group on 3 factors, last on 2. Rockets 2-2. Nuggets 3-3. TWolves last on one.

Both Rockets and Nuggets had a last place on one on the efg%s, Rockets on offense, Nuggets on defense.

Rockets got Durant to improve shooting and hopefully offense overall. Time will tell how much and to what other changes.

Nuggets swapped SFs and improved depth but impacts remain to seen, especially on defense.

Wolves are relying on continuity and internal improvement.

Rockets have the advantage over the other 3 on rebounding. Nuggets lead on shooting. Wolves balance is notable but not being best in group at anything is also notable.

Thunder have 3 factors in 20s. None of the others have more than 1.

In the playoff series, Thunder always won turnover comparison, usually efg%. Lost rebounding. FT factor was mixed. Won against Pacers and essential as efg% was lost.

Timberwolves offense did best against Thunder but the defense fell well short of the goal.

Against the Pacers, lineups that had all 3 of SGA, JDub and Holmgren were just 1.3 pts / 100p better than team average results or about 3 pts better than all other trios / lineups.

For 11 trios over 100 minutes in series, it was 5 positive, 6 negative. Barely enough.




In regular season, SGA without the other two and the other two without SGA were both 10 pts better than the trio. JDub without the other two was by far worst combo and slightly negative in 600 minutes.

The trio is very good but not best or absolutely required.

Against Pacers, 3 of 5 most used quads were negative. All involved Caruso with 3 of SGA, Dort, JDub and Holmgren. The only version of Caruso and three of these that was positive (a small one) was when JDub was not named. Caruso with all 4 was negative.

Caruso wasn't in, wasn't tested in any of 35 most used regular season quads.

Barely survived those minutes, that lack of adequate testing.

"All these guys were solid, all those quads were reasonable." Maybe in theory, maybe in long-run. But mostly not positive actually in finals.

In regular season, the bigger minute Caruso quads almost always involved Joe, Wiggins or Wallace and were almost always positive. But change the pattern at the end and miss... and escape.

Thunder handled Clippers 4 - 0 in regular season. Still, Clippers did a few things well: defensive rebounding, 3pt shooting and overall ts%. Off-season moves probably accentuate these strengths. Will they get worse anywhere? Shot defense?

The Topic prospect and non-SGA lineups:

https://x.com/bballstrategy/status/1947758085080289727

10 lineups, the best of the most used last season, with mostly 1 player changes as necessary to fully distribute minutes tealistically, could be a complete base rotation. Have 10, 20, 50 more variations as options if you like. That is enough.

In addition to the factor level comparison of Thunder to key western rivals, I would add sub-factor consideration of 3pt game (own & opponent), degree of rim protection and relative impacts at PF.

Rockets and Nuggets may look best overall but I see a lot of strength in the detail for Clippers and Timberwolves.

Not that high on Lakers but assume they aren't totally done. Warriors similar situation.

Important of Mavs to be determined but would not want to see in 1st round.

2 playoff runs with sub 30% 3pt shooting from Chet. Delivered on defense but does this change?

The starting lineups of the 3 playoff opponents after the first round all beat the Thunder. Pacers moderately, Nuggets and Timberwolves into + double digits / 100p. Griz starters were utterly destroyed.

The next 6 biggest lineups against the Thunder were all negative, all but one horrendous. Averaged 2 subs. Didn't work.

Nuggets used starters over 20 minutes per game. Could maybe have squeezed a little more out of it. Timberwolves and Pacers only went with them 11-12 minutes per game. Bad mistake by both. Should have been 75+% more. Go with what I'd working.

If a team beats Thunder next spring it probably play starters 20+ minutes per game and win those minutes. 3 teams proved it could be done, 1 tried hard but maybe not quite hard enough. Encouragement for the next round, for those paying attention and coaching in accordance.

Thunder won nearly 2/3rds of the other 50 most used lineups beyond the first 3. Better, less pummeled lineups beyond the starters would be nice but the data clearly indicates the successful path was with the starters.

What starting lineup do the Thunder use next playoffs? During the regular season? Test more than 1 adequately or just one again, barely?

Who else beat the Thunder with starters in the regular season? Rockets,,Celtics, Clippers and surprisingly some weak teams. Watch out for those contenders in the future too.

It is not an easy task to beat the Thunder. Nearly 75% of the 50 most used lineups failed. Coaches tried an average of nearly 40 lineups per team against them. More than 2/3rds of all those lineups were negative. Trying to win a micro-tactical dink bench lineup war with Thunder didn't work. At least this year by those who tried.

The best strategy would appear to be to use your best lineup to the very max and a few next best guesses over going scattershot. But go scattershot and lose if they want. The most successful teams against the Thunder in regular season went with starters for 15 minutes per game. Not bad, but more concentration might be better.

Concentration is the apparent best strategy but not a guarantee. Thibodeau went with starters 27 minutes per game but failed by a large margin. Wrong lineup and / or strategy & execution.




I developed a theory that the best way to attack the Thunder would ideally be with a low turnover playmaker, at least one inside threat, 2 tall shooters and a physical rim protecting big. The data produced by scattershot coaches is too scattershot to prove that or anything else but the best few lineups show some similarities to that, especially on the tall shooters aspect. 8 of the 10 biggest minute / positive lineups in regular season and playoffs shot better than league average from 3 to way better. There might be other common elements. About half of the best also did well on rebounding. Turnover rates were more often tolerable than not.

Caruso, 10 of 11 player pairs positive in playoffs. But somehow 4 of his 6 most used lineups by Daigneault were negative, 2 horribly. Coach D lineup management of Caruso was the main error. Caruso in 6 of 8 worst lineups in 50 team lineups most used, from -12pts / 100p to -82. All tiny to super tiny samples but harmful and barely survived.

In general, lesser team performance when with Jalen Williams, Dort and Hartenstein and quite negative with Jaylin Williams. Biggest issue was when with both Dort and Jalen Williams. Not problems in regular season.

Noise or signal or how much of each? More data next time may improve guesses. Team study before then?

Concentration on top 5 lineups last season declined in regular season by almost 30% from previous season, down around 10% in playoffs. Wrong direction imo. Obviously other stuff worked, but this is still worth noting as a common source of edge little tapped.


In regular season, Thunder only had 1 of 50 most used trios. An average team would have nearly 1.7. SGA - Dort - JDub. It was very good but only 27th most used. 3 in second 50. All good, all could / should be used more. The biggest was used 30% of time in regular season, the rest in low 20s. Some went above 40% in playoffs but more was possible.

The goal is to win across the sum of all lineups, but getting a lot or more from the top is the major path toward that.

Hartenstein had third straight playoff drop from regular season on total rebound rate and 2nd straight drop on block rate but overall team defense and more efficient shooting carried him to his best and a very good playoff BPM.