Skip to content

MLB Hall of Fame Voting Trajectories

Kenny Shirley sends along this interactive data visualization:

Screen Shot 2013-01-23 at 9.09.23 PM

What I learned from this was that Jim Rice is in the Hall of Fame! I remember watching him play. Whenever he struck out with a man on first base, we were just so relieved that he hadn’t hit into a double play.


  1. Steve Sailer says:

    That’s hilarious.

    I see at Baseball Reference that Rice is 6th all time in GIDP, and he did it in just 16 years, while everybody ahead of him played at least 21 years.

  2. Chris G says:

    >Whenever he struck out with a man on first base, we were just so relieved that he hadn’t hit into a double play.

    Me too! He hit some towering homeruns and I once heard a story that he was so strong that he could snap a bat when he checked his swing but, sadly, more than anything else I remember him a GIDP machine.

  3. Phil says:

    I remember a story about Rice breaking his bat on a checked swing, and the umpire saying “We’ll just call that a ball, OK Jim?”…I assume it’s apocryphal. I’m no expert on baseball, but based on Rice’s stats he’s hardly stinking up the Hall of Fame, although leaving him out would have been OK too. He had 5 truly excellent years (one of them at age 33).

    For me, as a casual baseball fan with nearly no memory or knowledge of player performances, the surprising thing is how many famous players are NOT in the hall. Johnny Sain (of “Spahn and Sain and pray for rain” fame)! Luis Tiant (a pitcher I know nearly nothing about, except that, when stopped by a cop for going something like 90 mph, he told the officer he was “bringing some heat.”) Gil Hodges! Steve Garvey! And Roger Maris! Let me say that again: Roger Maris! I know only two things about Roger Maris: (1) he played for the Yankees (for at least part of his career) and (2) he hit 61 home runs in a single season, a record that held until two steroid-assisted sluggers broke it something like 35 years later. I would have thought that if you hit 61 home runs in a season you must be terrific, but evidently not. Maybe he only had the one good year, or maybe he struck out a lot. A few moments on the intertubes will tell me, and I will indeed check. At any rate it’s surprising to me.

    Garvey’s trajectory is interesting (easiest to see if you just look at first basemen by clicking the 1B box in the menu at the right side): he started out with a lot of support, which fell off steadily with time. His voting window coincides with the sabermetric revolution, so I’m guessing he looks worse when analyzed with today’s understanding than he did when he was playing.

    • Kenny Shirley says:

      Andrew — thanks for posting this – we’d love lots of feedback.

      Phil: Great points, I think these borderline cases are the most interesting ones to look at. I was also surprised about Maris, but it turns out he really only had 3-4 good years. Garvey was interesting — he currently has the highest first-year vote percentage (41.6%) of anybody who did not eventually get elected (since 1967, the “modern era” of BBWAA voting), although Lee Smith (42.3% initial vote, now 11 years on ballot) will probably take his place soon, and a few still-eligible players also had higher first-year vote percentages (Biggio, Piazza, Bagwell), although these three will all probably get inducted eventually, leaving Garvey and Smith as the 2 main tough-luck cases. I looked at these guys by selecting First-Year-%, First Year on Ballot, and Number of Years on Ballot as histograms, and varying the selection based on the “Inducted by” legend.

      On the flip side, of those who have been inducted since 1967, the lowest first-year vote percentage belongs to Duke Snider (17.0%, 1970), and Bery Blyleven also had a very low first year percentage (17.55%, 1998). I selected First-year-% as the y-variable in the scatterplot on the right to see it’s distribution a little more clearly.

      So if the “zone of uncertainty” (based on the first year vote percentage, at least) is between about 15% and 45%, say, is there anything that predicts which of these players will make the hall, and which will never get in? There are 39 such players since 1967:

      12 are in, 14 are out, and 13 are still eligible. We thought batter vs. pitcher might be a good predictor, but it’s not (Of the 12 inductees, 6 are batters and 6 are pitchers. Of the 14 non-inductees, 9 are batters and 5 are pitchers).

      I think these trajectories are dying for a pretty simple, varying-intercepts, varying slopes model as a function of career statistics. Although steroids will surely cause a lack of fit for the next few years…

  4. Jim Bouldin says:

    That Tiant quote is hilarious, never heard that one. Tiant was the AL ERA leader in 1968, then with the Indians, the “Year of the Pitcher”. Same year that Gibson in the NL had the amazing 1.12 season ERA. Tiant’s was higher, but below 2, around 1.8 I think. He had a number of good years with the Red Sox in the 1970s, but that 1968 season was his best

  5. Steve Sailer says:

    First baseman Steve Garvey of the 1970s Dodgers didn’t walk much, so the rise of sabermetricians, with their emphasis on on-base percentage rather than batting average, hurt his reputation.

    Also, he had a poor throwing arm, so he looks bad defensively, but I think sabermetricians haven’t done enough work that would show that he actually was extremely useful defensively in a strategic sense. He was an absolute vacuum cleaner at grabbing throws in the dirt. This allowed the Dodgers to play three strong offensive players, Ron Cey, Davey Lopes, and Bill Russell at the other infield positions despite their scatter-shot arms (they were instructed to aim low and let Garvey do the work), and play the same infield for eight straight years (an MLB record), during which they went to four World Series. This gave the Dodgers a huge advantage in that their infield was set with four potential All-Stars for most of a decade, letting them wheel and deal to get good outfielders and pitchers. But without Garvey there to suck up anything thrown in the dirt, there would have been defensive crises that wreck teams.

    Interestingly, Garvey was a bigger star across the country, where his gaudy string of 200 hit seasons and RBI totals was impressive, than with Dodger fans who could watch him and his teammates every day. For example, in the pennant seasons of 1977-78, Dodger fans voted Reggie Smith (who walked more) as team MVP over Garvey for proto-sabermetric reasons.

    This raises an important point: If you listened to most Dodger games on the radio, you didn’t need sabermetric statistics to tell you that Reggie was contributing more offensively to winning than Steve was. As Yogi Berra said, you could observe a lot just by watching. Loyal fans who listened to 100 games or more per year could develop a gestalt sense of who was more valuable that would include obscure things like the fact that Reggie grounded into double plays less often and hit more sacrifice flies than Steve.

    Where the pre-Sabermetric statistics tended to fall down was in sportswriters voting for the league MVP award for players whom they didn’t follow on a daily basis. For that, they’d get obsessed with dumb statistics like RBIs.

    Anyway, I’ve gone on at length because I think it’s an important point that sabermetricians mostly didn’t come up with ways of measuring value that weren’t already visible to smart, attentive fans.

    • Chris G says:

      Thanks for the overview of the Dodgers. I really never saw anything of them until the playoffs – would get a glimpse of them then but not enough of a view to develop a well-informed opinion. I recall that infield being together for awhile but didn’t remember it was eight years. That’s amazing.

      >As Yogi Berra said, you could observe a lot just by watching.

      That’s what gets me about Sabermetrics. You can pick up a lot of things by watching – for example, when someone’s hot or when someone’s in a slump. Sabermetrics doesn’t give you insight into effects which fluctuate in the short-ish term. And those short-ish term fluctuations can make all the difference in the world – particularly come playoff time. Also, it strikes me that many Sabermetrics aficionados don’t understand confidence limits – particularly the limits inherent in drawing conclusions from small samples – or, for that matter, non-stationary data. For a case study of scouting vs Sabermetrics, check out, where the subject is Red Sox minor leaguers. Some people who post there provide incredibly good scouting reports. They’re keen observers. They answer address the question, “What does the data tell me?” in a qualitative but highly-informative manner. I can read their scouting reports then go to games and pick up on the subtle things they reported – sometimes notice changes over time. (Stuff I probably wouldn’t have picked up on my own if I hadn’t read the reports.) In contrast, the statistical analyses presented on the site… eh, I don’t generally find them as compelling. People look up and/or create numbers but don’t seem to spend as much time thinking about the limits of interpretation. Sabermetrics doesn’t address why someone can’t hit fastballs on the inside third of the plate and what adjustments they’re attempting to make. Good scouting can provide some insight into that.