January 6, 2015
The Changing Hall of Fame Criteria
Every year, the Hall of Fame vote brings a great deal of vitriol to baseball. With each year’s ballot, we are confronted by the specter of the steroid era, always a sore subject. But even neglecting the steroid era candidates, the BBWAA voters manage to produce a handful of idiotic ballots, defended with harebrained rationales, sometimes leading to obvious omissions.
It would be easy to pin the Hall’s recent mismanagement solely on the steroid issue, but the problems do not stop there. There’s a clear backlog of players that’s been developing for more than ten years, leaving deserving stars (with no steroid evidence against them) like Tim Raines and Curt Schilling on the outside. The situation is especially dire for pitchers, where the voters seem to rely upon outdated benchmarks like 300 wins, which even the best modern pitchers simply cannot hope to reach. This failure is through no fault of their own—pitcher usage patterns and injuries have changed the game. Clinging to the old milestones has the effect of artificially increasing the standards for induction, so that only the most inner-circle, obvious Hall players can make it.
In truth, the Hall’s problems go back a long way. As long as the BBWAA has been voting, the standards for induction seem to have been changing:
This graph shows the Hall of Fame Votes per Ballot as a function of time (data from Baseball-Reference). As should be obvious, the votes per ballot have been falling since the beginning. Part of that can be explained away by the fact that the early ballots were stuffed with deserving candidates from the past 50 years of baseball, during which the Hall had not existed.
But by the '60s, that backlog had been worked through. From about 1961 to 1982, one sees a relatively steady number of votes per ballot, hovering between seven and eight. Then the bottom drops out: from about 1983 onwards, the voting falls to about five votes per ballot.
This decrease in the voting pattern doesn’t coincide with the obvious beginning of the steroid era. In the later part of the '90s, voters were evaluating players like Phil Niekro, Mike Schmidt, and Don Sutton, players who, to my knowledge, are untainted by any suspicions of steroids. It does coincide, however, with a profound change in how pitchers had been used. From 1970-1979, the average league-leading IP total was 342.3; from 1980-1989, it had dropped down to 281.7, an extraordinary decrease.
Just as the votes were falling, the ballots were increasing:
The extent to which the BBWAA’s voter pool has increased is astonishing. From the beginning, when the ballots sat near 200, the number of returned ballots has nearly tripled to 571 last year. There certainly aren’t three times as many baseball teams playing nowadays as there were in 1940, nor is there three times as much baseball being played, so the cause of the increase is somewhat befuddling.
Without ballot-by-ballot information, we can’t say much about all the newly minted Hall of Fame voters. Ryan Thibs and others have done yeoman’s work to collect the publicly available ballot information, but voters are under no imperative to reveal their ballots. If they do not wish to provide transparency, they can remain anonymous, and we have no idea what kind of ballot they submit.
One small clue comes from the BBWAA’s own website, which lists the publication with which each BBWAA voter is affiliated. Some voters are listed as “Honorary”: these are voters who are no longer actively covering baseball, but have been granted lifetime voting rights. Interestingly, these voters skew towards fewer votes per ballot than the non-honorary members. In 2013, for example, honorary voters wrote down 6.25 votes/ballot, versus 7.02 votes/ballot for non-honorary members (a weakly statistically significant difference, p = .045). The difference is much smaller for 2014, so perhaps the gap between honorary and non-honorary members is shrinking.
While the votes per ballot are decreasing, the decrease could be affecting voting totals in one of two ways: 1) all finishers could be losing votes, regardless of position on the ballot; or 2) the top of the ballot could be remaining steady, while the bottom of the ballot declines. In the first case, this might indicate a general distaste for the Hall, a feeling that no one deserves inclusion, no matter their resume (plausible, for example, if a voter felt inclined to dismiss all steroid-era candidates). The second case seems to be a stronger statement on the kind of Hall the voters want, whether it should be Large or Small, encompassing the marginal candidates or only the very greatest. To test these possibilities, I made the following graph, which charts the percent of ballots for the 1st-8th place finishers over time (using LOESS smoothing):
The first, second, and third place finishers have suffered no loss in votes at all, continuing steadily onward. Indeed, the candidates who have lost votes have been primarily down-ballot, with the fifth place and lower finishers suffering most substantially.
To put it another way, the loss in votes has been extremely variable, depending on a player’s position on a ballot. The top of the ballot, which is most often occupied by the inner-circle caliber players, are still getting elected at their usual rate. But the more marginal players, often those who fail to achieve the now-obsolete career milestones, are being truncated. It is as though a conscious decision has been made on the part of the voters that the Hall should be smaller, that it should be composed of only the most elite players.
There’s several problems at work in the Hall of Fame voting, all of which may be interrelated. Votes per ballot have fallen at an alarming rate, even while the number of ballots skyrockets to tremendous heights. Some of the vote per ballot decrease can be pinned upon the honorary voters, who no longer cover baseball and seem to vote much more sparingly. What’s more, the loss in votes is extremely heterogeneous, affecting the top of the ballot minimally, instead changing the standards for the down-ballot finishers.
Intriguingly, most of these problems predate the steroid-era ballots. The falling tide of votes began soon after the Hall opened, and reached its present nadir in the mid-90s, long before PEDs were (overtly) on the voters’ minds. The meteoric increase in ballots has been going on as long as the BBWAA has been charged with voting for the Hall.
The new set of steroid-using candidates, most especially Barry Bonds and Roger Clemens, have driven a wedge between the voters that is clearly visible in their voting preferences. It has exacerbated these issues by stratifying the electorate, causing some writers to spend votes on two or more steroid users, while others avoid them entirely. The result has been a sort of electoral constipation, in which the divide between PED voters and non-PED voters increases the backlog, further enlarging the pool of possible candidates, thus dividing the support more evenly between a larger and larger number of players. Consequently, fewer and fewer players are elected.
Therefore we arrive at the present year, in which every shade of Hall-eligible players exists, from the transcendent, first-ballot types, some of whom are ostensibly clean, to the marginal candidates, deserving perhaps only a recognition vote from their local beat writer. Early returns suggest that the results of this year’s election will be much the same as the recent pattern: a handful of obvious honorees (Pedro Martinez, Randy Johnson, John Smoltz), combined with one or two who sneak over the line (Craig Biggio, Jeff Bagwell, or maybe Schilling). The rest will likely see their totals shuffle around a bit, without making substantial progress towards election or off the ballot. Without an obvious change to the voting or ballot-granting procedures upcoming, and with years of eligibility remaining for the confirmed PED users, we should get used to this state of affairs, because there’s no sign of it changing any time soon.