Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Hall of Merit > Discussion
Hall of Merit
— A Look at Baseball's All-Time Best

Tuesday, May 03, 2005

Term Limits or Perpetual Eligibility for Candidates?

DanG sent this to some of us about a week and a half ago:

I thought I would run this idea by youse guys.

I feel compelled to take another stab at limiting candidacies.  Perpetual eligibility is great in theory, but in practice it has its problems.

As we near the halfway mark, the cautions that Marc (Sunnyday2) and I brought forth in the formation of the project are bearing forth.  Only four candidates received 33% of possible points, a new record low.  69 players received votes, a new record high.  These records will soon be broken and the fragmentary balloting will worsen.  Players will be elected that will make Pike and Terry’s support appear overwhelming, especially as we get into the elect-3 years.

This is not really a huge problem yet, but I have a suggestion to subtly relieve the pressure in this direction.  I call it 70-and-Out.  For players last playing after 1891 (or eligible for our first election), they get 70 years on the ballot.  For players retiring before 1892, they are eligible until 75 years after their last game.  This would mean that Jim Creighton is already ineligible for election, but nobody has voted for him for quite a while anyway.

But.  Along with this I suggest a 10% rule.  A candidate will only be retired after he is named on less than 10% of the ballots (less than 6 votes, as things currently run) in two consecutive elections.  So Friends Of Charley Jones, all ten of them, have no worries.  In time, this will reduce the votes being thrown away on pet candidates (there were 63 votes wasted on candidates receiving less then 6 votes in 1949).  By the time we reach the present in 2007 we will have retired all the bad candidates retiring before 1932.

It’s not too late to enact this rule.  The first “significant” candidates to expire would be Harry Wright and Candy Cummings, after the 1952 election.  Unless they get a sudden revival of support.  Which leads to the chief benefit of this idea.

It’s a perfect way to highlight the old time candidates.  I would provide lists of players due to get the blade.  Discussion would proceed from this; Do we really want to let Harry Wright disappear forever?  When we get to 1959 and Tommy Bond and Levi Meyerle are on the bubble, this may very well revive their candidacies.  Or not; as it stands now they are almost entirely forgotten.

I expect the usual cries of, “But these aren’t my 15 best candidates, you’re forcing me to vote for someone I don’t want,” or similar drivel.  Look, we already know that the difference between your #6 and #16 player (or #26) is negligible, and becoming even less as time goes on.  So we’re not dredging up refuse by doing this.  Focusing attention on The Candidates and away from the fantasies is a good thing.  You’re saying 140 weeks is not enough time to study and stump for a player?  “Yes, we know you love Fred Dunlap/Tommy Bond/Tony Mullane, but it’s time to go now.”  Why can’t we say that?

To me, the point of this project is reaching consensus through informed discussion.  To those who stubbornly reject the values of the majority, we can help them.  Yes, there is virtue in open discussion of all candidates, and we are doing this.  But there is vice in anarchy and chaos, and this is a small attempt to ensure order because I think it ultimately leads to a better result.

 

What do you think?  Should we put this idea before the group?

 

John (You Can Call Me Grandma) Murphy Posted: May 03, 2005 at 09:42 PM | 158 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

Page 2 of 2 pages  < 1 2
   101. John (You Can Call Me Grandma) Murphy Posted: May 05, 2005 at 07:46 PM (#1314717)
My only problem with the run-off ballot is again administrative. Keeping track of having to vote a run-off ballot in an off-week wouldn't be so bad unless we find there are lots of run-off ballots. Then it can get confusing (voting every week). Also, there may be some voter dropout between the real ballot and the run-off, which could skew results.

How often is it going to happen? I know we wouldn't be having run-offs for every election. It would only happen if a player doesn't meet 50%.
   102. Daryn Posted: May 05, 2005 at 08:02 PM (#1314870)
I'm not in favour of run-offs, but you could do it automatically without reballoting. In the original ballot, everybody could be requested to order all the returning top 10 and supported newbies that fall outside of their top 15. This is done on most ballots right now in any event. You could use that information to create the run-off.
   103. Howie Menckel Posted: May 05, 2005 at 08:04 PM (#1314879)
DanG,
When I talk of Welch 'slowly sliding down the ballot,' I don't mean at all that I tend to prefer the newer candidates. It's just that HIS better contemporaries already have been elected, while the new guys' versions have not.
There are many pretty decent candidates nowadays who never get on my ballot; I like Welch a lot, and you have to be better than Welch to get on the ballot. That's hard to do, in my mind.
Over a few decades, a half-dozen or more unelected guys pull ahead of Welch. But most of them do not.

P.S. I'd like to see a little swing toward bailing out on hopeless causes, but probably it's best not to force such a thing.
I just wonder if some guy who was 6th on a ballot 25 years ago sensibly can still be placed 8th or 10th now. A ton of subsequent candidates, and only 1 or 2 unelecteds in all those years are better than that mid-ballot guy? That's a head-scratcher for me.
It's different if a guy becomes upgraded in a voter's mind. But if not, I'd just wonder if it was just reflex that kept getting the guy on a ballot, a la concerns that maybe "PHOMers" might get picked as a rationalization by that voter.
To be fair, I haven't really researched whether my gut instinct even has 'merit.'

Ultimately I think this is a good discussion which may in itself solve the 'problem' - if one even exists.
   104. DavidFoss Posted: May 05, 2005 at 08:13 PM (#1314937)
When I talk of Welch 'slowly sliding down the ballot,' I don't mean at all that I tend to prefer the newer candidates.

Yes, ranking 6th of a backlog of 60 is basically the equivalent of ranking 10th of a backlog of 100.

The #15 guy on my ballot is a lot stronger than he was when I first joined in the mid-1920s.
   105. John (You Can Call Me Grandma) Murphy Posted: May 05, 2005 at 08:23 PM (#1314967)
I'm not in favour of run-offs, but you could do it automatically without reballoting. In the original ballot, everybody could be requested to order all the returning top 10 and supported newbies that fall outside of their top 15. This is done on most ballots right now in any event. You could use that information to create the run-off.

Hmmm...that might work, Daryn. One could even say it has...uh...value!

What, you thought I was going to say "merit?" :-)
   106. Mark Shirk (jsch) Posted: May 05, 2005 at 11:07 PM (#1315305)
If we must do a run off I would rather do it Daryn's way than a yes or no vote. If a makes less than 50% of the ballots are 50% or 60% or whatever really going to vote yes on him? And if they vote no who takes his slot in the HOM? I think that a run off woudl be unfair to guys like Jennings, Beckley, Welch or any other candidate that has a relatively small but very loyal following.
   107. PhillyBooster Posted: May 06, 2005 at 02:25 PM (#1316336)
I agree with jschmeagol. A "run off" for less-than-50% candidates essentially create a "majority veto" than would eviscerate the whole point of awarding extra points to top candidates.

Consider, as an example, the 1940 election where Lip Pike finished second with about half of the voters and 40% of the votes. Assume there was one or two more voters, so that a Pike runoff was necessary. Pike had 496 points, but appeared on fewer ballots than Joe Sewell (483), Eppa Rixey (463 poins), Clark Griffith (411), George Sisler (406), Jack Beckley (402), and Tommy Leach (352). (Hughie Jennings was in there too, with one fewer ballot appearance).

Is a run-off really effectuating the purpose of our ballot structure if it is conceivable that Pike would finish "in the money" six times, but nonetheless lose all six run-offs to players with up to 140 fewer points?

In a future scenario where players are routinely receiving fewer than 50 percent of the votes, we might as well eliminate the points altogether and just count how many ballots a player appears on. That might end up a better proxy of who would get elected.
   108. PhillyBooster Posted: May 06, 2005 at 02:43 PM (#1316377)
Let me say also that I appreciate the goal that Dan is trying to accomplish -- as I understand it -- which is to increase the opportunities for those with both of two candidates off-ballot to register their preferences between to two. It is, assumedly, not merely an aesthetic preference against a long row of "6 points" at the bottom of the election results, but an assumption that absent that long row, much information would be imparted up-ballot.

Assumedly, for example, Hughie Jennings could be elected narrowly over George Sisler, with both appearing only on "peak voter" candidates, despite the fact that career voters (who are voting for neither) would all prefer Sisler.

Similarly, Jake Beckley could be elected narrowly over Tommy Leach by "career voters", despite the fact that peak voters (who are voting for neither) would all prefer Leach.

My view, however, is that the viewpoint of a voter who votes A 6th and B 8th should be weighed much more heavily than the view of a voter who prefers A 18th and B 16th.

Extending the ballot to 20 votes or having a run-off election would give the two voters equal weight, while limiting the candidates such that the second voter's ballot has candidate B pushed onto it actually gives the second voter more of a voice.

To the extent that Dan's suggestions "work" (and I have still not seen a good explanation for why they would), they would work by changing exactly those structures ("elect me" bonus and "on ballot" bonus) that we debated and decided on at the beginning.
   109. DavidFoss Posted: May 06, 2005 at 03:12 PM (#1316446)
Let me say also that I appreciate the goal that Dan is trying to accomplish -- as I understand it -- which is to increase the opportunities for those with both of two candidates off-ballot to register their preferences between to two.

Well said. This is the main bonus that would come from the type of change proposed. Otherwise, nothing is broken here.
   110. DanG Posted: May 06, 2005 at 03:59 PM (#1316556)
This shows the value of discussion. The ramifications of the proposal raised by PB had not quite occurred to me. But, having said that, I don't see that as having any significant effect.

At this point in the project, in most elections there is about the same difference in quality between #6 and #8 as there is between #16 and #18. So, the viewpoint that the former pair should be weighed more heavily seems unsupportable.

One thing to remember about those long-debated structures is that there were compromises made in the interest of simplicity and practicality. Ideally, voters would be free to assign different points to express their intensity of support, beyond simply ranking. As John pointed out, this becomes a headache come tallying time. There is also more opportunity for strategic voting.
   111. Chris Cobb Posted: May 06, 2005 at 04:16 PM (#1316594)
My view, however, is that the viewpoint of a voter who votes A 6th and B 8th should be weighed much more heavily than the view of a voter who prefers A 18th and B 16th.

Well, that wouldn't change by extending the ballot to 20. The ballot is designed with the idea that distinctions at the top of the ballot will tend to be more meaningful than distinctions at the bottom or below the ballot. This principle is correct, both on the basis of talent distribution among players and on the basis of voter care with the top of the ballot.

But with the pool of eligibles as deep as it has become and will remain, is it the case that all distinctions below #15 in rank should continue to be considered meaningless for the purposes of election?

That is not a rhetorical question. As I recall, jimd did a study back when we were thinking about expanding the ballot that indicated that in few elections were the down-ballot votes actually decisive: most of the time a shorter ballot would have elected the same players.

It may therefore be the case that with the basic weighting system in place, expanding the ballot, even with the deeper candidate pool, would make no difference in the outcome of elections at all: those lower distinctions would continue to be meaningless, even if they were counted.

Would that continue to hold true in the fragmentary future we are imagining for our election results?

If it would not hold true, is it still a better policy to ignore those lower-tier distinctions and hold that only distinctions among the top 15 players _really matter_, even if the lower-down votes, if counted would be affecting outcomes?

A run-off, obviously, depending upon how it would be weighted, might have considerable effect on outcomes because it would give weight to _all voters'_ distinctions among a top set of candidates. I'd like to see some projections of how run-offs could have affected past close elections before making any decision myself on the fairness of adding a run-off election if a candidate would be elected without appearing on a plurality or some other agreed-upon minimum of ballots or receiving an agreed-upon minimum of points.
   112. Chris Cobb Posted: May 06, 2005 at 04:51 PM (#1316662)
Quick results of a study of past elections:

Out of 52 elections for the HoM, there have been 18 in which one of the electees either received fewer elect-me votes than a non-electee or appeared on fewer ballots than a non-electee.

So that's 1 election in 3. In virtually all of those cases, one of a pair of electees led in both categories, so we're talking about 1 electee in 6 winning by virtue of elect-me votes or by virtue of broad support but not both.

In those cases, 8 elections went in favor of the candidate with broader support and 10 elections went in favor of the candidate with more elect-me support.

These results strongly suggests that the ballot design is well-balanced. No candidate has been elected who appeared on fewer than 50% of the ballots, and no candidate has been elected without a fair number of elect-me votes. The most extreme outliers are these. For ballot-appearances, Lip Pike was elected while appearing on only 50% of ballots in 1940, appearing on 75% as many ballots as the top non-electee in b-app. In 1939, Red Faber was elected with half as many elect-me votes as the top vote-getters in that category (Pike and Jennings), and only 6 overall; in 1905, Hardy Richardson was elected with 36% as many elect-me votes as the top vote-getter in that category (Ezra Sutton with 14), and only 5 overall.

My guess is that these outcomes would likely become more extreme in a situation in which it is routine for few players to appear on a majority of ballots. In those cases, a candidate who did gather broad support, even if not much was at the top of the ballot would be easily elected (Faber), and a candidate who gathered a significant group of top support would also be readily elected (Pike).

I don't see that the situation would favor one configuration of support over the other. HOWEVER, the fact that an elected candidate might be _far_ behind a non-elected candidate in one or the other registers of support seems likely to lead to less confidence in the outcomes (certainly that has been the case in the all the elections mentioned).

If this sort of outcome were to begin to happen on a regular basis, I think that it could be harmful to the HoM. If we can predict that the development of the ballot pool will lead to this sort of outcome regularly, I think it would be good to take proactive steps against it.

In the history of the HoM so far, these sorts of outcomes have been quite rare. Even in the cases where electees did not lead in both ballot-appearances and elect-me votes, the amount by which they trailed other candidates in one measure or the other has typically been small.

I mself don't foresee these sorts of outcomes becoming common before 1970, which is as far ahead as I feel confident making any predictions, especially because our deliberative processes are so strong. However, the candidates whose merit lands around the in-out line will only increase, so sooner or later it is hard to believe that the problem will not arise.
   113. David C. Jones Posted: May 06, 2005 at 05:06 PM (#1316685)
I think Chris is right in his analysis, and I actually think the runoff election is a good idea. To take two candidates who are perpetually in the top ten returnees but whom I rarely vote for, I have very clear opinions on whether Earl Averill or Eppa Rixey is more "meritorious." Ditto for Cool Papa Bell and Jake Beckley. Since I rank the top 100 players for my ballot, it would not be difficult at all for me to supply the information needed for an instanteous runoff, and I imagine the same could be said for most if not all other voters.

All we would need to do is decide what the threshold would be that would trigger a runoff.
   114. Daryn Posted: May 06, 2005 at 06:44 PM (#1316936)
I think we have to take Chris' analysis one step further. Faber beat Pike and Richardson beat Sutton in questionable elections, but in the end they all went in. I'm not sure there are many or any instances where a run-off or extended ballot would actually change the complement of the HoM. Even as future fragmentation comes upon us, I'm not convinced these changes would change the complement of the HoM and I'm further not convinced that if they do change the complement, it will be a change for the better. Lastly, determining whether any change of the sort contemplated is a change for the better is an extremely difficult exercise to undertake objectively. No matter what rules we impose, there is a certain arbitrary nature to this type of process. I have trouble seeing how one arbitrary disitinction is going to be objectively better than another, particularly where (as Dan said) an immense amount of thought went into the original structure.
   115. David C. Jones Posted: May 06, 2005 at 06:52 PM (#1316962)
I think that if a threshold were to be established for a runoff, it should be lower than what has happened in any past election. It should only be done in anticipation that a possible future election might be so fragmented that a candidate who appeared on, say, 35 percent of the ballots was elected. If that never comes to pass, then great, the runoff mechanism wouldn't need to be used.

To set the standard higher than what has happened in any previous election would implicitly suggest that the HOM believed that election's results were a mistake. This is the type of thing we should avoid. We shouldn't implement a new measure AFTER a problem occurs, because that immediately shades the result and indicates that the HOM believes one of its inductees doesn't really belong there, or didn't deserve his election. This is the problem that the HOF has experienced throughout its history. One can think whatever one wants to about the election of Mazeroski by the Veterans Committee, but the fact that the HOF took steps immediately after his selection to change the process suggests that the institution itself was not happy with the result. Sometimes I wonder how that makes Mazeroski feel; as if he's only a quasi-member of the HOF and most of his colleagues probably don't think he belongs there.

So what I'm saying is, if we are going to make a change (like constructing a run-off mechanism) we should do so well before any potential problem will arise, so it will not seem as if we are responding to the results of any particular election.
   116. Dr. Chaleeko Posted: May 06, 2005 at 07:04 PM (#1316995)
I'm generally in the "don't fix it" camp. And I agree with Daryn that the longer-range view of controversial elections shows that basically these guys were likely to be elected one year or another anyway.

I'm against limiting eligibility, I don't think expanding the ballot is any great shakes, though I do think it's possible that a run-off could be useful in cases of extreme pluralism.

Ultimately I'm unconvinced by the argument that things will get 'worse' in terms of pet candidates and fragmentation. The reasons for fragmentation are not altogether clear, however, I don't think they are to do with ballot structure at all. Here's what I think they have to do with:
-confidence in analysis of NgLs
-confidence in analysis of 1860s guys
-peak v. career
-pitching v. hitting
-relative importance of position
-importance of defense v. hitting

In other words, fragmentation occurs because each individual voter has differing opinions on each of these issues. For this reason, no system will reduce fragmentation without resorting to eliminating voter preferences.

If one believed that fragmentation was a problem, then the most effective way to reduce fragmentation would not be to expand ballots or reduce eligiblity, it would be to SHORTEN ballots to ten or fewer.
   117. DanG Posted: May 06, 2005 at 07:09 PM (#1317005)
So what I'm saying is, if we are going to make a change...we should do so well before any potential problem will arise, so it will not seem as if we are responding to the results of any particular election.

Yes, a good point to keep in mind.

And let's not be assuming we've never made a mistake. Where this becomes most likely is in an election immediately preceding a flood of great newbies. Someone gets left on the doorstep in these elections, getting knocked back down the line in following elections, sometimes never to rise again.

So we have to always question: Was this a mistake?

-in the early 30's, Rube Foster's mad dash past VanHaltren
-in 1939, newbie Faber grabbing the most votes
-in 1940, Pike edging Sewell
-in 1942, newbie Terry topping Rixey

These are the first examples that come to mind.
   118. Chris Cobb Posted: May 06, 2005 at 07:23 PM (#1317050)
I'm generally in the "don't fix it" camp. And I agree with Daryn that the longer-range view of controversial elections shows that basically these guys were likely to be elected one year or another anyway.

DanG rightly points out that the "well, they all got in anyway" argument, which works for all close elections prior to 1930, doesn't apply so neatly after 1930.

I think a fuller examination of the 1939, 40, and 42 elections especially would be instructive. Extrapolation from the trends there could help us decide if there is a threshold at which we would want to see the balloting changed by lengthening the ballot or incorporating a run-off trigger (which, if we go that way, should have both ballot-appearance and elect-me-vote thresholds).

There was a bit of debriefing immediately after those elections, but our last really serious retrospective analysis of balloting took place before those elections occurred.
   119. Daryn Posted: May 06, 2005 at 07:29 PM (#1317078)
in the early 30's, Rube Foster's mad dash past Van Haltren

This reminds me that we all come into this debate with our own biases. Perhaps one of the reasons I am so satiated by the current system is that the Foster in, Van Haltren out result might be the only one out there that would be permanently changed if a different system had been in place. I had Foster #1 on my ballot for several years.
   120. Michael Bass Posted: May 06, 2005 at 08:09 PM (#1317284)
-in the early 30's, Rube Foster's mad dash past VanHaltren

While we have no idea what would have happened to Foster, we do know what has happened to Van Haltren, so one would be hard pressed to declare this election a mistake given that GVH has essentially fallen off the election queue, including now being behind a number of candidates who were also on that 1932 ballot.
   121. Chris Cobb Posted: May 06, 2005 at 08:22 PM (#1317332)
I think the Foster/Van Haltren election has little to show about the voting system, actually. It was a close election, and the difference between Foster's and Van Haltren's vote profiles was not large (my notes are at home, so I can't give figures now).

No systemic reform will eliminate close elections :-) .

That close election has historical significance because that was the first close election in which, as of our present moment, the runner-up still has not been elected. But it doesn't raise issues about the voting system.

The elections of the late 1930s and early 1940s are instructive because of the strong divergence between the vote profiles of the candidates elected, which amounted to historically low totals of ballot appearances or elect-me votes for the winners. These historical lows raise the question of whether the voting system begins to lose validity when the pool of candidates becomes large enough that a candidate appearing on a large majority of ballots becomes the exception, rather than the rule.

The ballot design for the HoM is _amazingly good_, so it may well be that, if it does begin to lose some validity under fragmented voting conditions, we just have to accept that we are doing the best we can. I certainly don't have a run-off system in mind that I have confidence in, and no system can be perfect.

Nevertheless, it is clear that as the votes become more heavily split, the results for any given candidate reflect a smaller portion of the voters' considered judgments about the candidates in question. Would it be beneficial, would it be possible and fair to get more of the voters' judgments into the voting in these circumstances?
   122. Daryn Posted: May 06, 2005 at 08:31 PM (#1317371)
I think it is a mistake to think that, for example, a run-off vote would "get more of the voters' judgements into the voting" in a meaningful way. Expanded the ballot might. But I wouldn't want to see an election decided because 3 people who had Jennings and Beckley both out of their top 40 each preferred Jennings.
   123. Chris Cobb Posted: May 06, 2005 at 09:11 PM (#1317553)
I think it is a mistake to think that, for example, a run-off vote would "get more of the voters' judgements into the voting" in a meaningful way. Expanded the ballot might. But I wouldn't want to see an election decided because 3 people who had Jennings and Beckley both out of their top 40 each preferred Jennings.

A response pertaining to preferences:

Why not? Help me to see why this is bad.

A response relating to the principles of balloting:

This objection to a run-off system is conflating a close election with an election that triggers a run-off. These would necessarily be the same thing. If we did establish a run-off system, there could still be close elections that would not go to the run-off, because the winning candidate was above a certain miminal threshold of support, but that could be decided in Beckley's favor because somebody had him at 15 and Jennings at 16.

_Any_ election that's very close will be decided on some small, probably arbitrary difference in the voting.

The question is, if a candidate is elected without, say, a single "elect-me" vote or, say, while appearing on only 33% of the ballots, should there be a mechanism that reconsiders that outcome on the basis of a somewhat different way of registering voters' judgments in voting, given that in the one case _no one_ thinks the elected candidate is the most deserving candidate, and in the other case, the majority of the electorate's opinion of the candidate has been registered only in the most basic of ways.
   124. Michael Bass Posted: May 06, 2005 at 09:27 PM (#1317645)
Why not? Help me to see why this is bad.

Tossing out a couple quick reasons...

Why should someone who sees Jennings as 80 and Beckley as 81 have the same level vote in a runoff as someone who sees Beckley 1 and Jennings 80? Or Beckley 18 and Jennings 150th for that matter?

More importantly, and the group has argued about this before, I would much rather elect the people the group is most enthusiastic about. If we're honest, at this point most of us are in one way or another enthusiastic about everyone on our ballot, not just the elect me slots. So whether someone squeeks in because they made 15 #1s and were no where else to be seen or was everyone's #7, this is fine with me. What is less fine with me is the decision-making process becoming the lesser of two evils.

To give a personal example, if Rixey and Beckley (two candidates I don't support at all) are separated by 10 points for an election slot, I don't want my vote cluttering up the decision making between karl and PhillyBooster, who actually like one or both of these guys.
   125. Mark Shirk (jsch) Posted: May 06, 2005 at 09:34 PM (#1317693)
If we are going to do a runoff, and I am not really in favor of this, then there should be a threshold not just of the percentage of voters who voted for that palyer but also how close the voting was. If say Hughie Jennings gets 700 points and gets on 35% of the ballots and Joe Sewell gest 595 points while being on 45% of the ballots should we do a runoff? In this case one player beat the other by 105 points, a difference big enougorballots to overcome. However, the 'results' above show that Sewell gets more widespread support and would most likely win a run off.

And of course the players could be switched, just picking two players, one that I know gets a lot of his support by a smaller number of enthusiastic voters.
   126. DavidFoss Posted: May 06, 2005 at 10:16 PM (#1317933)
Why should someone who sees Jennings as 80 and Beckley as 81 have the same level vote in a runoff as someone who sees Beckley 1 and Jennings 80? Or Beckley 18 and Jennings 150th for that matter?

Yes, well *IF* we did a run-off, I would hope that we do it will the same ballot-counting. It wouldn't be a giant Beckley vs. Sisler YEA or NAY, it would be a re-count or re-vote with only the top 15 or so eligible for points. You could still have a guy ranking Beckley 1 and Sisler 15 and another ranking Sisler 14 and Beckley 15. Those two would NOT cancel each other out.

Of course, in that scenario the #3 guy might swoop in and take a victory. :-) There are indeed several ways we could do this. :-)

At any rate, its good that we are getting this discussion out in the open early while nothing is at all broken. If I recall, the issue with the 1939-40 & 1942 elections was one more of surprise than of sour grapes. (FWIW -- I was anti-Faber, pro-Pike and ambivalent about Terry). There was a notion in the de-briefing that followed those elections that people were not expecting certain ballot effects to occur.
   127. jimd Posted: May 06, 2005 at 11:15 PM (#1318115)
Anti-Runoff soapbox speech...

Why do we need a "runoff" election? We have a voting system. It says that whoever gets the most points, wins! Final answer. It also was purposely designed to say that some preferences are more important than others, that strong opinions count more than weak opinions. Adding a "runoff" reverses that and says that those preferences are now to be treated as equal, that a preference between #51 and #53 is the same as one between #1 and #53. In a runoff, I bet the "wishy-washy" candidate will usually beat the "polarizing" candidate.

Also, what is so special about a winner not being in the top "15" of the majority of the voters that it should trigger a runoff? That "special" number could be the top 5, or the top 25, and it would be just as irrelevant.

People are casting this potential situation as a "lack of consensus". It really indicates the equivalence of the candidates, that they are so close together in "merit" that the differences in our evaluation systems are what yield the different ballot orders. The originally agreed upon balloting system then produces a decision, but we now determine that there was not enough "input" to that decision (based on an arbitrary criterion), reweight the original preferences, and ask for a preference from those that had indicated no official preference (because they don't think either candidate is strongly qualified). This new result may "validate" the original decision (which doesn't need validation), or "overturn" it (causing much controversy?), but there will be no more "consensus" after the runoff than there was before.

A runoff election is not a good idea.
   128. Dr. Chaleeko Posted: May 07, 2005 at 02:47 AM (#1318952)
jimd,

i think i agree now with you about runoffs. keep it as it is, no runoff, no expansion, no terminated eligibility.
   129. Chris Cobb Posted: May 07, 2005 at 04:15 AM (#1319060)
jimd makes a good case against the run-off, but I remain unconvinced by a few points, mostly because the case does not consider that it is some congruence between the ballot design and voter judgments that underpins the validity of election outcomes.

Adding a "runoff" reverses that and says that those preferences are now to be treated as equal, that a preference between #51 and #53 is the same as one between #1 and #53. In a runoff, I bet the "wishy-washy" candidate will usually beat the "polarizing" candidate.

This is a problem, and a simple, unconsidered run-off system would certainly fall into it, but it is one that could be addressed in the design of the run-off voting, just as it has been addressed in the design of the current voting system, which as an examination of outcomes so far shows, tips in the direction of high-elect me support about as often as it tips in the direction of broad ballot-appearance. If no run-off system can be devised that continues to reflect this balance, then obviously a run-off system would be a poor addition to the voting system and should not be implemented.

Also, what is so special about a winner not being in the top "15" of the majority of the voters that it should trigger a runoff? That "special" number could be the top 5, or the top 25, and it would be just as irrelevant.

There's nothing inherent in 15 or 5 or 25, but there is a actual number of candidates that each voter considers "strong" or "serious" candidates, and the number of candidates that the ballot length enables a voter to support, need to be relatively close for the ballot to be a sensitive register of the voter's judgments about inclusion in the HoM. If, because of permanent changes in the size of the pool of good candidates, the voter's own set of "serious" candidates is no longer remotely congruent to the size of the ballot, it seems to me that the validity of the election could be diminished. The current system is good not only because it is weighted in such as way as to be balanced between top-ballot support and broad support, but because 15 is a good number of candidates to include. It has so far proven to be a little long in weak years and a little short in strong years to satisfy everyone's tastes. It is not a magic number, but its value is not arbitrary, either.

we now determine that there was not enough "input" to that decision (based on an arbitrary criterion), reweight the original preferences, and ask for a preference from those that had indicated no official preference (because they don't think either candidate is strongly qualified).

Given that many people are now saying that they see a number of candidates whom they cannot fit onto their ballots as well-qualified players for whom they would like to vote--or who are even in their personal Halls of Merit!--I think there is clear evidence that "no official preference" for a player can no longer be assumed to mean "not strongly qualified."

Here's the thing that would hurt: a player I support for election but cannot vote for because of the ballot design rules loses an election to a player I do not support at all and also did not vote for. What exactly did my casting a ballot accomplish, in practical terms, other than the satisfaction it gave me to cast it?

If we come to the point of electing players who appear on 33% of the ballots cast, I believe that this sort of of situation will occur for lots of people, and I wonder if that will be healthy for the election process.

Perhaps this problem of fragmentation can only be addressed by reasoned argument: if so, I still feel pretty good about our chances of avoiding a degree of fragmentation that will progressively disenfranchise voters who are not at the shaky center of opinion.

People are casting this potential situation as a "lack of consensus". It really indicates the equivalence of the candidates, that they are so close together in "merit" that the differences in our evaluation systems are what yield the different ballot orders.

As with Daryn's objection, I am concerned that this conflates close elections with unreliable ones. I agree that this is what close elections have meant so far. Perhaps that is what they will always mean, but as the pool grows, the degree of difference in ballot construction created by differences in evaluation systems is likely also to grow. When that degree of difference reaches the point of filling most of a 15-position ballot, voters will be faced with a choice between following the rules and voting for their 15 top picks or getting a say on the candidates who actually stand a chance of election, about whom they probably have considered judgments on who is preferable. When a particular kind of voting system is excluded from registering meaningfully within the system, the reliability of the election suffers.

We are not near that point; this is a theoretical argument. These concerns might be refuted by better theoretical arguments, particularly ones with numbers :-) . But at present I remain unconvinced that ballot fragmentation would not undermine the reliability of our elections, were it to reach a certain magnitude.

Ballot expansion is a mild remedy for the effects of fragmentation, perhaps too mild. A run-off ballot would be a strong remedy, perhaps too strong.
   130. Dr. Chaleeko Posted: May 09, 2005 at 02:49 AM (#1322580)
I'm growing increasingly convinced that ballot contraction is the best solution to the problem of fragmentation (if it is a problem). It forces a sharper decision-making process, and it reduces the likely number of one-off, tail-end stragglers that seem to be vexxing everyone.

Expansion seems likely to lead to more 1 vote stragglers, while candidate dormancy has, from the discussion above, enough issues that it doesn't yet seem tenable. Ballot contraction allows for the sharper focus Dan G seems interested in, and it also gets rid of the stragglers. I'm not in favor of making any change right now, but I think ballot contraction makes the most sense.
   131. Dr. Chaleeko Posted: May 09, 2005 at 03:49 AM (#1322658)
Before getting too much further down this road, I wanted to take a really serious look at who’s coming up for election, when, and what the backlog will look like in the near future. To start with, let’s consider any candidate receiving 100 points plus Jose Mendez (96 points in 1950) as “the backlog.”
Cronin
Suttles
Beckwith
Averill
Rixey
Ferrell
Mackey
Jennings
Sisler
Griffith
Beckley
Papa
Sewell
Van
Duffy
Welch
Roush
Waddell
Redding
Grimes
Leach
Childs
Lundy
Browning
Jones
Schang
Cravath
Bresnahan
Méndez

Looking at the elections through 1957, we can reasonably predict that the following players will be elected:
1951
Foxx and Cronin

1952
Gibson and Ott

1953
Greenberg and Dickey

1954
Vaughan and Wells, Suttles, Hack, or Beckwith

1955
Leonard and Wells, Suttles, Hack, or Beckwith

1956
Appling and Wells, Suttles, Beckwith, Hack, or Ray Brown

1957
Dimaggio and Wells, Suttles, Beckwith, Hack, or Ray Brown

That brings us with the potentially divisive 1958 backlog election. Before getting into that one, however, I want to quickly review who will enter the backlog from 1951 through 1957, listing anyone who I think could get even one vote

1951
Bob Johnson, Clift

1952
Brewer, Hughes, Bankhead, Radcliffe

1953
Ruffing, Hack, Herman, Lombardi, Mort Cooper, Bill Byrd

1954
Medwick, Walters, Hilton Smith

1955
Dixie Walker, Ray Brown, Gene Benson

1956
Joe Gordon, Tommy Henrich

1957
Boudreau, Doerr, Moses, Keller, Buck O’Neill, Max Manning

1958
Willard Brown, Dizzy Trout, Dom Dimaggio

Two of these guys plus our earlier backlog will be elected in 1958. My money is on whichever one of the Wells, Suttles, Beckwith, Hack, or Brown isn’t elected by 1957 plus one among Averill, Rixey, Hilton Smith, and Red Ruffing.

So then moving ahead from 1959 through 1975 the situation looks, to me, like this (electees at the top, backlog additions beneath in parens):


1959
Paige
Mize
(Dandridge)
(Elliott)
(Wild Bill Wright)

1960
Dandridge
Day or Newhouser
(Newhouser)
(Jethroe)

1961
Newhouser or Day
BACKLOG
(V Stephens)
(Kiner)

1962
Feller
J Robinson
(Irvin)
(Rizzuto)
(A Rosen)
(H Thompson)
(Kinder)

1963
Campanella
Irvin
(Kell)
(Artie Wilson)

1964
Reese
BACKLOG
(Lemon)
(Trucks)
(Maglie)
(Serrell)

1965
Slaughter
BACKLOG
(Vernon)
(Doby)

1966
Williams
BACKLOG
(Newcombe)
(Easter)
(Hodges)

1967
BACKLOG
BACKLOG

1968
BACKLOG
BACKLOG
(Ashburn)

1969
Musial
Berra
(E Wynn)

1970
Snider
E Wynn
(Minoso)
(Pierce)

1971
Spahn
BACKLOG
(Fox)

1972
Roberts
BACKLOG
BACKLOG
(Koufax)

1973
Ford
BACKLOG

1974
Mantle
Matthews
(Colavito)
(E Howard)

1975
BACKLOG
BACKLOG
(Drysdale)
(Boyer)

So by my count, as many as 14 backloggers could be elected between 1959 and 1975, or about 1 a year. It is possible, therefore, that the upward pressure on the ballot that causes much of the fragmentation we currently fear will be alleviated first by electing several backloggers through 1958, then again through the normal election processes that will likely clear the backlog in the late 60s and early 70s.

As candidates like Hack, Suttles, Beckwith, Averill, and Rixey move up the rankings and eventually are elected, voters will be forced to redistribute their support elsewhere within the backlog, at which point our currently successful deliberative processes could help sharpen the electorate’s focus adequately.

That's my interpretation of it, I don't see things as direly as others, but I am likely just as wrong as they are right. Perhaps moreso.
   132. David C. Jones Posted: May 09, 2005 at 05:48 AM (#1322740)
Dr. Chaleeko,

Thanks for the analysis. Overall I'm on the fence as to whether specific changes are needed at this point. I have a visceral reaction against ballot-contraction, however, as I feel there are well more than 15 candidates who deserve election to the HOM, so picking the top 15 is already sharpening my decision-making process quite well, I think.

I think at this point I would be in favor of putting together a run-off mechanism, if a candidate fails to reach a certain threshold but is nonetheless in an "elect-me" spot.

Two points that I have already made on this but that I think bear reiterating:

1. The threshold should be below the lowest support received by an elected candidate to date. We don't want to implement a rule change that would challenge the legitimacy of someone already elected.

2. The time to set up such a mechanism is right now, or sometime in the near future. You don't want to wait until a problem actually occurs: that, again, would cast doubt on the legitimacy of past elections.
   133. Joey Numbaz (Scruff) Posted: May 09, 2005 at 11:55 AM (#1322863)
Wow, my head is spinning. Great arguments on all sides.

At this point, the only things I would support would be:

1) Expanding the ballot to 20. I think this will help with having more elected candidates on the ballot of more voters. Although I realize these spots probably won't have much impact on an election, they do allow a voter to express a preference for these players, which is nice. I doubt anyone will add an 'unlisted' player there, if you have a guy 16-20, odds are someone else has him in their top 15. Maybe I'm wrong there.

2) A runoff, in the case of an extremely fragmented election. I'm not sure what the thresholds would be, but I would think if we did a runoff, we'd allow the top 10 in. This is not so a 10th place guy could sneak off to a win. What this would do is allow for the range of preferences we are used to. Giving the 'elect me spots' a bonus in the runoff would be important too.

I'm not sold on either of these, but I do see their merit, especially expanding the ballot to 20. I never dreamed our backlog would be this deep already.

I also agree that any changes (if made) should be below previously established threshholds.

I guess the question is, are we talking in circles now? Going over the same things over and over? Should we maybe set up a straw poll or something? Perhaps through the yahoo group (so as to avoid someone having to count yay or nay votes).

Emphasis on straw poll - it would just be to see if any consensus is forming, nothing binding yet.

Thoughts?
   134. PhillyBooster Posted: May 09, 2005 at 03:44 PM (#1323142)
First point: The voting structure was set up with two bonuses -- the "Elect Me" bonus and the "On Ballot" bonus. Having a run-off essentially removes the first of these bonuses. Expanding the ballot essentially removes the second of these bonuses. Maybe these are changes the group wishes to make, but I think that either one changes the inherent structure of the balloting, and will not be merely cosmetic.

I think a straw poll is a good idea to decide whether we are spinning our wheels or not. Like the HoM election is general, I'm supportive of the view of the electorate, even when I disagree. I will be voting "No" to both proposals, however.

----

My question regarding a potential run-off is whether there would be any limits to how often a guy could lose a run-off. Could he finish "in the money" (top 2 or 3) annually, and lose every run-off? My concern here is the "Lip Pike leads the ballot for the twentieth year in a row, and once again loses the run-off, this time to Keith Hernandez" scenario.

Assuming an occassional fragmented ballot, the run-off "exception" could make sense, but I'm of the view to it'll either be happening "hardly ever" or "all the time" (depending on whether we set the ballot limit at 33% or 50% or somewhere in between. To the extent it happens hardly ever, why bother? Those'll just be close elections and we'll all live with the results. To the extent that it happens all the time, then our elections essentially become the regular season in the NHL. Who cares? Just sit back and wait for the playoffs. To the extent this happens, the exception eats the rule and the voting scheme we have in place becomes meaningless.

----

Regarding expanding the ballot to 20 places, can I pre-emptively suggest that -- if that change is implemented -- we set up a run-off anytime the ballot is changed by the new down-ballot votes?

Unlike the mindset that agrees with these changes, my mindset is much more supportive of a guy who gets in with 15 first place votes, rather than a guy who gets in with 45 seventeenth place votes.

Looking at the collective view of the 16-20 place candidate in 1898 and 1950, I see that Mickey Welch makes both lists (17th in 1898, 18th in 1950). The other four from 1898 are C. Jones, who has slipped to 27th but still has some support, and Dunlap, McCormick, and Orr, who have all fallen by the wayside.

For those who favor expanding the ballots, is Welch worth more points now than he was in 1898 for the same placement? If anything, to the extent that Welch is now with players who are closer in value to him than they were in 1898, does it make sense to give them different points? Why not give them all 3-points for being in the 16th to 20th group.

Some are saying that there just not that much of a gap between 15 and 16 to warrant the 6-point gap. In my view, there's even less gap between 16 and 21, so why should 16-20 split 15 points while 21 gets none at all?
   135. Chris Cobb Posted: May 09, 2005 at 04:05 PM (#1323199)
I guess the question is, are we talking in circles now?

Well, some of us may be talking past each other, but I don't think we're talking in circles yet.

A straw poll at this point wouldn't hurt, but I think we need some more concrete modelings of ballot expansion and run-offs. I myself don't know whether I'd vote in favor of either one right now. I'm arguing the pro- side, because (1) I think ballot fragmentation is a potential problem and (2) I haven't found the arguments against the changes sufficiently cogent, but I don't think I'd want to vote to favor implementation of a change yet. I'd want to vote for continued discussion, I guess . . .
   136. Chris Cobb Posted: May 09, 2005 at 04:21 PM (#1323234)
Phillybooster wrote:

First point: The voting structure was set up with two bonuses -- the "Elect Me" bonus and the "On Ballot" bonus. Having a run-off essentially removes the first of these bonuses. Expanding the ballot essentially removes the second of these bonuses. Maybe these are changes the group wishes to make, but I think that either one changes the inherent structure of the balloting, and will not be merely cosmetic.

Not necessarily. It depends how each is structured. A run-off could be weighted to retain an "elect-me" bonus element. Expanding the ballot could involve shifting the numbers for the entire ballot upward so that there was still an "on ballot" bonus. However, I think those arguing for the expanded ballot tend to be making the case that the distinctions between players from 15-20 are so small, yet all in that range seem like good HoM candidates, that the "on ballot" bonus no longer makes sense given the nature of the pool.

My question regarding a potential run-off is whether there would be any limits to how often a guy could lose a run-off. Could he finish "in the money" (top 2 or 3) annually, and lose every run-off? My concern here is the "Lip Pike leads the ballot for the twentieth year in a row, and once again loses the run-off, this time to Keith Hernandez" scenario.

In a well-balanced run-off situation, this ought not to occur. There seems to be a presumption here that a run-off will favor the candidate with broad support. Consider, however, that if an "elect-me" bonus is retained in the run-off, a candidate with a strong but narrower base of support could _benefit_ in a run-off, because that candidate would possibly have more support to tap into in the just-off-the-ballot range than a candidate who already has his cards on the table, as it were.

For those who favor expanding the ballots, is Welch worth more points now than he was in 1898 for the same placement?.

While I think there are problems with comparing the voting resutls to individual rankings, I would say that, yes, a 17th-place finish in 1950 is worth more than a 17th-place finish in 1898, though that was the strongest of the early ballots. That is shown by the way in which the nearby candidates from 1898 have declined, while Welch has not.

If anything, to the extent that Welch is now with players who are closer in value to him than they were in 1898, does it make sense to give them different points? Why not give them all 3-points for being in the 16th to 20th group.

Andrew Siegel suggested something along these lines, 3 points for all 16-20 finishers, 1 point for 21-25. That makes a lot of sense, in that it fits a widely-held view of these candidates: worthy of some support, but not readily distinguishable at present.
   137. TomH Posted: May 09, 2005 at 06:38 PM (#1323519)
Fascinating discussion, of which I have only skimmed less than half, and don't intend to make the time to do more. Good proposals, but not crucial diffs IMHO, so I will go with whatever the majority wishes.
   138. jimd Posted: May 09, 2005 at 08:08 PM (#1323738)
Some data for those looking for interesting prior elections:

Elections where losing candidates got more "votes to elect" than a winner:
1900 11/10 Wright/Ward
1904 11/10 Radbourn/Rusie
1905 7-8-13-14-9/5 Galvin-McPhee-Spalding-Sutton-Start/Richardson
1910 12/9 Start/Galvin
1918 8-7-12/6 Keeler-Kelley-Bennett/Flick
1919 11/7 Bennett/Kelley
1926 10-10/9 Wallace-Jackson/Magee
1927 11/10 Wallace/Hill
1932 7-8-7-8/6 VanHaltren-Pike-Beckley-Welch/Foster
1939 12-12-9-10/6-10 Pike-Jennings-VanHaltren-Welch/Faber-Carey
1942 7-7-9-9/6 Sewell-Beckwith-VanHaltren-Jennings/Terry

Elections where losing candidates were on more ballots than a winner:
1901 35-34-34/33 Glasscock-Radbourn-Richardson/Wright
1903 44-44-43-40-40-40-43-39/38 Glasscock-Radbourn-Richardson-
xxxx -Start-Galvin-Sutton-Thompson-Stovey/Anson
1904 43/42 Richardson/Rusie
1912 39-39/38 McPhee-Stovey/Start
1913 40/37 Stovey/McVey
1916 42/41 Kelley/Stovey
1920 45/44 Collins/Walsh
1921 41-39-40/37 McGinnity-Johnson-Wallace/Bennett
1931 41-41-41/40 Foster-VanHaltren-Griffith/Pearce
1932 41/40 Griffith/Foster
1940 35-35-29-30-27-28/26 Sewell-Rixey-Griffith-Sisler-Beckley-Leach/Pike

Candidates elected due to the "votes to elect" bonus (4 extra points):
1901 640-619/16-10 Wright-Glasscock
1913 751-747/16-13 McPhee-McVey
1916 667-665/9-7 Stovey-Kelley
1929 629-616/15-9 Thompson-Sheckard
1934 1102-1088/23-19 Speaker-Collins
1940 496-483/12-7 Pike-Sewell

Candidates elected due to the "on ballot" bonus (5 extra points):
1929 629-616/38-35 Thompson-Sheckard
1939 589-522/44-29 Faber-Pike

Note: eliminate either bonus and Sheckard edges Thompson in 1929.
Note: eliminate both bonuses and 1939/1940 stay the same.
(That is, Pike benefits from the "vote to elect" bonus, his opponent benefits from the "on ballot" bonus. Eliminate one bonus and the result may change; eliminate both and the result does not change.)

Candidates elected due to the combination of the two bonuses:
1906 758-749.5/14-12/42-41 Spalding-Sutton
   139. Joey Numbaz (Scruff) Posted: May 10, 2005 at 11:28 AM (#1325762)
"For those who favor expanding the ballots, is Welch worth more points now than he was in 1898 for the same placement? If anything, to the extent that Welch is now with players who are closer in value to him than they were in 1898, does it make sense to give them different points? Why not give them all 3-points for being in the 16th to 20th group.

Some are saying that there just not that much of a gap between 15 and 16 to warrant the 6-point gap. In my view, there's even less gap between 16 and 21, so why should 16-20 split 15 points while 21 gets none at all? "

One thing - if we expand the ballot - I would not want it to remove the on ballot bonus. I would bump the whole scale up 5 points, not give 16-20, 5-4-3-2-1. That's how I always envisioned it, but this post makes me think that others see it different . . .
   140. Joey Numbaz (Scruff) Posted: May 10, 2005 at 11:31 AM (#1325765)
"Andrew Siegel suggested something along these lines, 3 points for all 16-20 finishers, 1 point for 21-25. That makes a lot of sense, in that it fits a widely-held view of these candidates: worthy of some support, but not readily distinguishable at present."

I kind of really like that idea. Very interesting.
   141. Joey Numbaz (Scruff) Posted: May 10, 2005 at 11:32 AM (#1325767)
Actually, I'd like 4 and 2. 16-20 get 4 points and 21-25 get 2 points. So there is still an 'on-ballot' bonus . . . 6-4-2 is a little more scaled than 6-3-1 and a 15 is still worth 3x a 21st or 25th place vote, not 6x.
   142. andrew siegel Posted: May 10, 2005 at 02:08 PM (#1325889)
I think I'm with Joe. I initially suggested either 6-4-2 or 6-3-1. (Chris picked up on the 6-3-1.) My only reservation is that we might be imposing a lot of work on voters and ballot counters for point totals too small to have any effect on the results.
   143. Mark Shirk (jsch) Posted: May 10, 2005 at 08:58 PM (#1326743)
However, it result in a 16-25 section that wasn't ranked as meticulously because there is no difference betwee 16 and 20 and 21 and 25. I try and make every spot on my ballot and just off my ballot improtant and I playe with them a lot. In this system it just wouldn't matter. I think every spot should have a different value in order to make every ranking at least somewhat important.
   144. karlmagnus Posted: May 10, 2005 at 09:09 PM (#1326772)
16 -5, 17-4.5, 18-4....25-0.5 would leave a gap with 15 and scale nicely, ie half the slope between 16-25 as between 1-15. But I tell you, if you go to 25, Levi Meyerle will WALK AGAIN -- I think it fragments the ballot more, not less.
   145. jimd Posted: May 10, 2005 at 10:23 PM (#1326956)
I think it fragments the ballot more, not less.

Depends on the metric used to measure fragmentation. Number of candidates on the ballot results page? Going up, absolutely. Number of ballots on which leading candidates appear? Also going up. Which means that "fragmentation" is going down, at least to some.

There will be more "pet" candidates getting throw-away votes, and less chance of the dreaded "winner below 50%".
   146. jimd Posted: May 10, 2005 at 10:33 PM (#1326974)
Increasing the total points on the ballot will dilute the value of a #1/#2 vote, the "vote to elect". Currently, the #1 and #2 votes add to 47 points on a 203 point ballot (23.2%). Compare to #1 on an MVP ballot, which is 14 out of 59 (23.7%). Under the current 25-man ballot proposal, they should be boosted to 55 points on a 241 point ballot (22.8%), an addition of 4 points each.
   147. PhillyBooster Posted: May 18, 2005 at 07:09 PM (#1346096)
Okay. I think I figured out how we "shoulda" done it. It's several years too late, but I think it would have addressed everyone's issues.

Step 1: In Week One, elect a "Hall of Very Good." The election would be exactly like the current one, with no eligibility requirements and no "waiting period". Everyone would be eligible forever. The number of electees per year would be exactly twice the number of HoM electees (scaled back five years).

After five weeks of building up a Hall of Very Good, (so that it would have about 20 players), move to

Step 2: Hold the first HoM election. The election would be exactly like it is now, except there would be an additional eligibility requirement that the player already have been elected to the Hall of Very Good. In the alternating weeks, hold an election to replenish the Hall of Very Good.

By now, there would be 100 members in the HoM and another 100 in the HoVG. Everyone would remain free to vote for their pet candidates in the HoVG elections for eternity, but would be constrained to a slowly-increasing pool of consensus players in the HoM election.
   148. jimd Posted: May 18, 2005 at 07:32 PM (#1346220)
The catch: practically everybody that is being mentioned on the most recent ballots would already be in the Hall of the Very Good (or be in immediate danger of being elected ;-).
   149. Rob_Wood Posted: May 18, 2005 at 09:54 PM (#1346734)
Wow. I just saw this thread for the first time! Terrific discussion.

My two cents. I favor expanding the ballot to 20 or 25 spots with decreasing points to each spot. I am indifferent whether we keep an "on ballot" bonus or let the points trickle down to one for the last spot.
   150. karlmagnus Posted: May 19, 2005 at 12:23 AM (#1347020)
Be very interesting to repeat the exercise using a HOVG/HOM dual election method; it might produce oddly different results.
   151. PhillyBooster Posted: May 19, 2005 at 01:54 AM (#1347145)
The catch: practically everybody that is being mentioned on the most recent ballots would already be in the Hall of the Very Good (or be in immediate danger of being elected ;-).

Not sure that is true, especially down by the <10% crowd that Dan's plan would eliminate. (Lon Warneke? Donie Bush?) Also, it would limit the new players you could vote for. There would be a HoVG backlog, so that you wouldn't be able to vote for some more recent players until they had passed through the queue.

It would also minimize the "elect a guy who retires to years before seven guys who were better than him" problem, because we'd be considering the HoVG candidates five years ahead.
   152. jimd Posted: May 19, 2005 at 02:15 AM (#1347185)
Hall of the Very Good (inferred from HOM voting)
Note: Please, don't take this too seriously. Our ballot isn't deep enough to make this reliable. During the 1920's, for example, anybody who got on a HOM ballot made the HOVG. It's more to prove a point as to who just might have been elected to this hall.

1893 4 BARNES, WRIGHT, SUTTON, SPALDING
1894 4 Start, McVey, Pike, McCormick
1895 4 WHITE, WILLIAMSON, DUNLAP, Pearce
1896 4 HINES, RADBOURN, WELCH, Jones
1897 3 GORE, RICHARDSON, GALVIN
1898 2 O'ROURKE, KELLY
1899 2 CLARKSON, WARD
1900 2 Keefe, GLASSCOCK
1901 2 BROUTHERS, EWING
1902 2 CONNOR, ANSON
1903 3 RUSIE, Stovey, Thompson
1904 4 MCPHEE, Bennett, Browning, Tiernan
1905 2 Caruthers, Griffin
1906 2 HAMILTON, DUFFY
1907 2 JENNINGS, Childs
1908 2 DELAHANTY, RYAN
1909 2 Grant, VanHaltren
1910 2 NICHOLS, BURKETT
1911 4 GRIFFITH, McGraw, H.Wright, Mullane
1912 2 BECKLEY, CROSS
1913 2 KELLEY, J.COLLINS
1914 4 DAVIS, DAHLEN, McGinnity, F.Jones
1915 2 FLICK, KEELER
1916 4 YOUNG, CLARKE, Waddell, Joss
1917 2 Willis, Chance
1918 4 SHECKARD, S.White, Long, York
1919 2 WALSH, WALLACE
1920 4 G.JOHNSON, BRESNAHAN, Monroe, LEACH
1921 4 LAJOIE, MATHEWSON, BROWN, J.Whitney
1922 2 WAGNER, CRAWFORD
1923 4 Plank, R.Foster, S.King, Bond
1924 4 MAGEE, Meyerle, Evers, Buffinton
1925 4 JACKSON, CRAVATH, DOYLE, CICOTTE
1926 4 HILL, KONETCHY, Chapman, J.Williams
1927 4 BAKER, Bush, Vaughn, Tinker
1928 4 POLES, PETWAY, Beaumont, Nash
1929 4 PRATT, DONALDSON, Daubert, R.Thomas
1930 2 HOOPER, VEACH
1931 4 SANTOP, MENDEZ, Burns, MOORE
1932 4 JOHNSON, WHEAT, GROH, SHOCKER
1933 4 COBB, SPEAKER, COLLINS, LLOYD
1934 4 SJ.Williams, Torriente, Coveleski, Carey
1935 4 ALEXANDER, HEILMANN, SISLER, Taylor
1936 4 ROUSH, REDDING, SCHANG, Mays
1937 4 Cooper, Bancroft, Schalk, Fournier
1938 4 FABER, SEWELL, RIXEY, Maranville
1939 4 ROGAN, BECKWITH, GRIMES, RICE
1940 4 RUTH, HORNSBY, VANCE, TRAYNOR
1941 4 TERRY, H.Wilson, ARLETT, Luque
1942 4 CHARLESTON, COCHRANE, FRISCH, FOSTER
1943 4 GEHRIG, GOSLIN, FERRELL, Lundy
1944 4 Cuyler, LAZZERI, MANUSH, Hoyt
1945 4 STEARNES, SIMMONS, SUTTLES, AVERILL
1946 4 GROVE, HARTNETT, J.WILSON, Dean
1947 4 GEHRINGER, LYONS, BELL, GOMEZ
1948 4 HUBBELL, MACKEY, KLEIN, BRIDGES
1949 4 WANER, DIHIGO, CRONIN, Berger
1950 4 FOXX, Leever, Warneke, Myer

Note:
CAPS indicate first year on ballot.
Players are eligible for HOVG one year before HOM eligible.
1893-97: "Priming the pump"; building candidate backlog for 1st HOM election
1898-02: "Pipeline", supplying enough candidates for next HOM election
1903: "Transistion" 29:14 plus 3 more becomes 32:16
1904-onward: Electing double the following HOM election number.
   153. Brent Posted: May 28, 2005 at 03:02 AM (#1367584)
1952 ballot thread # 62:

I'm dropping my official support for some candidates as lost causes. When I am the only supporter two years in a row, that's it, until someone else brings them back. They will remain on my ballot, but deprecated as a "lost cause" (see Fred Dunlap).

While I think expanding the ballot to 20 or 25 names would help somewhat in obtaining consensus (it would be more likely that electees would appear on a larger share of the ballots), I think the greatest benefit would be to the voters. When we make the effort to vote, we like to feel that our votes our meaningful, so as the number of viable candidates has grown it becomes difficult to continue to support "lost causes" when we have strong preferences among viable candidates who are just off ballot. I think the elections are stengthened when voters continue to make the case for forgotten or less viable candidates. But when we reach the backlog again, the choice may be between someone we think is # 17 and someone we think is # 57, and we would much prefer that # 17 win. Consequently, there is the temptation to "reevaluate" one of our lost causes to allow # 17 to move up to where our vote counts.

I think quite a number of voters have said they would support expanding the ballot to 20 or 25 candidates, while others have expressed opposition. How will this question be resolved? Can we hold a vote to officially consider a change?
   154. Mark Shirk (jsch) Posted: May 28, 2005 at 02:07 PM (#1367913)
In the ballot thread dan b said that he thought the one flaw in our system is the mandatory continued support of lost causes. I disagree. I really like the fact taht everyonenot in the HOM is eligible every single ballot. We are to rank who we think are the 15 best. If you are the only one who thinks that say, Fred Dunlap is HOM material it should be a cause to re evaluate him. If he still passes the test, then I dont' think it is a weakness that voters should be allowed to vote for him. I think it is a strength.
   155. John (You Can Call Me Grandma) Murphy Posted: May 28, 2005 at 02:32 PM (#1367938)
If you are the only one who thinks that say, Fred Dunlap is HOM material it should be a cause to re evaluate him.

I personally did that myself because I thought I had too many 1870 guys on my ballot a while back, so I changed a part of my system which dealt with era and positional durablity (I feel now I was crediting too much in that department back then for the early NL stars). But I revised my system because I thought my analysis was flawed, not because I was worried about ballot fragmentaion.
   156. Mark Shirk (jsch) Posted: May 28, 2005 at 06:14 PM (#1368264)
I will also say that when I see a guy like Bond or Dunlap on a ballot I tend to think, "Wonder why so and so put him on his ballot? Maybe I should find out." It causes me to look at guys again. Though to date I ahve yet to vote for a candidate like that.
   157. Michael Bass Posted: May 28, 2005 at 08:10 PM (#1368414)
I'd also add that for newcomers, the "voting results" list is also often where they start when forming their consideration set (at least it was for me). The one vote for Bond or Dunlap (both of whom, BTW, I like, though I'm not currently voting for either) may make a big difference if a newcomer that might not otherwise have looked at them likes their results.
   158. Rob_Wood Posted: May 28, 2005 at 09:45 PM (#1368518)
Here's an idea that I hope the group will consider to be an improvement to the current voting system.

Issue: Perpetual eligibility leads to voting fragmentation.

Solution approach: Collect more information from each voter, especially for the players at the top of the group consensus. (Note that while I support the idea of expanding the ballot to 25, I fear that this will not significantly eradicate voting fragmentation.)

Proposal: Maintain perpetual eligibility. Each voter continues to vote for his top 15 players on each ballot with the current vote points for each slot. In addition, each voter is given 10 additional vote-points to allocate among two or more players with the proviso that each of the non-elected group top ten from the previous ballot must appear on the ballot (either on the "top-ballot" or the "under-ballot).

I believe the current regime gives 6 points to a player appearing in 15th spot on a ballot. If so, an additional rule reflecting monotonicity would require that no more than 6 points can be allocated to any of the "under-ballot" players (below the top 15 spots).

The idea is that this "under-ballot" gives (forces) each voter the chance to reveal their rank ordering, and to some limited extent their strength of preference, among the top players (according to the group) that they did not include on their "top-ballot".

I will use my 1952 ballot as an example. My top 15 players would appear as on my submitted ballot. I then have 10 additional vote-points that I must allocate among other players. Since I did not vote for group top-tenors Ferrell, Mackey, Jennings, or Griffith, these four must be among those allocated these additional 10 vote-points.

Since I rank Griffith ahead of Jennings who is ahead of Ferrell and Mackey, I would allocate my 10 additional vote-points as follows:
5 Griffith
3 Jennings
1 Ferrell
1 Mackey.

I am thereby allowed (forced) to reveal information about these key players that would otherwise not appear on my ballot.

If a voter has voted for all of the top tenors (or all but one), then he is allowed (forced) to vote for additional players. In that way no voter is rewarded or penalized in his voting contribution based upon his degree of agreement with the group consensus.

I fully realize that this is a bit of a departure to our current system. However, I think it is an improvement in that it addresses the issue that concerns many of us. The key to the proposal is that it provides more information from each voter about the most "important" players so that we can be sure that the group tally reflects every voter's thoughts on these top players.

The downside to this proposal is that vote tallying will be more of a challenge. I would be happy to volunteer to help out in the vote tallying so that this would not be a hurdle in fairly evaluating the merits of the proposal.

Thanks much for your consideration.
Page 2 of 2 pages  < 1 2

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Dock Ellis on Acid
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.0350 seconds
49 querie(s) executed