11th April 2011
Here's an idea sent my way courtesy of BBR reader Rob P.:
"Can you think of players who had excellent 'per-36-minute' stat lines on limited
minutes, and who either outperformed or seriously underperformed those 'per-36'
numbers once given an increase in minutes?
I'm a Celtics fan, so Glen Davis comes to mind as being a good example of
someone who produced close to their per-36 averages upon being given a larger
I'm curious about some of the extremes; players whose averages were seriously
impacted by an increase in minutes. Basically examples that make you think, 'it
was a bad idea to give this guy more minutes' OR 'I can't believe he's been
coming off the bench all this time instead of starting!'"
One of the big early battlegrounds of APBRmetrics was the philosophical debate between per-minute and per-game statistics. Per-game was the traditional standard, but analysts like John Hollinger began to tear that way of thinking down after realizing per-minute performance held over for most players who received more playing time. From Hollinger's seminal 2004-05 Pro Basketball Forecast:
"It's a pretty simple concept, but one that has largely escaped most NBA front offices: The idea that what a player does on a per-minute basis is far more important than his per-game stats. The latter tend to be influenced more by playing time than by the quality of play, yet remain the most common metric of player performance.
Unfortunately, many NBA execs and fans still believe that somebody can be a '20 minute player' -- that he's only useful in short stretches but can't play a full game. With the exception of the rare few who are scandalously out of shape (Oliver Miller, for example), this is profoundly untrue. [Michael] Redd was the perfect example -- he was thought of as a bench player simple because that's what he'd always been, but there was no reason he couldn't play 40 minutes a night. There's a supposition that some players' production will decrease with increased minutes, but within reason that's completely untrue. The first Prospectus emphatically proved this with research showing that most player's [sic] performance improves with greater playing time."
Hollinger's examples of predictable "breakouts" from per-minute stats included Redd, Zach Randolph, Carlos Boozer, and Andrei Kirilenko, all of whom held onto their low-MPG production when thrust into bigger roles. In fact, Hollinger featured Redd on the cover of his 2nd book as an example of a player with great per-minute stats who was underrated because of a lack of playing time.
So, to answer Rob's original question, and in honor of Hollinger's early per-minute darlings, here are the "Redd-Randolph All-Stars". To qualify, a player had to:
- play in the "Hollinger Era" (the 1990s, 2000s, or 2010s)
- play at least 41 games in back-to-back seasons
- play less than 24 MPG in the first of the back-to-back seasons, and more than 24 MPG in the second
- see an increase of at least 7 MPG between the two seasons
Of that group (which included 320 players since 1990), I'll list 3 top-5 lists: players who improved their PERs the most when given increased playing time, players whose PERs were the closest to what they had been before when given increased playing time, and players whose PERs declined the most with an increase in PT. This will capture all of the possible extremes Rob mentioned, plus the Hollinger prototype of players whose PERs didn't change at all.