Quote:
Originally posted by Malleus Dei
Let me do this without flaming.
Holyroller, you are making an important error.
Bill James invents statistsics.
Bill james does analysis.
Bill James writes opinions.
When Bill James invents a statistic, like, say, Range Factor, it is unarguable because it is simply data (in this case, defensive plays divided by games played).
When Bill James does a statistical analysis, it is unarguable as long as (1) the data is correct; (2) the model is valid. Both conditions must be met. You are welcome to argue that his model is wrong or that his data is wrong, but if his data and his model are both correct and there are no math errors then you cannot argue with the results unless they are inapplicable to the question at hand, a statistical crime that I have yet to see James commit. If James says that Julio Franco created more runs per 27 outs (RC/27) during a season than Erubiel Durazo did, as long as his data is correct and the RC/27 model is correct and he didn't make any calculation errors, then you cannot argue with that no matter how much you don't want to believe it.
But Bill James also writes informed opinions, which can be argued with as long as you are also informed about the subject as well. Starting in the 1977 Abstract, James has written an amazing number of opinions. Some of them are brilliant. His condemnations of some players and managers are classic (he once wrote that the only way one player could field his position any worse would be if he used a machete) and many of his essays which stray into editorials/opinion pieces are very edifying and amusing. But some of his opinions - on Pete Rose not being guilty, for instance - are so much hogwash. (James mainatined that the betting slips that damned Rose were bogus because they had an Atlanta-Philadelphia game listed on them on a day that the Braves and Phillies did not play, which was true, but anyone who followed basketball, as he does not, would know that there was an Atlanta-Philadelphia basketball ganme on that date.) You can argue with Bill James informed opinion, as I have, but to do so requires that you also be informed on the subject; otherwise, you are just spouting "I-like-milkshakes" nonsense.
Do you see the difference? The point of all this is that there some things Bill James writes that you can argue with - his opinions, like whether baseball is better or worse today or the innocence of Pete Rose - and there are some things that you cannot argue with - his statistics or analysis - unless you can show that his data or models are bad or inapplicable or an error was made. James's statistics and models have taught me things about the game over the years that I would never have believed and did not want to accept, and that process has not been a painless one, as my previous and cherished beliefs often failed to survive contact with statistical reality. But as to his opinions...well, sometimes I know as much or more about one of the subjects of one of his opinion pieces than he does, though admittedly not often, and sometimes I can point out where he is wrong, as with the Rose betting slips...and you might know enough to do the same.
|
I agree 100%. Except you neglect to mention one very important fact. While the model of analysis can be correct, and the statistical evidence can be correct, the conclusions based upon the evidence are influenced by subjective factors.
Take his historical rankings of players in the new abstract. The player's win shares for each season, per 162 games, career total, and etc. are all there and cannot be disputed. His rankings, on the other hand, can be disputed because, like anything, they are partly based on subjective judgements. So, while the list of the top 100 catchers based on total win shares cannot be debated as long as the data is correct, the list of who the top 100 catchers of all time are CAN be debated.
I would think that his list of the top managers in the game would tend to resemble his list of the 100 best catchers rather than the top 100 career win shares by catchers. The data he bases his conclusions on cannot be debated, but his conclusions? Absolutely they can be debated. I am also 100 percent certain that Bill James would agree with me, rather than you, on this point. Even if your conclusion is sound and informed, based on correct data, it is still an opinion and can be debated. To complete the analogy, the list of the top managers based solely on wins above expected level cannot be debated, but the list of the best managers can be because it is based on several different factors, and each person will put different weight on each factor.
RC/27 are simply a fact. Win Shares are simply a fact. Linear Weights (by Pete Palmer) are simply a fact. Team wins above expected are simply a fact. However, the conclusions based on these facts are opinions, not fact. Any list of the top players of all time are an opinion. By the same token, any list of the top managers in the game are also an opinion.