By Ben Golliver
Whatever you do, do not tell Hall of Fame guard Reggie Miller that he averaged an 18.4 Player Efficiency Rating or that he accumulated 174.4 win shares during his 18-year career with the Indiana Pacers.
That much became clear this week, when Miller, now a broadcaster for TNT, critiqued the use of advanced statistics in basketball analysis. During a Thursday night broadcast of a preseason game between the Los Angeles Clippers and the Denver Nuggets, Miller unloaded on the numbers-friendly Basketball Prospectus for projecting the Denver Nuggets to win the Western Conference this season.
"One of those, lab geek rats, that are somewhere with their pocket squares, with their little laptop computers, punching numbers. You look at this Denver Nuggets team, on paper, because of their youth, No. 1 in points scored last season, No. 1 in assists, No. 1 in points in the paint, it's very impressive... Those things don't play basketball games for you. Humans play basketball for you. Those geeks. I'm telling you, never played a game in their life."
Presumably, Miller meant "pocket protectors," a geek fashion staple, at least in stereotype, and not "pocket squares," a high-fashion item he himself has been known to sport on occasion.
Miller makes a common error here, confusing statistical projections with predictions. Projections, generally speaking, are the result of impartial mathematical models run consistently year after year. Predictions, generally, are more subjective conclusions formed with, or without, the aid of projections.
Basketball Prospectus' Kevin Pelton reinforced that point in Twitter remarks made in response to Miller on Thursday.
Saying we "picked" Denver to win the West is not really accurate. Our picks will come out next week. I am not picking Denver. ONE version of our projection system had Denver on top of the West, and another had them second. Kevin Harlan had the correct interpretation, which is that we see Denver as ONE of the teams that could win the West.There are many possible responses to being informed that computer models like the Nuggets. "Nerds don't know anything about basketball," is about the least useful. Here, the proper conclusion was an easy one: The computer liked the Nuggets more than virtually every human analyst, so there's a possibility they are being meaningfully underrated by the non-robot intelligentsia. That could prove to be a useful tidbit, or not, but pursuing reliable and consistent data to inform opinions, generally speaking, is better than sitting content in blissful ignorance.