To keep everything in one place, here are the posts that appeared in the Newsblog thread:

1. OCF Posted: April 19, 2011 at 11:19 AM (#3801854)
This thread seems to be posted in the wrong place - it's on the BBTF Newblog, when it should be inside the Hall of Merit section.
2. sunnyday2 Posted: April 19, 2011 at 11:31 AM (#3801861)
Ditto.

My most basic understanding of this is that there is one ballot covering both leagues.

One ballot slot per team = 18 in 1961. I'd like to see a little deeper ballot. For 1961 to the present, maybe it escalates from 25 to 30. Or maybe it's always 25. My question would be, how deep of a ballot do you need to successfully differentiate our collective #10 from our collective #15 or #20, statistically speaking? That's not a function of the number of teams. I guess it depends on the number of voters. But anyway, I'd like a little deeper. I don't see how that would be any more work, since I'm going to eval a lot more than 18 to get to 18.

But if it's 18 and in other years it's 16 to 30, the value of a #1 vote should be the same every year, say 100. With 18 you'd be approx. 100-95-90, etc. With 30 you'd be 100-97-94, etc. Something like that. Then the number of career points would be comparable among all players.
3. snapper Posted: April 19, 2011 at 11:37 AM (#3801869)
Can anyone vote in this? Or is it restricted to previous HoM voters?
4. DanG Posted: April 19, 2011 at 01:35 PM (#3802024)
the length of the ballot should be determined by the number of teams in the major leagues each season, with one place per team
--I have always argued for a constant-sized ballot, the general idea being that the number of teams in the league has very little impact on the names/rankings of the players at the peak of the talent pyramid that we will be assessing.

--The thread linked at the top here doesn't seem to have much discussion of this new project. My original MMP proposal from four years ago, and much subsequent discussion, is in the "Once We Catch-Up" thread.
5. fra paolo Posted: April 19, 2011 at 01:42 PM (#3802032)
This is indeed in the wrong place. We are having some teething troubles, but I imagine it is better to get it going somehow, that wait a while longer. More apologies.

Snapper, anyone can vote in this as long as they are willing to follow the philosophy of the HoM. You have to defend a statistical case as to why, for example, Frank Robinson had a more meritorious season than Hank Aaron in 1961.
6. sunnyday2 Posted: April 19, 2011 at 06:30 PM (#3802359)
--I have always argued for a constant-sized ballot, the general idea being that the number of teams in the league has very little impact on the names/rankings of the players at the peak of the talent pyramid that we will be assessing.


Dan is of course correct. A 25 man ballot would vary over the past 100 years from about 3 percent to about 10 percent of ML rosters. I'm good with a constant ballot number, but if the consensus it should slide, I can slide with that, too.

I just don't want it to be too short.
7. sunnyday2 Posted: April 20, 2011 at 03:42 PM (#3803339)
hello hello hellp
is there anybody in there
8. sunnyday2 Posted: April 25, 2011 at 11:30 PM (#3808988)
Let's see where this little sucker goes when it gets a new hit! Purgatory, most likely.

EDIT; Well it did NOT go onto the Newsblog Hot Topics. But it's in the Newsblog! WTF! We are screwed.
9. Yardape Posted: April 25, 2011 at 11:40 PM (#3808995)
Well it did NOT go onto the Newsblog Hot Topics


I found it from the Newsblog Hot Topics...

But if it's 18 and in other years it's 16 to 30, the value of a #1 vote should be the same every year, say 100. With 18 you'd be approx. 100-95-90, etc. With 30 you'd be 100-97-94, etc. Something like that. Then the number of career points would be comparable among all players.

Other than comparability, is there a reason not to use the lowest points total like in the BBWAA MVP votes? 1st place is one point, 2nd is 2 points, and so on? Wouldn't the number of voters make player comparisons in either system difficult? If there are 20 voters for one year in which there's a unanimous MMP and 30 for the next unanimously selected player, they will have vastly different point totals in either system, even though they were both unanimous choices. Unless we mean to average the totals, in which case those players who are non-unanimous selections get shafted (only when making comparisons) if there are more voters with widely varying opinions. Sorry for this question. I'm a newbie.

Just so you know this is being read I would like to participate in this though I have not been a Hall of Merit voter. Snapper's question in the previous posting (thanks fp) was my big one.

I would prefer a smaller ballot - probably around 20 people for the early years, maybe growing on the order of 1 spot per team with expansion.

Alternatively for a larger ballot, I think the points need to be on a larger spread at the top of the ballot; e.g. 100-90-82-75...3-2-1. Intuitively it seems logical that the variance between the first few players should be more strongly rewarded than the variances between the 29th and 30th players.

Re #2:

I think the problem is if we have different ballots. A player who appears on a limited number of ballots will score better than a player who appears on many. For example, using a 1-2-3 system;

Jose:
Pujols - 1
Hernandez - 2
Longoria - 3

Piehole
Longoria - 1
Pujols - 2
Utley - 3

If I'm following you correctly Hernandez wins the MMP despite appearing down ballot on just one ballot;

Hernandez - 2
Pujols - 3
Utley - 3
Longoria - 4

Like you I'm a newbie here so I may be missing something obvious.

How many players do we vote for?
What is each place on the ballot worth?

I offered the following proposal in #587 in "Once We Catch Up":

"As DL points out, 10 is not enough; sure, you might end up naming 20-30 players (more if yest and karl are voting) but you won't get a good ranking of them. But Patrick is also right that we don't need to split hairs to separate the down-ballot guys.

I offered the solution in #146, what I called "Quadrant Voting". The idea is to individually rank the top guys and group-rank the rest.

You could rank 10 and then have groups of five players following that. A 30-man ballot might get points something like this:

Top 10: (40-38-36-34-32-30-28-26-24-22)
2nd 10: (18-18-18-18-18-13-13-13-13-13)
3rd 10: (8-8-8-8-8-3-3-3-3-3)

Or you could rank 10, followed by pyramiding quadrants:

Top 10: (40-38-36-34-32-30-28-26-24-22)
next 2: (20-20)
next 3: (17-17-17)
next 4: (14-14-14-14)
next 5: (9-9-9-9-9)
next 6: (3-3-3-3-3-3)

Or a million other ways to do it."

One thing I meant to point out about these two schemes is that the points assigned for the last 20 positions gives the same total points (210) as a sliding scale of 20-19-...-1.

Also, thinking about it a bit more, the top 10 would be better to start at 30 pts for #1 and decrease by one, down to 21 pts for #10.

DanG keeps raising Quadrant Voting as a possibility, but no-one has expressed an opinion about it.

Do people think it is too radical a notion? Or does silence indicate indifference?

For newbies, the concept of the MMP is given a brief treatment in this post:

The Most Meritorious Player (MMP) Project – Take 2

I think I like quadrant voting if the idea is to save time and thereby make the bar for participation lower.

Been hanging around the HoM for a couple years now - great stuff. I'd be really interested in being involved in this project. Is registration needed or does one just show up for discussions and voting?

Re new voters: it's pretty clear to me that we won't get all, or even most, of the 50-60 people who were the core of the HoM project between about 2004 and 2008. For one thing, that's already some time ago. And the MMP project is different - different rhythms, different major points of contention, far less continuity from one ballot to the next. And it won't do to have just 15-20 people voting. So we need new voters, quite a few of them.

In my opinion the right answer to Nolan should be "just show up for discussions and voting" - as long as that includes showing up for the discussions and not springing any surprises on us. And to accept that other voters will question some of your choices, and to be willing to defend yourself (or change your mind if necessary).

For my own sake, the longer the ballot, the less attractive I would find participation. Anything to make it easier would be a positive.

How about self-selected quadrants? A voter could choose to write a ballot something like

1. -
2. -
3. -
4. -
5. -
6 though 9 the following four: -. -, -, -
10 through 16 the following seven -, -, -, -, -, -, -,
17 through ... you get the idea.

Then the ballot tabulators would have to find a way to throw the designated points for slots 6 though 9 together and average them. It might take a little effort to construct the ballot counting spreadsheet to do that, but it might be possible.

Interesting project and ideas.

I'm not sure I completely understand Quadrant voting... Is the idea only to rank the top 10 (or whatever) with individual ranks and the ones after that get assigned to groups? So... Well, maybe an example would help. So a real ballot would look something like this then?

(Not a real 1871 ballot. I won't necessarily defend these choices.)
1 Rynie Wolters
2 Ross Barnes
3 Al Spalding
4 Jimmy Wood
5 Levi Meyerle
6 Fred Treacey
7 Cal McVey
8 George Zettlein
9 Lip Pike
10 George Wright
11 Al Reach
11 Steve King
13 Dick McBride
13 Fergy Malone
13 Dave Eggler
16 George Hall
16 Ed Pinkham
16 Ezra Sutton
16 Joe Start
20 Ned Cuthbert
20 John Radcliffe
20 John Bass
20 Deacon White
20 Jim Foran
20 Davy Force

If so, hmm... Not sure how I feel about that.

I think the point of quadrant voting is to reduce the burden of voting, right? If you want to participate in this project, as I understand it, you're basically saying you'll vote weekly for several years. (The project will still take several years to complete for MMPs back to the 1870s.)

Quadrant voting allows you to put guys who are all roughly equivalent in the same category but when your votes are combined with others, all of the other voters' votes sort out the (probably) minor differences between them. Since you're not voting in a vacuum, other voters' inputs will make easy what appear to you to be tough and/or unimportant decisions.

Please let me know if I'm putting my foot in my mouth. I think it's important that I (we) understand the process.

That's a 25-man ballot. DanG was talking about a 30 man ballot.

Let's suppose the ballot size is 25. (I personally think 30 is too big.) Let's say the votes scale as
35-33-31-29-27-25-23-21-19-17-15-14-13-12-11-10-9-8-7-6-5-4-3-2-1

Then I'd take Nate's vote in #11 and do this with it:

Wolters gets 35 points
Barnes gets 33 points
...
Wright gets 17 points
Reach and King get 14.5 points each.
McBride, Malone, and Egger get 12 points each (as if they were 14th)
Hall through Start get 8.5 points each (as if they were 17.5th)
Cuthbert through Force get 3.5 points each

I might want to prohibit any combinations of ranks 1-10 with ranks 11-25, just to prevent us from having point totals like 15.3333 or 14.857143.

Whether I could build a usable tallying spreadsheet out of this idea - um, I haven't tried setting it up yet.

I guess I'm more or less indifferent as to quandrant voting. I don't have a problem with just ranking them, but if everyone else wants to do quadrants, that's fine by me.

I think all the ballots being bandied about are way, way, way too large. The HOM had 15 person ballots in elect 2 and elect 3 years, right? So if we're 'electing' one person per year, seems to me that a ballot should be 5 to 10.

I see no reason why a large ballot would add anything in terms of voting accuracy, and it would be a huge time-suck and vote-deterrent if enforced strictly (meaning, even with quadrant voting, if you had to defend all your selections and made them rigorously, you're talking about hell of a commitment of time and effort).

5 slots per season till expansion. Maybe 7 per season now. Make this simple and easy, not a chore that chases away anyone who can't utterly and totally geek out on it.

I see no reason why a large ballot would add anything in terms of voting accuracy

I think DanG's rationale in the link he posted above in #7 convincingly explains why you need to rank more than 5 or 10 players. Maybe what should change is having to defend all of the selections. Defending 25 or 30 picks each week would probably be a burden. Why not just defend the top 10 choices and anything unusual in 11-30? Does every choice (1-25 or 1-30) deserve a comment? Particularly if we're going to use quadrant voting?

I was interested in joining in on the HOM but was always overwhelmed by the amount of work that would have been necessary to catch up. I would really like to participate here, particularly given OCF's comments in #10.

From my perspective, ranking 25-30 sounds like a LOT. When I've voted in the Internet Baseball Awards I found 10 slots to be somewhat limiting but would have found it very difficult to go beyond 20. And that's for a season I closely monitored while it was going on.

Obviously, I'll defer to the will of the existing electorate, but just thought I'd toss in my two cents.

Quadrant voting seems fine. My feeling is that the best approach would be to mandate clear 1-10 rankings, but allow ties (at the discretion of the voter) beyond that.

I don't think DanG's rationale is at all convincing. IMO, this is just a mechanism to keep out all but the most devoted 'voters'. I could - maybe- see the logic for a larger ballot in certain backlog elections in the HoM, but that was with a larger pool of candidates, more 'elect me' spots, and with two wildly different concepts of value (peak v. career) that were in competition.

Not to put too fine a point on it, but if 15 players versus 5 players on the ballot is determinative in an MMP election, then the system is broken. There simply shouldn't be that kind of variance.

EDIT to add: Think of it this way. In a theoretical world where the number of spots on a ballot does not effect the number of voters, there would still be a point of diminishing returns where adding names to the ballot wouldn't increase the accuracy of the election result. I think that point is well under 25-30. Then, when you add the inverse relationship between ballot size and expected voter participation, I think you're actually losing accuracy as you increase ballot size. And if you have a large ballot but don't require defending it, then that just invites formula voting or WARP voting or some other blackbox, ill-considered approach at the bottom of the ballot, and all the large ballot does is introduce false precision.

Quick hits:

--Quadrant voting is simply an alternative to make it easier for those voters who feel daunted by the prospect of filling out a large ballot. Personally, I don't see it as necessary. OCF's idea of each voter being free to group players in any way they like is great, but goes against the KISS principle.
--No, we're not looking to elect "one person per year". The general concept is to create an ordered ranking of the greats for every season. Follow the link in [#7] or in the lead above to understand better the aims of the project.
--As for justifying/defending your ballot. You don't really need a big writeup for every player. It's more a matter of explaining your system, what you put importance on and why. In the HoM we were very tolerant of idiosyncrasies and I suggest we continue that philosophy here. As long as we get more than 20 well-thought-out ballots, the madcap ballots will go out with the wash.
--For the record, after some discussion, the year 1961 was settled upon as the starting year for the project. I refer you to the later postings in the "Once We Catch Up" thread, if you're interested.
--Voting is easy; analysis is difficult. This is why I advocate creating a wave of discussion threads, one for each season (say, 1961-70) before we ever begin voting. If you're like me, you do research when the time and the mood converge. For instance, maybe I'll post three years of preliminary ballots in one day, then go a month without giving the project much thought. By allowing a few months for each year to be open to discussion we have a much better chance to gather worthwhile analysis from people. So I think it behooves us to always have the next 5-to-10 years already under discussion. And, as Piehole points out, "other voters' inputs will make easy" the work of sorting things out.
--At the top of each discussion thread should be listings of basic data for that season, such as the top 40 players in WAR, win shares, etc.; links to MVP voting and historical narratives; TSN all-star teams, and so on. IOW, since not all voters have the same resources, we'll give them whatever help we can.

I'm (yet another) person who has followed the Hall Of Merit but never voted in it as it just seemed too difficult (for me personally) to ever come up with a ballot that I thought was fair. Anyway, I think this yearly project is much more approachable and I'm hoping to participate.

As for the question about quadrant voting... I'm fine with it or without it... Although I might prefer it if the ballots do end up being 20+ players.

And I would be very pleased if (as DanG suggests in comment #19) there are listings of the top 40 (or so) players in WAR, win shares, etc. for each year in question.

If this project is not about ranking the most meritorious player, then why call it that? Call it "best 20 guys in a season".

I think that the notion of ranking the twentieth best guy in each season may have appeal to a core group of 10 or 15 guys. It has zero - none - zip - chance of attracting a critical mass of voters to create a vibrant electorate.

I say this as someone who both (a) adored the HoM and (b) has a full-time job - there's no way I'd participate in this project. It's an order of magnitude more of a commitment than the HoM, since each ballot is (a)new, and not largely recycled/updated and (b) huge, larger than an HoM ballot. And in the end, the effort is wasted, since the diff b/w spots 15 and 20 are not significant with respect to the precision with which we can analyze value in the distant past and, more importantly, no one gives a #### about who was the 25th best player in 1957. The HoM had a purpose and a connection to a real life problem - a MMP project that was more than MMP in name only would also have a purpose. What you propose is ranking for rankings sake.

The HOM was at it's best when it was accessible. That's what made the initial incarnation so great- it was simple, easy to understand, and practically invited participation and discussion from as many people as possible. A 20 person ballot manages to lose all the attractive aspects from the original project while catering to its geeked

... to it's geeked out, exclusionary element. Think about this from the perspective of an outsider, and you will realize it is a total nonstarter.

I'm agreeing with 'zop here. Shorter ballot. Maybe 15, but not beyond that.

I've still got my post-season memos to some friends from the mid-90's. The structure: choose a 25-man all-star squad for each league, with starting lineups and reserves. And then I did include both a CY and and MVP vote in each - but the CY list never went beyond 5 deep, and the MVP list never went beyond 7 or 8.

This sounds like a lot of informative fun, so I'm in no matter what. But, having participated in what amounts to quadrant voting in the old Baseball Maniacs, I am in favor of it. QV keeps the amount of work down for each voter, and what's more, it keeps it down in the place where the workload produces the least return on investment - the weakest guys, the bottom of the ballot. The Baseball Maniacs project fell apart largely because we did not have ENOUGH quadrant voting. It just got to be too much work for too little return to deal with the guys who were getting zero, 1, 2, or 3 points each, none of whom was going to get into even the Outer Circle. If we'd just assigned all those guys 1 point each, I think the project would have eventually completed. But when you've spent just a week separating out all the guys who have any real chance of getting into any circle, and you then spend 7 weeks of your two months sorting out the 0-3 guys, you eventually wear down and/or just get bored. A short ballot, of course, also helps with the workload. QV is a good way of reducing the workload without having to shorten the ballot.

- Brock Hanke

To summarize after a day's discussion, people seem not to mind Quadrant Voting, especially for a longer ballot. It seems likely that from the point of view of the people doing the count we would want to fix the quadrants, rather than allow voters to individualize them — unless people don't mind a longer gap between closing the balloting and releasing results.

The main debating point at the moment is over the length of ballot. It's fair to say that the consensus view from the 'what next' thread is a ballot at least as long as the 15-player HoM one. Yet here we've had a clear statement in favour of a shorter ballot. It could be Quadrant Voting is a requirement for a longer ballot.

In which case, what we might need is a ballot that ranks a top ten, then quadrants whose aggregate points values divided by the number of players don't lead to fractional results (just to keep counting simple).

A secondary topic that has emerged is whether to have multiple elections going on at the same time. If I post the 1961 thread at the end of this week, do I then post the 1962 thread a week later? Or do we want to get the ballot thread up before we post another discussion thread? Or do we post nine discussion threads and then start balloting, probably in the autumn? Or some other permutation?

whether to have multiple elections going on at the same time

Not multiple elections; multiple discussions. Don't open another election thread until the previous election is over.

Or do we post nine discussion threads and then start balloting, probably in the autumn?

Something like that. Each season has 3 threads: Discussion, Election, Results. I suggest posting ten Discussion threads now (1961-70); maybe a few days gap between them, whenever you can pull the season's data together.

Whenever folks are chomping at the bit to get voting (be it June, August, or whenever) you should post the 1961 Election thread. Leave it open for two weeks. After the first week, post a thread for the 1971 Discussion.

Establish a set time to close the elections, to establish a routine for the voters. HoM elections always ended at 8 pm ET on Monday, IIRC; you may want to do likewise. Within 48 hours after posting the 1961 Results thread, a 1962 Election thread should be up. One week later, post a thread for the 1972 Discussion. And so on.

The idea is to always have the next ten elections under discussion while running biweekly Elections. With biweekly elections we would reach the present two years from now (52 elections, 1961-2012).

I like the idea of a longer ballot but understand how it could be time consuming. I would be willing to accept ballots of varying length in order to accommodate those people who don't have time to rank all the players.

Here's a compromise proposal:

At minimum, rank the top N/2 players each year (where N is the number of teams in the league). At maximum rank the top N players each year. Points will be assigned in descending order by ballot position

N
N-2
N-4
N-6
N-8...
until half the ballot has been reached. The remaining slots will be worth 1 point each.

So for the 1961 ballot the slots would be worth

1st - 18
2nd - 16
3rd - 14
4th - 12
5th - 10
6th - 8
7th - 6
8th - 4
9th - 2
10th-18th - 1

Any ballot with 9-18 players listed would be considered a valid ballot.

We'll leave the question of how many "teams" there are during segregation for another day. For 1961-present the answer is pretty clear.

I agreed to help run the project (though I still haven't tried posting a thread). I can't run it any faster than 1 ballot per month. That means we're caught up to the present in 4 years and finished in 12. I'm fine with starting a project that outlasts my participation. We've been treading water for 4 years around here, slow progress is better than none.

I still think the wrong question is being asked - the question should not be, "What ballot length is most appealing to the guys who post on the 'whats next thread'?". It should be, "How do we construct a project that has a minimum of 30 active voters?".

Edit - I mean, you could take the most compelling question in historical baseball - whatever it is - and if it's just 10-15 people voting regularly, the project would die quickly. You could take an inane question - "who were the best 2nd basemen in the 1937 NL?", and if you have 60 intellignet people debating it, it would be awesome. The HoM was awesome and long-lived because of the discussion - the product was nice, but that's not what makes participation worthwhile.

I can't run it any faster than 1 ballot per month.

Why? If we follow what I've proposed, and open discussion for each season 20 weeks prior to voting, casting ballots is the easy part. Two weeks is more than enough time - why a month?

How do we construct a project that has a minimum of 30 active voters?

A legitimate question. IMO, if we build it, they will come. The challenge, as I see it, is to have a broad appeal; to make it accessible not only to math majors and statisticians, but to cultivate a diverse electorate with imagination and knowledge of baseball history.

if it's just 10-15 people voting regularly, the project would die quickly

This has not been my experience. I have run long-term projects with 7-10 dedicated, intelligent voters and have gotten very satisfying results.

Each discussion is theoretically "open" for 20 weeks but I'm going to have to time share on each of them. The time it takes between votes is the time I'm going to be paying attention to that year's particular discussion, creating a ballot, etc. I want to participate, not just count votes.

If things move smoothly and we want to speed things up later, that's fine. Currently we're at 1 ballot every 4.5 years and counting. I think 1/month is a massive improvement.

This has not been my experience. I have run long-term projects with 7-10 dedicated, intelligent voters and have gotten very satisfying results.


I think this would be a really, really bad outcome for any future HoM-descendant project. If you view this as an acceptable future, it explains why you're OK with 30 person ballots and what-not.

How about something like this:

1st = 18 or 17
2nd-3rd = 17 or 16 or 15
4th-6th = 15, 14, 13, or 12
7th-10th - 12, 11, 10, 9, or 8
11th-15th = 8, 7, 6, 5, 4, or 3
16th-18th = 3, 2, 1, or zero

Constraint 1 - You can have more than one player with the same number of votes, as long as all players get vote totals that are within their group's range.

Constraint 2 - You do NOT have to assign a player to each number. If you think there are 3 5-pointers tied, but no 7 or 8 above them, that's OK.

Constraint 3 - You can have players from the endpoints of different groups with the same number of points. For example, you could decide that you have an 8-pointer at the end of the 7th-10th group and also a tied 8 pointer at the top of the 11th-15th group.

Constraint 4 - Your total points have to add up to 18+17+...+3+2+1+0 = 171

This gives you a little wiggle room, just in case you really do think that two players are tied at a group dividing line or something. But it also enforces a ballot that looks like a pyramid, except at the bottom. That seems pretty realistic to me.

- Brock

I don't like people assigning their own points to their ballot slots. That leaves potential for gaming the balloting. Just rank them.

There are so many problems with the proposal in [#27] that I don't know where to begin.

I think that under that system you would need a strong "ballot committee" to closely monitor for ballots that attempt to manipulate/skew the results. Ideally, we want to devise a system where intervention is rarely needed.

But what I really want to talk about is the benefits of having a large ballot. In the link in [#7] I made a quick list of things we could discover by doing the MMP Project:

1 The best players in baseball for every season, ranked
2 The best players in each league for every season
3 The best players at each position for every season
4 Relative “star power” of each league
5 Which teams in history had the most star power; pennant winners with little star power; also-rans with mega star power
6 See exactly in what eras a position was weaker or stronger relative to the other positions
7 The top rookies in baseball history
8 Years with no good rookies
9 Create a peak value system for everyone in history, true MVP Shares
10 How many years, exactly, was a player among the top 10, 20, 30 in the game?
11 An automatic normalizing mechanism, e.g., no matter how much you stand out from the league you can’t be higher than #1

In order to reliably achieve the benefits of #3 thru #10 you need to have voters list at least 25 players on their ballot. If the project is worth doing it should accomplish all these things and more. I think that it's worthwhile to try and rank all the "star" players each season. Small ballots won't get 'er done.

Quadrant Voting looks like the way to go. As Brock said in [#24]: QV is a good way of reducing the workload without having to shorten the ballot. Keep it simple? OK. Rank 1 thru 10, then three groups of five. Assign points 25-24-...-16 for the top ten, then 13 points each for 11 thru 15, 8 points for 16 thru 20, and 3 points for 21 thru 25.

Each discussion is theoretically "open" for 20 weeks but I'm going to have to time share on each of them. The time it takes between votes is the time I'm going to be paying attention to that year's particular discussion, creating a ballot, etc. I want to participate, not just count votes.

If things move smoothly and we want to speed things up later, that's fine. Currently we're at 1 ballot every 4.5 years and counting. I think 1/month is a massive improvement.

I understand that it's your game and you design whatever will work for you and your partner. I would hope that you and fra paolo are also able to consider what can be done to enhance interest and participation in the project. A month between elections runs contrary to this ideal, IMO.

But what I really want to talk about is the benefits of having a large ballot. In the link in [#7] I made a quick list of things we could discover by doing the MMP Project:



Again, you miss the point. If you can't get a vibrant electorate it doesn't matter what you can discover.

Separately, even if we assume that the most important part of a project is the results, I think 3-10 can be achieved with a 10 person ballot, easily. I simply don't think the added value by going over 15 (let alone to over 25!) justifies the effort, even if it wouldn't reduce voter participation. Since it will, in fact, guarantee that the project is ~10 people rather than the HoM's 40+, you basically kill the quality of data by shrinking your voter pool. Dan, if you were going to undertake this project YOURSELF, like to write a book about it, you'd have the right idea, but its totally inconsistent with running an inclusive project.

In order to reliably achieve the benefits of #3 thru #10 you need to have voters list at least 25 players on their ballot. If the project is worth doing it should accomplish all these things and more. I think that it's worthwhile to try and rank all the "star" players each season. Small ballots won't get 'er done.

I think my big objection to this (besides the work involved) is that I don't think a big ballot will get it done either. Here are the positions of the top 30 in MLB by Fangraphs WAR last year:
CF, 1B, 1B, 3B, 3B, P, 3B/OF, 3B, LF, LF, P, SS, 2B, P, P, P, 1B/3B/OF, P, 2B, P, OF, 2B, OF, P, P, 1B, RF, P, CF.

Brian McAnn, best catcher by WAR is 31st overall in MLB. SS and C seem to be 2 positions that will be disproportionately affected by a small ballot, but neither will they necessarily be well-represented by a large ballot. I don't think the large ballot necessarily accomplishes all of what you say it will accomplish.

Edit: Rather, I don't know that it's worth it to rank so many players to find out that the best catcher was the 31st best player, or that the next best shortstop is 52nd overall.

Here's my proposed compromise:

Elections are in two parts. In the first part, a "nominating committee" does the DanG proposed election with huge ballots and such.

This election produces a 25-player list. From that list, voters choose players in the MMP "election", with a 7 player ballot. You must explain why any player in the top 10 of the nominating committee results are not on your ballot.

Thoughts?

This election produces a 25-player list. From that list, voters choose players in the MMP "election", with a 7 player ballot. You must explain why any player in the top 10 of the nominating committee results are not on your ballot.

I LOVE this idea. The best of both worlds.

vibrant electorate

I still believe if we build it, they will come. The HoM started with 29 voters, eventually growing to a maximum of 56. I believe that a worthwhile aim is what will attract quality voters. You believe a simple process would be the big draw. Wouldn't it be best to have both?

don't think the added value by going over 15 (let alone to over 25!) justifies the effort

You're overstating the difficulty. I don't think it's worthwhile for anyone to agonize over trying to precisely distinguish player #23 from #24. But it's not hard to see a group of players who belong in the general range of #21-25 and whittle it down to five guys.

Since it will, in fact, guarantee that the project is ~10 people

What experience is this assertion based on?

I'm fine with either a regular 10-15 player ballot (one or the other, not ballots that can contain a range) or a larger ballot that has quadrant voting, as long as the quadrants are defined before hand. I don't think there's a lot of value in trying to determine whether Jim Foran or Davy Force was really the 25th best player in 1871 (to use my earlier example), so either cutting the ballot down, or just having group rankings after a certain point, seems to me to the be best way to address that. Between the two choices I'm fairly indifferent, but if you put a gun to my head (and please don't :) ) I'd go with quadrant voting, because I'd find who people ranked farther down moderately interesting.

Some of the other ideas, like varying length ballots, voter defined quadrants, or multi-stage elections, just seem like needless complications to me. I think the process should be fairly simple, both for voters and for ballot counters, and not have a lot of hoops to jump through.

Anyway, while there's value in discussing this, no process is ever going to get full agreement, and while I'm just posting my thoughts now I realize this has been discussed, in one form or another, for at least four years. The Hall of Merit was a success, it seems to me, but the process wasn't perfect (in looking back over the threads, there was some definite discussion of how people would have changed the process/voting procedures if they could have done it over), and neither will be whatever we pick here. The worst case scenario is what, we start and everyone hates the voting process, or not enough people participate to keep it going? Well, that'd be unfortunate, but not the end of the world, and there's no reason why the process couldn't be changed in such a scenario.

I don't like Zop's compromise at #40. It makes a huge project last even longer (twice as many elections), and I think some voters will be turned off that they are limited in who they can vote for.

(non-participant opinion)

I'd like to briefly say im my 8+ years of debating various procedural issues at the HOM, I have never more thoroughly agreed with anyone than I agree with 'zop in this thread. If there are more than 10 people on the ballot I don't see how I could participate and I can't imagine that 'zop and I are alone or close to alone on this. And, of course, this is coming from a person who didn't miss a ballot from 1915 onwards.

I believe that a worthwhile aim is what will attract quality voters.

Yeah, but be real - the reason this project has stalled for so many years is that we don't have a compelling aim like with the HoM. You list of 10 purposes was, particularly once you get below #3, pretty esoteric.

I don't think you can say "if we build a process that interests me and the 9 guys who are most interested in this sort of thing, 40 other people are going to say, "hey, this is a great idea!". I'll concede that a good project will grow over time but I submit to you that if you don't think that you'll be consistently drawing 30 voters to the first wave of elections (like the HoM did), then you should be adjusting to do that.

"If you build it, they will come" is a line from a movie, and at best an aphorism. Its not how you actually build a project in real life. Create something that appeals to a lot of people, and a lot of people will come. Create something with a high barrier to entry and that's exclusive, and people wont come.

OK, I'll chime even though I rarely do outside of the HOM election time. An MMP ballot ideal for me would be a 15 player ballot. Beyond 15 is quite a bit of work to build the consideration player dataset, having to start over each year. Also, once down to #15 are you really showing a contender for the overall MMP award or are you just passing out a feel good vote? Many years you'll have 3-5 strong contenders, the rest are also rans. They'll be many different players getting their name in the final voting results, that's expected with our diverse voting bloc. Any pet player will be shown to have been a good player in year X.

Next each player separately slotted i.e. no quadrants. We're deluding ourselves if we suppose that player values will be easily separated at all times. But even with the HOM you have players basically tied for value; somehow voters made a tough choice and selected someone to have on/off ballot. I'd continue to request voters do their best and split players throughout the ballot.

I'll probably try to continue participation regardless of structure but those be my preferences.

It makes a huge project last even longer (twice as many elections)

Why does this need to be true?

and I think some voters will be turned off that they are limited in who they can vote for.

Huh? What the two stage process is doing is actually reducing the amount of work that you need to do to explain your choices. You're not limited at all. What the nominating committee is doing for you is separating the goats from the sheep. If you choose to vote for all goats, it's not a problem, you just need to explain why you prefer these goats to the nominated sheep.

I may be one of those y'all are trying to attract to the project. OTOH I am so ignorant you may be thinking of ways to deter me :-D Here's a stray idea:

no one gives a #### about who was the 25th best player in 1957

Not per se, but maybe there's a sense in which an analysis of "MMP Shares" would be interesting. It might not tell one much that one didn't already know, but if the intent is to do a better job of MVP voting in retrospect than the annual voters did, then knowing things like "how many years did a player get [a certain level of] MMP votes" is at least as interesting as knowing their Award Shares at the moment.

So the idea of a fairly large ballot doesn't deter me, as a potential newbie. (Maybe it should, for the "full-time job" reason 'zop mentions.) I never voted in the HOM because my knowledge of the whole span of baseball history was too weak for me to make a good contribution. But 1967 through 2011, I've been paying attention in real time, and I don't think it would be as hideous an effort to rank, say, 20 players per year as to comb through all of history to put together an HOM ballot. In fact, I used to fill out imaginary MVP ballots for my own amusement, years ago, 10 names in each league – 20 for both leagues doesn't seem a deterrent.

That said, I may just be too distracted by that full-time job to contribute anyway. Which will surprise those who see that I have 9,000 comments or some such number on BBTF ...

I've been lurking on BTF for at least six years, and this is the first time I've been moved to comment.

Anything more than a 15-player ballot is a VERY bad idea, for exactly the reasons Zop lays out. If the projects demands a 40 player ballot every year, the people will most decidedly not come, quadrant voting or no quadrant voting.

If you demand a ten-player ballot, I have every intention of voting myself (yay for not having to wrap my puny mind around Negro League Equivalencies to ethically contribute!).

I'm skeptical of a nomination committee, if only because the openness of the Hom to non-conventional candidates and approaches to value has always been part of its charm (the yest factor). But that's much less important than presenting the general voting public with a manageable and appealing task.

My nomination committee idea SHOULD NOT be construed to mean that someone cannot vote for someone who was not selected by the nominating committee. Rather, we just add some provision that if you want to vote for someone off the list, you have to "provide extensive, reasoned support for why you believe the committee was in error". You can be yest, but you have to be able to explain yourself!

The idea is that the nomination committee satisfies people who want to increase the depth of information ( eg, "Roberto Clemente has [X] MMP shares and [X] MMP nominations"), while maintaining electorate size to maximize the diversity of opinion and accuracy of selection at the top of the ballot.

Also, I did not anticipate the nominating committee idea to lengthen the process, but rather, to be an intermediate step, in the existing proposed election process. Since, in theory, no additional work is required to make a final ballot compared to a nominating ballot (just cut off your original ballot at #10), the elections can be fairly closely spaced. Discussion would begin prior to the nomination election and continue until the "final" election. If the big ballot idea is compelling, the proportion of participants voting in the nomination election will grow and at some point, the "final" election could be eliminated.

I fail to see how you would game the system with 1 point votes but maybe you can explain it to me. My next proposal is to set the ballot length at N/2 with N the number of baseball teams. I strongly believe the number of ballot slots has to be tied to league size. This would lead to ballots of 9-15 players for the period of interest. If the number of available slots isn't tied to league size you end up with players getting "MMP votes" based on league size instead of relative percentile accomplishments.

The two stage process is a non-starter. There is no nominating committee for HoM voting, why would it help for MMP voting? The pool of eligible players is anyone who played that MLB season with perhaps a cutoff for playing time (>160PA/40IP?). I want diverse opinions on what contributions provide value to a team in a given season.

I will say that DanG's goals of a long ballot and a 2 week time period oppose each other.

There is no nominating committee for HoM voting, why would it help for MMP voting?

Because we don't have a "returning top 10". Aren't we going to want some system to determine which players require either a vote or discussion? I thought that was a greta feature of the HoM - it forced people to reevaluate players that had fallen from their consideration sets once the players cracked the top 10. Without that feature, I doubt many of the guys who languished on the ballot for many years would have been elected in the late backlog elections.

Why would you forget which ballplayers played in 1961? You can look it up quickly. The returning top 10 helps when you are evaluating 150 years of baseball history including players with spotty or non-existent documentation. When you're looking at 1 well documented season it is a lot more difficult to overlook people. We plan on posting relevant WAR leader boards in the yearly threads.

I don't think it was a forget/remember thing, I think there was value in having ot articulate why you didn't support a guy who a substantial fraction of the electorate thought was Meritorious. I mean, lets take the following situation: someone posts a well-reasoned 12 person ballot . . . and leaves off the #3 player by WAR, without mentioning the player. Is the ballot valid? What if the ballot was submitted right before the deadline?

In the contemporary era, sure, use top-ten-WAR as a proxy, but what happens when we get back to where defensive stats are iffy? What happens when we get back pre-integration? I leave Josh Gibson off my ballot in a year where he was clearly a top 5 player alive - what then? Shouldn't I have to explain myself?

a long ballot and a 2 week time period oppose each other

There should be no difficulty in putting together a long ballot after 20 weeks worth of discussion of a season.

and leaves off the #3 player by WAR, without mentioning the player

Make it a requirement for voters to comment on anyone in the top 10 (or top whatever, depending on ballot size) in WAR, even if you didn't vote for them.

Just a note from the peanut gallery, but I think DanG's Field of Dreams approach to the project is flat wrong. I'm interested in this idea, but if I'm expected to go thirty deep every year, I'm just not going to do it. Far exceeds my attention span. You'll either end up with a lot of (A) straight WAR or Win Shares ballots, or (B) nobody voting. And I don't think this is a project whose results will hold much interest for anybody if there are only 10 voters.

If the number of available slots isn't tied to league size you end up with players getting "MMP votes" based on league size instead of relative percentile accomplishments.

A player's "relative percentile accomplishments" are not relevant. The 10th best player is the same guy whether the league has 12 teams or 24, although his percentile rank is very different.

> There should be no difficulty in putting together a long ballot after 20 weeks worth of discussion of a season.

20 weeks of discussion with 10 ballot threads open at a time would tend to lead to 2 weeks of decent discussion per thread. That method just dilutes attention from the most important season - the one up for election. I do not have time to keep up on 20 different discussions at once and I doubt anyone else has that kind of bandwidth.

A player's percentile rank is the only thing interesting to me when it comes to informing my HoM voting. Otherwise how do you compare MMP voting results across years? The value of an MMP vote should be constant if MMP vote "shares" are to be of any value for informing HoM voting.

20 weeks of discussion with 10 ballot threads open at a time

No, no. From [#26]: Not multiple elections; multiple discussions. Don't open another election thread until the previous election is over.

The idea is to perpetually have ONE ballot Election thread open along with the next ten Discussion threads. From [#19]: Voting is easy; analysis is difficult. This is why I advocate creating a wave of discussion threads, one for each season (say, 1961-70) before we ever begin voting. If you're like me, you do research when the time and the mood converge. For instance, maybe I'll post three years of preliminary ballots in one day, then go a month without giving the project much thought. By allowing a few months for each year to be open to discussion we have a much better chance to gather worthwhile analysis from people. So I think it behooves us to always have the next 5-to-10 years already under discussion. And, as Piehole points out, "other voters' inputs will make easy" the work of sorting things out.

percentile rank is the only thing interesting to me when it comes to informing my HoM voting. Otherwise how do you compare MMP voting results across years? The value of an MMP vote should be constant if MMP vote "shares" are to be of any value for informing HoM voting.

Since the project deals only with the peak of the talent period, absolute ranking is what matters, not percentile. The guys at the top are the same no matter the size of the league, so league size is irrelevant here.

In MLB they don't expand the size of the MVP ballot as the league expands; it stays at a constant vote-for-10. Likewise in this project, the ballot size should be a constant throughout the project. The 10th best player in 1879 should get as many points, as many MMP vote "shares", as the 10th best player in 1979. As with the HoM, "a season is a season".

Well, here are the thoughts of someone who was around at the beginning of the HOM and did not miss a single HOM vote. There is no "best" process, especially since it seems we cannot agree on what the objective(s) of the MMP project are.

If we are trying to determine the best player in each season, then a short ballot would work fine. If we are trying to determine the best 5 players in each season, then a medium ballot would work fine. If we are trying to determine the best 10 players in each season, then a longer ballot would be required.

Personally, I am extremely comfortable with a longish ballot. I plan on looking at many players each season and it would not be difficult to rank the top 20, say, especially if quadrant-voting was allowed.

By the way, I thought we were combining all major leagues into one seasonal vote, so looking carefully at 15-20 players each season does not seem like too much work. It may be interesting to see how many players on average receive MVP or Cy Young votes each season across both leagues.

I am hopelessly behind, but to piggyback on similar "Ripken-esque" Rob Wood's point...

I thought the original idea was just to vote on the true BEST PLAYER each year.
I figured that would be easy enough, as in many years either there is one obvious choice or 2-4 to analyze. So we'd rubber-stamp some previous picks, sure, but also thoroughly research the close calls. That seemed like a manageable effort, and not to take 10+ years or whatever.

I admire the ideas of trying to formalize a top 10/20/30 pecking order, but for me it's most interesting to grasp which players, or 2-3-4-5, dominated in a given year. And of course a point system can show which are true tossups and which are blowouts.

I thought the original idea was just to vote on the true BEST PLAYER each year.
I figured that would be easy enough, as in many years either there is one obvious choice or 2-4 to analyze. So we'd rubber-stamp some previous picks, sure, but also thoroughly research the close calls. That seemed like a manageable effort, and not to take 10+ years or whatever.

This is what I would sign up for. Well, I'll try to participate no matter what gets decided. I don't mind a long ballot, but anywhere from 10-20 sounds good to me.

A key issue seems to be the relationship between the MMP and the HoM, with some seeing the MMP as a supplement to the HoM; others view them as two separate projects.

The question is, in that case, whether a standard-length ballot (eg, 20 players every season, regardless of the number of teams) actually inhibits the ability to import MMP 'points' into a HoM ranking. I would have thought it was fairly easy for an individual HoM voter to construct a system that equalized the value of being 8th on a 20-player ballot based on a 16-team 1958 season and the same place on a 20-player ballot based on a 30-team 2008 season.

I don't think we are any closer to getting a consensus on the length of the ballot. However, if it's going to be fairly clear who the MMP is in many cases, then we are more likely to find the fun in working out who the Top Five, say, are. This does have implications about the length of the ballot. Perhaps we should look at it that way.

It does seem clear that there is no problem, in principle, with having multiple discussions going on at once. This is something we ought to work out through experience, I think. So I'll try to post three or four discussion threads in relatively quick succession, starting with 1961 tomorrow.

> multiple discussion threads

This proposal sounds like a person with attention deficit (like my son) would really enjoy it. Everyone else would be confused and the discussion would be diluted. Philosophical value discussions (pitchers versus hitters, rate versus playing time, etc) would be in 10 threads at once. I would rather not have more than 2 yearly discussions at a time - the current year and the next year. I do not believe having more discussion threads open for longer periods adds any value.

The writers should have expanded the ballot with league size. Otherwise you get people judging players based on whether a player ever received an MVP vote without adjusting for the fact that an MVP vote was easier to obtain in 1955 than in 2005. In my opinion this is a flaw to be corrected not a feature. You can't equalize for the fact that modern players didn't receive a vote at all. A 10th place vote in 1955 is equivalent to not receiving a vote at all in 2005. Conversely, if you truly only need 10 ballot slots to determine an MMP in 2005 then you only needed 6 to figure out the best player in 1955.

> a season is a season

This rings hollow. The purpose of the voting is to determine #1. That goal isn't compromised at all by a ballot length tied to the number of teams.

Philosophically, if we're trying to do something similar to the HoM - correct the BBWAA MVP voting using similar but tweaked guidelines - we should keep our ballot length approximately the same as theirs. BBWAA MVP voting has been locked in at 10 votes for many years. For this reason I'm opposed to a long ballot. I think a long ballot is trying to wrap up MVP, CY, AS and ROY voting into one vote. The best way to determine those winners is to have separate ballots for each of those items.

For this reason I'm opposed to a long ballot. I think a long ballot is trying to wrap up MVP, CY, AS and ROY voting into one vote.

Without looking into it at all, my intuition is that a 10-15 player ballot will capture both the MVP and the CY for every year. It will not necessarily capture the ROY. Part of DanG's rationale for the long ballot was that we'd see if a rookie was good enough to merit an ROY award. I'm fine with saying if a rookie doesn't show up in the top 25 players that he doesn't deserve an award. Voters can always indicate at the end of their ballot the top rookie of the year if they don't appear in the actual ballot. I'm sure that--whatever the length of the ballot--we'll be evaluating more players than we're actually voting for. In that case, just tell us who you think the best rookie is.

The response in [#66] shows general incomprehension to what I've proposed. If this is the state of things four years after I first presented it, I'm inclined to leave this alone and let the mods create whatever makes them happy. IMO, this enterprise is heading for an outcome that's only a step removed from the level of the BB-Ref Fan EloRater.

The response in [#66] shows general incomprehension to what I've proposed.

I don't think its incomrpehension, rather, a movement to take your idea and apply it in a feasible way.

I'm not sure HOW exactly it should tie in, but I agree with DL from MN that the ballot length should be tied in to the number of teams. I think (for example) being the 8th best player in a league of 16 teams is more impressive than being the 8th best player in a league of 6 teams (or whatever). Actually, if a short ballot is used I understand a fixed 10 every year (or whatever)... but if a long ballot is used I REALLY think it should be tied in with the number of teams.

As for DanG and 'zop... so far I think you're both in a tie for snarkiest comments on this thread.

I think (for example) being the 8th best player in a league of 16 teams is more impressive than being the 8th best player in a league of 6 teams (or whatever

Why? If MLB suddenly contracted to 8 teams, I don't think any MVP candidates would lose their jobs. It's the s rubs that would get the axe.

If MLB suddenly contracted to 8 teams, would Jose Bautista have had a job last season?

I can see two different factions here.

1) The purpose of this project is to determine the most meritorious player in a given year.

2) The purpose of this project is to rank the top 40 to 50 baseball players in a given year in order to determine the best player, best pitcher, best third baseman, etc.

I don't think #2 gets enough participants to make it worthwhile. I don't think it is a corollary to the MVP voting. I don't think you'll get better information than if you just did a weighted average of the various WAR calculators. #1 requires a consideration set of about 20 players. #2 requires a consideration set of about 100 players.

"The 10th best player in 1879 should get as many points, as many MMP vote "shares", as the 10th best player in 1979."

Why? The 10th best player in 1979 was a better ballplayer relative to his league than the 10th best player in 1879. Is the project about determining who was the 10th best player or who was the overall best player?

Maybe we should run a test. I'll volunteer to rank some players from 1961 over the next week or two. Say 25. I'll keep track of my time to see how long it took me. But I'm a newbie, so maybe someone who has a lot of experience should do it as well.

If the difference between ranking 10 or 15 players and 30 players is only an additional few hours, I won't object. If it's 5-10 hours, then I'll start to object. If it's an hour per player, I don't know that I can commit 25-30 hours of research for every ballot.

Complaining about all the extra work or saying the opposite--that it's not much more work--is probably pointless until we know how much work it will actually be.

Anyone up for the challenge? Or, barring that, can anyone chime in with how much time they actually spent on HoM elections?

1) The purpose of this project is to determine the most meritorious player in a given year.

2) The purpose of this project is to rank the top 40 to 50 baseball players in a given year in order to determine the best player, best pitcher, best third baseman, etc.

As a peanut gallery member, I think #1 is definitely the way to go.

I like a longer ballot for the reasons Dan has laid out.

I want a large electorate.

Incompatible? I don't know.

But if the idea is an inclusive project with "lots" of voters rather than "< lots," then a long ballot is not the only way to accomplish < lots of voters. Snobbery will do it as well. One of the really discouraging things in the later stages of the HoM project was the idea that if you don't use this or that specific version of WARP, you're a moron.

ie. If I don't get my way (e.g. 15) then I won't play. (God bless Brock who says, "I'm in no matter what.")
eg.

someone posts a well-reasoned 12 person ballot . . . and leaves off the #3 player by WAR, without mentioning the player

And which WAR is that?

Since I take the opposite POV--somebody said above that we want not only a large electorate but a diverse one in terms of how people think about and measure merit--and I agree with that. People can use a method that allows them to construct a long ballot, whatever that method is. If people are going to get jumped upon if it's not the politically correct flavor of WARP that month, well, we can force the long ballot to be a problem.

But, geez, in constructing a 15-man ballot, say, how many players do you have to realistically rate? 30? So how is it so hard to report back on 25 of the 30 instead of 15? If you're only rating 15, how do you know you have the right 15?

And do I really care to id. the best 2B and the best RF every year? No. But hopefully we can id. the best pitcher, at least. Do we need 25 to do that? Probably not. So in the last analysis, if it has to be 15, fine. I like 25. If it has to be 15, fine. But if I have to use some specific methodology (as opposed to the one I prefer) that is a powerful discouragement as well.

Yeah, but be real - the reason this project has stalled for so many years is that we don't have a compelling aim like with the HoM.

I don't agree with this. This project has stalled for 4 years because it did not have leadership. We couldn't get threads opened. That is a fact. Whether the idea is compelling or not remains to be seen.

BBWAA MVP voting has been locked in at 10 votes for many years. For this reason I'm opposed to a long ballot. I think a long ballot is trying to wrap up MVP, CY, AS and ROY voting into one vote. The best way to determine those winners is to have separate ballots for each of those items.BBWAA MVP voting has been locked in at 10 votes for many years. For this reason I'm opposed to a long ballot. I think a long ballot is trying to wrap up MVP, CY, AS and ROY voting into one vote. The best way to determine those winners is to have separate ballots for each of those items.

And I really disagree with this. The whole point of the MMP is to say that value is value and merit is merit and everybody should be in one bucket rather than having all of these specialized awards (players, pitchers, firemen, etc. etc.)

I'm not certain that a large ballot is necessarily incompatible with a large electorate. I am sure that a large ballot and a 2 week time period between elections is incompatible with a large electorate.

I agree that the project has stalled because it's difficult to open a thread around here. Compared to the SB Nation sites (for example) it is byzantine. I still haven't figured out how to do it.

I am standing behind Sunnyday when he mentions we want diverse viewpoints about value. People should be expected to explain their ballot choices and methods. There should be less controversy because we aren't comparing across eras but I expect good discussions on:

the relative value of pitching, defense and offense
league strength
replacement value in a given year
average value in a given year
rate versus counting stats
whether or not to consider postseason accomplishments
the relative location of various positions on the defensive spectrum
ballpark effects
how to handle the gaps and differences in the value calculators (catcher defense, relief pitching, etc)

For me, THAT is the reason I'm interested in the project. There should be plenty of points of view available to inform voting, none of them "wrong".

Well said, Marc. Let's hope it’s taken to heart.

I just want to clarify one item. From [#73]:

I can see two different factions here.

1) The purpose of this project is to determine the most meritorious player in a given year.

2) The purpose of this project is to rank the top 40 to 50 baseball players in a given year in order to determine the best player, best pitcher, best third baseman, etc.

IMO, neither one of these should be the primary purpose. The goal I envisioned is related to my 9th point in [#36]:

9 Create a peak value system for everyone in history, true MVP Shares

Above and beyond the two factions, we have the potential here to create an alternative career value system.

The idea is similar to the method used in Joel Whitburn's Record Research. He ranks records and recording artists across time simply by combining the weekly rankings from Billboard magazine.

I’m picturing we’ll eventually be able to make a chart like this for every player; Jim Wynn for example:


The dashes are years when he was not among the best players. This seems like a strong record, but until we have many players to compare him to we don’t really know if that’s a HOFer or not. And if we have a 15-man ballot we won’t be accurate (or even reflect) his standing in 1967, 1970 and 1975.

The more people we decide to rank each year, the more useful our results will be for this primary aim, because we will have accurate rankings of more player-seasons. The more data we compile, the better rankings we will create. We will then be able to make valid HOF arguments based on the wisdom of our crowd.

As usual, it seems that mine is a minority position, that this is a more ambitious goal than most potential voters want. If that’s true, that "utility of results" is being measured by the number of ballots cast, then the desire for something simple and easy is definitely in order.

sunnyday's post is well-written and logical. But I think it distorts the case against a long ballot. Responding piece-by-piece:

"a long ballot is not the only way to accomplish < lots of voters. Snobbery will do it as well. One of the really discouraging things in the later stages of the HoM project was the idea that if you don't use this or that specific version of WARP, you're a moron."

This is well put, and is precisely the reason why the long ballot is problematic. The longer the ballot, the greater the risk that people will take shortcuts and vote either on WARP or by plugging in numbers to their home-brew formula. Shorter ballots facilitate careful analysis.

I am not convinced that stretching out the time between elections fixes this problem, because the HoM history suggests that a very high percentage of voters make their selections at the last minute, regardless of how much time they're given to vote. (I know that no matter how much I want to vote in advance, I'm invariably preparing my ballot on the last day.)

Sunnyday quotes selectively from my post to make it seem like I'm advocating WARP voting, but I'm not - its actually the opposite.

He misinterprets my second point: that we need some sort of mechanism to identify players that must either receive a vote or a comment. This was a wonderful feature of the HoM - it allowed us to identify ballots where players were excluded in error, it often lead voters to consider players they would otherwise have ignored, and it shed light on each voter's philosophy.

In essence, for the HoM the electorate "self-nominated" the players that required comment in the next election. we don't have that luxury here because each election is quite independent from the last. WAR is a perfectly good proxy to use for purposes of determining which players are serious candidates for MMP in the contemporary era, but it breaks down as you go further into the past, particularly pre-integration. No one has proposed a solution to this problem. I think that's because some people envisioned a relatively small group of devoted voters discussing ballots extensively before a vote, with no last-minute ballots from people who lurk, but don't often participate, in the discussion threads. If you allow participation from outside the intimate HoM cocoon, the lack of a "requires comment" mechanism becomes probelmatic.

But, geez, in constructing a 15-man ballot, say, how many players do you have to realistically rate? 30? So how is it so hard to report back on 25 of the 30 instead of 15? If you're only rating 15, how do you know you have the right 15?


To construct any ballot requires rating more players than on the ballot. To vote 15 guys correctly probably requires rating 30. To vote 25 guys requires rating 40. And the real time-consuming part in voting, for me at least, is comparing each guy to the guys above and below him, considering if there are unsual factors that might lead me to change the order. That task scales up with the length of the ballot.
The only way a longer ballot does not become more time-consuming is if you create a formula, plug in numbers, and the formula spits out a list of players. If you vote that way - and I know many of the HoM electorate did- then it doesn't matter how long the ballot is because you just copy-paste a longer list from your spreadsheet, and all the time is invested upfront when developing/tweaking the system.

This project has stalled for 4 years because it did not have leadership. We couldn't get threads opened. That is a fact. Whether the idea is compelling or not remains to be seen.

Why did the project not have leadership for 4 years? Its a bit chicken-egg, but you don't think if there was some pressing project that people were chomping at the bit to vote on, that there would have been more momentum to reestablish leadership and get things moving? There's an element of DanG's "build it and they will come" here - in this case "start the project and we'll see if its compelling". Shouldn't we be trying to craft a project that is as compelling as possible?

And if we have a 15-man ballot we won’t be accurate (or even reflect) [Wynn's] standing in 1967, 1970 and 1975.

That's not quite right. The MLB MVP is a 10-man ballot, and usually ends up with about 25-30 players getting votes. If you have a 15-man ballot, you could end up with 40 players getting votes (and therefore, able to be ranked). (Of course, this depends on the number of voters involved.)

you could end up with 40 players getting votes (and therefore, able to be ranked).

Of course. But even if Wynn got a few tail-end mentions on a few 15-player ballots, it would result in a meaningless ranking. I would not put much reliability in the rankings below the early 20's under a 15-vote ballot scheme.

I agree with zop here about the ballot length. In my mind, the main problem with a long ballot is that it will require voters to make increasingly arbitrary distinctions as they get down to the lower-ranked spots--it is simply not possible to say with accuracy whether the 23rd best season or the 24th best season was the best in any given year. In the same vein, I don't think that there's much point in producing a group ranking of the top 20 or 30 seasons, just because once you get beyond 10 or so, the differences between each player are swamped by the margin of error. And of course, a shorter (<= 15) ballot will encourage more voters; I may have trouble participating if there is a longer ballot that takes a significantly longer amount of time (I'd like to participate in either case, I'm just worried about how much time I'll have to spend on each year's MMP).

I am a HoM voter who is very interested in participating in the MMP project. I haven't followed this discussion intently so I may have missed a few points.

1. The longer the ballot, the less likely I am to participate. I don't mind ranking 10 or even 15 players per year. But once we get beyond that, I have a hard time seeing myself putting in the effort or even submitting a ballot.

I understand that one of the arguments in favor of the longer ballot is that it would also reveal a definitive all-star team per season (we would know who was the highest ranked player at each position). In my opinion, that is a bad idea. Whenever you try to accomplish two goals with one project, you are likely to to accomplish neither of them well. Make it an annual All-Star project. Or make it an MMP project. But don't try to do both at once.

2. I notice that we're going with one ballot, rather than separate ballots for the AL and NL. Question: does that include all of organized baseball (in North America at least)? I would like to see Negro League players be eligible, even if it's unlikely that enough voters would embrace their stats or MLEs for then to beat out players from the majors.

3. I like the idea of a ballot that grows with expansion. I wouldn't want to match the exact number of teams for the reason stated above. A reasonable number might be half: 8, 10, 12, 16. However, if we adopt that approach and Negro League players are eligible, I would recommend that we add a couple of slots with the advent of organized Negro Leagues rather than waiting for 1961 expansion. That issue doesn't have to settled immediately, as that wouldn't effect ballots from '61 on.

4. Another question: I noticed that we're starting with 1961. Having not followed the discussion closely, are we moving forward from that year on an annual basis (61, 62, 63)? I saw some suggestions that we jump around more (61, 71, 81). I am heavily in favor of the former.

I like the idea of ranking just 10 or 15. That still leaves 25+ guys probably getting mentioned, and not sure how many more I need to see. Also not that intrigued by whether a guy with the 17th-most votes pts really would have finished 22nd if only we had 25 slots instead of 15.

I thought originally we might be "vetting" the actual MVP and Cy Young choices, often rubber-stamping and often having spirited debate and often rejecting the actual selection.

I'm ok with this variation, in part because we can illuminate some little-known facts about league strength (1950s AL? lol imo). I guess 15 better ensures acknowledging a couple of pitchers, too.

But simple seems good. 15 matches the HOM ballot, or with 10 we could vote MVP style of 14-9-8-7-6-5-4-3-2-1.

2. I notice that we're going with one ballot, rather than separate ballots for the AL and NL. Question: does that include all of organized baseball (in North America at least)? I would like to see Negro League players be eligible, even if it's unlikely that enough voters would embrace their stats or MLEs for then to beat out players from the majors.

From the 1961 thread and from DanG's proposed rules:

3) Criteria – Value to your MLB team(s) in that season only. If part of the season was spent outside MLB, that value should be considered as well, so long as most of his value was accrued in MLB (or top Negro leagues).
4) In other words, no consideration for players spending the bulk of the year in the minor leagues or Japan. There may be gray areas here regarding identification of certain leagues as “minor”: the AL in 1900, certain Negro leagues, the IA in the years around 1880, maybe the PCL for a time, the Mexican league in the 1940’s, maybe others.
5) Negro Leaguers – Yes. We will have to determine at what point Black ball ceased to be MLB caliber. We may have to allow for great players on barnstorming teams. Maybe credit for winter league play, but that’s a gray area, for sure.
6) War Credit – None. Ditto for injury credit. “Value in that season,” no what-ifs. The one exception might be for collusion credit, a la Raines in 1987.

I agree with this proposal, though I think that "collusion credit" and other such in-season credit (I think Urban Shocker might qualify due to a salary dispute) should definitely be given, provided that the player played most of the season. We don't want anyone playing half the year to win the MMP, regardless of whatever credit they deserve. Also, we need to determine exactly which leagues should be eligible among those in DanG's "gray area," though since DanG doesn't mention any minor leagues from 1961 forward, we may be able to hold off on deciding until later.

Negro Leaguers should definitely be included; this should be a requirement for voting. Hopefully, since we are doing the Negro League era after the present day, by the time we get to the Negro Leagues, we will have better Negro League stats available, allowing for better MLEs. Some Negro Leaguers did receive MLEs in 1961 (from the threads, WAR estimated by me based on batting line and defensive reputation):
Marvin Williams .262/.344/.409 3B/1B 1.9 WAR
Luis Marquez .158/.191/.222 OF -1.1
Luke Easter .258/.329/.420 1B 0.3

I also favor a 15-man ballot, as I think that 15 should be doable in terms of time, without being too small. And I agree with Chris Fluit that the ballot should expand with the expanding league.

I understand that one of the arguments in favor of the longer ballot is that it would also reveal a definitive all-star team per season (we would know who was the highest ranked player at each position). In my opinion, that is a bad idea. Whenever you try to accomplish two goals with one project, you are likely to to accomplish neither of them well. Make it an annual All-Star project. Or make it an MMP project. But don't try to do both at once.

This. I like the idea of an MMP project, and I also like the idea of an all-star project. But the voting should be separate, IMO.

In another vein, I gather we are going to continue to ignore Japanese baseball. Is that the consensus?

I think it would be a good idea to establish a constitution, similar to that of the Hall of Merit. Since the discussion on eligibility has garnered relatively little controversy, as compared to the ballot length issue (though that may just be a function of a weekend lull), here's a draft of the eligibility section of the constitution, based on DanG's post, and with some of my own ideas thrown in:

Any MLB player is eligible for the MMP award, as well as players in the top Negro Leagues and on the top Negro League barnstorming teams. Voters should consider the player's value to his MLB team(s) in that season only. If part of the season was spent outside MLB, that value should be considered as well, so long as most of the player's value was accrued in MLB (or top Negro leagues). In other words, voters should not consider players who spent the bulk of the year in the minor leagues or Japan. The one exception to this rule is players who were excluded from the majors in the 1950's and early 1960's due to MLB's gradual integration process (think Ray Dandridge, Bus Clarkson, Marvin Williams). No war or injury credit will be given, though voters may give "collusion credit," a la Raines in 1987, or otherwise credit players for labor situations outside of their control (Urban Shocker).

Obviously this isn't final, as I don't expect everyone to agree with me, but I think we should be looking to create something like this before we begin voting.

To settle the issue of ballot size, perhaps we should put it to a vote.

How many players should be listed on each ballot for the 1961 election? (List your top 3 preferences)

a) 10
b) 15
c) 20
d) 25
e) 30

Should the size of the ballot be constant or variable?

a) We should always vote for the same number every year.
b) We should consider the number of MLB teams and expand the number of players on the ballot as MLB expands.
c) We should explore a compromise between a and b. For example, one ballot size for 1871-1900 and a larger ballot size for post-1900.

For the first question my preferences, in order, are: e-d-c
For the second question I've mostly favored option a.

Any MLB player is eligible for the MMP award, as well as players in the top Negro Leagues and on the top Negro League barnstorming teams. Voters should consider the player's value to his MLB team(s) in that season only. If part of the season was spent outside MLB, that value should be considered as well, so long as most of the player's value was accrued in MLB (or top Negro leagues).

The second sentence seems contradictory with the first and third. Also, what's the rationale for excluding non-MLB players except in the case of the NgL and black barnstorming teams? I doubt many non-MLB performances are worthy of being on a ballot, especially if we end up with a shorter one, but if someone thinks one is and wants to vote for it, I don't see the harm. Yes, black players were kept out of MLB by discrimination, but some dark skinned Latinos (like Pedro Cepeda) choose to stay home and play in the leagues in their home countries rather than deal with American discrimination and play in the NgL. It's the same discrimination, so I don't see the point in legitimizing one brand of play and not the other. And, for what it's worth (and not to suggest it's on the same level of wrongness), Japanese players were essentially banned from MLB before Nomo exploited a loophole to come here.

For that matter, if you're looking to exclude all minor league performances, I think your proposal paints way too broad a brush. Under those rules, a voter wouldn't be able to vote for George Stovey's performances in white minor leagues (his 1887 performance in the International League looks quite good, without a deeper analysis), for Jose Mendez in the Cuban Winter League (or against barn-storming MLB teams) or Sal Maglie in the Mexican League (not a minor league at the time). Japan (and South Korea and Taiwan) have been so separate from MLB, until recent times, that I can see putting them off in their own area, even though I don't completely agree. That was the rationale in excluding them (along with post-Revolutionary Cuban players who played in the Cuban "amateur" league) from the HOM, along with the promise of a later International Wing, which never came. Plus, there was at least the idea there that the HOM was matching the HOF, which had allowed NgL players in, and even seemed to consider Latin performances in the case of players who had spent considerable time in the NgL (in the case of Dihigo and later Mendez), but not solely careers spent in Latin America or performances in any other leagues. But in this case, there's no precedent in the MVP for allowing non-MLB performance, and I think the minors/NgL/LA Leagues are too connected, especially pre-full integration, to easily separate without excluding worthy performances.

I'd much prefer to leave voting open to whoever anyone wants to vote for. I doubt many voters are going to want to look up NPB stats, and I don't blame them, but if someone does and wants to make a case for a candidate, I say go ahead. But if MLB performance is what we're supposed to be considering here, then that's fine. Then I think it should be the same eligibility rules as the MVP, and only MLB performance should be considered. No exceptions for NgL teams. Because I really don't see the point in allowing someone to vote for Josh Gibson in '37 for his performance for Homestead, but not vote for his performance in '41 for Veracruz. And if someone can vote for the later, why not Maglie 5 years later in the same league? Or even Fielder with the Hanshin Tigers?

To settle the issue of ballot size, perhaps we should put it to a vote.

How many players should be listed on each ballot for the 1961 election? (List your top 3 preferences)

a) 10
b) 15
c) 20
d) 25
e) 30

Should the size of the ballot be constant or variable?

a) We should always vote for the same number every year.
b) We should consider the number of MLB teams and expand the number of players on the ballot as MLB expands.
c) We should explore a compromise between a and b. For example, one ballot size for 1871-1900 and a larger ballot size for post-1900.

I'd vote for B on question 1, A on question 2, and include another warning that you're going to gut your electorate if you ask people to rank 25 or 30 players.

Another issue I have with the proposed rules is the language about considering the players "value" in "that season only". I intend to vote, as I did in the HoM, based upon what I believe to be "true talent" in the year in question - this means I plan to use seasons before and after the ballot year to smooth out flukes based on luck (my plan is to vote intelligently to avoid smoothing "real" improvement from season x-1 to season x).

It seems to me that the proposed rules, like the ballot size issue, represent another effort to narrow the new project to limit diverse participation.

I think 'zop fell off the truck in 91. This isn't a "most talented player" award. It is _all_ about which player had the best season. To think you can tell which seasons are a fluke and which are not from this distance is incredulous. I don't know anyone else who would consider voting like this so I am skeptical that this would narrow participation. If a player has a fantastic season, good for them. They should be recognized accordingly.

2. I notice that we're going with one ballot, rather than separate ballots for the AL and NL. Question: does that include all of organized baseball (in North America at least)? I would like to see Negro League players be eligible, even if it's unlikely that enough voters would embrace their stats or MLEs for then to beat out players from the majors.

I think it is clear NgL players are eligible but we've tabled the discussion of how to include them at this time. I don't believe we are considering Japanese players. I would like to limit the scope to North American baseball in the interests of making it simpler to complete.

3. I like the idea of a ballot that grows with expansion. I wouldn't want to match the exact number of teams for the reason stated above. A reasonable number might be half: 8, 10, 12, 16. However, if we adopt that approach and Negro League players are eligible, I would recommend that we add a couple of slots with the advent of organized Negro Leagues rather than waiting for 1961 expansion. That issue doesn't have to settled immediately, as that wouldn't effect ballots from '61 on.

If we decide on variable length ballots we should expand and contract based on the size of the Negro Leagues. There are some seasons with 4 major NgL teams and some seasons with 16.

4. Another question: I noticed that we're starting with 1961. Having not followed the discussion closely, are we moving forward from that year on an annual basis (61, 62, 63)? I saw some suggestions that we jump around more (61, 71, 81). I am heavily in favor of the former.

61, 62, 63, 64...

I think 'zop fell off the truck in 91. This isn't a "most talented player" award. It is _all_ about which player had the best season. To think you can tell which seasons are a fluke and which are not from this distance is incredulous.

Ok, then what about correcting for component park effects? Are we allowed to boost RHB in Yankee Stadium because of Death Alley? How do you handle the Gavvy Cravath types?

I can't tell -all- of what is a fluke and what is not, but I can sure take a stab at it. There are obvious BABIP flukes, both on the pitcher and hitting side, than you can easily identify and correct. Then there are more difficult seasons to analyze, like Norm Cash 1961.

I don't get how "I want to rank the top 15 players, as if I were conducting a draft to play the 1961 season over again" is inconsistent with the notion of a Most Meritorious Player. That a bunch of balls found holes for Jorge Posada in 2007 doesn't make him any more meritorious IMO.

If part of the season was spent outside MLB, that value should be considered as well

Nope. I'm not interested in enshrining that in a constitution. If people want to include minor league credit I guess I am okay with that. The player did play baseball to get those numbers. I don't want to require it of everyone. I'm fine with voters who insist on only crediting major league play.

Edited:
Any North American major league baseball player who has accrued a minimum of 120 plate appearances or 30 innings pitched in a major league is eligible for the MMP award including players in the top Negro Leagues and on the top Negro League barnstorming teams. Voters should consider the player's value to his MLB team(s) in that season only. If part of the season was spent outside MLB, that value may be considered as well, so long as the player meets the minimum playing time requirements.

How many players should be listed on each ballot for the 1961 election? (List your top 3 preferences)

a) 10
b) 15
c) 20
d) 25
e) 30

Should the size of the ballot be constant or variable?

a) We should always vote for the same number every year.
b) We should consider the number of MLB teams and expand the number of players on the ballot as MLB expands.
c) We should explore a compromise between a and b. For example, one ballot size for 1871-1900 and a larger ballot size for post-1900.

Question 1 go with option c, probably can handle that much research. Backup preferences are then b and d. Question 2 I'd prefer option a with consistent #'s but not a strong feeling in general.

Correcting for component park effects makes some sense though it will be difficult to come up with a good 1 year park effect. Docking a player because they took unusual advantage of their park doesn't make sense. Teams win ballgames by finding players with skills suited for their ballpark. It's better to adjust their contribution to the overall run environment.

How can you be you so certain that a season is a "BABIP fluke"? Why can't the answer be that a player was really good at making contact for a given season? Some years a player does a great job with a repeatable swing and sees the ball really well. Some years the opponents have lousy defenses. If you can adjust while using the context of just that season I'm fine with it. I see it as opposing the purpose of the project to dismiss a player's good year because they didn't do well in a different year. I also would see it as equally wrong to boost a player's season because he was really good the previous year and the year after. Why would you want to smooth out the variation? Vote for the player in his good years and don't vote for him in the others.

Can't say I'm exactly aboard with 'zop's idea of voting for "true talent" either. It's very much on the ability side of the ability vs value spectrum, and I certainly wouldn't vote that way. But if that's how he wants to vote... well, what's the harm in him doing so? If he's the only person that's going to do things that way, then it won't have a large effect on the vote. But if, on the other hand, others vote that way as well, then clearly it's an idea that's gaining some traction, and should be represented.

Voters are going to have different ideas on what merit is, or even if it's possible for certain players (such as minor league ones) to be meritorious. I say let the electorate decide on an election by election basis, rather than legislating certain unpopular ideas of merit out of the running.

Here'd be my rules on voting and eligiblity:

Ballots will consist of the 15 [or whatever] best baseball players in the world for the given election year. Anyone who appeared in at least one game during the calender year is elgibile, regardless of the level or quality of the league in which they played. Voters are free to come to their own conclusions on whether specific non-MLB play occured in a setting of high enough quality to merit placement on a ballot. Voters must be able to coherently defend their choices when questioned.

That strikes me as inclusionary as possible. And then if someone wants to vote for, say, Phil Rizzuto playing ball in the Navy and can present a good case for it, I'm all for it. I mean, the current rules being proposed are more restrictive than the HOM ones, which strikes me as moving in exactly the wrong direction.

I think it is clear NgL players are eligible but we've tabled the discussion of how to include them at this time. I don't believe we are considering Japanese players. I would like to limit the scope to North American baseball in the interests of making it simpler to complete.

Looking up recent NPB stats seems a lot simpler to me than trying to figure out what exactly Frank Grant did in 1902. To give one example. Anything other than "only MLB stats" is going to make things more complex. Is the average voter going to know if there's, say, someone who put up 55% of a really good MLB season in 1963 along with 45% of a great PCL season? I'm certainly not scanning through MiL encycopledias for every year to see if that's the case. But if a voter can present a case for a player like that, and win some other voters over, then cool.

Whether it's minor league credit, or NgL play, or play in the MxL, or play in Japan, it's a lot harder to pick out great seasons then it is great careers. Realistically, most voters won't do the research on their own every year to pick out players like that (most voters didn't do that for the HOM, from what I can tell, and that was an easier situation). It'll be up to voters who feel strongly about those cases to do the digging and present them to the rest of the voters. Or not. If no one is inclined to do that much work, that's fine, I wouldn't demand it of anyone. But I don't see why voters should be allowed to do that extra work for their pet projects in some cases and barred in others, when it's all going to make things more complex.

One of the cool things about a project like this is the different ideas people have of value or merit, and the things we can learn about baseball history in the process. I don't know jack about the NPB currently, and maybe I still won't after this project is over, but why eliminate even the possibility of it before the project begins?

Any North American major league baseball player who has accrued a minimum of 120 plate appearances or 30 innings pitched in a major league is eligible for the MMP award including players in the top Negro Leagues and on the top Negro League barnstorming teams.

This would exclude black players caught in the minors during the quota system of the 50s. And as Alex King pointed out up thread, Marvin Williams played in the Texas League in '61. Now, that was the last year of his career, he was pretty much done, and I certainly don't think he deserves to be on a '61 ballot, but I don't see why he shouldn't be eligible (especially if NgL performance counts) when some scrub back-up infielder in MLB qualifies.

I vote for 25 and constant size.

Voters are free to come to their own conclusions on whether specific non-MLB play occured in a setting of high enough quality to merit placement on a ballot.

You could run a project where everyone is free to make up their own rules. Usually this results in a mish-mash of methods often at cross purposes. (In the HOM we had a few of these discussions of freedom versus order.) In game design, it translates to an issue of detail versus playability. If you're aiming for a user-friendly design, the rules should be clear and straight-forward, not malleable.

IMO, it will be quite enough challenge for most voters to accurately rate the MLB players. Including everybody who played ball anywhere distracts focus from this effort. The consideration set should be strictly defined.

I'm fine with allowing any ballplayer but I'm still not interested in Japan. That works against the idea of doing this as a peak value exercise for the HoM by bringing in players ineligible for the HoM. MVP has no playing time requirement so I'm striking it:

"Candidate Eligibility: Any North American baseball player is eligible for the MMP award including players in the top Negro Leagues or independent teams. Voters should consider the player's on-field contribution to MLB team(s) in that season only. If part of the season was spent outside MLB, that value may be considered as well. However, the player's on-field contribution should be judged in relation to the highest level major league, not relative to a minor league. A season may include playoff or World Series games and but does not include spring training or exhibition games. No credit will be given for games not played due to injury, wartime service or contract disputes."

Two proposals - one variable and one fixed
"Ballot Length: The length of the ballot will be proportional to the number of major league teams with a minimum ballot length of 10. The ballot length shall be N/2 where N is the number of major league teams for seasons where there are more than 20 major league teams (1969-present)."

OR

"Ballot Length: The ballot length shall be 10 players."

I think a minimum of 10 folds in the excluded Negro League players with the smaller number of teams about as well as any other method. That gives enough slots to account for an additional 20% of baseball during segregation. I'm willing to hear more debate about a minimum size of 10 but it conveniently makes an MMP share about the same size as an MVP share.

"Voter eligibility: All voters must post a preliminary ballot in the ballot discussion thread at least 2 days before voting ends. All voters must fill out a complete ballot. Voters must briefly explain their ballot choices. One person, one vote; anyone determined to have voted with multiple accounts will be banned and their votes will be disallowed. The MMP ballot committee has authority to exclude any ballot that does not meet these requirements."

"Scoring: Points will be given in descending order with the #1 player receiving the same number of points as the number of ballot slots. For example, with a 10 player ballot the points will be 10-9-8-7-6-5-4-3-2-1. The player with the highest point total will be named the MMP for a particular year. In case of a tie, the tiebreaker will be number of 1st place votes. If the first tiebreaker does not determine a winner the player will be considered co-MMPs."