Like the Most Valuable Player award, the Coach of the Year award falls victim to many contending philosophies. Your choice for the Coach of the Year will depend very strongly on how you define the term.
The differing rationales given for each award can be very astonishing particularly when combined in contradictory manner. For example, many MVP voters wouldn't even consider a player like the Chicago Sky's Sylvia Fowles. Why? Because in the voters minds, the team finished sixth in the Eastern Conference this year and by the artificial definition given by some advocates, this precludes her. "How valuable can she be," goes the argument, "if her team doesn't win?" (*)
When it comes to determining the WNBA'sbest coach, however, that rationale goes out the window. For a coach, having the best team in the league actually works against you. "With players so good on (Team X)" goes the new argument, "does it really matter who the coach is?" Voters tend to seek out up-and-comers for this honor. For example, Marynell Meadors of the Atlanta Dream received the award last year primarily because the 2008 Dream were one of the worst teams in WNBA history and to go 18-16 and reach second place in the Eastern Conference in the following year was a major accomplishment. Perhaps the performance of the Indiana Fever in 2009 worked against Lin Dunn and not for her.
So how do you measure how good a coach is? Part of the problem with coaching is that it is very difficult to separate a coach's contribution from the record of the team. The record of the coach is obscured in part by the talent of the players. A good coach will look bad with bad players and a bad coach will look good with good players.
One way to separate a coach from his or her surrounding talent was suggested - facetiously, one suspects - by Bill Simmons. Rather that look strictly at win-loss records, Simmons suggested that one look at the statistics where a coach might have the most impact. These parts included, but were not limited to:
1. Close games. A good coach should do well in close games. Part of this depends on the "trust factor". There was a line in Training Rules where one of Rene Portland's former players claimed that the Lady Lions of Penn State could never progress beyond a certain point in the NCAA Tournament because whenever a crucial game was on the line, they feared Portland but did not trust her. Compare Portland to Pat Summitt of Tennessee who is praised by her players for having command of the situation when the team needs her guidance.
2. Turnovers. The team of a good coach should not throw away the ball needlessly. Plays should not be so hard to execute that they fall apart. The passing game should be competent.
3. Offensive rebounding allowed. A team might not be a good offensive rebounding team, but it should at least win the battle for defensive rebounds at the other end of the court. Well-coached teams don't permit second chance shots.
4. Momentum. This is very hard to qualify. In essence, a good coach builds upon what happened positively in the previous game and the lessons are not forgotten by the team in the upcoming games.
5. Opponent 3-point shooting: A well-coached team should competently defend the perimeter.
6. Road record: A well-coached team should be able to win under adverse conditions. Playing on the road is a test of adversity.
7. Rotation consistency: A well-coached team should not be trying out weird rotations in mid-season. A coach should know who his or her go-to players are and stick with them. There should be no mindless jumbling of the roster.
(* * *)
The next step was to build a metric encompassing all of these factors. It was decided to apply an equal weight to all aspects of the metric. Granted, teams that had good players would do better in some aspects of the metric than others, but this is a characteristic of how difficult it is to separate coaching talent from player talent.
Each coach would be giving a rating from 1 (league leader) to 12 (worst in the league). In case of ties, the sum of the ratings among tied coaches would be split. Since having a low rating is a good thing, the best coach would have the lowest cumulative rating among all seven categories.
For close games the metric used was team performance in games decided by five points or less - games where two baskets by the losing team could have tied or won the game in the final seconds. Those statistics could be found from individual team results.
Turnovers, opponent 3-point shooting, road record, and offensive rebounds allowed were all straightforward. Those statistics could be found at the WNBA website.
Momentum was defined by a stat I like to call streak wins. Streak wins is the total number of wins by a team only counting win streaks of three games or more. For example, if a team had winning streaks of 5, 1, 3, 2, 1 and 1 for the year it would have eight total streak wins (5 + 3).
Rotation consistency is defined by the Herfindahl index, which has been discussed in a previous post. Herfindahl index measures the distribution of minutes among individual players of the team. Teams that depend on the same players over and over will have low Herfindahl indices.
Here are the final totals for the 2010 regular season:
When creating a new metric, the first question is "do the results pass the smell test?" In other words, if someone presented this list to you as a list of WNBA coaches from best to worst this year, would you believe that the person presenting it knew at least something about basketball? (As opposed to a list generated by random chance.) My answer is "yes" - it appears that someone could present a good argument for this list as being an accurate one.
In which case, the WNBA Coach of the Year (excluding playoffs) should be Julie Plank. All kinds of good arguments could be made for Plank. The Mystics won a tough Eastern Conference with Alana Beard out for the season, and she won the #1 seed with a reduced roster. Many times, she only had nine players available to suit up. Plank was second in close games with a 7-3 record. Her team allowed the fewest offensive rebounds of any team in the league.
It might be a surprise to some readers, but I would have picked Brian Agler to be named the best coach by the metric. He came close, just a half-point away. Seattle was 8-1 in games decided by five points or less. Imagine, say, Steven Key at the head of the Storm (who had a 3-7 record in such games) and the Storm might have hit double-digits in losses.
Marynell Meadors ended up fourth this year, due to the Dream being 10th overall in turnovers and a sub-par Herfindahl index. It seems that Meadorscan get no respect from some quarters of the league, but clearly the Dream's performance had much to do with Meadors being at the helm.
Mike Thibault is under fire this year for the poor performance of the Sun, given that Tina Charles is such a great player and that Connecticut fans expected better. Connecticut only had three streak wins this year, a model of inconsistency.
For those who think that having a male coach is the key to WNBA success, the metric begs to differ. There are five male coaches in the league, and the metric puts four of them - Thibault, Gaines, Key and Richardson - right at the bottom. It might not be Richardson's fault, but Key is one of the usual suspects, finishing 11th out of 12 coaches.
As for Nolan Richardson and his first year with the Shock, it might be one he might wish to forget. His team finished either 11th or 12th in six of the seven categories and breaks the record set by the 2000 Seattle Storm for highest Herfindahl index of any team in WNBA history.
Is the above method the perfect metric to be used to evaluate coaching prowess? Not likely, but it at least provides a springboard for discussion. Most people have penciled in either Anne Donovan or Julie Plank for the honor this year, and they are great choices. The metric merely supports the previous contention and should not be used as a substitute for judgment - but sometimes even crude metrics yield intuitive results.
* My contention is that the MVP award shouldn't depend on what team a player is on but on the individual ability of the player involved. One wouldn't refuse to award a Nobel Prize in physics merely because the Ph. D. in question came from an otherwise weak university physics department, or claim that someone wasn't funny because they were born in a town not known for humorous people. It's Most Valuable Player, not Most Fortuitous Circumstance.