(More a curiousity than an argument.)

Though when I first noticed it, it persuaded me to vote Pearce up high.

... and not just one species, but an entire ecosystem in the bullpen. (Think 1990's.)

What sort of animals live in an Eckersley?

Or maybe you mean all those Thigs running around out in the Thigpen?

I have to say, this will be a fun election. First one since before Big Train and Wheat hit the ballot that I didn't know exactly who would be elected.

A question about Spotswood Poles -

His support peaked in 1931 when he placed in 17th place and received votes on 25 of 53 ballots. By 1938, however, he has dropped to 37th place and appeared on only 5 of 53 ballots. While we might attribute this drop to the arrival of new candidates over the last 7 years, I think there is more going on. Other candidates who placed just above or below him in 1931 have also lost support and now appear on fewer ballots, but not to the same degree as Poles:

pos('31) Name     bal('31) pos('38) bal('38)

15       T Leach  26       16       26
16       B Monroe 24       24       16 
17       S Poles  25       37        5  
18       L Doyle  19       22       13
19       H Hooper 19       25       12
20       C Jones  15       21       13
21       B Veach  19       26       13

Also, in spot checking a few people's ballots, I noticed that in 1931 several voters ranked Poles ahead of other candidates who still appear on their ballots in 1938, even though Poles no longer appears. So several voters have apparently purposely downgraded him.

So what's the reason? Has there been new analysis of Poles that has weakened his standing? Or is it simply the case that he has been especially hurt by the arrival of two relatively strong center field candidates, Carey and Roush?

I'm curious because I've been carrying him fairly high on my ballot and I'm wondering if I may have missed out on some new analysis that may have hurt his case.

I just finished this. Does this detail help voters, before too many more years?

team game # / opposing starter / result for Luque or Rixey:
Luque vs. Rixey 1921
Luque v. Bos tm finish: 4th 79-74
031 Fillinghim L
046 McQuinlan L
074 Scott L
089 Fillinghim W
125 Fillinghim L
143 Fillinghim W

Rixey vs. Bos
047 Fillinghim W
075 McQuinlan L
093 Scott L
124 McQuinlan W
144 Scott (starter) w/ McQuinlan (dec'n) L

Luque vs. Brooklyn 5th 77-75
023 Ruether L
051 Smith W
077 Cadore L
098 Cadore L
117 Cadore W
148 Miljus W

Rixey vs. Brooklyn
026 Grimes W
054 Pfeffer W
101 Mitchell L
116 Ruether L

Luque vs. Chicago 7th 64-89
013 Freeman W
019 Hanson W
062 Cheeves L
070 Martin W

Rixey vs. Chicago
018 Freeman L
072 Cheeves L
088 Cheeves W
128 Alexander L
149 Cheeves W
152 Freeman/Jones L

Luque vs. New York 1st 94-59
027 Benton L
085 Douglas W
094 Barnes L
113 Douglas W
138 Nehf/Shea L

Rixey vs. New York
050 Toney W
084 Toney W
095 Douglas W
114 Barnes L
139 Douglas L

Luque vs. Philadelphia 8th 51-103
035 Meadows L
055 Smith W
059 Baugartner L
102 Smith W
122 Meadows W

Rixey vs. Philadelphia
079 Smith W
103 Hubbell/Meadows L
119 Hubbell L

Luque vs. Pittsburgh 2nd 90-63
001 Adams W
009 Zinn L
028 Glazner/ Adams L
066 Yellowhorse L
130 Glazner L

Rixey vs. Pittsburgh
003 Hamilton L
011 Adams L
021 Adams W
037 Cooper L
060 Cooper L
067 John Morrison L
131 John Morrison W

Luque vs. St. Louis 3rd 87-66
005 May W
016 Pertica L
042 Walker W
106 Pfeffer L
135 Pfeffer W

Rixey vs. St. Louis
007 Haines W
015 Riviere W
043 Haines L
064 Walker W
108 Doak W
111 Walker W
136 Haines W

Cincinnati finished 6th 70-83

Since this is the second-to-last of the backlog elections before about 13 years of almost automatics, I wanted to add my voice to those who are asking every voter to deeply consider your vote.
Second, I urge caution toward the newly eligible. See how they compare to long-time eligibles. For example, Red Faber and Vic Willis have similar records - is it because they are very comparable pitchers or does the difference in era wipe out any superficial similarities? A second example would be the comparisons done above relating to Bancroft, Sewell, Maranville, and earlier shortstops.
If one why not the other? I know the constitution does not require this and I am not asking voters to defend everything to the minutest detail, but just to reconsider all the candidates - not just the newer ones.

Third, my voting philosophy is Mickey Welch or Death. Just kidding. I consider career win shares, win shares in 3 cons. years, the seven best win shares years, and win shares per 162 games/40 starts with the 7 year period weighted the heaviest, followed by about equally 3 yr and per year measures. In addition, I consider the number of times a player is the best at their position in the league and the majors and their competition. Lastly, I look at black ink/grey ink scores and in what categories they did best.

When I make my ballot I try to follow this passage from the Politics of Glory, pp. 130-132. I am paraphrasing a bit.
1. We are looking for the BEST candidate, not merely a qualified candidate.
Ask, is he the best pitcher who is not in the Hall of Merit? is he the best shortstop not in the Hall of Merit? Once the best have been identified, ask who is the Best of the best. Also, care about the Hall of Merit - keep it, its history, and its future in mind.

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 Hall of Merit is a point in favor of another candidate.

4. The fact that several comparable players are Hall of Meriters is an important element of proof for any Hall of Merit candidate.

5. It is important, in evaluating a Hall of Merit candidate, to show awareness of comparable players who are not in the Hall of Merit.

6. The HIGHEST Common Denominator Argument: This combines 4 and 5. There are many players with comparable records who are in the Hall of Merit, and there are NO players with comparable records who are NOT in the Hall of Merit.

7. The fact that a player meets the statistical standards of previous Hall of Merit selections should be counted in his favor. What standards should do is shift the ground of the argument. Start by asking, "Why should he be in the Hall of Merit?" Then, look at the established standards, if the candidate is at or above them, then ask, "Why shouldn't he be in the Hall of Merit?"

8. If a player is truly in a group of Hall of Meriters - in the middle of the group - that should be counted in his favor.

Thank you for your time reading this. There are a small group of us doing this. Whether you post a lot or just read and post a ballot, you have committed to making the most informed selections you can.

Why do I feel like Tom Cruise in Jerry Maguire where he has the epiphany in the middle of the night and then morning happens??

To take Mr Welch and the above standards:

1. I feel he is the best pitcher who has not been elected followed by Griffith (after some more recent thinking). I feel the best first baseman is (wait, let me draw a card ;)) Frank Chance. I feel the best second baseman is Cupid Childs. I feel the best shortstop is Joe Sewell. I feel the best thirdbaseman is Tommy Leach. I feel the best outfielders include Pete Browning, Charley Jones, George Burns, and Hugh Duffy. I feel the best catcher is Bresnahan. Based on everything, I think Mickey Welch is the best qualified.

2. The combined weight of the evidence. First off, I don't understand WARP. I understand many people here favor it as a catch-all stat. Welch does not do well in WARP (I don't know what the most recent iteration says for Welch.) Second, ERA+. Welch's is a 113. Among all HoMers, only Galvin's is worse. My personal issues with ERA+ are two: that era was not an official stat until 1913. I don't know how early before that it was calculated. I believe (without pure facts to back me up-just anecdotal) that some or most pitchers would coast in games - especially blowouts in either direction - from the start of baseball up to about 1920. Combining these two, I don't know how much stock to put into ERA or ERA+. It is a factor, but for me, it does not have the same relevance as it does in later years.
On the other hand, Welch's record against other HoMer /HoMer caliber pitchers is incredible. He went 62-38 or 62-36 (depending on which count of his games against Pud Galvin I use/find). NO pitcher/HoMer comes close to this. (I still have to do Johnson, Young, Plank, Matty, and Nichols.) Only Brown and Walsh had similar percentages but those were in many fewer decisions.
Also, he won 309 games. And unlike his contemporaries, he did not have the benefit of great offensive support. For whatever reason, over his career he did not receive as many runs as his teammate Keefe or his competitors Clarkson and Radbourn. I trust the information on Chris J.'s website that shows the Welch overachieved based on his run support and defensive support while the others underachieved.
In response to those who detail that Welch had a lower number of appearances against 1st/2nd place teams compared with Keefe, I bring up that Welch has the same number as Clarkson and 6 fewer than Radbourn. And he did this while pitching with another HoMer for most of his career.
I have frequently compared what Welch did while teamed with Keefe and vice-versa. The numbers have always had a slight edge to Keefe. The point is that Keefe was easily elected at the start of the process. When one is teammates with a HoMer and your numbers when surrounded by the same teammates are just slightly less, shouldn't that count in your favor as a demonstration of your ability and merit?

Regarding arguments about the number of innings or win shares which would be pitched in the National League during the 1880s if Welch were to be elected: Most comments have been in the vein of "it would be about 25% of the innings/win shares/etc., if Welch were to be elected." But why is this number wrong? It may feel high. I think it depends on how you feel greatness is distributed and then realizing the incredible differences in how pitchers were used in pre-93 baseball versus current day (1939).
The great number of innings pitched in the National League comes from 4 pitchers - Keefe, Galvin, Radbourn, and Clarkson plus 2/3 of Ward
(1500 IP) and the start of Rusie (225 IP in 89). And Caruthers should be included. If one only considers the National League, from 1880-89 there were 77105 IP and 15871 IP (20.6%) by HoMers. Including the AA which we should because Caruthers, Galvin and Keefe all pitched in it, there is 146,201 IP and 13.4% is HoMer (16.4% with Welch). Playing with the numbers of fun, but it shows that the high percentages are due to the ways the pitchers were used. 1-3 man pitching staffs were the norm during the decade. OF COURSE the innings are going to be concentrated in a few pitchers. Of course making one additional pitcher a HoMer while drastically add to the percentages. It is a direct result of how the game was played.
Because there were fewer pitchers needed, there were fewer pitchers used. That does not mean there were great pitchers waiting to be found. On the contrary, that means that great pitchers would pitch a higher percentage of innings. No one throws 400 or 500 innings now (1939). 200-300 innings is the norm. So more pitchers are being used. Do we see double the great pitchers of the 1880s? Do we have 10-14 great pitchers who belong in the HoM?

Regarding Comparable players inside and out of the HoM: Of the 7 most comparable pitchers by similarity scores, 5 are in the HoM: number 1: Radbourn, 3: Keefe, 4: Clarkson, 5: Nichols, 7: Plank and one is not eligible yet, number 6 Grimes. Only Tony Mullane is not in the HoM. I think when that many of one most similar players are HoMers, it is a definite positive.

The only eligible, but not elected player, spent his much of his career in the league with less talent. His 6 best years came in the AA and 2 of 3 biggest seasons happened in the 2nd and 3rd years of the AA, before it reached near-parity in quality with the NL.

Previously, someone commented that Welch had to be better than the other pitchers of his era already in the HoM. That has NEVER been the standard of the HoM. Groh was not better than Baker. Coveleski is not better than Alexander, Williams, or Johnson - the 3 most recent pitcher elected. Let alone Thompson to other leftfielders, or Wallace to Wagner, or presumably, Carey to Cobb and Speaker. To ask Welch to be better than those already in is applying a higher standard to him than we have applied to many of those we have elected.

Provisional:

1. Red Faber -- I came into the week thinking Eppa would be #1, but I've changed my mind since.

2. Jake Beckley
3. Eppa Rixey
4. Mickey Welch
5. Jose Mendez -- Let's see anyone call me a non-friend of the pitcher after his week!

6. Gavy Cravath
7. Roger Bresnahan
8. Lip Pike
9. Max Carey
10. Pete Browning
11. Bill Monroe
12. Clark Griffith
13. Dick Redding
14. Vic Willis
15. Cupid Childs
---
16-20 Ed Cicotte; George Sisler; Hughie Jennings; Ed Williamson; Bruce Petway

21-25 Rube Waddell; Wally Schang; Ed Konetchy; Tommy Leach; Ben Taylor

26-30 Joe Tinker; Joe Sewell; Larry Doyle; Urban Shocker; George van Haltren;

31-35 Bobby Veach; George Burns; Carl Mays; Ray Schalk; Edd Roush

36-40 Dave Bancroft; John McGraw; Hugh Duffy; Jimmy Ryan; Johnny Evers

... and not just one species, but an entire ecosystem in the bullpen. (Think 1990's.)

What sort of animals live in an Eckersley?

Or maybe you mean all those Thigs running around out in the Thigpen?

I assume you're kidding. But just to get it print, I was thinking of Wetteland - who has a plausible argument for best reliever of the 90's, not that that's worth very much.

To ask Welch to be better than those already in is applying a higher standard to him than we have applied to many of those we have elected.

It is a higher standard, but it is an appropriate standard for any candidate from the period with the heaviest representation in the HoM. It's readily demonstrable that we have elected more players from the 1880s than from any other period in baseball history. Players that we are electing who are not as good as others at their _position_ are clearly filling in the middle tier of HoMers for the teens and twenties. Within their period groups, they are not borderline candidates.

There's a strong argument to be made that we should not induct any more players from the 1880s at all. For Welch to have a good case, it ought to be clearly be demonstrated that he is not just similar to those already elected, but _better_. If he is similar (but maybe slightly worse), then he should probably be judged to be just on the wrong side of the in-out line. Only if he is clearly _better_ can his addition to the already well-filled 1880s wing of the HoM be justified.

Is anyone else getting this error when they try to get into primer/btf? Or is it my machine? Thanks!

MySQL ERROR:

Query: SELECT exp_members.weblog_id, exp_members.tmpl_group_id, exp_members.username, exp_members.screen_name, exp_members.member_id, exp_members.email, exp_members.url, exp_members.location, exp_members.last_visit, exp_members.total_entries, exp_members.total_comments, exp_members.language, exp_members.timezone, exp_members.daylight_savings, exp_members.time_format, exp_members.last_email_date, exp_members.notify_by_default, exp_member_groups.* FROM exp_members, exp_member_groups WHERE unique_id = '2adfe18589209566933b4a133f925a9d1dfff6a2' AND password = 'f49e42dc0b906db3d3fa6bc9e411c9fac67b54d6' AND exp_members.group_id = exp_member_groups.group_id

I've been getting something like that; I assume it's a BTF problem. But deleting cookies may help - try that.

"It's readily demonstrable that we have elected more players from the 1880s than from any other period in baseball history."

Average HoMers playing per year for each decade:

1870s: 9.8
1880s: 21.2
1890s: 22.5
1900s: 22.8
1910s: 21.5

Unless you are a timeliner or a big NA fan, I see no reason to believe any decade is over or underrepresented.

Further detail on a previous chart:
(these are all 10-game minimums)

HOM ELECTEES
1881-84 - 20 to 22
1885-89 - 23 to 25
1890-92 - 29 to 30
1893 - 26
1894-1903 - 20 to 22
1904-14 - 24 to 27
1915 - 21
1916 - 26
(then down to 20 in 1917, and declining further due to lack of time to elect sufficient players so far)

Welch's main years are 1880-90. Those years don't appear to be all that out of whack. The discrepancy, if one wants to call it that, would be 1890-93 vs 1894-1903 vs the burgeoning 1904-14.

I think anyone who is downgrading Welch for being in a 'glut' should take another look, unless they are timelining or whatever it is we call downgrading due to perceived inferior earlier competition.
That is, unless someone takes the time to look even closer, and finds more part-time players in certain years skewing the figures. I haven't gotten that far yet.

With a study similar to Howie's, I offer more on Welch and HoM gluts.

The decade-only breakdown is a bit too blunt an instrument to show the degree of representation well. Based on year-by-year HoMer counts, I’ve calculated the Been-there-done-that Index (Btdti) for each of the players among the top 20 returning eligibles whose careers fall between 1880 and 1915.

Pete Browning 24.67 HoMers per season
Rube Waddell 23.83
Hugh Duffy 23.81
Jake Beckley 23.75
George Van Haltren 23.47
Cupid Childs 23.33
Mickey Welch 23.25
Hughie Jennings 22.83
Clark Griffith 22.46

Welch is fairly low on this index, but much of his advantage comes from his play in a single-league environment in 1880-1, when there were 17 and 20 HoMers active in an eight-team league. That seems like pretty generous representation. If Welch is ranked only on 1882-1891, his Btdti is 24.2.

I would also argue that the total number of players admitted in the 1890s and 1900s should be 1-2 players _higher_ than the 1880s, because of the expansion in MLB’s population base with the spread of the game in the South and among African-Americans.

Conclusion: the raw Btdti does not show Welch as playing in seasons that are more heavily represented than most of the long-time backlog candidates. It does, however, confirm that at least two of the remaining 1890s candidates come from a period that is notably less well represented.

Preliminary ballot

I'm now a Win Shares voter. I understand its mechanics and flaws better than WARP's. I’m also balancing pitching and career with hitting and peak more than ever.

1. GVH: Best career total on the board.

2. Duffy: Best peak (non-Jennings division) on the board.

3. Poles: Good combination of peak and career.

4. Mendez: Best pitching peak on the board, plus enough shoulder to make a career out of it.

5. Rixey: A long, productive, if not peaky, career; top five in league 6 times.

6. Burns: Really sweet peak/prime, not enough career to get higher.

7. Roush: Could be swapped with Burns quite easily, and might be because he was a CF.

8. Jennings: That peak is big enough to merit a mid-ballot position, but no higher.

9. Leach: The more I reexamine him, the more I like. He’s not peakless, and he’s got a nice, long career with plenty of value at two key positions.

10. Cooper: Like Leach, the more I see, the more I like, and perhaps he'll move up. Led his league’s pitchers in WS once and was in the top five another six times. He would be close to automatic if he’d had just a little more career value. As is, it’s quite possible that both me and the HOM are underrating him.

11. Childs: Nothing new to offer except to say he’s enough better than Dobie to make my ballot.

12. Monroe: He’s just a smidge better than Doyle, who is just off the end of the ballot.

13. Willis: Led his league twice in WS, finished top five three times, but underachieved his projected W/L record. A couple more years and he, too, would have been nearly an auto-pilot choice. Or else Mickey Welch.

14. Faber: I’d guess most people will have him higher than this, and probably higher than all the other pitchers on my ballot. It’s not like there’s scads of difference between any of them really. Something that seems telling about Faber is that he led the league in WS twice but finished among the top five only one other time. He combines a really nice peak with plenty of above-average value, value that’s important. But ultimately, his peak isn’t high enough to get him near Mendez, and his career value not quite as good as Rixey’s, and his prime not quite as good as Willis’s.

15. Ryan: Man, I hate to exhume Old Backlog Ryan, but his totals, adjusted to 162 games, are undeniably better than his nearest competitor (in my consideration set), Max Carey:

best 3any5any10any 15career
Ryan98151259342368
Carey88141258342370

Now you could suggest that I’m giving Leach a benefit of the doubt that I’m not giving Carey (as Leach’s, Carey’s, and Ryan’s totals are all very close). My reasoning is this for Leach: a) He was more often among the best players in the league (6 times by WS, to 3 each for Ryan and Carey) b) he played half his career at 3B which was then a tougher position c) slight timeline advantage over Ryan. Carey only benefits from the timeline advantage, and given that Ryan’s peak/prime advantage over Carey is fairly substantial (3 WS/yr in best 3, 2/yr in best 5), the best Carey can do is draw even. Given that Ryan actually did lead his league in WS once to Carey’s zero times, it’s enough to tip it back to Ryan. Carey, like Moore, Doyle, and others are sitting just on the other side of the razor’s edge.

Falling Down….
I’ve rethought two pitchers in particular. I’m no longer convinced that Dick Redding was quite the pitcher I’d originally thought, with fewer big seasons than I’d like to see. And, here comes the controversial one, Clark Griffith. I don’t think that Griffith was, compared to his league, as dominant as Cooper, nor as often an All–Star-level pitcher as Rixey was. He was not as dominant as Willis, and in one of his best seasons, 1901, there are questions about the caliber of his league.

My perception of Griffith was largely based on his peak, featuring 3 30+ WS seasons. But that was a much more normative total for a SP of his time. Wilbur Cooper, to toss out one name, threw a 30 WS season once in his career, yet was among the league leaders twice as often as Clark was. This gave me pause to consider how I had been ranking pitchers relative to one another, and I think it’s led me astray on Griffith.

To validate what was really an obvious conclusion in hindsight, I inputted all the league-leading, unadjusted pitching totals among the various major leagues (no UA) from 1876 through today. I used three-year rolling averages until about 1894 due to the rapid evolution of rules, then five-year rolling averages from then on. I did the same thing for the fifth-best WS total in the league.

In the 1890s, these are the rolling averages of league-leading pitchers:
189018911892189318941895
565453504848

18961897189818991900
4745464239

And of the fifth-best pitcher:
189018911892189318941895
363533323332

18961897189818991900
3332323128

OK, now the 1920s

Leaders:
192019211922192319241925
363432323331

19261927192819291930
3030292829

Fifth-placers:
192019211922192319241925
232323242323

19261927192819291930
2222222222

I think that these contextual differences are causing me an awful lot of grief as I sift through pitchers, and it’s causing me to look at Griffith in a better light than I would otherwise—and conversely looking at the Coopers, Rixeys, and Fabers (and Grimeses!) in a worse light than I normally would. Their career totals look much more impressive than Griffith’s now despite being in the same general area.

Therefore, I’m very concerned that I’m over-crediting Griffith as compared to those that followed him, and I’ve adjusted myself accordingly. Could be an over-adjustment, I’m not sure, but this is where I’m at today.

Ugh, I hate when tabs from MS Word don't come through...repositing the numbers from above with spaces and apologies!

In the 1890s, these are the rolling averages of league-leading pitchers:
1890 1891 1892 1893 1894 1895
56 54 53 50 48 48

1896 1897 1898 1899 1900
47 45 46 42 39

And of the fifth-best pitcher:
1890 1891 1892 1893 1894 1895
36 35 33 32 33 32

1896 1897 1898 1899 1900
33 32 32 31 28

OK, now the 1920s

Leaders:
1920 1921 1922 1923 1924 1925
36 34 32 32 33 31

1926 1927 1928 1929 1930
30 30 29 28 29

Fifth-placers:
1920 1921 1922 1923 1924 1925
23 23 23 24 23 23

1926 1927 1928 1929 1930
22 22 22 22 22

I've said it before and I guess I just have to say it again. How is it wrong to eval. guys on ERA and/or ERA+ before 1913? Do we throw out WS before 1999 (or whatever) or WARP before (whenever)? I mean what the hell is FRAR and PRAR? You think Mickey Welch was pitching to that? You're saying we should basically throw out all of our tools. All of 'em. I think. Help me on this!?

MarcSunnyday2,
Last year's ballot thread clearly showed that Welch pitched numerous complete games when he allowed 15 or more runs, sometimes more than 20. I believe that even accounts for the biggest discrepancy in his comparison to Keefe.

Let's think about why, and how, that may have happened.
First, someone a lot more knowledgeable than me noted that Welch's teams didn't use an 'alternate' P for mopping up. Welch did it himself, and clearly he sometimes did it even when the score was 15-1 in the 3rd inning or whatever.

Now, feel free to count many of those runs as an indicator of his having sucked that day. He caused the team to lose the game, after all.
But at what point does the law of diminishing returns kick in?
Who cares if he allows an 11th, 18th, or 23rd run that day? I doubt that he or his teammates did; he's just finishing out the game.
Yet ALL of these runs - there must be more than 100 of 'em, no? - count just as much toward Welch's ERA+ as that one run he allows in a 2-1 win.
Meanwhile, say that a 'newcomer' pitcher like Rixey doesn't have it one day. He allows nine runs in five innings, and the team trails, 9-1.
Do you think he finishes THAT game? How about if it's 15-1? No, at some point he leaves (though probably not in the first inning after he allows four runs; that will happen decades later).

Now, you could say, 'But Welch could have helped his ERA+ if he had thrown three (however meaningless) scoreless innings late in one of those games, for instance?'
True, but I don't think that not doing so tells you much about his ability to help his team win games.

I don't think it's unreasonable to discount incremental runs allowed in games where the team already trails by 10 runs, for instance, or any runs beyond 15 or so.
We get the point: The guy sucked, and his ERA+ takes a deserved beating. But we don't learn much about his ability by acting as if the incremental runs tell us just as much as the first few.
And consider that we're not taking this info to put him ahead of Rusie or Clarkson or anyone, without checking to see if THEY had similar results (wild guess is no, anyway). We're comparing him now to pitchers who surely have only a tiny fraction of 'meaningless' runs compared to Welch.

As for the other tools, I don't see what they have to do with the argument. This is just about ERA+.
Now, if some hitter turned out to have a much lower batting average in the tail end of a 15-to-20 run wins and losses, with a substantial number of such at-bats, then yes, that definitely would interest me.
I'd suspect that he, too, might have just been looking to make sure his team caught the train out of town on time. And even if that wasn't Welch's intent, I just don't care about that 25th run when I evaluate his HOM ability.

From Dr. Chaleeko's cts on Clark Griffith:

Wilbur Cooper, to toss out one name, threw a 30 WS season once in his career, yet was among the league leaders twice as often as Clark was. This gave me pause to consider how I had been ranking pitchers relative to one another, and I think it’s led me astray on Griffith.

Griffith will do worse than pitchers from later eras on league-leader measures because most of his best years came in a twelve-team league. If you compensate him for the larger pool against which he is competing, he rises back up to parity (in ERA+ and IP, at least. I haven't examined WS this way, but I expect the effect would be consistent).

Should we really count it against Rixey (or give credit to Welch because) that his manager or whomever would decide to take him out down 9-1 in the third whereas, thirty to forty years later this wasn't the norm? In other words, giving Welch credit for finishing these games is the same as giving him credit for pitching in the 1880's as opposed to the 10's or 20's. It is not Rixey's fault he was born later than Welch and vice versa.

Chris Cobb wrote:

The electorate has mostly let peak vs. career voting habits pass unexamined for quite a long time now, and I think we need to start having those debates again, as voters seem to be gradually letting their personal predelictions outweigh their commitment to identifying value in weighing peak vs. career and considering the shape of peak, etc.

Ok, here’s my

Defense of Peak Voting

The career voter focuses on career totals. For those who use sabermetric summary measures like WS or WARP, the idea is to measure the contribution of the player to winning games over the course of his career. If one player contributes 9 wins one year and 1 win the second year, it is valued the same as the player who contributes 5 wins both years. Peak voters, on the other hand, argue that the player with the bigger season is – at least potentially – more valuable because he is more likely to have helped his team win a pennant. The basic idea of the peak voters is that being good enough to win a pennant or a championship is more important than merely being good enough to help win games.

1. Pennants added

First, I’ll admit that I haven’t read the original article by Michael Wolverton that explains this measure, but from perusing the HOM thread on the subject and googling a couple of discussions on other sites, I think I understand the basic idea, though not the details.

Pennants added is closely related to an idea discussed by Bill James in The Politics of Glory in a chapter entitled “Drysdale and Pappas.” James ran a series of computer simulations and found that adding a high-peak pitcher like Drysdale to a team would result in many more pennants than adding a consistent, low-peak pitcher like Pappas, even if their career statistics were identical. Apparently the reason is that in his peak years, the high-peak player is much more likely to push the team over the top. For example, in a peak year a player may provide 9 additional wins, which would be enough to push the team to victory if they otherwise would have been up to 8 games behind.

My understanding is that the Wolverton measure is based on a similar thought experiment – randomly take a player and add him to a randomly selected team and see how often the team wins a pennant that it otherwise wouldn’t have. The results are then accumulated over the player’s career to yield a measure of the player’s “pennants added.”

I have two comments on this measure. First, if you take a look at the rankings that Joe Dimino provided, it’s apparent that pennants added is not what most peak-value voters would consider to be a measure of peak value. The players are mostly arranged in the same order as their career WARP or WS statistics, with a player occasionally rising or falling a few places based on the effects of his big seasons. For example, in the first table (based on adjusted WARP3) Beckley with .64 pennants added is ranked above Jennings with .61. This ranking represents only a slight adjustment from their adjusted WARP3 scores (now several vintages out of date), which were 79.7 and 65.5.

My other comment is more substantive. By randomly assigning a player to a team, and then valuing his contribution relative to a replacement player, pennants added overlooks a critical question in valuing a player relative to winning a pennant – is the player even good enough to play for a pennant-contending team? To take an example from 1939, suppose the St. Louis Browns had offered their center fielder, Chet Laabs, in trade to the New York Yankees. Don’t you think the Yankees’ response would have been, “No thanks, we’re already well taken care of in the outfield.”

2. Contributing to a pennant contender

To contribute to a pennant contender, a player first must be good enough to play on the team. Pennant contending teams are already good – most of their players are above average – so their “pennant replacement level” (that is, the level of play that matches the performance of the player the contending team is already using) will be quite high, usually much higher than the usual replacement level (the level of play that matches the performance of “freely available” talent, such as backup or minor league players). Value to a pennant contender should be based on the wins above this “pennant replacement level.”

Unlike usual replacement level, “pennant replacement level” is not a fixed constant – it will vary from team to team and from position to position. In addition to several above-average or even great players, most pennant winners have at least one below-average regular.

To motivate this idea, consider the following thought experiment. Suppose team managers knew before-hand how each player and team would do that season. Which players would the pennant contenders want to add to their rosters to give them enough victories for the pennant? Let me illustrate this idea using the famous 1908 NL pennant race.

I will first list the WS of the regulars on the three contending teams (Chicago, New York, and Pittsburgh). (These will establish the pennant replacement level at each position.) We will then list the other NL players at the position who had at least 7 WS and see which ones could have helped a contender. (I limit this to the NL because trades between the leagues were uncommon at the time.) We will see that only a few of the players on the other 5 teams played well enough to have helped one of the contenders. (Note - for positions where the top player played fewer than 130 games I’ve added the secondary players needed to bring the total up to at least 130 games.)

C – Bresnahan 27, NY, Kling 22/Moran 7 Chi, Gibson 12 Pit
Dooin 16 Phi, Graham 8 Bos, Schlei 7 Cin

Dooin could have helped Pittsburgh. None of the other catchers would have helped a contender.

1B – Chance 20 Chi, Tenney 19 NY, Storke 5/Kane 4/Swacina 2 Pit
Bransfield 20 Phi, Konetchy 18 StL, Jordan 16 Bro, Ganzel 11 Cin, McGann 11 Bos

Pittsburgh had a major hole at 1B, so Bransfield, Konetchy, Jordan, and possibly Ganzel and McGann would have helped them.

2B – Evers 28/Zimmerman 3 Chi, Abbaticchio 22 Pit, Doyle 17/Herzog 12 NY
Huggins 17 Cin, Ritchey 17 Bos, Knabe 15 Phi

None of the available second basemen would have helped a pennant contender.

3B – Leach 31 Pit, Devlin 24 NY, Steinfeldt 16 Chi
Lobert 32 Cin, Grant 15 Pit, Sweeney 13 Bos, Sheehan 10 Bro

Hans Lobert, having a great year, could have helped all three teams, though he was only slightly better than Leach. None of the other third basemen were better than what the pennant contenders already had.

SS – Wagner 59 Pit, Tinker 32 Chi, Bridwell 24 NY
Dahlen 16 Bos, Doolan 11 Phi, Hulswitt 11 Cin

The pennant replacement level, 24 WS, is very high at shortstop; none of the available players would have helped a contender.

LF – Clarke 28 Pit, Sheckard 15 Chi, McCormick 8/Shannon 6 NY
Magee 26 Phi, Hummel 14 Bro, Delehanty 14 StL, Bates 13 Bos, Paskert 12 Cin, Burch 11 Bro, Kelley 8 Bos

Magee would have helped either Chicago or New York, and Hummel, Delahanty, and possibly Bates or Paskert might have helped New York.

CF – Thomas 19/Shannon 2 Pit, Seymour 19 NY, Hofman 14/Slagle 8 Chi,
Murray 27 StL, Osborn 19 Phi, Beaumont 17 Bos, Kane 14 Cin, Shaw 11 StL

Murray would have helped any of the contending teams. Osborn and Beaumont might have helped Chicago.

RF –Donlin 31 NY, Howard 13/Schulte 11 Chi, Wilson 9 Pit
Titus 23 Phi, Mitchell 10 Cin, Browne 8 Bos, Lumley 8 Bro

Right field has the lowest pennant replacement level (9 WS), but Titus and Mitchell are the only players on other teams who exceed that level.

We see that the pennant replacement levels varied between 9 for right field and 24 for shortstop. For other positions, we have C – 12, 1B – 11, 2B – 22, 3B – 16, LF – 14, CF – 19. It looks like for this particular season the “average” pennant replacement value is about 16 or 17.

3. Implications for measuring pennant value

If the pennant replacement level were a constant, we could measure a player’s contribution to a pennant contending team as his win shares above that level. For example, if the pennant replacement level were 16, then the “pennant win value” of a player with 27 WS (e.g., Murray) would be 11, while the pennant win value of players with 16 or fewer WS would be zero. These pennant win values could then be added over the player’s seasons; the result would essentially be the “total peak” measure that Chris Cobb described in post # 76 of the 1938 discussion thread.

If the pennant replacement level is random, then the pennant win value can be calculated from the player’s WS as follows:

PWV = Sum over distribution of x of: (WS – x) * Prob (WS > x)

where x is the pennant replacement level. For example, suppose the pennant replacement value is uniformly distributed over the integers from 12 to 21. (We’ve seen that the actual range of variation is somewhat larger than this). Then for a player with the following WS, the PWV would be calculated as follows:

WS PWV
12 0
13 0.1
14 0.3
15 0.6
16 1.0
17 1.5
18 2.1
19 2.8
20 3.6
21 4.5
22 5.5
23 6.5
24 7.5
…

The PWV can then be added across seasons to obtain the player’s career peak value score. The relationship between WS and PWV is nonlinear between 13 and 21, but becomes linear for values above 21. The seasons when the player has 20 or more WS receive most of the weight in this system. For the all-time greats like Wagner, Cobb, and Johnson who exceed 20 WS every season, this system is essentially a career value measure. But for peak-type candidates like Jennings and Duffy, this system emphasizes their above-average performances. I use a slightly more complicated version of this system as the basis for my own ratings.

In my opinion, this type of peak value system comes closer than the career value systems to the idea of merit that we seek to honor. In sports it is natural to measure champions, as well as players who perform at championship levels but may not have had the fortune to be on a championship team. What we see is that most of the value associated with winning pennants resides in a few of the best players each season.

Brent, excellent idea and fleshing out of it. My main criticsm would be that in 1908, you had a third-place team that won 98 games, whihc is a high-water mark. Had you picked the 1914 or 1915 NL, where the 3rd place team won 80 or 81, I'm pretty sure the pennant-replacement value for many positions would be lower.

"Should we really count it against Rixey (or give credit to Welch because) that his manager or whomever would decide to take him out down 9-1 in the third whereas, thirty to forty years later this wasn't the norm? In other words, giving Welch credit for finishing these games is the same as giving him credit for pitching in the 1880's as opposed to the 10's or 20's."

NO NO NO NO NO NO

Welch gets NO credit for finishing that game. Hell, I'd even take the innings away from him, since he didn't contribute anything with them. This case is NOT about how many complete games Welch had. It's about the pct of his IP that came after a game was long decided. I'm saying those innings and runs tell you next to nothing about his skills, and pretending they are as legit as other IP leads people to think incorrectly, "Wow, even with good teams, this guy still was lucky to win more than 300 games with a 113 ERA+!"

Except I'm saying he wasn't lucky.

Welch gets NO credit for finishing that game. Hell, I'd even take the innings away from him, since he didn't contribute anything with them.

A-ha! Here's something I haven't seen mentioned yet.

So, its not enough to simply cap Welch's daily run-totals, we should also be lowering his IP, too... assuming that he appropriately got the "hook". The capping of his daily run-totals was effectly crediting him with "shutout innings" after the game was already decided... that's not exactly fair either.

Does his 24 run game turn into a 13-runs-in-5-innings effort?

This method of demonstrating Welch's ability to "pitch to the score" might not help him as much as I originally thought.

Yes it will help him, because Welch's ERA is around 3, not 23.4 (13 runs in 5 innings.) The argument against paying too much attention to Welch's ERA+ (which in any case is close to all these 20s bozos now swarming out of the woodwork) is that if he's made to mop up a game when he doesn't have it, you are inflating his ERA by giving him 9 innings of 23.4-ERA pitching, rather than 5, as would have been done in the 20s.

Welch's best year was 1885, not 1889; only 4 out of 8 HOM'er starters on that team, the others were stooges.

you are inflating his ERA by giving him 9 innings of 23.4-ERA pitching, rather than 5

Yup... I thought it over during breakfast and came to the same conclusion. It helps him... but not as much as I first thought.

The argument against paying too much attention to Welch's ERA+ (which in any case is close to all these 20s bozos now swarming out of the woodwork) is that if he's made to mop up a game when he doesn't have it, you are inflating his ERA by giving him 9 innings of 23.4-ERA pitching, rather than 5, as would have been done in the 20s.

But wasn't Welch operating under conditions that every other pitcher during his time had to face? IOW, if we're making adjustments to Welch's record, don't we have to do the same for all of the other pitchers of his era (and then compare Welch to them)?

As for the twenties pitchers, the odds are a few more of them would have had 300 or more wins if they had played during the 1880s. Certainly, without question, their peak numbers would appear to be much improved with 40 and possibly 50 wins showing up on their resumes.

Here's a comp I didn't see before, and I haven't heard mentioned:

Larry Doyle and Lip Pike

Both good hitters, but Pike has a much better adjusted EqA, .302 to .285 (unless WARP has changed again), so even with a timeline difference, I think Pike was better. Plus his blinding speed has to have helped score some runs in the 1870s.

Doyle played a medicore 2B. Pike played CF, but played 2B and other infield spots in his younger pre-NA years. Don't see much defensive difference there.

Career length? Kinda depends on if you give Pike much credit for playing Western ball. It's either even, or advantage to Lipman.

To me, I find it difficult to justify Doyle over Pike, who is begging to be moved up to my #2 slot this ballot.

But wasn't Welch operating under conditions that every other pitcher during his time had to face? IOW, if we're making adjustments to Welch's record, don't we have to do the same for all of the other pitchers of his era (and then compare Welch to them)?

Exactly, the hypothesis is that this will help Welch's standing relative to his contemporaries. Welch certainly deserves a big "L" for these bad games, but just the one L. Those 20+ RA games maybe have been acting like multiple losses in the ERA+/Pythag analyses.

Playing with game-log data is tricky though. From ERA+/Pythag, I usually like to jump all the way to SNWLR instead of getting cute with all sorts of questionable adjustments. You need full boxes for that, though.

Well, since basically all games are complete games for this era, we may not need the full SNWLR to do this. Has anyone thought of doing some sort of additive probability on these game logs -- Say giving up 0 runs give you a 1-0 record. Giving up league average runs gives you a .5-.5 record. Giving up 15 runs gives you something like .05-.95 and giving up 25 runs give you something like .01-.99 ? This type of analysis would prevent super-negative effect of the super-blowouts. I don't know you would determine the probabilities for the non-trivial cases and the probabilities would be different for each season (and probably need to be park-adjusted as well). Sounds like a fair amount of work.

My main criticsm would be that in 1908, you had a third-place team that won 98 games, whihc is a high-water mark. Had you picked the 1914 or 1915 NL, where the 3rd place team won 80 or 81, I'm pretty sure the pennant-replacement value for many positions would be lower.

On the other hand, Brent's statistic is intended to show the replacement level for a "pennant contender", not a "top three team". A third place team with a losing record is not really a pennant contender.

Some changes in my evaluation this year. I've started giving bonus points for players who were among the best at their position for a given year. as well as a career bonus. This helps resolve the mid-ballot 3b glut I was developing over the last few years. It also helps me to evaluate comparable players that have recently come on the ballot. (Such as Bresnahan-Schang-Schalk and Bancroft-Sewell) I've only done this with hitters so far, but hopefully I'll get to do this for pitchers in the near future.

Preliminary Ballot for 1939
1. Charley Jones
2. Lip Pike
3. Pete Browning
4. Cupid Childs
5. Hugie Jennings
6. Heinie Groh
7. Eppa Rixey
8. Hugh Duffy
9. George Sisler
10. Joe Sewell
11. Tommy Leach
12. Wally Schang
13. Jose Mendez
14. Ed Williamson
15. Clark Griffith

16-20 Redding, Monroe, Willis, Mays, Faber
21-25 Roush, Carey, Tiernan, Bresnahan, Doyle
26-30 Van Haltren, Poles, Moore, McGraw, Bancroft
31-35 F. Jones, Cooper, Waddell, Bond, Taylor
36-40 Griffin, Long, Welch, Chance, R. Thomas
41-45 Burns, Konetchy, Cravath, Fournier, Beckley
46-50 Tinker, Schalk, Evers, Maranville, Veach

re: new WARP.

WTF? The top 3 pitchers on my ballot just saw their career values sink like they were Enron stock. Maybe that's OK since they weren't exactly consensus HOMers, but sometimes it's hard to take this stuff seriously.

Must be a technical glitch in the WARP thing. I lookd this morning and laughed. Dave Kingman has almost as much career value as Lefty Grove...but you already knew that.

Must be a technical glitch in the WARP thing. I lookd this morning and laughed. Dave Kingman has almost as much career value as Lefty Grove...but you already knew that.

Chortle, chortle.

Tom Seaver now has considerably less career value than Joe Tinker.

There's clearly some major technical glitch in the system right now. They've added an extra, unlabeled column in the "Advanced Pitching Statistics section between PRAA and PRAR, and they've put that column into the "Advanced Batting Statistics" section in place of the PRAR column, and are using that column, which is close in value to PRAA, to calculate WARP rather than PRAR. When you make replacement level equal to average, then, yes, the values are going to drop considerably.

Bad timing for them to screw up their system just as we're trying to get the 1920s pitching cohort sorted out.

Is it too late to mount a “Stop Lip Pike” campaign? A decade by decade look at percentage of AB by HOMer’s suggests strongly that we have already maxed out on Pike’s era and need to do some catching up in the deadball era:

1870’s – 16.7%
1880’s – 13.3%
1890’s – 13.2%
1900’s – 9.9%
1910’s – 6.3%

Add Pike and the 70’s jumps to 17.9%. Let Pike be his era’s guy to be just below the in/out line. Was baseball's Golden Age truly the 1870's?

Add Pike and the 70’s jumps to 17.9%. Let Pike be his era’s guy to be just below the in/out line. Was baseball's Golden Age truly the 1870's?

Will the 1990s be baseball's Golden Age in 2245? I hope they don't throw them all out of the HOF because they won't compare with the 23rd Century guys.

Besides, Pike played a good chunk of his career during the 1860s, which is underrepresented.

I say this every ten years or so:

If you measure the value over a player's career and give extra credit for truly outstanding seasons b/c/ those seasons are more valuable from a pennants-added perspective, you are a CAREER voter, a sophisticated, non-linear career voter, but a career voter nonetheless.

A true peak voter votes for the guys with the best 4 or 5 seasons not b/c/ those seasons are so valuable when aggregating the players' careers, but b/c/ they are asking a different question. You can ask the question any number of ways, but essentially you are asking "Take all the players in the history of the game at their highest established level and line them up. Who would I pick to be on my team?" It's not a different way of measuring overall value, it's an alternative definition of what we mean when we ask who the greatest players of All-Time were.

A decade by decade look at percentage of AB by HOMer’s suggests strongly that we have already maxed out on Pike’s era and need to do some catching up in the deadball era.

Yet DanG's ballot shows that the big jump in HOM-ers occurred after Pike left the game. The early-to-mid 70s are not overrepresented at all. Yearly HOM counts are about 9-11 in the Lip Pike era (71-78).

A couple things are going on in the 1870s. First, the NL contracted a bit for 77-78. Also, the 80s stars start showing up right about then as well causing your typical pre-expansion bulge.

Actually, I don't think there's an extra column in the WARP Pitching statistics - there's one column that's tabbed wrong, so it doesn't look like it's under anything, but that should be the "NERA". I think the problem might be that they're weighting the pitcher's Batting Runs more than they did before, but I don't really know how they did it in the past.

John – my post had nothing to do with timelining. If 17.9% of all at bats featured HoMer’s swinging away, how big would the HoM have to be? I am sure it would be a whole lot bigger than what we have in mind. That being the case, why should the 1870’s merit that kind of representation?

re: WARP

Man, I'm glad I went through it on Monday night before this disaster happened. Can we assume an e-mail has already been sent to BP?

After further review, I don't think it's the Batting Runs. They did SOMETHING screwy with the pitchers, though. I don't care if an e-mail has been sent, though, I'm sending another. (And the numbers don't match what Paul Wendt has in the file on the Yahoo site.)

I heard back from Clay Davenport (very nice job on the response time!) and there's an error in there right now. They take about 8 hours to generate, so the correct ones won't be back up until tomorrow morning.

Another response to Dan b:

A couple things are going on in the 1870s. First, the NL contracted a bit for 77-78. Also, the 80s stars start showing up right about then as well causing your typical pre-expansion bulge.

Two additional factors are at play here. 1) In the NA years, schedules were wildly uneven, and the best teams (esp. Boston) played more games, so their percentage of at bats is raised. 2) Boston's offense was _so_ much better than anybody else's that they were averaging considerably more ab/game than anybody else.

Player-seasons is a better measure than % of at-bats.

Percentage of AB by HoMers by year:

1871 – 11.6%, with Pike – 12.8%
1872 – 12.9%, with Pike – 14.7%
1873 – 17.3%, with Pike – 19.0%
1874 – 17.3%, with Pike – 18.5%
1875 – 14.3%, with Pike – 15.5%
1876 – 15.4%, with Pike – 16.7%
1877 – 19.0%, with Pike – 21.0%
1878 – 19.5%, with Pike – 20.7%
1879 – 21.2%, with Pike – 21.2%

Highest year of the deadball era:
1905 – 9.6%

The 1870’s are over-represented already. Vote NO on Proposition Lip Pike.

Dan's statistic reflects a point-of-view that says that the number of HOMers should be proportional to the number of teams. There were half as many teams in the 1870's, relative to the 1880's, therefore there should be only half as many HOMers. The 1890's reach the same percentage level with 3/4'ths of the HOMers.

There are some obvious problems with it; pitchers impact the stat less and less as time goes by, Negro Leaguers not at all. Maybe others.

IIRC, in the formative stages of this project team seasons were the basis for how many players would be elected each year and that all eras would be equitably represented.

Actually, I don't think there's an extra column in the WARP Pitching statistics

Sorry, Devin, there is an extra column. If you look at the team sheets, it's between RA and DHT (the old Delta Hits) and labeled DHL (Delta Hits League?, I assume a measurement of how well the pitcher/defense combo did on hit suppression).

It looks like the data transfer from the pitcher tableau (Advanced Pitching Statistics) to the summary tableau (Advanced Batting Statistics) wasn't changed to accomodate the new column, so PRAA got transfered instead, labeled PRAR, and then used to calculate the various WARPs, hurting all pitchers of every era. So Chris is right on there. It's a programming glitch, not a baseball logic flaw.

IIRC, in the formative stages of this project team seasons were the basis for how many players would be elected each year and that all eras would be equitably represented.

I think dan b is correct. While I also think (though I haven't checked my numbers lately), that player-seasons by HoMers in relation to team-seasons in the 1870s indicate that the period is not under-represented. Neither, however, is it obviously over-represented.

The AB argument against Lip Pike is arguably even stronger for Mickey Welch, as the % of National League ABs by HoMers is over 20% every year from 1881 to 1891 except 1888 (18.2) and 1890 (14.9). (Of course, if you include the AA, it's between 10.5 and 13.9 between 1883-1891.)

I'll post the updated (through 1938) version of the AB database on the Yahoo site later tonight.

How many Negro Leaguers do we have from the 00s and 10s now? How would include them into the percentages that dan_b is presenting?

When determining the number of eligibles, how many teams were "added" to MLB to account for the Negro Leagues?

When determining the number of eligibles, how many teams were "added" to MLB to account for the Negro Leagues?

IIRC, the same number of teams pre-NA, none.

Two related-unrelated topics.

1. Re. andrew's definition of a peak voter: By this definition, I doubt if we have any peak voters. Maybe we're really prime voters. And/or andrew is being a little unfair to peak voters there.

2. As to the linkage of the numbers of players we elect PER year/decade/era/whatever with the number of players we elect FROM a given year/decade/era/whatever: These are two different things.

a. There is the obvious fact that while there is a measurable change in the total opportunity (games per season X teams per season = total value opportunity) over time. BUT the distribution of all those opportunity values to individual players does not exactly mirror the grosser totals due to chance and other factors in the structure of the game. So to some extent, we would expect career and even peak values to deviate from the model. The odds that the two would mirror one another are very small.

b. The bigger yet more subtle problem is this. The player pool is of a different size than the value opportunity pool, though the player pool also has grown over the years. But the total amount of skill in the player pool is what it is, regardless of whether you give them 16 or 32 teams, or 16 x 154 or 32 x 162 opportunities to use those skills. More opportunities is not the same as more skill.

But in any event I think you can argue that the size of the talent pool should be more important than the size of the opportunity for value. The notion that the two have tracked one another all these years...well, it would again just be too much of a coincidence. There may very well have been "more skill" per opportunity at certain times than at other times, and those times should be overrepresented as compared to the numbers of players we elect IN a given year.

Did that make any sense?

When determining the number of eligibles, how many teams were "added" to MLB to account for the Negro Leagues?

IIRC, the same number of teams pre-NA, none.

I just checked and I think there is a small increase in the eligibles due to the Negro Leagues. Its the reason why there's an elect-3 year in the 50s and then a dip back to a final elect-1 year in the 60s... because there was effectively a brief period of "contraction" after integration and before expansion.

Anyhow, the original point was that Dead Ball Era is also represented by Foster, Lloyd, Williams, Santop, Hill, Johnson & Torriente. But, I don't have a mechanism for translating their contribution in some sort of AB or IP percentage.

Footnote: Its not going to effect the 1870s data much, but we should switch from an AB % to a PA percentage at some point.

1. Re. andrew's definition of a peak voter: By this definition, I doubt if we have any peak voters. Maybe we're really prime voters. And/or andrew is being a little unfair to peak voters there.

Nothing unfair here. Just semantics. I vote the way I vote, and it doesn't really matter to me what that is called.

There is a indeed career element involved in how I vote. Although McGraw was a better player at his best, I rated Groh higher than him because of issues with the McGraw's career length.

OK, the latest version of the At-Bats database is up on Yahoo. As far as the AB/PA question, I just used ABs because it's what Bill James used in The Politics of Glory, so there was more of a frame of reference there. I honestly doubt it will make much of a difference for our purposes.

I don't really feel there's an underrepresented era at this point. As noted, there's more Negro Leaguers in that era. I think there's an argument that the NL in the 1900s is underrepresented - was it really significantly worse than the AL then? (In the 1910s, I don't disagree.) And it's a little too soon to judge the 1910s representation, as there are still "no-brainer" candidates who played then (Hornsby and Ruth).

Does anyone think it would be worth the effort to try and estimate ABs for Negro Leaguers?

On my "defense of peak voting" TomH wrote:

My main criticsm would be that in 1908, you had a third-place team that won 98 games, whihc is a high-water mark. Had you picked the 1914 or 1915 NL, where the 3rd place team won 80 or 81, I'm pretty sure the pennant-replacement value for many positions would be lower.

It's certainly true that the pennant-replacement value bumps around from year to year, which is why I use a distribution taken from a sample of seasons, rather than actually using a single year.

Regarding the specific question of whether the pennant-replacement values were unusually high for 1908, the answer (perhaps surprisingly) is no. The reason is the Pirates, who probably had more deficiencies in their lineup than any other team that has won 98 games. They were average at catcher and well below average at 1B and RF. (Their pitching was also somewhat weak.) You figure they had to have had some serious problems to waste Wagner's greatest season--arguably the greatest season ever--as well as superb performances from Leach, Clarke, and Abbaticchio.

The average pennant-replacement value actually turns out to be somewhat higher in 1914 than in 1908, though in 1915 it's lower. Though I've only calculated it for about a dozen seasons, the average is usually around 16 or 17.

PhillyBooster wrote:

On the other hand, Brent's statistic is intended to show the replacement level for a "pennant contender", not a "top three team". A third place team with a losing record is not really a pennant contender.

Exactly. As a rule of thumb, I take a "pennant contender" to be the top two teams plus any other teams that finished within 10 games of the winner.

andrew siegel wrote:

If you measure the value over a player's career and give extra credit for truly outstanding seasons b/c/ those seasons are more valuable from a pennants-added perspective, you are a CAREER voter, a sophisticated, non-linear career voter, but a career voter nonetheless.

You're right. My system is not purely a peak voting system. (I also agree with sunnyday2 that there are probably not any pure peak voters here.) But it does capture some of the special value of high-peak, short-career candidates like Jennings. I think of it as a "career peak" measure. But there are lots of ways to evaluate a player's career - I just wanted to present the method that I find most convincing, though I'm certainly interested in looking at the question from other perspectives as well.

Well, I guess a "pure peak" voter would have to start with John Paciorek and work his way down from there.

I think "peak" vs. "career" is silly. The question is who provided the most value. Some players, like McGraw & Jennings, can provide tremendous value in a short career. Others, like Beckley, provided value over a longer career. Both should be able to co-exist on the same ballot.

McGraw
_XXX
XXXXX
XXXXX
XXXXX
XXXXX
XXXXX
XXXXX
XXXXX

Beckley
_______________XXX
X________XX_XXXXXXXXX___XXXXXXX_XXXX
XXXXXXXXXXX_XXXXXXXXXXX_XXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

Different shapes, but who cares? If McGraw has more "x"'s (WAA, WARP, WIN Srares, whatever the measure) compared to his contemporaries, he should rank higher....

Well, I guess a "pure peak" voter would have to start with John Paciorek and work his way down from there.

:-)

The most generous estimate of John Paciorek's contribution to his team is one win. That's why I'm a bit cautious about using only rate stats. A single season is the smallest increment I look at. This is a rehash of the WSAR discussion a couple of "years" ago, but WS/162 does not tell the whole story if the player has in-season durability issues.

Anyhow, the labels are just semantics. We use labels like "peak voter" and "career voter" as abbreviations for more lengthy descriptions. People in general know what we mean when we say those things.

Team by team breakdowns for Rixey and Luque in 1922 and 1923 (all information courtesy of Retrosheet and Baseball Reference):

game #, opposing pitcher, result, score
1922
Luque 13-23, Rixey 25-13, team 86-68 2nd
Luque v. Boston 53-100 last
026 Fillingim L 3-9
084 Miller W 3-2
100 Miller W 6-3
124 win in relief
147 Watson L 2-5

Rixey v. Boston
087 Oeschger W 9-3
098 Watson W 5-4
146 Cooney W 4-2

Luque v. Brooklyn 76-78 6th
029 Ruether L 5-6
051 Vance W 6-2
080 Ruether W 6-4
104 Vance L 0-5
116 Ruether W 9-4

Rixey v. Brooklyn
030 Cadore W 6-2
052 Ruether W 7-2
083 Cadore W 3-0
102 Grimes L 0-4
118 Cadore W 4-3
121 Hulihan W 5-3
137 Vance L 2-3

Luque v. Chicago 80-74
002 Aldridge L 1-5
009 Osborne L 1-3
042 Alexander/Cheeves (RP) L 7-8
047 Alexander W 6-1
126 Cheeves L 4-7
131 Osborne L 1-3

Rixey v. Chicago
001 Alexander L 3-7
008 Cheeves L 3-8
038 Cheeves L 2-5
049 Aldridge W 4-2
067 Stueland W 9-5

Luque v. New York 93-61 1st
037 Douglas L 1-2
055 Ryan L 1-2
088 Nehf L 2-5
092 Ryan L 1-4
120 J Barnes L 1-2
141 McQuinlan W 8-4

Rixey v. New York
035 Nehf W 7-2
056 J Barnes L 3-9
091 Douglas W 3-2
107 Nehf W 7-3
142 Nehf W 4-3

Luque v. Philadelphia 57-96 7th
033 Meadows W 2-1
059 Meadows/Ring L 2-7
075 Meadows L 1-3
096 Meadows L 5-6
152 win out of bullpen

Rixey v. Philadelphia
060 Singleton W 8-4
078 Hubbell W 9-3
094 Weinert/Ring L 7-12
114 Smith W 3-1
151 Behan W 10-3

Luque v. Pittsburgh 85-69 Tied 3rd
005 Morrison L 0-1
013 Morrison W 8-5
017 Morrison/Adams L 6-7
063 Glazner W 7-3
110 Cooper L 1-7
153 Cooper ND

Rixey v. Pittsburgh
004 Cooper L 3-4
015 Cooper L 3-7
044 Cooper W 9-3
062 Morrison W 6-2
111 Adams L 0-6
125 Glazner L 0-2
154 Glazner W 5-1

Luque v. St. Louis 85-69 Tied 3rd
021 Pfeffer L 2-3
070 Pertica L 1-2

Rixey v. St.Louis
012 Walker W 3-0
019 Doak L 5-6
023 Doak L 5-8
072 Doak W 11-9
128 North L 3-9
133 North W 10-6

1923
Luque 27-8, Rixey 20-15, Team 91-63 2nd
Luque v. Boston 54-100 7th
022 Marquard L 4-5
043 Miller W 7-1
079 McNamara W 4-3
080 Oeschger W 9-5 (both ends of doubleheader)
093 Benton W 2-1
121 Oeschger W 4-1
138 Oeschger W 9-1

Rixey v. Boston
021 McNamara W
045 Oeschger W
105 Genewich L
119 Genewich W
141 Cooney W

Luque v. Brooklyn 76-78 6th
054 Grimes W 1-0
084 Grimes W 10-6
089 Vance L 3-6
117 Grimes W 4-0
148 Vance W 5-1

Rixey v. Brooklyn
030 Cadore L
053 Dickerman No Decision
083 Ruether W
091 Grimes ND
116 Smith W
146 Decatur W

Luque v. Chicago 83-71 4th
015 Osborne L 1-2
038 Cheeves W 3-2
058 Alexander L 0-2
067 Keen W 6-3
124 Alexander W 4-3

Rixey v. Chicago
013 Cheeves W
032 Osborne/Cheeves L
042 Kaufman L
060 Aldridge ND
123 Aldridge L
136 Alexander W

Luque v. New York 95-58 1st
026 Bentley W 7-0
050 McQuillan W 3-0
070 Scott W 5-3
100 Ryan L 4-14
104 Watson L 2-6
109 McQuillan W 6-3
150 Watson L 2-3

Rixey v. New York
027 McQuillan L
049 Bentley W
069 McQuillan W
101 McQuillan L
109 Ryan W
112 McQuillan W

Luque v. Philadelphia 50-104 last
047 Glazner W 2-1
075 Weinert W 2-0
113 Couch W 6-3
144 Glazner L 0-2

Rixey v. Philadelphia
024 Head W
073 Glazner W
086 Ring L
142 Betts W

Luque v. Pittsburgh 87-67 3rd
008 Kunz W 5-4
019 ND
061 Morrison W 2-0
128 Meadows L 2-6
133 Cooper W 8-3

Rixey v. Pittsburgh
006 Morrison L
017 Glazner L
057 Cooper W
063 Adams L
127 Cooper W
153 Meadows L

Luque v. St. Louis 79-74 5th
004 Pertica W 10-2
011 Toney L 3-4
033 Toney W 2-1
151 Toney W 11-1

Rixey v. St. Louis
002 Toney L
037 Haines W
087 Pfeffer L
131 Doak/Pfeffer L

I hope these are helpful. Sorry no scores for Rixey. Too tired. g'nite.

HOM Pitchers by year
minimum 10 G pitched in that year

1868-76 (1) - Spalding
1877 (0)
1878 (1) - Ward
1879 (2) - Ward Galvin
1880 (3) - Ward Galvin Keefe
1881-83 (4) - Ward Galvin Keefe Radbourn
1884-88 (5) - Galvin Keefe Radbourn Clarkson Caruthers
1889 (6) - Galvin Keefe Radbourn Clarkson Caruthers Rusie
1890-91 (8) - Galvin Keefe Radbourn Clarkson Caruthers Rusie Young Nichols
1892 (7) - Galvin Keefe Clarkson Caruthers Rusie Young Nichols
1893 (5) - Keefe Clarkson Rusie Young Nichols
1894 (4) - Clarkson Rusie Young Nichols
1895 (3) - Rusie Young Nichols
1896 (2) - Young Nichols
1897-98 (3) - Rusie Young Nichols
1899-00 (3) - Young Nichols McGinnity
1901 (5) - Young Nichols McGinnity Plank Mathewson
1902 (5) - Young McGinnity Plank Mathewson Foster
1903 (6) - Young McGinnity Plank Mathewson Foster Brown
1904 (7) - Young Nichols McGinnity Plank Mathewson Foster Brown
1905 (8) - Young Nichols McGinnity Plank Mathewson Foster Brown Walsh
1906 (7) - Young McGinnity Plank Mathewson Foster Brown Walsh
1907-08 (8) - Young McGinnity Plank Mathewson Foster Brown Walsh Johnson
1909 (7) - Young Plank Mathewson Foster Brown Walsh Johnson
1910 (8) - Young Plank Mathewson Foster Brown Walsh Johnson Williams
1911 (9) - Young Plank Mathewson Foster Brown Walsh Johnson Williams Alexander
1912-13 (8) - Plank Mathewson Foster Brown Walsh Johnson Williams Alexander
1914-15 (7) - Plank Mathewson Foster Brown Johnson Williams Alexander
1916 (8) - Plank Mathewson Foster Brown Johnson Williams Alexander Covaleski
1917 (5) - Plank Johnson Williams Alexander Covaleski
1918 (3) - Johnson Williams Covaleski
1919-26 (4) - Johnson Williams Alexander Covaleski
1927 (3) - Johnson Williams Alexander
1928 (3) - Williams Alexander Covaleski
1929 (2) - Williams Alexander
1930-32 (1) - Williams

(Wallace actually qualifies in 1895-96 and there may be another stray season somewhere from someone, but I drew the line at listing Ward).

Observations: After Spalding gets things going, Ward and Galvin kick off a "golden 1880s age" with Keefe, Radbourn, Clarkson, and Caruthers.

Rusie, Young, and Nichols then take over as the 1890s representatives, as the mound change and old age turn those other guys into dinosaurs.

McGinnity, Plank, Mathewson, Foster, and Brown are your 1900s guys, along with ageless Young. Most steam forward well into the 1910s, while Walsh shows his skills only in part of each decade.

Johnson, Williams, and Alexander are the next wave so far, and they'll remain the class of the 1920s even after we start catching up on that decade.

Covaleski is our last arrival so far, in 1916.

Welch would be 1880-91
Griffith would be 1891 and 1894-1906
Waddell would be 1899-1910
Rixey would be 1912-17 and 1919-33
Faber would be 1914-33

This is an updated version of something I did awhile back. For each season, I averaged the OPS+ of the starting catchers in each league, found how much each player was above or below the average, and then added that up for their career. If you think it's useful, great.

I have everything done up through 1945. For Bresnahan and Schalk, there's a 2nd number which includes their seasons as starting OFs, compared to that position (Bresnahan, 1903 and 1904 CF, Schalk, 1916 LF). The numbers in parentheses are the seasons above and below average, followed by the per-season difference.

All Cs with totals above 200:
Bill Dickey 496 (14,0) 35.4
Gabby Hartnett 465 (14,0) 33.2
Mickey Cochrane 422 (11,0) 38.4
Ernie Lombardi 402 (13,0) 30.9
Deacon White 385 (8,0) 48.1
Wally Schang (all years) 351 (11,1) 29.3
Wally Schang (C only) 334 (10,1) 30.4
Buck Ewing 319 (7,1) 39.9
Jack Clements 313 (9,3) 26.1
Roger Bresnahan (all years) 306 (8,0) 38.3
Charlie Bennett 305 (10,3) 23.5
Deacon McGuire 246 (9,5) 17.6
Roger Bresnahan (C only) 235 (6,0) 39.2
Mike Grady 218 (4,0) 54.5
Chief Meyers 215 (7,0) 30.7
Johnny Kling 209 (8,3) 19
Cal McVey 208 (4,0) 52

Some others of note:
Rick Ferrell 61 (10,4)
King Kelly 50 (2,1)
Ray Schalk -27 (4,9) - I don't think he could have been that good a fielder to overcome the difference.
Connie Mack -140 (0,6)
Al Lopez -185 (3,11)

And the bottom 5:
Pop Snyder -214 (6,7)
Mike O'Neil -216 (1,5)
Doc Bushong -220 (1,6)
Luke Sewell -274 (0,12)
Bill Bergen -484 (0,8)

BTW, that's a -60.5 average for Mr. Bergen. Bresnahan did benefit a little bit from his numbers, but only 2 seasons, and one of those was Bill's best, with an OPS+ of 31. Roger didn't get a boost from the 6, the 1, and the -4.

I was reading about some of our new pitchers today in that Neyer/James book on Pitchers.
Faber said chewing gum and other things never worked for him, so he used only chewing tobacco saliva even though he didn't like it. He said he never chewed tobacco except in games, calling it a 'tool of the trade' like a carpenter's collection of stuff.

Grimes, coming soon, used slippery elm as his 'foreign substance'; he used to peel the bark right off the tree, and use the gooey stuff underneath.

Rixey gave an interview in 1927, shaking his head and saying that for 15 years, every time he went to 2-0 or 3-1, the hitters always looked for a fastball. "And they NEVER got it!" he laughed.

Fun note on Joe Sewell, beyond the usual 'didn't strike out much' declarations:

May 26, 1930: The Indians take a pair from the White Sox, winning 7–3 and 3–2. In the nitecap, White Sox lefty Pat Caraway twice strikes out Joe Sewell, an occurrence that Joe later blames on the white shirts in the CF bleachers. It is the last time the Indians 3B will fan this season, and he will end the year by striking out only three times in 353 at bats. Once before, on May 13, 1923, Sewell was fanned twice in a game.

Red Faber:
Really we're looking at 1915 and 1920-22 to make his peak case. Those are the only years he ever was in the top 10 in IP or Ks, or in top 7 in Ws.
He did lead the league in ERA in 1921-22, and he's got that sizzling 171 ERA+ in 1921, when he went 25-15 and the rest of the Ps on his team went 37-77.
Also pretty good in 1917, 1923, and 1925, too, and he's better than the W-Ls look late in his career because those were just awful White Sox teams in spite of the Faber-Lyons combo.
Overall, not quite as good as I'd thought, but obviously on the ballot and arguably better than most in a weak year.

Eppa Rixey:
1916 and 1921-25 look the nicest, but a little more supplement as well than Faber had.
Had impressive 10 top-9 finishes in IP, six top-6s in ERA, and seven top 10s in Ws. Also played on some crappy teams.
Weird that we have so many of these types coming up, after lots of Mathewson-Smokey Joe types previously. I can see Covaleski packed a lot more wallop in a shorter career; tough to judge what to make of these guys' extra modest seasons.

Joe Sewell:
8 SS seasons, 5 3B seasons.
Slugged .413 over his career, exactly the same as his league did.
Not quite a 100 OPS+ for his career after age 30.
Only one OPS+ over 117, a 146 in 1923. Earned his 4th in MVP voting that year; oddly got 3rd in 1925, a much weaker offensive season. Even then I guess voters would ignore a large drop in walk rate.
Rated similar to Cronin in his late 20s, but to Billy Herman, another early decliner, in his early 30s.
Somebody sell me on this guy's fielding; I see a short total career and an even shorter amount of time as any kind of offensive force at all. Good player, but is he better than a dozen other 2B-SS-3Bs in our mix?

Rabbit Maranville:
Only in 1917 and 1919 (110 and 112) did the Rabbit ever hit 100 on the OPS+ meter. Makes Sewell look like Rogers Hornsby. Still, the 1915-24 era is mostly in the 90s in OPS+ with a slick glove, so he's very much a quality player there overall. Yet never hit 80 in his last decade of play, which is in effect stats-padding, as his defense also presumably weakens.

Jack Quinn:
Debuted late at age 25 in 1909 with Willie Keeler and Jack Chesbro; said goodbye a week shy of his 50th birthday in 1933 with Ernie Lombardi and Leo Durocher. Played with 31 Hall of Famers, and greeted Babe Ruth's arrival to his Yanks in 1920 with one of his finest pitching seasons. Collected 10.5 pct of his career wins in the Federal League in 1914; went 9-22 in the other FL season. I love his 1928 sweet season at age 44, but there's little else to get excited about even if it is a fascinating career. Was league's 2nd-oldest player by 1921, and played 12 more years.

A slight boost to most of these guys for 1918 credit; that was more of a decimated MLB year than I realized before this project started.

BP appears to be back to normal now. Quick spot check of a few WARPs have it where it was earlier this week.

A slight boost to most of these guys for 1918 credit; that was more of a decimated MLB year than I realized before this project started.

What hitters currently on the ballot should I be looking at for potential WWI boosts? Right now, Maranville (and a Negro Leaguer or two) are the only ones getting any from me, but I suspect that's probably shortchanging some guys. FWIW, I'm looking for hitters here, for reasons best stated elsewhere, I don't give war credit to pitchers (I'm sure a situation could arise where I'd change my mind on that, but guys like Rixey and Faber, nope).

posted 1924 and 1925 opposing pitchers information for rixey and luque on their thread.

I then started wondering if there was a way to get a corner outfielder fielding scale.

What I did from the outset (for purposes of this project) was separate actual HoF outfielders into their primary outfield positions. I then calculated the mean and standard deviation for each outfield position's WS per 1,000 IP (for HoFers) as published in the Win Shares book.

I used the data to develop HoF ranges for defense, and I compare the candidates to those ranges to produce a defensive rating for HoM.

So, for instance, the average HoF rightfielder has 2.54 WS/1000 innings with a standard deviation of .38. My range for a "B" outfielder is half a standard deviation on either side of the mean (a generous system). So my range for a "B" is 2.35-2.73. Anything north of 2.73 is an "A"...and so on.

My rating scale measures HoFers only, because that's what I personally compare the candidates to. I want the HoMers to compare favorably to the best players of all time, not the general baseball population. Of course, a player could conceivably be a "C" on my rating scale (vis-a-vis HoFers) but be a "B" fielder relative to the rest of the baseball world, although that is less likely since my "B" range is a full standard deviation and I see no reason to believe HoFers are better overall defensively.

Anyway, this was an easier system than doing the same thing for every RF who ever played. :) Because the defensive abilities of HoF outfielders is not uniformly high, I bet the numbers from the HoFers aren't significantly different than for baseball as a whole, but that's just a hunch.

(BTW, I use a similar system for Total WS, 3-year peak WS, 5-year consec peak WS, 7-year peak WS, WS/162g Total WARP1, 3-year peak WARP1, 5-year consec peak WARP1, 7-year peak WARP1, WARP1/162g and, for pitchers only, Wins Above Team and Linear Weights).

John – my post had nothing to do with timelining. If 17.9% of all at bats featured HoMer’s swinging away, how big would the HoM have to be? I am sure it would be a whole lot bigger than what we have in mind. That being the case, why should the 1870’s merit that kind of representation?

I don't subscribe to the AB by Homer analysis for the 1870s, Dan. NA incompetence and the economic problems that the NA and the NL (as well as every fledgling business) had to endure has artificially created the impression that the '70s are overrepresnted. By the 1880s, the NL had solved many of the financial problems that it had and became a model for up-and-coming competitors (many whose owners originally had minor league teams prior to joining the AA). IOW, while I feel that the NA and NL snatched up the vast majority of the HoM types, there were still players out there to fill multiple team rosters for the NA and NL during the 1870s, just as there were during the 1880s.

First Basemen

Some have pointed out a lack of first sackers in the HoM. Indeed, the four we’ve elected are all dead, the most recent being born in 1858 (Brouthers). Is it that
a. there ain’t any good 1Bmen, or
b. our analysis underrates them?
Either way, I re-looked at a few of the guys on our ballots this week; Jake Beckley (JB), George Sisler (GS) and Frank Chance. Since Sisler has such an incredible break in his career after 1922, I analyzed him as 2 men; GS-peak (1916-22), and GS-career (1915-30). I also chopped off the last few poor years of JB and FC’s careers, so their data is for JB(1888-1905) and FC (1898-1911).

Offense only. Using James’ runs created per game, normalized to a 5 r/g league, where PA is normalized to FC’s career, so JB was twice as long, GS-peak 85% as long:

Player…. RC/G .…PA
FC98-01.. 8.07 ...1.0x
GS-peak.. 8.95 0.85x
GS-career 6.55 ...1.8x
JB88-05.. 6.18 ...2.0x

GS-peak was the best hitter, but about as valuable as Chance when you account for playing time.
JB is to GS-career as FC is to GS-peak; slightly worse hitters, slightly less playing time. Very nice match actually.

So depending on where you draw the ‘replacement level’ line, I believe you can make a case for any of these four. What else do we know?

League strength: Sisler has an advantage here, and Beckley just a small edge over Chance.
Defense: Sisler’s rep was excellent, but Win Shares “failed to show it” according to Bill James. Chance’s rep was very good, Beckley’s not bad either. The BP stats show them all with a +40 to +50 FRAA for their primes.
Sisler and Beckley played almost their entire careers at first base. Chance played almost 15% of his career at catcher! I'd have to say a slight edge on position to Chance.

A little deeper look at Chance, because he is getting no support on our ballots: in the ‘38 election, Beckley finished 6th with 466 pts, Sisler 8th with 415, and Chance 37th with 68.

Chance was on one of the best defensive teams of all time; indeed one of the greatest dynasties of all time. He was considered one of the prime parts of that defense, and probably The Key Player on that dynasty. His team won 4 pennants in 5 years, losing only in 08 to the Pirates who won 110 games. He was the manager, for which we don’t give him credit, but if anything, you sure can’t knock his influence on the team. Chance has very few ‘big seasons’ (WS or WARP), and the prime reason is that he missed games every year. But look at it for a minute; his teams often coasted to a pennant, winning by 20 games in 06, and 13 in 07. So he only played 111 games in 07; did they really Need him every day? Did Chance the manager sit himself out to concentrate on managing duties, and keeping himself rested for a World Series? Chance played every game of the four W.S. form 1906-10; he was always there when needed. Kinda like Yogi Berra taking days off during the Yankee seasons, but no way Stengel would ever sit him down against the Dodgers. Would Chance have played more if he wasn’t the manager? Maybe. Did his taking games off hurt the team’s chances to win? Maybe they lost an extra game or three, but pennants is what mattered, more than 110 victories instead of 108, right??

Frank Chance was a truly great prime player, an OWP of .720, masked by the true deadball conditions, and his strengths of OBA and speed. He captained a truly great team. He started as a backstop, and then played an excellent first base. The only downside to his game was missing a lot of time, but this often didn’t hurt his team any. He was a better hitter than the ‘average’ at his position every year from 1898 to 1911. I understand career voters going for Beckley, but if you are a ‘career voter above average’, taking into account all we know about the men, Chance looks awfully good. Yes, Sisler had a few greater years, but the difference in stats is MUCH more about playing conditions than performance.

I'm not campaigning to elect Chance; but it seems as a group we have been grossly undervaluing him. He will land somewhere between 11 and 17 on my ballot this week.

I'm not campaigning to elect Chance; but it seems as a group we have been grossly undervaluing him. He will land somewhere between 11 and 17 on my ballot this week.

Chance barely made my ballot this time. I also like him better than Sisler.

I used to be a pretty big Chance backer, but more or less gave up on him.
Is there any evidence that Chance benched himself during those blowout seasons? It makes sense intuuitively, and it would boost his status significantly in some eyes.
I never did understand why he played in so few games.

Looking once more at Griffith:
Check out his 1894-99, then consider it's in a 1-league era. Over 300 IP in each year, 1895-99.
Averaged over 300 IP for 1894-1901, when he also averaged 22 wins in those eight seasons.

I almost wonder: If that's all he had, would people more clearly see his case? I realize he has only one off-the-charts ERA+, and he wasn't THE best pitcher around, but his long dropoff and his one-league context seem to have given him a disadvantage.
And this is a guy who used to say it was 'bad luck' to throw a shutout, though he did have 22 overall. He also had three seasons in which he won 20 games and completed at least 35 without a single shutout.

From 1904 on, consider like a single final season of:
18-14 in 277 IP with an ERA+ around 120.
Do that, and his career looks a heckuva lot more like Covaleski's, under arguably tougher conditions.

Clark is in the hunt for my top ballot spot...

Is there any evidence that Chance benched himself during those blowout seasons? It makes sense intuuitively, and it would boost his status significantly in some eyes.
I never did understand why he played in so few games.

There was some discussion about 10 "years" back about Chance's high HBP totals causing him to miss quite a few games. I don't remember who said that or where they got it -- Deadball Stars of the National League perhaps?

Anyhow, that's my 'hearsay' recollection of what that discussion said. Anyone remember more specifically?

as close as I found........

May 30, 1904: In an a.m.-p.m. doubleheader in Cincinnati, the first-place Cubs take on the 3rd place Reds, with just a few percentage points separating the team. The two split the holiday twin bill, the Reds taking the opener 7-4, despite a 9th-inning grand slam by Davy Jones. The Reds then lose, 5-2. Frank Chance of the Cubs is the real loser as he is hit three times by P Jack Harper of the Reds in the morning game, once reportedly losing consciousness when hit in the head. He continues to play and in the 2nd game, he is hit once by Win Kellum, giving him a record four hit by pitched balls for the day.

July 1, 1911: In a 3–0 Chicago win over the host Reds, Cubs player-manager Frank Chance leaves the game suffering from a blood clot in the brain. Except for 11 brief appearances at 1B over the next three years, his playing days are over.

I have my toy that I use to look at offense - a scaled-to-run-environment (using a sliding exponent) Runs Created above average, with an extra bonus for good years, and second calculation with a 75% of league baseline, and a composite of the three numbers. It's a distinctly peak-leaing measure, and I must also add knowledge both of the length of career and of the quaility of defense. Here are the results for two consideration sets for this year - one set for outfielders and first basemen; the other for non-1B infielders and catchers. I included Leach in the infield group.

The outfield-1B group, ranked by:

Scaled RCAA:
Chance, Tiernan, Van Haltren, Fournier, Duffy, Thomas, Browning, Roush, Ryan, Sisler, Cravath, Beckley. (CF only: VH, Duffy, Thomas, Roush, Ryan, Burns, Carey, F. Jones).

Big-year bonus score:
Chance, Sisler, Tiernan, Fournier, Cravath, Youngs, Duffy, Burns, Stone, Thomas, Kauff, Donlin, Browning. (CF only: Kauff, Roush, VH, Ryan, F. Jones, Carey)

Scaled RC above 75%:
Beckley, Van Haltren, Ryan, Sisler, Duffy, Tiernan, Hooper, Roush, Carey, Chance, Burns, Konetchy, Fournier. (CF only: VH, Ryan, Duffy, Roush, Carey)

Composite of the three scores:
Chance, Tiernan, Sisler, Van Haltren, Duffy, Fournier, Ryan, Roush, Burns, Thomas, Beckley, Cravath, Browning. (CF only: VH, Duffy, Ryan, Roush, Thomas)

The infield-catcher group, ranked by:

Scaled RCAA:
Doyle, McGraw, Evers, Bresnahan, Childs, Schang, Leach, Sewell, Jennings, Pratt, Bancroft, Tinker.

Big-year bonus score:
McGraw, Doyle, Childs, Bresnahan, Jennings, Evers, Leach, Schang, Sewell, Pratt, Bancroft, Kling, Tinker, Bush.

Scaled RC above 75%:
Doyle, Leach, McGraw, Evers, Childs, Bresnahan, Sewell, Schang, Pratt, Jennings, Bancroft, Long, Tinker, Bush.

Composite of the three scores:
Doyle, McGraw, Childs, Evers, Bresnahan, Leach, Schang, Jennings, Sewell, Pratt, Bancroft, Tinker, Long, Bush.

Thoughts now that I line it up this way:

I've made no attempt to adjust for league strength. If you believe the AL was superior to the NL, that knocks down such players as Doyle and Carey.

By my own standards, I have probably been rating both Childs and Jennings too low.

McGraw looks fantastic, and so does Chance. This isn't a totally career-insensitive rate stats approach (else Stone, Donlin, and Youngs would rank higher), but it doesn't really take into account games played. And McGraw played fewer than 1000 games and Chance fewer than 1300 - that does matter.

I've got some players popping up here who are just inconvenient, since they have already been dismissed by the electorate. Most prominently: Mike Tiernan (offensively as good as - maybe better than - Thompson; defensively a butcher?). Also: Johnny Evers, who by this measure looks about as good as Childs, and he's got defensive credentials.

There's been a sort of mini-boom in support for Ed Konetchy. My problem with that: other than career length, what does he have over Jack Fournier? I take Chance, Sisler, and Beckley as our best eligible first basemen. (Lou Gehrig is still an active player, as far as we know - we'd be very shocked to hear that his career will last just 8 more games. Of course, none of these guys is Gehrig. The Hall of Fame went ahead and considered Gehrig immediately. We'll follow our rules and wait the five years because that's what we do.)

Re: Frank Chance and HitByPitches:

From BaseballReference: He was in the Top10 every year from 1900-1908 plus 1910.

From Deadball Stars of the National League: He was "plagued by injuries throughout his career, most notably a barrage of beanings due to his propensity for crowding the plate." Ring Lardner wrote, "There are pains that shoot through his head and other parts of his body too numerous to mention, and make it just about impossible for him to take an active, playing role in the game at which he is so expert."
By 1911, he had lost the hearing in one ear and part of it in the other.
In 1912, he had to be hospitalized for blood clots on the brain.

From the first MacMillan Encyclopedia: 1901 - broken wrist, 1909 - broken shoulder, 1911 - ankle injury. It doesn't list the causes.

Hope this helps...

Congrats on passing the bar Matt!

I had just written a pretty lengthy response to the series of posts regarding Campanella (TomH #29), etc.. Damn IE crashing while I was doing a 'find' on the thread, flipping brutal.

Anyway, Tom's explanation, about Campanella being steady hypothetically being more valauble than his big year/bad year career is one of the big reasons why big years are in general overrated. They are slightly more valuable on average (10-15% for a "10" and a "0", as opposed to two "5's"). There can be diminishing returns on the big years (sometimes they aren't needed). Everyone seems think of the year when the big year puts a team over the top, but they forget the year where an average year as opposed to bad year would also have put a team over the top - my favorite example is the 2002 Mariners - replace Jeff Cirillo with an average 3B and we may have had a different champion.

I also disagree with Michael that 31 vs. 26 is significantly more valuable than 20 vs. 15. I need to take another look at the formula, and I don't have the time right now, but I believe I posted the formula on the Pennants Added thread, and the numbers could easily be plugged in if someone wants to take a crack at it using hypotheticals.

As for the individual candidates, I cannot fathom Sisler getting more support than Roush. I think Tom is right that the peak guys and career guys get support and the tweeners are getting shafted.

I also think we are systematically underrating 1B defense from this era. Charles Comiskey and Hal Chase were considered great players. While they obviously, they weren't, I think that should give us a hint that before Babe Ruth started blasting homers, first base was a much tougher defensive position. Bad gloves made the position much tougher on the hands, and the deadball era made the position much more involved in the game. Just catching that darkened ball with a small poorly padded glove probably required more skill than we generally give credit for, for example.

Still catching up . . . more to come.

Kelly - #107 - phenomenal post.

If I can muster the time, I will answer all of your questions for each candidate on this ballot. I encourage others to do the same. Maybe we should post that in the comment when we launch the voting (hint John, if I run out of time reading this thread and you end up posting it, which has become the trend :-).

"Now, feel free to count many of those runs as an indicator of his having sucked that day. He caused the team to lose the game, after all.
But at what point does the law of diminishing returns kick in?
Who cares if he allows an 11th, 18th, or 23rd run that day? I doubt that he or his teammates did; he's just finishing out the game."

Very important for evaluating a pitcher - a pitcher that can be shown to give up many 'blowout' runs has an advantage over another that gave up the same number of runs, that is an absolute certainty. But the question with Welch is - was he atypical of his era in this respect. If so, it's a point in his favor. If other guys were doing the same thing, then it doesn't make him any better than them.

Pitcher A gives up 11 runs and 1 run in two starts. Pitcher B gives up 4 runs in each of two starts. Who helped his team win more?

Depends on the league average of runs/game.

If the league average is 2 runs per game, A would be expected to win game 1 .028 of the time. Game two they'd win .721 of the time, total wins, .749.

Pither B wins each game .239 of the time, for a total of .478 wins.

But again, it has to be proven that Welch was out of the ordinary in this respect.

Someone like Chris J, who has the data for each start of all of these guys careers (I think) could conceivably come up with a Pythagorean expectation for every start (based on park adjusted league average run support), add them up and get something like Win Value Added. This would be a better approximation than strictly using ERA+ or W-L, etc..

But it's a ton of work :-(

Brent:

"My understanding is that the Wolverton measure is based on a similar thought experiment – randomly take a player and add him to a randomly selected team and see how often the team wins a pennant that it otherwise wouldn’t have. The results are then accumulated over the player’s career to yield a measure of the player’s “pennants added.”"

Close - not adding him a random team - adding him to EVERY team ever (not exactly, he uses standard deviations to do the statistical equivalent), and seeing how many that didn't win the pennant would otherwise have won the pennant, with this player as opposed to a replacement player. Very important diffference though.

A key stage is taking out the teams that would have won the pennant anyway (the Campanella dilemma).

I wasn't pro or against peak or career. Wolverton didn't come up with this method to defend the types of players he liked. He developed the theory, put it into practice and it told us what it told us. It just happened that the metric more closely mimicks a straight career value rating as opposed to a straight peak value measure. There wasn't any bias built into it. The theory behind the metric, if you ask me is flawless, indisputable. The only things that are debatable are the things that always are, where to set the replacement level, and the metrics used to rate each season in the first place.

James' study showed a slight advantage for the Carlton/Drysdale vs. the Sutton Pappas. Many people have taken this too far, and read it to say that the Drysdale type was more valuable than Sutton which the study clearly shows to be false. All James study shows is that all other things equal the pitcher with somewhat bigger seasons is slightly more valuable. To the tune of about 1/5 of a pennant over 20 years IIRC.

Also, every time James made the study more realistic - like varying the amount of wins required for a pennant, and the quality of the pitchers teammates - the gap shrunk - to the point where it almost disappeared. The James study and Pennants Added both show the exact same thing. Big years are slightly more valuable, all other things equal - that's all.

"To contribute to a pennant contender, a player first must be good enough to play on the team."

Miguel Cairo played for the Yankees this year. Even great teams have weaknesses, most contenders are not great teams. Again, Jeff Cirillo was good enough to play for the 2002 Mariners. I think this factor is an extremely minor. Many teams failed to win a pennant because they could not find an average player to fill a hole. History is littered with them. The Cubs could have been the ones to take St. Louis to a Game 7 this season if they had found an average player to play SS for 4 months this year, as another example.

"You're right. My system is not purely a peak voting system. (I also agree with sunnyday2 that there are probably not any pure peak voters here.) But it does capture some of the special value of high-peak, short-career candidates like Jennings. I think of it as a "career peak" measure. But there are lots of ways to evaluate a player's career - I just wanted to present the method that I find most convincing, though I'm certainly interested in looking at the question from other perspectives as well."

The problem is that Jennings isn't any more likely to end up on a great team than on a team that wins 65 games a year, but for him - see Banks, Ernie. Especially in the era before free agency. Well, I guess that was too strong, the best teams will do a better job of identifying talent, etc. - but I think general premise still holds. I think the only fair way to do it is to add the player to all teams and see how many he pushes over the top. I suppose I could see using a slightly higher replacement level for this purpose, but I still feel funny about it. I need to stew on this one for a bit . . . any other thoughts?

Two part post:
I thought I would examine Jennings under the 8 Arguments from post 107. I don't think I have ever had Jennings on my ballot. Mainly because I thought his career was too short and his value was concentrated into too small a window as I generally emphasize 7 year "prime" which does not have to be consecutive. Anyway, I thought I would take a look at him again to see if I should have him on my ballot.

1. Best Candidate, not merely qualified.
Among Shortstops, I posted above that I thought Sewell was the best candidate at the position. I could definitely be wrong about that. I was basing it on the longer time as the best player in his league at the position, better "ink" scores, and significant durability carrying a little more weight than Jennings' much higher 3 yr peak and, by win shares, slightly better defense. Definitely, I can see the argument for Jennings here as the best candidate at shortstop.

2. Combined Weight of Evidence.
Should probably be looked after the other arguments. We'll rejoin later.

3. Comparable player in the Hall of Merit.
This is a key topic as how you define "comparable" may decide how you see Jennings as a candidate. What constitutes a comparable player to Jennings? A person with similar career totals? A player with a similar career shape?
Looking at similarity scores for his career, his top 10 look like this:
Cecil Travis (898)
Ezra Sutton (888) HoM
Cupid Childs (877)
Johnny Pesky (875)
Jimmy Williams (867)
Bill Hallman (866)
Lou Bierbauer (864)
Buck Ewing (863) HoM
Tom Daly (862)
Billy Shindle (856)
This list is a plus, 2 HoMers among the Top 10. However, there is a caveat. One of them was a catcher (mostly) and both played much of their careers when schedules were shorter than when Jennings played. Also, it is doubtful any other will end up in the HoM (as much as I vote for Childs...)
On the other hand, "comparables" to his 5 years of excellence, 1894-1898, place him among the best of his peers - all of whom are HoMers: WinShares totals for position players 1894-98:
1. Billy Hamilton 150
1. Hughie Jennings 150
3. Ed Delahanty 140
4. Jesse Burkett 137
5. Joe Kelley 136
Keeler 126, Duffy 123, GVH 119, George Davis 118, Dahlen 113.
Even expanding to 7 years (93-99), Jennings is 6th though his fragility is now a factor:
1. Delahanty 209
2. Burkett 191
3. Kelley 186
4. Hamilton 185
5. Duffy 168
6. Jennings 161
7. Davis 160
8. GVH 157
8. Keeler 157
10. Dahlen 153
So, if "comparable" means career totals, there are some definite doubts. But if you focus on the seasons of excellence, he has some definite points in his favor.
But looking at comparables based on his career path of average or injury-shortened periods surrounding an intense period of greatness, I don't think there is another HoMer with such a career path who is not a pitcher.

4. Several Comparable Players are HoMers.
Again, this depends on your definition of "comparable". See the previous question. To look at it in other ways... When WS was published, Jennings was tied with 4 others for 469th most win shares. There were 49 players with 5 WinShares of him. Looking at the list, Jennings may be the best qualified for the HoM, though Charlie Keller, Bill Mazeroski, Ernie Lombardi, Frank White, and Dave Stieb will have their supporters.
Or, looking at the most similar players with similar defensive achievements: Jennings ranks as an A+ fielder, so do the following shortstops from 1876-1930: Bob Allen 1890-1900, Bill Dahlen 1891-1911, Mickely Doolan 1905-1916, Art Fletcher 1909-1922, Herman Long 1889-1903, Rabbit Maranville 1912-31, Everett Scott 1914-26, Germany Smith 1884-98, Joe Tinker 1902-16, Honus Wagner 1901-17, Monte Ward 1881-91. While they were all "A+" in the field, there was a large difference in the estimated innings in the field:
Maranville 19191
Dahlen 18754 homer
Wagner 16971 h
Long 15916
Tinker 15406
Smith 14658
Doolan 14468
Scott 14189
Fletcher 12694
Jennings 07845
Ward 07253 h
Allen 05396
What they did at the plate also was quite different - ranked by OPS+:
Wagner 150 h
Jennings 117
Dahlen 110 h
Fletcher 100
Tinker 096
Long 094
Ward 093 h
Allen 088
Maranville 082
Smith 074
Doolan 072
Scott 066

So do the above numbers help Jennings? He was the second best hitter among great defensive shortstops. But he played the 3rd, and probably 2nd if Ward had played in longer seasons, fewest defensive innings.

part 2
5. Comparables not in the HoM?
Comparables with his 3 or 5 year peak? Hugh Duffy had 90 ws vs. Jennings 97. Cravath 92. Seymour 93. Fournier 90. No position player who has more in 3 straight years is NOT in the HoM. But there are several who are very close to Jennings who are NOT IN. Over a 5 straight year run, I don't know, I don't have the numbers, but I don't think there are any players who would rank as the best over 5 yrs (even if tied) who are not in the HoM.

On the other hand, players with similar career totals who are HoMers number only two players whose career numbers were restricted by the lengths of seasons they played. And looking into the future, there are very few if any candidates with similar career total worth as measured by Win Shares.

6 Highest Common Denominator Argument:
Difficult to apply because of how odd Jennings' career is. But there are definitely comparable players who are not in the HoM. But those players had a more consistent spread of value season-by-season. Who do you think is the right set of "comparables" for Jennings? Just on peak, just on career shape, just on total career value?

7. Previous HoM standards at the position.
Jennings did not have a career anywhere the length of previous HoM inductees at short - Davis, Dahlen, Wagner, Ward, Lloyd, Glasscock, and Wallace all had longer careers. Wagner and Lloyd were better hitters while Dahlen, Davis, and Glasscock are comparable. As a fielder, George Davis is the only one without an excellent reputation.

8. Middle of group?
Do you weigh rate more heavily or career totals?
If you weigh rate than he is in the middle, but by career, no.

2. Weight of all evidence/redux.
Jennings is a keystone for the debate over career totals vs. peak/prime contributions. Here is a player who ranked in the top 5 of all position players for 5 years in a row. Yet in only one other year was he even in the top half of shortstops. Here is a player who tied with Billy Hamilton for most win shares for 1894-98 with 150 and played on 5 teams that finished with the best record in the league. Yet the other players who were close to him in win shares for the period finished with 337 ws career - Hamilton, 355 - Delahanty, 295 - Duffy, 344 - Van Haltren, 389 - Burkett, Davis - 398, Dahlen - 394, Kelley - 305, Keeler - 333 while Jennings finished with 214. He hit better than most shortstops and fielded better than most. Just not for very long...

So, who is he "comparable" to? How much weight do you place on 5 years vs. the career as a whole?

Friends of Jennings/Enemies of Jennings - comments?

when Max Carey came up with the Pirates, he played LF form 1912-15 before moving to CF. I've heard that this was largely because his speed was needed in the spacious LF area in Pittsburgh. If this is so, do WS and WARP under-rate his fielding contributions for those years, assigning him the LFers' share of credit, when it's obvious he could have played a fine CF?

another Carey note: probably would have been the MVP of the 1925 Series, won by Pitts in 7 games. He hit .458, stole 3 bases, and scored 6 runs.

I don't think there is another HoMer with such a career path who is not a pitcher.

Ross Barnes.

"If I can muster the time, I will answer all of your questions for each candidate on this ballot."

I meant each candidate I vote for, not every player that retired prior to to 1934 :-) I'm not a lunatic.

Kelly, nice job with Jennings. I'm going to maybe try to do two per night at least tonight through Wednesday.

Kelly #189
3-yr peak Win Shares:
Hugh Duffy had 90 ws vs. Jennings 97. Cravath 92. Seymour 93. Fournier 90.

This is unfair to Jennings. Number of games scheduled during 3-yr/5-yr peak seasons: Duffy 139/138, Jennings 132/136+, Seymour 144+/151+, Cravath and Fournier 154/154.

So, who is he "comparable" to? How much weight do you place on 5 years vs. the career as a whole?

Yes, Ross Barnes,
although Barnes was more dominant in his six seasons than was Jennings in his five. And the difference between six and five is significant. Barnes may be a paradigm rather than a threshold case.

"Comparable" players may be identified by career pattern rather than career peak or sum. Justin Kubatko published something like that for all candidates, in support of the "Baseball Survivor" project. I suppose that TomH can provide a web reference. But Kubatko used all seasons ordered by value rather than chronological order, I think. For example,

OPS+, Cy Seymour 1901-1910
181 137 134 134 132 121 110 102 94 92

(OPS+ is a hand example. Kubatko didn't use OPS+ and Jennings isn't similar to Barnes by OPS+. Ten seasons matches Seymour's career as a semi-regular non-pitcher. Kubatko used multiple numbers of seasons, akin to the use of 3-yr and 5-yr peak in this forum.)

In regards to Frank Chance and being beaned, I found an article in an April, 1918 issue of Baseball Magazine. It is titled "Reminiscences of the Old-Timers" and the three old-timers in question are Frank Bancroft, Fred Pfeffer and Bid McPhee. It is McPhee's account I will bring forth. It is found on p. 358, 383-384.


THREE HUMAN TARGETS
“THERE have been three men in the
history of the big leagues,” says
good old Bid McPhee, the famous
second baseman, “who stood out preeminent
above all others when it came to
being hit by pitched balls. Those three
men were Curtis Welch, Hughey Jennings
and Frank Chance. They have never had
any rivals, and I don’t suppose any play-
(Continued on Page 383)
358 BASEBALL MAGAZINE for AUGUST
——————————
REMINISCENCES OF
THE OLD-TIMERS
(Continued from Page 358)
ers are really anxious to equal the records
that they made.
“The strangest thing of all, so far as
two of these men were concerned, was the
fact that they didn’t mean it. They were
game, all right, but not game enough to
voluntarily run the risks they seemed to
take, and the countless bruises they received
were not endured to help their
teams, but because they couldn’t dodge.
It’s a fact—Hugh Jennings could not
dodge a ball, and Frank Chance couldn’t
duck one either. As for Curtis Welch, he
was a different proposition. He got hit
intentionally time after time.
“The box scores of the old Baltimore
games seldom went into print without the
words, ‘Hit by pitched ball, Jennings.’
Time after time Hughey was cannonaded,
and, as a rule, was bumped hard. His
nerve and gameness were widely praised,
while some of the critics said he was simply
foolhardy. The latter opinion was
almost universal when, after Baltimore
had safely won the flag, Hughey continued
to get thumped, and to get hit in games
where the Orioles were miles ahead.
“Not until late in his career was it discovered
that Jennings never meant to get
soaked at all. All these years, he had been
unable to dodge. He seemed unable to
convey the sense of danger from his brain
to his limbs quick enough to spring aside.
”Frank Chance was hit in the head no
less than 38 times, and stunned on twenty
occasions, while he received innumerable
smashes on shoulders, ribs or legs. Chance
wanted to get out of the way, but couldn’t.
He stood flat-footed, resting heavily on his
pins when batting, almost imbedding them
in the ground, and he couldn’t sidestep or
make his feet move in time to save him.
I doubt if he could even dodge a slow ball,
his feet moved so rebelliously.
“Curtis Welch stood lightly set upon
his feet, and could spring away from the
fastest pitching, if he desired. But he
didn’t desire, and so agile, so snake-like
was he, that he could seem to be grazed by
purest accident every time. He even developed
a way of glancing the ball off his
forearm, apparently dodging, yet leaving
the arm exposed and letting the ball tick
against it. Finally, the National League
introduced a rule by which a batter got
no base if hit on arm or hand, and this
rule, created especially to check Curtis
Welch, was rescinded when that great and
tricky player died.”

Thank you Esteban; love to read that stuff.

OPS+ is "at hand" as I write, a "handy" example.

Elsewhere here at BTF.org, there is a 2-year old article and blog on the Baseball Survivor project.
http://www.baseballthinkfactory.org/files/main/article/malcolm_2002-02-20_0/

scruff #91 says, "I love player data pages, especially the most similar at peak years tables." (emphasis mine) Those are the tables that I recalled above, regarding "comparable" careers. I don't find them on the web now.

Modified "player by position" list below.

Numbers are the percentage of total games played by each HOMer at that position. 10 pct minimum.

C (3.15 'players') - BENNETT 88, SANTOP 75, EWING 47, Kelly 36, McVey 30, White 28, O'Rourke 11

1B (5.69) - START 100, BROUTHERS 98, CONNOR 88, ANSON 83, Stovey 37, MCVEY 31, Lloyd 25, Heilmann 22, Ewing 19, Kelley 16, Delahanty 15, Hines 12, Lajoie 12, Spalding 11

2B (5.64) - MCPHEE 100, E COLLINS 98, LAJOIE 83, GRANT 70, BARNES 69, RICHARDSON 43, Ward 26, Johnson 25, Groh 20, Hill 20, Wright 10

3B (4.64) - BAKER 100, J COLLINS 98, GROH 79, SUTTON 69, WHITE 51, Davis 22, Wallace 18, McVey 14, Richardson 13

SS (8.21) - PEARCE 96, GLASSCOCK 94, WRIGHT 89, DAHLEN 88, WALLACE 77, JOHNSON 70, LLOYD 70, WAGNER 68, DAVIS 58, Ward 44, Barnes 28, Grant 20, Sutton 19

OF (22.18) - CLARKE 100, HAMILTON 100, THOMPSON 100, WHEAT 100, BURKETT 99, COBB 99, FLICK 99, GORE 99, SHECKARD 99, SPEAKER 99, JACKSON 98, KEELER 97, CRAWFORD 94, MAGEE 91, HINES 82, TORRIENTE 80, KELLEY 79, HEILMANN 77, DELAHANTY 72, HILL 70, O'ROURKE 69, STOVEY 63, Caruthers 50, KELLY 47, Richardson 40, Santop 20, McVey 18, Ewing 17, Davis 13, Spalding 13, Wagner 13, Ward 11, White 10


Major caveats: Doesn't account for change in length of season for early-era players. Also the Negro Leaguer numbers are guesswork, part-time pitcher numbers aren't ideal, not sure on pre-1871 positions, and so on. Still, it's a lot sharper than before. 'Corrections' welcomed on any and all players.

After the refinement, we see more there have been 1B and OF 'seasons' than by the more primitive method, while SS looks a little less 'oversold' and C looks even more scarce.

Your welcome, jimd. What really jumps at me is that Chance was hit in the head 38 times! With no batting helmet! No wonder he went deaf, suffered a brain clot, and died sooner than you would expect. As an aside, the Pfeffer and Bancroft anecdotes were humorous. I'll include them here:

THE TRUSTFUL MR. HOTALING
“REMEMBER Pete Hotaling?”
queried Fred Pfeffer, in a reminiscent
mood. “He was one of the
old birds—good batsman, sturdy all-round
player—fine type of the bygone generation
—and yet a type that was wholly different
from the tricky, fast-thinking stars like
Mike Kelly, Buck Ewing and Arlie Latham.
Old Pete thought of things at a leisurely
pace, reflected over them before he
decided to do them, and then solemnly executed
them when he got around to it.
Incidentally, Pete was a trustful and confiding
individual. He had implicit faith in
the honesty and good intentions of his
fellow-men, and hated to think that anybody
would deceive him. That is, Pete
used to think that way—but, after he had
been bumped around the big circuit several
seasons, I am inclined to believe that some
of his ideas must have changed. Those
were days of sinfulness and deceit; days
when they’d do anything short of murder
if the one umpire was looking the other
way, and it must have been hard for any
one as innocent and trustful as Pete Hotaling
to get by.
“If I live a century, I’ll never forget
the richest example Hotaling ever gave of
his confiding disposition. Playing against
us one afternoon, Pete slammed a snorting
single, and broke for second a moment
later. Mike Kelly, who could throw them
high and far when he did happen to toss
one astray, shot way over my head, and
the ball went out in center field. Here, it
got past George Gore, and rolled towards
the end of the park, with Gore in desperate
pursuit. Hotaling, who had slid headfirst
to second, never saw the ball go by, and I
made a despairing effort to keep him on
second by slamming my empty hand on his
back as he reposed across the bag.
” ‘That’s right, Fred!’ chirped Ed
Williamson. ‘Keep that ball, and tag
Pete if he edges off that base!’
“ ‘Leave it to me,’ I answered. ‘Nice
slide, Pete, but that’s as far as you’ll get
today!’
“Gore was still chasing the ball, and
every member of Hotaling’s crowd was
screeching for him to get up and come
home. Pete rose, keeping one foot carefully
on the base, and shook his head in
answer to the wild demands that he score.
At last the ball came back, and I saw a
look of absolute bewilderment and sorrow
on Hotaling’s face as Clarkson regained
the leather. The next man popped out,
Hotaling walked off the base to resume
his field position, and at least seven of his
crowd surrounded him to ask him why in
blazes he hadn’t scored.
“ ‘I thought the ball was on second’,”
he moaned in misery. ‘Anyway, Pfeffer
and Williamson told me it was, and I
didn’t think they’d lie to a fellow!’ “
TWO NOBLE SLEEPERS
“BALL CLUBS,” says good old Frank
Bancroft, “have been very fortunate
in escaping railroad accidents.
Only a few instances are recalled where
serious injury has been inflicted on any
number of ballplayers while traveling, and
when you consider the number of the boys
who are making the circuits from April
to October their good luck seems almost to
justify the belief that they have charmed
lives. Charlie Bennett, of course, lost his
legs in a railroad wreck, and Jimmy Ryan
was so badly battered, on one occasion,
that he was out of the game most of the
season. Taken on the whole, however, the
boys have been the luckiest of all the people
who make numerous journeys on the
rails.
“I was in one smashup, 25 years ago,
that was a corker, and yet had its funny
side. The Reds—with Comiskey in command
and old Pete Browning as the star
slugger—were coming across from St.
Louis, and, in the middle of the night, a
freight crashed into us. A scene of horror
and confusion followed, doubly augmented
when the two sleepers stood on end and
then pitched down a deep incline.
“Somehow, some way, we scrambled out,
in pajamas or without them, and took a
census of our numbers. Bid McPhee had
a smashed nose, Frank Dwyer had a
skinned elbow, and there were a few minor
injuries, but nobody was killed or even
crippled. And then, to our utter horror,
we found that Charles Comiskey and Pete
Browning were missing!
“We remembered that they were in the
forward end of the second sleeper, which
was now on fire. If they had not been
killed by the fall, they would soon be
burned, and there was no time to lose.
With the train crew, we grabbed axes and
handspikes. We toiled madly, and in a
jiffy we had cut a road into the side of
that blazing car.
“We struggled along the aisle of the
car, almost tilted on end, and consequently
a tough place to travel. The green curtains
had fallen out, bedding and splinters
were heaped in the aisle, and the
flames were gaining ground. Onward we
struggled, and at last we pulled aside the
last obstructions, dreading what might be
revealed.
“It was some reveal, too, for there lay
Charles Comiskey and Pete Browning,
standing on their heads in the berths,
with their feet pointing heavenward in the
upturned car, and both of them snoring
like slumbering bulls! The crash, the
shock, the tilting of the car, had never even
roused them from their happy sleep, and
they were mad as hornets when we woke
them up. Sleepers? There never were
their equals since the world began!”

Moving along to a different subject, that of the abilities and disabilities of using win shares for fielding evaluations. I was looking over the shortstop group of Tinker, Long, Bancroft, et al, and I started wondering if any of them were being underrated by win shares. It has been mentioned here many times about the cap on fielding win shares per game (no more than .32375 per game played) and about certain teams that hit the cap each year (1890 's Boston Beaneaters and Chance's Chicago Cubs). So I decided to find out which teams actually hit the cap. The following is a list of all the significant teams from 1876 to 1940 that are capped in fielding win shares. Maybe by studying these teams patterns could be found. Maybe by looking at player turnover each year one might figure out which players on that team were responsible for the cap. Maybe some teams reached the cap because they had two great middle infielders. Maybe they reached the cap because the pitching was exceptional and the fielders are credited with the excess that the pitchers are not, or vice versa. Maybe they had really great hitting and the excess spilled over to pitching and fielding. Or maybe there is nothing there. But, in the vain hope that I didn't invest my time for naught, here is the list.

1876 - Chicago, Louisville, St. Louis, Hartford
1877 - Boston, Louisville
1878 - Boston
1879 - Chicago, Buffalo, Boston
1880 - Chicago, Cleveland, Providence
1881 - Chicago
1882 - Providence (NL), Cincinatti (AA)
1883 - Boston (NL), Cleveland (NL), Providence (NL), Cincinatti (AA), New York (AA), St. Louis (AA)
1884 - Boston (NL), Providence (NL), Cincinatti (UA), St. Louis (UA), Columbus (AA), Lousiville (AA), New York (AA), St. Louis (AA)
1885 - Chicago (NL), New York (NL), Pittsburgh (AA), St. Louis (AA)
1886 - Chicago (NL), Detroit (NL), Philadelphia (NL), St. Louis (AA)
1887 - Chicago (NL), Cincinatti (AA), St. Louis (AA)
1888 - Philadelphia (NL), Brooklyn (AA), Cincinatti (AA), St. Louis (AA)
1889 - Boston (NL), Cleveland (NL), St. Louis (AA)
1890 - Boston (NL), Chicago (NL), Cincinatti (NL), Philadelphia (NL), Louisville (AA), Boston (PL), Chicago (PL)
1891 - Boston (NL), Chicago (NL), St. Louis (AA)
1892 - Boston, Cincinatti, Cleveland, Philadelphia
1893 - Boston, Cincinatti, Pittsburgh
1894 - Baltimore, New York, Cleveland
1895 - Baltimore, Cleveland
1896 - Baltimore, Cincinatti, Cleveland
1897 - Boston, Cincinatti, Cleveland
1898 - Boston, Cincinatti
1899 - Baltimore, Boston, Brooklyn
1900 - Boston, Pittsburgh
1901 - Boston (NL), Philadelphia (NL)
1902 - St. Louis (AL), Pittsburgh (NL)
1903 - Boston (AL), New York (NL)
1904 - Boston (AL), Chicago (NL), Cincinatti (NL), New York (NL)
1905 - Philadelphia (AL), Chicago (NL), Pittsburgh (NL)
1906 - Cleveland (AL), New York (AL), Chicago (NL)
1907 - Chicago (NL)
1908 - Cleveland (AL), Chicago (NL)
1909 - Philadelphia (AL), Chicago (NL), Pittsburgh (NL)
1910 - Philadelphia (AL), Chicago (NL)
1911 - None
1912 - Boston (AL), Washington (AL), Pittsburgh (NL)
1913 - Cleveland (AL), Washington (AL), New York (NL)
1914 - Boston (NL), Buffalo (FL)
1915 - Boston (AL), Washington (AL), Pittsburgh (FL)
1916 - None
1917 - Boston (AL), Cleveland (AL)
1918 - Chicago (NL)
1919 - New York (AL), Chicago (NL), Cincinatti (NL)
1920 - Brooklyn (NL), Pittsburgh (NL)
1921 - Boston (AL), Pittsburgh (NL)
1922 - New York (AL)
1923 - New York (AL)
1924 - Washington (AL), Pittsburgh (NL)
1925 - Washington (AL)
1926 - Cleveland (AL), Philadelphia (AL), Chicago (NL)
1927 - None
1928 - Chicago (NL)
1929 - Philadelphia (AL)
1930 - Washington (AL)
1931 - Philadelphia (AL), Washington (AL), St. Louis (NL)
1932 - Washington (AL)
1933 - Washington (AL)
1934 - St. Louis (NL)
1935 - Chicago (NL), St. Louis (NL)
1936 - Chicago (NL)
1937 - Chicago (AL)
1938 - Chicago (NL)
1939 - Chicago (AL), New York (AL), Cincinatti (NL)
1940 - Brooklyn (NL), Cincinatti (NL)

Esteban, great post. There are some very interesting pieces of information on it. Thank you for taking the time.