Next Week’s Cooperstown Results Todayby Chris Jaffe
January 05, 2009
Last year here at THT, I wrote a column predicting how the BBWAA vote would go based on 10 guidelines. I was pleased to find out my approach actually worked pretty well (though certainly not perfectly). So I thought I'd take another whack at it this year. Since Hall will make its announcement on Monday, Jan. 12, this is the time to do it.
Before getting into my guidelines for predicting the vote, let's see how I did last year in my inaugural Cooperstown guesses. For context, I'll include my predictions and the actual BBWAA results with the work of Keith Law and Darren "Repoz" Viola of Baseball Think Factory.
Neither made predictions, but they did surveys. For the last few years, Repoz has collected all ballots posted online and tabulated their results. Law took it a step further, privately contacting individual BBWAA members who don't publicly print their results to gain a larger sample size.
Here are our estimated/predicted percentages, along with the BBWAA's actual results:
Player BBWAA Law Repoz Me Goose Gossage 86 90 87 83 Jim Rice 72 68 71 74 Andre Dawson 66 66 65 62 Bert Blyleven 62 66 67 56 Lee Smith 43 37 34 35 Jack Morris 43 48 47 42 Tommy John 29 18 20 26 Tim Raines 24 35 31 45 Mark McGwire 24 24 26 32 Alan Trammell 18 24 23 17 Dave Concepcion 16 13 11 17 Don Mattingley 16 5 6 13 Dave Parker 15 9 8 15 Dale Murphy 14 11 10 12 Harold Baines 5 3 4 3
Average margin of error for all three of us:
Law 5.10% Repoz 4.70% Me 4.30%
It's a narrow victory, but I'll take it.
Actually, it's amazing I came out ahead at all when you realize I made what was by far the single worst prediction of the bunch. I was off by 21 percent on Tim Raines. Neither Law nor Repoz were off by more than 11 percent with anyone.
Raines was also the only first-timer among the 15 players who broke 5 percent. It's not a coincidence that I was so wildly off on the newbie because my guidelines for forecasting the BBWAA vote primarily deal with the backlog. If you look only at the predictions for people who returned to the BBWAA ballot last year, the margin of error shifts in my favor. Law and Repoze both scored at a 4.6 percent average error, while I had a typical error of 3.1 percent.
That is a small lead, but then again the competition didn't leave me much room for improvement. Under the circumstances, it went as well as I could reasonably hope. Of course, that just puts pressure on me to duplicate the success this year (dammit). May as well explain the system then.
Since it's pointless to completely regurgitate what I wrote last year, you can read last year's column, which gives the math justifying most of the guidelines. Below are the most important parts, as the guidelines pertain to understanding the 2009 ballot.
Guideline No. 1: Consistency
Unless you have a reason to think otherwise, assume the candidate will do as well as he did the year before.
Guideline No. 2: Strength of ballot
This is the most important for determining how the backlog will do. If a bunch of really strong candidates first appear on the ballot, the backlog will almost universally see their support dip. If it's a weak crop of first-timers, the backlog members should rise up.
This year's HoF vote contains one great candidate in Rickey Henderson and no other really serious threats. Over the last 20 years, the BBWAA averages approximately 1.55 newbie votes per ballot. It will be lower than that this year, indicating a good year for the backlog.
Guideline No. 3: Comparable candidates
The second and third guidelines are very similar; the former looks at the ballot on a macro-level while this one looks at the micro-version.
Specific candidates can be helped or hurt by arriving and departing similar candidates. If a newbie who is similar to but distinctly superior to someone on the backlog, the backlogger's vote will suffer. Alternately, a departing similar candidate can boost someone in the backlog. It's common sense, but it really rings true when you look at the results.
This year, Rickey Henderson's arrival means a dark year for Tim Raines. Since Goose Gossage gained entry last year and Dave Concepcion ran out of time, one should expect possible rises for Lee Smith and Alan Trammell.
Guideline No. 4: The "Over the top!" surge
This is one guideline that needs significant retooling from last year. In 2008, I simply noted that anyone getting really close to the summit will see a rather considerable uptick in his votes. This is especially remarkable because they should have the hardest time getting a noteworthy uptick, because there is a smaller number of voters left to win the support of.
This rule works, but I badly underestimated it last year. Dawson and Blyleven each were among my five most miscued predictions, and I guessed low on both of them. I knew it should be applied to Goosage and to a lesser extent Rice, but Dawson and Blyleven also had some of the largest rises in the entire backlog.
In response to my error in prognosticating Dawson, and Blyleven, I further investigated where this surge begins. Short version: 50 percent is the breakpoint. If you can get over that, the logs start rolling. Below that, upward movement is far less certain. Once over half the BBWAA decrees someone worthy, the rest of the body looks for reasons to be talked into voting for someone. The closer to 75 percent one is the larger the surge, but it begins at 50 percent, far lower than I would've guessed.
This year, Rice, Dawson and Blyleven should all benefit from this surge to varying extents.
Guideline No. 5:
Relief pitcher wackiness Primordial conversations
Here is another guideline that needs retooling.
One basic assumption I have is that BBWAA voting can be predicting in part because of the Hall's institutional history. Once upon a time a sizable chunk of the writers could legitimately make the claim that Jimmie Foxx wasn't worthy of the Hall because not enough men had been elected to make clear what is the cut off for enshrinement.
It's still quite hazy, but not totally indiscernible. With over 200 inductees, including over 100 by the BBWAA, standards have formed. That's why the voting can often fall into patterns—though some writers might be big hall or small hall guys, the actual Hall itself puts pressure on them to put those who measure up to Cooperstown's historic standards on their ballot.
But there are no historic standards for relievers, making those candidates a bit different. Their situation is becoming clearer due to a recent spate of reliever inductions (Dennis Eckersley, Bruce Sutter, and Gossage in this millennium), but it's still more vague compared to other positions.
Steroids are another area still being sorted out. The BBWAA's real steroid wars won't begin for a few years still, but McGwire is the provisional shot.
Last year, I was off by 8 percent on both Smith and McGwire; only my Raines prediction was worse. I guessed that Smith would be hurt by Hoffman's eclipsing his save record combined with Gossage's surge. I was wrong. I thought McGwire would benefit last year from some writers holding him off for one year. That didn't happen.
This year I assume McGwire will roughly stay the same, but am guessing on Smith. Gossage is off the ballot, so Smith should go up. Then again, Smith is now third in career saves, and that stat has always drove his candidacy, and he's never attained the 50 percent surge-marker. I assume he'll remain stuck between 40 and 50 percent, but I have less faith in that prediction than any other backlogger.
Guideline No. 6: Last year on the ballot
Players in their 15th and final year under consideration from the BBWAA typically receive a modest bump. It's not huge, but can be worth an extra 3 percent on average. They only go down if the ballot becomes much more difficult.
Rice and John make their last stand in 2009. Both should experience a rise.
Guideline No. 7: Candidates per ballot
The BBWAA has averaged fewer names per ballot over the last 20 years than at any point previously. It first ducked under seven per ballot in 1987, and has never gone back above. Typically, it's around six per ballot, bouncing around 5.5 and 6.5 names. In 2008, the BBWAA set a new all-time low with 5.35 names per voter.
What does this mean for 2009? Well, let's start with last year's score of 5.35. That included Goosage's 85.8 percent vote and Concepcion's 16.2 percent. Their departures leave surviving backlog at 4.33 names/ballot.
Henderson should be on virtually all the ballots, but my hunch is that the next best candidates will struggle to reach 5 percent. Total, the newbies will come in at 1.1 votes per ballot, by my guesstimate. That's 5.43 names per ballot.
Ah, but that's only if the backlog stays the same, which is unlikely. For perspective, here are the five lowest votes/ballot in BBWAA history, and how much support went to newbies on average in those years:
Year Vote Newbies 2008 5.35 0.26 1998 5.41 0.65 1997 5.60 0.30 2000 5.64 0.68 2006 5.64 0.32
It's really difficult to score historically low if you have a player like Rickey Henderson on the ballot. That doesn't mean you can't have a really low overall vote total despite strong first-balloters. (The seventh-lowest ballot ever occurred in 1993, when Reggie Jackson and Phil Niekro debuted.) It's trickier, however.
Based on history, my hunch is that the 2009 ballot will have at least 5.60 names. That's only a rough guess, though.
Guideline No. 8: Repoz (and formerly Keith Law)
For newbies, these guidelines don't work, so your best bet is to find those you are tallying the BBWAA ballots. Last year both Repoz and Law did. Law has opted not to do it this year, leaving the field clear for Repoz.
Time out: if my system for tracking newbies depends on following Repoz and/or Law, then how come my Raines prediction was so wrong last year?
Three things went wrong: 1) the writers Repoz and Law tracked were more supportive of Raines than the BBWAA as a whole, 2) I have to write these columns a few days in advance due to real world stuff and couldn't use the full sample size of Repoz/Law ballots. (A lot of writers release their ballots on the days immediately prior to the election.) Last year, the early batch of ballots Repoz and Law tracked were especially pro-Raines. 3) Most importantly, I just plain screwed up. My guess was actually a bit higher than anything they ever posted.
For 2009, Repoz's survey has shown surprisingly little support for Mark Grace. I didn't expect huge support for him, but it looks like the former first baseman might not clear the five percent barrier that determines who returns to the ballot.
Guideline No. 9: Beware five percent
This is the flipside of the 50 percent logrolling. Guys who spend too long with too little support eventually fall under five percent and thus off the ballot. No one is looking for reasons to put them on, and just enough of their supporters fall away. Harold Baines may survive 2009, but he won't last 15 years.
Guideline No. 10: Guidelines ain't laws
None of this is perfect. This is an art, not a science.
FINALLY: my predictions
First, 2009 is a harder year to predict that 2008. The overall ballot was so weak last year I could simultaneously predict all backloggers except Harold Baines would experience a rise in their support and the overall ballot would set a record for fewest names per voter—and be right on both counts.
This year, most of the backlog should rise, but it's not clear to me that they all will. The first-timer will receive far more support than last year's crowd, though still a below-average total. Though my exact predictions may not be as accurate as last year's, I feel confident I have the general picture. See for yourself—here are my guesses alongside last year's BBWAA results.
Player 2009 2008 Rickey! 99 XX Jim Rice 85 72 Andre Dawson 71 66 Bert Blyleven 69 62 Lee Smith 47 43 Jack Morris 45 43 Tommy John 33 29 Mark McGwire 23 24 Alan Trammell 21 18 Tim Raines 19 24 Don Mattingley 17 16 Dave Parker 16 15 Dale Murphy 15 14 Harold Baines 5 5 Mark Grace 4 XX Others 4 XX
That's 5.73 names per ballot, which might be a little high. Or not—who knows?
Though no one has ever received 100 percent, eight of the 10 highest percentages of all time have come since 1992. Only three men have ever topped 98.5 percent, but all are recent ones: Tom Seaver, Nolan Ryan, and Cal Ripken. Henderson should join them. After all, there is only one Rickey.
Rice has all arrows pointing up for him: he is top of the backlog, it's his last year, and the overall ballot strength is weak. Going purely by the guidelines listed above, Rice should do better than I'm predicting. Heck, just last year Gossage jumped from 71 percent to 86 percent without having the final ballot working in his favor. However, I think there's a larger anti-Rice sentiment than normally exists for a bubble candidate. Regardless, Rice will going in and it won't be close.
Dawson and Blyleven should continue to surge. Both have an outside shot at winning election next Monday, but I don't quite see it. In the last half-century, the BBWAA elected three players in only four elections. None of those votes (1972, 1984, 1991, and 1999) are good comps for 2009.
On top of that, it's very difficult for two backloggers to win a plaque in the same year, so either is unlikely to join Rice.
In the last 30 years, there have been only four times more than one backlogger made it in. In 2000, Carlton Fisk and Tony Perez got over 75%. Then again, Fisk was only a backlogger because he debuted on the toughest ballot in the last generation, alongside Nolan Ryan, George Brett, and Robin Yount in 1999. Perez also might have gone in during 1999; he was at 68 percent in 1998, which was his third straight year over 60 percent. Both were in better standing in 2000 than either Dawson or Blyleven in 2009.
In 1991, Gaylord Perry and Fergie Jenkins both made it. Again, it hadn't been for the strength of immediately preceding ballots, neither would've still been still hoping for election in 1991. Both scored higher in 1990 than either Dawson or Blyleven last year (Perry had 72 percent, Jenkins 67 percent). Plus, in 1990, Jim Palmer won election, freeing up more votes for starting pitchers (after all, comparable candidates are one of the guidelines).
The other two times came in 1987 and 1984. Those years also had, not so coincidentally, the two weakest showings by first-year eligibles in the last 30 years. (In 1987, all newbies combined to appear on 0.08 ballots; in 1984 in was 0.07.)
I think Blyleven's surge will be bigger than that of Dawson because all the other top candidates are outfielders, but both should come up short. Rice's induction in 2009 should pave the way for Dawson's plaque in 2010. Blyleven will either enter alongside Dawson or shortly thereafter.
I have little faith in my Lee Smith prediction for reasons mentioned above. I could also be wrong about Morris. That 40 percent range is a place from which candidates sometimes have trouble escaping, but then again last year was Morris's best showing ever.
Raines will experience the largest drop. He's the second-greatest leadoff hitter of all time, but #1 is on the ballot this year. Raines backers shouldn't lose heart. This should be a one-year blurb for him. Fun fact: in history of BBWAA voting, only seven players had a better debut than Raines's 24.3 percent last year without eventually gaining a plaque for himself. (Yes, really.) That seven includes four guys still on the ballot, one of whom I'm predicting gains entry this year.
I think (or is it hope?) Trammell goes up ever so slightly now that he's the only shortstop on the ballot. The rest are just marking their time with no real hope for election from the BBWAA.
Bookmark this article so you can mock my forecast when the Hall of Fame announces the results on January 12.
References and Resources
For this project, I went to the Hall of Fame's website and put the results of every Hall of Fame election into an Excel file, then played around with it. (By "playing around" I mean I added a flock of info from b-ref about player's last year, position, etc. That is the basis of my guidelines.)
As far as I know, this is the first time Repoz's full, completed survey from last year has been put online. He put the full survey several times when gathering votes, but after the election he didn't post the results all the way down to Baines (unless I missed it). I e-mailed him last year and got his results for this column.
I have to admit, the fun fact about Raines at the end is a tad misleading. Due to structural problems (discussed here every single player who ever received 10% of the vote in their ballot debut from 1936 until the mid-1950s has gained entry. Thus saying "in the history of BBWAA voting" while true, gives a perhaps overblown point of view. Then again, that's still only seven guys (including four still on the ballot - Rice, Dawson, Lee Smith, and Mattingley) from the last half-century of voting. That's still impressive, and hence why I'm willing to make the statement above, even with this disclaimer below. (The other three, if you're curious, are Steve Garvey, Luis Tiant, and Maury Wills.
If anyone wants to see a different approach, Sean Lahman has predicted how this year's vote will go.
History instructor by day, statnerd by night, Chris Jaffe leads one of the most exciting double lives imaginable; with the exception of every other double life possible to imagine. Despite his lack of comic-book-hero-worthiness, Chris enjoys farting around with this stuff. His new book, Evaluating Baseball's Managers is available for order. Chris welcomes responses to his articles via e-mail. Oh, and now he's on twitter.