THE NUMBERS GAME
I respect Bill James' hard and enthusiastic work (He Does It by the Numbers, May 25), but still, statistics are used to serve a purpose and often ignore external factors. My father used to say that a statistician is the fellow who will tell you that if the cook has one foot in the oven and the other in the freezer, on the average he is very comfortable.
As a lifelong Phillies fan, I was hurt and surprised by Mr. James' assessment of Philadelphia as a probable fifth-place team last year. The Phillies organization is recognized far and wide as a class act; it has developed and acquired outstanding player personnel. The Phillies didn't win the World Series by accident: no team does.
I'm afraid the Phillies were losers for so many years that the stigma stuck even after three straight divisional championships. This year, even after winning a world championship, they aren't generally picked to win their division despite fielding a stronger club.
I suggest that Mr. James remember that statistics prove too little. The intangibles that make a winner cannot be measured.
GEORGE H. LATHAM
I enjoyed Daniel Okrent's article on Bill James, but I take exception to his observation that James can analyze fielding statistically, resolving the "age-old problem" once and for all. James' assumption is that a ball not handled by a fielder is the fielder's fault. James obviously bases his analysis on the fielding of Larry Bowa. The fact that Bowa handled fewer chances than some of his contemporaries may mean to James that Bowa has "the range of the Birdman of Alcatraz," but it can also be interpreted that Bowa simply had fewer balls hit anywhere near him. Keep in mind that he is playing alongside Mike Schmidt, who keeps many balls from going into the hole.
We shouldn't be so awed by James' calculations that we overlook another age-old problem—that statistics can be made to say whatever we want them to.
Elkins Park, Pa.
I enjoyed the article on Bill James so much that I stayed up most of the night working on some statistics of my own. I took his Runs Created formula and applied it to the career totals of each of the 100 best offensive players I could think of. I divided each result by the number of games played by that person to arrive at a Runs Created Per Game statistic. This number (or fraction) predicts, plus or minus 3%, what a given player would contribute to the number of runs his team would score per game.
What is the best hypothetical offensive team of all time? In arriving at mine, in some cases I had to estimate the "caught stealing" variable in the formula; a variation in my estimate wouldn't change the player selected, but it would slightly alter his predicted runs created per game:
JOHN A. BLEYLE
Upper Arlington, Ohio
Address editorial mail to SPORTS ILLUSTRATED, Time & Life Building, Rockefeller Center, New York, New York, 10020.