It’s the NBA Playoffs, the perfect time to peruse sports columns, blogs, and general comments from fans and haters. I had forgotten how much of the game – specifically, the hype and rumor mill that the media drives to keep the chatter alive – is about statistic overanalysis and unfounded conclusions. Well, that, and the third-grade trash talk.
At some point in modern sporting history, though, someone came up with the ingenious idea of trying to quantify sports with stats. Now, I think it started as a good idea; in basketball, other than the points scored, people first started caring about rebounds, assists, steals, and other metrics that correlate with a team’s success on the court.
It makes for a pointed study in using stats to say whatever the hell you’d want.
Even with a relatively few measures it’s already not easy to tell the impact of one or two stat categories, and how the numbers interplay (e.g., more rebounds can take away steal opportunities) are totally unknown. Granted, it wasn’t meant to be a numeric replacement for the action, especially when there are so many more things happening in the game that aren’t accounted for anyway.
Then somebody built an entire set of stats built around arguably the only objective measure of a team’s performance, in points scored and games won. Again, the reader’s led to believe that some series of numbers – a win streak, maybe the average winning differential in points, or historic losses in a season – are an indication of something and that it’s totally worthwhile to sift through and divine some correlation that justified their existence. Trying to draw a comparison between a ’95 Bulls team (holding the NBA record for wins at 72) and, say, the ’69 Lakers (69 wins) is futile when so many other variables need to be factored anyway.
Nowadays, what was already borderline arbitrary is now pretty much completely removed from the ballgame. Sports commentators go even deeper by massaging official records with their own formulas to come up with stuff like efficiency ratings and true shooting percentages, and with a clever database query or two, they can trot out some impressive sounding, convoluted stat lines like “the first player under 23 years old who went 15 points, 5 rebounds, and 3 steals in 30 minutes or less in his first 10 games with a new team”.
Since the entire point of sports is determining who’s best, it’s not much of a surprise that all the numeric masturbatory tools are used for comparisons between players, coaches, teams, and sometimes even eras of the sport. The sad part is, for all the objectiveness of hard numbers, nobody has come up with a reliable way to make use of them for predictions (1) or even as a comparative measure. Sports arguments go along the lines of “Jordan’s better than Kobe, he won more championships!” versus “Kobe’s got a higher shooting percentage at the same stage in his career!” ad nauseum. I also like justifications that try to sum up a game as a single, magical, overarching number, stuff like “they lost this game because Lebron shot 2-14, he won’t shoot this bad again, it’s humanly impossible“. Of course, sports fans also happen to have severe myopia and can’t see their way past the last game anyway.
The cynic in me says the sports-o-sphere can’t be this dumb, and that a real statistician probably drop by at one point and told them the system is terribly inaccurate and broken, but the sports media needs to keep the conversation going so they’ll keep peddling the same misconceptions. I’m just happy watchin’ the games.
- then again, that’d completely break the sports betting industry (↩)
Try saying that to the day traders that hold their formulas in their economic bibles.
Oh hello anamolys
Go Bulls!