Chris has an excellent track record on this, but I'm inclined to bet the under on his predictions for Morris, Smith and Walker.

On Morris, Chris over-estimated Morris' gain by 5% last year. I think positions on Morris are unusually entrenched, with a historically high proportion of "no" voters believing their peers are mistaken to vote for him. He will therefore see less of a momentum bounce than we would historically expect.

Smith went nowhere last year, sticking at 45%. Expecting him to jump 7% looks a shade too high.

Finally, while Walker should get a bit of a second year boost I can't see being to 27%. The Repoz tally is, I think, a little more stat friendly than the electorate as a whole. And I would expect any Walker surge to be predicated on stat-friendly writers working out he is deserving notwithstanding Coors. Not showing up so far so I struggle to see him getting into the high twenties.

I have a theory to explain this. Years ago, ex-THT writer Maury Brown posted the complete list of BBWAA members. About a fifth were based out of New York City. Less than a fifth of Repoz’s tally comes from there. I think there’s a definite Big Apple effect going on, and I expect Bernie Williams to top five percent with no problem.


This is commonly known in the industry as The Pepe Problem.

Minor point- Chris says in the article that Juan Gonzalez should poll over 5%, but he's not listed in the predictions.

Finally, while Walker should get a bit of a second year boost I can't see being to 27%. The Repoz tally is, I think, a little more stat friendly than the electorate as a whole. And I would expect any Walker surge to be predicated on stat-friendly writers working out he is deserving notwithstanding Coors. Not showing up so far so I struggle to see him getting into the high twenties.

Walker actually outperformed his Repoz tally by a large amount his first year. It could just have been a fluke, but I have a theory that his career .313 BA got him quite a few votes from the ignorant, the retired, and the lazy - folks who are the least likely to file a column. Further, Walker's likely to get a few votes from the "not a first balloter" crowd. I think 27% is more likely low than high, as an estimate.

I think he's right on with Morris, and I think you're right that Smith has plateaued.

High on Smith and Morris and, sadly, Raines and Trammell.

Bagwell and Larkin look right.

I think he is right on Morris and Smith. I think they are the sort of players who will be helped by the weak ballot.

I think Palmeiro is the player he is wrong on. I think he's going backwards this year.

I think he is right on Morris and Smith. I think they are the sort of players who will be helped by the weak ballot.

I think Palmeiro is the player he is wrong on. I think he's going backwards this year.

I think Smith might have a ceiling at or right below 50%.

I agree with you on Palmeiro and would add McGwire. McGwire's peak vote total (not %) was in his 1st year (and 2nd and 4th). For the most part, once guys decide to disqualify a guy for known steroid use, he stays disqualified regardless of the strength of the overall ballot.

I have a theory to explain this. Years ago, ex-THT writer Maury Brown posted the complete list of BBWAA members. About a fifth were based out of New York City. Less than a fifth of Repoz’s tally comes from there. I think there’s a definite Big Apple effect going on, and I expect Bernie Williams to top five percent with no problem.

I have a hard time believing there is much of a NYC bias. I am guessing there might have been before. Graig Nettles I think has a pretty similar profile as Bernie Williams, and he lasted on the ballot a few years. But then there's David Cone, who belongs in the same HoVG as Nettles, didn't pass a first ballot. Has the NY-bias been watered down more recently? With the internet, I wouldn't be surprised if there were a lot of writers based in NYC who simply don't follow the NY teams and have more of a national beat? Or maybe that's always been the case. I guess I don't automatically assume that "based in NYC" = "bias towards Yankees and Mets." Don Mattingly, Ron Guidry, Tommy John and Keith Hernandez are all decent enough candidates that it seems like a NY bias should buoy them closer to contention, but this doesn't seem to be the case.

I have a hard time believing there is much of a NYC bias. I am guessing there might have been before. Graig Nettles I think has a pretty similar profile as Bernie Williams, and he lasted on the ballot a few years. But then there's David Cone, who belongs in the same HoVG as Nettles, didn't pass a first ballot. Has the NY-bias been watered down more recently? With the internet, I wouldn't be surprised if there were a lot of writers based in NYC who simply don't follow the NY teams and have more of a national beat? Or maybe that's always been the case. I guess I don't automatically assume that "based in NYC" = "bias towards Yankees and Mets." Don Mattingly, Ron Guidry, Tommy John and Keith Hernandez are all decent enough candidates that it seems like a NY bias should buoy them closer to contention, but this doesn't seem to be the case.

I don't think the suggestion is that the NYC bias improves a candidate's chances of induction/staying on the ballot (I personally think it works in both directions. The pro-Yankee/Met bias of the New York guys is offset by the anti-NY sentiment of the rest of the country). Only that the bias presents itself in the Repoz count.

I think you're misunderstanding. Chris isn't saying that the voters have a New York bias.

Bernie Williams actually deserves serious consideration, and has the sort of resume (high batting average, ringzzz) that usually produces serious consideration. The question is why, if Williams has a resume that should draw at least 10-20% of the vote, Williams isn't polling higher in the Repoz tally.

Chris is saying that the Repoz vote tally is underrating Williams' expected HoF vote because a large percentage of the non-publishing voters are New York area writers. A similar effect has been seen with Mattingly, who usually doubles his Repoz tally in his actual voting.

As such, Chris is arguing that Williams should end up with a voting tally that is reasonable for his resume - and by implication, that there isn't a New York bias overall.

Aha, gotcha.

Chris, did Repoz actually give you a breakdown by region (or at least by NYC) as you reference in TFA? (sorry, it takes fewer keystrokes to say "article" that way--when you're not explaining it, that is...).

Here is the data from last year--official vote first, with the final Repoz percentages tacked on after the % sign:

Roberto Alomar 523 90.0% 93.4
Bert Blyleven 463 79.7% 79.7
Barry Larkin 361 62.1% 49.2
Jack Morris 311 53.5% 47.1
Lee Smith 263 45.3% 31.9
Jeff Bagwell 242 41.7% 39.1
Tim Raines 218 37.5% 40.6
Edgar Martinez 191 32.9% 26.8
Alan Trammell 141 24.3% 19.6
Larry Walker 118 20.3% 12.3
Mark McGwire 115 19.8% 16.6
Fred McGriff 104 17.9% 13.8
Dave Parker 89 15.3% 4.3
Don Mattingly 79 13.6% 5.1
Dale Murphy 73 12.6% 5.8
Rafael Palmeiro 64 11.0% 8.7
Juan Gonzalez 30 5.2% 2.2

There is a lot of variation here, much of it seemingly irrelevant to a NYC bias issue. Almost all the percentages were low in '11, some much more so than others. Only Raines and Alomar were high.

The big surprise in the Repoz numbers this year is the strong showing for Raines, but we have to remember that we only have about 15% of the vote and there is no "exit polling" avaiable here. Did we have a crush of ballots in the last week? With 81 right now, it seems we're down ~60 from what we had last year.

The big surprise in the Repoz numbers this year is the strong showing for Raines, but we have to remember that we only have about 15% of the vote and there is no "exit polling" avaiable here. Did we have a crush of ballots in the last week? With 81 right now, it seems we're down ~60 from what we had last year.

One thing that stands out in my memory about Raines is not just that he typically overperforms in the Repoz total than in the actual, he also steadily declines in the Repoz count as more precincts begin to come in. A year ago, for instance, he was above 60 percent on the Repoz count on December 29, but slipped to the aforementioned 40 percent.

I hope that trend doesn't continue, but I wouldn't be surprised if it did.

Did we have a crush of ballots in the last week? With 81 right now, it seems we're down ~60 from what we had last year.

I think Repoz answered this in his general ballot-gathering thread, but I think he said that MLB.com and ESPN released a tally of all of their voters a day or two before the vote was released last year, which I think was a total of about 30 or so ballots.

Minor point- Chris says in the article that Juan Gonzalez should poll over 5%, but he's not listed in the predictions.

CRAP!! Yeah, that's bad. DAMMIT. I'd peg him around 7%. Of course, since I think the existing prediction of 5.18 is more likely too high than too low, that means I really think that 5.25 is too high. I'd probably knock several guys down by a point or two to adjust for it. But too late - those are the article's predictions and it's out that.

One thing that stands out in my memory about Raines is not just that he typically overperforms in the Repoz total than in the actual, he also steadily declines in the Repoz count as more precincts begin to come in. A year ago, for instance, he was above 60 percent on the Repoz count on December 29, but slipped to the aforementioned 40 percent.

I hope that trend doesn't continue, but I wouldn't be surprised if it did.

This is what normally happens with Raines. (Actually, in 2010 his votes went up a tad as the predictions went on, but he was always high).

I didn't really pay too much attention to the repoz count this year, in part because the newbie class is so weak I just assumed everyone goes up, and in part because there's a full week between when I wrote this and when the BBWAA results come out. The closer you get to announcement day, the larger repoz's sample size (obviously) and thus the better it is.

Right now, Raines is at 61.7%. He won't get that high.

I assume that Repoz's sample (biased thought it likely is) overlaps greatly from one year to the next. When that's the case, estimates of change are much more precise. Or, in this case and more directly, somebody doing better in the Repoz count from one year to the next is likely to be due to voters adding him to their ballots.

That's why I'm pretty sure Walker will have a nice bump this year. His Repoz total is currently about what his 2011 overall percentage was but he is running well ahead of his Repoz percentage (although it's been going down). All told, in last year's Repoz count, he got only about 20-21 votes out of the 150+ ballots Repoz had; he's up around 20 votes already with (hopefully) another 70+ ballots to go.

Now if the Repoz sample actually shifts a lot more from year-to-year than I think ...

I'll take the over on Larkin and Trammell. I'll take the under on Smith, Mac and Palmeiro. And under on Mattingly I guess.

Minor nit on names/ballot. You also "lose" the small number of votes that went to <5%ers. I suppose in general you assume these "wasted" votes are pretty constant from one year to the next but (1) last year's included holdover Baines at nearly 5%; (2) this year's new crop is not just weak at the top, it's a pretty horrible collection throughout. Franco, Brown and Tino used up nearly 9% last year, doesn't look like anybody's gonna come close to that this year even though Salmon or Javy or Radke you'd think would score a couple percent.

But I'm not sure if that results in more votes for the backlog or even fewer names on the average ballot.

Anyway Dag is predicting 5.2 names per ballot down from about 6. Might go lower than that but ... about 200 percentage points left last year's ballot so even 5 names per ballot gives you 100 percentage points to hand out. Almost everybody has to go up ... or we have to see about 4.5 name per ballot.

What would've really impressed me would have been if Chris had posted his predictions for 2013---and wound up within 10 percentage points on his predictions for the Joy Boys.

I assume that Repoz's sample (biased thought it likely is) overlaps greatly from one year to the next. When that's the case, estimates of change are much more precise. Or, in this case and more directly, somebody doing better in the Repoz count from one year to the next is likely to be due to voters adding him to their ballots.

That's why I'm pretty sure Walker will have a nice bump this year. His Repoz total is currently about what his 2011 overall percentage was but he is running well ahead of his Repoz percentage (although it's been going down). All told, in last year's Repoz count, he got only about 20-21 votes out of the 150+ ballots Repoz had; he's up around 20 votes already with (hopefully) another 70+ ballots to go.

Let's see. I do a bit of repoz tracking and here's what I can tell you.

Last year, Walker was at 15% through 60 ballots. When it was all done, Walker scored at 13.8% among the full 138 ballots in the repoz count. Meanwhile, the full BBWAA gave Walker 20.3%.

One thing to keep in mind: last year, on the very eve of the ballot ESPN posted the ballots of their 18 voters & MLB.com did likewise for their 13 voters. Those 31 guys gave Walker a total one vote. Let's see if there's any movement from those guys this year.

What would've really impressed me would have been if Chris had posted his predictions for 2013---and wound up within 10 percentage points on his predictions for the Joy Boys.

I'm not looking forward to next year. It will be impossible to predict. Just impossible.

I'm not looking forward to next year. It will be impossible to predict. Just impossible.

I completely agree with that honest admission of cluelessness, which I'll see and double. It's like trying to predict a presidential election based on a poll in January. What I'm going to enjoy is to see how many people here miss by double digit percentage points and then blame the writers instead of themselves.

I assume that Repoz's sample (biased thought it likely is) overlaps greatly from one year to the next. When that's the case, estimates of change are much more precise. Or, in this case and more directly, somebody doing better in the Repoz count from one year to the next is likely to be due to voters adding him to their ballots.

Even if the sample of voters that make up Repoz' numbers is relatively stable, I wouldn't be too quick to assume that trends within that sample necessarily mirror those in the full pool of voters. Bias might manifest itself not just in different levels of support, but also in different changes in support.

Even if the sample of voters that make up Repoz' numbers is relatively stable, I wouldn't be too quick to assume that trends within that sample necessarily mirror those in the full pool of voters. Bias might manifest itself not just in different levels of support, but also in different changes in support.

It's possible but there's no obvious reason why non-Repoz voters would drop in Walker support in the same year that Repoz voters jump in Walker support.

Dag's depressing warning aside, the most likely outcome is "if Walker up in Repoz, Walker also up in non-Repoz, Walker gets big jump". The 2nd most likely is "Walker up in Repoz, steady in non-Repoz, Walker gets a jump." "Walker up in Repoz but down by as much or more in non-Repoz" seems unlikeliest of all. That last seems especially unlikely given the lack of good new candidates -- a 2011 Walker voter really has no reason not to tick the box in 2012.*

I'm more concerned that Repoz voters aren't as stable a group as I expect ... or that the late non-Walker surge Dag mentioned is coming.

In short I'm just wishing Repoz would get off his lazy butt :-) and also track gained/lost votes in his common sample. :-)

* It's always possible. They may not have had a particularly good reason to vote for him the first time ("hey, his BA is 319") and have rethought it since then. They may have thought they were voting for Harry the Hat Walker. They might now look at Bernie Williams and think "Walker's no better than that and Bernie sure isn't an HoFer". But we'd expect to see at least some of that in the Repoz count.