Posted by Trentonjoe on 3/11/2011 3:45:00 AM (view original):
Is there enough variation to be significant? I don't know the math well enough. I also don't know if I used the right vocabulary.
The correlation coefficient was .31 which by itself is not significant at least not with the limited data. (Statistically one looks for an R^2 value of .9; practically this is more like 0.7.) However, without being able to adjust for other factors: drafting, managing, pitcher to defense matching, this is pretty much what I expected.
Keep in mind no one has ever claimed that LuE is the only way to win in the sim. Many of us have said it was important and thus far the evidence back it up. If you and Boog are correct and honest about your team's 110% LuE, and the average is around 100%, that extra 10% is quite significant: on average 81 runs per season or ~$4million player. However, this pales in comparsion to the surplus value of a smart draft. '08 Addie Joss is priced ~$2M too low; '08 Brown is ~$1M too low and Gary Carter, everbody's favorite cookie catcher offers over $1M in surplus value. So just those three cookies offer more surplus as the best owners can create by creating remarkably efficient line-ups. I haven't finalized the math yet, but my data so far suggests that matching pitching to defense is worth $2M to $4M.