What are “Wins Above Bench”?
The “Wins Above Bench” statistic is my attempt to translate the baseball statistic “Wins Above Replacement” to basketball. Let me explain why the stat is valuable, then I will explain what I mean by the first sentence.
“Replacement level” comparisons basically compare a player’s production against the production the team could expect at the same position from easily acquired talent. Stat people like this comparison because its more realistic than comparing a player against “the average”. In sports, “average” tends to be very good, and hard to obtain.
So, in other words, what the stat is asking is, “what are you providing us with that we couldn’t have gotten for basically nothing last summer?”
The statistic also captures the value a player adds to the team by playing minutes that might otherwise go to a crappy player. Bucks fans who watched Dan Gadzuric get dusted off and added to the rotation following Bogut’s injury can relate to that value.
What is a “Easily Acquired” talent?
In baseball, “easily acquired” talent is easy to define. Its the average guy sitting on the bench or the average guy in Triple A waiting to be called up.
But in basketball, there is a different rotational reality. Starters do not generally play entire games as they do in baseball. Basketball teams playing minutes are basically distributed as follows: A. Starters, who play about 67% of the time; B. Second Stringers, who play about 30% of the time; and then C. Deep Bench Players, whom the coach never wants to play, but who are on the team for practice reasons and for “emergency” situations. Those players tend to be either big men, point guards, or young players not ready for prime time. They play about 3% of the time.
So in basketball there are two “replacements” that need to be considered. The first is for the player himself, the second is for the Second Stringer who was the player’s main substitute. (For example: When Andrew Bogut was injured, it not only made Second String center Kurt Thomas a Starter, it also made Deep Bench Player Dan Gadzuric a Second Stringer). You must account for that secondary reality. That is why you cannot tabulate “replacement level” in basketball by merely tabulating the production of the average Second Stringer, like you can in baseball. In basketball, you must also account for the Second Stringer’s replacement.
Therefore I developed “Wins Above Bench”. Its based on a team of “easily acquired” NBA talent and it is made up of average Second Stringers serving as Starters and playing 67% of the time, and average Deep Bench Players serving as Second Stringers and playing 33% of the time.
Why did I construct it this way? Because I imagined a situation where an owner came to a general manager and said, “We’re broke. I need you to cut every player and then put together the best team you can for minimum wage.”
Under such conditions, I imagined that the GM could attract average second string talent for little money by offering to promote them to starter. He couldn’t attract the very best second stringers because other teams might want to bid on them, but he should be able to attract an average second stringer.
But I concluded that the “thrifty GM” could not also replace his Second Stringers with the same caliber player. After all, why would an average Second Stringer on a different team be attracted by an offer of a parallel position on the thrifty GMs team for no money? He wouldn’t. So to replace the Second String, the Thrifty GM would have to recruit average Deep Bench Players.
Thus “Wins Above Bench” is the production you could receive at the position a combination of: (67%average Second String) + (33%averageDeepBench).
The Wins Above Bench stat then tells you, basically, what did we get from this guy that we couldn’t have gotten from a team of ordinary joes? Therefore you could almost conclude that a player who provides negative Wins Above Bench is basically stealing wins from the team.
Note also that the stat will tend to add value to centers and diminish the value of perimeter players, because of supply and demand. There aren’t a lot of really tall people in the world, so the production level of backup big men tends to slide faster than the production of perimeter players.
Here are the winning percentages I came up with at each position for the imaginary “Bench Team” using the logic above:
SG – .405%
SF – .379%
PF – .333%
C – .239%
If you add those together, the Bench team could be expected to win in the range of 28 games.
Footnote: Interestingly, the very best expansion teams in NBA history have won about 28 games in their inaugural seasons (the Bucks did). Since expansion teams are genrally made up of “easily acquired” talent, this makes total sense.