Davenport Translations

Baseball Prospectus 1996

One of the most important features in this book is a system that has become known as the Davenport Translations, or DTs. I'm Clay Davenport, and the DTs are a result of baseball-related studies that I've been doing for about ten years now. In the reports, you'll see a line of statistics for every player, but be warned: they are NOT his real stats. Instead, they are a detailed analysis of those statistics, where I've tried to isolate actual player performance from everything else in the statistics. The key idea in them is that performance is not the same thing as results.

Think about it for a minute: what goes into a player's statistics, the results of his play? His performance, obviously. The better you play, the better your stats are, the more likely your team is to score (if you're a hitter) or keep the other team from scoring (if a pitcher). This is what we want to know.

But performance isn't the only thing in the stats. Who you're playing against makes a big difference. If Frank Thomas were to play against AAA pitchers, his stats would be better than in the majors, other things being equal. They'd be even better against AA pitchers, better still against A ball pitchers, and yet better still against high school pitchers. The quality of your opponents determines whether the exact same performance, in terms of tracking a thrown ball and bat speed, results in a batting average of .700, .400 or .100.

Where you play makes a difference as well. Different parks determine whether a fly ball is caught or goes over the fence; whether a ground ball gets through on turf or is held up by tall grass; whether a popup is caught by the third baseman or his biggest fan in the seats; whether the wind tends to blow in or out; whether the ball is easy or tough to see. All of these factors combine to create 'hitter's parks' and 'pitcher's parks': parks which tend, repeatedly, to favor or hinder offense. This is as true in the minor leagues as it is in the majors.

And when the majority of the parks in a league are either friendly to hitting, like they are in the Pacific Coast League, or to pitching, like in the Florida State League, you get hitter's leagues and pitcher's leagues. This increase in batter performance in the PCL has nothing to do with either better hitting or lousier pitching in that league compared to, say, the International League. Measuring players in those leagues only by their traditional statistics is likecomparing the heights of two players by measuring from sea level instead of the local ground. Sure, hitters in the PCL put up great numbers: but they are getting, in essence, a 20 to 30 point bonus in batting average because of where they are playing.

Unfortunately, all of these things change every year. League-wide offense goes up and down (the last three years have seen considerably more offense in the majors than the previous ten years). Park factors change somewhat over time, as the weather and other stadiums in the league change. League skill levels evolve as well: Dick Cramer published a study some 15 years ago which strongly suggested a gradual increase in average skill over baseball history (a trend which matches our experience in all other athletic events), a trend which I believe has continued to the present, although expansion has tempered it somewhat.

All of this means that it is very difficult to tell whether a player really improves from one year to the next: just because the stats get better doesn't mean the player really performed better. You have to take into account the context of each year's performance: the competition, the park, the league. Most people don't do that: they have one set of standards that they recognize as 'good', standards that were probably set by major league performances during the first ten years of their fanhood. And when they apply that standard to a league where the real standard is very different (like the 1994 PCL, where an average player hit .295) they reach an erroneous conclusion: overrating players from a league where their personal standards are too low, and underrating players when their standards are too high.

That's what the DTs are for. The effects addressed above can be measured: not perfectly, but approximately. And when you know roughly how much impact everything is having on player stats, you can take them out or put them back in at will. The DTs change the statistical lines of every single player - from any year, from any park in any league - so that they are all being held to the exact same standard. The biases have been removed.

Perfectly? Of course not. The size of the biases is based on large scale averages; the biases in a player's own statistics are based on his personal response to the conditions around him. Some players are better at taking advantage of a hitter-friendly park than others, and the translation procedure won't pick that up. But I am very confident that they are a valuable analytical tool. Translated statistics behave very much like major league statistics, for players of the same age, regardless of what league a player is currently in. There are any number of players whose major league performance is exactly what was expected from their minor league performance; many more who showed steady improvement in the minors which continued into the majors; and only a few whose performance radically changes: but radical changes in performance happen at both the minor and major league levels, at roughly the same frequency for a given age.

How is it done? I break down a player's offensive production into its component parts, asking questions like: how important are doubles to this player's total offense? Walks? What is the ratio of doubles to home runs? I make all possible comparisons between singles, doubles, triples, home runs, walks, and basestealing. And then I do the same thing again, for the league in which he actually played and for the league to which I am trying to translate the statistics. I generally choose a single target league and location and translate everybody to that it: for this book, my reference league is a neutral park in the 1995 National League.

With the information above, I also have an estimate of a player's total offensive value, adjusted for the offensive level of his league, his home park, and the level of competition in the league. I measure it through a statistic called Equivalent Average (EqA; more about that later), although you could use others. This 'total offensive value' is conserved during the translation. All that is left is to produce a line of statistics that satisfies two goals: it results in the same EqA in the new league as measured in the old league, and preserves the relative performance of each offensive component for that player. It is not a trivial task, because it is usually impossible to perfectly satisfy both goals. I treat the EqA goal as paramount, and use an iterative routine to minimize the differences in the second goal.

You may be thinking that this sounds a lot like Bill James' Major League Equivalencies, or MLEs. The concept is similar; the execution is completely different. While I happen to like DTs better, I'm hardly unbiased. As I see them, these are the main differences between my work and James':

A Typical DT Line for Players


1991	BAK	451	110	13	1	18	24	0	1	.244	.282	.397 	.235	106	46
1992	SAN	118	44	5	0	7	12	0	0	.373	.431	.593 	.347	41	26
1992	ABQ	344	104	11	2	15	28	1	2	.302	.355	.477 	.287	99	53
1992	LAD	70	18	1	0	2	5	0	0	.257	.307	.357 	.234	16	7
1993	LAD	549	179	23	2	32	47	3	4	.326	.379	.550 	.314	172	103
1994	LAD	408	137	16	1	22	34	1	3	.336	.387	.542 	.315	128	76
1995	LAD	439	160	17	1	29	39	1	0	.364	.416	.606 	.344	151	97
(OK, so Mike Piazza is hardly typical.)

Top line: the player's name, baseball birth year, and primary position(s) in 1995. The 'baseball birth year' relies on the convention that a player's age on July 1 is considered his age for the season. For a player born in the first six months of the year, this will be the same as the calendar year. But for a player born in the last six months, like Mike Piazza (September 8, 1968), it will be listed as the following calendar year. Mike was considered to be 26 in 1995, and 1995 minus 1969 is 25.

YEAR: is obvious.

TEAM: A list of all the teams and their abbreviations is provided on page X. All of the stats are translated, not the real thing. Remember that!

AB: Atbats

H: Hits

DB: Doubles

TP: Triples

HR: Home Runs

BB: Walks

SB: Stolen bases

CS: Caught stealing

BA: Batting average; league average, .263.

OBA: On-base average. League, .328.

SA: Slugging average. League, .408.

EQA: Equivalent Average. See below. League average hitter: .260.

EQH: Equivalent Hits. See below.

EQR: Equivalent Runs. See below.

For space and significance purposes, seasons with less than 30 PAs have been omitted.

A Typical DT Line for Pitchers


1991	CHC	261.3	245	107	24	69	213	3.68	16	13	8.44	2.38 	7.34
1992	CHC	266.7	227	76	11	77	226	2.57	21	9	7.66	2.60 	7.63
1993	ATL	263.0	254	92	17	60	213	3.15	18	11	8.69	2.05 	7.29
1994	ATL	201.0	155	41	4	33	161	1.84	18	4	6.94	1.48 	7.21
1995	ATL	208.0	155	37	8	23	179	1.60	20	3	6.71	1.00 	7.75
(I know; he's not typical either.)

Pitcher's lines are in some ways similar to the hitter's lines; the first line shows the pitcher's name, 'baseball birth year', and position. The 'position' for pitchers is a three-letter code. The first letter is L or R, depending on whether the pitcher is left-or right-handed; Greg Harris could be listed as S, for switch, if he changed hands more often. The second letter is S for a pitcher who started in 80% of his games, R for a pitcher who relieved in at least 80% of his games, and B for a pitcher who started and relieved, but did neither 80% of the time. The third letter is simply P, for pitcher. RSP indicates that Maddux is a right-handed starting pitcher.

Once again, all of the listed statistics have been translated; they are not the genuine numbers.

YEAR and TEAM are once again obvious.

IP: Innings Pitched.

H: Hits.

ER: Earned runs.

HR: Home runs.

BB: Walks.

SO: Strikeouts.

ERA: Earned run average. Set up so that an ERA of 4.00 is perfectly average.

W: Wins, calculated from ERA, innings pitched, and average offensive support.

L: Losses, calculated with wins.

H/9: Hits per nine innings. League average is 9.06.

BB/9: Walks per nine innings. League average, 3.32.

K/9: Strikeouts per nine innings. League average, 6.63.

Seasons with less than 10 innings have been omitted.

Equivalent Averages: What Are They?

Equivalent average and equivalent runs are one of my independent creations. Equivalent runs, like runs created, are a measure of how many runs a player provided to his team. There are several reasons why we are using them here:

1) I created them.

2) Equivalent runs are slightly more accurate at projecting team run scoring from raw statistics than any other statistic I have tested, including Bill James' Runs Created and Pete Palmer's Linear Weights (although in the last ~30 years the difference between EqA and LW is virtually nil).

Table 1. Root mean square errors for predicted runs from various statistics.

1901-1992		  AL	  NL	Majors	
Equivalent Runs		23.99	23.00	 23.51	
Linear Weights		24.94	23.96	 24.46	
Runs Created		25.62	24.27	 24.97	
Onbase plus slugging	27.52	26.37	 26.96	
Batting Average		47.32	42.97	 45.24	

1960-1992		  AL	  NL	Majors	
Equivalent Runs		21.41	22.08	 21.73	
Linear Weights		21.54	21.82	 21.68	
Runs Created		22.85	22.68	 22.79	
Onbase plus slugging	23.42	23.47	 23.44	
Batting Average		41.77	39.85	 40.87	
3) At extremes of performance--players or teams who are either far above r far below the league average--the runs created and linear weights formulas become less accurate, a result which isn't shown by team comparisons like the above. Equivalent Runs remain on target.

4) Equivalent Average adjusts easily to corrections for park, league offensive level, and league quality, giving a uniform measure of batting skill that can be used across time. It is adjusted so that the league average is always .260. EqA measures rate of performance; EqR measures the total contribution.

5) Perhaps the most important and useful feature of EqA is that the resulting number is easy to understand. Equivalent Average comes very close to matching the historical scale of batting average. A .300 equivalent average is just as common, historically, as a .300 batting average. (see Figure 1). There have been 153 players with a .300 batting average; there have been 148 who had a .300 EQA. Since even a casual baseball fan has a good intuitive feel for the BA scale, anyone can quickly tell whether a player's EqA is good or bad. League-leading figures and alltime records are similarly close to the historical batting average record. This gives EQA a huge advantage over Runs Created or Linear Weights, because you don't have to learn a new scale. Consider a player described as having an RC/27 of 8.75 per game. How good is that? League-leading good? All-time great? What about if I help you out by saying he also rates a +64 linear weights? Still not sure where he stands? Well, I would say he had a .356 Equivalent Average, and you can think about what a .356 batting average means to decide how good it is. (It was Kevin Mitchell's league-leading 1989 season.)

As a further example, compare the leaders in Equivalent Average and Batting Average over the last ten years:

National League

Equivalent Average		Batting Average
1986 Tim Raines		.338	Tim Raines	.334
1987 Jack Clark		.356 	Tony Gwynn	.370
1988 Darryl Strawberry	.337 	Tony Gwynn	.313
1989 Kevin Mitchell 	.356 	Tony Gwynn	.336
1990 Barry Bonds 	.345 	Willie McGee	.335
1991 Barry Bonds 	.342 	Terry Pendleton	.319
1992 Barry Bonds 	.382 	Gary Sheffield	.330
1993 Barry Bonds 	.382 	A. Galarraga	.370
1994 Jeff Bagwell 	.388 	Tony Gwynn	.394
1995 Barry Bonds	.345 	Tony Gwynn	.368

American League

Equivalent Average		Batting Average
1986 Wade Boggs 	.340	Wade Boggs	.357
1987 Wade Boggs 	.358	Wade Boggs	.363
1988 Wade Boggs 	.349	Wade Boggs	.366
1989 Fred McGriff 	.335	Kirby Puckett	.339
1990 Rickey Henderson	.379	George Brett	.329
1991 Frank Thomas 	.356	Julio Franco	.341
1992 Frank Thomas 	.356	Edgar Martinez	.343
1993 John Olerud 	.363	John Olerud	.363
1994 Frank Thomas 	.390	Paul O'Neill	.359
1995 Edgar Martinez	.363	Edgar Martinez	.356

Average NL		.357			.347
Average AL 		.359			.352
Average Majors		.358			.349
Park and League Adjustments

Park adjustments are a notoriously controversial area in baseball statistics. First, there is no one park factor that can be truly accurate for every player. The park may favor left-handed hitters or right-handed, flyball hitters or groundball hitters, contact hitters or sluggers, or any combination therein. I don't even try to assess how an individual is benefitting from his park: it is too difficult to try for every hitter, and the data is usually insufficient anyway. What I do instead is determine how much the park helps the typical player who was there. If you take better than average advantage of your park, the adjustment won't remove that extra. You'll get a better ranking than you deserve, but that's OK: if you really are taking better advantage of your park, your team is getting an extra benefit. In the real world, changing parks alters the number of runs you produce; I estimate that by altering the value of a run in each park. A run in Houston is not the same as a run in Denver, any more than a dollar in the US is the same as a dollar in Canada. In Houston, fewer runs tend to score, which means that each one is more valuable: you don't need as many to win a typical game. As with the league adjustments, I am adjusting the value of the product, not the quantity.

The other argument with park factors is over how long they should be averaged. I feel comfortable using one season as my baseline, but many have argued that this is too short: that two, or even three seasons must be used to get a valid sample. I feel that there are enough fluctuations from one season to the next to legitimately cause park factors to jump around, and that this is primarily due to changes in weather. Even for an indoor stadium, this is true: a park factor is always implicitly compared to the league average, and weather effects at the outdoor parks affect everybody's baseline.

The formulas I prefer to use for a park adjustment are rather complex, because I distinguish between the effect of the park itself and the team that plays its games there. Each park receives a preliminary rating based on the total number of runs per game scored in that park, divided by the runs per game scored in road games. Both the home club and visiting club are included. The Rockies, for instance, scored 485 runs, and allowed 490, in 72 games in Coors Field: 13.54 runs per game. They scored 300 and allowed 293 in 72 away games: 8.24 runs per game. The Coors Field park factor is 13.54/8.24, or 1.643: the park seems to cause 64.3% more runs to score than average.

The adjustment given to the Colorado Rockies, however, is different from that given to Coors Field, because the Rockies don't play every game at home. I use the weighted average of the fields in which they did play, so their team rating is calculated by taking 72 times 1.643, for the games at Coors Field, plus 7 times 1.100, for the games in Atlanta, plus all of the other games they played at every other stadium in the NL, and dividing by total games played, or 144. That tends to be about half the distance between the pure park factor and the team factor; for Colorado, that comes out to 1.300.

For minor leagues, I wasn't always able to get the game-by-game data needed. There, I usually used simply (home games times home park factor) plus (road games times road park factor) divided by (total games). The road park factor is found by

where HPF is the home park factor and N is the number of teams in the league.

For some years prior to 1992, I wasn't able to get some minor league data at all. Park factors for 1992 and earlier have sometimes been estimated by knowing the peak factors in later years. A full list of park factors used appears in the appendix.

League adjustment is more straightforward and less controversial, and involves two corrections: one for the average level of offense being different, and another for having different numbers of runs score from the same set of offensive statistics.

It should be obvious that different leagues, or the same league at different times, has different offensive standards. Changes in the liveliness of the ball, the size of the strike zone, the tactics of hitters and pitchers, the parks in the league, and more cause offense to fluctuate. And changes from one year to the next can very rarely be attributed to changes in the quality of pitchers or hitters. An entire league pitchers does not suddenly get worse one year or better, as apparently happened in 1992-93. Even expansion cannot explain why pitchers who were already in the league saw their ERAs increase by an average of half a run between those years. Pitchers and hitters, taken as a whole, don't get better or worse between seasons; but conditions can change that make them look better or worse.

The second adjustment is really a deficit in statistics. The EQA formula uses just eight commonly available statistics, but there is a fair amount of offense that isn't covered by those eight. Some have statistics themselves: hit by pitch, balks, sacrifices. Some don't: outs made on the basepaths, and probably the most important, reaching base on errors. These ROEs look like outs in the statistics, but act like hits in the game. The more ROEs there are, the more runs will score from the same 1set of apparent statistics. This is a very real effect when comparing across long periods of time (comparing todays major leagues to, say, the 1920s), or across leagues at different levels (between the American and Texas Leagues).

There is also a third adjustment needed: for the DH. A league using a designated hitter has a league offensive level that is higher than one without, other things being equal, simply because they replace very poor hitters with at least average hitters for about 1/18 of their plate appearances. This makes a difference of about 5% in run scoring, which equates to a difference of 2.5% in EQA.

This section looks at the top 100 performances in a variety of categories related to EQA and EQR. A number of active players play prominent roles in these lists.

Career EPEQA

Career EPEQA is the player's rate of run production per out over the course of the player's entire career. Because it is based on a how the player compares to the average player in his league, it favors old-time players, since it makes no accounting of increased player skill or the reduced difference between the standout and average player that has undoubtedly occurred. At the same time, it also has a bias in favor of active players, who have yet to finish their careers and experience the almost inevitable decline in career averages that the last few wretched seasons bring.

A minimum of 4000 plate appearances is required to appear. Otherwise, Frank Thomas would appear on this list in third place.

Player			EqA

Babe Ruth		.3776
Ted Williams		.3716
Lou Gehrig		.3530
Mickey Mantle		.3506
Ty Cobb			.3505
Rogers Hornsby		.3488
Dan Brouthers		.3447
Joe Jackson		.3432
Pete Browning		.3383
Stan Musial		.3372
Tris Speaker		.3365
Jimmie Foxx		.3362
Barry Bonds		.3337*
Billy Hamilton		.3336
Willie Mays		.3325
Joe DiMaggio		.3311
Hank Greenberg		.3307
Hank Aaron		.3305
Honus Wagner		.3305
Charlie Keller		.3304
Dick Allen		.3297
Mel Ott			.3297
Johnny Mize		.3287
Eddie Collins		.3275
Roger Connor		.3275
Frank Robinson		.3271
Ed Delahanty		.3259
Elmer Flick		.3259
John McGraw		.3245
Mike Donlin		.3224
Rickey Henderson	.3223*
Nap Lajoie		.3219
Harry Heilmann		.3205
Gavvy Cravath		.3196
Willie McCovey		.3193
Ralph Kiner		.3192
Frank Chance		.3189
Fred McGriff		.3188*
Harry Stovey		.3187
Willie Stargell		.3179
Eddie Mathews		.3176
King Kelly		.3175
Jesse Burkett		.3168
Wade Boggs		.3166*
Hack Wilson		.3164
Bill Joyce		.3160
Mike Schmidt		.3160
Cap Anson		.3157
Will Clark		.3157*
Babe Herman		.3150
Sam Crawford		.3149
Harmon Killebrew	.3148
Kevin Mitchell		.3144*
Mark McGwire		.3142*
Joe Morgan		.3142
Arky Vaughan		.3140
Duke Snider		.3138
Frank Howard		.3136
Mike Tiernan		.3136
Tip O'Neill		.3131
Jackie Robinson		.3131
Sam Thompson		.3131
Bill Terry		.3130
George Gore		.3125
Pedro Guerrero		.3116
Al Rosen		.3111
Jake Fournier		.3110
Sherry Magee		.3108
Tim Raines		.3106*
Norm Cash		.3105
Tony Gwynn		.3104*
Paul Waner		.3103
Jeff Heath		.3102
John Kruk		.3101*
Al Kaline		.3100
Larry Doby		.3099
Daryl Strawberry	.3091*
Roy Cullenbine		.3088
Reggie Smith		.3088
Jack Clark		.3087
Bob Johnson		.3086
Fred Clarke		.3079
Frank Baker		.3077
Dolph Camilli		.3077
Rod Carew		.3075
George Brett		.3074
Jim O'Rourke		.3074
Wally Berger		.3073
Fred Dunlap		.3073
Roberto Clemente	.3072
Buck Ewing		.3071
Topsy Hartsel		.3071
Reggie Jackson		.3071
Boog Powell		.3070
Denny Lyons		.3069
Ross Youngs		.3069
Joe Kelley		.3067
Mickey Cochrane		.3065
Riggs Stephenson	.3062
Chuck Klein		.3057
Career EPER

EPER are adjusted runs; whereas EPEQA measures a rate of production (i.e., runs per out), EPER measure how much is produced. The 2000-EPER plateau is a signal achievement in baseball history, with only 14 members, although two active players are within striking distance. It rewards both a high rate of production and longevity. It tends to favor recent hitters, who played in a 162-game season, and is strongly biased against 19th-century stars because of the short seasons in which they played.

The de facto standard for induction to the Hall of Fame would seem to be somewhere around 1675 EPER. Every eligible player above that line, regardless of position, is in. Three active players are already over that line, and several more appear likely to make it.

Player			EPER

Hank Aaron		2721
Ty Cobb			2685
Babe Ruth		2470
Willie Mays		2454
Stan Musial		2434
Pete Rose		2292
Tris Speaker		2225
Honus Wagner		2207
Carl Yastrzemski	2187
Frank Robinson		2185
Ted Williams		2134
Eddie Collins		2083
Mickey Mantle		2083
Mel Ott			2054
Lou Gehrig		1990
Dave Winfield		1980*
Eddie Murray		1952*
Reggie Jackson		1932
Al Kaline		1919
Joe Morgan		1915
George Brett		1910
Rogers Hornsby		1906
Nap Lajoie		1873
Sam Crawford		1847
Jimmie Foxx		1822
Eddie Mathews		1779
Rickey Henderson	1766*
Mike Schmidt		1743
Willie McCovey		1730
Robin Yount		1730
Billy Williams		1714
Cap Anson		1709
Roberto Clemente	1700
Lou Brock		1697
Rod Carew		1691
Paul Waner		1691
Harmon Killebrew	1690
Rusty Staub		1668
Tony Perez		1642
Roger Connor		1634
Willie Stargell		1628
Andre Dawson		1610*
Paul Molitor		1609*
Jesse Burkett		1573
Dwight Evans		1565
Ernie Banks		1562
Zach Wheat		1560
Brooks Robinson		1553
Fred Clarke		1551
Dan Brouthers		1540
Max Carey		1515
Dave Parker		1515
Darrell Evans		1513
Harry Heilmann		1508
Tim Raines		1508*
Al Simmons		1503
George Davis		1501
Goose Goslin		1500
Charlie Gehringer	1492
Ed Delahanty		1488
Vada Pinson		1487
Jake Beckley		1484
Al Oliver		1462
Joe DiMaggio		1456
Orlando Cepeda		1453
Wade Boggs		1435*
Willie Keeler		1427
Ron Santo		1424
Sherry Magee		1423
Sam Rice		1422
Dick Allen		1415
Harry Hooper		1414
Frankie Frisch		1407
Duke Snider		1405
Willie Davis		1402
Steve Garvey		1394
Luke Appling		1393
Joe Torre		1390
Jim Rice		1388
Mickey Vernon		1386
Ted Simmons		1382
Lou Whitaker		1382*
Johnny Mize		1373
Carlton Fisk		1370
Jose Cruz		1367
George Sisler		1362
Cal Ripken		1357*
Enos Slaughter		1357
Johnny Bench		1352
Greg Nettles		1351
Reggie Smith		1351
Bill Dahlen		1348
Jack Clark		1346
Jim O'Rourke		1344
George Van Haltren	1343
Ken Singleton		1342
Joe Medwick		1334
Jimmy Ryan		1334
Keith Hernandez		1326
Jimmy Sheckard		1326
Career EQH

EQH are really something of a freak-show stat; they don't quite measure EPEQA well enough to be genuinely useful, but they serve adequately as an estimator. I've included it here to highlight the equivalent 3000-hit club, and note with some despair that Pete Rose made it to the top of this hit chart, too.

Player			EqA

Pete Rose		4090
Hank Aaron		4078
Ty Cobb			3996
Stan Musial		3694
Willie Mays		3612
Carl Yastrzemski	3611
Honus Wagner		3433
Tris Speaker		3423
Dave Winfield		3300*
Frank Robinson		3268
Eddie Collins		3246
Eddie Murray		3231*
Babe Ruth		3165
George Brett		3160
Al Kaline		3124
Robin Yount		3124
Mel Ott			3096
Nap Lajoie		3077
Reggie Jackson		3014
Sam Crawford		3009

The near misses:
Lou Brock		2964
Paul Waner		2932
Brooks Robinson		2912
Joe Morgan		2905
Roberto Clemente	2889
Cap Anson		2862
Rusty Staub		2862
Rod Carew		2859
Ted Williams		2858
Tony Perez		2847
Rogers Hornsby		2844
Billy Williams		2844
Andre Dawson		2837*
Mickey Mantle		2835
Lou Gehrig		2819
Five Year EPEQA

Five year EPEQA is one attempt to measure a player's peak. It is the highest EPEQA achieved by a player in any five consecutive years in which he had at least 2000 PA or 400 EPER. The year given indicates the start of the five year run. This list is very strongly weighted to players from before 1920, which makes the accomplishments of Barry Bonds and Frank Thomas all the more amazing.

Player			EPEQA

Babe Ruth		.3986	1920
Ted Williams		.3913	1953
Ty Cobb			.3847	1909
Rogers Hornsby		.3792	1921
Mickey Mantle		.3719	1954
Honus Wagner		.3708	1904
Joe Jackson		.3702	1909
Dan Brouthers		.3684	1882
Pete Browning		.3669	1881
Lou Gehrig		.3666	1927
Barry Bonds		.3615	1991*
Nap Lajoie		.3615	1901
Frank Thomas		.3599	1990*
Jimmie Foxx		.3597	1932
Stan Musial		.3582	1948
Tris Speaker		.3577	1912
Ed Delahanty		.3570	1895
Roger Connor		.3545	1882
Willie McCovey		.3543	1966
Billy Hamilton		.3537	1891
Eddie Collins		.3527	1911
Joe DiMaggio		.3509	1939
Joe Morgan		.3508	1972
Dave Orr		.3500	1883
John McGraw		.3496	1897
Frank Robinson		.3474	1966
King Kelly		.3471	1884
Johnny Mize		.3470	1936
Hank Aaron		.3460	1959
Al Simmons		.3460	1927
Harry Heilmann		.3458	1921
Willie Mays		.3456	1961
Mel Ott			.3435	1935
Frank Chance		.3430	1902
Willie Stargell		.3423	1971
Charlie Keller		.3422	1941
George Sisler		.3415	1919
Dick Allen		.3412	1963
Fred Dunlap		.3407	1880
Hank Greenberg		.3406	1934
Cap Anson		.3401	1880
Wade Boggs		.3396	1985*
Joe Kelley		.3390	1894
Roberto Clemente	.3387	1967
Harmon Killebrew	.3386	1966
Eddie Mathews		.3382	1953
Ralph Kiner		.3381	1947
George Stone		.3377	1903
Duke Snider		.3366	1952
Benny Kauff		.3362	1913
Tip O'Neill		.3362	1885
Mike Schmidt		.3354	1979
Rickey Henderson	.3353	1989*
Elmer Flick		.3352	1904
Rod Carew		.3350	1973
Frank Baker		.3347	1911
Jesse Burkett		.3346	1897
Carl Yastrzemski	.3338	1967
Arky Vaughan		.3336	1934
Frank Howard		.3335	1967
Tim Raines		.3334	1983*
Hack Wilson		.3332	1926
Harry Stovey		.3330	1882
Mike Tiernan		.3327	1888
Chuck Klein		.3319	1929
Al Kaline		.3314	1964
Pedro Guerrero		.3311	1985
Sherry Magee		.3311	1906
George Gore		.3310	1882
Mike Donlin		.3309	1901
Jackie Robinson		.3304	1949
Gavvy Cravath		.3302	1913
Babe Herman		.3301	1929
George Brett		.3298	1979
Willie Keeler		.3292	1895
Ken Singleton		.3285	1975
Eddie Murray		.3280	1981*
Denny Lyons		.3279	1887
Jake Stenzel		.3278	1893
Lefty O'Doul		.3274	1928
Norm Cash		.3270	1959
Jeff Bagwell		.3268	1990*
Fred McGriff		.3268	1988*
Joe Medwick		.3266	1935
Orlando Cepeda		.3265	1963
Will Clark		.3264	1988*
Ken Griffey, Jr.	.3263	1991*
Bill Terry		.3263	1930
Dolph Camilli		.3261	1937
Chick Hafey		.3261	1927
Ron Santo		.3259	1963
Augie Galan		.3257	1943
Sam Crawford		.3256	1907
Eric Davis		.3255	1985
Reggie Jackson		.3255	1973
Don Mattingly		.3250	1982*
Paul Waner		.3250	1925
Ken Williams		.3247	1921
Al Rosen		.3246	1950
Edgar Martinez		.3244	1991*
Five-Year EPER

Like the five year EPEQA, the five year EPER is the highest total EPER a player accumulated in any five consecutive seasons, with the year indicating the start of the run. The strike of 1994-95 almost certainly cost Bonds and Thomas spots in the very-exclusive 700 Club; the effects may be such that no new player will be able to get onto the list until 2000.

Player			EPER

Babe Ruth		809	1920
Lou Gehrig		789	1930
Ty Cobb			755	1908
Mickey Mantle		741	1954
Stan Musial		737	1948
Hank Aaron		736	1959
Rogers Hornsby		730	1920
Jimmie Foxx		720	1932
Willie Mays		719	1961
Honus Wagner		716	1904
Tris Speaker		699	1912
Ted Williams		696	1946
Barry Bonds		691	1989*
Joe Morgan		686	1972
Frank Thomas		679	1991*
Frank Robinson		663	1962
Carl Yastrzemski	662	1966
Eddie Collins		657	1909
Frank Howard		656	1967
Tim Raines		655	1983*
Dick Allen		654	1964
Wade Boggs		652	1985*
Willie McCovey		652	1966
Mel Ott			652	1932
Ralph Kiner		645	1947
Eddie Mathews		637	1953
Ron Santo		637	1963
Rod Carew		636	1973
Duke Snider		635	1952
Chuck Klein		634	1929
Joe DiMaggio		630	1937
Johnny Mize		629	1937
Bobby Bonds		626	1969
Don Mattingly		625	1984*
Billy Williams		624	1963
Roberto Clemente	623	1963
Pete Rose		623	1969
Ed Delahanty		619	1895
Joe Medwick		617	1935
Elmer Flick		613	1903
Joe Jackson		612	1911
Harmon Killebrew	612	1966
Will Clark		608	1988*
Lou Brock		607	1967
Harry Heilmann		605	1921
Dale Murphy		604	1983
Willie Stargell		603	1971
Sam Crawford		599	1907
Frank Baker		598	1910
Ernie Banks		598	1956
Mike Schmidt		598	1979
Rickey Henderson	597	1982*
George Sisler		597	1918
Fred McGriff		596	1988*
Tony Perez		596	1969
Ken Singleton		596	1975
Dan Brouthers		594	1885
Babe Herman		593	1929
Bill Terry		592	1928
Orlando Cepeda		591	1959
Hack Wilson		591	1926
Roger Connor		590	1885
Reggie Jackson		590	1973
Eddie Murray		590	1982*
Al Rosen		589	1950
Rusty Staub		588	1967
Billy Hamilton		587	1891
Nap Lajoie		587	1906
Bobby Murcer		583	1969
Dave Parker		583	1975
Jesse Burkett		582	1897
Bill Nicholson		582	1940
Paul Waner		581	1932
Steve Garvey		580	1974
Al Simmons		579	1929
Rocky Colavito		578	1961
Sherry Magee		578	1906
George Burns		577	1913
Cesar Cedeno		577	1972
Jim Rice		577	1975
George Brett		576	1976
Charlie Gehringer	576	1934
Ken Griffey, Jr.	575	1990*
Charlie Keller		575	1939
Tony Gwynn		571	1984*
Vada Pinson		571	1959
Dave Winfield		570	1976*
Cecil Cooper		569	1979
Paul Molitor		569	1989*
Wally Berger		568	1931
George Foster		568	1976
Tony Oliva		568	1964
Arky Vaughan		566	1933
Jim Wynn		566	1965
George Stone		565	1905
Jackie Robinson		564	1948
Ryne Sandberg		564	1988*
Roy White		563	1968
Harry Stovey		562	1885
Dolph Camilli		561	1938
High Career EPEQA

High Career EPEQA is a different take at measuring a player's peak. It tells us how high a player's career EPEQA got, at the end of any season, after he was eligible for the list. In other words, look at the player's EPEQA after he reached 4000 PA. The highest EPEQA at the end of any season thereafter is his high career EPEQA. It essentially removes the last few years of the player's career. It has the virtue, which the career EPEQA list does not have, is that you cannot move backwards on the list; your own high-water mark is there to stay, regardless of how long you continue to play out the string.

Player			EPEQA

Babe Ruth		.3861
Ted Williams		.3773
Ty Cobb			.3677
Lou Gehrig		.3573
Mickey Mantle		.3569
Dan Brouthers		.3544
Rogers Hornsby		.3540
Stan Musial		.3518
Jimmie Foxx		.3488
Honus Wagner		.3483
Eddie Collins		.3465
Tris Speaker		.3457
Pete Browning		.3456
Joe Jackson		.3434
Johnny Mize		.3431
Roger Connor		.3422
Billy Hamilton		.3418
Nap Lajoie		.3398
Willie Mays		.3396
Dick Allen		.3383
Joe DiMaggio		.3365
Hank Aaron		.3361
Charlie Keller		.3361
Cap Anson		.3353
Willie McCovey		.3351
Mel Ott			.3345
Barry Bonds		.3337*
Hank Greenberg		.3334
Wade Boggs		.3328*
George Sisler		.3316
Frank Robinson		.3311
Elmer Flick		.3291
Eddie Mathews		.3290
Chuck Klein		.3281
Al Simmons		.3275
Ralph Kiner		.3272
King Kelly		.3269
Ed Delahanty		.3262
John McGraw		.3262
Arky Vaughan		.3262
Rickey Henderson	.3256*
Tim Raines		.3246*
Jesse Burkett		.3245
Joe Morgan		.3243
Harry Stovey		.3239
Mike Donlin		.3232
Pedro Guerrero		.3231
Joe Kelley		.3231
Fred McGriff		.3231
Frank Baker		.3229
Harmon Killebrew	.3229
Frank Chance		.3223
Willie Stargell		.3222
Harry Heilmann		.3220
Hack Wilson		.3216
Jim O'Rourke		.3211
Paul Waner		.3211
Babe Herman		.3210
Reggie Jackson		.3208
Willie Keeler		.3208
Will Clark		.3204*
Sherry Magee		.3202
Duke Snider		.3200
Joe Medwick		.3198
Gavvy Cravath		.3197
Mike Schmidt		.3193
Jackie Robinson		.3190
Mike Tiernan		.3184
Sam Thompson		.3180
Norm Cash		.3178
Sam Crawford		.3178
Tip O'Neill		.3174
Eddie Murray		.3171*
Carl Yastrzemski	.3170
Don Mattingly		.3168*
Orlando Cepeda		.3162
Frank Howard		.3162
Bill Joyce		.3160
Larry Doby		.3158
Buck Ewing		.3158
George Gore		.3157
Tony Oliva		.3156
Ross Youngs		.3154
George Brett		.3153
Bill Terry		.3152
Rod Carew		.3151
Daryl Strawberry	.3151*
Al Kaline		.3150
Paul Hines		.3149
Jim Bottomly		.3147
Hugh Duffy		.3145
Ken Singleton		.3145
Kevin Mitchell		.3144*
Jake Fournier		.3142
Mark McGwire		.3142*
Edd Roush		.3141
Goose Goslin		.3139
Jim Wynn		.3130
Roy Thomas		.3129
Denny Lyons		.3128

BP 1996 Info Center | Sample Entries | Baseball Prospectus Home