Deep Dives: More on Matchups

facebooktwitterreddit
NC Deep Dives
NC Deep Dives /

Last week, I examined individual players who performed better or worse versus bench units than against starters. While it only stands to reason working against starters would be more difficult, the raw numbers don’t show it, at least not in terms of shooting. leaguewide, eFG against “starting” units was slightly higher than against “bench” units. Of course, I was missing half the equation, as who one plays with often as if not more important than who one plays against.

Further, a frequent point of coaching criticism is losing the matchup better, allowing bench-heavy units to get overwhelmed by the opposition’s top five. I was curious to see how different combos of starting and bench lineups bumped up against each other. Obviously, there are a lot of confounding factors as not all bench units are created equal, with a big difference between a lineup including LeBron and several Cavs reserves and that same lineup including Timo Mozgov and a similar number of bench players. Still, dealing with the aggregate of the entire league, the general magnitude of lineup-based advantages should be discernible.

To prove this, I used combined play-by-play and lineup data to count the number of starters on the floor for both teams at all times and tracked net scoring differential, here’s some of what I found,

Starters do play better:

That why they’re starters. But seriously, here’s the time spent on the floor with 0-5 starters and the net +/- per 48 minutes[1. I used 48 minutes rather than the more common per 100 possessions to simplify the coding, as counting “possessions” can get complicated in these terms with substitutions in the middle of a possession and before/between free throws.] for each category:

Dashboard 1 (7)
Dashboard 1 (7) /

Mostly as expected, the fewer staters in the game, the worse a team plays, aside from the little weirdness between 3 starters and 2 starters. Perhaps there slightly different use patterns so that 3 starter lineups play more against all starters than more bench-heavy units? This matters because…

Having More Starters on the Floor is an Advantage

Again, not earth shattering, but here’s the detail. Note that for “zero” advantage, situations where teams have an equal number of starters in the game, the net rating is the amount per 48 minutes the home team outscores the visiting team, while the others are the advantage enjoyed by the team with more starters on the floor:

Dashboard 1 (8)
Dashboard 1 (8) /

A few things to point out – first, nearly 80% of minutes accrue with lineups reasonable balanced in terms of starters. This perhaps indicates NBA coaches are already pretty good at matching lineups, NBA substitution patterns are somewhat rote, or some combination of both. In any event, given the returns to starter heavy units working against benches, maybe there are some advantages to be gained here through clever in-game lineup management?

More Detail

Breaking it down by specific matchup types, comparing number of home starters to number of visitor starters. The +/- indicated is the net points per 48 minutes for the home team, so negative numbers indicate a road team advantage:

Dashboard 1 (9)
Dashboard 1 (9) /

As a reminder from above, situations with more than a “2 starter” advantage are comparatively rare, so some of the oddities in the chart are probably explainable by small[1. And perhaps biased samples as well as teams which spend a lot of time with large “starter” disadvantages might be teams bringing one or more of their preferred “closing” lineup off the bench.] sample sizes.

To repeat one last time, this is all from a 30,000 foot view. Individual team circumstances almost certainly overwhelm these general rules. Still, it does give a possible basis for evaluation if a team is consistently getting outscored in a certain segment of the game, perhaps they are giving up too much of a lineup advantage away without bench players of sufficient quality to make it work? But mostly, it’s confirmation that minute and starter distribution[3. As a digression, I hope this analysis becomes obsolete if an as the laudatory effect of “starting” declines. With luck this trend will emerge in response to all the positive press and Finals MVP received by Andre Iguodala last season. Especially in the regular season, there is no better way of controlling the minutes of an aging and perhaps fragile player than by bringing him off the bench. Well, aside from giving him Popovichian DNP-old’s. It could also be a way of better taking advantage of certain players better abilities to take advantage of lesser opponents by maximizing their floor time in those situations by keeping them away from starters, even if it means starting an ‘overall’ inferior player.] is reasonably efficient already.