APBRmetrics

At trio level:

https://twitter.com/bballstrategy/statu ... yMiV3KAnwA

One could argue that Giddey SGA Jalen Williams Dort Holmgren might have some surface design similarities to a Draymond Green Curry Thompson Payton Looney lineup. Thunder Results? Check back late fall 2023.

Warriors results? Super in 48 regular season minutes. Just 1 minute in playoffs. My focus was on Thunder but that is interesting on Warriors. Didn't need it, but should they have used it anyways? Celtics ran 2 big lineups. May look into that playoff issue later.

Warriors used 32 combinations of those 5 players in regular season. This was the 5th best performer but only 23rd in minutes. Anybody in Warriors organization highlight that lineup for more use at any point? Didn't have much impact if they did.

The 2 best of the bigger minute combos each had Curry-Payton-Thompson so that design element is not a one off success. If Thunder emulated that trio it probably would be SGA-Dort-Holmgren.

26 of the 32 Warrior combos were positive, so they could get away with not highlighting or optimizing use of this trio or 5 man. They most used combo had Green-Looney so there wasn't a general aversion to using that. It did fine.

Other future lineup to real lineup comparisons may also come later.

SGA Dort Holmgren would put Holmgren as the Draymond and make the defensive leadership more important than the passing. Could Holmgren pass like Draymond? I dunno, maybe from Draymond catch places.


It doesn't have to be Holmgren or Giddey but it might be, if wanting to maximize Holmgren (and SGA)

Could also look at Thunder lineup possibilities thru the lens of main lineups run by Mavs and Jazz.

Mavs could put out a 2nd creator, 1-2 big wing defenders / 3 ball shooters or 1-2 big defenders or combos. SGA is the Luka, not Giddey. Giddey is the poor man Brunson here, in terms of shooting, defense, fit with star and team success. Actually Mavs were pretty confused. 6 of 10 most used and 11 of 20 most used lineups were negative in regular season. A conference finals does not necessarily mean everything is figured out or that more is coming. Virtually all the big minute positive lineups had Brunson, who they didn't retain and might not replicate.

Thunder could be as confused about what to do around SGA.

With Jazz comparison, Holmgren is hypothetically Gobert with offense. SGA is probably Mitchell. It comes down to if Holmgren meshes with SGA as well or better than the Jazz pair and completions with role players, following Jazz selections or not. And having an above average coach.

Warriors did not retain Payton. For different reasons, perhaps including his size. They still have Wiggins, as a bigger defender to fit with their guard combos.

Thunder probably need a bigger, playable defender, though it could be K Williams if intending to stick with him long-term (vs. using as a trade piece). I don't have a lot of confidence in other options (mainly Dieng) but we'll see if there is development.

Since I brought up Mavs, I'll continue to say that their mixed regular season results continues in playoffs with 5 positives, 5 negatives in 10 most used lineups. Only one positive still possible, with Dinwiddie. It was actually most effective lineups in the 10 most used. Just a bit over 3 minutes per game. Another example of ridiculous lack of optimization based on actual data.

Winning the Finals may indeed come down to who had the best, or least bad, lineup management rather than talent advantage.

Lots of things to watch this season and next. Who clicks with who? There is design and then there is reality. Ultimately the on-going design should be selected based on reality, not the design concepts.

I’m really high on Giddey and Holmgren, but I would still like to also see how some of the Thunder’s older developmental players would look on the court with some better quality ones like some of the more experienced veterans on the roster.

I think a Gilgeous-Alexander/Dort/Williams, K./Favors/Muscala lineup would be really great all round.

That lineup was never tried last season under the tank. Only 1 one change variation was tried- without Favors. Way better than team average (not desired by team) but still almost -3pts / 100p. Got 117 minutes, more than I would have expected.

I'd emphasize trying these lineups:

Giddey JDub KW JRE JWill
G SGA. JDub Dieng JWill
G SGA Dort KW Musc
G Mann KW Poku JWill
G Dort Baz KW Musc
G Mann Baz Poku JRE
SGA Dort Baz KW Musc
SGA Mann Baz KW Musc
SGA Dort JDub Poku Musc
SGA Dort JDub KW Musc
SGA Mann KW Baz Poku
SGA Mann Dort Dieng Musc
G JDub Dort Dieng JRE
Mann JDub KW JRE Musc
G JDub Dieng Baz Musc
Maledon Krejci Waters Poku JWill

Not all every game. Room for some choice, matching up.

If Muscala is overworked, substitute in somebody else.

Giddey - SGA de-emphasized but play the listed options higher minutes and / or substitute one for Mann or JDub some, if / where you want.

Only a few cases of more than 1 rookie on the court. That is my preference for learning.

I could agonize & fine tune these lineups for weeks... but that is someone else's job.

They could play these 16 lineups 200 minutes (give or take) and a few for a lot more... or use close to 800 for an average of 5 minutes each and very very few over 100 minutes. I know what makes more sense to me.


(See my Twitter stream for other recent lineup commentary leading up to this post.)

Lots of ways to run a lineup rotation. One game into pre-season I have a new short set:

Giddey SGA K Williams Poku JRE 12 minutes / gm.

JDub Mann Dort Dieng JWill 4

Giddey Mann Dort Muscala Bazley 16

JDub SGA K Williams Poku JRE 16

.

Fwiw comments on data from first 2 games:

https://twitter.com/bballstrategy/statu ... 1qNocZ4eCQ

Crow wrote: ↑Tue Oct 04, 2022 5:58 am Lots of ways to run a lineup rotation. One game into pre-season I have a new short set:

Giddey SGA K Williams Poku JRE 12 minutes

JDub Mann Dort Dieng JWill 4

Giddey Mann Dort Muscala Bazley 16

JDub SGA K Williams Poku JRE 16

Oh hey, none of these lineups used a single second so far in regular season. (Dunno about pre-season.)

Too good? Or just not using data in same way?

That earlier list of 16 proposed lineups? Checked half and found 4 minutes of use. May or may not check the rest later.

woohoo! Found 1 more minute of test in final 8.

So for those 16 lineups and the 4 later ones, a total of 5 minutes test or about 5% of total time available.

Clearly Coach D and company have completely different lineup ideas than I do. They are failing thru 2 games. Ok with that? Probably but we'll see after say 20 games.

Can't say a lot about my 20 if they don't play them, except maybe a little indirectly thru sub-lineup data. May try that much later if they continue to ignore these suggested lineups.