06/21/06
The thing with sports is that it draws out the stats geniuses. There are so many professional statisticians crunching sports numbers for teams, leagues, commentators, etc. These guys know their stats for the most part. However, all this talk of statistics makes Average Joe only remember the interesting sounding stuff.
So some people like to present interesting stats. Of course, this phenomenon also makes people really want to challenge conventional wisdom about sports statistics. See this Salon article written by King Kaufman which has a part here (you will have to view an ad to view that article for free) that appears on first glance to rip apart ESPN for touting the stat that the team that scored the first goals is 21-2-3. It starts of pretty reasonable, but throws this out near the end:
If a goal is scored and ESPN flashes a graphic saying, "Teams that have scored first are 22-3-3," I, the typical American sports fan who doesn't care about soccer, will think, "Well, there's about a four-in-five chance that this baby's over. I believe I will turn off the TV, kick my dog, curse some foreigners and play with my assault rifle."
But if that graphic said, "Teams that have scored second are 17-2-3," I'm going to want to stick around to see which team can come up with that all-important tally. Better for me, better for ESPN and way better for the dog.
Sounds pretty damning, especially if you read the whole thing. But wait. Let's look a little bit deeper. Let's crunch some numbers real quick. The team that scored first went 22-3-3 as of that article (games played through June 19, 2006) and teams that scored the second goal went 17-2-3. Some more stats that Salon isn't telling you about that set of 34 games:
- 6 games went 1-0, where the team that scores goal 1 wins.
- Of the 22 remaining games where more than one goal was scored, the team that scored the first goal also scored the second 14 times. These 14 games were all won by the team to score the first (and second).
- Of the remaining 8 games where the team that scored the first goal DID NOT score the second, team that scored the second went 3-2-3. Not that bad, but significantly worse than the 17-2-3 the article states in a misleading fashion.
So if you're going to "stick around to see which team can come up with that all-important [second goal]," let me kill some suspense for you - the team that did NOT score first will score the second goal only 21% of the time, and even then will only win slightly over a third of those games. In fact, the more you crunch the numbers the more you realize that the original stat of where the first team wins over 82% of the time is not at all a misleading indicator of what will happen next.
Kaufman's older article investigating the exact same phenomenon in hockey is actually ok. It correctly argues that the first goal is not significantly more important than any other goal. It tries to demystify the importance of the first goal in proportion to any other goals. However, this more recent article argues that the first goal is not that important and implies that the first goal is not that good of an indicator of who will win a game, which is absolutely wrong.
I don't actually have a problem with most of the article. But when you put these paragraphs as your conclusion (the only things after the paragraphs I quoted were a brief preview of coming games and a zippy quote), I start to wonder if you understand your own argument. In fact, upon reading the previous article, I realize that he had relied almost exclusively on the arguments of another person more knowledgable about statistics.
