by Chris Marshall:
I planned to write this article about a month ago; the Royal Rumble1 was just about to take place, so it actually made sense to do it back then. Of course, if you want to make God laugh, tell Him your plans. I kept pushing it off because it seemed like a lot of work for such little reward. After doing the data analysis, however, I think I might have found something interesting.
The original idea was to treat the Rumble as if it were an actual athletic competition and perform a statistical analysis2 in much the same way as I would for a legitimate sport. For example, how much does starting position affect a wrestler's finishing position? Does the number of opponents eliminated by a wrestler have any bearing on how far they make it?
The first draft of this article addresses these issues, and it will eventually become another part of my Royal Rumble Sabermetrics series, but I decided that they could wait. At first I had no demographic variables, so I added two: race and age3. In order to make it possible to run a t-test, I dichotomized the racial variable into white and non-white. I ended up creating a second racial variable, black and non-black, for reasons I will explain later. Anyway, the results were a little troubling.
