moving this to the hot topics . . .

If I understand the concept right, this discussion is about how we each vote, how we incorporate new information, what new information we would like to have, timelining, position "quotas", etc.
If this is right, I am very interested to know how people balance peak and career, how people incorporate eligible players who are statistics deficient (1860s players and NegroLeaguers), how people rank pitchers vs. everyday players.
If I'm wrong, please ignore this.

Oh, and what statistics do you consider?

I don't know how anyone got the idea that what we are doing here is "scientific". As I said several years ago when we were discussing voting structure, the economist Kenneth Arrow scientifically proved that when there are more than 2 choices, none of which garner 50%, it is simply impossible to express a majority will (or, rather, that there are numerous ways to express a majority will, none of which is more valid than any other.)

To say what we are doing requires some sort of scientific accuracy gives more credit than I think is possible. No one doubts that, say, Dickey Pearce was more valuable to his team than Bobby Wallace was to his. No one doubts that, if Pearce was magically transported ahead in time 40 years, he would not have contributed as much as Wallace did. I am sure everyone understands both sides of the timelining issue -- they just reach different, reasonable conclusions about it.

Marc suggested that "New information is quite often completely ignored." I do not see evidence of that. In terms of integrating data, here are the number of ballots "controversial" candidates like Pearce and Caruthers appeared on at regular intervals (starting in 1899, the first year both were eligible):

1899: Caruthers: 23/31 = 74%
Pearce 3/31 = 10%

1909: Caruthers: 10/43 = 23%
Pearce: 10/43 = 23%

1919: Caruthers: 35/47 = 74%
Pearce: 20/47 = 43%

1928: Caruthers: 34/47 = 74%
Pearce: 30/47 = 64%

So, far from ignoring the stats presented, it looks to me like 3 proponents of Dickey Pearce managed to sway over half of the electorate to their point of view. Caruthers, meanwhile, got stricken from over a dozen ballots before being put back (or put anew) later on.

That looks to me like not only can the electorate be convinced, but they can be re-convinced.

Marc also suggests: "What's needed therefore is information, more information and especially new information. Advocacy is not welcome."

In my view, full treatises cannot be put together in a couple of weeks. The top carry-overs for 1929 are Bobby Wallace and Jimmy Sheckard. They are both strong candidates (otherwise they wouldn't have the support that they do), and it is certainly reasonable to suppor them. Jimmy Sheckard, however, is not on my ballot (and reasonably, I think, since he is also not on the ballot of 10 other voters). When faced with 7 days until the next election, should the goal be to present a full flowchart showing the pros and cons of Jimmy Sheckard? I believe that the best course is to shoot out 10 reasons NOT to vote for him (e.g., his stats are indistinguishable from Jose Cruz's) to get people to reconsider (and, as I showed above, a decisive portion of voters DO reconsider.) If he is elected this year (and he very well may be), then discussion is cut off.

While this is certainly a form of advocacy, it's not as though any of us actually SAW any of these players play. You want to see REAL advocacy, wait until the 1990 election and watch someone try to tell me that Ken Singleton wasn't the greatest left fielder of 1970-1984.

watch someone try to tell me that Ken Singleton wasn't the greatest left right fielder of 1970-1984.

To quote post #55 from the '29 Ballot Discussion:

David Foss wrote: "I welcome debate... but at times it appears obvious that a FO-Candidates and EO-Candidates really twist the intrepretations of the data and only present what supports their case. Personally, I'm more swayed by a balanced argument where poster is well aware of any weaknesses that their candidate has."

And I wrote: "I'm the one that used the term 'science project' but the quote above is the gist of what I meant."

----

My original statement was that this project is like a science project, and the statement was then followed by a lot of stuff about how players are elected. In other words, the origins of the "science project" concept were never meant to imply that there were correct answers to who should be in the HoM or that we could attain a HoM that is scientifically accurate. My "science project" comment was based on procedure (and was a bit utopian).

I think the quote from David Foss above does a good job of explaining the concept I expressed when the term "science project" was originally used.

So we probably shouldn't latch on to the science concept in all its potential iterations.

I am sure everyone understands both sides...they just reach different, reasonable conclusions about it.

I think that's a great way to look at our voting process here. There simply aren't any unreasonable ballots submitted. We all put at least a modicum of though into who we think should go in, and in what order.

I do think that there's a bit of a "soft dichotomy" between the voters who are constantly posting new information about the candidates--more or less original research--such as Chris Cobb and John Murphy; and voters like me, who don't really have the time, expertise, or resources to do such in-depth analysis--and simply stand on the shoulders of others. So we have leaders and followers, which I suppose could be expected in any such exercise. I call it "soft" because the ranks of the "leader group" and the "follower group" shift from week to week, with followers not necessarily being influenced by the data provided by only one of the leaders.

It's not necessarily a good or a bad thing, simply an observation.

3 points:

(1) It might be helpful if a mod copied sunnyday's original post and inserted it in the beginning of this thread (otherwise, in future months, newbies won't be able to understand what we're taling about)

(2) sunnyday said:

New information is quite often completely ignored. DavidFoss' info about the 1860s is an obvious case in point, but you can all think of others.

I can personally state that I changed my vote for Pearce because of what Foss posted. Not to miss your bigger point about new information being generally ignored, but I'm wondering what made you think this.

(3) Not to reduce this to absurd levels, but isn't the big issue here simply that people have very different definitions of "merit"?

Not to reduce this to absurd levels, but isn't the big issue here simply that people have very different definitions of "merit"?

That's an interesting question: do we have significantly different standards or do we simply disagree about whether players meet those standards?

Obviously it is some of both, but I would have guessed more the latter than the former. I'd be interested to hear what other people think about that.

we have different standards, at least I think I do. I place a high value on uniqueness, and a fairly high value on traditional stats, either career or a very high peak, but not so high on the 7-10 year "extended peak" if it's not supreme and doesn't result in really good career stats.

Also, I take sabermetric exercises and fancy "one big number" stats like WS and WARP with a big pinch of salt, because in both my banking and journalist incarnations I've seen over and over again what people can do with massaged statistics.

And, by the way, I'm definitely a "follower" in terms of new information, but I attmpt to be a follower who uses his critical and analytical faculties and tells people about the result, when interesting or relevant. But at least 2/3 of you know more about this subject than I do.

I prefer the terms "teachers" and "students" to "leaders and followers" and enjoy being a student in this classroom, particularly since I spend the rest of my life as a lawyer and father being a teacher. I think we full-time students in this class (those who rarely take on a teaching role) are in the minority but help to teach the teachers what is persuasive and what isn't.

It takes a whole lot of theories to build a ballot. Here are some of my theories (or rather, hypotheses?).

1. Baseball, we now know, pre-dates even the year 1800. I am willing to consider anybody who played baseball, anytime, anywhere. Assuming, that is, that we can measure and evaluate and rate and rank their "value." I once "considered" (or maybe pre-considered) Doc Adams, but we cannot measure him. But we can begin to measure players about 18567 or so, so why not?

2. A pennant is a pennant. What counts is the value a player had for his team in his day. It doesn't matter if he could step onto the time machine and come to 2004 or any other time and play.

3. And that value is indeed measured by its contribution to a pennant. If a season was 20 games, or 80 games or 162 games, they're all one pennant (so you adjust for season-length).

4. However, a pennant is not always a pennant. A 1904 AL or NL pennant is the same as a 2004 AL or NL pennant, and an 1890 NL pennant is the same as an 1990 NL pennant. But an 1890 AA pennant is not equal to an 1890 NL pennant, nor the 1884 UA pennant equal to an 1884 NL pennant, nor a Negro League pennant equal to an AL or NL pennant in the same year, nor of course is any minor league pennant equal to a major league pennant. So I make quality adjustments for players who did not play the best competition *that* year. (Again, doesn't matter if they didn't face as good of competition as was available 50 years later.)

5. When i say that we "measure" value, that means quite explicitly that this is not about tools. Joe DiMaggio had 5 tools, Ted Williams clearly did not. Ted Williams was still more valuable.

6. And when I say "measure"(and this is a really important one), I do not insist on quantitative measures. I am willing to consider qualitative information. We now have seen more and more statistical information about the 1860s and Negro Leaguers than I ever thought I would see, but our evaluation of these players is still very qualitative. And the fact is that our evaluation of all players is qualitative. Bill James calls it the "BS dump," and I think we all have one. I sure do.

7. Similarly, I am OK with a small sample. As I said above maybe that means a 30 game season in 1871, maybe it means 50 AB for Pete Hill in 1915, whatever. Use whatever you can get.

8. I strive for a "reasonable" position mix, but who knows what is "reasonable." To me, we have not yet run up against the outer boundaries of what is reasonable, because finally it comes down to Player A vs. Player B. And if we already have 7 LFers and 3 C, but Player A is a LF and is more worthy than Player B, I want the more worthy guy.

9. The same is true returning to the issue of timing. I don't care, really, how many players from the 1880s we've enshrined versus the 1920s. I want the most worthy player. There's an outer limit to that but, again, I doubt if we will every approach it.

10. Finally to some more specifics. I use the ultra-mega stats like WS and WARP (WARP1) adjusted for season length and for league for reasons that are clear from all of the above. I don't use ERA+ and OPS+ other than as a sanity check, but not in any sort of calculations, because they overstate the value of players with short careers, injuries, etc. They confuse me.

11. I like peak and prime, not much for career. I think the idea of a "great" player (and this is probably the most important point of all) is that when he stepped out on the field, he was more often than not the best player on the field, at least for some minimal period of time. 3-5 years is good, or rather "great." If he was never really the best player on the field for any significant period of time, then he probably wasn't a great player. And this goes back to point #1--if Dickey Pearce or even Jim Creighton was clearly the best player on the field every time he stepped on the field for 3-5-7-9 years, boy, that's a great player, and the HoM is here to honor great players.

I'm sure I've left out a couple more steps in ballot construction, but these are some of my theories or hypotheses, almost all of which have counter-theories. In a perfect world, we could test at least some of these in some way such that we all could (or would have to) adopt a similar position on some of them, and thus focus our voting a little bit. But that ain't gonna happen.

PS. When I said that new information is often ignored, I also followed by saying that some of it should be ignored. I also meant that we ignore some information because it appeals to this or that theory that we don't agree to. That is natural. My point, again, was can we focus on two or three of these points, construct a hypothesis, test it in such a way that everybody could agree that it was a fair test, and then everybody agree to abide by the outcome. Again, that is probably not possible here for lots of reasons. But that is what my comment was meant to get at.

Oh and my final voting theory. Vote early, vote often. Especially if you're in Chicago and you're dead (Or, the new version: If you're in Florida and you're white and under 65 years of age.)

(Or, the new version: If you're in Florida and you're white and under 65 years of age.)

...and vote in Democratically controlled counties. :-)

9. The same is true returning to the issue of timing. I don't care, really, how many players from the 1880s we've enshrined versus the 1920s. I want the most worthy player. There's an outer limit to that but, again, I doubt if we will every approach it.

I wonder if having more leagues in the 1880s somehow objectively created "more value" to go around. I mean, in 1878 we had 6 teams and 11 HoMers. I certainly have no problem going up to 12 or 13 in the right circumstances, but beyond that it simply seems silly to look at a league with only 45 regulars and say that almost half of them were the "Best". It almost seems silly now to say that 1/4 of all the starters were the "best."

Baseball in 1884 or 1890 wasn't much different emprically, but there were a lot more teams, so that 20 players could be the top 5% instead of the top 50%.

The 45th best player today makes the All-Star team, the 45th best player in 1878 was likely below replacement level.

Mark,

I think you and I are actually on the same page as far as theories. The only one you've listed that I don't really do in my system is #11. I tend to use a combination of prime, above average value and career value rather than peak. I just feel that peak value is too arbitrary. (Why 5 years, why not 6 or 8 or 10, etc..)

Another thing I look at is number of times a player is in the top 5 in certain categories(ex. OPS+, IP, ERA+, SO/9, BB/9, etc.) This shows how a player would compared to his contemporaries, as well as diminshing some of the differences when conditions or rules changed (Ex. Innings pitched pre- and post-1893)

One thing on the "pennant is a pennant" argument. I think (all other things being equal, which they arent but we wont worry about that right now) that the % of HoMers playing at any given time should be a constant. Thus, a pennant in a 16 team league is worth twice a pennant in an 8 team league. In fact, this concept is the reason that we vote in more players in later years. If we "as a whole" accepted the pennant is a pennant then from the founding of the AL league on, we would have had a constant number of electees. I hope the number of 1990s HoMers is ~= to twice the 1950s HoMers even without a timeline.

This is a small adjustment to what sunnday2 said.
I have other disagreements but those arent ever going to be settled, and thats okay.

Here are some of my theories (or rather, hypotheses?).

This may surprise you, but I agree with every one of those theories or hypotheses.

Since the start of professional baseball, the North American population has doubled roughly every 40 years. The number of major league teams has doubled roughly every 80 years.

I've been aiming to split the difference and have the proportion of players that I vote for double every 60 years (without any strict quotas). Many here will vote in proportion to the number of major league teams per year. Some have proposed inducting about the same number of players from every year. While all of these philosophies are justified, as the "years" go on, the effect will be huge differences in which kind of borderline candidates (i.e. oldtimers vs. modern ones) different voters will be voting for.

Anyway, my point is that one can justify having more players from later years on the basis of population growth, a reason having nothing to do with quality of play. If one believes this principle, but uses a rating methodology that does not yield this result, then one can and should consider using a "timelining" bonus to create such a balance.

In general, the sender of the data weighs heavily on how much weight I give to the new information. For a poster like Chris Cobb, who does a very admirable job of separating information, analysis and opinion, his posts often give me cause to rethink the subject at hand. For others, who are heavy advocates for certain players, I’ll usually just read the opinion and move on – I guess I’m biased against those whose comments (IMO) are obviously biased in favor of a certain player or players.

[Of course I’m lying if I state this is a universal rule for me – opinions from those voters whose ballots appear similar to mine get a warmer reception from me (e.g. I think Joe D. & I had the same 1-2 last year. When he changes his opinion, I should probably look at mine a little closer). Human nature I think.]

It is for this reason that I rarely have tried to persuade others on the discussion thread to vote for players like Thompson or Cicotte, or against players like McGinnity or Pearce. All I feel comfortable doing is stating my opinion on the ballot, and if it convinces you great. But I don’t think that my opinions should be the opinions of the group anymore than I think your opinions should be.

[My exceptions:
1.I feel it's reasonable to question another voter's ballot when that voter supports a player receiving only 1-3 votes for an extended period.
2.new voters offering radical opinions are free game. :^) ]

My Hall of Merit will have fewer high-peak, short career players than our Hall of Merit. Your (not referring to any specific ‘you’) Hall of Merit might have more catchers than our Hall of Merit, based on a conscious decision to elevate their relative value. My Hall of Merit could overvalue the Negro Leagues more than our HoM does. Your Hall of Merit might have more NABBP-era players than our HoM. But your Hall of Merit & my Hall of Merit is not Our Hall of Merit. I think only ‘your’ HoM will exactly correspond to the standards and theories you proscribe. And only my HoM will agree with mine.

Something to keep in mind (my opinion): the point of the project is not that these 213 (+/-) guys we elect are better than every other player; it’s entirely possible (even likely) that someone more worthy will not make it and someone less worthy will (what happens near the in/out borderline is almost by its’ nature random or murky). The point is that our 213 are easily better than the HoF’s 213.

I mean no offense to anyone else’s opinions with these comments. Just trying to offer my honest thoughts on the subject. Apologies for the rambling.

My philosophy/biases.

I'm treating this project as a kind of scientific experiment, with my personal interest to see how well an objective formula can distinguish the kind of careers that get players elected to Cooperstown.

While this is a work in progress, I've found that a formula of the form

Player rating = "strength of prime" plus "length of prime"

works remarkably well. It seems to be the only way to objectively get 8+2 players like Jennings and 2+8 players like Beckley on the same ballot. Strength of prime is determined by rate of performance relative to all peers and rate of performance relative to peers at the same position, weighted more heavily toward whichever strength rating favors the player more.

For position players, I use Win Shares rates. I'm happy with them, probably because the balance between different positions appears to have been "cooked" by Bill James to reflect the judgement of Cooperstown. I don't like pitcher WS rates, because of the reliever bonus, and because of the zero level chosen, so I use ERA+ instead, with all of its problems.

To be honest, the comments that others write do not influence my ratings. I have tweaked the parameters in my system many times, however, when I have deviated too far from consensus opinion. I do enjoy reading the comments here to learn more about baseball history, and I'm participating as much to learn about baseball history as to see who gets elected.

Well, I said I had probably forgotten a couple of "theories," and robc hit on what is probably the biggest one is quota not by position but over time. This certainly related to timelining, but timelining might be a method geared toward an outcome but not guaranteeing it. A quota would guarantee the outcome.

Of course this project at a macro level is quota-driven, by virtue of how many we elect per year. And I don't remember now the rationale for the number we elect each year. Is it the pool (population) size? Or is it the opportunity to earn value (number of teams)? There's a couple competing theories, and I would assume that between 1901 and 1961 it would make a difference which one you choose, unless that is you figure Negro League teams into the opportunities. Another theory to think about.

Just off the top, it makes more sense to me that the quota would be opportunity driven, not population driven, but I don't know.

Anyway, maybe I left this one out of my list (above) because it is already built into this project by virtue of electing 3 a year later on vs. 1 per year in some of the early years. So as a voter I don't really need a quota, it will probably come out that way in the wash.

Or will it? I have the opportunity to elect (PHoM) or try to elect (HoM) more than the alledged quota of old-timers by voting for, say, Mickey Welch in 1985 rather than, oh, I dunno, Tony Oliva? Whereas I can't elect Tony today. So maybe as a voter I do need a quota...someday.

Actually I have struggled with how many 19th century players to put in my PHoM and to continue voting for against the 20th century newcomers of the "borderline" type, but I won't get into the details of that now.

I guess where a quota over time is concerned, I am happy thus far just to compare Player A to Player B with time not a variable in the mix. Who had more value, Mickey Welch or Eddie Cicotte? Cupid Childs or Larry Doyle? When did they play, again?

That might work short term, but maybe not longer term. But OTOH, I think that I might want more old-timers than new-timers in 1960, 1985, 2000. It may have been easier to dominate, but value is value. Easier to dominate really gets at ability.

So it's all circular and I'm still going around in circles after 30 years. Oh well.

The number elected per year was based on a team-years formula (elect 1 player every so many team-years) with an adjustment made for the negro leagues (and I believe, maybe a negative adjustment for WW2). pre NA received 0 team-years and NA team-years were weighted (at .5?).

I just picked up, for the second time, Bill James' Politics of Glory. In it he discusses what are the eight valid arguments for Hall of Fame election. This is Chapter 11, pp 119-36. Do these arguments make sense for the HoM? Do you (general you) use them? And would you find any of them convincing?

1. Look for the Best Candidate, not merely a qualified candidate.
James puts it loosely as "I think [insert player name] should be in the HoM and I believe he is the very best player who isn't in."
2. No one argument places a man at that pinnacle. It's the weight of the evidence; it's always the combined weight.
3. The fact that a comparable player is in the [HoM] is a point in favor of another candidate.
4. The fact that several comparable players are [HoMers] is an important element of proof for any [HoM] candidate.
5. It is important, in evaluating a [HoM] candidate, to show awareness of comparable players who are not in the [HoM].
6. The Highest Common Denominator Argument.
James says, "One of the best arguments that can be made ... is this: that there are many players with comparable records who are in the [HoM], and there are no players with comparable records who are NOT in the [HoM].
7. The fact that a player meets the statistical standards of previous [HoM] selections should be counted in his favor.
James says, "It's not proof, ... but it counts. ... When you show that [player x] is above the standards of [HoM pitchers], the question becomes 'Why shouldn't [player x] be in the [HoM]."
8. If a player is truly in a group of [HoMers] - in the middle of the group - that should be counted in his favor.

I find these arguments to be helpful and they have changed my opinion on how to configure my ballot - from not just best available 1 to 30, but best available and how do they compare to those already inside.
Now, I am not going so far as to compare eligibles to a straight average of elected players. I realize the problem of outliers in a limited sample size. For example, in post 93 pitchers, Young, Mathewson, and Nichols sort of blow the curve when there are only 7.5 elected pitchers (counting Rusie as a half). Also, the context of a player's career must be factored in - such as the general offensive drought from 1903-1919 with the Marianas Trench of 1908 and the cork-centered ball mini-explosion of 1911-12.

Just to present the other side of the coin…

A pennant is a pennant only in that a pennant is an artificially constructed vehicle to proclaim a league’s best team through a less than optimal number of games, designed to approximate what might happen if an infinite number of games were allowed.

A pennant race is entirely enjoyable, but as all good Reds fans know, it’s one thing to be in the race in late June; it’s quite another to be in the race in late September. As such, the longer a season, the more reliable the results and the more accurate the statistics as a reflection of theoretical value.

To take this thought down a level to the individual players, it is also clear that statistical accomplishments have more weight and accuracy in longer seasons. Put more specifically: Hitting 20 home runs in 50 games does not have the same value as hitting 60 home runs in 150 games, because the list of players who have done the former is considerably longer.

Yes, but hitting 20 home runs in 50 games is more valuable than hitting 20 home runs in 150 games.

Dolf Lucky, are you saying that pennants won in longer seasons should carry more weight?

Yes, but hitting 20 home runs in 50 games is more valuable than hitting 20 home runs in 150 games.

Yes and no. For the shorter schedule seasons, you should use standard deviation to draw the outliers back toward the mean. By doing this, Levi Meyerle's 1871 season is more reasonable (for example).

While I agree with Mark in this regard, the same guys were always at the top of the heap in the NA and early NL. While the shorter schedule created a greater spread between the best and worst players, the schedule was long enough to show who were the best players in the league. The sample size was large enough that a Tony Suck couldn't be the top hitter in the league.

I would like some input from the group on a toy I'm putting together.

I thought I could do Bill James-style Sim Scores with Win Shares, not to measure which player is most like another player in terms of tools, but which player is most like another player in terms of value as measured by WS.

For hitters, I was thinking of the following:

Start with 1000
Subtract the difference in 3-year peak
Subtract the difference in 5-year peak
Subtract the difference in 7-year peak
Subtract the difference in 9-year peak
Subtract 1/2 the difference in career
Subtract 3 times the difference in WS/162 games.

I would do this <u>separately</u> for batting win shares and fielding win shares, so that the players' defensive value would be directly compared and accounted for. Bill James' original system subtracted a fixed number of points based on the two primary positions of the players being compared. I think this will work better, because it will by its nature take into account that some positions are more valuable defensively, and it will also account for players who play more than one position in their careers (e.g., Ernie Banks).

The reason I'm doing the 3, 5, 7, and 9 year peaks is it helps to compare the arc of the career, whereas the original SimScore just measures career.

I have to admit I've randomly chosen the "points" to be assigned, and haven't tested it yet, but I'm wondering if anyone has any thoughts. For instance, should I subtract the full difference between career WS, instead of only half? Should I raise or lower the multiplier for WS/162 games?

By the way, if any of you want a hitting and pitching spreadsheet that calculates the Bill James Sim Scores based on career statistical totals, I have them. The data is good through about 2000. The batting simscore macro was created by someone else (so the credit goes there). I got the macro off the Web years ago. I adapted it to create a pitching macro too.

It has hitters with 3,000+ ABs, starting pitchers with at least 1,500 IP, and relief pitchers (measured by GS/G < 40%) with either 1,000 IP or 500 appearances.

With the hitting version, you can toggle on and off the positional adjustment and the adjustment for era played in. You can also calculate it based on career totals, or by rate (per AB).

The pitching version has a toggle for the era adjustment (as in "era" not "E.R.A.").

Baseball reference has the simscores of course, so you may not care about the spreadsheets, but if you do, just e-mail me. (By the way, there are slight differences in B-R's results and my spreadsheet b/c B-R has changed the simscore formula a little bit for its purposes).

I don't adjust seasons like Meyerle's 1871 downward at all, but I don't prorate them either. I don't compare that season to a season in 1927 or 2003, I compare it to the other players who played that year. If there is a league quality issue, i only make modifications as compared to other leagues that year, not other leagues in the past or future. Then in comparing players who played in different times I give equal "points" for being the best in 1871 and the best in 1927. And extra points if they lap they're own fields.

On theory in general, I think it is interesting to note that i may be the only one who entered the project thinking that career accomplishments were the ONLY barometer for Hall of Fame/Merit voting. That is how I saw it -- what did you do over the course of your career. Peak was given zero value by me, except in the sense that players with high peaks would generally have good careers. I have changed quite a bit (I like the all peak Walsh and Jackson-type careers) but I still would not elect a Hughie Jennings or someone whose entire career consisted of Bonds' last 2000 to 2003 or Ruth's 4 year peak. A power hitter with a career OPS of 1.300, 240 homeruns and 500 rbis over 2500 plate appearances just doesn't cut it for me.

Dolf Lucky, are you saying that pennants won in longer seasons should carry more weight?

More or less. What I'm really saying is that everything should be looked at in the proper context. I mean, we already discount the (for example) '87 Twins because their RS/RA was so bad. What we're really saying is that if the season were 300 games long, the Twins would not have won their division. I think that is an extension of giving more weight to longer seasoned pennants.

To do the same thing with player seasons is natural, and correct. You can't put a Lip Pike 200 OPS+ season on the same pedestal as a 200 OPS+ season from Barry Bonds, because the longer the season goes, the more exclusive the "200 OPS+ Club" becomes.

To do the same thing with player seasons is natural, and correct. You can't put a Lip Pike 200 OPS+ season on the same pedestal as a 200 OPS+ season from Barry Bonds, because the longer the season goes, the more exclusive the "200 OPS+ Club" becomes.

I totally agree. Of course, the flip side applies for counter stats from later generations. Bonds (born in 1831) doesn't hit anywhere near 73 homers, Hack Wilson can't smell 191 RBI and even the Old Hoss Radbourne of 1884 wouldn't even be able to win 59 games in the NA of 1871.

Given his age and reliability, Bonds would presumably be known as "Old Hoss Bonds" by now if he played in 1880!

Given his age and reliability, Bonds would presumably be known as "Old Hoss Bonds" by now if he played in 1880!

:-)

The subject of comparing and ranking players from different eras has been raised here and in Jeff M & Max P's discussion of Lip Pike, so I thought I would post the results of my attempt to work out decade-by-decade shares of the HoM places available through the 2010 election.

Assuming that places should be distributed in proportion to the number of major-league teams (or major-league team equivalents), as that's the idea Joe D. used to develop the election schedule,

I added up all places to be elected through 2010, by which point about half of the top stars whose careers were centered in the 1990s will be eligible -- 240 places.

I divided by # of decades of baseball under consideration (13.5) to find avg. # per decade--17.78 places per decade.

I calculated avg. number of teams per decade, using estimates for pre-1871 team-equivalents and for negro-league major-league-team equivalents and also found overall average, which is 18 teams active per decade.

I then prorated avg. # of places per decade by # of teams in each decade to determine each decade’s share of the 240 places, assuming that the 1990s would ultimately have at least twice as many HoMers as would have been elected by 2010 (as most Bonds, Biggio, F. Thomas, Bagwell, Griffey, Jr., Palmiero, R. Alomar, Maddux, Sheffield, Piazza, Clemens, Glavine, R. Johnson, Brown, Smoltz, I. Rodriguez will not yet be eligible). By assuming that half the 1990s share would be filled by 2010, the estimate arrived at will be more accurate.

Those results, rounded to nearest multiple of .5
1860s = 2
1870s = 7
1880s = 14.5
1890s = 14.5
1900s = 17
1910s = 20
1920s = 20.5
1930s = 20.5
1940s = 20
1950s = 16
1960s = 20.5
1970s = 24.5
1980s = 25.5
1990s = 27.5 (13.5 filled)

I then assigned electees to a single decade on the following system. Player counts in decade in which he earned the most career win shares, except for nineteenth century, where player counts in decade in which he ranks highest in win shares for that decade. Players ranked for one decade only unless player’s rank in win shares earned per decade clearly places him among the electable group for two decades, e.g. Cap Anson for 1870s and 1880s. (There's a complex system to this that I won't describe here in full.) The goal is to map players' careers fairly into decade placements.

I then compared electees by decade through 1928 to shares of HoM places per decade.

The results (above/below share)

1860s = 1 (-1)
1870s = 7 (0)
1880s = 16 (+1.5)
1890s = 10.5 (-4)
1900s = 14 (-3)
1910s too far from complete for meaningful results.

ANALYSIS

Given that we’re likely to elect Sheckard and Wallace, both 1900s stars, this year, I make the following observations.

Our distribution of electees in relation to decade shares is pretty good, so far (1880s a bit high, 1890s a bit low, but partly explicable as an effect of contraction, which cut short some players who might have shared their value with the 1890s otherwise).

If the system of decade shares seems a reasonable guide to dividing HoM places between eras, we should consider our selection of players for the half century from 1860 to 1910 as nearly complete. There’s a bit of room left, but not much. After Sheckard and Wallace, I think we’d be right to elect about 5 more players from baseball’s first professional half-century through the 2010 elections. Any more than that and we will be cramping later eras somewhat. (I've done a quick study of the 1980s candidates in relation to the "decade shares" system, which I think corroborates this assertion, which I'll post shortly.)

It looks like the HoM mostly finishes up a decade’s players about 20 years after the end of the decade. We will be substantially finished with the 1900s decade after the 1929 election. With 3.5 players representing the 10s elected so far (let’s call it 4), we have 57 players more to elect from the 10s, 20s, and 30s. We’ll have enough slots to have elected those 57 players, after 1929, at the completion of the 1958 election. Assuming we elect a few more from the 1860-1910 period during those years, we should finish the 1930s in the early 1960s.

These posts aren’t meant to advocate for strict quotas, especially on a decade-by-decade basis, but I think large-scale departures from decade's or a half-century's shares of HoM places would be a sign of poor choices by the electorate.

Ranking players within a designated period and comparing lists from different periods to discern "relative excellence" seems to me to be a reasonable approach to dealing with the problem of comparing players from different periods.

As a follow-up to my post on a "decade-shares" system to assign HoM places per decade as a basis for ranking players from different eras against each other, here's a quick overview of what the "decade shares" system suggests about the merit of 1980s players, taken as a group.

"Decade-shares" system assigns 25.5 HoM places to 1980s through 2010 election, based on 26 teams for the decade and an average of 18 teams per decade.

I would divide these places, in the abstract, between position players and pitchers, like this:

18 pos players
7.5 pitchers
(using my usual ratio of 3 pitchers per 10 players, total)

Does this share provide enough places for the players who should be elected? Does it provide so many that unworthy players might be eleccted by adhering to it? Here's a quick review of the eligibles. (Players who also rank in another decade count for half a place, as indicated.)

Position Players

Shoo-ins
.5 Henderson ( shared w/ 90s)
.5 Schmidt (70s)
.5 Ripken (90s)
.5 Molitor (90s)
.5 Gwynn (90s)
.5 Fisk (70s)
1 Yount
1 Murray
1 Boggs
1 G. Carter
1 Winfield
8 total places

Good Candidates
1 Dw. Evans
1 Trammell
1 Whitaker
1 Dawson
1 O. Smith
1 Sandberg
8 total places

If one agrees that those 16 pretty clearly deserve election (obviously there won't be unanimity about all the "good candidates"), that leaves 2-4 places for some of this group of borderline candidates (I think pos players will cut into pitchers' share for this decade, because the pitcher pool is _very_ weak):

Borderline Candidates
.5 Da. Evens (shared with 70s)
1 Murphy
1 Hernandez
1 Clark
1 Randolph
1 Parrish
1 Mattingly
1 B. Bell
1 Cruz, Sr.
1 Rice
1 Parker
1 Puckett

The outcome from the decade-share system for position players seems pretty close to the currents standards of the Hall of Fame, when they are electing sensibly.

Now a look at pitchers. Tweaking the share pitchers receive based on my sense of the strength of the pool, I'd suggest we look at 5-7 places to be filled by a selection of the following pitchers, generated by listing all the pitchers with at least 150 career WS who earned more than 50% of those WS in the 1980s, plus Blyleven and Ryan, who were among the top 8 in WS for the decade, even though they earned more WS in the 1970s:

Pitchers
.5 Blyleven (70s)
.5 Ryan (70s)
1 Eckersley
1 Tanana
1 Hough
1 Gossage
1 Morris
1 Hershiser
1 Stieb
1 L. Smith
1 Saberhagen
1 D. Alexander
1 Reuss
1 Welch
1 Gooden
1 Viola
1 Langston
1 Darwin
1 Guidry
1 Sutter
1 Valenzuela
1 Flanagan
1 Tekulve
1 Quisenberry
1 Reardon
1 Sutcliffe
1 Rhoden

It's a weak decade for pitchers. Even at 5 from this group we would be electing more than the Hall of Fame has so far. Maybe some more slots would go to position players, or maybe we need to establish a more reasonable standard than the HoF for post-1980 pitchers. In any case, in my view there is neither an obvious excess nor shortage of places assigned for the 1980s by the decade-shares.

In case you are concerned that some stars of the 1980s have been overlooked, here's a list of some stars who are counted only for other decades in the system:

Reggie Jackson
Bobby Grich
Barry Bonds
Fred McGriff
Will Clark
Roger Clemens
Steve Carlton
Tom Seaver
Gaylord Perry
Phil Niekro
Don Sutton

In sum, this projection of 25.5 slots for the 1980s looks to me like it would keep HoM standards about where they are now. There would easily be room for the obvious choices, and there would be room for a part of the borderline group, whom we’ll debate and sort. To keep room open for these players, we need to pay some attention to not overfilling earlier years. So far our collective judgment seems to have been good, but it’s worthwhile, I think, to check them against a numerically derived quota. The farther we proceed in the project, the more challenging it will be to keep the different eras fairly represented.

Interesting analysis. I am wondering if the 1980s, though, may be unrepresentative of baseball overall. In a decade of parity among teams, maybe there was a lot more "parity" among players, with a whole lot of very-goods, but not a whole lot of greats.

I wouldn't be suprised if, by 2010, we have elected substantially more stars of the 1970s than the 1980s.

I also note that you have overlooked Dennis Martinez, who was one of the winningest pitchers of the 1980s (and Top 50 all time).

Very nice analysis, Chris. It supports my suspicion that the elections will generally be playing "catch up" with the most recent couple decades of players, and that the election of oldtimers will be a relatively rare event. Quite a contrast from a BBWAA/Veterans Committee setup.

By the way, will you have to recuse yourself when Ty Cobb comes up for election? :-)

Interesting analysis. I am wondering if the 1980s, though, may be unrepresentative of baseball overall. In a decade of parity among teams, maybe there was a lot more "parity" among players, with a whole lot of very-goods, but not a whole lot of greats.

Ah, here we get into a real question of voting theory! I would make the argument that it's more like stronger competition diminishes the apparent superiority of the best players, but that the players are still just as good. On this basis, I'd argue that the 1890s stars ought to be getting a notable boost in the rankings relative to their 1880s peers, and even a slight boost against 1900s stars; I'm planning some small adjustments to my 1929 ballot on that basis.

Whatever the theory, it looks to me like in practice the 1980s don't look much different from the preceding decades in terms of player merit. I've done similar quick scans of the 50s, 60s, and 70s now, and, at a rapid glance, I don't think any of three decades appears to have more players worthy of election than the 80s. There's a concentration of great _pitchers_ in the 70s, but the position players pool looks weaker.

I also note that you have overlooked Dennis Martinez, who was one of the winningest pitchers of the 1980s (and Top 50 all time).

Actually, D. Martinez counts for the 1990s; he earned 103 win shares in that decade, as opposed to 84 in the 1980s. He has a very unusual career shape.

A somewhat more glaring omission from my quick pass through the decade is George Brett, who belongs in the shoo-in category . . .

By the way, will you have to recuse yourself when Ty Cobb comes up for election? :-)

He's no relation, so I think I can trust myself to be unbiased in my evaluation of him :-) .

El Presidente will be elected to the "Dizzy Gillespie" wing of the HoM, with Don Zimmer, Greg Luzinski and Tug McGraw.

Ah, here we get into a real question of voting theory! I would make the argument that it's more like stronger competition diminishes the apparent superiority of the best players, but that the players are still just as good. On this basis, I'd argue that the 1890s stars ought to be getting a notable boost in the rankings relative to their 1880s peers, and even a slight boost against 1900s stars;

I'm in full agreement with you, Chris.

This isn't really voting theory, but mostly a public announcement because I figured some folks might be interested. ESPN Classic is going to show a "Vintage Game" Saturday night, played with 1886 rules, uniforms, and equipment. The game will be at Wahconah Park in Pittsfield, MA.

ESPN Classic

Thanks mbd1^2!

A voting theory question. Several of the next group of eligible pitchers were legally allowed to use a spitball after it was outlawed for other pitchers. When measuring them against league average or replacement level pitchers, we realize that they had a legislated advantage over those other pitchers. Do we discount for it in any way? Or is this another of those ability versus results issues?

Also can someone provide the complete list of the grandfathered spitballers? Who among plausible candidates is effected - Coveleski? Faber?

OCF:

Go to http://www.baseball-almanac.com/legendary/lispit.shtml

It looks like Coveleski, Faber, Shocker and Grimes are the leading contenders who benefitted from the grandfathering.

Do we discount for it in any way?

I would say absolutely not.

Chris, great stuff. Your conclusions are very carefully stated as always and you say that you are not advocating for a strict quota...(but). It is a very weak but and so I don't want to put words in your mouth.

But you conclude by saying that any large scale departures from the decade shares would be a sign of poor choices. I think that's a tad strong, except maybe the fact that you said "large scale." Setting aside the definition of large scale, I would hesitate to assign quite such importance to a departure.

Or at the bare minimum I would hope that nobody would miss your statement that you are not advocating for quotas, and I hope that nobody would conclude from your information that a quota would be the way to go.

Rather I think that what you show is just the opposite--that we are doing great and not getting off track at all *without* a quota, and therefore we *should* continue without a quota.

I think that what you've shown is that this whole exercise is tautalogical. We are about value, not ability or tools, and value is opportunity-driven and not ability-driven. And our "quotas" are totally driven by opportunity (how many teams, how many pennants).

But what is my point, then, if we're on track and it is in fact almost inevitable and tautalogical that we are on track.

Well my point is that we could easily depart from the quotas without really being off track. We *will* elect 3-4 more 19th century players by 1934, I think, and by your numbers that should be the end of it. I think it would be a bigger mistake to accept that that is the end of the 19th century than it would be to elect 3-4 more 19th century players after these next 3-4 (the first 3-4 now queued up being Wallace in part, Caruthers, Thompson, Pearce and maybe Pike, and then part of Beckley, Van Haltren, maybe Griffith, maybe Ryan or Bresnahan after that).

All of them would be OK regardless of the quotas for two reasons.

1. It comes first and foremost down to individuals--Pearce or Tinker, Beckley or Chance?

2. The end point of this project as 2004 or 2005 or 2007 is arbitary. In theory we will go on foreover, and 20th century players will keep on getting elected (19th too, but few of them after 1934) through 2050 or whenever. So the evening out of the quotas needn'e happen by 2004 or 2007. It can, in theory, occur later and the further out we go the more favorable the distribution will become for the later models.

In light of this latter point, what would be more of a problem, more irreversible, would be if we got too many 20th century players compared to the "quotas" any time soon.

sunnyday,

I thought this project was about merit, not about value. Maybe value = merit to you, but to someone else ability = merit, and Im not going to tell them they are wrong.

Well, okay, I will. To me, merit=merit. I will stick with my happy little tautology and slap down those that try to decree something else as the definition of merit.

Robc, I don't think sunnyday2's thoughtful post depended on whether you define merit as value, but on balancing between eras. To some, ability is merit, they are peak voters. To others, value is merit, they are career voters. To me, unique talent has more merit than either, particularly for these early players -- my curiosity to see ESPN Classic's 1886 baseball game is only moderate, but to watch (say) Caruthers, Anson and Old Hoss (yes, and Beckley -- all those triples) in action would truly be a privilege.

I thought this project was about merit, not about value.

Agreed, but taking into account the differences in eras, medicine, etc., doesn't deviate from that. It's no different from using park factors to pinpoint merit more clearly.

We don't want to get to the point that we start (not us, I assume :-) electing mediocre players centuries from now because they happened to have had the luxury to have been born in a time where they were bigger, stronger, faster, etc., than 19th, 20th or 21st century greats (which that version of WARP3 will surely indicate). That's not highlighting merit, but silliness.

robc, well, you've defined the question of the ultra-mega-theory of everything. All the other theories I listed the other day are just means toward an end, and the end is to measure...what?

OK, merit.

But what is that? I mean if the HoF is measuring fame...no they're not. So I guess it's not obvious to me that we're about "merit." Merit is just a word. What is "merit"? What are the components? How do you define it?

I guess I think it's value, not ability, there I said it. It's value as measured in some kind of units such as or like or similar to WS or WARP or LWTS.

If it were ability (as I understand it) then Joe DiMaggio would rank higher than Ted Williams. More tools = more ability. But not more value.

That's what I think it is. I don't think I said everybody had to agree. What do you think it is?

The happy little tautology of merit = merit isn't the tautology I was talking about anyway. What is tautalogical here is that our voting is based on a quota system and then, voila!, the people we elect appear to be quota'd out over time.

Merit = merit is a different tautalogy. But I am curious as to how people define merit (either the one before the = or after, doesn't matter ;-)

I guess I think it's value, not ability, there I said it. It's value as measured in some kind of units such as or like or similar to WS or WARP or LWTS.

That's my definition of merit, vis-a-vis this project's.

For me, the definition of "merit" is rarity of achievement. If a player achieves what only one in a hundred million people could do, it doesn't matter whether it's a five year run of pitching in the 1960s or a twenty year run at first base centered in the 1890s.

TOP TEN SIGNS YOU'VE BEEN SPENDING WAY TOO MUCH TIME ON THESE PAGES (or, Get Back to Work!)

10. You have actually read each comment on this "Voting Theory" thread.

9. When watching the talking heads during the Jayson Williams trial, you've speculated as to which reporter is poster "Howie Menckel."

8. Jeff M has ever commented on your ballot, and added a smiley face.

7. You have ever posted something negative about Dickey Pearce, and John Murphy <u>didn't</u> respond within three minutes.

6. You have a very strong opinion about the merit of Bob Caruthers, or you can point out the link between Bob Caruthers and his biggest advocate, karlmagnus. (Hint: it has to do with a certain region of the world with a really big art gallery, wine, a famous tower, and the first Holy Roman Emperor.)

5. You know that sunnyday2's real name is Marc.

4. You fully understand every one of Chris Cobb's posts, or like to think that you do.

3. You miss the days when Hall of Famer Happy Jack Chesbro checked in to see if he had won that year's election.

2. Without checking, you can name all five voters who put Dickey Pearce on their ballots in the 1898 election.

1. When thinking of a player who completed his career before the current election, you know longer think "Hall of Famer" or "not Hall of Famer", but rather "He's in the Hall of Merit" or "He's not in the Hall of Merit."

Sorry, 2 should be "the only voter who put Dickey Pearce on their ballot in the 1898 election."

2.a You've ever hit "submit" before editing your post.

I think some of my post was misunderstood (probably my fault but Im blaming ya'll anyway). The tautology I was referring to was mine, not sd2s. And marc, you wrote "We are about value", unless that was a royal we, you seemed to imply that everyone was in agreement, but I knew you didnt mean it. Just wanted to knock it down in case any newbies read it.

As to what I consider merit.

Mostly, its value. But ability does play a part. Although, if it were ability, Jose Rijo would be an eventual HoMer, and he wont be. Mainly because one of the abilities he was missing was the ability to stay healthy. And the lack of that ability prevented him from using his other abilities to put up value.

There are other things that go into merit. Our rules limit them to on the field activities, which I fully supported. I consider throwing the world series to be a huge act of negative merit (and it is an on-field activity) and adjust accordingly. Jackson got elected without a vote from me, and Cicotte wont ever make my ballot (not from an absolute ban, but from the size of the negative merit penalty lowering him to a point he wont ever make it).

Finally, I timeline. Actually, no I dont, but everyone calls it timelining. Bill James goofy year of birth adjustment is timelining. Adjusting "value" based on the quality of opposition is something different that happens to mostly follow a timeline.

I wonder, are the pennant is a pennant people going to treat 1944 the same as 1941 or 1946? I wont, it will have less value than 1941 even though it is later in time. Also, does Rizzuto get any credit for his war years from the "value" people? He put up zero value during those years so a pure value voter should give him no credit.
And if you are willing to give him credit because he was drafted and had no choice (reasonable) what about someone who wasnt drafted but volunteered for a war?

One of my rationales for timelining is our voting methodology. If we were voting for the best 210 players ever, I might not timeline. Since our voting follows the history of baseball, it seems reasonable that the reason we are doing that is due to the changes in ability over time. By starting in 1898, we give all players a chance to be elected against their peers without having to go up against the supermen of the future. We aint ever voting anyone out, no matter how good baseball gets in the future.

Anyway, its late, Im done for the evening (I saw my first steal of home in person tonight at the Louisville-Indy game. Home needs to be stolen more often, it was an exciting play). Everyone have a happy July 4th. Read a copy of the declaration this weekend if you get a chance.

7. You have ever posted something negative about Dickey Pearce, and John Murphy didn't respond within three minutes.

LOL

It's sad, but true. My advocacy for Pearce makes my earlier one for Sutton seem small in comparison.

"the only voter who put Dickey Pearce on their ballot in the 1898 election."

Wasn't that Dan? Whoever it was, he was the person who got the ball rolling for me about Pearce. You EODP can blame him now. :-)

I wonder, are the pennant is a pennant people going to treat 1944 the same as 1941 or 1946?

I'm not, but that's because we know that the best players were "sitting it out" in Europe or Asia. However, Hal Newhouser was still great regardless of the inferior competition as Dickey Pearce was during the Civil Ware era (see Ron, nobody even mentioned him and I got him into the post :-).

And if you are willing to give him credit because he was drafted and had no choice (reasonable) what about someone who wasnt drafted but volunteered for a war?

They both get credit from me(Greenberg volunteered, I believe). They were serving their country in a time of crisis. That's good enough for me.

Interesting discussion of merit, value, and ability.

Going back to Sunnday2's tautology point about my quota analysis, with which I mostly agree:

But what is my point, then, if we're on track and it is in fact almost inevitable and tautalogical that we are on track.

I agree that it is _almost_ inevitable that we will stay on track -- if by on track we mean fairly honoring the most meritorious players from every era of the game -- if we use value as our primary indicator of merit (and find it accurately--but that's not a problem of _voting_ theory). The schedule of elections pushes us towards electing players in proportion to the opportunities to create value in each era, but it doesn't force us to do so. I think that if many voters did not use some view of value as their main criterion (and I think that even "ability"-oriented voters who rely more heavily on rate stats are still giving a significant nod to value if they contextualize those stats, as most do), then our selections might have given less proportionally balanced shares to the different eras, or there could be massive positional inequities within eras (an aspect of representation I didn't study).

So I would say that an important implication of the study is that if proportional representation of the best players from all eras of baseball is a goal of the HoM and its electorate, emphasizing value helps to achieve that goal.

I'm not saying that value should be the _only_ criterion for assigning merit, but I think it is a necessary part of assigning merit.

But you conclude by saying that any large scale departures from the decade shares would be a sign of poor choices. I think that's a tad strong, except maybe the fact that you said "large scale." Setting aside the definition of large scale, I would hesitate to assign quite such importance to a departure.

"Large scale" is a fuzzy term, I will readily admit, and I don't have any specifics to apply. "Poor choices" is a strong term, and, in retrospect, I admit to using somewhat inflated rhetoric to push the point.

I think it's a point worth pushing for three practical reasons, right now in the history of our elections:

1) We're going to be making a big transition from dead-ball to lively ball, and in which the level of competition continues to rise. I'm not certain that there won't be a tendency to privilege earlier players from a time with the the best players accrued a larger share of the available value. There may be a contrary tendency to privilege later players too much by excessive timelining. So far I think that, collectively, we've kept the balance right. But I believe that checking ourselves against a numerical quota projection by decade or era can be useful when trying to decide how to compare an 1890s outfielder to a 1920s outfielder. Not that we have to adhere to a quota, but we should know and be convinced that we're right when we decide not to.

2) The HoM's stance against the HoF is going to change as we hit the 20s and 30s. So far we've mostly been including players who have been left out of the HoF. When we hit that period, there's going to be a lot more excluding, even of players that right now we believe are pretty darn good. I don't know how that's going to affect our perception of the process, but again a look at quotas could be a useful way to remind ourselves about the vision of fairness for all eras that's built into the process.

3) I'm concerned about the 1880s vs. the 1890s. There are some legitimate reasons why we will be right to honor a few more 1880s players than 1890s players, but it looks to me like the higher level of competition and the shortening of careers by contraction has caused us to give a bit less credit to the 1890s players than they deserve. I wasn't sure of that before I did the study, but I'm convinced now. I think Mickey Welch and Bob Caruthers merit election, but I also think Clark Griffith merits election, and I think he is more likely to be overlooked or kept on the doorstep because the difficult conditions for pitchers made it much harder for them to accumulate value. So for me, checking our results against the quota system (and noting that we have a bunch of 1890s players who've been just below the election threshold for a long time), is evidence that we've been undervaluing these players. Thus I go for Griffith before Welch and Jennings before Caruthers.

but it looks to me like the higher level of competition and the shortening of careers by contraction has caused us to give a bit less credit to the 1890s players than they deserve.

This is a point I've referred to for awhile without clearly detailing. In my comments for Griffith and Van Haltren and now Jennings, that they excelled during the contraction era 1892-1900. When using a metric such as OPS+, ERA+ or win shares, players of this era have an inherent disadvantage compared to adjacent eras.

I'm glad Chris has become the point man on this topic and, hopefully, specific numeric adjustments will result.

And yes, I was the original FODP in 1898 (and take pride in saying so). Maybe I'm just a sucker for an underdog.

This is a point I've referred to for awhile without clearly detailing. In my comments for Griffith and Van Haltren and now Jennings, that they excelled during the contraction era 1892-1900. When using a metric such as OPS+, ERA+ or win shares, players of this era have an inherent disadvantage compared to adjacent eras.

I also have trumpeted this same argument myself on occasion.

And yes, I was the original FODP in 1898 (and take pride in saying so).

As you should, Dan. Before your placement of him, he wasn't even on my

I don't buy this "superiority of the 90s" argument because of the economics of it. Today anyone with a decent chance to become an ML star or even a regular would do so, unless his career alternative was to be a top Wall Street derivatives trader.

In the 1880s and 1890s, things were different; for those players whose alternatives were skilled working class or better, baseball didn't offer a huge amount more money (presumably why Thompson was a carpenter till 25 -- the expansion of the game in the early 80s, perhaps the successful 84 season, presumably convinced him there was a good living to be made.)

Thus the universe of players would have been less in the early 1870s, when it wasn't clear there was decent stable money in the game (a marginal argument against Pike and Pearce -- maybe it wasn't just the lack of good scouting.)

After the Players League and the AA folded, the NL instituted a salary cap, of $2,500 from memory, and of course the number of opportunities declined. Just as some aging players were forced out of the game by this, so some potential new players who had other career opportunities (in carpentry, for example) will have given up on the baseball idea and done something else.

This salary drought lasted until the AL came along in 1900-01. Conseqently, in 1892-99 a lot of good players, some of them maybe potentially top stars, never took up the game. (with embryonic scouting, neither the player nor the scouts KNEW who the potential stars were in the 1890s -- they don't now really, look how many #1 draft picks wash out.)

The players missing from 1890s baseball were therefore NOT necessarily the marginal ones, and the decline in HOM members may in fact be a genuine reflection of this.

After 1901-1910, the money in baseball improved by an order of magnitude (a top star made $10,000 maximum in the 1880s, Ruth made $80,000 in 1930, with very little price inflation between the two.) So future periods of non-expansion, like the 1950s, may indeed have missed out only on the marginal players, with few top stars never taking up the game. But baseball salaries in 1880-1900 were close enough to a skilled working man's earnings, let alone a laweyer's earnings, that the economic effect MUST have been significant.

karlmagnus:

IIRC, when taking into account the rise in inflation, ballplayers of the early seventies made less money than their counterparts from the twenties. I don't think there are too many people here that think the pre-Jackie Robinson era was more competitive than the post-Robinson era.

John, there's a supply and demand factor here, too. First, baseball in the 20s was extremely well paid, by comparison with blue collar and many white collar occupations, as it still was in the 1970s -- as you probably remember but not everybody else here does (:-)) there were very few people in 1970 making the $100-150,000 that top baseball stars made. So a modest decline in relative salaries wouldnt have much effect, unlike in the 1890s when baseball's salary strcuture was much more down to earth (since there wasn't rapid inflation till after 1970, the $2,500 maximum of 1892-1900 was just about the same as the $7,000 minimum (from memory) of 1966-70.)

Second, on the supply side, between 1925 and 1970 you had the influx of black and Hispanic players into the game, which increased supply by about half (I seem to remember about 1/3 of players in the early 1970s were black or Hispanic.) In any case, the country's population increased by about 50% between the two eras. So even if a few stars in 1965-1970 were lost to the joys of being a trial lawyer, the counterbalancing effects were much greater. Except for the population increase (but roster sizes were increasing too, with more ptichers) there was no such supply-side effect from the 1880s to the 1890s.

PS. I don't agree with the idea that value = career and ability = peak, BTW. I mean, I understand that ability is something that you don't accumulate over time. But I'm a "peak value" and "prime value" voter.

A couple of days late, but that "Top 10 Signs".. list by Ron Wargo is hysterical!!