Guilty as charged...Nonetheless:

The original suggestion to look at McGwire's performance over TRAD leagues was from the Commish in response to a post from someone complaining about his McGwire stats. I just did follow through as was suggested and the results do not look to favor this SIM.
I don't think the little data I looked at was very meaningful as you pointed out. I just pulled out data from the TRAD leagues I could identify that seemed to have a reasonable number of PA's for Mac. It still took a while to do that.

What would be nice would be to look at a large number of different players over several TRAD leagues. Even if the SIM was working perfectly it would be possible to find individual player's whose stats across many leagues looked out of whack. It seems to me that the real question is whether these aberrant performances are the exception or the norm.

I do NOT think 7 TRAD leagues is too few to be meaningful statistically. Even in a single league, after 400-500 PA's there has been plenty of sampling to allow for a meaningful conclusion if you look at the stats of many players and the players are a random selection, not ones cherry picked because their stats looked weird.


I agree with you that looking at standard deviations from 7 observations is inadequate to come to a conclusion - and I didn't. What I suggested was that IF the same trend was found over many players, that would be disturbing.

Unfortunately, the player stats are not presented in a way that would make it easy to put together stats for multiple players across many leagues.


BTW - As everyone seems to realize, it seems obvious that any comparisons need to be limited to TRAD leagues where a player's performance didn't vary throughout the season. Trying to do this with data from RT leagues would be nearly impossible.

I think this discussion is a good one and needs to get serious consideration, not just a brush off of "small sample size". There is plenty of data available.

DaveF