Editor's note: This story was published prior to the reports that Derek Fisher accepted the Knicks' coaching job.
So you're looking to hire Derek Fisher as the Knicks' head coach? Maybe you shouldn't have said that while he was still playing. Next time, just leak it through agents like everyone else.
How do you know that Fisher is going to be good? How does anyone know whether a coach is going to be good?
Evaluating coaches is very hard to do -- that's especially true for a coach with no track record. Evaluation usually takes the form of a simple rule: Keep a coach whose team exceeds expectations, but release a coach whose team doesn't. That simple rule doesn't even acknowledge whether the coach actually had a significant role in the team's accomplishment.
In contrast to players, who put up statistics and whose performance is very visible on the court, coaches don't show a lot of things reflecting their performance. Good players make shots during a game, but good coaches sit, stand, cross their arms, scowl, yell, whistle and sometimes intentionally spill water on the court (tsk, tsk, Jason Kidd) -- none of which is recorded systematically because no one knows if any of that actually matters.
What does matter?
During my years in a front office, I had some opportunity to build models of how coaches contribute to wins and losses. I also worked with people who had been around basketball for many decades and we talked a lot about what makes up good coaches. Those conversations went a lot of different directions, reflecting the numerous roles a coach has. I tried to record what we talked about and what other people have said over the years to form some model of what a coach needs to do to be successful. This is relevant to you, as you try to figure out who to bring in to coach the Knicks.
The basic framework is that an organization's success is based upon knowledge, communication and execution. None of these will ever be perfect, but if one of these pieces is deficient, it drags down the entire organization. A head coach is central to linking knowledge to execution. A coach has to have knowledge, he has to communicate both up to management and down to players, and he ultimately is responsible for players' execution.
Let's start with knowledge. The stereotypical knowledgeable coach is an encyclopedia of plays and defensive schemes (think of Jon Gruden). But that is only a small part of the job.
A good coach knows how to choose the right plays and the right schemes from that encyclopedia. I have been around coaches with the entire encyclopedia in their heads, but who can't tell you when to use what or why. A good coach knows how to fit schemes to players and vice versa. A good coach knows which schemes will counter tactics made by opponents, plus how those opponents will likely adjust. In a regular-season game, certain tactics may be effective, but in a long playoff series, a coach needs to know how to counter a series of adjustments by opponents. Players see how a coach does this, and coaches gain or lose respect based on these situations.
Another vital aspect of coaching knowledge is evaluating personnel. George Karl was very easy to work with in Denver because he understood the quality of his players. We didn't have battles over playing time because, even when J.R. Smith drove him crazy, George knew how many minutes to play him and what role to give him so that he was helping the team. Fitting players into roles that help the team and making them feel valuable for doing those roles -- that was one of your major strengths as a coach, Phil.
A good coach understands the difference between how good a player is now and how good he can be. A coach needs to know how to give young players non-garbage minutes during the season, so that they're ready for competitive minutes later in a season or later in their career. In doing so, a coach should have an answer to the question, should we improve this player's weaknesses or his strengths?
After knowledge is communication. "Communication" is usually just a buzzword, used far too generally as a code word for "I like him" when hiring a coach or "I hate him" when firing a coach.
But beyond buzzwords, communication for a coach is turning his own knowledge into player and management knowledge.
For example, what plays and schemes to run has to be conveyed day in and day out to players. A head coach and his staff have to tailor and narrow down the game plan to the players to convey what's important. You can't expect players to take in a comprehensive scouting report before Game 37 in the middle of January. In the playoffs, when the players know their matchups a lot better after Game 1, a coach needs to listen a lot more to what his players see and feel on the court and put it together with his own study.
A coach should admit mistakes, but not make so many of them that he loses credibility. This is basic.
To management, it is really helpful when the coach communicates information on his schemes, system and perceptions of good/bad players. Management that knows what is going through a coach's head can much better separate execution on the floor from what the coach is communicating. It helps set expectations that often define how long a coach will keep his job.
In my discussions with old-time basketball people, some surprising aspects of coach communication came up. One of them was that the actual sound of the voice of a coach is very important. If it is too bland, too excitable or too annoying, players will tune it out. This makes some sense, but still surprises me a bit.
The other important note is that a coach should "act human," having a life outside of basketball and recognizing that players also have lives outside of the game. Giving players multiple dimensions in which to relate to a coach helps get through the doldrums of a long season. Phil, you were known for having a broader life and talked about it in "Sacred Hoops," so I'm sure you appreciate this.
A coach doesn't shoot, pass, dribble or defend, but the players' execution is something for which a coach does take responsibility. Publicly, a coach takes responsibility for failures and gives credit for success.
There have been analytical attempts using on-court execution to measure coaching performance. Some of those attempts have been pretty good, and some methods suggest that the best coaches, like you, can add double-digit wins per season. Of course, these methods are attributing player improvement (or decline) to a head coach, when actually it could be due to assistant coaches, athletic trainers or, according to what I watch in commercials, brands of shoes or sports drink.
None of these analytic methods really explain why. That is, what is it that a coach is doing to make the team good? How much of it is knowledge versus communication? Do the best coaches work well with superstars or develop supporting players? That type of analytics -- prescriptive analytics -- can tell you not only how to hire a coach, but how to make a coach better. Unfortunately, until that kind of analytics is done on coaching, hiring coaches really requires asking a lot of questions that get at their knowledge and communication. Or, according to common wisdom in the NBA, it means hiring people that you know and respect, like Steve Kerr or, shhhh, Derek Fisher.
Dean Oliver is the ESPN Director of Production Analytics and the author of "Basketball on Paper." He was previously director of analytics for the Denver Nuggets.