Dan Uggla and Creationism

Lest you think (as I did) that this is the first page on the web with both "Dan Uggla" and "creationism," you will be surprised to know that there are actually 1,420,000 at this very moment.

Anyway, congratulations to Dan Uggla on his 33-game hitting streak, if only because I have him on my fantasy team and he was fucking killing me up to that point. I am ashamed to say that I only kept him in my lineup because I had no viable alternatives.

Unfortunately, whenever someone stitches together a long hitting streak, writers invariably scramble to throw up stories that convey the rarity of such a streak by multiplying the odds of a one-game outcome by however many games the streak lasted. This year's offender is Joe Posnanski, who crunches some numbers and comes up with odds of 3 million to one for a Dan Uggla 33-game streak.

That sounds like an incredible number. And indeed, Dan Uggla is one of the less likely candidates due to his lack of speed and poor contact ability. But looking back in hindsight and calculating the odds to come to the conclusion that we've witnessed some sort of millenial event seriously conflates a priori and post hoc probabilities. The a priori probability may indeed be 3 million to one, but the post hoc (observed) probability is one. To his credit, Posnanski does word it correctly ("The odds of Dan Uggla on July 5, 2011, beginning a 33-game hitting streak were more than three million to one against"), but then incorrectly uses that figure to justify his awe that it occurred. In short, yes, it's cool that it was Dan Uggla and nobody could have predicted it beforehand, but honestly, it's not that big a deal that it was achieved. Indeed, 55 players have had a 30+ game streak in baseball history - Andre Ethier did it just this year!

It's like winning the lottery - the odds of a given player winning are astronomically low. But SOMEONE usually wins. We correctly are not awestruck that Joe Schmo wins the Powerball jackpot, even though his chances were one in 100 million. It's easy to see whey we can't take those a priori odds and then use them to say "Wow, I can't believe someone won that lottery, it was so unlikely."

That seems like an academic distinction (and in many ways it is), but it has serious implications for a frequent argument I see bandied about in favor of creationism. It goes something like this: "The odds of all the wonderful things in the world arising through random mutation are so low, it couldn't have happened, a creator or designer was necessary."

Um, no. Another conflation of a priori and post hoc probabilities.