Cooperstown Notebook: Insights from the Spreadsheets, Part 2

Our story so far: At the end of every Hall of Fame election cycle, I have a set of spreadsheets that I update that help me track voting patterns and other long-term trends, as well as some demographics regarding any honorees. Because the 2021 election cycle yielded no honorees — BBWAA voters pitched a shutout, and the two Era Committee votes were postponed — I realized while going through this year’s post-election exercise that I had yet to reckon with the impact of Major League Baseball’s 2020 decision to recognize seven Negro Leagues that operated from 1920 to ’48 as major leagues. Not that I haven’t covered various angles of that decision, particularly as they pertain to the Hall; just that my tools of the trade haven’t kept pace.

I concluded my previous installment with a timeline illustrating the number of active Hall of Famers per team per season, using the Hall’s definition of one game being enough to represent one season played. It’s a display that illustrates the saturation of the immediate pre-World War II era via a very generous Veterans Committee and the extent to which voters haven’t kept pace with the later waves of expansion.

As previously noted, the above version does not include the 28 Hall of Famers elected for their playing careers in the Negro Leagues, a few of whom (Willard Brown, Monte Irvin, and Satchel Paige) had stints in the American and/or National League once they integrated, but not the 10 years needed to wind up on a BBWAA ballot. Adding those players compresses the pre-war peak, quite noticeably:

The scales are the same on the two graphs, but the broad peak in the middle is lower, with the space in the 2.5–3.0 range nearly empty. I did away with the BBWAA/committee distinction on this one, because I can’t stack the values if the denominators are different; the white Hall of Famers are coming from one player pool that for the 1920–48 period was constant at 32 teams, and the Negro Leagues Hall of Famers are coming from another that for the period in question ranged between six and 19 teams.

(Note that I’ve counted every team, even though some were very short-lived, loosely affiliated, and/or lacking in data that’s up to the Seamheads/Baseball Reference standard. I did draw the line by excluding the 1933 Cleveland Giants, who have data for just Negro National League games on B-Ref; they were apparently the replacement for the Columbus Blue Birds, who dropped out, and all 10 of their players appeared with other teams in the league during the same year, so leaving them out seems appropriate. Within the 49 Negro League-seasons now counted as majors, there might be a few other such instances of teams that shouldn’t be double-counted for the purposes of this exercise, but that will require closer study.)

Here’s a table separating these periods into buckets. I’ve used the same cutoffs as for the graph I’ve re-presented from the first installment:

What’s changed from the first graph is that the 1920–41 average is way down from a high of 2.76 players per team per season, and the 1946–68 average is slightly down from 2.01. Beyond that, it’s worth noting that by these measures, the Negro Leagues’ representation within the Hall is pretty light compared to their white contemporaries, coming in at about 80% of the BBWAA average for the 1920–48 period but only about 42% of the average of the BBWAA plus the committees. There are other ways to measure this, and it’s possible that my lack of vetting those other league-seasons might reduce the count of teams and raise the rates somewhat. But this gets at the efforts of the “42 for 21” Committee to press for the elections of more Negro Leagues players to correct a demographic disparity among the honorees — one that would require the Hall of Fame to consider such candidates more than once every 10 years.

Looking at the Hall breakdown in what may be a more familiar way, a few years back, Hall of Fame chairman Jane Forbes Clark and president Jeff Idelson often used the talking point that only the top 1% of major leaguers wind up in the Hall. “In the long first of professional baseball there have been just over 18,500 players who have been privileged enough to wear a major league uniform,” said Idelson at the 2016 induction ceremony. “And only one percent, one out of a 100, make it to Cooperstown.”

The point did not go unexamined. Updating a table that appeared in my work at SI.com and then in the Casebook, the percentage is higher than that — significantly higher except for players born after 1970:

For certain decades, the percentages are more than two and a half times as high as Clark’s and Idelson’s 1%, and for all of the others through 1960, they’re at least 18% higher and sometimes over 50% higher; the overall average is closer to 1.6% than 1%. While Clark and Idelson included over 1,000 more recent players in their denominator — a count much-inflated by expanding pitching staffs — that’s still a significant difference that overstates the selectivity of the voters, as they’ve been much more generous than that. A Hall of Fame with 1% of the players for the above period would have 170 players; instead, there are nearly 100 more than that.

The percentage for the 1970s group is particularly low in part because the book on that period is still open, so to speak. Of the 15 returning candidates on next year’s BBWAA ballot, the top four — Scott Rolen, Todd Helton, Billy Wagner, and Andruw Jones — were all born in the 1970s, as were eight of the other 11 candidates (all but Jeff Kent, Gary Sheffield, and Omar Vizquel). The top newcomers for 2023 and ’24, Carlos Beltran and Adrian Beltré, were born in that decade as well.

It would take all of those players being elected and then some just to get to 1% for the 1970–79 period, which points to two other issues suppressing that rate: a hesitance or refusal to elect players linked to performance-enhancing drugs (Manny Ramirez and Alex Rodriguez were born in the ’70s, and as you can imagine, the 1960s percentage has taken its hits as well), and the aforementioned proliferation of pitchers. The latter matter has had a longer-term impact on rates, which helps to explain why the percentages from 1940 onward are substantially lower than those prior.

Thus it makes sense to look at those rates in a couple of other ways, though before we do, it’s also worth noting the high rates at which Negro Leagues players from 1909 and before were elected — a finding that’s somewhat surprising in light of what I noted above. Smaller rosters had something to do with that, and perhaps this methodology overcounts some transient teams. The matter deserves a closer look at some point down the road if we’re to continue to make the case for the elections of significant numbers of Negro Leagues players.

Moving along, since the Hall’s voting requirements include the need for players to have participated in 10 seasons, we can look at the percentages of those players:

Note that this tally includes only the non-Negro Leagues players. Six of the 28 Hall of Famers from the Negro Leagues have fewer than 10 years in the 1920–48 majors, but they were elected before that classification came about; players such as John Henry Lloyd and Joe Williams were credited with meeting the 10-year minimum based upon their starring in pre-1920 leagues in the United States (the same was true for 2022 Early Baseball candidates such as Grant Johnson and Dick Redding). The same was true for Martín Dihigo and others with respect to foreign leagues. I’ll save that aspect of the accounting for another day. Meanwhile, one AL player has been excluded from the count (hence 239 Hall of Famers instead of 240): Addie Joss, who died of tubercular meningitis in 1911 ahead of what would have been his 10th season. In 1977, the Hall’s board of directors waived the 10-year requirement, and he was elected on the following year’s ballot by the Veterans Committee.

From here you can see that nearly 12% of players born in 1939 or earlier reached 10-year status, compared to just over 4% of those born later. Note that the ratio here for the older set of players is much higher than that of the first table, 2.74 versus 1.66. Again, the impacts of PEDs and extra pitchers have something to do with the discrepancy, as does the lesser amount of time for smaller committees to consider these candidates, though those committees have done a better job of holding candidates to higher standards. You can point to a handful of selections from the late 1960s and early 70s Veterans Committees featuring Frankie Frisch and/or Bill Terry for every Jack Morris or Harold Baines that grinds one’s gears.

The percentages are higher but the ratios similar when one limits the player pool to the practical cutoffs of 5,000 plate appearances or 2,000 innings pitched, which cover the vast majority of Hall of Famers from the NL, AL and bygone white leagues, but only Cool Papa Bell from among the Negro Leagues honorees. Dizzy Dean and all of the elected relievers save for Hoyt Wilhelm and Dennis Eckersley fall short of 2,000 innings, and only Roy Campanella is short of 5,000 PA — though he crosses that threshold once his eight Negro Leagues seasons (the first one of which is his age-15 season!) are included.

For most of those decades, the percentages of Hall of Famers from among the playing time qualifiers are more than double those of the 10-year qualifiers; even after weeding out the relievers and the role players, you can see that there’s been a significant downturn in the rates at which more recent players have been elected. Nearly a quarter of those who reach those thresholds and were born prior to 1939 are enshrined, but that’s true of just under 10% from 1940 onward, a ratio of about 2.5.

This shouldn’t be taken as a literal call for current and future voters to open the floodgates and double the overall rate of election for post-1939 players, so much as to point out that the hand-wringing about “too many” recent honorees is off base. Even with the recent record-setting run of BBWAA honorees (22 from 2014 to ’20) and the Era Committees electing living ex-players after a drought of nearly two decades, the more modern rates have a long way to go before they’re in the ballpark of those born in 1939 or earlier. Likewise when it comes to voters using all 10 spots or something close to it; from 2014 to ’19, 41–51% of voters maxed out their ballots annually, and new post-1966 records for players per ballot were set three times.

All of this is worth thinking about as we circle back to the topic that sparked this series: the election rates of recent starting pitchers and the new light that my experimental S-JAWS metric casts upon them. I’ll take up that topic in my next installment.