Chad’s got a post up about how Baseball Statistics Are Crap. I’ve got a different beef.
(There are, certainly, a lot of dubious statistics in baseball. I just don’t agree that things are as bad as Chad says but maybe it’s just that I’m used to the idiosyncrasies. I do understand the infield fly rule, after all. If that weirdness makes sense, maybe the weird statistics do, too.)
Anyway, my objection is that even with these simplified statistics, the sportscasters and writers read too much into them. They don’t understand what the statistics are saying, and the value of statistics is to be able to compare players. In baseball it’s not so bad — even if the stats are flawed, a player hitting .356 is objectively a better hitter, by this measure than one who is hitting .290. But what does “by this measure” mean? In baseball, you can hit for average or for power — there are different skills and abilities useful to the team, and you want to find the statistic that is appropriate to the skill you are trying to quantify.
In this regard, I think, football is an example where the reporters are a great abuser of statistics. And this goes beyond saying “turnover ratio” when “differential” is meant (one thing that’s gotten better over the years). The main abuse, I think, is saying that accuracy is measured by completion percentage, and this seemingly happens all the time.
Accuracy is your ability to hit a target, and if you want to compare apples-to-apples, the target should be the same one. A stationary target at 10 yards is easier to hit than a moving one at 40 yards. A better receiver, who can get open, is easier to hit, and also affects the ability for other receivers to get open. You can have a receiver who drops the ball even though it’s “right at the numbers,” or one who catches everything thrown his way. When nobody’s open and he’s trapped, a quarterback can take a sack or throw the ball away, giving him an incompletion. All of that affects completion percentage, and none of it reflects accuracy.
Chad Pennington is touted as an “accurate quarterback” by many sports journalists, who, in the next breath, mention he has a weak arm and dumps the ball off quite often. Short passes. Connection? I think so!
My favorite example is Donovan McNabb. When Terrell Owens was about to join the team, analysts were all cautious about how Owens would tolerate the inaccurate McNabb, who had never completed 59% of his passes. Until that year, when he completed 64%, and everyone was saying how accurate he had become. Owens leaves, and the completions percentage drops back down. (It’s up again this season, and last — he’s got better receivers, and he dumps the ball off to Westbrook when he has to)