APBRmetrics

http://bkref.com/tiny/CdNCt
Miami with 2 of the 3 best lineups used over 50 minutes in playoffs. Indiana has a nice one good for 7th. Memphis 8th.

Spurs' best is only 12th and only modestly positive. Most of their positive results are coming from small minute lineups with a healthy dose of bench guys in them. http://www.basketballreference.com/team ... 3/lineups/
Last season they had the best playoff starting lineup, with Diaw instead of Splitter.

Thunder's starting lineup with Perkins was by far the worst performing. That is 4 straight years that the Thunder have had a starting lineup that was in the bottom 25% of lineups used over 50 minutes in the playoffs and the second time they were the clear cut worst.

No team has won the title in last 12 years without a best lineup used over 50 minutes in the top 6. Spurs' best is currently 11th, improved by last game or two to almost +7 per 100 possessions. But they are still bucking the recent history. Their starting lineup was great in the regular season (+19 per 100 possessions). Just not performing near as well now and / or better opposed.

Something about small sample sizes here. Watching the Spurs play I would say they have yet to kick into the high gears, having more or less coasted through to the finals off the back of Westbrook's injury.

Sure small sample size is a consideration. But if this lineup data were totally meaningless and random you might expect a wider / wilder pattern than every single title winner in 12 years in the top 6 and most in the top 3.

It appears that it is typical that about 10 lineups overall get over 100 minutes in the playoffs and one or a few get over 200 or 300 minutes. For the title winner the minute level is often considerably higher than a mere 50 minutes. That was just the screening cutoff chosen to try to fairly include teams that only played one round or two.

If the Spurs win, it may be that either their top lineup performance gets better or it may that they continue to rely on the top lineup less than other winners did. By noting the data, I made this fork in the road known so one can watch to see whether they take one of these winning paths or lose and possibly lose to some significant degree because of their weaker than normal top lineup performance ranking.

In the season, Spurs principal starters -- Duncan, Parker, Leonard, Green, Splitter -- played 52% of the team's minutes. This was 22nd in the league, in this ratio.
These 5 started 81.5% of the 82 games, which was 14th.
They totaled 32.9 eWins, which was 5th among starting 5's; this was 67% of team eWins, ranking 14th.
Their starters' eWins per 484 minutes totaled 1.54, which was 3rd (after Mia and OKC)
Team total eWins (and Pyth wins) ranked 4th, behind also the Clippers.
So the Clipps were even less of a starter-dependent team.

Looking at the regular season data with possessions played together > 400 (which I consider to be a good starting point for a credible non-bayesian estimate, and it's fine if you disagree) we find the Spurs starting five at number two, only behind the Anthony, Chandler, Felton, Kidd and Smith monster. Below the Spurs is the Miami starting five, neck and neck. There is then a serious drop-off to any other lineup with a decent sample size after those two.

There is more of the regular season data of course and it should be considered. I just wanted to see what the playoff data said since I had previously looked at the regular season data some time ago.

Crow wrote:No team has won the title in last 12 years without a best lineup used over 50 minutes in the top 6.

Is this regular season top 6 or postseason top 6?

Postseason top 6.

I had some months ago found that the title winner's best lineup used over 250 minutes averages about the 4th-5th best from the regular season.

Crow wrote:Postseason top 6.

I had some months ago found that the title winner's best lineup used over 250 minutes averages about the 4th-5th best from the regular season.

That wouldn't mean much since we would expect the champions top lineup to do well hence the reason why they won a title. I'm more curious is if you can use regular season lineup data to be more predictive of playoff success than SRS. Or if you could predict the finals better by comparing the top playoffs lineup plus/minus for each finals team.