APBRmetrics

Crow wrote: ↑Thu Dec 26, 2019 7:59 am Payton - Ntilikina? 4 minutes of testing for season. Come on. Try it. They try so many other things but not this? A mistake not to test it.

No testing since.

Crow wrote: ↑Thu Dec 26, 2019 6:47 am No Hawks lineup tested even 4 minutes per game.

None over 5 minutes per game in last 5 games either. Top 5 lineups all negative, 4 sucked. At least they are "new". All pretty much insufficiently tested before and insufficiently tested recently too though. Knowledge gained? Not much probably.

Crow wrote: ↑Thu Dec 26, 2019 8:32 am Mills - White and Mills - Murray are both essentially neutral. I am not seeing any testing of all 3 together. I'd at least try it some to check whether to test further. Probably with 2 of Gay, Aldridge and Poltl but possibly with other interior combos.

Since comment, Mills-White went from 10 min /gm to 13 / gm and from neutral to good.

Mills-Murray from 7 min /g to 4.5 but from neutral to fabulous. Do they know or care? Not strong evidence.

Crow wrote: ↑Fri Dec 06, 2019 8:28 pm Wizards with by far the worst lineup of the 20 most used in league.

Rui goes out 10 games ago. What did they replace that lineup with? Near-dink and dink lineups. One of the near dinks was good, one bad. The one with basically all new starters was good.. in so far very very limited testing. The one with basically all previous / frequent starters was bad.

Do they know what they are doing? Doesn't much seem like it overall, now or for season.

Crow wrote: ↑Sun Dec 15, 2019 12:34 am Spurs starting lineups and lineups with a lot of starters continue to struggle. It is the bench players and bench heavy lineups that are the main source of team edge, including in today's win.

They are doing better lately but they aren't fully optimized.

Spurs went back to original starting lineup 11 games ago and it was doing great, far far better than early in the season. So what does Pop do last game? Make 2 changes to it. It worked for that game. one of the few lineups that did in a loss. But what happens next? Will it work better than what it replaced over the long-run?

Pop goes back to original starting lineup and probably uses the full super sub lineup.

Hawks have done little to no testing of Young Huerter Parker Collins quad. Only modest testing of trio of YHC this season. I would do more of both. YHC was most tested last season and essentially the most successful bigger minute trio but there are a lot of new faces to sort thru to complete a lineup. Miss Dedmon. Could they get back? Probably need new 4th and 5th guys. Probably not Hunter or Reddish.

It took me 1-2 minutes to identify Rose Galloway Kennard Wood Drummond as a lineup I would test a lot. Pistons have tried it just 13 minutes. Those guys with Griffin over Wood just 36 minutes. The first super great in those trivial dink minutes fwiw, the latter also great. Not testing these extensively seems very rigid or poor oversights. The quad without the PF specified is like the 2nd best bigger minute quad. Under 100 minutes. Should have been over 500 minutes probably. Maybe way over.

82games.com data suggests Kennard is fine individually and team-wise at SF. Galloway does fine at SG both ways. There are no apparent clunker player pairs involved.

Stevens has not played any 5 man lineup over 4 minutes per game. One quad over 7 minutes / gm. No trio over 13 min / g. Is that enough testing of the top? I don't think so.

Conley has been a disaster individually. They play a few lineups a lot and well though with him. They probably need to go even further in that direction in the playoffs.

If they want to win more, the Griz should probably play starters twice as much and try various ways with Hill more.

Bulls have a fine lineup that they use on average 8 minutes per game. Why not use it 12-15-18 minutes per game? Why not indeed.

9 of 10 trios are positive and the 10th is barely negative. But when they move to 5 man lineups only half are positive. Too much futzing around try to customize. 213 lineups used is not that bad compared to other teams but it is still ridiculous, inefficient and unnecessary.

Bulls probably could be better. Perhaps a lot better. With better lineup management.

I now see that the good Bulls lineup has been used more recently, when available.

League-wide at about the halfway point (average of 43 games):

8624 lineups used, or about 288 per team on average. 4.2% more than at similar point last season. 40.55% cumulatively positive (just slightly less than last season). So overall lineup use is modestly more diffuse and slightly worse than last season (by this measure).

Celtics, 273 lineups used, 44% positive.
Sixers, 269 lineups used, 41% positive.
Heat, 293 lineups used, 42% positive.
Raptors, 302 lineups used, 44% positive.
Bucks, 372 lineups used, 46.5% positive.

Lakers, 306 lineups used, 43% positive.
Clippers, 203 lineups used, 40% positive.
Rockets, 285 lineups used, 41% positive.
Nuggets, 188 lineups used, 45% positive.
Jazz, 286 lineups used, 40% positive.
Mavs, 205 lineups used, 43% positive.

Across the 2 criteria, I'd call the Jazz's numbers t h e least desirable. Best? Bucks and Nuggets with quite different lineup counts. Bucks getting results but the lineup count concerns me. More selections, more ways to go wrong.


Budenholzer only picked 33% winning lineups in 2 losses to Celtics and picked fewer total lineups. May need to try different there. Only 23% winners in loss to Heat.

Vogel used 33 lineups against Clippers in 2 losses. Just 2 used in both games. 27% winners. May need to try different there.