The Sacramento Kings fired their head coach, Mike Malone, on Sunday after the team got off to a sub-.500 start this season. If that doesn’t sound earth-shattering, it’s because the Kings have been lousy for a while. They haven’t finished a season above .500 since 2006 and have had seven head coaches (including Malone and his interim replacement, Tyrone Corbin) over that span. Yes, the Kings fired another coach. So what?
The Kings’ abrupt decision to fire Malone is a big deal because it flies in the face of what tends to keep coaches employed. In April, FiveThirtyEight’s editor in chief, Nate Silver, looked at NBA coaching dismissals and found a strong relationship between firings and whether the team underperformed the preseason Las Vegas over/under win totals. Win just 41 games with a team expected by Vegas to win 50? There’s a roughly 50 percent chance you’ll get fired after the season. (Sorry!) But win 41 games with a team Vegas expects to win 30? Nate’s model says there’s just an 8.2 percent probability of being canned for that performance.
The latter is what Malone was on pace to do. In October, the Westgate Las Vegas Resort and Casino listed the Kings’ over/under as 30.5 wins, while the offshore sportsbook Bovada set that number at 29.5 wins. (Our projections also called for 29 wins.) In other words, the Kings were essentially expected to be a 30-win team. And yet Basketball-Reference.com’s playoff forecast projects them to win 41.2 games by the end of the season, based on a combination of their point differential, schedule strength and remaining opponents.
(It’s also worth noting that the past nine games of Malone’s Kings tenure were spent without his best player, DeMarcus Cousins. Cousins was having a lights-out season before coming down with viral meningitis in late November. It’s likely the Kings’ win projection would be even stronger had Cousins been active in the team’s most recent games.)
Piloting a team with 30-win talent to 41 wins typically earns a head coach job security 92 percent of the time, but that wasn’t enough for Malone. Of course, this eventuality is built into the model, which would be wrong if 8 percent of coaches who exceeded expectations, like Malone, didn’t get the ax. That 8 percent can, in part, be explained by other aspects of a coach’s performance that don’t involve simply winning more than was expected.
For instance, the Kings’ brain trust (owner Vivek Ranadive and general manager Pete D’Alessandro) reportedly took issue with some of Malone’s stylistic decisions, such as the pace at which the team played and the emphasis on defense over offense. Some have even suggested that Malone’s not having been hand-picked by D’Alessandro hurt his cause substantially.
For what it’s worth, I couldn’t find a statistically significant effect for how being hired by a different GM affects coaching dismissal rates. Using Basketball-Reference’s database of NBA executives, I looked at everyone who was an NBA coach (but not in a dual role as GM and coach) to begin a season since 1985-86, setting up a regression to predict whether someone would be fired before the following season began. As Nate found in his research, wins against expectation (in this case, set using a regressed-to-the-mean version of the team’s Pythagorean record from the previous season) was a highly significant predictor of employment. But a dummy variable for whether the team’s current GM was the same GM who originally hired the coach was nowhere near significant.
In any event, Malone’s firing was nothing if not an outlier, according to the metrics that tend to predict whether a coach will hold onto his job. It goes to prove that, in the NBA, no job is truly safe.