APBRmetrics

Last season 4 of 5 most used lineups were negative. Starters at 9 minutes per game and slightly negative. 2 of the 5 most used badly negative. 55% of the 20 most used negative, 30% badly.

One of few teams at 20 contracts.

Lineup management will again be critical and last season raises significant concerns.

Only strong player pairs were VanVleet with Thompson and with Tate and both were very very far from first tier in use.

The only good first tier in use pair was VanVleet - Sengun.

Jalen Green and Jabari Smith in a couple of the worst bigger minute pairs with other starters.

Pair data generates interest- to me- in the quad of VanVleet Thompson Tate Sengun but Udoka only gave it 29 minutes. +13 / 100p though. 151st on use. How many of the 150 more used had a better case?

The trio without Tate got 108 minutes and was even better at +21. 27th most used. WAY TOO LOW.

The trio without Sengun was barely half as good as without Tate but +12 suggests more use / test of that too.

F. VanVleet - J. Green - J. Smith Jr. - A. Thompson
was +10 but only 10th on use. What the heck? Just the best performer in 27 most used.

The strong performance of VanVleet - Thompson is encouraging. Is it considered the core of core for next season? Would have to convince me if putting something else forward.

Which lineup unit will Sheppard do best and worst with? Which will be most used? Important choices to make and begin to judge on the results.

More Rockets talk:

https://x.com/bballstrategy/status/1814698211489292682

https://x.com/bballstrategy/status/1843822992297717775

And some stuff out there between the last 2 posts. Maybe I'll find the link.

Starting lineup doing very well.

Lineups with Thompson or Sheppard mostly negative. Landale all negative.

More probably a bit later.

5 of 7 most used badly negative. But 10 of 10 most used pairs and trios positive overall. Issues with exact lineups. Begins to break down some at quads.

Lots of bad bigger minute Thompson quads. All of the 7 most used. All of the 4 biggest Landale quads. 5 of 6 biggest Eason quads. 3 of 4 biggest Sheppard quads.

Lots to improve or replace.


Sheppard with poor early shooting and a 25% turnover rate.

It is super early but this is 1 way to run the Rockets rotation:

VV JG BR JS AS 20 min
VV AT BR E A 4
AT JG Tate E A 8
VV RS W E JS 4
VV JG W Tate AS 4
RS AT W E AS 8

Main messages are: play the heck out of the starters, find a few others (these or different) and rarely play dinks (and have a good reason).

Whittle 11 man rotation to 8-9 by playoffs based on team performance on, not draft pick number or salary or place on design template.

Rockets with 4th best SRS, tied for 6th on W-L. Where from here on both?

Thompson - Eason pair is +17 pts per 100 possessions. Best pair in 15 most used but only 11th in use. Only pairs better in 20 most used have one of them and Brooks.

I suggested 20 minutes per game for pair. Actually use is a bit less.

Starters used 13.5 minutes per game. I suggested 20. Mildly positive.

VanVleet Brooks Thompson Eason Adams. I suggested 4 minutes per game. Only 9 minutes actual use so far. Did great.

Lots going well.

Still wonder about optimizing design / fit, but not much inclined to pick thru the detail today. So instead I'll say... the low level of Moreyball is a massive issue that will probably hurt them in the end.

For season 6 of 7 most used lineups positive, with starters and Thompson instead if Smith an under or not discussed total failure. Starters with Thompson over Green a tremendous success in a far too small test. Starters with Thompson and Eason over Green and Smith twice as good as with just Thompson but again in ridiculously small test.

Too weded to Green and Smith? There are no big minute pairs in 20 most used beyond meh with Green. Only good pairs are with Whitmore and Tate.

Smith team results good with most. With Green is least good big minute pair.

Sengun is in nearly every strong trio in 20 most used. Start there, use the good trios and fill out the other spots. I can do do that and probably will but maybe not exactly this moment. It is pretty straightforward. Turn the best trios into the best 10 quads and then finish them, hopefully better than currently. Divide the minutes mostly by past success or possibly strong feelings if they exist, if you have a basis for trusting them. Reduce or eliminate almost everything else.

382 total lineups used. Not the highest. not the lowest.

Should be way lower imo.

Sheppard plays pretty often with Eason and Thompson and pretty well. Appropriate for defensive help and finishing. 36% with Sengun. Would benefit from more time with him but isn't going to happen this season. More time with Adams is realistic and desirable for screens, rim protection and some finishing.

Starters, 9 min / gm last season. Modestly over 5 minutes this season. This while the effectiveness went from -1 to +5 / 100p. Odd. But 6 of 7 most used being positive makes it better.

VanVleet - Sengun has a handful of rival strong pairs but remains near top. VanVleet Thompson is also up there.

VanVleet -Green, one of the weaker biggest minute pairs (slight positive) with Green-Brooks and Green-Thompson actually crossing into the negative

F. VanVleet - J. Green - J. Smith Jr. - A. Thompson remains 10th most used with now a mild / middling positive.

Thompson - Eason pair is still one of best but less than half as strong as it was early.

Thompson Eason Sengun finally got real test this season after barely playing last season(and me calling that out) and is best trio in 20 most used.

5 good big minute quads, 4 weaker ones.




Designed a new 6 lineup rotation based on the lineup and sublineup data. First draft but way way more concentration than present. Every lineup would get more time than current starters.


Udoka in last 5 games, 69 lineups. 4 lineups over 2 minutes per game... if used evenly but only 1 of 4 used more than twice. 3 of 4 positive though, without VanVleet. In fact only 4 of the 69 even used twice. Near unique circumstances for each call or just excess change? 12 unique lineups in each game? Looks like chaos to me.

5W-11L in last 16. Downturn almost completely coincides for VanVleet absence.

3 of 4 most used lineups positive for season but 3 of 4 negative over last 15 games.

6th seed is a bad case possibility.