Figured I'd make a separate thread, rather than continue to hijack the Clash of the Titans thread.

Okay, here's how these work. Schell adjusts for 4 things:

1. Mean performance of the era. You all know what this means.

2. Park effects. His work really excels here. Rather than applying one broad park adjustment, as we see so often, he reasons that some parks are better for home runs than for batting average, and vice versa; some are good home run parks for lefties and some are good home run parks for righties; etc. So he has separate park adjustments for each offensive event analyzed, both from the right and left side!

3. Talent pool. He uses the standard deviation of the offensive event in question to adjust for the talent pool of the league at the time.

4. Late career declines. For each offensive event, the average productive career length is determined and becomes the basis for the number of at bats used to determine the fully adjusted stat for each player.

Then, the numbers are transformed into something that is easily understood: their equivalents for the period 1977-1992. In other words, what follows is what each player's career average per 550 NFP's (number of times facing a pitcher) is worth in 1977-1992 currency, assuming equal ballpark effects. Obviously, had he translated the numbers into 1992-2004 stats, we'd see bigger power numbers; into 1900-1910 stats, smaller numbers, power-wise, but bigger numbers in other categories like steals. He chose 1977-1992 because this period had a nice balance of power and speed, and no one single baseball strategy predominating to the exclusion of all others.

Note that the work rests on some assumptions, as all statistics do. He assumes that the quality of play has changed for the average player but not for the top performers - for eaxmple, the assumption is that a player in the 95th percentile in one era is the same as a 95th percentile in another era. He also assumes that universal changes in equipment or rules will affect all players equally (as do we

Only players with 4000 or more at-bats plus walks and plus hit-by-pitches were chosen for analysis. So 1140 players qualified for analysis. And the book has the numbers listed for each of them. Here's a juicy selection:

Hank Aaron

.310 BA; 28 HR; .378 OBP; .546 SLG; 23 SB

Ty Cobb

.336 BA; 22 HR; .406 OBP; .550 SLG; 63 SB

Joe DiMaggio

.309 BA; 31 HR; .369 OBP; .582 SLG; 4 SB

Barry Bonds

.294 BA; 36 HR; .429 OBP; .584 SLG; 33 SB

Mickey Mantle

.302 BA; 32 HR; .416 OBP; .554 SLG; 28 SB

Willie Mays

.308 BA; 29 HR; .391 OBP; .566 SLG; 49 SB

Honus Wagner

.324 BA; 22 HR; .395 OBP; .551 SLG; 45 SB

Babe Ruth

.309 BA; 50 HR; .439 OBP; .673 SLG; 10 SB

Ted Wiliams

.322 BA; 39 HR; .448 OBP; .628 SLG; 3 SB

Tony Gwynn

.338 BA; 6 HR; .392 OBP; .451 SLG; 25 SB

Rogers Hornsby

.327 BA; 31 HR; .430 OBP; .597 SLG; 10 SB

Despite the flaws in the assumptions, this is still light years better than the bogus "relative stats" drivel currently being paraded around as legitimate basis for cross-era comparisons.

Okay, here's how these work. Schell adjusts for 4 things:

1. Mean performance of the era. You all know what this means.

2. Park effects. His work really excels here. Rather than applying one broad park adjustment, as we see so often, he reasons that some parks are better for home runs than for batting average, and vice versa; some are good home run parks for lefties and some are good home run parks for righties; etc. So he has separate park adjustments for each offensive event analyzed, both from the right and left side!

3. Talent pool. He uses the standard deviation of the offensive event in question to adjust for the talent pool of the league at the time.

4. Late career declines. For each offensive event, the average productive career length is determined and becomes the basis for the number of at bats used to determine the fully adjusted stat for each player.

Then, the numbers are transformed into something that is easily understood: their equivalents for the period 1977-1992. In other words, what follows is what each player's career average per 550 NFP's (number of times facing a pitcher) is worth in 1977-1992 currency, assuming equal ballpark effects. Obviously, had he translated the numbers into 1992-2004 stats, we'd see bigger power numbers; into 1900-1910 stats, smaller numbers, power-wise, but bigger numbers in other categories like steals. He chose 1977-1992 because this period had a nice balance of power and speed, and no one single baseball strategy predominating to the exclusion of all others.

Note that the work rests on some assumptions, as all statistics do. He assumes that the quality of play has changed for the average player but not for the top performers - for eaxmple, the assumption is that a player in the 95th percentile in one era is the same as a 95th percentile in another era. He also assumes that universal changes in equipment or rules will affect all players equally (as do we

*all*, when using mean-adjusted stats) when intuition says that this isn't true.Only players with 4000 or more at-bats plus walks and plus hit-by-pitches were chosen for analysis. So 1140 players qualified for analysis. And the book has the numbers listed for each of them. Here's a juicy selection:

Hank Aaron

.310 BA; 28 HR; .378 OBP; .546 SLG; 23 SB

Ty Cobb

.336 BA; 22 HR; .406 OBP; .550 SLG; 63 SB

Joe DiMaggio

.309 BA; 31 HR; .369 OBP; .582 SLG; 4 SB

Barry Bonds

.294 BA; 36 HR; .429 OBP; .584 SLG; 33 SB

Mickey Mantle

.302 BA; 32 HR; .416 OBP; .554 SLG; 28 SB

Willie Mays

.308 BA; 29 HR; .391 OBP; .566 SLG; 49 SB

Honus Wagner

.324 BA; 22 HR; .395 OBP; .551 SLG; 45 SB

Babe Ruth

.309 BA; 50 HR; .439 OBP; .673 SLG; 10 SB

(Ruth leads all in fully adjusted HR and SLG.)(Ruth leads all in fully adjusted HR and SLG.)

Ted Wiliams

.322 BA; 39 HR; .448 OBP; .628 SLG; 3 SB

(Williams leads all in fully adjusted OBP.)(Williams leads all in fully adjusted OBP.)

Tony Gwynn

.338 BA; 6 HR; .392 OBP; .451 SLG; 25 SB

*(Gwynn leads all in fully adjusted BA.)*Rogers Hornsby

.327 BA; 31 HR; .430 OBP; .597 SLG; 10 SB

Despite the flaws in the assumptions, this is still light years better than the bogus "relative stats" drivel currently being paraded around as legitimate basis for cross-era comparisons.

## Comment