Sabermetric-toothed tiger: The dangers of advanced sports statistics


The release of “Moneyball” marks an interesting trend in sports. Increasingly pundits are releasing new, advanced statistics with fun to say acronyms. PER, QBR, EFGP, VORP. It sounds like “Star-Trek,” all power to the vorp.

However these stats outside of baseball are misguided at best. The reason why is that they are yardsticks, attempts to fully look at a player’s value, and those are flawed. Instead, we should build microscopes.

The heartwarming story of “Moneyball” proved that advanced statistics worked in baseball. Following this, other sports’ advanced statistics became more prevalent. Basketball mad scientist John Hollinger released his culminating stat PER, an attempt to make a basketball player’s influence on the court into one single statistic with football following suit. But all of these stats are flawed, and this is because baseball isn’t a team sport.

Baseball is entirely about individual accomplishment. Everything a player does is dependant on his own ability. Basketball is not like this. Neither is football. In both of these sports, teammates are a significant contributor to the outcome. Can you create a defensive stat for basketball that somehow ignores all the other people on the court? I wouldn’t bet on it. Football is more extreme. We should practically tally rush yards for the o-line, not the running back.

So what should we be doing? The answer actually lay in the world of basketball general managing. I call the big formula “stats yardsticks” as they try to measure a player completely against his peers. But yardsticks are always imprecise. Basketball general managers understand this, and have instead gone towards microscopes, extremely detailed narrow stats with things like corner three point percentage in the last four minutes of a game for example. These stats paint a very clear picture, just a little bit at a time

No matter how advanced, a flashy algorithm can never accurately capture the impact of a player on a game in the true team sports (hockey, basketball, and football primarily). We instead need detailed specific-situational stats, the microscopes, which paint give a very detailed, very narrow picture, like corner three pointers made for example. Hollinger, you are a madman, but you make me think.