“I’m not smart enough to do the math.”
- Jeff Van Gundy
From an analytic expert's perspective, the ideal coach understands all the numbers. He gives minutes to the team’s most productive players in the ideal lineup combinations. He plays the odds, regardless of potential public backlash, every time.
From a player’s perspective, the ideal coach relates. He pushes the right buttons to inspire and motivate. He understands the unique challenges facing professional athletes.
The ideal coach, to both groups, also wins. In a perfect world, there are coaches who would satisfy both camps. But I’m not sure how many people have the technical prowess to use analytics properly and the charisma to get their players behind him.
TrueHoop at MIT Sloan Sports Analytics Conference
Jeff Van Gundy has been the star of Day 1 of the Sloan Sports Analytics Conference. During his two panels, he had the crowd smiling and laughing. Daryl Morey, a co-chair of the event, joked that he needed to get Van Gundy on every panel. It’s easy to see why players would like him. I can’t imagine players happily spending hours and hours during the season answering to some of the presenters of the conference’s more in-depth papers.
This is not to lump people exclusively into the nerd-or-macho paradigm. All the paper presenters I saw spoke eloquently. And many who’ve spent time playing in the NBA -- often a perquisite for coaching -- are very smart. But spending the hours necessary to excel in statistical analysis or playing basketball rarely leaves enough time to thrive in the other field. Leaving this merged task up to a single person is asking coaches to fail.
As Van Gundy explained after the panel, he didn’t run numbers when he coached the Rockets. He just got them from Morey -- using them when they fit his message, and when they didn’t, lying to his players about the numbers. To Van Gundy, statistics served his means rather than informing his ends. Even though Van Gundy didn’t completely trust the relevance of all Morey’s numbers, their relationship might illustrate the framework of the ideal setup, where numbers, from their origin, are passed down and simplified and passed down and simplified until they reach the coach.
But even that scenario contains a number of potential challenges. The chain must contain trust at every link, and the longer the chain, the more opportunities for a breakdown. Even when a coach completely trusts the analyst, the data can be dangerous in the wrong hands.
Once, Van Gundy asked Morey for the analytical answer to how many minutes Yao Ming should play. Morey determined, statistically, Yao’s ideal minutes per game was 48. No matter how tired Yao got, he was better than the alternative. Mindful of other factors -- such as Yao’s long-term health -- the Rockets obviously didn’t act on that finding.
Detroit Lions defensive end Lawrence Jackson said he spends time on his own analyzing plays, looking for the big trends, the offensive lineman who put their foot back on 90 percent of pass plays from a given formation. But what about the lineman who puts his foot back 54 percent of the time? Areas like that are where a more-trained statistician could flesh out which stats are significant and, ideally, improve a team's chances of winning.
Someday, will that statistician be Jackson’s coach? Would that make a team more likely to win?
By the way, Jeff Van Gundy enrolled at Yale out of high school. Whether or not he was being slightly self-aggrandizing in that lead quote, basketball statistics have reached such a level of dizzying complexity that they sail above the heads of Ivy Leaguers.
Professional basketball requires leaders who can deliver the message in a way players respond to, but it also requires leaders who can use numbers to determine the right message. Even a coach could tell you finding someone who satisfies both traits isn’t statistically likely.