APBRmetrics

Spurs should try a modest number of lineups to a meaningful level. That should include an all young guy lineup and maybe an all "old guy" lineup. Against starters and bench units. I would not assume which goes against which and not test the alternatives. Young perimeter / old interior is most likely but will it be best? We will how fairly and rigorously they test and what that shows and what that leaves vague.

TWolves did not play any lineup for 4 minutes per game for season last year.

I'd probably play Teague Culver Wiggins Covington Towns 20 minutes per game. Maybe remove Wiggins but try him to insulate him in this group. Probably best place for him if you have him (by foolish huge extension).

If coaches are not going to test and prioritize big minute lineups as much I think they should (and even if they do), they should get help from analytics in picking situational lineups / matchups with a rationale for being above average or outstanding. Using RAPM to find lineups / matchup that jump out or types of lineups / matchups that do. It will be even harder than the work with lineups with more testing but if they are going to guess a lot, guess with analytics. May want to look beyond one year, one team. I'd think it comes down to factor strengths, not totally unique performances trends. Either you get something from this or you have even more incentive to find and utilize a handful or two of the bigger minute lineups.

if the coaches and / or analytics can find good to great lineups / Marcus based on knowledge of how plays work, great. But the current success rate is generally not impressive imo.

Celtics with no lineup used over 5 minutes per game? Everything a unique situation or don't know what is "better / best"?

They lost their starting center between game 1 and 2, and their backup center between game 2 and 3.

If you look at their 4-man lineups the top 9 are a combination of the same 6 players.

https://stats.nba.com/lineups/advanced/ ... 1610612738

Thanks for the new lineup information. There is a bit more concentration at 4 man level between now and when I commented before game 3. So Stevens went the direction I suggested to some degree. Lineup concentration is usually the better approach. At 4 man level is good / easier, at 5 man level not as common but probably even more important. Concentrate, injuries or no injuries. We'll see where they are later.

Of the 24 lineups with more than 20 minutes play before today, more 60% were positive.

Strongest among these were lineups of the Nets, Nuggets, Blazers, Spurs, Rockets, Heat, Sixers, Jazz, Lakers and Knicks.

Weakest included Pelicans, TWolves, Bulls, Pistons, Wizards and Kings. Kings by far the worst performing so far.

Bucks starters were slightly below neutral so far.

Clippers are only team with 2 positives over 20 minutes each.

Some of these lineups are in the range of 40% of available minutes. Probably few maintain that level but there is some rationale to try.

Subject to change of course.

As of today, Raptors are getting the most edge from their starting lineup along with Rockets. Raptors doing it with more minutes, Rockets with twice the pop in half the minutes. Blazers and Clippers next but about 50% less minutes than Raptors and 50% less impact.

Knicks and TWolves hurting themselves among the most with their starting unit. Kings and Magic at same level of damage with twice the hurt level in half the minutes (an insanely bad -55 points per 100 possessions each).

Knicks fell back a lot since a few days ago.


On average teams have used 43 lineups. Most lineups used by Atlanta, Charlotte, Phoenix and Detroit with about 60-70.

A surprising 40% are positive. Probably influenced by small sample size. Expect that to fall but worth watching.

The average lineup has been used about 1 minute per game. That will fall massively if coaches follow last season and try about 15 times more new lineups the rest of the way from here than the ones they started to toy with. The average lineup was used about 6 minutes for the entire season last season.

Rockets, 5 of 7 most used lineups cumulatively negative. 82 lineups used. Only 3 used over 2 minutes per game. 38% positive.

Heat, 8 of 10 most used positive. 63 lineups used. 6 over 2 minutes per game. 48% positive.

Lakers, 8 of 10 most used positive. 66 lineups used. 5 over 2 minutes per game. 48% positive.

Mavs, 7 most used negative, most horrible per 100 possessions. 69 lineups used. 4 over 2 minutes per game. 48% positive.

Nuggets, 5 of 6 most used negative, bad to horrible per 100 possessions. 37 lineups used. 5 over 2 minutes per game. 41% positive.

Spurs, 7 of 8 most used positive. 56 lineups used. 5 over 2 minutes per game. 43% positive.

Blazers, 10 of 12 most used positive. 52 lineups used. 5 over 2 minutes per game. 42% positive.

Sixers, 5 of 5 most used positive. 51 lineups used. 4 over 2 minutes per game. 45% positive.

Raptors, 4 of 8 most used positive. 43 lineups used. 5 over 2 minutes per game. 47% positive.

Clippers, 6 of 8 most used positive. 56 lineups used. 5 over 2 minutes per game. 36% positive.

Celtics, 6 of 7 most used positive. 85 lineups used. 2 over 2 minutes per game. 46% positive.

Bucks, 5 of 10 most used positive. 74 lineups used. 3 over 2 minutes per game. 42% positive.

Jazz, 5 of 10 most used positive. 50 lineups used. 6 over 2 minutes per game. 36% positive.





Thunder, 7 of 10 most used negative. 58 lineups used. 4 over 2 minutes per game. 31% positive.

Rockets with one of the worst performances sets for their most used lineups among these top teams (OKC included as an aside). One of the highest number of lineups. Fewest lineups over minutes. Lowest positive rate.

They do have a good big minute starting lineup. Though it is median in performance for lineups over 50 minutes.



Wizards with new players and a new assistant coach and maybe a new approach? 3 most used lineups terrible performing so far and 5 of 7 terrible. 85% of the 20 most used player pairs are negative and most are quite bad.

Raptors, Nuggets, Clippers and Bucks with among the best big minute units for good or potential good teams.

Lakers and Twolves with among the worst big minute units.

Doncic - Porzingis -4.5 pts per 100 possessions overall.

No Mavs lineup used over 24 minutes so far. That is whack but I guess Carlisle does not know what is best (yet). Now it is 8 of most used 9 are negative.

Make it 9 of 10 most used negative for Mavs now. Bench players mostly have the positive plus minus so far. Doncic 6th on that at plus 2.4 pts / 100 possessions.

Knicks, 92 lineups used. Only 8 more than twice in 8 games. 6 of the most used 8 were bad to extra horrible.

Best by far, far performing bigger minute lineup is R. Barrett | M. Morris | E. Payton | B. Portis | J. Randle. Has not been used as a starting lineup yet. Should when Payton comes back.

Lot of emphasis on Robinson. Bad results so far.

Main positive is offensive rebounding.

Cavs starters playing 15 minutes per game. That should not be considered that remarkable but it is 3rd most total minutes in league. Up and down but doing well on average, surprising me. Almost entirely based on advantages at ft line.

On average teams have played 1.4 lineups for more than 30 minutes (or about 4 minutes per game). Basketball is going to be chaotic with 5 * 5, even more when is almost all low frequency, low understanding how to work together lineups. And in a hurry.

Barely half of these lineups are cumulative positive so far, but that is better than the lower minute lineups alternative. They are doing it with better 2 & 3 point shooting than opponents. More fts made but at a a bit lower ft%. Playing at a pace almost 3 possessions slower than league average. About 2.6 guys 6-8 or taller on court on average. (Some 2s, a few 4s.)

These lineups average +4.1 per 48 minutes. Everything else average -2.2. Which group are you going to pick a lineup out of to put on the floor now? I'd put the better performing bigger minute lineups on the court a lot more. Coaches? Nah, lets mostly try to win the hyper match-up lineup game. Lets try to be the mystical magician, not some slave to the analytics. The pay is better. More drama. Can use all the ad-hoc, subjective, thinking hard on the fly faces.