I for one do not object to provocateurs....

Nor I. I agree 100 percent. We need to break new ground, we need to be doing some original thinking, something that is a worthy followup to the HoM.

What does that mean? Well, the HoM is a really terrific thing, why? Because we replaced 225 of their guys with 225 of ours? No, that's not it. It's because of how we did it. IOW, we could have elected 150 20C MLers. And then we coulda elected some old-timers, separately. And then we could have elected X number of Negro Leaguers. But we didn't do that, we did it our way. We threw them all in one bucket and forced ourselves to find out how Alejandro Oms compared to Dom DiMaggio and how Lip Pike compared to Fred Lynn. It's because we went to the trouble to explore the Negro Leaguers beyond anything I've ever seen, and the 19C, and because we figured out that giving special credit for time missed while serving in the military, especially in WWII, wasn't special at all, it was just business as usual, if you wanted to be fair.

We set the bar very very high. And so just picking an MVP and a Cy Young winner for each league? Been there, done that. That's not what the world needs from the folks who brought you the HoM.

Sunnyday2 gets it right. Very very admirable project.

Homer:It's just a little dirty. It's still good, it's still good!
[Passes traffic, jumps a bridge and lands in the water.]
It's just a little slimy, it's still good, it's still good!
[It gets caught in a dam spillway, and when the pressure builds,
it shoots into the sky.]
It's just a little airborne, it's still good, it's still good!
Bart: [Crestfallen.] It's gone.
Homer: I know.

We set the bar very very high. And so just picking an MVP and a Cy Young winner for each league? Been there, done that. That's not what the world needs from the folks who brought you the HoM.

Well, if that's what people are thinking, let me throw a completely self-serving idea out there. . . . .If you're going to start the next elections come next year, why not try tackling managers for the Hall of Merit. Hey, my book should come out in December. . . .

There are 19 HoF managers (the official count is sometimes different based on how they clssify Hanlon/Selee or Wright, but it's 19 guys). You could look for the 19 best managers in baseball history using the same approach as the HoM did when voting for players.

I recognize this isn't the first time voting for managers has been pitched to this bunch, but I figure it can't hurt to just throw the idea out there.

It would still be cool to get all the voting results up on the site. Nobody has that data, huh?

Dag Nabbit's idea isn't a bad one, if you want an interim project while you work out exactly what to do next.

Create a 'manager's wing' of the Hall of Merit, with the same number of managers as are in the Hall of Fame. (Don't worry about guys who managed but were 'elected as players', just pick HoFers who managed for 'n' years.)

Then there will be debates over metrics, and it will be fun!

Whatever route you take, whether MVP/Cy Young or Most Meritorious Player or whatever, I feel obliged to point out that I found the two-week cycle too intense, as a parent with young(ish) children, and dropped out after about ten 'years'. You might get a better response with a three-week cycle, or even once a month. But that might be too infrequent for other types of voters.

Well, this debate is obviously not going to wait for my opus which was promised for November 1. So I'll throw some stuff out there myself.

Most Meritorious Player. That is the name. It's the Hall of Merit. Let's be consistent here and honor the Most Meritorious Player each year.

Again, I don't want to be seen as a naysayer or a provocateur. I'm just trying to create focus. Why do this?

Like we did in our attempt to outdo the BBWAA/Vet's Committee regarding the HOF, we would now try to best the BBWAA's efforts in regard to their seasonal awards voting, Dan. Sounds like a HoM thing to do to me.

Most Meritorious Player. That is the name. It's the Hall of Merit. Let's be consistent here and honor the Most Meritorious Player each year.

Going that route, we would have the Most Meritorious Player (MMP), the Most Meritorious Moundsman (MMM) and also the Most Meritorious Position Player (MMPP). Need to think about that a little bit more.

As for the Manager wing, we have talked about it in the past and I'm all for it.

Going that route, we would have the Most Meritorious Player (MMP), the Most Meritorious Moundsman (MMM) and also the Most Meritorious Position Player (MMPP).

Why all the divisions? In the HoM, did we vote for pitchers separately? Did we vote for Black players separately? Did we have a National League HoM and an American League HoM? No, we threw them all together and dealt with positional differences and league differences. Why not do that now? Because it's too hard? As sunnyday2 wrote: "We need to break new ground".

Why all the divisions? In the HoM, did we vote for pitchers separately?

Don't you want to know who we think are the best pitchers and position players for each season?

No, we threw them all together and dealt with positional differences and league differences. Why not do that now? Because it's too hard? As sunnyday2 wrote: "We need to break new ground".

Isn't that what the MMP is? With that said, I wouldn't mind a MMP, MMM and MMPP taken from all leagues for any given year. In fact, that would be awesome.

Don't you want to know who we think are the best pitchers and position players for each season?

Yes. Maybe we're talking about the same thing.

I'm saying that for each season we should have one ballot that ranks our top X players (number to be determined by this discussion) in MLB, giving us our MMP rankings. From that you could see the MMM and the MMPP. I thought that you were saying that for each season we have multiple ballots, splitting off pitchers over here, splitting off leagues over there, etc. Which is it?

It would still be cool to get all the voting results up on the site. Nobody has that data, huh?

I have an excel file with each year's voting results on its own sheet. If that's what you need, tell me where to send it.

I'm saying that for each season we should have one ballot that ranks our top X players (number to be determined by this discussion) in MLB, giving us our MMP rankings. From that you could see the MMM and the MMPP. I thought that you were saying that for each season we have multiple ballots, splitting off pitchers over here, splitting off leagues over there, etc. Which is it?

I was talking about multiple ballots like the BBWAA has, but I understand where you're coming from, Dan.

DanG #113
I'm saying that for each season we should have one ballot that ranks our top X players (number to be determined by this discussion) in MLB, giving us our MMP rankings. From that you could see the MMM and the MMPP. I thought that you were saying that for each season we have multiple ballots, splitting off pitchers over here, splitting off leagues over there, etc. Which is it?

Dan has made this point before in other contexts and I think he is right on. His way seems appropriate here although I warn that the group should be prepared to increase the ballot size X for recent seasons if it's important that MM Pitcher meet a fairly low standard of quality. (Not to mention if it's important to generate something like top three pitchers.)


107. fra paolo Posted: October 14, 2009 at 08:16 AM (#3351538)
Whatever route you take, whether MVP/Cy Young or Most Meritorious Player or whatever, I feel obliged to point out that I found the two-week cycle too intense, as a parent with young(ish) children, and dropped out after about ten 'years'. You might get a better response with a three-week cycle, or even once a month. But that might be too infrequent for other types of voters.

This may be a related issue.
Perpetual eligibility for a timeless Hall meant that regular HOM voters (not me) were refining their rankings of the same performances (ballplayer careers) over years of real time. What's distinctive and hard about ranking something like top ten season performances is that completing the task for 1893 doesn't go far toward completion for 1894.

I don't know if I like the one combined ballot all that much.

Is it going to tell us anything about the position players if we vote a bunch of pitchers at the top every year in the 1880s?

Is it going to tell us a lot about the pitchers if no one is willing to vote any of them higher than 10th once they start maxing out around 225 IP a year?

Will anyone even bother putting a reliever on a MMP ballot in the 2000s?

I'm more in favor of a MMPP, MMP and once we hit the 1940s maybe a MMR. I think we'd learn a lot more that way.

As John said earlier, we could always have a combined ballot also.

There will be a lot of work with this, in terms of tallying, but I think it'd be worth it.

As for the value of the project, I think there's a ton of value in saying this guy scored X career MMPP points. This guy was top X in the majors Y times during his career.

No one will vote simply by win shares but these data on win shares ratings of pitchers relative to non-pitchers may be relevant to the discussion

Pitchers stature by win shares

I have an excel file with each year's voting results on its own sheet. If that's what you need, tell me where to send it.

Holy Rizzuto! Please send to ryanjlind->gmail and I can get this website going as much as I can.

I don't know if I like the one combined ballot all that much.

Is it going to tell us anything about the position players if we vote a bunch of pitchers at the top every year in the 1880s?

Is it going to tell us a lot about the pitchers if no one is willing to vote any of them higher than 10th once they start maxing out around 225 IP a year?

Will anyone even bother putting a reliever on a MMP ballot in the 2000s?

Well, it seems some here are talking in terms of a 10 or 15 player ballot. I'm thinking in terms of a 25 or 30 player ballot. Since the players near the bottom have such small differences, they could be voted by quadrants. For example, on a 30-player ballot, rank the top 10 in order, then the rest in 5 groups of 4 (or 4 groups of 5) (or in ascending groups of 2, 3, 4, 5, 6 players). This way, every position is likely to be represented; there is sure to be a good number of both position players and pitchers no matter what era we're in.

I agree with Joe in #117. Though I'm not too interested in a MMR.

Joe has suggested that the depth of the ballot be flexible over time and be = to the number of "major" teams. There will be some work to determine how many major teams there are some years especially NgL years but it can be done. With a 30-32 man ballot some years I think you're going to get a MMP and a MMPP.

I sent a doc via Yahoo Groups. John and Joe, let me know if you don't get it.

I would argue for a constant ballot size, rather than one that fluctuates with the number of “MLB” teams. What does the number of teams have to do with identifying the best players in any one year? We’re talking about the top 20 or 30 players, the outliers. These would be basically the same guys whether MLB had 8 teams or 50 teams, right? The cream rises to the top no matter the width of the container.

In the HoM we’ve always been big on treating every era with equity. The 10th best player in 1879 should be worth the same number of points as the 10th best player in 2009. Under Joe’s scheme in post #26 this does not happen. The 10th best player in 1969 is awarded more points than the 10th best player in 1968 simply because there are more teams. How does that make sense? Doesn’t Joe’s scheme have us assume that the number of great players expands and contracts to match the number of teams?

In maintaining a constant ballot size we avoid the issue of identifying what teams are “major” and instead put the focus where it belongs, on what players are major. Hardy Richardson was a great player in 1878; the NL was a contracted six team league that year, so Hardy got his gig playing for Utica in the International Association. Was Utica a “major” team? Who cares, but Richardson was a “major” player and ought to be under consideration for MMP that year.

You might say, “Wouldn’t this open the can of worms of having to consider minor leaguers?” Ideally, yes; practically, no. We’re far from having good MLEs for most minor leaguers (not to mention Japanese, Cuban, Negro Leagues, etc). So, for practical matters, I would generally exclude minor leaguers from consideration. I would favor four exceptions: 1) for players prior to the founding of the AA in 1882; 2) for AL players in 1900; 3) for Black players in the periods they were allowed in the minors but largely barred from MLB, the approximate periods being1884-91 and 1947-53. 4) for players who earned significant MLB value in a season, their minor league play (MLE) could be considered also. I’m thinking here of players like Willie McCovey in 1959.

Joe Dimino: I'd like to go from the past to the present - but that definitely isn't a unanimous viewpoint. Heck it might even be a minority one.

I agree. I think past-to-present is preferable.

Joe Dimino: I would also like to see the voting be 'scalable' IE you vote for X number per year, based on the number of teams in the league, etc.. Not sure how well that idea goes over either, etc..

I see that some others are coming out in opposition to this idea but I like it. I think it makes sense to list a 8-person ballot in 1905, then slowly increasing to 14 and 16-person ballots for the present (or possibly 15 for each, even though the size of the two leagues are not equivalent). I disagree with the suggestion that that's somehow unfair to earlier eras.

However, we may want to smooth things out in the early going. There was a lot of fluctuation and we may prefer a steadier number of ballot spots even as the number of franchises bounced up and down.

Joe Dimino: Another question, do we want pitchers to be eligible for the MVP or just have a Position Player of the Year or MVPP and a Pitcher of the Year or MVP(itcher)?

I like the idea of having an MMP and an MMPP for each league and then one Most Outstanding Player award covering all of baseball. The MOP would then be the place where we would debate pitcher vs. hitter. It would also be the place where we would have arguments like Honus Wagner or Nap Lajoie, Mickey Mantle or Willie Mays and Barry Bonds or Alex Rodriguez. We would even have the opportunity in some years to argue Lou Gehrig or Josh Gibson.

That would give us five awards to hand out most years (and seven-to-nine for years with Negro Leagues and additional major leagues). One question I haven't resolved in my own mind is whether we would have one MMP and MMPP for all of the Negro Leagues throughout or if we would increase it to one for each league during the eras when there were two Negro Leagues with six teams a piece.

DanG: With the Hall of Merit we wanted to establish an alternative to the Coop that was built on reasoning and evidence, rather than ignorance and cronyism; to identify, as objectively as possible, who really are the greatest players,

I think that our rationale is very much the same. With the Hall of Merit, we inducted roughly 80% of the same players as the Hall of Fame. As you noted, it was the other 20% that was interesting. I think we would find the same insights and benefits in this same process. I wouldn't be surprised to find out that 80% of our MMPs are the same as the Cy Young winners. But that other 20% would be revealing.

The process would also demonstrate that the BBWAA has been as imperfect with their award selections as with their Hall inductions. Whereas the BBWAA often focuses on intangibles, clubhouse leadership, new presence on the team, playoff participants and the like, we would again look at reasoning, evidence and statistical analysis. We would be a corrective to the times that the BBWAA sometimes voted for the best story instead of the best player, focused on less-important statistics (like batting champions and win totals) or resisted giving great players multiple awards (one MVP each for Hank Aaron and Willie Mays?!?). I think it would be especially rewarding to honor Negro League MMPs and MMPPs and to put them up against white major league competitors as we did for the Hall of Merit.

Dag Nabbit: There are 19 HoF managers (the official count is sometimes different based on how they clssify Hanlon/Selee or Wright, but it's 19 guys). You could look for the 19 best managers in baseball history using the same approach as the HoM did when voting for players.

I'm all for it, but I don't think it makes sense as a side-project while we debate what shape the MMP project will take place. It probably only postpones any serious discussion. But I'm up for voting for managers whenever you are.

Hi Folks,

I realize I have been away for a while. But a full-time job, a night time Masters program, and an almost 4 year old just make time go away.
I like the single ballot idea the best. I have seen several "retro" MVP/CY/ROY tables already. Breaking new ground would be integrating players and pitchers from the different leagues.
I know there was significant debate about the value of pitchers independently compared with the value attributable to their defense during the 1870s and 1880s. While I wasn't a voter during the early days (missed the first year), I have read most of the threads thanks to archive.org. There was a lot of disagreement over how valuable a pitcher was at that time. I could see a voter making the argument that a pitcher who was a poor fielder and hitter and didn't strike many out (relatively) wouldn't be very valuable if there was a good defense behind him, even if he had giant IP totals. The game was changing so fast during the 19th C and yearly MMP votes could be a great way of tracking that.
Also, I think it would be interesting to see how different voters incorporated the changing defensive spectrum.
And there would be some interesting discussions about what was Meritorious in the first place. Is it the player who contributed the most to his team? What if the team won less than 35% of its games?
It would be interesting to see how voters handle disparities in league strength. Should it matter when you are determining merit? Does a great year in a bad league give you the most merit if no one has a great year in the better league?
Plus, there is the debate over how to measure merit - traditional stats, older advanced stats, BPRO, DanR., WPA, WS, etc.
Lastly, trying to integrate Cuban League, NEL, with white players would really need some deep thought.
I think a 15 person ballot would work at the beginning. There weren't 15 good pitchers in any one year so hitters would definitely show up on a ballot.
Anyway, those are my thoughts about why I like the MMP project the best. 2010 would be difficult for me to participate, but after that, it shouldn't be a problem because the Masters program will be finished.
Thanks for reading,
Kelly

I'm more in favor of a MMPP, MMP and once we hit the 1940s maybe a MMR.

We could start it in the 19th century instead. Don't know if we need one in 1871, however. ;-)

Glad to see that you're back, Kelly!

Thanks, John.

I can see the thought behind starting the elections with 1901 for various reasons - "Modern" baseball records dividing line, AL and NL both established, etc. It would give voters some time to get accustomed to the process while there isn't much controversy - no established Cuban or Negro Leagues to try to incorporate into rankings and the game is not changing as radically as it did in the 19th C. Also, people are more familiar with the greats of the Oughts and Teens - Mathewson, Cobb, Speaker, Lajoie, Wagner, Johnson, and Alexander, etc. than they are with Spalding, Wright, Barnes, and Ewing.
But, I think it makes sense to start with 1876 because baseball changes move along a continuum. The changes during the early 20th C started 25 years before. Also, it feels strange to start voting about a player in the middle of their career.
Also, I am not in favor of a separate MMR vote.

Kelly

Where will you be in 2018? Finishing up the MMP project! Assuming a triweekly election cycle and 147 elections (1871-2017).

I suggest we start with the first complete Retrosheet year, 1954 and work to the present. That's 59 elections (1954-2012). In 2013 we'll pause to catch our breath and see who has the stamina to do the rest.

I'm just sayin'.

As John said earlier, we could always have a combined ballot also.

We definitely need a combined ballot so we can include NeLers and certain MiL seasons and should be the focal point of the project. Still like setting up ballots for each league season, though, since we will gather more info regarding players left off the combined ballot. Besides, I'm in no rush to speed through the process. :-)

Whatever route you take, whether MVP/Cy Young or Most Meritorious Player or whatever, I feel obliged to point out that I found the two-week cycle too intense, as a parent with young(ish) children, and dropped out after about ten 'years'. You might get a better response with a three-week cycle, or even once a month. But that might be too infrequent for other types of voters.

No objections from me, fra paolo. If we can gain more voters that way, it would be foolish not to expand the length of the election then.

Even better, since the recent years are easier to deal with, use a biweekly election schedule for them.

Where will you be in 2018? Finishing up the MMP project! Assuming a triweekly election cycle and 147 elections (1871-2017).

If Joe or I are not here, someone else will take over then, Dan. Besides, I'll only be 53 then. IOW, I'm not leaving the planet anytime soon. :-)

I suggest we start with the first complete Retrosheet year, 1954 and work to the present. That's 59 elections (1954-2012). In 2013 we'll pause to catch our breath and see who has the stamina to do the rest.

The problem is that the more recent seasons hold less interest for me.

The problem is that the more recent seasons hold less interest for me.

I hear ya, John. But we may be in the minority.

There is a real world reason why we should do the recent years first: It might help to get someone elected to the HOF while they're still alive. There are two eras where most of the HoMer-not-HOFers come from, the really early years and the really recent years. The sealing up of the VC back door to the Hall after 2001 has left a lot of guys out in the cold. If our efforts serve to highlight the greatness of a Santo, Minoso or Simmons for everyone to see, we can feel gratified when we see them standing at the podium holding their plaque.

There is a real world reason why we should do the recent years first: It might help to get someone elected to the HOF while they're still alive. There are two eras where most of the HoMer-not-HOFers come from, the really early years and the really recent years. The sealing up of the VC back door to the Hall after 2001 has left a lot of guys out in the cold. If our efforts serve to highlight the greatness of a Santo, Minoso or Simmons for everyone to see, we can feel gratified when we see them standing at the podium holding their plaque.

You make a valid point, Dan.

I sent a 3 page prospectus for a MMP to John and Joe this morning. But this debate ain't waitin' around for nobody, so I'm going to post it here.

Part 1

Prospectus
Most Meritorious Player (MMP)

This is a proposal or prospectus for a new project under the auspices of the Hall of Merit.

Title/Concept

Identify the Most Meritorious Player (MMP) in baseball for each year from 1871 to the present in a manner that is befitting of the people who conducted the Hall of Merit (HoM)

Rationale

Using a wide variety of contemporary (and traditional) tools of evaluation, some 50+ voters have selected a HoM as an alternative to the Cooperstown Hall of Fame (HoF). After the Hall of Fame, baseballs’ annual awards—especially the MVP and Cy Young awards—are the game’s most prestigious individual honors. It is logical and reasonable, then, that we would address these annual awards as our second major project. In keeping with the Hall of Merit terminology, it makes sense for this project to identify the Most Meritorious Player or players of each year.

In thinking about the format of the MMP project, it is worth revisiting what made and makes the HoM so special. Already mentioned is the fact that it was a joint effort of 50 or 60 or more voters with a wide range of knowledge, perspectives and skills. As a result, the outcomes represent a consensus among a community of some size and of significant diversity of interests and perspectives. The results are sometimes controversial, but rank bias largely has been eradicated.

Second, the HoM project did more than just select a different set of honorees than the HoF. The HoF has selected players through a variety of disjointed processes and committees over the years—first tier twentieth century major leaguers through one process, old-timers and/or second tier honorees through others, Negro Leaguers through yet another. The HoM, on the other hand, integrated all potential honorees into one process with the result that players from the earliest days of “major league” baseball—Dickey Pearce, for example, and George Wright—are compared and compete directly for enshrinement with players from 100 years hence—such as Ozzie Smith and Alan Trammell. The Negro Leaguers go head-to-head with their Caucasian contemporaries—Home Run Johnson, John Henry Lloyd, Dick Lundy, Dobie Moore and Willie Wells with Joe Cronin, Joe Sewell and Arky Vaughan. Certainly this is more interesting and more fun for the participants, but the HoM also has played a seminal role in placing under-analyzed players—from the nineteenth century and the Negro Leagues, specifically—within an integrated analytical framework. The HoM made a strong and unambiguous statement about the relative merits of players from different times and places, and players at different positions (e.g. pitchers, relief pitchers, position players, etc.).

Third, the HoM has with consistency utilized a variety of other innovative analytical tools and perspectives, designed for the most part to facilitate fair and accurate cross-era comparisons:

• Development of Negro League (NgL) equivalencies, adjusting NgL records for strength of opposition and season-length considerations

• The integration of credit for time missed in military service into the career records of selected players, in particular (but not only) those of the World War II era

• Seasonal adjustments for short nineteenth century seasons (to a standard 154 or 162 game season). Some participants also utilized discounts for the American Association and other leagues thought to be inferior to the best baseball being played at the time.

• Extrapolation of minor league and winter league records to Major League Equivalencies (MLEs), and their consideration in the evaluation of selected players

The result of all of this is the most highly integrated scheme for evaluating and rating the “meritoriousness” of baseball players—from 1860 to the present, and incorporating the NgL from the 1900s to the 1940s, as well as performances in the winter leagues—that has ever been attempted.

Summary

The key elements of a Most Meritorious Player project will be widespread participation. This helps assure that we have consensus plus innovation. In order to get widespread participation, the project needs to move at a reasonable pace so that participants can see a record of completion and accomplishment building. And it needs to start at a time that the most possible participants are interested in. I doubt that that is 1871.

Ballot Structure

This is the single most important decision to make. It is a waste of time to talk about any other aspect of the project until this is decided. My recommendation is:

Like the HoM ballot, the MMP ballot should be integrated. In HoM voting, integration meant that all eligible players are evaluated and ranked on one single ballot—across all variables, whether of league, or of position. That means pitchers and position players appear on the same ballot. American (AL), National (NL)and Negro Leaguers, too, all on one ballot.

So, too, the MMP ballot should have all players at all positions and in all “major” leagues on one single ballot.

This may not be easy. But the point, as in the HoM project, is to force the participants to make the difficult judgments about the relative merits of pitchers and position players, and of Josh Gibson and Satchel Paige, for example, as compared to Lou Gehrig and Lefty Grove. The point, as in the HoM project, is that at the end we will have made judgments that others often have been and remain uncomfortable in making.

Simply picking an AL MVP, an AL Cy Young award winner (CYY), an NL MVP, an NL CYY, an AA MVP, an AA CYY, a NgL MVP, a NgL CYY will add very little to the body of knowledge—nor force us to continue to innovate and add to the analytical toolbox—of sabermetrics or of baseball generally.

This is not to say that the project would not identify an AL MVP, an AL CYY, an NL MVP, an NL CYY, an AA MVP, an AA CYY, a NgL MVP, a NgL CYY, etc., (or better yet, Negro American League [NAL] and Negro National League [NNL] MVPs and CYYs), as appropriate to each year. But it is to say that those award winners will be identified simply by their standing within a single integrated ballot. The highest ranking position player in the AL is its MVP (or, better yet, its MMPP). The highest rated NAL pitcher is its CYY, and so on. But, there is no separate vote for those honors.

The only question is how long of a ballot this would require. How deep of a ballot will it take such that we would be likely to have at least one pitcher and one position player from each league who would appear on an average ballot? There is merit (pardoning the expression) to the suggestion that the ballot might consist of as many positions as there are “ML” baseball teams in any given year. A similar suggestion calls for ten players per recognized “major league” operating in any given year. Either way, decisions will have to be made about the “ML” status of some leagues and teams—the UA of 1884, the Federal League, etc.—but those would seem to be relatively easy problems to overcome. Under this scheme, in the twentieth century, the ballot would start out with 16 slots, then expand as the NgLs come online. It would contract back to 16, then expand again beginning in 1961 to a total today of 32 slots.

Also, with the number of ballot slots changing over time and in order for voting to be comparable across the years, ballot values could be pro-rated from a value of 100 (#1) to maybe 10 (bottom slot), with the intervening values varying depending on the number of slots.

Summary: Trade-Offs

The trade-off that is involved is clear enough:

Multiple Ballots: Assures that we can name the top pitcher and the top position player, potentially in each major league, because we will vote specifically for that purpose. It is also more likely that we can also name the top player at various, if not all, positions include relief pitcher.

Single Integrated Ballot: Forces judgments about the relative merits of players across categories—pitchers versus hitters, MLB versus NgLs, AL versus NL, etc. etc. There could be gaps in the identification of some awards—e.g. there may be no pitcher from a given league in some certain season; that is, no NL pitcher among the top X players in the ballot results.

In the final analysis, Multiple Ballots are sure to accomplish their purpose (of being able to name the winners in various award categories) and sure NOT to accomplish the goal of addressing relative merits across categories. A Single Integrated Ballot is sure to accomplish the latter, and is likely to accomplish the former. As such, it is a vastly more robust tool and represents a vastly greater contribution to the advancement of the understanding of baseball history.

Timing

There has been considerable discussion of what years to begin the project with. There are objections to starting the project in 1871 or 1876 because some potential participants are lacking in knowledge of nineteenth century baseball. In part as a response to this, some have suggested going backwards from some date more or less in the present time. Along with the nineteenth century, the NgL era will probably be the most difficult period for voters to deal with.

The recommendation here is to attack the voting as follows.

a. 1901-present
b. Then, 1871 or 1876-1900

Going backwards is an uncomfortable proposition. Frankly, it abuses the laws of thermodynamics. It seems more natural to be able to track a player’s career in the same order in which it actually occurred. But starting in 1901 rather than all the way back in the 1870s seems like a sound compromise: It enables us to address more than 100 years of baseball history in a single forward motion, but starts the project at a time that should not discourage too many potential participants. And by the time we get to the NgL years (about a year into the voting in real time), most of the more mundane procedural debates should have been more or less resolved.

Management

Joe and John are willing to continue in their roles a la the HoM. In the HoM project, a key function was the determination of the year in which players became eligible and the publication each year of a list of newly eligible players.

A key in the MMP project will be the gathering and publication of NgL, winter league and minor league MLEs from various HoM pages to a single page where participants can gain ready access to that information.

I would have thought we could do it year by year, but have an MMPP, Cy Young and MMP for each election -- one lot of study, 3 ballots, 1-10 like the other guys do it.

However starting in 1901 suffers the dual disadvantage of making it difficult/boring for new guys to enter and being very irritating for those of us whose greatest interest is the 19th C. (why stop at 1871?) So I suggest we run time backwards, starting at 2009. There's no Coop to jog for this one anyway, but we should pick up lots of voters in 2009-1980, and then hopefully keep them going back. And I LIKE time running backwards; it means a gradual approach towards greater civilization :-)

Not sure whether I have the time/data/IQ to participate,(much busier than in '03, thank goodness, as it means $$$) but will certainly cheer the project on.

PS. On the question of Ballot Structure, I'd need to see a proposal of how Multiple Ballots would work.

A Single Integrated Ballot could be discussed for 1-2 weeks, with voting in the 2nd or 3rd week.

What's the schedule for Multiple Ballots?

Is it going to tell us anything about the position players if we vote a bunch of pitchers at the top every year in the 1880s?

Is it going to tell us a lot about the pitchers if no one is willing to vote any of them higher than 10th once they start maxing out around 225 IP a year?

It will tell us about the structure of the game of baseball at that time.

Anything but backwards.

So anyway, how many multiple ballots are we talking about?

If this is just a Hypothetical MVP/Hypothetical Cy Young Award, then we’re talking 4 ballots per year. AL MVP, NL MVP, AL CYY, NL CYY.

Maybe there's NgL awards, maybe not. If so, NgL MVP and CYY? Or NAL MVP and CYY, and NNL MVP and CYY?

Or are we also talking a Rolaids Award? Silver Slugger Award? etc.

How long does it take? Is there a Final Four? A final ballot that unifies everything so we know whether the MMP is the "MVP" or whether the MMPP is the real "MVP"? If there's no unifying ballot, then we won't actually be picking an "MVP." If there is, then that's going to take another week?

There is going to be opposition to a Single Unified Ballot (call it the String Theory Ballot if you like). But what is the alternative? So far, the alternative is "Multiple Ballots," but I don't know what that means.

Is it going to tell us anything about the position players if we vote a bunch of pitchers at the top every year in the 1880s?

Is it going to tell us a lot about the pitchers if no one is willing to vote any of them higher than 10th once they start maxing out around 225 IP a year?

I actually plan on not doing either of those scenarios, Joe. My ballots will be pretty balanced year-to-year.

Regarding the 2018 completion date, that's fine with me . . . the Hall of Merit was first proposed in the fall of 2001 and we starting voting Spring 2003 and wrapped up December 2007.

Life's a journey, not a destination . . .

Also, multiple ballots can be done just as easily as a combined ballot - actually they'd be easier, since you don't have to answer the pitching vs. hitting value question each year.

We could still submit all of the ballots at once, so 2 or 3 or 4 votes a year doesn't mean an extra week for each.

I could see the 1901 start time that's reasonable for sure.

I haven't had the time to read Sunny's whole thing yet . . . working on that now.

I would add though, that we don't have to go 1901-present, then 1871-1899 . . . we could always decide midway that we have enough info to be comfortable going back then.

I would also say 1892 would be a better start time than 1901. There's kind of a clean break there, a 12-team league, etc.

Great work, sunnyday. You're bringing me around to the single integrated ballot theory.

A few comments: I would prefer to start at 1871 and move forward, but I'm open to 1901 forward instead. Though I do agree with this statement: "anything is preferable to backwards."

You wondered whether we should expand the ballot for third leagues such as the UA and the FL. I would suggest against it. Though UA and FL players would be eligible for the integrated ballot (and could conceivably win), I don't think that such a huge fluctuation (16 spots one year, 24 the next, then back down to 16) would be wise. The one down-side is that we may not identify a position player and a pitcher from those leagues. But I guess I'm willing to risk it.

We could still submit all of the ballots at once, so 2 or 3 or 4 votes a year doesn't mean an extra week for each.

That's with no MMP, no final ballot to rank the mutliple category winners. If there's an overall ballot to rank the other winners, then there's extra time.

Also, when you say 2-3-4, then, it appears you mean a pitcher and a position player and maybe a reliever, with each category encompassing all leagues? I think John said he would include the Negro Leagues? I wish I knew what the Multiple Ballot option actually was.

Simply picking an AL MVP, an AL Cy Young award winner (CYY), an NL MVP, an NL CYY, an AA MVP, an AA CYY, a NgL MVP, a NgL CYY will add very little to the body of knowledge—nor force us to continue to innovate and add to the analytical toolbox—of sabermetrics or of baseball generally.

This is very well said and has me switching tunes a bit . . . but if we were to do this, I'd want a very deep ballot, at least 1.5 players per team. I don't just want to ensure we get the best player in each league, I would like a few of each in each league. This would be important so that if anyone wanted to go back through the individual ballots and construct individual voting they could.

And of course, we would have to actually come up with metrics that show how important pitching was vs. fielding, and defense vs. offense as well. Meeting that challenge would be great and would add knowledge I think.

Also, when you say 2-3-4, then, it appears you mean a pitcher and a position player and maybe a reliever, with each category encompassing all leagues? I think John said he would include the Negro Leagues? I wish I knew what the Multiple Ballot option actually was.

Yes, I was thinking a MMPP, MMP, MMR, overall. All leagues on one ballot. Have a 'crack committee' determine how many ballot spots we add in a particular year for the Negro Leagues, and how to phase in expansion in terms of ballot places.

I would agree with not adding spots for the UA but making their players eligible.

As far as the Federal League, who were the players? Did they pull from good leagues (besides the major leaguers), thus expanding the talent pool? Or did they just water down the majors without adding anyone worthwhile? We know there were plenty of solid players in the PCL at the time, for example. Were they pull some of those guys, or was it just AA quality talent?

Based on that answer I'd think we could make the case as to whether or not to add spots for the league.

I don't think the going backwards thing is a good idea.

If we do separate position player/pitcher/overall votes, then I would want those ballots to be made up of the players from all the leagues combined and not separate AA/NL/AL/NA/NEL. I like trying to figure out how to combine the leagues' players.

I agree with not adding spots for UA/FL players.

I think a ballot of 12 to 20 would allow for good depth of different thoughts about how to measure various players' merit.

If anyone wants it, I have a spreadsheet with league-by-league WS leaders from 1876 to 2003 or 04 (I don't remember). Starting with 1876, it has at least 10 players including the top at each position (I think), then gradually grows in length to about 25 or 35 depending on ties. Even though it's Win Shares, it's a good start to putting together a candidate list.

Hello friends.

Sorry to interrupt the current discussion, but I have an update about the website.

HUGE thank you to DanG for emailing me the full voting data! I have completed putting this into the database, and setting up a very preliminary player page where you can see voting history of the candidates.

The data was a bit messy, and my code is no doubt messier, so I can't guarantee anything close to perfection, and please feel free to point out any errors you might find, but I think it's a cool place to start. Check it out for yourself. Go to hallofmerit.com and search for a player in the left-hand box to see his voting history.

I have also written one "fuller" page to show sort of a "proof of concept" of my vision for this portion of the website. See Reggie Smith.

Hopefully this is the starting point we can use to really get something cool going here that will help HOM followers (like myself) easily find information we're looking for (no offense to BBTF.)

Ryan

I like the single, combined ballot option myself. Even though I fell so far behind the HoM balloting that I haven't voted in forever, I'd try to participate in this project. Which, as an added plus, would be easier to catch up on if I missed a year or two, since it's not cumulative.
I would also ask if we are going to include Japanese leagues in the MMP voting? I think that's the one great area we overlooked in the HoM, and including them now would be of great interest, and a great learning opportunity. However, I realise there are some obstacles to that (lack of data and information, chiefly) that I don't know how to overcome. But perhaps it could be done?

I'm OK with injecting the Japanese, and more recently Cuban players. We would not of course add any ballot slots for that. But if somebody wanted to argue that Sadaharu Oh in 1973 (or whatever) was among the top X players in the world and wanted to put him on their ballot, why not? We'd consider whatever data anybody could come up with and we'd talk about and do some research around relative strength of the Japanese league. This, in fact, would be exactly the kind of innovation that this project would need in order to be as valuable as the HoM.

My views, thrown into the mix:

One ballot, from there you can get all the MM(PP,P,R) combinations. We said with the HOM throw everyone into the same pool when analyzing a career, lets do the same with each season.

Vary the ballot size based on # of "major" teams is fine, provide a somewhat lengthy ballot. If you allow enough time for each ballot you should be able to do ample study and fill all spots.

Start back in time, move to present day. Start time doesn't matter as much. For me 1901 might be better, by then the AL & NL had separated enough from lower leagues to make merit easier to judge. For example, whose to say in 1875 the best player wasn't a semi-pro guy playing Philly independent leagues who cricketed in his spare time? The lack of statistical accounting at the time outside the one league makes it extremely hard. Also black players don't really come into play until the 20th C. and that's where adding NeL MLEs make for fun investigation.

I would agree with not adding spots for the UA but making their players eligible.

I think that's reasonable.

This is very well said and has me switching tunes a bit . . . but if we were to do this, I'd want a very deep ballot, at least 1.5 players per team. I don't just want to ensure we get the best player in each league, I would like a few of each in each league. This would be important so that if anyone wanted to go back through the individual ballots and construct individual voting they could.

If it were a deep ballot, than I would also be on board for a single unified ballot. 1.5 players per team would be acceptable to me. Less than that wouldn't be.

Someone please explain to me how a fluctuating ballot size is in keeping with fairness to all eras. In #124 I asked:

What does the number of teams have to do with identifying the best players in any one year?

(There's some other good stuff in that post, go back and read it if you missed it.) Aren't we working with the HoM mantra of fairness to all eras? The idea that great players are the outliers, that their number does not fluctuate greatly from one era to another?

Joe disagrees with me. In #33 he said:

yes, the 6th place guy in 1878 should get about the same credit as the 30th place guy in 2008. Not because of the game getting better, but because there are 5x the current major league players.

Remember, we're not just trying to identify THE best player, as in a normal MVP vote. The main thing, as I see it, is we're developing MMP Award Shares. This creates a system that rates every career. The down-ballot points have to be equal, in fairness to all eras. This is most easily, and I believe most fairly accomplished with a ballot of constant size and a constant point scale for every season. In this way, George Gore's season as the 10th best player in the game earns the same number of points as Todd Helton's season as the 10th best player in the game. Nobody has to stop and think, "Wait, how many teams were there? We'll have to adjust the total to see them fairly." Tenth best is tenth best; a season is a season; a pennant is a pennant. We can add more teams to the league, adding more players. So what? The 10th (or 20th, or 30th) best player is still the 10th (or 20th, or 30th) best player. Why would that standing change much? Treat every season equal, the same point distribution every year; disregard the number of teams, it's irrelevant.

This is most easily, and I believe most fairly accomplished with a ballot of constant size and a constant point scale for every season. In this way, George Gore's season as the 10th best player in the game earns the same number of points as Todd Helton's season as the 10th best player in the game.

I could not disagree more.

Being the 10th best player in a league with 54 starting spots is nowhere near the same thing as being the 10th best player in a league with 434 starting spots (30 team league, 8 PP/team, 14 DH total, 5 SP/team, 1 CL/team).

They aren't remotely close.

Not to mention that if you went back, you could always recompile the points as if the ballot size were static, even if you expand it based on teams. You can't do the opposite, if you don't gather the data.

If it were a deep ballot, than I would also be on board for a single unified ballot. 1.5 players per team would be acceptable to me. Less than that wouldn't be.

That's great news. I'm fine with a long ballot. Somewhere between 1 and 1.5. Maybe we should do a test ballot. How about 2009? And we see what result we get. Maybe we do it with a 1 and a 1.5 (a 30 and a 45).

I'm fine with a fluctuating ballot length. I agree 100 percent with Dan G. on the goal, but I'm not opposed to a fluctuating ballot length. It still achieves the goal if executed properly.

Remember, we're not just trying to identify THE best player, as in a normal MVP vote. The main thing, as I see it, is we're developing MMP Award Shares. This creates a system that rates every career. The down-ballot points have to be equal, in fairness to all eras. This is most easily, and I believe most fairly accomplished with a ballot of constant size and a constant point scale for every season.

I love the idea of career MMP Award Shares. To be honest I had missed that as a goal, but it is a terrific idea. But a fluctuating ballot length is not incompatible. I wrote somewhere, and maybe I never posted it, the following:

Use a 100 point scale. The number 1 player each year (1871, 2009 and everything in between) gets 100 points. The bottom player on the ballot each year (1871, 2009 and everything in between) gets X number of points (5 points, 10 points, 3 points, whatever). Everybody else in between in pro-rated based on the number of ballot slots. Some years it's 100-95-90, etc., some years it's 100-90-80, etc., some years it's 100-92-84, etc., some years it's 100-98-96, whatever. Something like that. Somebody who is better at math than me can compile some workable version of this or do it better. Maybe it has to be to four figures or two decimal places or whatever.

What this does is enable you to increase the quantity of players in recognition of an increasing pool size in fairness to eras when there are more players. But hold the quality at the top of the heap constant in fairness to eras when there are fewer players.

I'm not gonna say that

the 6th place guy in 1878 should get about the same credit as the 30th place guy in 2008

but something like that. And Joe is right, it's

Not because of the game getting better, but because there are 5x the current major league players.

So whaddya say Joe and Dan. Does that bridge the gap here?

And P.S. buried in that last post is the idea.

Let's do a 2009 MMP ballot as a test. We can always do another "official" 2009 ballot "in time." This is just a test. But, say, a 45 man ballot covering all leagues and all positions. We tabulate both as a 45 and also as a 30 and look at what the two different ballot sizes give back as results. Are there any relief pitchers among the top 30? Is there enough diversity at all the various positions? Or is 45 more satisfying?

I'll start.

1. Pujols
2. Mauer
3. Grienke
.
.
.
.
.
.
.
14. Papelbon
.
.
.
.
.
.
45. Jeter

How am I doing? (Actually I'm not ready with a ballot but you get the idea.)(Just kidding about Jeter 45th.)

How would you tabulate the two ballots so that the 1st and last places were both of equal value? I mean, 1st on the two ballots are equal to one another; and 30th and 45th on the two ballots are equal to one another; and everything in between in pro-rated.

Being the 10th best player in a league with 54 starting spots is nowhere near the same thing as being the 10th best player in a league with 434 starting spots

It's the same thing. Suppose you're drafting six APBA teams from all the players in the current MLB. The 10th best player is still the 10th best player; we've just cut out all the lesser players, those who wouldn't get any votes in this project anyway.

Or suppose MLB decided to expand to 60 teams next year. Assuming everyone plays at the same level as they did in 2009, the 10th best player is still the 10th best player; we've just added a bunch of lesser players.

Again, it's because all we're looking at is the players near the peak of the talent pyramid. These are the same guys whether it's a large league or a small league.

Not to mention that if you went back, you could always recompile the points as if the ballot size were static, even if you expand it based on teams.

This cannot be assumed. Call it the "small ballot effect". Suppose you have an 8-player ballot. Only players that some voter sees as one of the top 8 will get points, right? So, a player who would rate as the 20th best player if we had a 30-player ballot may gete no votes at all on that 8-player ballot; the short ballot is insufficient to measure the group's actual consensus of him. Instead, the #20 player will quite possibly be a player with very little (but unmeasured) support among the group, someone with a handful of extreme supporters. We saw this effect in the HoM voting with a couple guys, where a player was elected by a small minority, despite more voters ranking him nowhere near their ballot. The system was insufficient to measure the group's true ranking of the player.

The larger ballot is gathering opinions on more players, better reflecting the opinion of the electorate, giving us a truer ranking.

You can't do the opposite, if you don't gather the data.

Sure you could. Just re-tally the results cutting off the "extraneous" votes you don't want.

The down-ballot points have to be equal, in fairness to all eras. This is most easily, and I believe most fairly accomplished with a ballot of constant size and a constant point scale for every season.

Hasn't the number of HOM selections regularly expanded or contracted based on the size of the league?

I love the idea of career MMP Award Shares. To be honest I had missed that as a goal, but it is a terrific idea.

Thanks, Marc. In the thread "Once We Catch-Up: The Hall of Merit After 2007" I posted a list of some benefits of the MMP Project:

"By doing this, we’ll see many things:

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

I also worked up a detailed proposal. Post #145 in that thread mentions a lot of ideas and issues for this project.

It's the same thing. Suppose you're drafting six APBA teams from all the players in the current MLB. The 10th best player is still the 10th best player; we've just cut out all the lesser players, those who wouldn't get any votes in this project anyway.

About 60 players a year get MVP votes right now, give or take. That's with a 10 man ballot in each league (so really a 20 man ballot) and 60 voters (but half the voters are precluded from voting for half the players).

Or suppose MLB decided to expand to 60 teams next year. Assuming everyone plays at the same level as they did in 2009, the 10th best player is still the 10th best player; we've just added a bunch of lesser players.

1. That's not what happened. MLB has expanded from 12 to 30 teams over 117 years. Big difference.

2. Bill James has said this, and I agree - if you were to suddenly expand to 60 teams, the talent level would naturally rise and in 10 years or so the overall level of talent (average level of talent per team, if you will) would be no different than it is now.

It's the number of jobs that drives the talent level, not vice-versa.

We saw this effect in the HoM voting with a couple guys, where a player was elected by a small minority, despite more voters ranking him nowhere near their ballot. The system was insufficient to measure the group's true ranking of the player.

I disagree. If a minority thinks a player was great and should be at the top of a ballot, there is nothing wrong with their strong opinion having an influence on an otherwise fractured ballot. I don't believe this does anything but reflect the group's true ranking of the player - they are part of the group too.

Not to mention that our 5 point on ballot bonus would help mitigate this impact. If a player overcomes this, he must have a VERY strong minority supporting him.

I don't entirely get #164 . . . maybe I'm missing something.

I'm arguing for a large ballot, and feel like you can skim off the top if you only want the top 10 each year to be voted on - meaning you think 1876 should have the same number of ballot spots at 1976.

But I also think it makes zero sense to have 36 players on the 1876 ballot - that's more than half the league.

Repeating my post #26 here, in terms of a ballot structure that equalizes points across eras:

Sunnyday (or anyone else) how would you feel about 1 per team. So in a 12 team (majors) league we'd vote for the top 12. In the present day we'd rank the top 30.

That would help make it easier across eras. We'd make the points proportional, something like this:




The formula is pretty simple, basically a first place vote is 30 points, and then you go down each place based on this:

30-(29/(teams-1)

Then take the points and normalize to 30 voters. So if we have 40 voters, divide the final total by 1.33; if we have 20 voters, multiply by 1.5, etc..

Don't worry about the complexity, the formulas can easily be worked into a tally spreadsheet.

The advantage of scoring this way is so that vote totals will be comparable across time. A player getting 200 points in 1878 would be directly scalable with a player getting 200 points in 1978.

We could also work in the 'on ballot' or 'elect me' (only #1 here) bonuses as well, in the Hall of Merit tradition, one that has served us quite well, IMO.

But I do think it would be valuable to have a system that is comparable across the years.

Obviously, going with one overall ballot, I'd want to go to 1.5 spots per team.

I would also say that's a 'fungible' 1.5 spots.

Say expansion, we could phase it in at 1/3 of a spot per year or something. So instead of going from a 24 man ballot to a 27 in 1961 and a 30 man ballot in 1962, we could go something like:

1960-24
1961-25 (1/3 AL)
1962-27 (1/3 AL, 1/3 NL)
1963-29 (1/3 AL, 1/3 NL)
1964-30 (1/3 NL)

We did something similar with the Hall of Merit, the numbers there phased in expansion as well.

I'm arguing for a large ballot

OK, so am I. But I do favor the same size ballot every year. I'm interested in where a player ranks in his time and making an apples-to-apples comparison across eras. A season is a season. Clearly, that is a minority view of this project.

In the HoM, wasn't the number of electees determined by an estimate of the pool of players, not the number of teams? This idea is not necessarily incompatible with being fair to all eras. OK, I could perhaps compromise into taking a similar approach in this project. Only I don't think the size should change very often; and the difference between smallest and largest ballot should be fairly small. Here's one simple scheme that demonstrates this idea and seems to be along the same lines as the HoM election schedule:

Ballot size 20 from 1871-1900
Ballot size 25 from 1901-1960
Ballot size 30 from 1961-present
(perhaps Ballot size 35 from 1993-present)

1. That's not what happened. MLB has expanded from 12 to 30 teams over 117 years. Big difference.

2. Bill James has said this, and I agree - if you were to suddenly expand to 60 teams, the talent level would naturally rise and in 10 years or so the overall level of talent (average level of talent per team, if you will) would be no different than it is now.

It's the number of jobs that drives the talent level, not vice-versa.

Except we're dealing with the outliers, the peak of the talent pyramid, not the general talent level. And these are the same guys whether it's a large league or a small league. The 30th best player is the 30th best player no matter if there's 12 teams or 30. Perhaps this is what you're not seeing: in the small league #30 will be just an above average player; in the large league he will rank as an elite player. He's the same player, one of the 30 best; it's the league around him that's changed. That's why I don't see the number of teams in MLB as being relevant to anything here.

If a player overcomes this, he must have a VERY strong minority supporting him.

Only if the VERY strong negative opinions are not being measured. And I believe this did happen in the HoM, because the 50th place votes counted the same as the 16th place votes - zero. Players were elected despite having a lot of +50th place votes because this negative intensity was unmeasured by the format.

Frankly I like 1.5 better than 1. You'd be evaluating 45 players to figure out your top 30 anyway.

Like I said, why don't we do a 2009 MVP right now, soliciting 45 man ballots. We could tabulate the 45 but also tabulate a 30 to see if we're happy with the shorter ballot or not.

As to #171, yeah, that's what I was thinking. I think that makes lots of sense. Obviously we're not rating players all the way down to replacement, but in theory what we're saying is all #1s are equal (1871, 1971, whatever) and replacement is always equal (1871, 1971, whatever). It's just that sometimes there are more players in between. That's pretty defensible.

161. Joe Dimino Posted: October 16, 2009 at 12:44 PM (#3354032)
Not to mention that if you went back, you could always recompile the points as if the ballot size were static, even if you expand it based on teams. You can't do the opposite, if you don't gather the data.

164. DanG
> Not to mention that if you went back, you could always recompile the points as if the ballot size were static, even if you expand it based on teams.

This cannot be assumed. Call it the "small ballot effect". ...

... The larger ballot is gathering opinions on more players, better reflecting the opinion of the electorate, giving us a truer ranking.

> You can't do the opposite, if you don't gather the data.
<

Sure you could. Just re-tally the results cutting off the "extraneous" votes you don't want.


Gentlemen,
You both make the same point because the situation is symmetric. It isn't possible to square this circle. (except perhaps by ranking every player-season, which is very far from both suggestions)

Dan,
Joe has suggested ballot size 1 per team but he is now entertaining proportion 1.5. That will be 45-deep for 1998 to 2009. Your 30-deep ballot will not generate the data Joe wants after 1968, the last season with 20 major league teams.

Joe,
You say you support a "large" ballot but Dan has suggested constant size 30. You are now entertaining 1.5 slots per mlb team, which will be merely 18-deep in 1899 and 24-deep for the next 60 years. Your 1.5T-deep ballot will not generate the data Dan wants before 1962, the first season with 20 major league teams.

How about 1.5 with a minimum, whichever is greater?

I feel obliged to point out that I found the two-week cycle too intense, as a parent with young(ish) children, and dropped out after about ten 'years'. You might get a better response with a three-week cycle, or even once a month. But that might be too infrequent for other types of voters.

How about biweekly elections with an extended discussion cycle? Unlike the HoM project, the next election doesn't depend on the results of the current one. Each MMP election stands alone. That means we have the option of having more than one election discussion going on. I think somewhere I suggested that we post all sorts of stats and leader lists at the start of each discussion thread; and to post the discussion thread for each year many months ahead of the election thread for the year. Having them up for an extended period enables others to contribute information and studies and dicoveries as they're uncovered, rather than trying to cram eveything into a two or three week discussion. Some seasons won't need much discussion to sort out the leaders; other years will have all sorts of complications to consider and will benefit greatly from longer discussion. In some cases, such as league quality adjustments, one season will have a relationship to surrounding years that will be better served by having discussion for those years taking place simultaneously.

I think we should always have discussion threads up for the next ten elections, giving us nearly five months to analyze a season. For those concerned about dividing our focus, I don't think that's a problem. I think that focus will tend to go where it's most needed; on the year of the current elecion and on years with particularly interesting issues.

As soon as we decide on a starting year, we can start posting discussion threads for seasons. I still say start in 1954. We have the situational breakdowns starting in that year, we don't have to deal with Negro leagues, and it's relevant for dealing with the cases of most living HOF candidates. Recent years are more likely to draw interest from those outside of the BBTF universe. Also, it gives us time to best decide how to deal with the complications and data shortages of the earlier eras.

I'm certainly more interested in developing a method based on more recent years and backwards applying it. Maybe the solution if there are going to be different years discussed is to split the eras and have recent years on a faster election timeline than those 100 years ago. It is far easier to determine a list of the 25 MMP in 1990 than in 1890. Perhaps work from both ends and meet in the middle? 2009->1954 could have a 2 week election cycle and 1873->1953 would have 4 week cycles. The whole project would be done in 80 months but you'd have recent (and ancient) history done in 2 years. Plus, I could opt out of discussions on pre-modern baseball from 1873-1900 and still participate in the project. That would leave the trickiest part (1901 to 1953) with the Negro Leaguers until the end. By that time the HoF might have released the stats.

I dunno about TEN elections at a time, but certainly there can be more than one. If we did 3 week cycles (2 weeks of discussion and 1 week to vote), we could certainly start the next three week cycle during voting. Many of us voted on day 1 anyway.

Week 1--Discuss year 1 (e.g. 1954)
Week 2--Discuss 1954
Week 3--Vote 1954, Discuss 1955
Week 4--Discuss 1955
Week 5--Vote 1955, Discuss 1956
.
Week 7--Vote 1956, Discuss 1957
etc.

You'd be knocking off years at a 2 week cycle but with 3 weeks to address that year.

The alternative is to introduce a new year every week. e.g.:

Week 5 Vote 1955, Discuss 1956 and 1957
Week 6 Vote 1956, Discuss 1957 and 1958
Week 7 Vote 1957, Discuss 1958 and 1959
Week 8 Vote 1958, Discuss 1959 and 1960

Now you've still got 3 weeks to address each year but you're doing 52 years in one real year. I think this is probably too fast, however.

But I don't think voting for more than 1 year at a time, or going backwards is very appealing.

I know you don't like going backwards, it probably doesn't matter but I think we could do two different elections. One starting with 1954 forward and the other from 1871-1900 with the modern elections on a faster timeline and 1901-1953 covered last.

I dunno about TEN elections at a time

No, one ELECTION at a time, open for two weeks. But the next ten discussions would be going on along with it.

DL's idea works for me, but I can understand objections to it.

This bears on segmentation and pertains to DL's middle segment, which he would postpone to the end of the schedule, 1901-1954.

The biggest problem I see, posing a challenge maybe impossible to overcome, is the meager data available for analysis of Negro Leagues and pre-Negro Leagues blackball. By ingenuity and hard work Chris Cobb used those data to assess careers remarkably well. He presented that work in terms of single-season MLEs but its adequacy for the HOM elections did not depend on adequacy at the season level and there is no reason to suppose the single-season MLEs are remarkably good in the same sense as the career MLEs. On the contrary, the crucial method is extrapolation of a featured season historical record (h.r.) by regression to the mean of several surrounding seasons h.r. Almost everyone here is familiar with one consequence: the Cobb MLEs show career patterns that are remarkably consistent in comparison with the incomplete historical record (which fits the purpose of regression) or with complete major league careers (which practically disables the "non-consecutive peak" approach for blackball players). The MLEs lack peaks and troughs, especially scattered excellent seasons or scattered dismal seasons except for known injury.

What methods may be available for assessing single-seasons adequately for the purpose of rank-order MM ballots?

What methods may be available for assessing single-seasons adequately for the purpose of rank-order MM ballots?

This is exactly why it's a good idea to save these seasons for last, to get people thinking about and developing tools and approaches for dealing with these players. Impossible challenge? We do the best we can, as before. Hey, fools rush in....

Some thoughts:

1) Clearly, we would want to devise MLE's for a lot more guys. Probably a multi-year project in itself.
2) Review the MLE's that have been done to better ascertain value in prime seasons.
3) I think we'd have to make more assumptions based on age, especially as regards fielding values.
4) It would behoove us to undertake a thorough review of the anecedotal record to inform the analysis.

Of course, we always have the option of setting aside the Negro leagues for this project. Assigning value to players without having numbers could rightly be seen as a creative exercise that has no real place in an analytical process such as this. Or perhaps we would only want to consider them in years where there were actual "Negro leagues" in operation.

A dilemma for sure. If we can't get Chris Cobb back in here at something nearing his level of involvement previously, I admit we're in trouble. But putting aside the NgL is not an option IMO. That is precisely what makes this just another Hypothetical MVP project. Ho Hum.

Horrifying I know, but it seems to me that the answer to Paul's comments is anecdotal evidence, as Dan said. If the MLEs show that Satchel Paige was just the same for 5 years, the anecdotal evidence will not corroborate that, it will clarify the peaks and troughs. So we're talking supplemental methods, new thinking. But frankly, I find that to be a lot less daunting than what we did in the HoM was (daunting) at the outset of that project.

I'm OK with 1954-present, then either 1871fff, or then 1901-53, and then 1871-1900. Starting in 1954 has a variety of benefits--probably more participants at the start, more data include Retrosheets, etc. etc. But then I'm willing to start anytime, but just let's start!

I've said it before, let's do a 2009 MVP ballot and we can tabulate the votes in various different ways, just to uncover any effects that we may like/not like of counting in those ways. I haven't done one myself yet.

OTOH how about a 1934 ballot for the same purpose, but also meant to re-familiarize everybody with the NeL data. Following is 1934 data for NeLers who are in the HoM (only). Just for the record, most of the NeLers we've elected primed in the 1920s and a bunch of them "retired" or were getting a tad long in the tooth by 1934.

Pitchers

Ray Brown 13-4, 6.4 WAT; MLE 23-10, 2.86, 157 ERA+, 34 WS
Willie Foster 6-5, 2.92, 1-2, 2.00 in playoff; MLE 12-12, 4.56 in 204 IP
Satchel Paige 20-5, 1.85 or 13-3, 1.99 or 12-1, 2.36; MLE 20-12, 3.52, 128, 25 WS

Position

Josh Gibson 16 HR (1st), 2.95; MLE .289-.381-485, 130, 17 WS
Biz Mackey .299 as utility player; MLE .277-.350-.338, 85
Buck Leonard .330 in rookie year; MLE .326-.424-.463, 138, 22.5 WS
Mule Suttles .254, 12 HR (3rd); MLE .260-.446 SA, 15 WS
John Beckwith .286; MLE 3 WS
Jud Wilson .333, 10 HR (5th), 1B at age 37; .312-.411-.408, 128, 20.5 WS
Dick Lundy .279 at utility player; no MLE (would have been out of MLB)
Willie Wells .264, 10 2B (3rd), 8 3B (1st); MLE .266-.348-.396, 99, 17.5 WS
Cool Papa Bell .364, 5th in BA, 1st in SB; MLE .332-.407-.378, 112, 22.5 WS
Oscar Charleston .285, 12 HR (4th), 1B; .313-.384-.497, 134, 20 WS
Turkey Stearnes .398, 13 HR (2nd), 13 2B (2nd), 5 3B (2nd), CF, MVP; MLE .371-.444-.612, 176, +9.2 WARP, no WS

The only guys who had among their better years were Stearnes and Brown. Paige and Bell might be deserving of some support. Otherwise, one would have to dig into some of the guys whom we have not elected to find some more candidates. But I think Stearnes would be top 5. Meanwhile Brown and Paige would have to compete with Dizzy Dean 30-7, 2.66 in 311.2 IP.

I agree, a 2009 MMP ballot is timely and lets us work out some bugs. Can someone open a thread?

I'm with DL. Joe or John, could you open a 2009 MVP thread, or an MVP Project Test 2009. I'll transfer this over when it's there. Nobody's been clamoring for more than 1.5 players per ballot so 45 names it is.

2009 MVP Ballot All Leagues

1. Albert Pujols
2. Joe Mauer
3. Prince Fielder
4. Zack Grienke--I don't usually have a pitcher up this high due to low IP, but there were some great pitcher seasons this year
5. Tim Lincecum
6. Mark Tiexiera
7, Chase Utley
8. Derek Jeter
9. Hanley Ramirez
10. Evan Longoria

11. Adrian Gonzalez
12. Ryan Braun
13 (tie). Chris Carpenter
Adam Wainwright--there are plenty of arbitrary placements on this ballot but splitting up some of these guys would really be really arbitrary so, sorry, there are ties. Typically they're teammates or guys at the same position.
15. Ryan Howard
16, Miguel Cabrera
17. Kevin Youkilis
18 (tie). Andre Ethier
Matt Kemp
19. Jason Bay
20. F. Hernandez

21. Pablo Sandoval
22. Roy Halladay
23. Aaron Hill
24. Ben Zobrist
25 (tie). Bobby Abreu
Chone Figgins--this seems low for a playoff team MVP but tell that to the Rockies
27. Dan Haren
28. Derek Lee
29. Alex Rodriguez--this sure seems low now but no peeking! No playoff credit!
30. Troy Tulowitzki

31. Victor Martinez
32 (tie). Robinson Cano--you Yankee fans can split these hairs
Johnny Damon
Mariano Rivera
CC Sabathia
36. Henry Votto--who?
37. Ichiro Suzuki--! to you, too
38 (tie). John Jurrjens
Javier Vazquez
40 (tie). Joe Nathan
Jonathon Papelbon

42. Kendry Morales
43. Jayson Werth
44. Raul Ibanez
45. Matt Holliday

Note that I've got 4 AL and 6 NL starting pitchers, and 3 AL and 0 NL relievers. And there's no catcher from the NL and no CF from the AL (unless you consider Ichiro! to be a CF!). Sorry, Torry. What kind of ballot construction would it take to be sure to get every position on there?

And there's no catcher from the NL and no CF from the AL (unless you consider Ichiro! to be a CF!). Sorry, Torry. What kind of ballot construction would it take to be sure to get every position on there?

I think it falls into the category of "things not worth worrying about". In this case, the finding is that there were no NL C and no AL CF among the outstanding players of 2009. I think it's enough in these cases that we say it was a bad crop; I don't think we need to be concerned about identifying the best of a bad lot.

I think it falls into the category of "things not worth worrying about".

Actually, I agree. I was being silly or sarcastic or...ironic, that's it. I was being ironic. But seriously, the perfect is the enemy of the good and I think we've been trying to build the perfect mousetrap. There's perfectly serviceable stuff out there to catch the little rodents with already.

Man, I'm dyin' here.

I agree too on the things not worth worrying about.

Do we want a 2009 test ballot to get in the way of the actual HoM election that is coming? Can that wait?

What is the timing of the HoM election? We could certainly do both before December 31. Let's just establish a schedule for the MVP 2009 vote (and the HoM 2010 vote) and open up a 2009 MVP discussion thread. I think we can multi-task our way through that.

We want the HoM election to finish the Monday before Thanksgiving (11/23). I'd rather focus solely on that until then, was planning to post a discussion thread this weekend with voting to start 11/9.

Are we really going to keep electing three guys a year, no matter how many the actual Hall takes??

Why wouldn't we elect 3 per year? The actual Hall will unblock that flow sometime later when today's players are old enough to have "played the game the right way, unlike those kids you see today".

DL is right. Not to mention that we established a pace that was in line with the historical Hall of Fame through 2002. If the Hall of Fame is going to all of the sudden change their own standards, much to the detriment of honoring the recent generation, I don't see how that impacts us.

Not to mention that they did elect 3 last year.

If anything our original plan was to be electing 4 every fourth year sometime around 2010, IIRC, so we already scaled that back slightly (although that wasn't in response to the HoF electing fewer).

197. Joe Dimino Posted: October 27, 2009 at 09:24 PM (#3368050)
If the Hall of Fame is going to all of the sudden change their own standards, much to the detriment of honoring the recent generation, I don't see how that impacts us.

The Hall didn't change its standards, only its constituency.
The living Hall of Fame players climbed a rung or two.

We want the HoM election to finish the Monday before Thanksgiving (11/23). I'd rather focus solely on that until then, was planning to post a discussion thread this weekend with voting to start 11/9.

Joe, OK, that's fine. I had kinda forgotten about that.

How about in keeping with our proposed or tentative timeline for the MVP project (3 weeks/ballot) that at least as a test case that we try to do a 2009 MVP in 3 weeks. To wit:

2009 MVP discussion thread open 11/23-12/7
2009 MVP ballots 12/7-12/14 to finish up well ahead of the December holidays
Discuss 2009 MVP ballot procedure, fine-tune ongoing procedures 12/14-1/2 3 weeks
Discussion thread for (what year?) 1/2-1/16
First (or second, counting 2009) MVP ballot 1/16-1/23

Obviously the timing piece proposed here is still open to discussion during that 12/14-1/2 period, same as any other procedural aspect.