by Ken Pomeroy on Monday, August 21, 2006
It’s time for the annual look at luck. In order to keep things fresh, I’m throwing my old buddy Pythagoras to the curb and leaning on a new pal, Karl Friedrich Gauss, to show us the way to the luckiest and unluckiest teams in the game last season.
You see, past versions of this post used the Pythagorean method to estimate what a team’s winning percentage should have been. But now, I’ll be using the correlated Gaussian method, invented by Dean Oliver. The difference between the two is that Pythagorean averages each of a team’s offensive and defensive efficiency to come up with a winning percentage, while the Gaussian uses the distribution of a team’s game efficiencies. Since most teams have a tendency to play to the level of the competition, the latter method is little more accurate.