Weighted WAR: Negro Leagues and Catcher Rankings

So far in the Weighted WAR series, we’ve covered the weighting system, schedule length adjustments, positional classification, timelining (part 1 and part 2), and negative seasons. Now, having covered all of the adjustments we’re making to WAR, it’s time to go through all of them again.

MLB now recognizes Negro League statistics from 1920-48. (This is not the totality of segregated black baseball, and it’s possible more will be added to the official record at some point, but for now I’m limiting my analysis to the MLB-recognized seasons.) WAR totals are also available for these years, via Baseball Reference and other sources. In this case, despite bWAR being my usual metric, I’ve decided to use Seamheads WAR instead; my understanding is that B-R doesn’t use park adjustments for NeL seasons, and Seamheads does.

Let’s go through how the previously explored WAR adjustments apply to the Negro Leagues, starting with the smallest changes. The weighting system is the same. Players who change teams in-season still have their overall yearly totals used. Negative seasons (after adjustments) are still dropped from the calculations.

Positional classification still follows the same basic principle; the slight difference is that Seamheads doesn’t have the same level of detail available in fielding data. Fortunately, Seamheads already classifies each player-season by a primary position, so I was able to use their results in that regard.

My tentative understanding is that the Negro Leagues never built a stable minor league feeder system, so the minor league control adjustment (half a win per 600 PA) is at full strength throughout their run. (I am not terribly well-informed on this topic and am open to being presented with information to the contrary.)

Expansion is something of a tricky issue with regard to Negro League seasons, because it so often involves existing independent teams joining a league (or several existing independent teams grouping together to form a league) rather than new franchises being created from scratch. As such, I have reduced this adjustment to half strength and removed some (but not all) of the counteradjustment for shrinking leagues. Here is a table of the applied adjustments (as usual, per 600 PA):

Year	Teams Added	Adj
1920	8	-0.4
1921	0	-0.35
1922	0	-0.3
1923	6	-0.55
1924	3	-0.65
1925	0	-0.6
1926	0	-0.55
1927	0	-0.5
1928	0	-0.45
1929	0	-0.4
1930	0	-0.35
1931	0	-0.3
1932	5	-0.5
1933	-6	-0.15
1934	0	-0.1
1935	0	-0.05
1936	0	0
1937	7	-0.35
1938	0	-0.3
1939	0	-0.25
1940	0	-0.2
1941	0	-0.15
1942	0	-0.1
1943	0	-0.05
1944-48	0	0

Segregation, obviously, is a major factor in the Negro Leagues, since it is the reason the leagues existed. NeL play can reasonably be held to have been more affected by segregation than MLB play since the pool of players was smaller, both in raw population base and in eventual fraction of the integrated major leagues. Without diving too deep into the kind of research that would make me feel like I need a long shower afterward, I settled on a segregation adjustment double the strength of the one I used for MLB (1 win per 600 PA) for 1920-46, and as the MLB adjustment starts to phase out in 1947, the NeL adjustment gets larger as the league loses players. The three timeline adjustments produce a total that varies between 1.5 wins per 600 (1936, 1944-46) and 2.15 wins per 600 (1924).

Schedule length seems like it should be straightforward – NeL teams played widely varied schedules, but so did early MLB teams, and we have a method to handle those. But actually, this is the trickiest modification of the bunch. The trouble is that, unlike in the white and integrated major leagues, we don’t have individual statistics available for every league game, even if the result of the game itself is known. This difference is often a huge factor – the 1944 Memphis Red Sox played 143 total games with known results (118 league games, 25 others), but we only have individual stats for 31. So, instead of just using total league games played by the team for the schedule length adjustment, we use the smaller value between league games and total games with individual stats (estimated by using total games started by the team’s pitchers). The latter total may include some non-league games, so it’s not a flawless solution, but it’s the best alternative I could find that wouldn’t have been likely to require months of research.

And with all of that, we still aren’t quite done, because we have to examine players who spent time in both MLB and the Negro Leagues. This is less straightforward than simply looking at players who appear in both databases, because not everyone who split time has a schedule-adjusted 3-WAR season in both contexts (which, as before, was the inclusion criteria for NeL analysis). Fortunately, Wikipedia provides a list of Negro Leaguers who made it to the integrated majors, and from there it's simple to figure out which ones had timeline-adjusted positive value in both leagues, therefore requiring a hybrid total to be generated. (Sorry, Willie Mays; the 83 PA of .239 hitting for the 1948 Birmingham Black Barons score at 0 WAR even before adjustment.)

It should be noted before we get to the lists that there are things missing. As noted above, we don’t have full statistical records even for the included seasons. Many of these players also spent their winters playing in places like Cuba or Mexico; any statistics accumulated there are not included at this time, nor are those accrued for independent teams that predated the formation of official black leagues, or played outside the available league structures. The lists that will follow are not intended to be comprehensive in capturing the best pre-integration black players, they’re just the best I know how to do with the officially recognized stats.

With that in mind, a couple of tables. First, every Negro Leagues-only player with a weighted WAR total of at least 23:

Player	Pos	Years	WAR	aWAR	wWAR
Josh Gibson	C	1930-46	40.0	74.9	58.2
Oscar Charleston	CF	1920-41	48.8	70.3	55.9
Turkey Stearnes	CF	1923-40	48.8	75.0	51.1
Willie Wells	SS	1924-48	43.0	59.3	45.4
Buck Leonard	1B	1935-48	31.6	60.1	45.1
Jud Wilson	3B	1923-45	32.7	51.3	41.1
Cristobal Torriente	CF	1920-32	30.3	42.8	38.7
Dobie Moore	SS	1920-26	29.3	38.6	34.6
Mule Suttles	1B	1924-44	32.9	44.9	32.6
John Beckwith	3B	1920-35	21.8	32.6	28.2
Newt Allen	2B	1923-47	25.3	31.6	27.1
Ted Strong	RF	1937-48	12.3	30.4	26.6
Dick Lundy	SS	1923-37	21.1	29.3	25.5
Bullet Rogan	CF	1920-38	24.7	30.2	25.4
George Scales	2B	1921-46	18.1	30.0	24.2
Biz Mackey	C	1920-47	25.2	28.5	23.4
Martin Dihigo	1B	1923-45	14.3	25.0	23.2

Second, every NeL-MLB hybrid career meeting the same standard:

Player	Pos	Years	WAR	aWAR	wWAR
Jackie Robinson	2B	1945-56	64.5	65.3	52.9
Larry Doby	CF	1942-59	57.8	64.3	47.3
Minnie Minoso	LF	1946-80	54.0	58.4	44.5
Willard Brown*	CF	1937-48	24.1	49.8	41.4
Roy Campanella	C	1937-57	42.2	47.0	38.0
Jim Gilliam	2B	1946-66	45.4	45.3	35.6
Monte Irvin	LF	1938-56	31.5	38.0	30.6
Hank Thompson	3B	1943-56	31.8	35.3	29.0
Sam Jethroe	CF	1938-54	16.7	31.4	26.7

Willard Brown gets an asterisk here for an upsetting reason, to put it mildly. Brown, unlike the other players listed above, did not post any positive WAR totals in his MLB career, which normally would mean it would be discarded. In this case, it factors in because his brief stint with the St. Louis Browns (in which his teammates treated him abominably even by the standards of the time) came in 1947, a year in which he also had a productive NeL season, so the “combine stats across multiple teams in the same season” rule comes into effect. If you want to waive this adjustment in Brown’s case, thereby bumping his score up to 41.9 (and pushing him to #23 in the CF rankings, ahead of Vada Pinson and Kirby Puckett), that is eminently understandable.

On to the usual tables for our position of the day, catcher. First, active players who are among the top 100 catchers of all time, or close enough to join with a good year:

Player	Rank	Years	WAR	aWAR	wWAR	2024 WAR	Rank Change
JT Realmuto	27	2014-24	36.4	37.9	31.0	3.0	+3
Salvador Perez	31	2011-24	35.7	38.1	29.6	2.5	+1
Willson Contreras	45	2016-24	27.2	28.3	23.5	3.0	+12
Yasmani Grandal	73	2012-24	20.3	23.4	19.3	0.6	0
Yan Gomes	77	2012-24	18.7	21.2	18.1	-0.7	0
Will Smith	79	2019-24	18.4	20.0	17.9	3.5	+17
Gary Sanchez	108	2015-24	15.0	15.3	13.8
Sean Murphy	122	2019-24	13.2	14.5	13.3
Adley Rutschman	133	2022-24	13.1	13.1	12.6
Cal Raleigh	149	2021-24	11.3	11.8	11.3
William Contreras	153	2020-24	11.5	11.6	11.1

Not a great selection of active players, with nobody in the top 25 and only three in the top 70. Fortunately, there is at least a reasonable class of young players who are 1-2 good years away from reinforcing the veterans.

And now, the overall top 25 catchers, plus numbers our standard multiple-of-10 extras:

Player	Rank	Years	WAR	aWAR	wWAR
Josh Gibson	1	1930-46	40.0	74.9	58.2
Johnny Bench	2	1967-83	75.4	75.3	55.3
Gary Carter	3	1974-92	70.2	71.7	54.1
Ivan Rodriguez	4	1991-2011	68.8	69.7	47.9
Mike Piazza	5	1992-2007	59.4	60.6	47.5
Carlton Fisk	6	1969-93	68.6	72.2	47.2
Joe Mauer	7	2004-18	55.1	55.0	42.1
Yogi Berra	8	1946-65	59.7	57.2	41.9
Joe Torre	9	1960-77	57.6	56.1	40.9
Ted Simmons	10	1968-88	50.3	53.2	39.7
Deacon White	11	1871-90	45.9	51.6	38.6
Roy Campanella	12	1937-57	42.2	47.0	38.0
Bill Dickey	13	1928-46	56.5	50.9	37.5
Thurman Munson	14	1969-79	46.1	45.4	37.3
Gene Tenace	15	1969-83	47.0	47.5	37.1
Buster Posey	16	2009-21	45.0	45.1	37.0
Mickey Cochrane	17	1925-37	49.9	44.8	35.5
Bill Freehan	18	1961-76	44.9	43.7	34.7
Gabby Hartnett	19	1922-41	55.4	49.6	34.3
Jorge Posada	20	1995-2011	42.7	43.2	34.3
Buck Ewing	21	1880-97	48.2	46.1	34.2
Charlie Bennett	22	1878-93	38.8	40.3	33.0
Jim Sundberg	23	1974-89	40.5	41.9	32.5
Jason Kendall	24	1996-2010	41.5	41.3	32.2
Darrell Porter	25	1971-87	40.9	41.7	32.0
Wally Schang	30	1913-31	48.0	41.6	29.7
Ernie Lombardi	40	1931-47	37.6	32.3	24.5
Mike Scioscia	50	1980-92	26.1	27.3	22.5
Walker Cooper	60	1940-57	27.3	25.0	20.6
Duke Farrell	70	1888-1905	31.3	25.0	19.5
Jack Clements	80	1884-1900	26.7	21.0	17.7
Earl Battey	90	1955-67	18.9	18.7	16.4
Brad Ausmus	100	1993-2010	16.4	17.5	14.9

So, umm, you might notice some slight differences between the numbers put up by catchers and the players at other positions – notably, the fact that Mike Scioscia’s #50 score behind the plate is slightly lower than Stan Javier’s #100 mark in center field. Catcher is very rough physically; it’s hard for catchers to amass as much value as their peers at other positions, either within a single season or over the course of a career. That’s part of the reason I’m posting positional lists rather than an overall top 100; it will be possible for the reader to compile a list of the overall top 100 at the end of the series (rounding notwithstanding), but I’m not entirely comfortable with the conclusion that there are no catchers among the 30 best position players ever, and only three in the top 95. And if you start applying bonuses, you immediately run into a tangle of how to handle catchers who moved around, both in-season and over the course of their careers. Positional lists allow me to avoid diving into that particular briar patch.

Speaking of players who move between positions, readers who are familiar with the Negro Leagues might have noticed a couple of names in the NeL weighted WAR table who showed up quite a bit lower than expected. In particular, Hall of Famers Bullet Rogan and Martin Dihigo seem to be shortchanged somewhat. Rogan is credited with 61 WAR by Baseball Reference; are park effects and missing games really enough to drop him to the 24 shown above?

They aren’t. Coming up next time, we’ll examine the best first basemen ever, and also discuss the two positions that don’t get a list to themselves in this series: DH and pitcher.