This will be fun. :-)

Looks like the nineteenth century stars will fill up the top and bottom of this ballot, with the 1900-1940 contingent mostly in the middle. To put it another way: the HoF's most glarig omissions are all from the 19th century, while all of our choices from the first half of the 20th century are pretty sound, with only Stan Hack being borderline, and our most questionable selections have come from the nineteenth century.

Here's my first pass at combining the 19th century and 20th century players.

Astonishing Oversights (comparable to with Blyleven, Raines, Grich, and Santo)

Bill Dahlen 366
Paul Hines 310
Deacon White 301

Clearly Deserving (comparable to Trammell, McGwire, Whitaker, Simmons, Allen)

Joe Start (call it 290)
Jimmy Sheckard 282
Ezra Sutton 273

Solidly Qualified

Joe Gordon 267
Heinie Groh 266
Jack Glasscock 261
Charlie Keller 261
Sherry Magee 260
Wes Ferrell 255
George Gore 252

Lower-Tier but In

Cupid Childs 247
Stan Hack 246
Hardy Richardson 241
Charley Jones 240

Questionable HoM selections.

Charlie Bennett 211
Harry Stovey 210
Bob Caruthers 205
Pete Browning 188

Ferrell, Magee, Sheckard and Bennett will be at the bottom of my ballot, while Dahlen, Sutton, White and Hines will be at the top. Beyond that? You got me.

Do you give a catcher bonus Chris? I can't see Bennett as being very questionable . . . I agree with the other three.

For me Jimmy Sheckard is the biggest surprise on Chris Cobb's list.
Where would you place Tommy Leach, Chris?

Offhand my bottom four are Richardson, Jones, Browning, and Childs in some order. I don't see Caruthers as questionable but I would be a McVey, Cravath, and Keller voter too.

Say, where's Beckley?
;-)

Full Seasons Equivalent games played at fielding position (now with LF-CF-RF)
ordered by mlb debut year
Start - 1b 14.39(#5) - Yes, that's fifth all-time in full seasons at first base, although he was an old 29 at the starting gun
, of 0.08; sum 14.40 (#100)
White - c 7.42(#77)
, 3b 7.12(#84), 1b 1.48 ( 18.14 (#15, second among eligibles)
Sutton - 3b 11.84(#18)
, ss 2.96; ; sum 16.32 #36, sixth among eligibles)
Hines - of 15.12(#11) and L-C-R 1.42 13.60(cf#6) 0.12
, 1b 2.27; sum 18.72 #10, first among eligibles)
Jones - of 10.12(#120) and L-C-R 7.43(lf#49) 2.68 0.07
, 1b 0.18; sum 10.27
Bennett - c 8.61(#45)
, of 0.76; sum 9.79
Glasscock - ss 13.72(#12)
, 2b 1.00; sum 14.94 #74)
Gore - of 11.50(#76) and L-C-R 0.71 10.56(cf#26) 0.26
, 1b 0.20; sum 11.65
Richardson - 2b 5.19(#144)
, of 4.58(#556) and L-C-R 2.87 1.63 0.07, 3b 2.09; sum 12.03 #267
Stovey - of 7.39(#272) and L-C-R 3.93 1.53 1.94
, 1b 5.24(#163); sum 12.56 #212
Browning - of 7.59(#258) and L-C-R 3.59 3.77 0.26
, 3b 0.74; sum 9.55
Caruthers, pitcher (no full seasons statistic)
, of 2.51 and L-C-R 0.44 0.00 1.97
Childs - 2b 10.45(#28)
, ss 0.01; sum 10.46
Dahlen - ss 14.72(#4)
, 3b 1.59 ( 16.85 (#29, fourth among eligibles)
Sheckard - of 13.75(#31) and L-C-R 12.24(lf#11) 0.15 1.41
, 3b 0.09 ; sum 13.94 #124
Magee - of 12.10(#55) and L-C-R 10.43(lf#18) 0.89 0.80
, 1b 0.96; sum 13.43 #154
Groh - 3b 8.62(#55)
, 2b 2.08; sum 10.82

Ferrell, pitcher (no full seasons statistic)
, of 0.086
Hack - 3b 11.88(#17)
, 1b 0.30; sum 12.17 #256
Gordon - 2b 9.80(#35)
, 1b 0.19; sum 9.96
Keller - of 6.60(#345) and L-C-R 5.70(lf#95) 0.00 0.90
; sum 6.60 (played only outfield)

That is a chronological gap between Groh and Ferrell. The latter pitched one inning in September 1927 when Groh finished up as a barely-player.

The first nine have major league debuts before 1880 and Stovey was an established pro in '79.

This will be a much, much more difficult vote than the previous two. OK, Dahlen on top is obvious, but beyond that ...

Gore, Hines, and White were elected before I joined the HoM electorate - in essence, I never considered them. Now I'll have to investigate their cases.

I first voted in 1904 - Glasscock's election. I remember being solidly convinced at the time, but of course he's no Dahlen.

There are some people here I was never much of a fan of: Browning, Jones, Caruthers, Gordon. I'll have to check my own ballot record.

Early on, I was the one of the strongest advocates for Harry Stovey. I'll have to dredge up some of my old material - but I'm not 100% sure I still feel as strongly about him now.

In Group 2 ballot notes I wrote
Grich has been number two to eight on my personal list of worthy players who are not in the Hall of Fame (number two on a very bad day this Winter). He is usually about four and number one with 20th century debut. . . .
Santo has been number four to eight on my personal list of eligible players not in the Hall of Fame.

Chris Cobb #2
Astonishing Oversights (comparable to with Blyleven, Raines, Grich, and Santo)
Bill Dahlen 366
Paul Hines 310
Deacon White 301

That's the trio, in different order. I was sick in the chest with a bad back too, that day last month when Grich was number two.

When I had a clear-cut eight the other three were Barnes, Glasscock, and Magee. That was several years ago, some time after the elections of Davis and McPhee and some Sherry Magee provocation by SABR Deadball Era Cmte founder and chair Tom Simon.
That has changed. Among other things I wouldn't have considered Grant Johnson or any of the "Negro Leagues" class inducted in 2006, or Blyleven, McGwire or Raines.
And I would have dismissed Joe Start from *those* heights before participating here.

Here is the quartet Chris Cobb tentatively (purely by rating system) ranks #16-19 of 21
Lower-Tier but In
Hardy Richardson 241
Charley Jones 240

Questionable HoM selections.
Charlie Bennett 211
Harry Stovey 210

That happens to be the quartet re whom I wish we had International Association and independent pro statistics for the late 1870s(*). We know the National League employed a variable subset of the best players before 1882 when the AA helped pick up a higher share.

While at it I would try to bring back data on Jack Rowe (for myself, a catcher-shortstop in the tradition of Dickey Pearce and George Wright), Levi Meyerle, and Jim McCormick (for all of you).


*1874-75 for Jones. Glasscock was almost 22 at his mlb debut in 1879 but his playing statistics make me doubt he entered as a weakling at bat.

Paul, don't you mean quintet? Browning was playing for Louisville before 1882.

Paul,

A long time ago, about 8 months, you mentioned you had some International Association stats that you would be willing to mail me. Are they still available?

Does anyone else know where IA stats or information may be found? What old newspapers or even what teams were involved? Any books? For awhile I tried doing some google research, but I didn't get much.

Thanks,

Kelly

actually, it likes there is more information at the wikipedia page than there used to be. there are even sources listed. i'm going to do some hunting. but if anyone has any other advice, please post. thanks

Nope, still pretty much a dead end.

I now think that it would be useful - and easy - to do a cumulative "top 15" at the end.

If we knew we were doing that beforehand, we might have been able to have a "Group 4 and Group 5," roughly like this:

Group 4
Wes Ferrell (1964)
Joe Gordon (1976)
Heinie Groh (1938)
Stan Hack (1958)
Charlie Keller (1996)
Sherry Magee (1926)
Jimmy Sheckard (1930)


Group 5
Charlie Bennett (1921)
Pete Browning (2005)
Bob Caruthers (1930)
Cupid Childs (1988)
Bill Dahlen (1915)
Jack Glasscock (1904)
George Gore (1898)
Paul Hines (1898)
Charley Jones (2003)
Hardy Richardson (1905)
Joe Start (1912)
Harry Stovey (1916)
Ezra Sutton (1908)
Deacon White (1898)

Possibly tossing Dahlen and/or Bennett into Group 4...

This is much the most fun list. I checked past ballots; it will be Caruthers 1, Dahlen 2, as I had them 2 and 4 on my 1915 ballot (thankfully preserved in HOM archives.) Davis was 1 that year, Mickey Welch 3, because of the 300 wins. I would put Dahlen above Welch today, but not above the unique Caruthers. For those who aren't convinced or post-dated his election, take another look; he is the only player to have operated at a HOM level as both pitcher and outfielder. (123 OPS+ and 218-99 would get him in as a pitcher, 135 OPS+ is well above borderline for outfielder, albeit in excessively short career.) His combined value in 1885-87 is thus among the top 10 in history.

I don't know that I would call Hines an astonishing oversight. He was arguably the 2nd best hitter at his position in an 8 team league, but obviously not the 2nd or 3rd best hitter in the league as Speaker was in his time. From the subjective reports, it can be taken that he was a very good defensive centerfielder, but more than that is subject to reasonable disagreement.

I suppose that you could say the same about Dahlen. He hit about as well as Trammell, and by reputation was a superior fielder, but reputation is a fairly weak marker and the error bar for any defensive metric going back into the 19th century is huge.

White is probably the easiest to describe as an "astonishing oversight" because of his offensive performance and position, although I suspect that Dahlen was the best of the three.

Well in Dahlen's case the numbers and the reputation match, so that's pretty strong in my view.

Both Dahlen and Hines, as well as White, had exceptionally long careers for their era, as well as top-notch peaks. Dahlen's career was longer than Trammell's, even without adjusting for the shorter seasons he played in. For his career, Cal Ripken hit about as well as Trammell, too, but that doesn't make Trammell his equal as a player.

White is actually the most excusable oversight for the HoF, since his best years were in the National Association, and it is not clear to what extent Cooperstown acknowledges the NA. Without his NA peak, White looks much less impressive.

Responding to Joe's question re Bennett: certainly I give catcher bonuses: without a bonus, would Bennett rank ahead of Stovey? I supported both Bennett and Stovey when they were elected, and I'm not sure that it was incorrect to elect either of them, given the pool when they were chosen. That's why I put them in a group called "questionable picks" rather than "mistakes," (which is how I would categorize Pete Browning, really). Viewed from an "all-time" perspective, however, they look very weak. I haven't reviewed their numbers in about "a hundred years," but as I recall, Bennett and Stovey both received small subjective boosts for value that was not adequately captured by the comprehensive metrics. That subjective boost doesn't show up in my system's numbers, which are what I used to set the order on my preliminary ballot. In any case, I'm going to be taking a close look at the pre-1893 candidates again. They are much harder to gain certainty about than the later players.

I realize this is probably too late, and it's my own fault for not paying attention to the earlier discussion, but why is Joe Start in this group? Shouldn't he be in Group 4 with Pike and McVey? I realize he does have 10 years in MLB, but his case really relies on including his NA time.

This is the list where my system breaks down. I can rank the players in Howie's "group 4" but without WARP I got a whole lotta nothing on most of these guys. I might sit this one out.

his case really relies on including his NA time

I might go further and say that Joe Start's case really relies on including his pre-NA time. His NA years are mostly a trough in the middle of his career: his period of stardom was the 1860s.

I was being kind to Dahlen in saying that he hit about as well as Trammell.

As for the defensive metrics of the 19th century, I'd like to see the way they work. When Dahlen was at his defensive peak in the early 1890s, league on-base percentages were very, very high and league slugging percentages were very low. The Cubs had a below average pitching staff. So, you had lots and lots of runners on first. Notwithstanding this, Dahlen's DP totals are totally ordinary. He's making lots of plays, but he doesn't have the spike in DPs that you might expect. It's possible that his second baseman doesn't have much of a pivot. Or it's possible that the routine flip to second with not much of an effort to turn two was part and parcel of the game then.

It is interesting to compare the Hall of Fame's treatment of Jimmy Collins and Bill Dahlen and think about the career path of Honus Wagner. It seems conceivable to me that in the baseball of the 1890s, shortstop and third base were much closer on the defensive spectrum than they were subsequently. A ground ball to third base with a runner on first was much more likely to result in an infield single in the conditions of the time, and the importance of the smooth catch and quick flip on the DP were much less significant for the shortstop because the DP was such a rare event.

I do think that Dahlen was an easy Hall of Meriter, but I don't have the same confidence about where he stands in relation to the greats of his position, as I do for Grich, say.

I was being kind to Dahlen in saying that he hit about as well as Trammell.

Trammell at his peak was certainly a better hitter than Dahlen, but Dahlen was better in their "off" years, so that their career rates are very similar.

Am I correct in assuming that Pearce isn't listed because he's pre-1871? Is he then going to be on his own list or with other non-HOMers of that very early era or....nothing?

It does seem a bit weird to be ranking Start and not Pearce. Or White/Sutton and not Barnes, but those are how the groups were defined. Group 4 consists of the ineligible people (less than 10 years of ML service), Pearce, Barnes, Grant Johnson, etc.

Prelim

Despite my 'outlier' status in the last group, we are mostly taking about an all-star batting order or the like, since I agreed that all but one of those guys should be in the Hall. In this group I stray a little but further from average (though the top guys are likely to be close to the norm), because there are seven guys on this list that I don't have in my PHoM.

IN

1. Bill Dahlen
2. Deacon White
3. Paul Hines
4. Stan Hack
5. Bob Caruthers
6. Jimmy Sheckard
7. Joe Start
8. Ezra Sutton
9. Sherry Magee
10. George Gore
11. Jack Glasscock
12. Cupid Childs
13. Heinie Groh
14. Harry Stovey

Not in but still sitting in my backlog

15. Pete Browning
16. Charlie Bennett

Not in and not getting in

17. Charley Jones
18. Joe Gordon
19. Charlie Keller
20. Hardy Richardson
21. Wes Ferrell

19. Devin McCullen in Flame Broiled...TO HELL! Posted: February 11, 2008 at 10:57 AM (#2688089)
I realize this is probably too late, and it's my own fault for not paying attention to the earlier discussion, but why is Joe Start in this group? Shouldn't he be in Group 4 with Pike and McVey? I realize he does have 10 years in MLB, but his case really relies on including his NA time.

Howie #14 and Devin,
I know from reading the discussion here that the rationale for sprawling Group 3 is to match the categories defined by recent changes in the Veterans Committees(plural). One committee will cover everyone with mlb debut before 1943, so group 3 matches the jurisdiction of that new VC. It speaks to the members of that committee, and anyone who prepares reference material prior to their meeting, and of course any fans who follow the particular VC elections.
Although the case for White and Sutton, at least, depends in part on considering their pre-NL play, the idea must be that that committee (and its assistants and fans) may consider their entire careers. Ten seasons as active mlb player defines players who are eligible not the basis for judging them.

The messy Group 4, on the other hand, is everyone in the Hall of Merit who cannot be considered for Cooperstown induction until there is further action by the NBHOFM Directors or the ED. Or by MLB reinstating Ete Prose.

--
10. Devin McCullen in Flame Broiled...TO HELL! Posted: February 11, 2008 at 01:16 AM (#2687973)
Paul, don't you mean quintet? Browning was playing for Louisville before 1882.

Yes and no.
Chris Cobb didn't list Browning consecutively with the others so I missed him.

Probably I should have said quartet --Glasscock (not prejudging his age 20 by his age 21 OPS+), Bennett, Richardson, Stovey-- because I was thinking of the late 1870s.

Jones would be 1874-75 and 1881-82, Browning ??-1881. I suppose they played mainly in the region of the two cities KY-IN-OH but from Al Spink's history I know that his club in St Louis, or two St Louis clubs, did travel. Some of the St Louis, Louisville, Cincinnati, and Pittsburgh(Allegheny) clubs that founded the AA were established before the end of the 1881 baseball season but there was semipro baseball in all four cities, I think.

Richard Hershberger has worked on baseball in Philadelphia between 1876 and 1882/83/84, the end and start dates for its four franchises in the major leagues. As I recall there were some good clubs but none ever joined the International/National Association, which was northern with late entries from Baltimore and Washington.
As far as I know, we have no HOM members who played in Philadelphia.

Back to the question: Browning was young like Glasscock, but he was a dominating batter in his mlb debut season. So yes it would be good to know more.
His article in Nineteenth Century Stars provides 1881 affiliation Eclipse (the club that joined AA 1882, I think), fielding position 3b, games 40, hits 57, batting average .333, fielding average .885. Earlier he played on the amateur Reccius club.

I see that I might have added George Gore too. Nineteenth Century Stars provides records such as I have just passed on for Browning 1881. What distinguishes 1876-1880 is not so much extending those meagre lines but the prospect of providing context. National League clubs played a lot more games against independents in 1876 than in 1881. The IA/NA provides alone provides some context for 1877-80.

And I see that Stovey was from Philadelphia, played for the newly-independent Athletics in 1877. Pitcher is the only listed position.

Robert L. Tiemann (Bob T) compiled the data in Nineteenth Century Stars.

Oops, I meant Start's pre-NA record. That's what I get for typing quickly.

I do think that if you're not going to have Dahlen #1, you should probably explain yourself here. (For karl, it's entirely consistent to have Caruthers #1, so that's not a big deal.)

11. mulder & scully Posted: February 11, 2008 at 02:57 AM (#2687989)
Paul,

A long time ago, about 8 months, you mentioned you had some International Association stats that you would be willing to mail me. Are they still available?

Does anyone else know where IA stats or information may be found? What old newspapers or even what teams were involved? Any books? For awhile I tried doing some google research, but I didn't get much.

Thanks,

Kelly

Yes. Please post your mailing address or send to pwendt [[at]] fas.harvard.edu


12. actually, it likes there is more information at the wikipedia page than there used to be. there are even sources listed. i'm going to do some hunting. but if anyone has any other advice, please post. thanks
13. Nope, still pretty much a dead end.

(at wikipedia) Bryce's Base Ball Guide, published in London Ontario. That would be interesting. Published prior to the 1876 and 1877 seasons, I presume, so *useful in this forum* only for any coverage of pre-IA independent baseball.

7. OCF Posted: February 10, 2008 at 11:30 PM (#2687937)
This will be a much, much more difficult vote than the previous two. OK, Dahlen on top is obvious, but beyond that ...

Gore, Hines, and White were elected before I joined the HoM electorate - in essence, I never considered them. Now I'll have to investigate their cases.

White, Hines, Gore finished 1-2-3 in the first election covering players who retired before 1893, the modern pitching distance.

Chris Cobb #12, elaborating himself:
White is actually the most excusable oversight for the HoF, since his best years were in the National Association, and it is not clear to what extent Cooperstown acknowledges the NA. Without his NA peak, White looks much less impressive.

Yes, much less impressive as a batter and he may also be mistaken for a thirdbaseman.
White is the most excusable oversight by Veterans Commmittees and special elections in our lifetimes. On the other hand, the blue ribbon advisory committee in 1993(?) urged election of Dahlen among others; committee chair and SABR founder Bob Davids prepared portfolios for them. (Maybe he prepared more.)
Total Baseball (Pete Palmer's linear weights as "TPR") floodlit Dahlen and Davis among players not in the Hall of Fame, ranking them both around all-time #50.

White was alive and well during the baseball centennial and NBHOFM planning, 1936-39 or so. George Wright passed away in 1937 and a committee soon elected him as one of five "builders of baseball". The grand opening came and went without honoring Deacon White. Connie Mack and Cy Young, 15 and 20 years younger, were the grand old men at the main event.

(George Wright may belong in halls of fame as an American builder of lawn tennis, ice hockey, and golf but not baseball.)

Sorry if this is a double post, BTF just kicked me out.

Paul,

Thanks for the information. I followed some links about the Bryce Book, but I don't think the Canadian Museum has scanned the pages like the Library of Congress did with many of the Spalding Guides.

I emailed you my mailing address.

Thanks,

Kelly

I haven't reviewed their numbers in about "a hundred years," but as I recall, Bennett and Stovey both received small subjective boosts for value that was not adequately captured by the comprehensive metrics.

IIRC the old debates, Stovey received a significant "subjective" boost for his baserunning abilities.

Bennett received a significant boost for his relative endurance playing a difficult position. Gloveless and armorless catching was extremely debilitating and when the schedules lengthened during the 1880's, catching was treated similarly to pitching, with every team playing two or often three catchers, so "games caught" does not scale well relative to the other positions. There is much more on this on the old Catchers thread. There's lots of good material before and after, but my work on normalizing catcher workloads is around post 52 and subsequent. I don't want to repeat the old arguments verbatim, but with both White and Bennett on this ballot, I think this is worth reviewing (and I'm making it easy to do by providing the link ;-).

Bennett: #12 on my 1904 ballot, behind Richardson and Start but ahead of McVey. Reached #2 by 1914 as those ahead of him got elected. Was #1 on my ballot in 1921 when elected, ahead of Sheckard.

Browning: #13 in 1904, ahead of Caruthers. Jumped to #7 in 1905 and #5 in 1906, ahead of Start and Bennett. Faded downward, back behind Start and Bennett. Was #19 in 1917, #24 in 1921, then I stopped mentioning him. Never reappeared on my ballot.

Caruthers: #15 in 1904, #14 in 1905. Disappeared for a while as I named only 15. "Honorable mention" in 1915-1917. I started naming a top 25 or to 30 around 1921; Caruthers was consistently in the low to mid 20's (behind Griffith and Waddell, among others.) Was #21 when elected in 1930, ahead of Pearce and behind Rube Foster.

Childs #13 in 1907, 1908, #15 in 1910, HM in 1913-1916. I apparently reevaluated him in 1939 and he suddenly jumped to 7th, then slid donwards, to #16 in 1950, #20 in 1957, #30 in 1983. Was #26 when elected in 1988. Behind most of the people in Group 2.

Dahlen #1 on my 1915 ballot. I had him #1 and George Davis #2, which is a strong endorsement of Dahlen.

Ferrell #11 on my 1944 ballot; stayed between #11 and #7. Was #8 when elected in 1964. Behind Hack and Beckwith, ahead of Suttles and ahead of Gordon.

Glasscock: #1 on my 1904 ballot, putting him ahead of Stovey, Sutton, Richardson and Start.

Gordon: #19 in 1956, stayed between #19 and #23. Was #21 when elected in 1975. I did have him ahead of Doerr and Sisler.

Gore: Elected before I joined.

Groh: Started at #6 in 1933; pushed down by the monster class of '34 but came back up to #4 when elected in 1938. Ahead of Sisler, Waddell, Roush, Beckley.

Hack: Started at #5 in 1953, rose to #1 when elected in 1958. Ahead of Medwick, Rixey, Freehan.

Hines: Elected before I joined.

Jones: I never supported him, not even in my top 30. Behind nearly everyone else here.

Keller: I never supported him.

Magee: #2 behind Grant Johnson on my 1925 ballot, #1 when elected - ahead of Sheckard.

Richardson: #6 on my 1904 and 1905 ballots. Behind Stovey and Sutton, ahead of Start.

Sheckard: #4 on my 1919 ballot, just behind Bennett but ahead of McGinnity and Brown. Never lower than #4, #1 when elected in 1930.

Start: #7 on my 1904 ballot; slipped to #9 and then climbed steadily (by attrition of those ahead of him) to #4 when elected in 1912. Behind Stovey, ahead of Bennett.

Stovey: My first cause (with my argument that his ability to score runs gives him value beyond his batting statistics). Never lower than #3 on my ballot from 1904 through 1916; #1 when elected. Always ahead of the backlog.

Sutton: #4 on my 1904 ballot (behind Glasscock, Radbourn, and Stovey); #2 (behind only Stovey) in 1908 when elected.

White: Elected before I joined.

That does not immediately give me a ranking, and my changing opinions, notably of Browning and Childs, complicate the situation considerably.

Did anybody answer the question "Where's Beckley?" yet?

He's in the HoF.

This is the most fun list, but frankly, if the point is the achieve clarity, I think the separate lists for 19C and 20C would be better at this stage, then integrate later.

This is the most fun list, but frankly, if the point is the achieve clarity, I think the separate lists for 19C and 20C would be better at this stage, then integrate later.

I echo this sentiment. Dahlen and Childs form an 1890's sub-group that I could see adding to the 20C list, splitting the group based on the 60'6" distance, before and after. (Endorsing both of Howie's suggestions.)

prelim list from me, will add coments later on ballot thread:

Bill Dahlen (1915)
Deacon White (1898) - start a petition drive for the first two.
Paul Hines (1898)
Jack Glasscock (1904)
Hardy Richardson (1905)
Joe Start (1912)
Ezra Sutton (1908)
George Gore (1898)
Harry Stovey (1916)
Wes Ferrell (1964)
Charlie Bennett (1921)
Jimmy Sheckard (1930)
Heinie Groh (1938)
Stan Hack (1958)
Sherry Magee (1926)
Bob Caruthers (1930)
Joe Gordon (1976) - still a clear HoMer at this point on the ballot
Cupid Childs (1988)
Charlie Keller (1996) - HoM borderline
Charley Jones (2003)
Pete Browning (2005)

I see 12 of the 21 debuted either on the original 1898 ballot or in 1899, followed by Childs and then Dahlen and then Sheckard and then Magee and then Groh and then 4 WW II-era guys.

That would be tough to split, as Childs-Dahlen-Sheckard seemingly are best off with the first group. Magee and Groh are sort of stuck in the middle, and there aren't enough WW II boys.

Maybe the first course was wisest after all; the bulk of these guys are contemporaries to some degree, so not as tough to order. Then it's just sprinkling in a half-dozen more into the soup.

Yes.

I see this election as an opportunity mostly to think through the 1870s and 1880s again, in light of more information and fuller knowledge about what comes after, with a handful of 20th-century players to slot in as well.

This is new to me so perhaps I am misunderstanding the criteria or purpose. But I am surprised not to find Indian Bob Johnson on your list. Seems to me he matches up well with many of those apparently enshrined.

Their debuts are fairly regular from Caruthers in 1884 to Gordon and Keller in 1938/39, with a long gap only between Groh and Ferrell. Pitchers Caruthers and Ferrell are a challenge anyway.

regular "position players" - none, 1925-1931
regulars in their prime seasons - none, 1922-1934 - that's a career, Pie Traynor as a regular

Bob Johnson does make for argument fodder, but the key issue is this: we have not elected Johnson to the Hall of Merit. (He finished 10th on our 2008 ballot.) To be eligible in this group, the candidate must have been elected to the HoM.

It's also worth noting that Bob Johnson finished 10th in the HoM's 2008 election, so he is certainly viewed as a player who is close in value to those already enshrined, and he stands a good chance of being elected in the next several years.

Arguments on Johnson's behalf are relevant to the 2009 ballot discussion thread, however, not to this thread, where we are ranking players elected to the Hall of Merit but not to the Hall of Fame, in order to establish most clearly where the HoF ought to start if it were to seriously attempt to address its most serious omissions.

18. Chris Cobb Posted: February 11, 2008 at 10:55 AM (#2688088)
Both Dahlen and Hines, as well as White, had exceptionally long careers for their era, as well as top-notch peaks. Dahlen's career was longer than Trammell's, even without adjusting for the shorter seasons he played in. For his career, Cal Ripken hit about as well as Trammell, too, but that doesn't make Trammell his equal as a player.

Exceptionally long for any era. Adjusted for the number of team games played in every season (full seasons equivalent) Hines, White and Dahlen rank #13-18-34 by all games and #7-11-20 by fielding games (no dh, ph, pr), estimate.

Rose, Henderson, and Bonds are in both top twenties. The only other listing for anyone not in the Hall of Fame is Rusty Staub #15, all games.

There are several shortstops and part-shortstops high on the list. By fielding games, Dahlen #20 trails Ripken, Wagner, Maranville, and Yount. Davis, Aparicio, and Smith are also among the 33 players with 16.00 or more full seasons in the field (that's 8 of 33). Wallace and Ward are in the top 40; Concepcion, Banks, and Appling in the top 50; Vizquel[thru 2006] and Corcoran among the 56 players with 15.00 full seasons (that's 15 of 56).

It looks like Vizquel thru 2007 makes 9 of 34 with 16.00 FSE in the field.

Glasscock is #58, . . . Trammell #96 (Whitaker #65)

Trammell and Whitaker have some non-fielding games but not enough; they are #107 and #78 counting all games. (Again, this is normalized to full seasons equivalent.)

See the latest 2008 Election Results

Stan Hack and Bob Johnson show that Cooperstown has picked the 1920s a little cleaner than the 1930s.
One might add pitchers Wes Ferrell and Bucky Walters (high in the queue).
There are several very good pitchers from the 1920s who are not in the Hall of Fame but in the end --so far plus high in the queue-- we rate them all merely Very Good.

I have no idea how one could rationally adjust for "durability" of the stars of the 1870s. To compare the game of 1927 with 1967 or 2007 involves a relatively modest adjustment for season length, but it's a whole different story for 1877. Huge season length adjustments, as well as some recognition of the differing travel demands, health care availability and level of competitiveness of the games, makes the whole process pretty arbitrary to me.

You can say that Dahlen, White and Hines were exceptionally durable for their time, with a great degree of confidence. It is also pretty clear that Dahlen was durable by any standard, basically playing full-time (75% of the games, totalling at least 115 per year) in every season from age 21-38 except for 1 year, but trying to compare Hines with a reasonably durable centerfielder of the 1970s or 1980s seems to me to be well nigh impossible.

The only method that makes any sense to me is to compare Hines (e.g.) to his contemporaries--and let's just say for the sake of example that he's 1.5X more "durable" than other "all-star caliber" CF (whatever that means). He would rank then about the same as a CF from the 1970s or 1980s who is 1.5X more durable than his peers. Obviously, you have to operationalize all the terms. In the meantime, I guess I would just eyeball 'em and say, okay, he's Max Carey or Kenny Lofton on career length.

And then you'd want to compare the quality of those careers as your second step.

That's how I'd approach it. It's complicated to execute, but pretty simple conceptually. Not "well nigh impossible" at all. Just the kind of thing that needed to be done and so we HoM voters have all pretty much made peace with it.

There is no clear interpretation of full seasons as any particular combination of durability or longevity or other factors but it is way far away from a focus on durability.

46. sunnyday2 Posted: February 12, 2008 at 03:22 PM (#2689145)
The only method that makes any sense to me is to compare Hines (e.g.) to his contemporaries--and let's just say for the sake of example that he's 1.5X more "durable" than other "all-star caliber" CF (whatever that means). He would rank then about the same as a CF from the 1970s or 1980s who is 1.5X more durable than his peers.

I'm not sure this would work if the major leagues were ten times larger than in fact.
--except as a pure durability measure, where one might be satisfied to compare Hines with different all-stars every season. That may pertain to #45 but not to the rest of this thread.

Can anyone update me on whether there were any significant defensive spectrum shifts from 1871 to 1892, leaving pitchers aside?

Can anyone update me on whether there were any significant defensive spectrum shifts from 1871 to 1892, leaving pitchers aside?

The period is too short to study from the point-of-view of average OPS+ at a position. Naturally, this didn't prevent us from speculating, particularly about the effects of rule-changes.

Fair-foul rule. IIRC, this was changed for 1877 to very near it's current form. In earlier games, a ball was deemed fair or foul based on where it landed; in later games a ball that landed fair could become foul if that happened between the plate and 1st/3rd base (same as today). Some ball-control batters (e.g. Barnes) could place the ball down fair so that it quickly went foul; fielding these quickly placed extra emphasis on mobility at 1st, 3rd, and Catcher.

(Lack of) Gloves: At catcher particularly, and 1b somewhat, playing gloveless did a number on the hands. This naturally affected the hitting of these players. During Deacon White's prime in the 1870's, his best batting season by far was when he didn't play catcher.

Substitution rules: About 1890, substitution reached its modern form. Before then, substitutions needed the permission of the umpire and the opponents. As manager, if you thought that you might require a substitute that day, it might be wise to include the sub in the starting lineup, usually in RF, so a quick position-swap could be done when necessary. Harry Wright (and others) liked to put a 2nd pitcher out there, just in case, while others put a 2nd catcher. Longer schedules made individual games less important and tempted some managers to put a hitter out there, like Thompson, Tiernan, or Shaffer. King Kelly allowed Anson to have both, a hitter and a backup catcher/utility IF. Caruthers & Foutz gave Comiskey a similar flexibility as pitchers who could also hit.

48. David Concepcion de la Desviacion Estandar (Dan R) Posted: February 13, 2008 at 02:54 PM (#2690338)
Can anyone update me on whether there were any significant defensive spectrum shifts from 1871 to 1892, leaving pitchers aside?

For the catcher there was a huge decrease in frequency of fielding batted balls for the catcher. Handling the pitched ball and controlling the baserunning game became more important, perhaps absolutely as well as relatively.

--
The rule changes governing foul hits --pops, flies, bounds, tips, bunts-- made mobility and sure hands in the catcher less and less important.

For bunts the fair-foul rule would have been directly related to fielding skills of catcher (and 1B, 3B, at least). The foul strike rules --first strikes one and two, then strike three-- would have been related to fielding skills only indirectly, through the disincentive to batters and presumable decrease in number of bunts.

For the other foul hits the effect of rules changes is largely measurable in aggregate, I suppose, by the decrease in number of catcher putouts that are not strikeouts. Successively more foul hits became dead balls rather than opportunities to put the batter out.

On the Cincinnati Red Stockings tour I suppose there were games where catcher Doug Allison put out ten men and the other catcher put out one.

--
The changes in pitching distance, one early and one defining the end of your time period, would have changed the nature of the pitcher's fielding responsibilities. You don't care about the pitcher's fielding, DanR, but presumably the 1892 pitcher handled some bunts and nubs that only the catcher or no one must handle in 1893. The 50' distance was 1881-92, up from 45'.

A foul hit caught on one bounce put the batter out thru 1882 except 1879. The NL required the fielder to catch a foul hit in the air off the bat beginning 1883 (plus 1879); the AA on 7 Jun 1885.

The foul tip differs from the fly or pop because the batter can be put out only with two strikes. (And when the catcher does get a two-strike foul tip in the air, it is scored as a strikeout rather than a flyout.) I don't find the foul tip rule in TB3 just now. I think it was after 1893.

Here is some data on batting hands, relevant to the fielding spectrum.

AtBats by Batting Hand
1900 NL
01204 missing (about 3%)
03476
16069
18383 (more right-hand batters lost NL jobs in the contraction)

1892 NL
02769 missing (about 4%)
04862 B
17722 L
38523 R

1881 NL
01239 missing (about 5%)
00738 B
06524 L
15876 R

1876 NL
02496 missing (about 25%)
00310 B
02491 L
14824 R

1871 NA
5271 missing (almost 50%)
0158 B
1340 L
4053 R

Beware, that's a lot of missing data.
This is from lahman5.4, a little out of date.

The share of atbats with batting hand missing drops from 5% to 3% in 1900.
After 1910 it is trivial even in the Federal League and in 1918.

Today the right-left-switch proportions of atbats are very close to 4-2-1.
The right-left ratio is about 2:1 since 1893.

For pitching hand there is no missing data in 1877 (R:L = 9369:354, about 4% left) and in 1900 (R:L = 23728:6014, about 20% left), almost none after 1910.


DECADE AGGREGATE DATA

Batting Hand 1870s
34462
2855 B
24770 L
98912 R

Batting Hand 1880s
66680
20989 B
130601 L
401336 R

Batting Hand 1890s
47934
55655 B
193675 L
365520 R

Pitching Hand 1870s
6746
3625 L
99388 R

Pitching Hand 1880s
38764
67415 L
354757 R

Pitching Hand 1890s
23930
87032 L
382798 R

Here i have supposed random matchups and no tactic except that switch-batters always bat left against rhp and bat right against lhp.

1890s, thousands of atbats, share estimates

_______ Pitching
Batting _R_ _L_ _?_ Batsum
_Right 283 _74 _17.7 : 375
_Left 193 _34 __9.4 : 237
_Unknown 37 __8.5 _5.0 : _50

Pitchsum 514 117 _32 : 663,000 atbats

OK, this is hard. Only Ferrell and Sheckard are not PHoM.

Prelim

1. Bill Dahlen--okay, feelin' pretty good about this. #2 on my ballot when elected in 1915.
2. Deacon White--#1 in 1898.

(gap)

3. Paul Hines--#2 in 1898.
4. Jack Glasscock--#2 in 1904.
5. Bob Caruthers--#2 in 1930.
6. Heinie Groh--#4 in 1938.
7. Joe Gordon--#9 in 1976 against tougher competition than any of the guys above him and most below.
8. Charley Keller--#2 in 1996. In hindsight, Gordon was more valuable.
9. George Gore--#7 in 1898.
10. Ezra Sutton--#3 in 1908.

11. Pete Browning--#7 in 2005.
12. Hardy Richardson--#7 in 1905. Wow, Pistol Pete only took 100 years even to follow Hardy into the HoM!
13. Charlie Bennett--#2 in 1921.
14. Joe Start--#1 in 1912. This is probably too low.
15. Sherry Magee--#13 in 1926.
16. Charley Jones--not rated in 2003 but peaked earlier at #5.
17. Cupid Childs--unrated in 1988 but peaked earlier at #7.
18. Stan Hack--#14 in 1958.
19. Harry Stovey--#12 in 1916.
20. Wes Ferrell--unrated in 1964, not PHoM.

21. Jimmy Sheckard--unrated in 1930, peaked earlier at #8, not PHoM. Only Ferrell and Sheckard ar not PHoM so they're at the end of the list for now, though that may be too low compared to some of the guys who faced weaker competition. Well, but for Sheckard at least, who would that be? Stovey and Childs are probably the only guys he could pass up.

Years that GROUP IIIs were on HOM ballots (* if elected that year)
in order of finish by year within this ballot group

1898 - DWhite*, Hines*, GGore*//ESutton, HRichardson, Start, CJones
1899 - HStovey, Richardson, Sutton, Start, Caruthers, Bennett, Browning, CJones
1900 - Richardson, Sutton, Stovey, Start, Bennett, Caruthers, Browning, CJones
1901 - Glasscock, Richardson, Sutton, Stovey, Start, Bennett, Caruthers, Browning, CJones
1902 - Glasscock, Richardson, Sutton, Start, Stovey, Bennett, Caruthers, Browning, CJones
1903 - Glasscock, Richardson, Start, Sutton, Stovey, Bennett, Caruthers, Browning, CJones
1904 - Glasscock*//Richardson, Start, Sutton, Stovey, Bennett, Browning, Caruthers, CJones
1905 - Richardson*//Sutton, Start, Stovey, Bennett, Browning, Caruthers, CJones
1906 - Sutton, Start, Stovey, Bennett, Browning, Caruthers, CJones
1907 - Sutton, Start, Stovey, Bennett, Childs, Browning, Caruthers, CJones
1908 - Sutton*//Start, Stovey, Bennett, Childs, Browning, Caruthers, CJones
1909 - Start, Bennett, Stovey, Childs, Browning, Caruthers, CJones
1910 - Start, Bennett, Stovey, Childs, Browning, Caruthers, CJones
1911 - Start, Stovey, Bennett, Childs, Browning, Caruthers, CJones
1912 - Start*//Bennett, Stovey, Childs, Browning, Caruthers, CJones
1913 - Stovey, Bennett, Childs, Browning, Caruthers, CJones
1914 - Stovey, Bennett, Caruthers, Childs, Browning, CJones
1915 - Dahlen*//Stovey, Bennett, Caruthers, Browning, Childs, CJones
1916 - Stovey*//Bennett, Caruthers, Browning, Childs, CJones
1917 - Bennett, Caruthers, Browning, Childs, CJones
1918 - Bennett, Caruthers, Browning, Childs, CJones
1919 - Bennett, Sheckard, Caruthers, Browning, Childs, CJones
1920 - Bennett, Sheckard, Caruthers, Browning, Childs, CJones
1921 - Bennett*//Sheckard, Caruthers, Browning, Childs, CJones
1922 - Sheckard, Caruthers, Browning, Childs, CJones
1923 - Sheckard, Caruthers, Browning, Childs, CJones
1924 - Sheckard, Caruthers, Browning, Childs, CJones
1925 - Magee, Sheckard, Caruthers, Browning, Childs, CJones
1926 - Magee*//Sheckard, Caruthers, Childs, Browning, CJones
1927 - Sheckard, Caruthers, Browning, Childs, CJones
1928 - Sheckard, Caruthers, Childs, Browning, CJones
1929 - Sheckard, Caruthers, Childs, Browning, CJones
1930 - Sheckard*, Caruthers*//Childs, Browning, CJones
1931 - Childs, Browning, CJones
1932 - Childs, Browning, CJones
1933 - Groh, Childs, Browning, CJones
1934 - Groh, Browning, Childs, CJones
1935 - Groh, Browning, Childs, CJones
1936 - Groh, Browning, Childs, CJones
1937 - Groh, Browning, Childs, CJones
1938 - Groh*//Childs, Browning, CJones
1939 - Childs, Browning, CJones
1940 - Browning, Childs, CJones
1941 - Childs, Browning, CJones
1942 - Childs, Browning, CJones
1943 - Childs, Browning, CJones
1944 - Childs, Browning, CJones
1945 - WFerrell, Childs, Browning, CJones
1946 - WFerrell, Childs, Browning, CJones
1947 - WFerrell, Browning, Childs, CJones
1948 - WFerrell, Childs, Browning, CJones
1949 - WFerrell, Childs, CJones, Browning
1950 - WFerrell, Childs, Browning, CJones
1951 - WFerrell, Childs, Browning, CJones
1952 - WFerrell, Childs, Browning, CJones
1953 - Hack, WFerrell, Browning, Childs, CJones
1954 - Hack, WFerrell, Browning, Childs, CJones
1955 - Hack, WFerrell, Browning, Childs, CJones
1956 - Hack, WFerrell, Browning, Childs, Gordon, CJones
1957 - Hack, WFerrell, Childs, Browning, CJones, Gordon, Keller
1958 - Hack*//WFerrell, Browning, Childs, CJones, Gordon, Keller
1959 - WFerrell, Browning, Childs, Gordon, CJones, Keller
1960 - WFerrell, Browning, Childs, Gordon, CJones, Keller
1961 - WFerrell, Childs, Browning, Gordon, CJones, Keller
1962 - WFerrell, Childs, Browning, Gordon, CJones, Keller
1963 - WFerrell, Childs, Browning, Gordon, CJones, Keller
1964 - WFerrell*//Childs, Browning, CJones, Gordon, Keller
1965 - Childs, Browning, Gordon, CJones, Keller
1966 - Childs, Browning, Gordon, CJones, Keller
1967 - Childs, Browning, CJones, Gordon, Keller
1968 - Childs, Browning, Gordon, CJones, Keller
1969 - Gordon, Childs, Browning, CJones, Keller
1970 - Gordon, Browning, Childs, CJones, Keller
1971 - Gordon, Browning, Childs, CJones, Keller
1972 - Gordon, Childs, Browning, CJones, Keller
1973 - Gordon, Browning, Childs, CJones, Keller
1974 - Gordon, Browning, Childs, CJones, Keller
1975 - Gordon, Browning, Childs, CJones, Keller
1976 - Gordon*//Browning, Childs, CJones, Keller
1977 - Childs, Browning, CJones, Keller
1978 - Childs, CJones, Browning, Keller
1979 - Childs, CJones, Browning, Keller
1980 - CJones, Childs, Keller, Browning
1981 - Childs, CJones, Keller, Browning
1982 - Childs, CJones, Keller, Browning
1983 - Childs, CJones, Keller, Browning
1984 - Childs, CJones, Browning, Keller
1985 - Childs, Keller, CJones, Browning
1986 - Childs, Keller, Browning, CJones
1987 - Childs, Keller, Browning, CJones
1988 - Childs*//CJones, Browning, Keller
1989 - Browning, CJones, Keller
1990 - Browning, Keller, CJones
1991 - Keller, Browning, CJones
1992 - Browning, Keller, CJones
1993 - Keller, Browning, CJones
1994 - Keller, Browning, CJones
1995 - Keller, Browning, CJones
1996 - Keller*//Browning, CJones
1997 - Browning, CJones
1998 - Browning, CJones
1999 - Browning, CJones
2000 - Browning, CJones
2001 - Browning, CJones
2002 - Browning, CJones
2003 - CJones*//Browning
2004 - Browning
2005 - Browning*

prelim

1 Dahlen
2 White
3 Hines
4 Gore
5 Start
6 Glasscock
7 Richardson
8 Sutton
9 Magee
10 Stovey
11 Bennett
12 Groh
13 Childs
14 Caruthers
15 Browning
16 Sheckard
17 Hack
18 Ferrell
19 Gordon
20 Keller
21 CJones

Seems a little rough on the "modern guys," I'll have to re-look that.....

Been working on the 1870s and 1880s guys, and a few changes are in the works from the prelim.

1) Sutton is dropping and Glasscock is rising. It looks like they were pretty similar offensively, with Sutton a bit better. On defensive, they are far apart: Glasscock was an excellent defensive shortstop (WARP sees him as, in his context, in the same league as Ozzie Smith and Bill Dahlen), while Sutton was, to all appearances, an average defensive third baseman. Sutton's career was a bit longer, but not enough to offset Glasscock's defensive advantage. I now have Glasscock as #4 in the pre-1893 group after White/Hines (still sorting them out) and Start. Sutton is next, but he could slip down below . . .

2) Gore and Richardson (George and Hardy, not Al and Bill), who make an interesting pair. Their careers started and ended at the same time, and they played in about the same number of games. Both were excellent offensive players and outstanding defensive players. Gore is in centerfield, and Richardson is at second base (mostly). Where do these two positions fall on the defensive spectrum during the 1880s? Is Gore's 7 point advantage in OPS+ (.308 vs. .298 EQA) great enough to outweigh Richardson's superior defensive value -- if it was superior defensive value? WARP1 has Richardson at 111 FRAA for his career, and Gore at 62, so Richardson was the superior glove at his position. But how do the positions weigh against each other?

I'd be very interested to hear what folks have to say about this question.

full seasons equivalent fielding games, Hardy Richardson
5.19 2B
4.58 OF (L-C-R 2.87 1.63 0.07)
2.09 3B
0.16 SS
0.11 1B
0.07 C
0.04 P
That is at least five games at every position (5 at pitcher)

Bill James gives him a grade at second (A), no grade in the outfield.

When did Richardson play each of his positions?
2B, 1882-84, half of 85 86 87, 1888-89

fielding games at position as percent of team games played
yr 79 80 81 82 83 84 85 86 87 88 89 90 91 92
P_ __ __ __ __ __ __ _1 _3 __ __ __ __ __ __
C_ _1 _6 __ __ __ __ __ __ __ __ __ __ __ __
1B __ __ __ __ __ _3 __ __ __ __ __ _1 _2 _6

2B __ __ _6 99 94 62 45 33 50 43 65 __ __ 22
SS __ __ _1 __ __ __ _1 _2 __ __ __ _5 _3 _4
3B 99 95 _1 __ __ _4 __ _1 __ __ __ __ _6 _1

OF __ __ 95 __ __ 21 43 63 46 __ 35 95 43 16

LF __ __ __ __ __ __ _1 63 45 __ 35 95 41 _6
CF __ __ 95 __ __ 21 42 __ _1 __ __ __ _1 _4
RF __ __ __ __ __ __ __ __ __ __ __ __ _1 _6

Richardson played one position essentially every day in six seasons, 1879-1883 and 1890 (PL). In eight other season he played more than half of team games at one position three times.

Total 12.02 seasons in the field, 14 years

Detroit purchased Fred Dunlap from St Louis for $4700, mid-1886. For Detroit Dunlap played 51 games in 1886 and, because he was hurt, 65 games in 1887. Together Richardson and Dunlap covered second base for every game in 1887;

second base, Detroit 1886-87, fielding games as percent of team games
86-87
40 50 Dunlap
33 50 Richardson
30 -- Crane

I have supposed that Gore was a superior fielder.
That +112 for Richardson is a surprise to me.

But for me Richardson is the greatest cipher among the 19th century HOMers. I don't recall reading or hearing anything about him, except that he was one of the Buffalo/Detroit Big Four. My supposition about his defense may be sheer prejudice.

--
According to Nineteenth Century Stars, biog by Joe OVerfield, Richardson played semipro in Philadelphia, 1876. He played part of 1876 and all of 1877 for Binghamton, 1878 for Utica (International Association '77-78); in 1878 "winning the New York Clipper award as the best fielding centerfielder."

I have read all of 19c Stars at least twice but that is news to me again.

Re Richardson's fielding:

I'll have to dig deeper to see if WS agrees, but WARP's assessment of Richardson defensively suggests that the spectrum was 3B-2B-CF. Richardson was slightly below average at third base in 1879-80, the only position at which he did not excel. He was great in CF in 1881, then excellent to very good at 2B 1882-84.

These shifts may have had to do with team needs only, but it looks like Richardson was not good enough defensively to excel at 3B, so he was moved off when a better opportunity came along. He was so good in center that he was then moved up the spectrum to 2B, where he remained, having found his optimal level, until he was moved off of second in favor of Dunlap in mid-1886 (his fielding at second looks by WARP1 to have declined to just a bit above average by 1886). He covered second when Dunlap was injured in 1887, and was back at second full time in 1888 (Dunlap having moved on), until he was himself injured.

In Boston, in 1889, it looks like he was the 2B solution until Tom Brown was injured, at which point there was a scramble involving several players that landed Richardson in left field, where he remained in the declining years of his career.

It looks to me like Richardson's fielding prowess was recognized when he joined Buffalo (why else move a CF to 3B?), but when that wasn't an ideal solution he was moved back to CF, then tried at second, which worked very well. His ability to play OF was known, however, so he covered CF as needed in 1884 and 1885, then got shifted out there, with the intention that it be permanent, when Dunlap, who was the preeminent defensive 2B of his generation, was acquired. Richardson was still able to play second, however, and Detroit reverted to him readily when Dunlap became unavailable. By 1889, however, his skills were slipping somewhat, so once he moved off of second base in 1889, he stayed off.

That's how I read Richardson's career and its implications about the defensive spectrum, so far.

I would agree intuitively to the 3B-2B-CF defensive spectrum in the late 1870s and early 1880s, though I haven't really studied it. Double plays weren't huge at this time, right, with all the running, bad gloves, etc?

Would anyone agree that 1B is closer to CF than LF/RF at this time also? RF is clearly the far end of the spectrum too, right? It's been quite awhile since I've looked at it, but that's how I 'remember' it.

Would anyone agree that 1B is closer to CF than LF/RF at this time also? RF is clearly the far end of the spectrum too, right? It's been quite awhile since I've looked at it, but that's how I 'remember' it.

My belief is you should begin (1871) saying "LF/CF than RF" but that is not grounded in or confirmed by any analysis of batting data that would pertain to scarcity and might suggest timing.

49. jimd Posted: February 13, 2008 at 05:08 PM (#2690506)
> [DanR:] Can anyone update me on whether there were any significant defensive spectrum shifts from 1871 to 1892, leaving pitchers aside?

The period is too short to study from the point-of-view of average OPS+ at a position.
Naturally, this didn't prevent us from speculating, particularly about the effects of rule-changes.

I'm not certain what jimd means by period too short. I would have guessed changes too rapid. No generalizations about 1871-92 as a whole are true.
But jimd is the one who knows from experience the method of "average OPS+ by position". His point may be grounded in some difficulties of measurement.

--
Regarding 1B,
I am certain that the change from 1871 to 1892 was huge. Before 1892 the gloves were good enough that first base must have been a good option for many many more players than in 1871. Alternatively, if captains simply put their biggest bashers at first base, the cost in below average fielding would have been much less in 1892 than in 1871.
But this certainty does not support any conclusion about value at the major league level. Only by data analysis can we know whether the change was big or small in terms of filling the rather small number of major league (and through the early 1880s other high-level professional) jobs.

--
Joe Start's record shows at a glance notable increase in the DP rate in 1877. That will be easy to study systematically but difficult to interpret regarding the fielding spectrum, I suppose.

Chris Cobb concludes
That's how I read Richardson's career and its implications about the defensive spectrum, so far.

It makes great sense to me, rings true, so far.

jimd #49
Fair-foul rule. IIRC, this was changed for 1877 [yes] to very near it's current form. In earlier games, a ball was deemed fair or foul based on where it landed; in later games a ball that landed fair could become foul if that happened between the plate and 1st/3rd base (same as today). Some ball-control batters (e.g. Barnes) could place the ball down fair so that it quickly went foul; fielding these quickly placed extra emphasis on mobility at 1st, 3rd, and Catcher.j

Even without the common spin and outward curve of the hard ground ball that is foul today, a batter can get many or most of them to hit the ground in fair territory by moving up in the batter's box. (With the spin and curve, I suppose that many hard fouls hit fair territory, even for the batter in the back of the box.)
So the early rule not only put premium on fielder mobility but it kept the third and first basemen near the foul lines.

One 1871 rule permitted the fielders to play anywhere, eg one or both feet in foul territory.
One 1874 rule introduced the batter's box.
I suppose that both rules regulated fair-foul hitting, the first making it easier to defend, the second making it harder to execute.
Then in 1877 the ball must remain in fair territory or pass the bases in fair territory -- the end of "fair-foul".

For now:

(Please note that my All-Time ratings include players not yet eligible for the HoM.)

Major Oversights--All-Time Top 110 or So Players
(1) Dahlen--Not sure, but I think he beats out Blyleven, Raines, Grich, Santon, and Home Run Johnson for the title "Best Player Not in the HoF (Non-Gambler Division)"
(2) White
(3) Hines

Signficant Oversights--All-Time 110-160
(4) Glasscock
(5) Sutton
(6) Magee--Not sure why other voters are knocking him; along with Dahlen, strikes me as one of the few candidates the HoF simply forgot.
(7) Groh
(8) Caruthers
(9) Keller--With appropriate credit for war and MiL, his 7-year prime is in the top 100 of All-Time.

Comfortably HoM--All-Time 160-210
(10) Gordon
(11) Richardson
(12) Gore
(13) Bennett
(14) Sheckard

Towards the Bottom of HoM--All-Time 210-230
(15) Ferrell
(16) Hack
(17) Stovey
(18) Childs--Looks like Gordon or Doerr until you adjust for shift in defensive spectrum.

Right on the Bubble--PHoM, but not All-Time Top 230
(19) Start--His performances in the NA and NL don't match up with the performances of elite 1B at those ages.
(20) Jones

Not PHOM
(21) Browning--A great hitter, but can't survive the triple wammy of (1) poor durability; (2) fielding ineptitude; and (3) AA deduction--every time I come up with a comprehensive system, he comes out about 300th.

Start--His performances in the NA and NL don't match up with the performances of elite 1B at those ages

And?

And we'll just pretend Joe Start wasn't the greatest baseball player in the world before 1871.

Right on the Bubble--PHoM, but not All-Time Top 230
(19) Start--His performances in the NA and NL don't match up with the performances of elite 1B at those ages.

NL - ages 33.7 to 43.9 yrs.mos??
He batted poorly in the first NL season, then returned to good form (or attained it). OPS+ 127 after age 33.

Note.
baseball-reference provides part-career records after age X.
See "Most Similar by Age" and select 'C' (compare) in the row for age X.
Then use the options "Display totals from age to end of career."


sunnyday2:
just pretend Joe Start wasn't the greatest baseball player in the world before 1871.

I'm not sure he was the greatest at any moment but he has a good case for the decade '61-70 or for entire pre-1871 careers.

elite 1B at those ages.

NL - ages 33.7 to 43.9 yrs.mos??

Who is in the comparison group?
Adrian Anson, Ete Prose, Astanasio Perez, Eddie Murray, Andres Galarraga

Among them Ete was in the outfield before age 38 (his credit), Tony was half-time or less after age 38 (Joe Start's credit) and Eddie was DH after age 37 (Joe credit).

Jake Beckley was only 39.2 at the end of his last full season.

Jake Beckley was only 39.2 at the end of his last full season.

My perennial Jake Beckley comment:

He’s Joe Start, but without a peak and retired four years sooner.

Would anyone agree that 1B is closer to CF than LF/RF at this time also?

Depends on the year you ask about. I'd agree with you for the 189?-192? period (of the Beckley debates). Also for the fair-foul period of the early 1870's. OTOH, the ABC generation hit as well as if not better than the FGG generation of the 1930's. 1B was the masher's spot then.

RF is clearly the far end of the spectrum too, right?

There was a genuine strategic debate going on there. Play an insurance substitute (UT or 2nd C/P) or a DH hitter. Different managers/teams doing different things. Some RF'ers mashed (e.g. Tiernan), some situations were scars based on the managers perceived substitution needs, the most successful/famous had players that combined both strategies (Kelly, Caruthers). Which of the two strategies was best is not at all obvious from our vantage point.

More on Richardson's fielding:

Win shares more or less agrees with WARP on Richardson's fielding at each position.

It grades him out at C- for his two years at third base, 1879-80.
He earns an A+ in CF for 1881.
He is A to A+ at 2B in 1882/83.

Win shares sees Richardson's most valuable defensive performance during this period as coming in centerfield in 1881. WARP disagrees, seeing his play at second as being more valuable, though he was farther above average for his position in CF.

On Richardson at 3rd base:

Bot WARP and WS sees Richardson as acceptable out there. But he must not have looked very good doing it, or he refused to continue there maybe. After being Buffalo's starter at 3b for 1879 and 1880, he was replaced by Jim O'Rourke, career OF who had never been a regular at 3rd (though he had played some emergency fill-in in both 1875 and 1880 for Boston; not very good at it either). In 1882, Force moved back to SS after two years at 2b, Richardson was placed at 2b, O'Rourke went back to CF, and they put Deacon White at 3B (even less experience than O'Rourke). It's difficult to determine exactly what is really going on here with all of the defensive shifts on the Buffalo team.

Joe Start was the best baseball player in the world at a time when not many people were playing baseball. Like any other player who dominated inferior competition, we need to see how well he did/could have done against better competition to see how good he really was those years. Lucky for us, we have some data to help us. When the game expanded beyon a few cities in 1871,he was only 28. Here's a crude comparison of him from that age on to the other elite 1B of his era and some of the marginal HoM bats from the era.

Times in Top 3/5/10 in League in OPS+, Ages 28 On

Start 0/1/4

Anson 3/6/11
Brouthers 6/6/7
Connor 7/7/7

Stovey 2/5/6
Jones 2/5/6 (plus blacklisted 2 years)
Thompson 2/3/4

I don't think even coming close to matching up to Anson/Brouthers/Connor is a requirement for HOM qualification.

Plus the fielding-value difference between a 1B and OF is absolutely enormous at that time. Why pick OFs?

I'm comfortable with electing Joe Start to the HoM. In fact, I PHoMed him sometime during the 1970s. He strikes me as very similar to Beckley--an extraordinarily long career, a good but not great glove at 1B when the position had more defensive value, and a bat that was in the top 20 or so in the league most seasons but ony only cracked the top 10 in its best years. All, I'm saying is that his performance from age 28 on suggests that he wasn't a spectacular hitter. When you take that into account and timeline a little bit (to deal with how underdeveloped the game was in its earliest years), I have him somewhere around the 25th best 1B of All-Time. On this ballot, that beats one or two guys.

On Howie's specific question, I agree that 1B had a lot more defensive value in the 19th century than in the modern game, but I'm surprised ny the suggestion that there was "an enromous difference" between 1B and OF. Do others agree? Does that apply to CF as well?

During most of his career, but no more than half of his league career, I'm sure there were serious baseball teams that were hobbled in competition because they couldn't find or train a firstbaseman who could catch the infield throws reliably, or who could sometimes catch the bad throws.

I don't know whether or when that affected the 5th best team or the 15th best team (eg, the weakest NL and IA teams in 1877 when Spalding moved to 1B in Chicago and used a glove). That is, I have no clue how much it affected them. After the war I suppose that no serious team would use someone who was three catches per game below the norm for serious teams. But how much did it cost them in diminished batting to use someone adequate at first? How many "good catches" enjoyed 10 and 15 year careers when they were fair batsmen for only 5 seasons?

Even with more data and good analysis it may be impossible to tease the captain effect from the fielding and batting. First base is a good position for the captain/manager. Much later Patsy Tebeau led a strong team from first base for seven seasons while woefully weak as a batter (1894-1900).

--
About "outfielders" see the left:right discussion above.

31. mulder & scully Posted: February 11, 2008 at 02:01 PM (#2688229)
Paul,

Thanks for the information. I followed some links about the Bryce Book, but I don't think the Canadian Museum has scanned the pages like the Library of Congress did with many of the Spalding Guides.

I emailed you my mailing address.

Thanks,

Kelly

That is IA/NA 1877-1880 playing statistics.
I thought I would find, fotocopy, and send by Saturday mail. Nope. Tuesday is far enough in the future to seem likely. Well I'm at it, would anyone else use a copy?

Others have focused on the chronological subsets.
Here our eight 3B-SS-2B candidates.
The first two columns give career and fifth-best season OPS+. The next two give fielding win shares grade and full seasons equivalent at all fielding positions (#rank in parentheses).

Group 3 candidates at 3B-SS-2B
all 5th FWS fsField -
109 123. A+ 16.85 Dahlen 1891-1911 (#20)
119 146. B+ 16.32 Sutton 1871-88 (#27)
112 118* A- 14.95 Glasscock 1879-95 (#58)
119 129. B- 12.17 Hack 1932-47 (#232)
130 134. Ao 12.03 Richardson 1879-92 (#244)
118 123. A- 10.82 Groh 1912-27 (#361)
119 123. B+ 10.46 Childs 1888-1901 (#411)
120 123. Ao _9.96 Gordon 1938-50 (#470)

Hack, Groh, Gordon, and Dahlen played about 0.4, 0.2, 0.1, and 0.1 seasons as pinch-hitter/runner.

*This does not count Glasscock's split 1884, one-third as a dominant player in the Union Association. His fourth best season by OPS+ is 137

The fielding win shares grades cover primary fielding position beginning in 1876.
Sutton loses roughly his first 4.5 of 12 seasons played at third.

professional play prior to major league careers

Sutton, 1870, 3B
Glasscock, 1878, 3B
Richardson, 1876-78, of 2b 3b c ("gold glove" cf in 1878)

Richardson was 2.2 years older than Glasscock. They batted roughly .325 and .265 in 1878.

There was a lot more professional baseball in Cupid Childs time. According to the biography in Baseball's First Stars (1996), he played for eight different clubs 1886-88 before a full 1889 season in the new International Association, the strongest minor league. That is the IA where Frank Grant excelled in 1886-88, also fielding 2B. They both batted about .350 (all four IA seasons plus Childs in AA 1890), Grant slugging about .530 and Childs .450 and .480.

Bill Dahlen played one season near home in New York State League 1890, a lower-level minor. He too fielded second base.

--
Sutton, Glasscock, and Childs entered the major leagues with their clubs in 1871, 1879, and 1890. Richardson signed with Buffalo one year after that club left the IA for the NL. Dahlen was invited to tour with Chicago in 1891 preseason, based on the recommendation of a local friend of Cap Anson.

http://www.baseballthinkfactory.org/btf/pages/essays/rickey/goodby_to_old_idea.htm

Dolph Camilli?

And we'll just pretend Joe Start wasn't the greatest baseball player in the world before 1871.

I can't see him being better than Pearce.

(argh, lost a few paragraphs. Don't compose in the composition window, Paul.)

Briefly, after a Saturday marathon of baseball player stuff I signed off feeling pretty good about
14. Charlie Bennett
and a bottom seven, not yet in order, composed of
Sheckard
Ferrell
Childs
Hack
Keller
Jones
Browning

I fell asleep mulling, "What does Chris Cobb see in Jimmy Sheckard? I can't post a prelim until I work that out."
Now John Murphy has posted the first ballot. He agrees with Sheckard and Ferrell near the bottom but he puts the other five of my seven at #5-6-7-11-12.
The highlit comment on Childs is most remarkable.
Best major league second baseman for 1890, (almost in 1891), 1892, 1893, 1894, 1895, 1896, and 1897.

If so, can I put Childs at the bottom?
No.

Is it so?
No, it isn't, right? :-)
Childs 1890 should be one of the great suspect seasons of the 19th century, with Browning 1882.
But that leaves "almost in 1891" followed by *six consecutive seasons as the best major league second baseman*, something that didn't sink in here when John said the same on his annual ballots.

(that's all for now)

I fell asleep mulling, "What does Chris Cobb see in Jimmy Sheckard? I can't post a prelim until I work that out."

FWIW, I have downgraded Sheckard from my prelim. Not to Murphyesque levels, mind you! I think my rating of him was pulled upward too much by my use of systems in which the replacement levels were too low. I now have him down to #8, behind Dahlen, White, Glasscock, Hines, Start, Gordon, and Groh, and just ahead of Keller, Ferrell, Magee, Gore, Bennett, and Richardson.

He's now at the top of the middle third. He may well move lower in that group, but I don't see him dropping into the lower tier.

A note about Sheckard, both for users of WS and of my WARP (which are dependent on Fielding WS for their defensive numbers): Fielding WS is absolutely in love with Sheckard. It has him as the greatest defensive corner outfielder of all time by *far*, leaving guys like Clemente, Barfield etc. in the dust. The equivalent of something like an average +25 a year. DRA agrees that Sheckard was downright superb, but no greater than, say, Kip Selbach. Perhaps a WS expert could explain why Sheckard's Fielding WS are so out of line with those of other oustanding corner outfielders.

Perhaps a WS expert could explain why Sheckard's Fielding WS are so out of line with those of other oustanding corner outfielders.

Sheckard played for most of his career on _outstanding_ defensive teams. Since WS basically gets half of its outfield rating from team defensive efficiency (reasoning that if the team is a good defensive team, it has a good defensive outfield), that is probably a contributing factor to his top-notch ranking.

Since we don't have reliable replacement levels to use for the 1870s and 1880s players, I have been looking at them from the perspective of BRAP (taken from Lee Sinins' encyclopedia) plus WARP's FRAA. Having nearly finished working up these players from that angle, I started looking at the later players from this perspective, too, and I was surprised when Sheckard came out a lot lower by this measure than by Dan R's WAR. The WS fielding assessment being a part of Dan R's WAR would explain it.

Even by a runs above average measure, I still think Sheckard is Hall-of-Merit-worthy, but when a gonzo assessment of his fielding is taken out of the picture, he looks much less impressive.

Btw, looking at BRAP from Sinins, its clear that standard deviations for the 1880s must have been huge.

86. David Concepcion de la Desviacion Estandar (Dan R) Posted: February 18, 2008 at 09:48 AM (#2693619)
A note about Sheckard, both for users of WS and of my WARP (which are dependent on Fielding WS for their defensive numbers): Fielding WS is absolutely in love with Sheckard. It has him as the greatest defensive corner outfielder of all time by *far*, leaving guys like Clemente, Barfield etc. in the dust.</i>

I believe it's first by a little over Fred Clarke, by a lot over all the moderns. (Are there any very-highly-rated modern LFs or very-highly-rated deadball era RFs?)

Fred Clarke was the captain/manager beginning in his age 24 season (1897-1911). In his last season as a player, he worked only 101 games in left; rookie Max Carey worked 46. But for the next four years, with Clarke as bench manager, Carey played left. By all accounts Carey was a superior center fielder, but he didn't move there until Clarke left the bench.
Pittsburgh had a big left field. So did Washington. I guess that a combination of factors put Carey in left: it was almost second CF in Pittsburgh; Clarke overrated his own position (either "behind the curve" or because he played there).

I don't know any of the details regarding Sheckard and Brooklyn or Chicago except what I can report from bb-ref. Former CF Ned Hanlon made the early decisions about his mlb career, not only where to play him in the outfield but whether to keep him in Brooklyn for 1899 or send him to Baltimore (to Baltimore, taking Keeler's place in right).

It's clear why he started in left, joining fixtures Mike Griffin and Fielder Jones (1898). Back in Brooklyn for 1900, it was left beside Jones and Keeler (1900). But why in 1901 and after did he remain in left field?

(Hey, that isn't all straight from bb-ref.)
I meant why in Brooklyn under Hanlon 1901-1905.
I don't know anything about the ballparks in Brooklyn or Chicago.
In Chicago for the 1906 season, the record-setting winners, Frank Chance slipped Sheckard and Steinfeldt into their customary places, perhaps looked for veteran LF and 3B to fill those places? Fielding was the strong suit of incumbent CF Jimmy Slagle, too.

At "Joe Start and Cal McVey" I have posted per game statistics for Joe Start 1860-70. The summary providing rank on team is #102, the detail is at #98. A mess intervenes.

During his first season with Atlantic, 1862 --when he ranked 9 in runs and 5 in outs-- the club played only five matches, none before August 11. Otherwise Atlantic always played the same order of magnitude as the leader.

During 1864-68, for two undefeated seasons and three as arguably the strongest team, Start ranked 5-1-1-1-1 in runs per game and 2-1-2-1-1 in (fewest) outs per game. The professional seasons of 1869-70 brought no great change for Atlantic except that other clubs further concentrated talent. Start led in runs, hits, and total bases (1869), then in hits (1870, no runs data).

Paul, there was a time in this project when I thought I knew who was regarded as the best player--the MVP, as it were--for the period 1860-1870. Now, I have no clue.

Jim Creighton would be in there somewhere.

Dickey Pearce for a few years.

Then Joe Start.

I think it would be George Wright by 1868-69-70?

Could you put a list together?

A runs-created-above position look at the pre-1893 players.

I have been bothered enough by the problem of the inaccurate replacement levels in WS and WARP for pre-1893 players—since I don’t have much more accurate rep levels from Dan R to work with—that I have taken a fresh look at this cohort using RCAP, taken from Lee Sinins’ encyclopedia, to look at these players’ value above average, and then make my own adjustments to a plausible replacement level. What I have done is to combine seasonal RCAP from Sinins with seasonal FRAA from WARP1 to get seasonal runs above positional average (I know RCAP are not normalized and FRAA are, so they’re not perfectly commensurate measures, but since we are comparing players from the same period, I think this should lead to little distortion of their relative values), which I have competition-adjusted for the AA (using league strength estimates derived from WARP) and season-adjusted to 162 games for all players.

The table below gives the season- and league-adjusted RCAP and FRAA career totals for the players under consideration, and their sum as RAA.

The table also includes a runs-above-replacement estimate. Dan R’s work suggests that in 1893 the average player, considered independently of position, was about 2.12 wins above replacement. This total would decline swiftly over the next ten years to about 1.7 wins above replacement. Extrapolating backwards, I have set replacement level for the 1880s seasons as 2.3 wins above rep, and 2.6 wins above rep for the 1870s. I have assumed a 4.5 r/g environment to be in keeping with WARP, so I have given players 9*2.3 and 9*2.6 runs per 162 adj. games played for each season and added that total to their RAA totals to get RAR. I have also calculated their rates of RAA for 5 consecutive seasons to find a peak rate of RAA/162 games. (When a player would have a better rate over 6 or 7 seasons, I have used the longer interval).

I have then combined their RAA, RAR, and Peak Rates according to my usual system, RAA + RAR + (5xPRate) to get a composite figure, which I have used to rank order the candidates. I am going to look at this ranking as well as the one generated by my old WARP/WS system when putting together my ballot. This table says nothing, of course, about how to integrate post-1893 players into this list. That’s my next project before I cast my ballot.

Because this is the early game and these are the players the Hall of Fame has overlooked, most of them have something odd that has to be estimated or adjusted for. Notes following the table explain what I have done in each case.

I was surprised by some of the results. RCAP is far from perfect, but in absence of certainty surrounding the defensive spectrum during this period, it seems like a measure that is worth a look.



I sure hope that formatted decently. I avoided spaces, but I'm not sure I remembered the proper tags . . .

A. White is an obvious #1 by this system, leading in every category except FRAA. These totals do not include any pre-1871 credit, mostly because there was no need to work it out in order to rank him properly.

B. Glasscock’s UA portion of his 1884 season is pro-rated here as if he had continued playing in the NL at his seasonal rates as established to the point where he jumped to the UA. His RCAP is better than I would have guessed, and his FRAA is eye-popping.

C. Hines’ totals do not include his final (1890-91) seasons, which were below replacement level. With those included, his composite drops about 100 points. If you were to drop off his below-replacement 1872 as well, he would gain 45 RAA and 34 RAR, totaling 1998 in his composite.

D. Start’s RCAP and FRAA #s cover 1871-86. His RAA and RAR numbers include estimated value for his pre-NA play, which I have simply set as equal to his first nine seasons in the NA and NL. This probably underrates him, but I can’t see any method based on raw data for ascribing post-1871 batting or fielding runs to him, so this seems a convenient way of making a conservative estimate.

E. Like Start, Bennett’s RCAP and FRAA are his actual, season-adjusted totals. His RAA and RAR include a catcher-bonus, basically increasing his playing time by a multiple. That multiple is 1.15 for 1880-83, 1.2 for 1884, and 1.3 for 1885-93.

F. Ezra Sutton’s totals include credit for 1870, for which I assigned the average of his 1871-75 seasons. He does poorly by this measure, mostly because WARP sees his fielding as well below average: say C or C+ in WS terms. WS, of course, sees him as above average: B+. Adding 60-70 FRAA to Sutton’s totals would make a big difference, as his RCAP are in the range of the other 1870s and 1880s HoMers.

G. Charley Jones’ totals do not include his 1887-88 seasons, which, when the AA is adjusted for quality of competition, were below replacement level. All of his AA years are competition adjusted, using the rates of adjustment employed by BP’s WARP. For his missing seasons, I have given him an average of the four surrounding seasons, league-context adjusted. I have also prorated his 1880 season as if the season had ended when he was kicked off the team.

H. Browning’s totals are competition-adjusted for the AA.

I. Stovey’s totals are competition-adjusted for the AA. His totals are absolutely killed by using RCAP rather than RCAA. He trails Browning by 100 runs in RCAA, which he makes on defense and playing time to top Browning in an RAR measure. But in RCAP, he trails Browning by 300 adjusted runs, and there’s no making that up. Looked at in terms of RCAP, he appears a clear mistake for the HoM.

J. Caruthers’ line is incomplete because I haven’t worked out replacement level for pitching yet, or a denominator for rate. The total in his FRAA column is actually the sum of his PRAA and his FRAA. I suspect, given the shortness of his career, that he will end up here. Certainly he looks to be in the Jones/Browning tier rather than any higher one, unless his rate is stupendous.

Overall, I was really surprised at how poorly Sutton and Stovey did and how well Glasscock, Richardson, Bennett, and Browning did. I think Sutton and Stovey are a bit better than shown here, and that Glasscock is probably a bit worse. I see no reason to doubt the results on Richardson, Bennett, and Browning, however.

Great work, Chris. I see you are still using version 1.0 of my WARP--in my updated numbers, I make an adjustment so that the overall replacement level across positions for hitters remains the same in any given year. Letting the overall level float, as I did in v 1.0, is by definition not being fair to all eras, no? The shift you are observing is just because of kurtosis--in the 1890's, teams were much more willing to hand large amounts of playing time to REALLY GODAWFUL PLAYERS than in later eras, so even though I adjust for standard deviation, the "fat tails" of the 1890's distribution lead to a deceptively low calculated replacement level. There is absolutely no way that, say, Farmer Weaver was really the best right fielder Louisville could come up with for the minimum wage in 1894--not close to it. The result is that 1890's players, particularly outfielders, were all overrated relative to subsequent eras in my initial WARP. Version 2.0 corrects this problem. I definitely would not use a lower rep level for the 1870's than for the 1890's.

Another thing to bear in mind is that the standard deviation of BP FRAA is much higher in the 19th century than it is later on, so your results will be biased towards good fielders and away from bad ones more than they would if you applied the same methodology to subsequent eras. Maybe you think that's right, but I just wanted to draw attention to it.

I'll start poring into your helpful work here now and see if I have anything to add.

Letting the overall level float, as I did in v 1.0, is by definition not being fair to all eras, no?

Certainly, and if I were comparing players across eras I would want to adjust for shifting replacement levels. For working within a single era, though, I thought it better to model small shifts in value of the average player. Not that I think it would make any difference in the study if I set the level constant through the whole period at, say 2.2 wins above replacement per 162 games. Where players are very close, like the Richardson/Start/Bennett trio, the order might change, but the noise in the system is already so great that I don't think one can place much weight on differences of a few runs.

I plan to upgrade to your updated numbers at some point, but I want to do it as a global change, and I haven't had time to for that. Possibly over the summer . . .

<i>I definitely would not use a lower rep level for the 1870's than for the 1890's.

I don't agree, based on what (little) I know. I would guess that 1877-83 rep levels were higher than in the 1890s, but before and after that time I think they were probably lower. The quality of team range was enormous in the NA and in the mid-1880s during the expansion period. You regularly had 3 and 4 teams with under .350 winning percentages in the NA, not counting teams that only played a part season. Even in the 1890s in a twelve-team league, there were never more than 1 or 2 teams that bad. In 1884-86, it is similar to the NA years: an eight team league with two teams below .350 and two above. I haven't looked at it on the level of individual players for these seasons, so the player data could surely prove me wrong, but it looks to me like there were more very bad players around in these seasons, so rep level was probably even lower than in the 1890s.

Another thing to bear in mind is that the standard deviation of BP FRAA is much higher in the 19th century than it is later on, so your results will be biased towards good fielders and away from bad ones more than they would if you applied the same methodology to subsequent eras. Maybe you think that's right, but I just wanted to draw attention to it.

And you are right to do so. I wouldn't use this method in any straight-up way to compare players even of the deadball era to 1871-92 players. All I am trying to do here is get a better handle on the implications of fielding position for players who are roughly contemporary. I think it does that. There are lots of ways this study could be improved by those with more sophisticated math skills, fuller data sets, and more computing power. I have been hoping to see more analysis and debate before the voting than has so far been forthcoming; maybe this will provoke some?

One other methodological point: Sinins' data only goes back to 1876. For the NA years (for Start, Sutton, Hines, and White), I set the difference between RCAA and RCAP based on the average difference between those values for that position 1876-79, took WARP1's BRAA as the base stat, and modified it accordingly. Just another piece of guesswork pastiche behind these numbers.

C. Hines’ totals do not include his final (1890-91) seasons, which were below replacement level. With those included, his composite drops about 100 points. If you were to drop off his below-replacement 1872 as well, he would gain 45 RAA and 34 RAR, totaling 1998 in his composite.

Hines in 1872 and Jones in 1875 entered the majors with clubs that played only about one quarter of the number of games played by the leading clubs. Neither played for a second NA club after his first one dropped out, so their NA seasons were short. Hines was a weak player and Jones a strong one, so how you handle the short seasons (eg, prorate to 162 team games with regression) probably makes a big difference for them, in opposite directions.


I. Stovey’s totals are competition-adjusted for the AA. His totals are absolutely killed by using RCAP rather than RCAA. He trails Browning by 100 runs in RCAA, which he makes on defense and playing time to top Browning in an RAR measure. But in RCAP, he trails Browning by 300 adjusted runs, and there’s no making that up. Looked at in terms of RCAP, he appears a clear mistake for the HoM.

Do you (Sinins) assign fielding position at the season level --compare every player-season batting record compared to that season's norm for one primary fielding position?

What is the scope of the norm? Eg, is it LF (distinct from CF and RF) in AA 1883 (distinct from NL)?

Neither Stovey nor Browning played anything resembling a full career at one position. Can you tell whether this 300-run difference derives mainly from some particular fielding positions, such as Browning at CF or Stovey at 1B?

Chris,
Re Stovey and Browning, I am wondering whether you can analyse 200-run difference between their lifetime batting norms. How much from assigning distinct norms to AA and NL seasons, etc.

Chris, I would consider that a standard deviation issue, not a replacement level issue. If there were plenty of sub-.350 teams, there were also plenty of teams above .650. That, in turn, means that a single win above replacement or average is worth fewer pennants in that league than it would be in a lower-stdev league. What I believe to be the correct methodology is to first correct for standard deviation, so that the wins above average-pennants relationship is the same in all league-seasons, and then subsequently calculate a flat *overall* (as opposed to position-specific) replacement level from those standard deviation-adjusted wins above average. You could theoretically do it the other way around, first using worst regulars to calculate a replacement level and then multiplying the resulting WARP by the standard deviation correction, but that would make it difficult if not impossible to track changes in replacement level between positions across eras, since you'd be measuring both changes in the floor and changes in the league stdev.

The effect of your setting an overall rep level lower than the one used for subsequent eras is to make these numbers more like BP WARP or Win Shares--rewarding "just showing up" at the expense of on-field excellence. Combined with the high FRAA stdev, you'd expect long-career players with great gloves to come out particularly strong.

I hope it didn't sound like I was criticizing your work at all--I really appreciate your taking a stab at this. I just wanted to chime in on the one part of the methodology I felt I had something substantive to comment on.

I myself have been scared off from the pre-1893 crowd due to my inability to find a run estimator for those league-seasons with a tolerable degree of accuracy. I have equations that work OK for some years (the early NA I can do fairly well, actually); but others (the 1883 AA I think drove me nuts) just seem impossible.

Hines in 1872 and Jones in 1875 entered the majors with clubs that played only about one quarter of the number of games played by the leading clubs. Neither played for a second NA club after his first one dropped out, so their NA seasons were short. Hines was a weak player and Jones a strong one, so how you handle the short seasons (eg, prorate to 162 team games with regression) probably makes a big difference for them, in opposite directions.

There's no regression in this little study. Except in the case of Jones 1880, I adjusted all seasons according to the "normal" season length of that year, so for Jones 1875 I considered him to have played 13 games out of a 75 game season, while I considered Hines to have played 11 games out of a 45 game season. (Reviewing this data shows that I entered Hines' season length incorrectly: I'll post adjusted numbers for him tonight). Jones actually shows up as a below-average player for 1875, because WARP sees his fielding as _terrible_ (-4 FRAA in 13 games!), and adjusting his BRAA to RCAP costs him a run created.

Neither Stovey nor Browning played anything resembling a full career at one position. Can you tell whether this 300-run difference derives mainly from some particular fielding positions, such as Browning at CF or Stovey at 1B?

I'll post their RCAA vs. RCAP vs. position breakdowns from Sinins tonight. The RCAP numbers are definitely for a single position in a single league season, e.g. RCAP for 1B for 1881. Whether Sinins prorates for seasons where a player split time among two or more positions (as Browning often and Stovey sometimes did) I don't know. A lot of the difference derives from Stovey's time at 1B vs. Browning's time at 2B/3B.

Chris, I would consider that a standard deviation issue, not a replacement level issue.

Well, it is both, as your fuller description shows. I agree that your approach to addressing it is appropriate, but, as it requires a lot of data and calculating, it's not an approach I can take myself.

The effect of your setting an overall rep level lower than the one used for subsequent eras is to make these numbers more like BP WARP or Win Shares--rewarding "just showing up" at the expense of on-field excellence. Combined with the high FRAA stdev, you'd expect long-career players with great gloves to come out particularly strong.

Well, if my replacement level estimates are closer to accurate than those of WARP and WS, which are clearly too high by several wins per season, then I think my results will be better than theirs. As to whether "just showing up" should be rewarded, that's a normative question. I think it is reasonable to look at contextual value, as long as it is calculated with reasonable accuracy, even if it does not lead to a measure that can be used unadjusted for cross-period comparisons. Tonight, I will post adjusted games played totals by decade for this cohort of players, so anyone who wants to adjust the replacement level can do so.

I hope it didn't sound like I was criticizing your work at all.

My work is certainly far from perfect, so it should be criticized. Insofar as I can implement changes in response to methodological criticism, I will, but many of the methodological limits of this study are not ones that I can solve, and so part of my response to criticism must be to distinguish between what I can do and what I can't do.

I myself have been scared off from the pre-1893 crowd due to my inability to find a run estimator for those league-seasons with a tolerable degree of accuracy.

Yes, that's another methodological rub. I would describe my numbers as an attempt to remove one source of error that WARP and WS obviously introduce, on the one hand, and to be inaccurate in a different way, on the other hand. I don't know what run estimator lies behind Sinins' RCAP, and RCAP is a volatile measure with its own issues as a measure of value or merit. To help give everyone a comparative view of the offensive estimates of the various available systems, I'll try also to post tonight an (unadjusted) comparison of WARP1, WARP2, WS, RCAA (Sinins), and RCAP (Sinins).

> Neither Stovey nor Browning played anything resembling a full career at one position.
> Can you tell whether this 300-run difference derives mainly from some particular fielding positions, such as Browning at CF or Stovey at 1B?
[later #97]
> Re Stovey and Browning, I am wondering whether you can analyse 200-run difference between their lifetime batting norms. How much from assigning distinct norms to AA and NL seasons, etc.
<<

I'll post their RCAA vs. RCAP vs. position breakdowns from Sinins tonight. The RCAP numbers are definitely for a single position in a single league season, e.g. RCAP for 1B for 1881. Whether Sinins prorates for seasons where a player split time among two or more positions (as Browning often and Stovey sometimes did) I don't know. A lot of the difference derives from Stovey's time at 1B vs. Browning's time at 2B/3B.

That is plausible. During Stovey's time at 1B, 1883-85 and half of 86-87, the position was covered by a lot of great batters. ABC in the NL, also John Morrill. Dave Orr and John Reilly in the AA, also Comiskey. For Morrill and and Comiskey, merely above average and below average batters by career, Stovey's time at 1B matches their peak seasons, OPS+ about 125 for Morrill and 100 for Comiskey. Those are the eight whom I recall covered by biographies in 19c Stars and First Stars. For the four who are not in the Hall of Merit, I just looked up their records at bb-ref. I invite all to do so. I haven't checked the other eight teams, where other short-term notable batters may be hiding

Henry Larkin is another very good batter in the AA but he and Stovey switched positions for the Athletics, Larkin LF to 1B and Stovey 1B to LF.