Bullets from the basketball analytics panel

Four years ago, I attended MIT's Sloan Sports Analytics Conference for the first time. At that point, I was embarrassed to tell my geeky friends (sorry guys) exactly what it was that I was attending. Since then the event has morphed into “dorkapalooza” and I have come to embrace this conference and basketball analytics.

Given that the founder of MIT Sloan Sports Analytics Conference is Houston Rockets GM Daryl Morey, I considered the “Basketball Analytics” panel to represent the central theme of the event, even though the conference has grown to embrace panelist from a range of fields and sciences. Here are the most salient talking points from this year's foray into advanced basketball statistics.

The Panelists consisted of a group very familiar to basketball the Sloan Conference:

  • Mark Stein(Moderator), ESPN Senior NBA Writer

  • Mark Cuban, Dallas Mavericks Owner

  • John Hollinger, ESPN Columnist

  • Kevin Pritchard, Former NBA General Manager

  • Mike Zarren, Boston Celtics Assistant Executive Director of Basketball Operations

General comments from the panel

It was reported that 20 of the 30 NBA teams utilized advanced statistical analysis.

TrueHoop at MIT Sloan Sports Analytics Conference

When asked how often coaches simply ignore statistics, Mark Cuban indicated that only Gregg Popovich and Phil Jackson were able to get away with this. I suspect he was simply stating that these were the only two coaches with enough power to get away with this.

Cuban went on to explain that analytics are predominantly useful for risk management. He stated the three things that his team analyzes for every trade:

1) How the deal helps the current team

2) How the deal impacts the team’s financial situation

3) How the deal impacts the future team


To what extent are analytics involved in NBA front office decision making?

  • The panel indicated that everyone knows who the league's best players are and the challenge with those players is figuring out how to acquire them. Analytics provides value by helping to find talented but less known players.

  • Pritchard estimated that five years ago about 5% of the decision making process involved statistical analysis, but that rate seems to continually increase, but that it has never replaced the visual aspect of personnel evaluations nor did he expect it to.

  • 5 years coincides with the first Sloan conference. One wonders how important the Sloan Conference has been in educating late arriving teams to the wonder of basketball analytics.

  • Mike Zarren observed that the first year the event was held consisted of an entirely different crowd, joking that 93% of the attendants from the first conference wore pocket protectors. I think that using 93% instead of 95% or 90% confirms that Zarren belongs in this group. I’ve attended for 4 straight years, so it should be pretty safe to include me as well, preferably minus the pocket protector.

Regarding Minnesota Timberwolves General Manager David Kahn recent comment: “Analytics are less important for rebuilding teams”:

  • Cuban replied “I’m glad they think that.”

  • Hollinger stated that he thought analytics might be more helpful since there were more potential solutions starting from that level.

  • In my opinion, this depends on how you look at it. Hollinger seems right to suggest more potential solutions, but there is less statistical accuracy the further you project in the future. Others shared this sentiment, but that doesn’t suggest you should limit your resources.

What is your preference regarding the one and done rule?

  • Zarren pointed out that high school stats are basically useless since the level of competition could be so drastically different.

  • Pritchard said that he believed that the longer most players stayed in college the better it would be for the game, but he admitted that he liked the potential to land a major talent by picking a high school kid.

  • Hollinger stated that draft picks become more valuable when a player has more years to develop since teams learn more about the potential pick. The pick becomes less risky.

  • Hollinger is right about more mature players carrying less risk. However, assuming the goal is a title, risk is required since only 1 in 30 win a championship.

When do you concede a decision was wrong?

  • Kevin Pritchard suggested it depends on age. A player coming out of high school might require 4-5 years before you can give up on them. A 4 year college player like Brandon Roy needed to show something practically immediately. Pritchard also stated that it seems big men take longer to develop.

In discussion of the Lakers and Heat’s struggles against elite teams, the group had different takes.

  • Pritchard stressed the importance of consistency in roles for closing out games. He mentioned that Wade, James and Bosh all shared the role of a last possession shot and that this hurt player comfort levels in these situations.

  • Hollinger quoted the saying that you if you torture a stat badly enough it will tell you anything. Hollinger concluded his thoughts with an amusing reference to Tobia Moskowitz’ presentation on “The Real Reasons Behind Home Field Advantage” when he joked that Cuban’s team has been able to pull out so many close games because of referee bias.