knowquestion :: Stephen Bounds

I was a little surprised that Tom Davenport's post about basketball analytics and the New York Times story it was based on didn't get better coverage in the KM community.

Essentially, the Houston Rockets and specifically their General Manager Daryl Morey have found a way to quantify what many people thought was unquantifiable: a person's contribution to the team.

Although their techniques are a pretty closely guarded trade secret, its clear that a key part of their method is to evaluate players based on their plus-minus score. A plus-minus score evaluates how a player's presence on or absence from the court affects the outcome of those few minutes.

This allows for even headline "bad" results to seen in a positive light. Perhaps the team loses a 10 point lead over the course of a quarter - but if it can be demontrated that without Player X, the net loss would have been 20 points instead, Player X has a +10 score.

On the other hand, even a brilliant player might have a low rating if their careless or selfish play leads to more turnovers.

By taking the focus from headline metrics (points scored, assists, possessions, steals, rebounds etc) and instead looking at adjusted benefits or losses realised through participation or absence, it is no longer possible to benefit a single entity at the expense of others. Only a truly team outcome-focused approach will increase the plus-minus metric.

Basketball is, of course, comparatively simple compared to real-life organisations. But just imagine the impact if we could evaluate hospitals on a plus-minus metric. No more buck-passing or statistics inflation to meet performance targets.