GotowarMissAgnes wrote:The best measure used is neither winners or play-off teams, but what's called the Noll-Scully measure (basically, it looks at how every team performs over the course of the season).
The book shows a number of things, including the fact that over the last 20 years football has had the most competitive balance. But, baseball is not far behind. Baseball's competitive balance has also been improving, not worsening. Furthermore, there's almost no relationship between the success of a league (in terms of revenue, attendance, etc.) and competitive balance.
GTWAR, I've just perused a couple snippets about this measure including here (http://www.wagesofwins.com/WOWCh5.htm), but can you tell me if I'm understanding it correctly?
It's the ratio of the std. dev. of the actual league versus what one would expect under a league of all equally talented teams?
Does it take account of the fact that low-scoring games such as baseball are subject to more variance than a high-scoring game such as basketball? In other words, a dominant baseball team dreams of a .600 winning percentage whereas the Bulls in the '90's won somewhere around .900 if I remember right. And I've always attributed that to a higher sample size, that if you have a better team in basketball, you have about 100 scoring chances, all with equal weight, of proving it. (And 100 chances to defend against an inferior team). Whereas in baseball, you might only have ten or so in any given game. Also, one missed pitch that is sent down the center of the plate can provide 4 runs which is a lead that is not insurmountable, but very significant. Whereas in basketball the most you can do on one random event is a 4-pt play where you hit a 3, get fouled, and then sink the free throw. Which is something you can make up quite easily.
I think this would be all folded into the standard deviation of scoring... that a team is expected to score say 5.4 runs with a std dev of 2.2 (to pull numbers out of my head). Which would make baseball games innately more unpredictable. Just want to see if I'm getting what they're talking about. Or is the std dev of expected wins? In which case you would still suffer from this league scoring bias.