Hall of Stats

For the longest time, the Hall of Stats “About” page said:

The Hall of Stats does not yet include Negro League players. With the emergence of the Seamheads Negro League Database, we’ll be looking into how we can incorporate these players as soon as possible.

As part of my professional work with Sports Reference, I was thrilled when we dramatically expanded our coverage of the Negro Leagues and historical Black major league players (powered by the Seamheads data). This included stats like WAR and WAA for all Negro Major League players between 1920 and 1948.

So, that solves it, right? I can now simply include Negro League players in the Hall of Stats and use their WAR and WAA numbers to power their Hall Rating. Mission accomplished.

Well, not so fast.

Baseball Reference is a historical record. It is important that Negro League teams and players are shown alongside their American League and National League counterparts (with as much context describing the different leagues as possible). The Hall of Stats is very different. The purpose of the Hall of Stats is to use a player’s statistics to show objectively how good a player’s Hall of Fame case is. For the most part, this works well for the American League, National League, and the other leagues that have long been considered “major” (such as the Federal League and American Association). But, the Negro Leagues are different.

First of all, the Negro Leagues played a shorter, more unbalanced schedule. Josh Gibson never played more than 69 games in an “official” Negro League season. But he was playing dozens of (or maybe even another 100-plus) games outside of the league schedule—barnstorming and exhibition games against all kinds of competition. The Seamheads database only includes the statistics played against other Negro Major League (or major-league quality independent Black Baseball) teams. This is why you see 165 home runs on Josh Gibson’s Baseball Reference page rather than the “almost 800” cited on his Hall of Fame plaque.

Additionally, the Negro Leagues data is not complete. Researchers have been working for decades to compile a more complete statistical record. They have done this by finding box scores and tabulating the results. They only include full box scores so that the data can be balanced between teams and between offense and pitching. Some seasons (like much of the 1920s) are pretty much complete. But for a player like Quincy Trouppe, we are missing quite a few box scores. For example, between 1945 and 1947, he played in five East-West All Star games, but we only have stats for 51 of his team’s 195 league games.

Players also played wherever they could (and where they could get the most money). This wasn’t always the Negro Leagues. Sometimes, they played with independent teams. Sometimes they played in Latin America. Sometimes they simply barnstormed. Many of the early stars of the Kansas City Monarchs got their start with the 25th Infantry Wreckers, an all-Black unit stationed in Hawaii. Hall of Famer Bullet Rogan spent his early years playing for the Wreckers before immediately dominating with the Monarchs. Other stars who deserve a closer look from Hall of Fame voters (Dobie Moore and Heavy Johnson, for example) also spent several years playing for the Wreckers. Their stats with the Wreckers are not part of any type of league system, but make no mistake—they were professional baseball players. And they were damn good.

Simply using traditional Hall Rating will not paint an accurate picture of the Hall-worthiness of Negro League players. And the last thing I want to do is mislead visitors about how amazing these players were.

That’s why I’m calling on the support of Negro Leagues researcher Eric Chalek (@EricChalek on Twitter). Eric has developed a series of Negro League MLEs (Major League Equivalencies). MLEs attempt to estimate what a player would do in a traditional major league environment. In this case, the MLEs specifically aim to estimate what a Negro Leagues player would do if he was placed in the AL/NL during his career.

It’s important to note that this does not estimate what the player would have done in a fully integrated AL/NL—just if the single player was dropped into the AL/NL. Calculating estimates for an integrated AL/NL would be a much larger project that would need to adjust the statistics of every player in the AL/NL as well (including removing many players who would have lost their jobs to better talent from the Negro Leagues). I’m not even sure how to attempt that type of project. The MLEs feel more manageable.

They also feel more accurate. The 2022 Early Baseball Era Committee ballot featured George Scales, a second baseman who played 20 years in the Negro Major Leagues. He had a 147 OPS+ and averaged 5.4 WAR per 162 games in the Negro Leagues. But due to the differences in Negro League data, his Hall Rating is 75—well short of the Hall of Fame borderline. Using Eric’s MLEs, his Hall Rating is 121 (more like a median Hall of Famer).

For more information about how Eric calculates his MLEs, view the methods for hitters or pitchers.

Here are some important things to know about how we’ll display MLEs on the site:

1. We do not have MLEs for every Negro League player. Since the focus here is on the Hall of Fame, I’ll try to have MLEs here for stars and other prominent players. Eric is calculating the MLEs himself, and I’ll be using his data. For the initial release, I’m including all players with an MLE Hall Rating of at least 50 (114 players).
2. By default, we display the Hall Rating calculated from the player’s stats on Baseball Reference. Then, if there are MLEs available for that player, we'll show a second Hall Rating.
3. This impacts four types of players.
   1. Players who only played in the Negro Leagues and never in the AL/NL (examples: Josh Gibson, Buck Leonard, Turkey Stearnes, Bullet Rogan): These players are not be included in overall and positional rankings on the Hall of Stats. This is because their Hall Ratings are not an accurate representation of their Hall-worthiness. There is a page for Negro League MLEs that ranks each player by their MLE Hall Rating. Positional rankings pages have a top five list of players by Negro League MLEs.
   2. Players who played in both the Negro Leagues and the AL/NL, but predominantly in the Negro Leagues (examples: Satchel Paige, Willard Brown, Quincy Trouppe, Sam Jethroe): These players similarly will not be included in overall and positional rankings.
   3. Players who played in both the Negro Leagues and the AL/NL, but predominantly in the AL/NL (examples: Jackie Robinson, Larry Doby, Minnie Miñoso, Roy Campanella): These players will be included in overall and positional rankings. This is because most of their career came in the AL/NL with more complete statistics. Their raw Negro League stats (not MLEs) will be included in their Hall Rating used in these rankings. Their player page will have both the standard Hall Rating and the Hall Rating supplemented by MLEs.
   4. Players who played in Black Baseball entirely before 1920 (examples: Grant Johnson, Julián Castillo, Rube Foster, Regino García): These players will only have an MLE Hall Rating as they do not have any “Major League” data from 1920–1948.
4. For two-way players, Eric calculates separate MLEs for them as a position player and a pitcher. In these cases, I use the higher MLE.
   1. Position Players: Bullet Rogan, Lázaro Salazar
   2. Pitchers: Martín Dihigo

Here’s one example from each of the types of players listed above:

1. Josh Gibson: Eric’s MLEs estimate that in the AL/NL Gibson would have moved from catcher to first base in his age-20 season (1932). He would accumulate 83.3 WAR in just over 9,000 plate appearances before passing away shortly after turning 35. Based on his Baseball Reference data, his Hall Rating is 144. But based on MLEs it is 170. The MLEs have Josh hitting .306/.375/.542 (160 OPS+) with 435 HR and 2498 H in his shortened AL/NL career. If anything, I think the MLEs are conservative.
2. Satchel Paige: Paige’s numbers from 1948 on are untouched because he was playing in the AL. But before that, the MLEs adjust his stats. He ends up with 95.8 WAR thanks to a 298-229 record, 3.28 ERA, and 127 ERA+. Again, I think these estimates are conservative. Using Baseball Reference data, his Hall Rating is 102. When calculated from MLEs, his Hall Rating rises to 188.
3. Minnie Miñoso: Miñoso’s numbers from 1951 onward are untouched. His three seasons in the Negro Leagues (1946–48) are adjusted, but so are the next two years. In 1949, he played nine games with Cleveland but was mostly stuck in the minors. In 1950, he spent the entire season in the minors despite excelling. The MLEs give him credit for these two seasons because all that was holding him back was race. You could argue that Cleveland was coming off a World Series championship and simply had no room for Miñoso—and you’d be partially right. But Miñoso had already established he was an All Star caliber player in a major league (149 OPS+ in 1947, 170 OPS+ in 1948, an East-West All Star both seasons). The Indians already had several Black players and in those early years of integration teams were reluctant to have “too many” Black players at a time. There were few options elsewhere for Miñoso as only two other teams had integrated at the time he made his AL debut. Miñoso is credited with 1.8 and 4.0 WAR seasons in 1949 and 1950. Along with the adjustments for his Negro League seasons, his career WAR increases from 53.8 to 65.6 and his Hall Rating from 111 to 127.
4. Grant Johnson: Johnson debuted in 1895 for the Paige Fence Giants and played through 1914 with the New York Lincoln Giants. In his 18 seasons, he was a dominant force at the plate and a steady middle infielder. Johnson’s Hall Rating based on his MLEs is 137.

I just want to reiterate that I feel these estimates are conservative. The highest MLE Hall Rating is Joe Williams at 218. That would not put him among the Top 25 players all time. I think we all know that Williams, Gibson, Paige, Oscar Charleston, and others would be right up there with the best. The fact that these MLEs are conservative should tell us even more about the ratings of some of the more overlooked players like John Beckwith, Heavy Johnson, and George Scales.

Nothing I’ve found has made it more clear how dominant Black and Latino players were than this tweet:

The Negro major leagues were in operation from 1920-1948 (29 years). How much top talent was kept out of the AL & NL during that time? The NEXT 29 years (1949-1977), these were the top 40 players by position player WAR. Players in orange would not have been allowed in the majors. pic.twitter.com/uko6ZBjE5J

— Adam Darowski (@baseballtwit) August 25, 2021

First and foremost, I just want to make sure I’m presenting Negro League players as respectfully and accurately as possible. While keeping Negro League players separate from their AL/NL counterparts and using estimates as data seem at first to run counter to these goals, the more I’ve thought about it, the more I think it is the best path forward. While the Negro Leagues were indeed major leagues, they were very different. Those differences need to be explained and celebrated rather than forced to exist within the AL/NL framework. I’m hoping that this presentation achieves this goal.

More on Building the Ballot

My Building the Ballot podcast has some conversations directly related to the Negro League MLE’s. First, on Episode 10 Eric Chalek joined to discuss his MLEs (this was before they were added to the site).

Then on Episode 13, Kevin Johnson of the Seamheads Negro League Database joins to discuss the top Negro League (and Latin) players outside the Hall of Fame, using the Negro League MLEs to frame the discussion.

Update (January 13, 2021): Thanks to Eric Chalek, we now have MLEs for 413 Black and Latin players.

Special thanks to all of the researchers who put in the work to make the Seamheads Negro League Database possible—particularly Larry Lester, Gary Ashwill, Scott Simkus, Mike Lynch and Kevin Johnson.