APBRmetrics

My first thought on a starting lineup for Knicks is Payton Ellington Dotson Morris Robinson. Some vets out of the gate. Ball control. Defense. Space & 3 point shooting.

I'd try 1-2 player substitutions into that as well as a second line of Smith Barrett Knox Randle Portis. I'd think they'd want to win as much as possible to get out of the basement and the costs of that.

Unclear what will work.

What didn't work was Fizdale using 856 lineups, with only 29% positive. 10 of the most used 11 lineups were losers. He has to do way better, take instruction or be removed.

Really try 5 to 20 lineups. Beyond that make selective use of another 30-80. 856 is just insane. It is changing every few minutes and rarely trying for more than a few minutes each for the entire season.

The team could be a gritty competitor but they need a design and tough choices on who plays how much.

Trades should be constantly considered. Few of their signings are sure long-term keepers, if any. A consolidating trade for a lead player would help. Robinson is the core guy but they need a lead scorer, preferably scorer / creator.

Lakers will be able to play James Green Kuzma Davis Cousins. They are also play many many variations with 3-4 of these... and hundreds of others with less. What happens with team in regular season and playoffs will be heavily dependent on lineup management.

Vogel practiced one big minute lineup early in his career with the Pacers. His best years. Up to 20 minutes per game, which is about as high as you ever see. Then he went away from it. Because of George injuries but even after. It continued in Orlando. His last season there no lineup was used 3 minute per game for season. Was this a conscious decision or unrecognized chaotic behavior? Either way it doesn't look good... cuz it didn't work. He probably should find a big minute lineup. The one I suggested or some modest variation. Dinking around with hundreds and hundreds of lineups sound like a good way to fall short of optimum, of expectations. A good, maybe quick way to get fired. Or to deserve to be fired.

If Wizards want to try to win games, they should have at least 4 of these guys on court as much as possible: Beal, Schofield, Brown, Hachimura, Bertrans and Bryant. Beal as their main PG. By as much as possible, I mean toward 40 minutes, not just 25.

If you want to tank or see no way to do better, then play less than 4 of these guys as much as possible or even never play 4 together. Play hundreds and hundreds of dink lineups and act like you are trying by giving everyone a chance.

Try the best in a very disciplined manner or it will probably be bull or just not work.

There is plenty to learn from number of starters matchup analysis. But a next stage of learning would be to look at performance of lineups based on quality of player in specific roles. Creators, shooters, interior scorers, etc. Where can you get by or skimp? What skills are the most important? It depends what plays you run, what shots you want / can get. The lineup should determine the plays or vice versa or both depending on stage of game and results to date.

If you have an elite creator, an elite shooter and an elite interior scorer on court with 2 bench guys, ran a play using the elite guys were can or use the other guys were possible or necessary. Maybe substitutions need to focus on roles.


Who shows more focus on roles in lineups? Coaches, lineup analysts, both or neither? It probably should get more recognition / attention. Functional lineups, not just 5 names or plus minus. Though they at important, they probably aren't sufficient. And again it is lineup / playcalling. Playcalling will affect efficiency and efficiency should affect playcalling.

Every play ever written up should probably be annotated for types and level of skills required by each player and then the lineup should match that well or at least pretty well.

No play type or specific play data (run & efficiency) by lineup exists publicly. Team analysts should have it and be working it hard. Are they?

Fix lineups by changing players, roles, plays or combinations.

What "should" work vs. actually works?

On defense, you can't predict plays... unless you try & can. Probability and sequencing. Decide where & how to disrupt or reduce efficiency by coaching instinct and video and analytics.

Mavs have a lot of lineup options, as do many teams. But the point , of course, will be to find the best and play them most. Teams on the playoff borderline have more urgency to get it more right and sooner than later.

Saw someone suggest I. Thomas as starter for Wiz. I am skeptical. I'd think they'd want to maintain Beal dominance out of the gate. Maybe Thomas finishes, if hot and / or playing team ball.

Hornets only had 2 lineups used much more than 1 minute per game for season next year. Neither still possible.

Batum - MKG were +3 per 100 possessions last season, one of their better pairs, and a lot presumably with Walker. But if they want to win games, they might see if they can find a way to fill around that for similar results.

If it starts going bad, and it likely will, they might want to make a big break from the past and trade older guys or just play the younger guys more and against tougher competition and see what they have.

For Lakers, LBJ- Rondo was one of James' worst raw plus minus player pairs. Playing James at PG MAY minimize that time this season. Will it? LBJ-Caruso was great on raw plus minus. How much will it be used? Rondo - Caruso was terrible. Will it persist or be reduced?

Adjusted plus minus pair data may be more reliable. Do the Lakers have it? How many teams do and give it consideration?

Best Celtics lineup?

Maybe Walker - Smart- Tatum - Williams - Theis.

Best Clippers lineup? Maybe Beverley - George - Leonard - Harkless - Harrell.

Kings should test Fox Hield somebody Bjelica Dedmon a lot. With Bodganovic, Barnes or Jackson and maybe others but not Bagley. And try these guys with somebody introduced for Hield.

I could do a whole rotation but it would be tentative and I'm not that motivated here and now.

But in terms of minutes, I'd say it should be something like (when healthy and not slumping):

Fox 35
Hield 30
Joseph 13
Bogdanovic 25
Bjelica 25
Barnes 20-25
Ariza 10-15
Bagley 15-20
Dedmon 25
Giles 10
Holmes 10
Rest of team 7-22

Full rotation featuring a starting lineup of Fox Hield Barnes Bjelica Dedmon achieved with 5 lineups. Different ways to finalize it. Later.

4 lineup rotation for Suns:

Rubio Booker Bridges Saric Ayton 20 minutes
Ty Jerome Tyler Johnson Oubre Diallo Baynes 14
Rubio Booker Bridges Diallo Ayton 8
Rubio Booker Oubre Saric Baynes 6

Rubio 34 Booker 34 Bridges 28 Saric 26 Ayton 28 minutes
Ty Jerome 14 Tyler Johnson 14 Oubre 20 Diallo 22 Baynes 20


Minutes are a little tight. Could easily switch in substitutes in a few spots for more minutes if someone doesn't fit / perform.

This isn't necessarily the very best, just feasible, worth considering. Modify with evaluation of new data.

Could add in Cam Johnson and / or Kaminsky for minutes in garbage time or beyond; but I'd start with a 10 man rotation and try to refine it before spending much time on the deep bench guys.

I might actually do different than this suggested rotation. I tried to stay within fairly likely acceptable to team range.

If they want to maintain some Booker at PG time, then things would have to change.

Ayton - Baynes might be considered. I didn't go there in this first draft.

One way to start out for Blazers:

Lillard McCollum Hood Tolliver Whiteside 14
Lillard Bazemore Little Collins Whiteside 8
Simons Bazemore Little Collins Whiteside 6
Simons McCollum Bazemore Hezonja Collins 10
Lillard McCollum Trent Tolliver Gasol 10

Lillard 32 McCollum 34 Hood 14 Tolliver 24 Whiteside 28
Bazemore 24 Little 14 Collins 24 Gasol 10 Simons 16 Trent 10 Hezonja 10

Adjust as needed, especially among the forwards.

Things will change as season progresses, when Nurkic comes back and due to performance assessments and trades.

James Green Kuzma Davis McGee could be a very good or great lineup for the Lakers. How much will it get used? Probably should be used very heavily. Good chance it will be used less than that.