Building an NBA team at the Sloan Sports Analytics Conference

BOSTON—Jeremy Lin is a 7. Ideally, the Knicks would have him playing with an 8 and a 12. Instead, he is with a 2 and a 5, and the result is likely to be a pretty good but not great Knicks team.

This is according to a research paper by Robert Ayer of MIT, presented here at the annual Sloan Sports Analytics Conference. Ayer went through every NBA season since 1977 and broke down what combinations of top players on a team’s roster had the biggest impact on a team’s winning percentage—you could have an incredibly talented Top 3—say, Rajon Rondo, Russell Westbrook and Derrick Rose on your team, Ayer pointed out, but that combination of three star point guards would not help you win more games than expected.

What Ayer found was the best combination for any team’s Top 3 players is a high-scoring, high-assist, high-turnover point guard (Ayer grouped players into “clusters,” and those players were Cluster 7); a, “multi-faceted, high scoring wing, with high assists for his position and a great 3-point shooter” (Cluster 8); and a dominant center (Cluster 12). Over the years, teams that had that grouping—Ayer points to the 2009 Orlando Magic team that went to the Finals with point guard Jameer Nelson, small forward Hedo Turkoglu and center Dwight Howard—tended to overachieve by a dramatic coefficient (plus-13.6).

The next best grouping was a Cluster 8 small forward, a Cluster 12 big man and a Cluster 10 player, a power forward focused on defense and rebounding, in the vein of Dennis Rodman or P.J. Brown. That group’s coefficient was plus-5.4, emphasizing just how much good the “7-8-12” group is. “If you have those three types as the best on your team, you’re going to be in the best position to win more games,” Ayer said.

Seeing team-building through the prism of Ayer’s research could also help explain the Lakers’ willingness to trade Pau Gasol. L.A. is hoping that its center, Andrew Bynum, will develop into a “12,” that rare dominant center. The best players to pair with a center of that type are Cluster 10 dirty-work power forward or a Cluster 8 versatile small forward of the Paul Pierce type. Keeping Bynum with an offense-minded power forward is fine because both players are talented, but that combination does not push a team above its expected win mark.

The same can be said for Lin’s Knicks. A point guard who can post the kind of numbers we’ve seen from Lin over the past month does best with a versatile small forward and a dominant center. Instead, Lin has a scoring small forward and a scoring power forward who is not a good rebounder. That combination does not register as one that has a significant impact on winning. That doesn’t mean the Knicks can’t have some success this year—it just means that a Big Three of Lin, Stoudemire and Anthony, based on statistical history, isn’t likely to exceed expectations.