Nebulous denoters of player quality litter basketball vocabulary. There are “stars”, “All-Stars”, “All-NBA” or “max contract” players, while lesser luminaries are “rotation” players or “solid”. Closer to the periphery are “end of the bench” or “fringe” players. Without uttering a number or a firm ranking, there is already an entire taxonomy of player quality terms.
At one point in time “starter” and “bench” were clear markers of quality, an acknowledgement of whether a player was one of the five best on his team or not. Perhaps it is fitting then, that in a post-positional revolution, modern analytics era that these terms no longer mean anything concrete, and are watered down with phrases like, “he wouldn’t be a starter on a playoff team”.
Starting your five best players and playing them the greatest amount of minutes may seem sound strategically, but it is also tactically inflexible and doesn’t take into account the myriad of interactions between individuals and groups of players. It also creates a huge drop-off in play near the end of the first and third quarters, when those starters inevitably need a rest. In response, borrowing from the team tennis concept of stacking—where a team will take a near-certain loss in order to create the most favorable cross-matches possible—an ever increasing number of teams are having some of their better players start the game on the bench.
This strategy is most commonly used by teams that want one of their better ball-handling scoring guards to run the second unit offense: think Manu Ginobili in San Antonio, James Harden in Oklahoma City or Jamal Crawford in Los Angeles. Sometimes teams will go to extreme lengths to keep these players from starting, as in the case for the Clippers where they upgraded Willie Green from purgatory to the starting unit when JJ Redick broke his hand in solely to keep Crawford on the bench.
But instant offense isn’t the only reason the starter/bench distinction is impossibly muddled. Bigs like Kendrick Perkins and 2010–13 DeAndre Jordan start games but rarely finish them for fears about how they drag down the offense and whether they will be fouled. The Miami Heat have won two championships taking advantage of their depth in three-point shooters to end games surrounding LeBron James, Dwayne Wade and Chris Bosh with whichever shooters happen to be rolling, whether that be Mario Chalmers, Norris Cole, Mike Miller, Shane Battier, Ray Allen or somebody else. Other teams start veterans out of deference, knowing that it will create a more harmonious locker room if those players are happier being “starter”, whether or not they play the most minutes.
These tactical maneuvers by individual teams are underscored by a more structural strategic shift in league-wide thinking. Earlier this year Henry Abbott wrote about the dearth of 20-point scorers, and posited that some of the explanation lied in a decrease in minutes for top players. It’s not just top players that are playing fewer minutes though: on the whole, Poppovich-ian rotations are on the rise everywhere. Throughout the late 1990s and most of the aughts, the players that started the game averaged around 32 minutes per game. In the last few years though, that has decreased to 30. Two minutes a game might not seem like much, but over the course of a full season that’s 164 fewer minutes per starter. If Michael Finley had played 164 fewer minutes over the course of the 1997–98 season, he would’ve played the fifth most minutes in the league, instead of leading it by 100 minutes played.
Though the per game impact of shifting minutes may seem small, this should be seen as a revolutionary—not evolutionary—transformation. Evolutionary describes slow, incremental changes, revealing a system not so different than the previously existed. But the league-wide decrease in minutes happened quite suddenly, and current tactics are blowing apart the old orthodoxy.
The old orthodoxy said that bench scoring was important, that you needed a scorer to lead the second unit. That is still the most prevalent bench role for good players because “scoring” is still the easiest and most widely understood metric, but an ever increasing emphasis is being placed on who closes games, top five-man units and other ways players fit together. That’s because there is absolutely no evidence that bench scoring has any relationship with wining basketball games. Going back to the 1997–98 season—the last season with good bench data—there is only a 3% correlation between winning percentage and the proportion of a team’s scoring that comes from players that don’t start. What’s important is scoring points, not necessarily where those points come from.
Across the league starters are playing fewer minutes. Across the league teams are de-emphasizing the importance of who is standing on the court for the opening tip. It raises the question: is it time we stop describing players as “starters” and “bench players” altogether?