Quick response:

Wallace looks like a #1 to me for just the reasons Joe lists, pending final assessment of Bill Monroe. Monroe looks like he belongs in the top third of the ballot, but not certain exactly where yet.

Walsh looks better than Waddell, not as good as McGinnity. I'll post numbers later, but consider that he threw almost the same number of innings as Waddell with _slightly_ higher effectiveness, and he did it in fewer seasons, at a time when average innings pitched have dropped a bit (5%, I think) from Waddell's peak. So Walsh's career value is about the same as Waddell's, but his peak is higher. But, like Waddell, he's 500 innings short of McGinnity. His inning-for-inning value is a little higher than McGinnity's, but that's not enough to make up the difference in innings pitched. Walsh beats Joss any way you look at him, I think.

I think I'll probably have Walsh leading the pitching pack in '20. I think he's comparable to McGinnity with regards to Win Shares, both Career and Peak; he averages over 3 WS/162 above the Iron Man; looks better in both WARP 1 and WARP 3; posts a 145 career ERA+ vs. McG's 121. Nothing certain yet, just a quick assessment.

Any hints as to who our 1919 electees will be?

Wallace and Walsh will both rate very highly.

Wallace is closer to Dahlen and Davis than to Tinker or Evers or Chance. The Beckley of SSs would have earned vastly more defensive value over a long career and therefore rank much higher than the equivalent 1B.

Walsh's peak is much better than McGinnity's, or anybody else's who is eligible. And since that is basically what the current crop of pitchers has to offer, Walsh has to go to the head of the class. (Caveat: Of course a pitcher with a very high peak and a HUGE hitting component could still rate ahead.)

Prelim Ballot

1. Pearce

Total subjectivity here:

Robert Peterson, in his book "Only the Ball was White..." says of Bill Monroe (paraphrased) that Rube Foster thought he (Monroe) was the greatest player he ever saw. John McGraw, though less enthusiastic in his praise, claimed that Monroe could be a star on several major league teams.

I agree that Clint that Bobby Wallace is the "Jake Beckley of Shortstops," but since I have Beckley on my ballot, Wallace will be there too.

According to my analysis, Beckley was the best of second best first baseman in his league 8 times. (Best: NL 1901; 2nd best: PL 1890, NL 1893, 1895, 1899, 1900, 1902, 1904).

Wallace was the best of second best shortstop in his league 8 times (Best: NL, 1901; 2nd Best: NL 1899, AL 1902, 1904, 1906, 1907, 1908, 1910).

Wallace was a good defender for a long time, and racked up defensive points the same way Beckley racked up hits. I will have them paired somewhere in the middle of my ballot.

From David Foss's post #66 on the 1919 ballot thread:

" Just got my Sabermetric Encyclopedia in the mail and thought I would try it out on OCF's question:

1871-1918

HOMERUNS YEAR HR

Ed Walsh's support neutral W/L record is 214-107 (.667 W%), which gives him 250 Neutral Fibonacci Wins.

Pitchers in the same "neighborhood" :

Name Supt. Neut. Fibonacci Win Points

I hate short careers, but Walsh is going to be very high on my ballot. So high, he is making me reconsider the absence of Joss and Waddell on my ballot.

#1 ERA of all-time combined with #7 ERA+ of all time is pretty darn good. It just surprises me that Joss is so close to him in both of those rankings and yet Joss isn't in my top 20. What separates these two for anyone else who has them, like me, 15-30 places apart?

I am having a similar dilemma as daryn.

I am also tempted to place Walsh high on my ballot, but he seems very much like the "Cupid Childs of pitchers," and I don't have Childs on my ballot.

Perhaps I should.

Interesting historical notes on Wallace. He began his career as a pitcher, doing very well for Cleveland in 1895-96. He switched to 3B for 1897-98. After that year, the Cleveland owner put all his best players (incl. Young, Burkett, Powell, Childs and Cuppy) on St. Louis, whom he also owned. Lave Cross was sent from St. Louis to Cleveland to play 3B and to manage.

A few weeks in to the 1899 season, washed-up (injured?) StL shortstop Ed McKean was stinking up the joint. Cross was snatched back from Cleveland and Wallace was installed at SS, where he played almost exclusively for the next decade.

For Walsh, I have a career RA+ - based equivalent record of 218-111, compared to 230-152 for McGinnity. That's the record you would get from a career RA+ of 140, which is higher than everyone other than Walter Johnson. In about the same number of innings as Waddell, that's a 140-128 advantage. In 500 fewer innings than McGinnity, it's 140-123.

Walsh will be the highest ranking pitcher on my ballot this year. Although it may not come to a comparison, I haven't yet decided whether he's behind or ahead of Mordecai Brown.

As for Wallace: His offensive value is in the same neighborhood as J. Collins. That offense looks more attractive if you just look at 1897-1910 rather than 1894-1918. Note that he does have some value as a pitcher at the beginning of his career, and that being a 60'6" pitcher probably depressed his hitting. I don't know where to place him yet - the key is to decide how much defensive value he had.

In what I hope to make an "annual" presentation, here are the players who died in 1920:

HoMers: None

Here's my list of % above average value for the major pitching candidates, 1890-1913, with Walsh added. Listing includes all seasons in which pitcher was a full-time starter. Value is calculated by comparing pitcher's record to the record an average pitcher would compile with the same offensive and defensive support for an average number of innings. For innings a pitcher throws above league average, his value is compared to a replacement level pitcher, whose winning percentage equals that of an average pitcher divided by 1.2.

McG -- Gri -- Wad -- Wil -- Joss ? Walsh

Assuming the inductees for 1919 are the runners-up from 1918 (I haven't calculated the results), my prelim ballot is as follows:

1. Bobby Wallace

Not really close (though it is a weak field)

2. Jimmy Sheckard

I humbly bow to the PC police.

In fact, I think we all ought to be hyper-cognizant of whom we offend. For instance, has it ever occurred to anyone how the descendants of Bob Caruthers must be suffering over our treatment of him here? It's time we either elected him or cancelled this project entirely.

We're not talking numbers here. It's about ignorance and bigotry. Wherever the potential for this exists we must all do our utmost to eradicate these purveyors of evilness and disharmony.

I'm sorry, that was about as relevant as #25. Trolls. Yeesh!

Considering it is fall of 1919, Im glad we arent using the preferred nomenclature.

Still tweaking but it looks like my 1920 ballot will begin:

1. Walsh

Walsh is going to be #1 for me as well, I believe. He looks a bit better than McGinnity, and I have McGinnity pretty high already.

Monroe will definitely make the ballot, although I think he'll be in the lower half, and below Frank Grant.

Wallace will not make my ballot this year, I don't think. He wouldn't be a bad choice, but I see him, like Clint, as a Jake Beckley-type. He'll be off the ballot, waiting to come on in weak years.

Good questions re. WARP1, WARP3 and league adjustments. I use adjustedWARP1. The adjustment is to *2/3 season length, so Lip Pike is not getting a 2.5X or more boost, and then I use my own league adjustment, rather than accepting WARP season adj that I don't understand.

Did any major league baseball players die of the flu between the 1918 and 1919 seasons?

Ed Walsh's RSIs & adjusted W/L records:

PS:

>3. When Jay Jaffe used this method in his HOF article, he calculated the averages for each position with

So Walsh's career value is about the same as Waddell's, but his peak is higher. But, like Waddell, he's 500 innings short of McGinnity. His inning-for-inning value is a little higher than McGinnity's

Walsh's ERA+ 142

So Walsh's career value is about the same as Waddell's, but his peak is higher. But, like Waddell, he's 500 innings short of McGinnity. His inning-for-inning value is a little higher than McGinnity's

Walsh's ERA+ 142

Did any major league baseball players die of the flu between the 1918 and 1919 seasons?

I believe Larry Chappell (homepage) is the only major leaguer to die of the flu pandemic, although he was serving in the Army so was not "active" at the time of his death.

Based on my initial 20-second scanning of his stats at b-ref he reminded me more of Herman Long than Jake Beckley

Perfect comp in my mind, Chris.

I have to question the initial comparisons of Bobby Wallace to Jake Beckley. Wallace has a better career and a better peak.

Jake Beckley's fielding adjusted, season-adjusted win shares, ordered high to low

Herman Long, same data

Chris J. (#40): Looking at RA+ instead of ERA+ hurts Walsh a tiny bit and helps McGinnity a tiny bit - but it's still something like 140-123, and your comment but that's a little more than a "little" higher inning-for-inning value is appropriate.

On another thread, John Murphy has argued that deadball infielders had it easier than nineteenth-century infielders and that may mean that Wallace's career value is somewhat less impressive than McPhee's (though I would argue that his peak value is somewhat more impressive than Bid's).

I also noted that McPhee's was damaged more during his time than Wallace's time. With healthier hands during the Deadball Era, he probably would have been able to raise his offensive production relative to the league as a whole.

Is a twenty year career in our time worth the same as one starting in 1880 (everything else equal)? Of course not. Context, context, context.

On another thread, John Murphy has argued that deadball infielders had it easier than nineteenth-century infielders and that may mean that Wallace's career value is somewhat less impressive than McPhee's (though I would argue that his peak value is somewhat more impressive than Bid's).

I also noted that McPhee's peak value was damaged more during his time than Wallace's time. With healthier hands during the Deadball Era, he probably would have been able to raise his offensive production relative to the league as a whole.

Is a twenty year career in our time worth the same as one starting in 1880 (everything else equal)? Of course not. Context, context, context.

context, context, context.

But wasn't everybody else gloveless, too in the 1880s? If McPhee went gloveless longer, that was his choice.

But I really don't want to argue about the relative merits of Bid McPhee and Bobby Wallace. By the numbers, they are _very_ similar players. I think that's meaningful, John doesn't. Others can form their own opinions. The more important comparisons are to Wallace's eligible contemporaries.

But wasn't everybody else gloveless, too in the 1880s? If McPhee went gloveless longer, that was his choice.

I think the infielders had a tougher time nabbing line drives during that time (not to mention all of the basepath shenanigans from opponents) than the outfielders were facing.

As for McPhee wearing his glove longer than others, I agree that was his choice. But that really has nothing to do with his prime during the 1880s.

looking at Marc's prelim ballot, I assume that Keeler and Kelley got elected last year? :)

Shhh! :-)

Tentative ballot:

1) Dickey Pearce-SS/C

Considering it is fall of 1919, Im glad we arent using the preferred nomenclature.

Unfortunately, this is an understatement. We're in the middle of an ugly, ugly time in the U.S. Try googling such phrases as "red summer of 1919" or "tulsa race riot" to see.

Howie:

Bennett NL Ranks among Catchers by raw WS:

1878-7

Just looked a little more at Wallace. . . . I thought he only had a 5-6 point edge on Long at OPS+, so that separates them some. Since he's being compared to Jake Beckley, I thought I'd do the same thing for him - that I already did for Beckley: compare his OPS+ to his rivals at his position in year he was the starter. The results:

1897..3/12

Looking at John Murphy's ballot, I assume Keeler and Kelley got elected last year? ;-)

>Posted 4:09 p.m., February 9, 2004 (#50) - John Murphy

PS (or should I say PC, or "not PC"). I don't think Dan was trying to inform us that Bill Monroe was a Negro. I thought that he was informing us that he was in the Negro Leagues, which is what they were called, and that therefore he did not have WS and WARP numbers for him. That's what I thought that line on his post meant.

I'm with Chris (#42) on Beckley and Wallace. Another way to look at it is peak and prime. Here's what I've got for peak.

3 year consecutive peak adjWS--Wallace 78 Beckley 64

Marc, as usual you know right where I'm coming from.

Perhaps instead of "Negro" some would prefer I leave it "N-----"?

Which of course stands for "NonMLB".

Still favor Beckley, in part because I think a SS with a career OPS+ of 105 should have a Win Shares defensive grade better than "B." Ranks 34th in career defensive Win Shares, as near as I can tell.

I've often found that it pays to look behind a WS career defensive letter grade. Because they are a career rating for a player's work at a single position, they can sometimes be a bit misleading. Here's more detail on Wallace's defensive record.

First, a little more detail on the grade itself. Wallace gets a B. The "B range" for shortstops, as James defines it, begins at 5.00 WS/1000 defensive innings and ends at 5.80 WS/1000, which is the bottom of the A range. Wallace's career average is 5.46, squarely in the middle.

Second are two notable features of Wallace's career. One, he started as a third baseman, playing 3.5 years there (and playing very well). Since most players peak in defensive value early in their careers, some of Wallace's best defensive seasons don't count towards his shortstop grade. Two, although he was a starter at short or third for 15 years (a considerable stretch), he continued to play part time for quite a few years after that at shortstop, bringing his career rate stats down. So, I hypothesize that Wallace's letter grade at shortstop may underrate his defensive value over his career as a starter, which is what mostly matters in evaluating him.

Third, Wallace's individual seasonal performances suggest that he was, at his best, a truly outstanding defensive player. He won a WS gold glove at 3rd base in 98, and he won 3 gold gloves at shortstop in 02, 03, and 08. His best defensive season, however, was 99, in which he split the season between 3rd and short. He accumulated enough defensive WS to win the gold glove at short, if all of his defensive win shares had been earned at that position. He wouldn't have won it at third because Collins and Cross had even better years, but they were all within one win share of one another (Collins 10.9, Cross 10.5, Wallace 10.1) and were the three best defensive players in the league. So if you look at Wallace and see "B Shortstop," you need to remember that this B shortstop was good enough to win 5 gold gloves.

Fourth, if one looks at Wallace's defensive win shares as a defensive regular as a whole, counting his 3.5 years at third base and dropping his long decline phase as a part-time player, his rate of defensive win shares jumps to 5.83 WS/1000 innings, which raises his grade as a full-time player to A-, a grade that makes more sense of his gold gloves.

Wallace wasn't the defensive shortstop that Dahlen was: 6.83 WS/1000 innings over 17 seasons as a regular. He wasn't the defensive shortstop that Long was: 6.40 WS/1000 innings over 14 seasons as a regular. But it looks to me like he was an excellent defensive player, at third base and at shortstop, not merely a good one.

Marc wrote:

"Posted 2:14 p.m., February 9, 2004 (#34) - Marc

I guess what I?m saying is that if you do a Keltner list, Wallace does pretty poorly. (As an aside, why doesn?t anyone ever do Keltner lists for the HOM discussions?)

I see the Keltner list as a tool to evaluate the chances of someone getting elected in the HOF and not to determine a player's merit. (Just like the HoF monitor and black ink tests)

I did a Keltner list on Caruthers on the Caruthers thread; he aced it, 10 and 4 doubtful out of 15.

This is to JoeDimino,

I've heard people say before that since this is "your" project that your views will be taken very seriously. I think there is some truth to that and that puts a little extra responsibility on you to walk the fine line between making arguments for a player and telling people they are stupid if they don't see your way.

If John Murphy had made your argument above (I pick on John because he appears to have a thick skin) I would have paid attention to it but not felt a moral imperative to change my views (I have wallace 11th). Noone wants this to be Joe's Hall of Merit (including you of course) and I think sticking to the facts and losing the "no brainer" comments might help everyone. On the other hand, maybe everybody treats you like everybody else and this post is unnecessary.

Here's my prelim:

I have 29 players under consideration: 20 hitters, 9 pitchers.

Philip, I adjust WARP1 for season length at the *2/3 rate (not out to 162 games for reasons that have been discussed) and I apply league adjustments (only for contemporaneous leagues that are not the best at the time, not as a timeline).

With these adjustments, Sheckard's raw WARP1 doesn't change much. But all players who played seasons of less than 150 games or so, their adjWARP1 is quite different, and usually higher, than the raw score. So Sheckard doesn't drop, others rise.

Second, I rate mostly for peak and prime, as the above comparison of Beckley and Wallace. Adding Sheckard:

3 year consecutive peak adjWS--Sheckard 106 Wallace 78 Beckley 64

I think Wallace's detractors are knocking him for the same reason Ernie Banks, Cal Ripken, Alan Trammell and other offensive players are defensive positions are knocked. Because they compare them to other hitters, and not other SS's.

I've compared him to other SS, and I think he comes in behind Long and Jennings, which puts him somewhere about #18 overall. He only made 1 STATS All-Star Team in his 22 years in the league.

Daryn:

While I agree with you that (1) Joe has some added responsibility with respect to this project, (2) voters tend to be swayed by the fact that this was conceived by Joe and (3) Joe's recent post was a bit testy, let's cut him a little slack given his earlier report that he has switched job schedules and can't get as much sleep as he needs.

:)

I consider myself a peak and prime voter, not a career voter, and yet I have Wallace ahead of Long. Go figure.

To daryn and his reply to Joe, I understand daryn's point. But to daryn's support, anybody who has Wallace at #11 has nothing to apologize for.

With the possible exception of Bob Caruthers ;-) there are no right or wrong answers. Wallace could be #1, he could be #22. That's not near the spread we're seeing with Ed Williamson.

Thanks Marc,

I'll look into the seasonal adjustments a little more. Like you, I don't adjust on a season-per-season basis but live by the "pennant is a pennant" argument.

Andrew wrote: I need to be convinced that WS systematically underrated defensive players during his era. While I buy that story for the pre-1894 world, I have not yet seen any evidence to that effect for 60'6" baseball. Sure, WS overrates pitchERS during that era, but that is only because pitchers still pitched lots of innings and WS doesn't properly discount for opportunity cost/replacement level performance. I see no evidence that WS overrates pitchING or underrates fielders during the years Wallace played.

Andrew, could you explain a little more fully why you see pitchERS but not pitchING as being overrated by WS in this era? It's not immediately clear to me why "opportunity cost/replacement level performance" would lead to pitchERS being overrated. Are you saying basically that replacement level was high in this era, significantly higher than the WS zero point, so that pitchers who threw a lot of innings get too much credit? Or am I misunderstanding your point?

WARP adjusts season length in WARP3 at a rate of *2/3 (i.e. to the power of 2/3). This is a compromise between a 100% adjustment--i.e. all the way out to 162 games, OTOH which for players playing only 50-100 games in probably not realistic (and esp. for pitchers and catchers--IOW, a pitcher who pitches half of an 80 game schedule probably could not pitch twice as many games without his arm falling off). Or, OTOH, no adjustment at all, which seriously penalizes any player before 1900 and especially 1890.

If anything the *2/3 seems a bit light to me. I would adjust upward a little more myself but I use *2/3 because there is at least some kind of consensus about it (or is there?). Lip Pike has dropped down a bit because I adjusted my adjustments from 162 games to *2/3 about 5-6 years ago. McCormick and Bond dropped because of vastly reduced raw WARP1s.

That clarifies, Marc.

That clarifies, Marc.

That clarifies, Marc.

The last statement is true, but surprising. I wonder why nobody's ever tried using a knuckleballer as an "opener" -- pitching just the first inning to set the hitters' timing off. If you had a stable of good but fragile fastball pitchers, it might well make sense.

Thanks, Andrew! I hadn't thought about #2 having different consequences for pitchers in different eras, but I see that you are right. My sense is that #1 is still operative in the period 1893-1920 and I've been basing my assessments on that assumption, but before I present a case that WS is indeed underrating pitchING versus fielding, I'm going to look again at the WS formula for dividing WS between pitchers and fielders, which I haven't studied in a while. I've learned a good deal from this project since I last examined my position on the question, so I'd better think more before I speak . . .

Is the simple way of saying what andrew is saying the following:

There are a certain amount of fluff points in WinShares. Over the course of time for any individual this doesn't matter too much because everybody is getting them. But when pitching staffs include 2 starters rather than 5 (to pick numbers out of a hat), the fluff points are compressed for those 2 starters and this throws the analysis of those players out of whack.

Or put more simply, when 10 Fluff Shares (TM) are spread among 10 pitchers, who cares, but when they are all given to 1 or 2 pitchers we have to account for it.

Seeking clarification. Andrew thinks pitching pre-'94 is overvalued, and maybe '97-'19.

Chris wonders if deadball pitchers aren't underrated.

Is that right? Come out with your hands up, may the best man win!

How can dead ball pitchers be underrated? We're talking about enshrining in the HOM a bunch of pitchers with less than 250 wins, some with less than 200, in an era wen there were no home runs, pichers were allowed to scuff the ball, and they didn't "Bear down" unless the game was on the line.

Surely these bozos were all considerably short of Tom Glavine (251-157, in a much more difficult era for pitchers) and Waddell/Walsh were probably not even better than John Burkett (166-136.) Young and Mathewson yes, and I'd argue for Mcginnity and Brown, but even the latter two are borderline.

Caruthers, when you add in his hitting, and Mickey Welch were better than any of them except Young/Matty. 1880s pitching distance may have been shorter, but offense was much stronger.

Marc,

The issue Andrew and I have been discussing is chiefly concerned with the value of fielding in the deadball era (though the value of pitching is the necessary flip-side to that issue). Does the win shares system systematically undervalue fielding and overvalue pitching in this period, as most of us agree it does prior to 1893, or not?

If you follow the thread of discussion back to Andrew's post at #69, you'll see that Bobby Wallace is the player whose value is most in question. Does he have a peak high enough to merit a solid ballot placement or not? That may depend on whether or not his defensive value is underrated by WS. He's either just below a a key line (30 WS per season for several seasons running) or just above it. Andrew currently has him below that line because he sees fielding as properly valued by WS during the deadball years; I have him above it because I see fielding as undervalued by WS during the deadball years.

There are implications for pitchers, of course, but since there are about 77.5 different ways of evaluating pitchers out there, most of which don't deal with the comprehensive metrics at all, I wouldn't say that the way voters are evaluating deadball pitchERS is really at issue.

OK. Having gotten home and having had a chance to look at _Win Shares_ again, here's my explanation of why the Win Shares system continues to overrate pitching in the deadball era. The system establishes a proportion between pitching and fielding credit for defensive win shares by adding up "claim points" for various events. In the numerator go all the things that pitchers get credit for; in the denominator go all the things that both pitchers and fielders get credit for. On both top and bottom are some constants, also. The overall formula is designed to center the split at 2/3 credit for pitchers, 1/3 credit for fielders, with adjustments on a team-by-team basis to move this split away from this norm in one direction or another. It's a reasonable system.

The problem with it is that the proportions are calibrated to achieve an accurate distribution only when the events that contribute to scoring are themselves arranged in proportions typical of the modern game. The three key elements here are walks, home runs, and errors. Pitchers get a larger share of the defensive credit if they allow fewer walks than a league average pitching staff, and if they allow fewer home runs than an average pitching staff; fielders if they make fewer errors. Well and good. But the value of each of these elements in the formula is calculated off of a constant, to which the team's variance from league average is added. "The claiming formula for walks is 200, plus one for each walk or hit batsman that the team is better than average." (_WS_, p. 28). "The Claim Points for Home Runs allowed are 200, plus five points for each home run allowed less than expected" (WS, p. 29). Claim points for errors work the same way as walks, figured off of a constant of 100. The formula assumes that the overall contribution of walk suppression, home run suppression, and avoidance of errors to defensive value on the team level is constant. This is how the 2/3 ratio is maintained. A pitching staff that matches the league average in walks or home runs allowed gets 200 claim points regardless of how many walks or home runs they actually gave up, and average-handed fielders get 100 claim points. Thus, in low-walk and low-home run environments (relative to the modern game) pitching is going to be overrated somewhat by the win share formula, and in high error environments, fielding will be underrated by the claim point system. Claim points for strikeouts, interestingly, are not built on the same model. They vary freely, and so they can more readily reflect historical variance in pitchers' contributions to defense. But any time walks, home runs, and errors vary significantly from the level of significance in scoring assumed by the formula, pitchers will be overrated, or underrated, and fielders conversely underrated or overrated.

It seems to me that pitching in the deadball era is less overvalued by WS than pitching pre-1893. Its walk levels are much closer to the modern game, as are its error levels. But its home run levels are still _far_ from the modern game, and its walk rates and error rates are still lower and higher, respectively, than in the modern game.

I suggest, therefore, that those using win shares should continue to adjust fielding win shares upwards somewhat for players in the deadball era and even in the 1920s. I adjust fielding win shares to 1900 upward by 1.3, fws 1901-10 upward by 1.25, 1911-20 upward by 1.2, and 1921-30 upward by 1.1. These are rough and conservative adjustments, and careful study might determine more exact ones tailored to changing conditions, or adjust different positions at different rates (I have tended to adjust first-base defensive win shares upwards by 1.5, for instance). But I think some adjustment at least to 1930 is appropriate.

Refutations and questions welcome!

Boy, do we love OFs. Pitchers now only 8(plus Ward, sort of) out of 36.

HOMers (36) by position, roughly chronologically

CATCHER (2): Cal McVey (C-1B), Buck Ewing (C-1/0); see also White, Kelly

[insert standard apology about post length here]

Ever have one of those weeks were all the sudden every player on your ballot seems to be flying in all kinds of directions - some up, some down, some both, some sideways, some inventing new directions - so that you're not sure where you're going to end up putting anyone? I've got a half-dozen different reasons upsetting all kinds of different players up or near my ballot & right now it's so dang unsorted that I feel like I could just throw 'em all up in the air & they'll come down in about as sensible an order as I can figure.

But that's what this thread is for right? To help us figure out how we should order players. . . .

First thing screwing me up is I lost my ballot's anchors - K & K. I look at the remainders & feel like I should vote for 2-16 instead of 1-15. But that's probably just me being a relatively new voter. I though I had a sure #1 in Beckley, but the last few days have seen a major change in how I view him.

That's as good a place as any to start. I already posted my shock at seeing Beckley compared to Bobby Wallace, but the more I look at Wallace, the more sense it makes. Both were really good for a really long time, never really the best at what they did, but hung around forever. This has had two results: it's caused Wallace to move up in my estimation & it's caused Beckley to move down. Or more accurately, it's caused, Wallace to move up, & caused me to re-consider just how much emphasis I place on career value.

Not that I really feel I'm becoming a peak-er. Lord, no. But one comment someone (forgot who, sorry) keeps making on his ballot is that he entered this project expecting to find himself a career guy, only to find himself becoming more of a prime-r (no website pun intended). In my mind at least there is a distinction between prime & peak.

Still not really a prime-r, though. Ultimately, I'd like a guy with a really strong career & a really strong prime. Failing that, in general, I'd guy for a strong career over a strong prime. In general. Then a strong prime over a peak-er. K & K were nice because they were nice primes & nice careers. Beckley's got good career value, but his prime's weak. Ditto Wallace.

Rambling on elsewhere, there's Frank Grant & Bill Monroe. With regard to these early Negro Leaguers, I'm letting the expert's votes in the "Cool Papas & Double Duties" be my guide. They put Grant above Monroe & though many posters here I respect are putting Monroe ahead of Grant, for now I haven't seen enough evidence to sway me. I am considered that even the experts are merely groping with Grant, though, because he's so far back. Are they voting him higher simply because he's first? Could be. Not convinced of that yet, so Grant's still higher, but I have enough jitters so that I really don't want to put him first. . . .

So far, this is looking like Jimmy Collins will end up last man standing by default. And maybe he deserves it. Nice hitter, great fielder, best at his position for a long time, not enough 3Bers in the HoF. He'd be more a place-holder #1, than a guy I see as a real #1, though. Let's see if I can do better.

Well, there's the other new guy. Ed Walsh. First off, I deal with pitchers differently. Since their career at this time tended to be shorter (15 years for a pitcher is like 20 years for a hitter) & they did so much pitching - especially Walsh & McGinnity - in those years that I do weigh in peak a lot more. Let's compare him to Iron Joe, my current top pitcher. Joe's got him by about 500 innings, but Walsh has around 20 points of ERA+. I like Walsh. So how's he do against Collins? Hmmmm, I like Walsh's peak & prime a lot more, but Collins in terms of career value. Normally, this is an easy one, for Collins, but 1) I'm paying more attention to prime/peak for pitchers, & 2) I don't want to get too dogmatically boxed into one system. Right now I like Walsh more than Collins. Frankly, I like McGinnity a little more than Collins, too, but that's provisional.

As long as I'm on pitchers, there's Rube Waddell. He's dropping. Will he fall off? I'm an ERA+ man, but he let so many unearned runs score that I don't trust his ERA+ much. His 134 ERA+ is more like a 126+, & that, more than his run support, explains in W-L, IMHO. He gets some points for his peak, but how impressive was it really? Not as good as Walsh's. Not as good as McGinnity's. Not as good as Joss's. Heck, not as good as Caruthers, & I'm an EOBC. - Or am I?

Bob Caruthers. I'll be kicked out of the Enemy Brotherhood for sure for saying this, but all the sudden, he's making a hard charge for my ballot. I've said his pitching ain't enough to get on, & he didn't have enough at bats to sway me otherwise. Then a few days ago, I was flipping through Win Shares - James gives his bat 27% of his total value. . ?!?!? 27%? Maybe he does have enough at bats. I'll have to think about it more though. Just what I need - yet another player moving around on my ballot.

Charlie Bennett - He'll likely get helped with an increased emphasis on peak. Still like to compare 19th century catchers to modern day closers, but in his best years he was being used like Goose Gossage. He may move up a bit. Or not. I'm awful tentative on him - & with three newbies on the ballot he better be careful.

Jimmy Sheckard. On next year's ballot - assuming Tinker & Brown are on it - I'll have a nice waaaay tooo long post on that old Cubs team (short version: maybe best team ever, so Brown's overrated, Tinker underrated). For the same reason, I think I should move JS up a bit. But that's pretty minor.

Cupid Childs gets torn too ways. Aided by increased interest in peak, but hurt for a weird reason. I believe in trying (but not too hard) in balancing the positions on the ballot, & he's been aided by being such a good 2Ber. . . . Um, Chris - psst! - Grant played 2B. So'd Monroe. Um, yeah. Duh. (doinks self in head). Not sure where he's going.

Back to Wallace for a second. As I understand Chris Cobb's comment on his fielding, he's essentially saying that while Wallace was a fielder peak-er. For some reason my mind has a hard time wrapping itself around the notion of a hitting career-er & a fielding peak-er coming in one package, but it makes sense. Not sure what to make of that, though. For now, what I want to do is to look at him & Beckley. Is Beckley's edge in hitting (& games - slight, but larger than it looks because JB had a larger number of shorter season-ed leagues), & hitting enough to make up for Wallace's advantages in fielding & positional adjustment. Yeah. For now at least, yeah. Maybe I'm just seduced by the near 3000 hits, or it's force of habit, but I like Beckley better. . . . . So where's this leave me?

Provisional ballot:

Career votes-points leaders

"... one comment someone (forgot who, sorry) keeps making on his ballot is that he entered this project expecting to find himself a career guy, only to find himself becoming more of a prime-r (no website pun intended)."

Thanks for noticing!

Sigh, off to court. Not quite Martha, but the Sports-y one in the news....

I'll be posting my provisional ballot pretty late this year, as I need to get a proper handle on Monroe (and Wallace) which means research.

Big Ed Walsh seems to reaping major benefits of the new calculations of WAPR3. He went from 68.9 to 88.3 when the change took place. That's a nice run of years he had to establish peak value but I'm not ready to bestow Cy Young-type support for him quite yet.

Question about Wallace: I take it he was batting middle of the order? He has quite high RBI totals and low sacrifice counts when compared to teammates like Jesse Burkett and George Stone.

So, Howie, you're on the Jayson "Big Teddy Bear" Williams case?

Does a government agency sign your paycheck or a private firm?

Prelim ballot - earlier in this thread I said Wallace would probably be third, he ended up 4th. Sheckard moved up. Only other interesting note is that Pike is now in my top 30. Its going to have to get really weak for him to make my ballot though.

1. Walsh

FWIW, just looked back in Win Shares . . . though Bobby Wallace has "only" 87.5 Def WS at shortstop, at third he posted another 24.7 DWS (he averaged 6.73 Def WS per 1000 innings - higher than any other 3Ber listed - hell, it's almost as high as Dahlen's mark at SS - & even though he's only got 3661 innings there, he's got more DWS at third than anyone under 4750 innings). This gives him a total of 112.2 Def WS, which would put him tied with Tinker at 9th all-time among the SS listed, & among all players tied with Maz & Tinker & Fisk for 11th. Assuming that some players (such as Tinker, Fisk, & Maz & others just barely behind him like Morgan or Collins or Reese or a few others can move ahead of him in total career DWS, he'd still end up in the top 15-20 of all-time defensive value. And apparently had a strong peak as a defensive player. And was one of the better hitters at his position for several years. And lasted 25 years. He may end up moving up my ballot some.

On next year's ballot - assuming Tinker & Brown are on it - I'll have a nice waaaay tooo long post on that old Cubs team (short version: maybe best team ever

Correction: the above sould say maybe best defensive team ever. And another correction:

As I understand Chris Cobb's comment on his fielding, he's essentially saying that while Wallace was a fielder peak-er

There's a clause missing in that sentence - while Wallace was a hitting career-er.

RobC, Bill Monroe doesn't make your top 30? What do you see against his candidacy?

Just to make sure I'm not missing anybody...the N**** Leaguers who we probably ought to be considering this year are Grant, White and Monroe, right? Unless we want to revive G. Stovey, e.g. But no other newly eligible, just now?

Marc - that's my understanding, yes.

Chris J. -

I've got a sense of where you're coming from with the Cubs. When I look at them, the things I see include the following pitchers:

Buttons Briggs, after an undistinguished trial in his early 20's, suddenly reappeared (probably from the minors) at the age of 28 for the 1904 Cubs - and immediately had ERA+ of 129 and 138 in 1904-05.

Carl Lundgren had a 212 ERA+ in > 200 IP in 1907. Who's Carl Lundgren?

Jack Pfiester washed out of brief trials with another team at ages 25-26. At ages 28-29, he was with the 1906-07 Cubs posting ERA+ of 174 and 216.

Orval Overall had ERA+ of 115 and 65 in a year and a half with another team. In his first three and a half years with the Cubs his ERA+ was 140, 148, 122, 179.

Chick Fraser was a long career pitcher with a losing record. Still, at the ages of 36 and 37 he managed better than league average ERA in spot use for the Cubs.

Rube Kroh had only one season with > 100 IP: for the 1909 Cubs, ERA+ 154.

King Cole had ERA+ of 159 and 105 in two years with the Cubs and was bad after he left them.

Not everyone follows quite this pattern. You can't make that kind of case from, say, Harry McIntire. Ed Reulbach had a long, good career and was good for other teams. But the overall sense is of interchangable parts - that many different pitchers could succeed with the Cubs.

At the moment I've got Frank Chance rated a little higher than the consensus. I'm prepared to see Tinker and Evers in a favorable light, and to be a little skeptical of Brown (although Brown was certainly very good.)

Among the old Cubs I've got it 6-4-3. Tinker-Evers-Chance. But Sheckard and Brown, in that order, top all three.

Brown had the best peak, though Sheckard's wasn't bad if you don't insist on consecutive seasons. Old 6-4-3 didn't have peaks, even though Evers won an MVP, and not an undeserving one though I wouldn't have picked him myself. Chance not only had a short career but, really, not much of a peak. Not enough games in his best years, at least not enough to overcome the lack of peak/prime values.

Tinker was the least famous of the three, but better than advertised. OTOH none of the three will ever make my ballot. Brown will, Sheckard already has.

Chris,

RobC, Bill Monroe doesn't make your top 30? What do you see against his candidacy?

He doesnt even really make my radar screen. I havent completely finished integrating (heh) ChrisJ's info into my Negro League rankings, but even from it, it seems that at the best, Monroe was maybe the 4th or 5th ranked 2nd baseman. My impression is that 2B is also the weakest position of the possible negro league candidates.

Should he be on my top 30? If I was voting for a top 30, he might be there. But that list is actually a top 30 amongst players still on my consideration list. Monroe isnt on my consideration list, so he doesnt get ranked. Everyone in the top 15 gets added to my consideration list, so he might make my top 30 next year if he makes the top 15 overall.

So, Howie's in my neck of the woods. I live about 15 minutes from Somerville, where the trial's taking place. Oh, and odd bit of trivia: I went to the same high school as the motorcycle accident Jason Williams, although 10 years before he did.

I guess the moral is if you're athletically inclined, live in Central New Jersey, and are named Jason Williams, pick a sport besides basketball.

Devin, don't you mean pick a sport besides motorcycles and guns.

Let me preface this by saying I'm sure I'm wrong, I'd just like someone to explain in short words I can understand.

Sorry to be late with this, but Chris Cobb in post #92 says that in Win Shares, James errs by not taking into account the league environment for walks, home runs, and errors. However, for strikeouts he says they vary freely, and therefore adjust to the league environment. I'm not sure this is the case, because as far as I can tell, James is trying to assign the same relative weight to strikeouts, walks, home runs, and so on. It seems counterintuitive to me to argue that the numbers that ARE league-adjusted do a worse job of reflecting the league environment than the ones that are. They're all based off a constant, it's just that James assumes a certain league-average rate for strikeouts. And double-plays and DER are also used, and I know double play rates have varied over time. I guess what I'm saying is that I don't understand, from Chris' explanation, why he's making the adjustments that he does, only considering some of the elements that James uses to split up defense and pitching.

Marc,

Devin,

Thanks for responding! Here's a fairly quick response.

I focused on walks, home runs, and errors because they were the elements whose variability seemed to me the most significant. I should perhaps have included double plays. James _does_ handle strikeouts differently (see p. 28 in _Win Shares_). Defensive efficiency's role in the formula is a complicated problem, but I don't think James' use of it is introduces serious historical distortions into his formula.

I'm not sure this is the case, because as far as I can tell, James is trying to assign the same relative weight to strikeouts, walks, home runs, and so on. It seems counterintuitive to me to argue that the numbers that ARE league-adjusted do a worse job of reflecting the league environment than the ones that are.

That James is trying to assign the same relative weight to each of these elements is exactly the problem. When teams give up something like 18 home runs in a season in 1905, is home run prevention as important a contributor to run prevention (a contribution that must be credited wholly to the pitcher) as it is in 1985 when teams give up something like 120 home runs in a season? The win shares formula says it is.

When James "league-adjusts" defensive numbers in his formulas, he is eliminatng the kind of error that would show all 1800s fielders to be terrible because they commit lots of errors and fail to turn lots of double plays. Fielders and pitchers are good or bad relative to their opponents. This sort of league adjustment is a necessary step, and I'm not arguing against his practice of assessing defensive performance by comparing teams to their league average. However, when James incorporates his league-adjusted elements into the larger formula by using constants that are not historically varied, he performs a second kind of normalization that assumes that all elements of defense have the same importance throughout baseball history.

The basic question is: given an average defense, how are you going to allocate shares in defensive credit between pitchers and fielders? Because all of James's measures except strikeouts are based off of constants that are not historically adjusted, the ratio of credit assigned to pitchers and assigned to fielders for an average defense will always be the same, except for the variance created in changes in strikeout patterns.

Can this be accurate?

Here are the defensive stats for average teams at 3 points in baseball history:

1879 NL

DE .7178

Chris, thanks, that does make things clearer. I guess the only thing I'd really disagree with you on is that I don't think the treatment of strikeouts is really that different than any others - they're all based around a certain-sized chunk of the claim points, which doesn't vary over time.

I agree with your confusion as to why James doesn't vary the proportions over time - it doesn't seem like it would add THAT much complexity to the formula.

Devin, strikeouts are a side-issue, really, but they are handled differently. Every other component has a lump sum -- 100 or 200 -- to which points are added or subtracted according to the team's variance from league average. Here's the formula for strikeouts:

1. Figure the team's strikeouts per 9 innings.

Provisional ballot, no discussion yet.

1. Walsh

FWIW, just took another glance at the expert voting in the "Cool Papas" book. In determining the best Negro Leaguers not in the Hall, 10 of the more distinguished & knowledgable of early Negro League ball have their ballots put to the side & turned into a special committee on early Negro Leaguers. Five of those ten had Grant on their ballot. Two had Monroe on their ballot.

Nine of those ten guys were invited to participate in the 2nd election (all-time Negro League team) & eight consented to it. None of them voted for Grant or Monroe for the all-time Negro League team.

It has been mentioned how pitchers' hitting took a tumble after the pitching changes 1892 to 1893. Some voters seem to place a premium on pitchers who hit well, while dunning those who hit poorly. I looked up the career OPS+ for 42 leading pitchers in the McGinnity-Walsh era, from the late 1890's to the mid teens. Here's the ranking:

100 George Mullin

The homepage is a link to a Primate Studies thead about a graph that Bob Mong put in his blog about average year of individual peak performance over the years. I made an effort at explaining some of the features in a post on that thread, although I could have gotten it wrong.

1920 prelim.

1. Ed Walsh Between 1907 and 1912: 5 seasons with huge workloads and 5 seasons with ERA+ in the AL top 3. Incidentally, has the all-time lowest career ERA. Per DanG's post #176 in the New Eligibles thread, is 3rd in career WARP3 (behind only Mathewson and Plank) among all pitchers eligible between 1919 and 1932. If he doesn't get elected relatively quickly, the standards for electing pitchers are too high.

2. Hughie Jennings

3. George "Rube" Waddell His career WARP3 is higher than McGinnity's, although McGinnity pitched 480 more innings. Keeping in mind that the zero-line in WARP is such that a 0-WARP peformance is actually below a real "replacement-level player" performance, this implies that Waddell plus 480 innings of anybody does more to help teams win than McGinnity alone. If true, it would make it hard to support McGinnity above Waddell on a ballot. This analysis is dependent on relatively large league factor splits between the (stronger) AL and the (weaker) NL, such as those that Davenport uses. I have done my own analysis, and find that his league factors are plausible. Admittedly, it ain't an exact science.

4. Lip Pike Pro: In documented years, almost always best at his position. Evidence that he was the fastest player in the game suggests that he was a defensive asset. Con: Making the ballot depends on credit for being an 1860's star, with the usual uncertainties.

5. Bobby Wallace The "Jake Beckley of SS" is not a bad comparison, but, then again, I have Beckley 6th. Pitching years help his case. Considering that his career WARP3 is above 100, I would have expected more overwhelming support here.

6. Jake Beckley

7. Addie Joss Addie Joss: 2327 IP, 142 ERA+ John McGraw: 4926 plate appearances, 30.51 WS/162. Frank Chance: 5099 PA; 29.86 WS/162. Obviously, I am not too concerned about a short career if the rate stats are historically great. McGraw and Chance are also good candidates because their position/decade combinations are not represented in the HoM. OTOH, it's perfectly fine with me if your voting philosophy disqualifies candidates whose careers were this short.

8. John McGraw

9. Dickey Pearce

10. Frank Chance

11. Jimmy Collins

12. Jimmy Ryan

13. George Van Haltren

14. Cupid Childs

15. Frank Grant

Previous Top 10, not on ballot:

Charlie Bennett. Great analyses and arguments in his favor in Marc's post #68 in the Catchers thread. All along, I've prorated catcher playing time relative to typical catcher usage, and also given a slight bonus to catchers for the fact that, like pitchers, there are effectively more of them in the roster than there are players at other positions. Bennett still, unfortunately, doesn't make my ballot, because his period of stardom was too short. The catcher shortage in my system is not severe enough at this time to compel me to give any additional positional bonus. Bresnahan had a longer period of stardom at C (and OF), and will enter my ballot around #4 in 1921.

Joe McGinnity. See Waddell comment. Since I'm one of the few people with McGinnity off the ballot, I rechecked my data, and didn't find any errors. Iron Man simply had too many seasons with less than spectacular ERA+, in the weaker of the two leagues, to rate well in my system.

17. Jimmy Sheckard. (behind 16. Hugh Duffy) Several factors: my positional balancing effectively prevents too many outfielders from making the ballot; JS falls short of Delahanty, Burkett, Clarke, Kelley, and maybe Magee among contemporary LF; the NL during his career was relatively weak. Maybe Cooperstown got this one right. Ought to compete head-to-head with Magee.

Bob Caruthers: The center of my rating system is to determine a player's established ability to win games, on an average team in a neutral league, in standard deviations above average for their position and era. Combining this value with career length gives a good indicator of Hall support, whether of Fame or of Merit. The problem with Bob is that his level of play/career length combination falls well below levels that typically make one a serious candidate. While any rating system has flaws and uncertainties, what I see is enough to make me worry that electing Caruthers to the Hall could be a mistake. I would advise extra caution and skepticism while looking at his numbers.

Sam Thompson: Again, my positional balancing prevents too many outfielders from making the ballot, and I rate Thompson as only the 9th greatest outfielder of the 1890s, behind Delahanty, Hamilton, Burkett, Kelley, Ryan, Van Haltren, Duffy, and Tiernan. An intriguing 2004 comparison is Juan Gonzalez, who also has great slugging stats and RBI totals, but may not stand out enough from contemporaries to make it to the HoM.

Heh. Big Six had the same career OPS+ as Rey Ordonez currently does...

>3. George "Rube" Waddell His career WARP3 is higher than McGinnity's.... This analysis is dependent on relatively large league factor splits between the (stronger) AL and the (weaker) NL, such as those that Davenport uses.

Thanks for pointing out the league factors, Eric. I can't offer anything other than skepticism--i.e. no alternative theorems or proofs. But I thought everybody should see the raw numbers.

CAREER

WARP1--McG 71.7 Waddell 66.3

As a follow-up to my WARP league-adjustments post disguised as a McGinnity-Waddell post, now here's a WARP league-adjustment post disguised as a look at the 1904 season. It is the one season in which both McGinnity-Waddell both approached their peaks (>10 WARP1) and it comes about midway in the first decade of the AL-NL era.

I created 2 sets of 1904 pitchers--one AL, one NL. It is made up of pitchers I have had in my HoM consideration set at one time or another, plus pitchers in the top 5 for the season in TPR, plus pitchers on annual all-star games I had created at one time. Total 14 NL and 15 AL pitchers, basically (but not quite precisely) the top 29 pitching seasons that year.

AL--Altrock, Bender, Bernhard, Chesbro, Dinneen, Joss, Mullin, Owen, Plank, Powell, F. Smith, Tannehill, Waddell, D. White, Young

Total WARP1 104.8 WARP3 92.9 Difference -11.4%

NL--Briggs, Brown, Flaherty, Hahn, Harper, Leever, Lundgren, Mathewson, McGinnity, Nichols, D. Taylor, J. Taylor, Weimer, Wicker

Total WARP1 99.8 WARP3 82.0 Difference -17.8%

The totals are about what a <inner circle but still HoM career might look like. But the difference between 104 and 99 is not much. The difference between 92 and 82 is a bunch. Why?

Here are my all-star teams for that year (3 deep; AL on left, NL on right)

C-Sullivan, Criger, Sugden Kling, Grady, Dooin

I see Walsh as a truly great player with a short career. The other guys I go back and forth on, none of them are what I would call compelling. I guess that's why they haven't been elected yet.

1) Walsh?Some voters think that his incredible run from 1906-1912 isn?t long enough. As I figure it, just 1907-1910 would be enough for my vote. Those are BIG seasons. HUGE seasons. A seven year peak by a pitcher is much more uncommon than a seven year peak for a hitter. Ed Walsh is one of the all time greats, everyone else should be fighting for #2 this year.

Al Peterson #98

Good stuff of late.

I do sense some lapses in critical thinking in some votes; the old 'I've found the Holy Grail stat, now I just have to make my list in the order that eligible players are in.'

Thus, says Clay Davenport concerning the old and new WARP3, Walsh "doesn't gain 19, he gains 13 (8 of it fielding)".

This got me to thinking. Can a pitcher's fielding really amount to +10% of his value? I don't know. What I did find was that in the early AL, especially thru 1907, pitcher's TC/G were at a historic peak. The NL 1902-05 had most of the other top years.

Top 15 Leagues in TC/G for pitchers

Prelim ballot sans old exhibits:

One new exhibit added, Walsh. McGinnity makes a move up over the OF glut. Pike and Pearce move to ?Clearly deserving? status. Bobby Wallace takes Kelley?s #2 spot and Ed Walsh moves into Keeler?s #5 place. In 1921 Bresnahan, Leach and Tinker represent the three great NL teams of the era 1903-13. In 1922 we?ll coronate Mathewson and Lajoie, while ?Three Finger? and ?Home Run? also debut.

1) Bennett (1,1,3)?

'Bob Caruthers?Say hitting was 25% of his value. His pitching career is not 75% of a HoM career, when you consider AA discount and the shortness of his career.'