attempting to re-think Dickey Pearce:

Pearce was, anecdotally, the best baseball player prior to the mid-late-1860s.
In addition, research shows him to often be among the top 3 hitters on his team in this period, and since he played shortstop, and played it "very well", as well as some catcher, he does look by the records to clearly be the best player on his team, and his team seemed to often be the best of the day.

How good was Pearce's defense at short? It's a real stretch to make him Ozzie Smith - while he apparently "revised" shortstop play, his comparison set in this regard is awfully light (when you revise something, by definition no one else is doing it). I give him credit for being a gold-glove caliber SS for some of his career, like Trammell or Larkin or Concepcion - near the best at their positions in a lrage league, not overwhelming like Ozzie/Brooks/Maz/Willie.

How much do I extend Pearce's best-on-best-team to a league, say, in 1885? In 1860, how many teams were there? How many guys even attempted to play baseball? What areas of the country did they draw from? Did they often play much longer than age 26 when many of them got married and got a real job? The data set from which Pearce is drawn seems to be Much smaller than that of 25 years later.

The research done on the pre-NL and pre-NA guys has been spectacular; I don't mean to be discounting it. George Wright and Joe Start, names unknown to most fans, have gotten their due. But in picking between some of the best players in a known strong era and chosing a 'probably the best' from a fuzzy murky era, I'm not convinced we should honor Dickey over the other consensus top few (Sheckard, Caruthers, Pike) on our ballots.

Admitting, like St Paul, that I only "know in part" and "see through a glass darkly"
Tom

I agree that pearce's pre-NA record and Charley Smith's are similar in quality, though Pearce's is much longer. But Pearce in the NA had a career that clearly resembles the latter years of a really great player. Charley Smith had 1 year in which he hit under 100 OPS, and then was out of the league at 30. I think pattern recognition techniques tell us Paerce was the one who is HOM-deserving, and probably not marginally so.

John Murphy's grandma need not start a campaign for the others; it won't work (now if it was his great great grandma, she'd presumably have seen these guys play, and be able to give us extra input!) Incidentally, so what if you were out on one bounce (which actually I think ended before 1860, but one of you guys will know the defintive truth)-- it will change the stats but not make shortstop any easier to play.

Catching up after 4 days away.

Some responses to comments and questions on John Donaldson:

1) Questions were asked about why he is eligible now.

I listed him as eligible in 1930 because he played at the top level of competition available to him only from 1913-1924. He played on mostly white semipro teams in the upper midwest and Canada from 1924-1930. He had what appear to be token appearances with the KC Monarchs in 1931 and 1934, but these shouldn't affect his eligibility any more than Jim O'Rourke's appearance with the Giants in 1904 affected his.

If this rationale for his eligibility this year seems doubtful, it could be discussed.

2) The eye-popping stats that jhwinfrey posted in #12 for Donaldson were for his years with semi-pro teams. If one were to give him credit for those years, one would need to apply a very steep competition discount to those numbers -- I'd guess a 25% discount to get to Negro-League equivalents, and 35-40% to reach major-league equivalents. Donaldson was done as a pitcher in the top Negro-Leagues after 1920, and he didn't hit quite enough to hold a starting role long-term as an outfielder.

3) KJOK's comparison of Donaldson to Rube Waddell in post 90 seems to me to be apt in terms of their _relative_ dominance at their peaks and in terms of the shortness of those peaks. Donaldson's life was very different from Waddell's, of course. I also think KJOK's ranking of Donaldson relative to contemporary negro-league pitchers is about right.

4) I don't see Donaldson as ballot-worthy; he'll place in my rankings somewhere between Waddell (currently around 30) and Frank Wickware (currently around 50).

On Mickey Welch:

In post 74, it was asked if Welch had a Griffith-like reputation as a shrewd pitcher.

I don't have the book in front of me, but the Neyer/James guide to pitchers suggests he did have some such reputation. He was fairly small for a pitcher in his day and didn't have over-powering stuff, but he was very cool on the mound. He got his nickname, "Smiling Mickey," not because he was reputed to be a happy fellow at all times, but because he was always smiling on the mound, no matter how tight the situation. Someone who actually owns the book could confirm these details.

Brief recaps of the Neyer/James bios for some other 1880s pitchers, as I remember them:

Radbourn -- had a reputation for being wily like Welch, but he was also an experimenter, a guy who mixed speeds, breaks, and arm-angles incessantly.

Galvin -- a pure fastball pitcher; relied on speed and pinpoint control throughout his career.

I regret to say that I can't remember Keefe's note. I don't think it was as informative as the others.

Hey guys,
I'm back. I've missed 2 elections because my workload has been insane. For my personal HoM, I'm going with Frank Baker and HR Johnson in 1928 and Ed Walsh and Joe Kelley in 1929. I'm still deciding in 1930. Here's my tentative ballot for the actual election, though:
1) Lip Pike- Great peak, and for his era a decent career.
2) Dickey Pearce-Must have had an incredible career. Unless you completely ignore career value and focus solely on Peak, he's the #1 eligible SS. He's been in my HoM since 1919.
3) Bob Caruthers-Great pitcher, as I see it the best eligble. His hitting doesn't hurt either.
4) Rube Foster- He's subjective, I'll admit, but I just wouldn't feel right if he doesn't make it eventually.
5) Bill Monroe- Ditto
6) Jimmy Sheckard- Strong Peak, strong career
7) Jimmy Ryan- Ditto, they're almost too close to call.
8) Ned Williamson- Still very underrated by thiselectorate. Supposedly was brilliant defensively. And a solid Off player to boot.
9) Hughie Jennings- Hard to ignore that incredible peak.
10) Roger Bresnahan- Easily the best catcher eligible.
11) Vic Willis- Also Underrated. Strong peak and career numbers.
12) Spotswood Poles- I expect Spotswood to move up later, I just don't feel like I've got a handle on him yet.
13) Jake Beckley- Yes, I'm finally adding Beckley to my ballot. This is more a lack of viable candidates than any new opinion of Beckley. He did have a strong career, though. Even that wasn't spectacular when WS are adjusted for Replacement level.
14) Bruce Petway- Ditto on Poles.
15) Hugh Duffy- Good career, decent peak. I like Hugh, just not for the Hall.

Close(In no particular order)- John McGraw, Herman Long, Charley Jones, Tommy Leach, Rube Waddell, Eddie Cicotte, and Cupid Childs

this might seem like a strange quiston on Pearce but how important is defense when teams were scoring on average 20 to 30 runs a game?

Chris Cobb,
Thanks for the info on Donaldson--I hadn't realized those were semi-pro numbers, though it should have been obvious to me by the locations he was playing in. Shows how little I know about the Negro Leagues, especially in the early 20th century. I will definitely take those 1924-30 numbers with a grain of salt. (And drop Donaldson significantly in my rankings.)

Didn't Pearce play when a ball that bounced once and then was caught was an out?

Didn't Ozzie Smith play with a glove and when they allowed the phantom tag? :-)

(when you revise something, by definition no one else is doing it).

You mean like Babe Ruth? Does any analytical system take into accounts his domination of an offensive weapon that took almost another generation of players to master? If you want to start doing so with Pearce, then we have to follow him with the Babe because there's no way that he would be as dominant (not to mention differences in competition) among today's players as he was back then.

But I doubt the "long knives" will touch Ruth on that point...

Keep throwing your best stuff on Pearce. I think I can hit the cheap without any problems all day long. :-)

John Murphy's grandma need not start a campaign for the others; it won't work (now if it was his great great grandma, she'd presumably have seen these guys play, and be able to give us extra input!)

Actually, I have changed my handle again (but keeping "Grandma" in it). For those not familiar, it's an homage to the great Johnny Murphy of the '30s and '40s Yankee era.

this might seem like a strange quiston on Pearce but how important is defense when teams were scoring on average 20 to 30 runs a game?

Defense was a much greater part of holding back the opposition from creating runs back in that time than it is today. Pitching was far less important than now, so the position players took up the slack.

Didn't Pearce play when a ball that bounced once and then was caught was an out?

Yup... debated since a few years before (1859) the "fly/catch" rule was narrowly defeated at the 1864 NABBP convention. Many teams played "fly games" as exhibitions during this period. The "fly-catch" rule was finally instated after the 1864 season.

Prelim: I am starting to lose faith in Lip Pike, especially looking at him in a world where Dickey Pearce is in the HoM. Do we really want close to 25% of all regulars enshrined?

1. Bob Caruthers (1)

2. Jake Beckley (2)

3. Dickey Pearce (4)

4. Rube Foster (3)

5. Roger Bresnahan (5)

6. Mickey Welch (8)

7. George van Haltren (9)

8. Gavvy Cravath (7)

9. Cupid Childs (10)

10. Pete Browning (12)

11. Clark Griffith (15)

12. Frank Chance (13)

13. Lip Pike (11)

14. Spotswood Poles (off)

15. Hughie Jennings (off)

16-20: Williamson, Monroe, Petway, Konetchy, Sheckard
21-25: Evers, Gardner, C. Jones, McCormick, Ryan
26-30: Doyle, Duffy, Willis, McGraw, Pratt

A bit off topic, but I thought the group would be interested in this:

http://www.baseballamerica.com/today/features/040709schwarz.html

It's an article about a book called "The Numbers Game" by Alan Schwartz, which purports to be a history of baseball statistics. The article itself focuses on George ("Don't Call Me Goober") Lindsey, who pioneered run expectancy tables.

Looks like you can order it on Amazon now, though the publication date says July 21, 2004.

Thanks Jeff!!! I'm putting it on my wish list!

Anyone who wants to to buy that book for me... feel free! :-)

1930 Preliminary Ballot

The drought continues. No new candidates make the ballot. Hoping to see Sheckard elected this year.

For more details, see preliminary ballot and other posts.

1. Jimmy Sheckard. See my post late in the 1928 ballot thread for more, but Sheckard has the best combination of career, peak value, and peak ability on the ballot. Because a lot of his offensive value comes from walks, because fielding makes a major contribution to his overall value, and because his career shape is odd, Sheckard’s greatness doesn’t show up strongly on some measures. But WARP and WS both support a high ranking for him, and I think that’s the most reliable data we have.
2. Dickey Pearce. (2) Clearly one of the top players of the 1860s.
3. Clark Griffith. (4) Career value about equal to the just elected McGinnity. Peak was quite a bit lower, but it was harder for a pitcher to have a great peak in the 1890s. He was the fourth-best pitcher in that decade. He deserves induction.
4. Mickey Welch. (5) Better than Keefe and Clarkson, contemporaries of his already deservedly elected.
5. Lip Pike. (6) Career wasn’t long, but he was a regular longer than Hugh Duffy, Pete Browning, or Sam Thompson. The last serious 1870s candidate, I think, though I've promised Tommy Bond another look before the 1931 election.
6. Hughie Jennings (7) Best peak on the ballot among position players, and one of the best we’ve seen so far. Among position players eligible through 1929, only Barnes, G. Wright, Wagner, and Lajoie have higher peak rates than Jennings. During his 1894-1898 peak, he was the best player in baseball, and better than a pair of contemporary first-ballot HoMers, Billy Hamilton and Ed Delahanty, who were also at their peaks during these years.
7. Rube Foster. (8) The biggest star in black baseball in the aughts. Hurt in my rankings by the shortness of his career, but had more career than Ed Walsh, and nearly as good a peak.
8. Bob Caruthers (9) His case is very similar to Jennings’. He looks to have been one of the top 3 players in baseball over a five year stretch, and a good player for a couple years after that. The last 1880s star I support for induction at present.

------------IN/OUT LINE-------------
Everyone above this line seems to me clearly worthy of induction into the HoM.. The rest of the players on my ballot fall slightly below the current standards of the HoM. They all should wait until after the next wave of all-time greats has passed, until we see if the standards of the HoM will need to lower just a touch as the game expands.

9. Hugh Duffy (10) Outstanding peak and prime. If the Hall looks short a center fielder come, say, 1948, Duffy could get the call.
10.George Van Haltren (11) Or it could be Van Haltren. Outstanding prime. Was an above average player for an exceptionally long time, but never one of the top players in the game.
11. Tommy Leach (12) A lot like Van Haltren in that he had a lot of above-average years, but no great peak.
12. Roger Bresnahan (13) Top catcher of the aughts. Genuinely great player (his peak rate trails only (Jennings, Chance, McGraw, and Pike among eligible position players), but not enough playing time or defensive value to be a definite HoMer.
13. Bill Monroe (14) Great reputation, but the data on his hitting doesn’t justify placing him ahead of Leach, who was an outstanding defender at 3rd and in center, had a career as long as Monroe’s, and who was clearly an above average major-league hitter.
14. Spotswood Poles (15) Just makes my ballot in his first year of eligibility. Career value is equal to Hugh Duffy’s, but more spread out, so he rates a bit lower than the Duffy/Van Haltren/Leach trio.
15. Larry Doyle (16) Makes my ballot for the first time. Fine player. Mediocre defensive play keeps him below the “should elect” line, just as it does Bresnahan.

1930 Preliminary Ballot, continued

The rest of the contenders

16-20. Cupid Childs, Charley Jones, Ned Williamson, Herman Long, Fielder Jones.
21-30. Jimmy Ryan, Jake Beckley, Gavvy Cravath, Tony Mullane, Jim McCormick, Lave Cross, Frank Chance, Rube Waddell, Joe Tinker, Johnny Evers,
31-40. Addie Joss, Bruce Petway, John McGraw, Pete Browning, Roy Thomas, Billy Nash, Harry Wright, Mike Tiernan, Ed Konetchy, Cy Seymour,

Comments on consensus top-ten candidates who don’t make my ballot

Beckley is a consensus top-ten candidate. He has no peak.

Comments on new arrivals worthy of note

42. John Donaldson – Brilliant pitcher for a few seasons, but career as a whole not impressive. A bit like Frank Wickware – a pitcher with great stuff and reputation who barnstormed more than he played for established teams, and burned out fairly quickly. He was a top-notch pitcher 1913-1920, losing most of 1918 & 1919 to WW1 service. Unlike Wickware, who drank his way out of baseball, Donaldson maintained his playing form and played the outfield successfully after his arm burned out. Played outfield for the KC Monarchs 1921-23, then pitched for white semipro teams in the upper midwest and Canada for the rest of the decade and into the 1930s, returning to play outfield for the Monarchs in 1931, after eight years out of organized black baseball.

52. Larry Gardner. A very fine player, probably 9th or 10th among 2b-3b-ss of the 1910s. Has enough career value to be a serious candidate among infielders, but his peak is not high enough, especially in terms of peak rate. More durable than Tinker or Evers, but not as fine a player at his best.

75. Jake Daubert Very fine hitter for average, and sound defensive first baseman. The third best first baseman of the 1910s, behind Ben Taylor and Ed Konetchy.

76. Del Pratt A fine player. Steadily slightly above average for most of his career, but his career isn't long enough for that level of play to make him a viable candidate.

89. Doc Wiley – 3rd best Negro-League catcher of the deadball era, after Santop and Petway. An excellent player during his peak, 1910-15; probably a better hitter than Petway, though not as good as Santop. Didn’t have Petway’s defensive reputation. Career shorter than both of the other two catching stars. He was a part-timer after WW1 – once he got his degree in dentistry, he began his practice, playing baseball on the side. Ranking is approximate.

Candy Jim Taylor – Would be a serious managerial candidate, along with older brother C. I. Taylor. Managed 28 years in the Negro Leagues, won several championships. As a player, good enough to have been a starter for about 10-12 years in the majors, I think. Probably in the 160-180 win share range, which would not place him in the top 100 eligible players.

"(when you revise something, by definition no one else is doing it).
You mean like Babe Ruth? Does any analytical system take into accounts his domination of an offensive weapon that took almost another generation of players to master?"

Uh...yes.
Does any analytical system NOT take his domination of a new weapon into account? They all say that he scored and drove in ('created') more runs than anyone, thus winning more games for his team than any player in history before or since.

Pearce invented a new (defensive) weapon. Is anyone claiming he dominated a generation with it, causing his team to win many games each season over many many years they otherwise would have lost? Hello out there (echo echo echo). I didn't think so.
Linking Pearce with the Babe is downright silly.

"Keep throwing your best stuff on Pearce."

I'm not even warmed up yet....

I got nothing against Pearce. He's on my ballot. He happens to be behind a bunch of other HoFers on my ballot, and ahead of many others. Because, when it comes to comparing 1860 baseball to 1910 baseball, I nope none of us Really believes "a pennant is a pennant". Or votes like it.

I do. I'm the reincarnation of a voter born in 1855, and firmly believe they just don't play the game like they did in my young day. Consequently, the stars of my boyhood and early manhood get preference, and as I get older, I get crankier. In 1930, I'm 75, and if you think that young fly-by-night Cobb is a patch on Dan Brouthers, you're mistaken!

By 1975 I shall be 120, and REALLY cranky!

Uh...yes.
Does any analytical system NOT take his domination of a new weapon into account? They all say that he scored and drove in ('created') more runs than anyone, thus winning more games for his team than any player in history before or since.

Exactly my point. Ruth created a new, improved way of playing the game. If he had been playing Cobb's style of baseball, he still might have been the best, but he wouldn't have stood out to the same degree as he did with the uppercut.

This is the same exact thing that Pearce did as a pioneer at short. Eventually, other shortstops were able to duplicate what Pearce created, but that doesn't mean that we negate the advantage (and wins) created by his innovations.

Because, when it comes to comparing 1860 baseball to 1910 baseball, I nope none of us Really believes "a pennant is a pennant". Or votes like it.

In this case, I do. Pearce played with and against the best. As if I have pointed out before, while Pearce may not have stood out in later years, the same could be said for Cobb, Grove, Mays and Piazza if they had been born in the 1830s. They would be smaller, less strong and devoid of many years of baseball wisdom that guys like Pearce had to invent for them.

By 1975 I shall be 120, and REALLY cranky!

But you don't look a day over 99. :-)

For those worried that the next three years are it for their favorite candidates (Pike, Griffith, Foster, etc) and for those who fear a Sheckard-itis for Caruthers and Pearce, Here's what we have for No-Brainers the rest of 2004 (and some early 2005). For those of you offended by the term No-Brainer, look at the list below. (They are, in chronological order, Johnson, Cobb, Speaker, Collins, Lloyd, Alexander, Ruth, Hornsby, Charleston and Gehrig.) I feel pretty comfortable saying that those candidates are No Brainers.

1930 - None
1931 - None
1932 - None, some might Santop is a N-B, but I won't
1933 - Walter Johnson
1934 - Ty Cobb, Tris Speaker, Eddie Collins, Pop Lloyd, some might say Joe Williams, and I would
1935 - None
1936 - Pete Alexander, some might say Harry Heilmann, but I won't
1937 - None
1938 - None
1939 - None
1940 - None, but some might say Bullet Rogan, and I don't know if I will yet
1941 - Babe Ruth, Rogers Hornsby
1942 - Oscar Charleston, some might say Bill Foster, and I don't know if I will yet
1943 - None, some might say Mickey Cochrane and Frankie Frisch, and I would agree with those people
1944 - Lou Gehrig, some might say Goose Goslin, but I won't
1945 - None

That's 10 N-B, 8 borderline N-B for 31 ballot spots. There is plenty of room for debate and backlog from now until 1945.

This is the same exact thing that Pearce did as a pioneer at short. Eventually, other shortstops were able to duplicate what Pearce created, but that doesn't mean that we negate the advantage (and wins) created by his innovations.
--
Show me the wins he created. Show me a clipping that say "Pearce, playing a different kind of shortstop, gave us a huge advantage for 10 years until everybody else copied thim". I need that before I credit him.


I nope none of us Really believes "a pennant is a pennant".
--
I do. Pearce played against the best.
--
At the risk of re-hashing an old argument....
"Best" is REALLY relative. There is probably some tribe who has a best javelin thrower. And if a contest were held today, he would win. Because most of the rest of us
a. don't practice this sport much , and
b. won't bother showing up for the contest.
Put a ten million dollar prize on it and promote it for fall of 2006, you'll get some better competition. Mister tribesman won't look so hot. Not because of timelining, but because his current definition of "best" wasn't quite robust.

IMHO, Babe Ruth would still be Barry Bonds today.
But Dickey Pearce would be....better than Phil Rizzuto, but not Ozzie Smith.

Brief recaps of the Neyer/James bios for some other 1880s pitchers, as I remember them:

I glanced at the book the other day & caught the comment on John Clarkson. He believed in putting the ball in play & letting his defenders take care of things. Given that, but my system, he had the 2nd best defensive support of any pitcher ever, that's pretty interesting. Give him Pud Galvin's defensive support & I bet he suddenly becomes a far far less impressive pitcher.

A quick point re. Dickey Pearce--when outs are very hard to come by (a la slow pitch softball) then great defense is even more, rather than less, valuable.

A not-so-quick point re. Dickey Pearce. I think the actual record has gotten a little muddied.

1857-1865--Surely had the most accumulated career value during this period, playing valuable defensive positions and being one of the top 2 hitters on what was cumulatively the best team in baseball. Probably didn't have the highest peak--that would be Jim Creighton--but Pearce would have been an MVP candidate pretty much every year. Call him Cal Ripken at Cal's peak (I won't say Honus).

1866-1870--Joe Start took over as the best player on the Atlantics, and G. Wright was probably the best player in baseball in '69-'70. But Pearce continued to play SS (and some C) and to be the #3 hitter (on average, sometimes #2, other times as low as #5 or #8) for what continued to be one of the top 3-5 teams in the game. He would probably no longer be an MVP candidate but would continue to be a perennial all-star, and surely had the most accumulated career value of any player in history as of the founding of the NA in 1870-71.

Now he is merely Ozzie Smith at or near Ozzie's peak.

1871ff--now he is a journeyman SS, still a GG but an average hitter at best and sometimes below average. Now in his decline phase (age 35-41) he is still comp to Ozzie in his decline phase.

For his career, Dickey is probably Ozzie while G. Wright is Ripken, relative to their contemporaries. I'd say Dickey was probably the MVP 3-4 X and an all-star 11-13 X. At the time of his retirement, he would have been viewed as the 2nd best shortstop of all-time, but more importantly he still probably was right behind Start, Wright and Barnes among position players all-time, or 4th best for a 20 year period. That makes him more Ripken than Oz. Everybody will have to decide for themselves how to discount his achievements. But over 20 years, he was #2 at his position, #2 on the best team in baseball, and maybe as high as #4 overall.

Ron Wargo wrote:

That's 10 N-B, 8 borderline N-B for 31 ballot spots. There is plenty of room for debate and backlog from now until 1945.

I agree that elections will not be totally closed to current candidates after 1933, but I would urge us to be careful about sentimentality for current candidates. Many of us have been supporting the candidacies of some of these players for a long time, and we want to see them elected to the Hall of Merit as well as to our various personal HoMs. But as we will have elected, by 1933, a proportionally representative number of players from 1860 to 1910, we need to be careful to give the new arrivals full and fair consideration.

Here's an attempt at a comprehensive list of the fair-to-good candidates for the 8 spots left by 10 shoo-ins and 8 probable shoo-ins that Ron listed who will become eligible from 1933 through 1945.

1933 -- Heinie Groh, Ray Schalk, Zack Wheat
1934 -- Stan Coveleski, Ben Taylor, Christobal Torriente
1935 -- Max Carey, Bingo DeMoss, Carl Mays
1936 -- Dave Bancroft, Oliver Marcelle, George Sisler
1937 -- Dick Redding, Edd Roush, Wally Schang
1938 -- (none)
1939 -- Red Faber, Rabbit Maranville, Jack Quinn, Eppa Rixey, Joe Sewell, Nip Winters
1940 -- John Beckwith, Burleigh Grimes, Joe Judge, Dolf Luque, Herb Pennock, Sam Rice
1941 -- Sad Sam Jones, Pie Traynor, Dazzy Vance
1942 -- Andy Cooper, Bill Terry
1943 -- Judy Johnson, Dick Lundy

That's 34 players for 8 spots, plus the backlog. That's to say: most of this group will not be gaining election, at least not for quite a while. The backlog is going to be receiving very stiff competition during the 1933-1945 period.

This is the same exact thing that Pearce did as a pioneer at short. Eventually, other shortstops were able to duplicate what Pearce created, but that doesn't mean that we negate the advantage (and wins) created by his innovations.
--
Show me the wins he created. Show me a clipping that say "Pearce, playing a different kind of shortstop, gave us a huge advantage for 10 years until everybody else copied thim". I need that before I credit him.

First, I didn't say Pearce had a ten-year advantage over all shortstops due to his creating the position, but all you have to is read his contemporaries' views about him to know that he knocked their socks off. The overall opinion was that he was clearly the best overall at short until George Wright makes the scene.

IMHO, Babe Ruth would still be Barry Bonds today.
But Dickey Pearce would be....better than Phil Rizzuto, but not Ozzie Smith.

Since Rizuuto was a great player in his own right and (at the very least) is a borderline HoMer, I don't consider that a real insult for Pearce. Pearce was still more durable, though.

At the risk of re-hashing an old argument....
"Best" is REALLY relative. There is probably some tribe who has a best javelin thrower. And if a contest were held today, he would win. Because most of the rest of us
a. don't practice this sport much , and
b. won't bother showing up for the contest.
Put a ten million dollar prize on it and promote it for fall of 2006, you'll get some better competition. Mister tribesman won't look so hot. Not because of timelining, but because his current definition of "best" wasn't quite robust.

When every country fields championship baseball teams, I'll be waiting to hear from you about all the players we should remove from the HoM because of competition concerns. Even though they played the "best" from their times, we know "best" is a relative term, right? :-D

...we need to be careful to give the new arrivals full and fair consideration.

For instance, just looking at pitchers: I currently have Waddell and Welch as my two highest ranked pitchers. I would place both Vance and Coveleski ahead of Waddell as good-rate, shorter-career candidates, with Babe Adams just behind Waddell. Some of the career cases - Rixie, Faber, Ruffing - I see as very similar in overall value as well. There are also 3 or 4 pitchers on Chris's list that I haven't really looked at yet and need to.

Groh, Wheat, Sisler, Traynor - I'll take them seriously. And we have the seriously perplexing (for vastly different reasons) cases of John Beckwith and Rabbit Maranville.

That's 34 players for 8 spots, plus the backlog. That's to say: most of this group will not be gaining election, at least not for quite a while. The backlog is going to be receiving very stiff competition during the 1933-1945 period.

Its going to happen gradually, with two weeks of debate per year. Just looking back at the 1916 results and Childs, Welch, and Pearce were way down the list and Duffy & Waddell have failed to climb.

Hard to predict what will happen.

Warning: rant coming up!

Will some of you (you know who you are) STFU with the "remove people from the HoM" argument. It aint going to happen, we set up the system differently, and no one has argued that the 210? players we have at the end are supposed to be the 210 most meritorious. If we had wanted that we would have had one vote, not voted over the course of time.

Some of you make very good arguments defending a position and then throw it away with that BS. Personally, I ignore the rest of the arguments from anyone who makes that one, even though that isnt logical.

End of Rant, have a happy weekend.

Will some of you (you know who you are) STFU with the "remove people from the HoM" argument.

No, I don't. My post was in no way advocating that at all. In fact, I would leave the project if that were the case. My point was to show the hole in Tom's argument. Until someone can point a hole in my argument, I'll keep using it and won't "STFU," too. Please feel free to ignore my later "BS."

BTW, unless someone brings up an entirely new point in the next few weeks (which hasn't happened so far), there will be no more defenses of Pearce from now on. He's on his way in regardless, so it's silly to keep rehashing the same arguments over and over.

Will some of you (you know who you are) STFU with the "remove people from the HoM" argument.

Inappropriate.

KJOK's comparison of Donaldson to Rube Waddell in post 90 seems to me to be apt in terms of their _relative_ dominance at their peaks and in terms of the shortness of those peaks. Donaldson's life was very different from Waddell's, of course. I also think KJOK's ranking of Donaldson relative to contemporary negro-league pitchers is about right.

Just want to clarify that it wasn't my comparision, but the article that I posted in the HOM egroup from the LA Times in 1917 specifically referred to Donaldson as the "black Waddell".

We could I suppose turn this project into one in which we never spoke to each other again, by holding a "sudden death" un-election in 2007, in which everybody voted for the HOM members they thought most spurious and the top 3 were booted out. Intellectually, this would be a fine way to correct mistakes and remove HOMers who slipped through in a down year. In terms of human relations, even in cyberspace, maybe not.

Two left fielders:

Jimmy Sheckard: 2122 games, 120 OPS+, 339 Win Shares, 130.7 WARP1, 93.5 WARP3.

Zack Wheat: 2410 games, 129 OPS+, 380 Win Shares, 123.9 WARP1, 85.0 WARP3.

If Sheckard gets in the HoM before 1933, shouldn't Wheat be a first-ballot HoMer in 1933? Discuss.

If Sheckard gets in the HoM before 1933, shouldn't Wheat be a first-ballot HoMer in 1933? Discuss.

Not necessarily... that's not how the project works. Once you are in, you are in. Wheat will not be running against Sheckard, he will be running against whoever is on the ballot at the time.

If they were on the ballot at the same time, I would probably rate Wheat higher, yes. But it looks like they won't be.

I imagine it depends on the level of support Groh will get.

We could I suppose turn this project into one in which we never spoke to each other again, by holding a "sudden death" un-election in 2007, in which everybody voted for the HOM members they thought most spurious and the top 3 were booted out.

Dickey Pearce, Bob Caruthers, and Joe Start.

*runs for his life*

everybody voted for the HOM members they thought most spurious and the top 3 were booted out.

Morgan Bulkeley
Tommy McCarthy
Jesse Haines

What's that? Oh, you meant HOM, not Hall of Fame. Until we make a selection as remotely bad as these three, let's not talk about removing anyone from the Hall of Merit.

DavidFoss wrote:

Hard to predict what will happen.

Amen!

However, that surely won't stop us from making predictions :-) .

EricC wrote:

If Sheckard gets in the HoM before 1933, shouldn't Wheat be a first-ballot HoMer in 1933? Discuss.

As DavidFoss said, it depends. I think it likely that Wheat will be elected in 1933, but Sheckard isn't making 13 ballots or something like that, and WARP sees Sheckard as better than Wheat, so support for Wheat might be mixed.

In my view, Wheat was a great player, _slightly_ better than Sheckard, and will probably be at #2 on my ballot in 1933, unless I see Santop or Mendez as better (haven't finished analyzing Santop or started analyzing Mendez), but they don't get elected in 1932.

It's likely Wheat makes it in 1933, but with the ballot being so fragmented already, who knows what it might look like after another 3 elections.
Also, it seems like a lot of us are rethinking how we rank 1890s players so GVH or Becks or someone else could move up in 33 and Wheat will have to wait until after the 1934 -37 explosion.

who knows what it might look like after another 3 elections.

I've got a pretty good idea what the top spot will look like. Beyond that - for all I know Jack Chesbro could get votes.

Also, it seems like a lot of us are rethinking how we rank 1890s players so GVH or Becks or someone else could move up

Looking back at the first election I took part in (1915) it's amazing how far Hugh Duffy's fallen. I don't think there's anybody who finished ahead of him on that list that ain't in yet - & he topped Caruthers, Beckley, GVH, Ryan . . .

Just got back from a minor league game. Anyone have a theory why minor league umpires are relatively small, relative to the major league counterparts?

A lower per diem maybe?

it's amazing how far Hugh Duffy's fallen

Just comparing the 1913 election, where Duffy finished in 5th with 1929 where he finished 17th.

McVey 2 to election
Stovey 3 to election
Bennett 4 to election
Thompson 6 to election
Grant 7 to election
Ryan 8 to 11
Beckley 9 to 7
Pike 10 to 6
Van Haltren 11 to 8
Jennings 12 to 15
Pearce 13 to 5
Childs 15 to 14
Caruthers 19 to 4
Welch 23 to 13

Others who became eligible and were elected:
Dahlen and Davis 1915
Young and Clarke 1917
Flick 1918
Keeler and Kelley 1919
Walsh 1920
JCollins 1921
Lajoie and Matty 1922
Wagner 1923
Crawford and Plank 1924
HR Johnson and TF Brown 1925
Magee 1926
Jackson and Hill 1927
Baker and McGinnity 1928
Wallace 1929

22 players have become eligible and elected since Duffy had reached number 5.
The 3 players ahead of him in 1913 are in and 2 behind him have been elected.

Did Duffy pass gas in an elevator full of HoM voters that got stuck between floors?
Different standards? Different analysis?

Note that Ryan and Van Haltren aren't doing much, either. We have a lot of OFs from that era, and most of us feel like we have enough. I was never a big Duffy fan anyway; a fluke year, not an incredibly lengthy career, etc.
Seems like he needed fielding for a boost, and I ain't giving big fielding credits to an 1890s OF. No one alive ever saw him play.

Now if you wonder how Sherry Magee makes it right away while Duffy now struggles to get a 15th-place vote, well, so do I. Later era, better player, yes, but geesh.

And of course, that elevator incident hardly helped, nor his choice of onion-bean garlic soup for lunch that day.

Well, I accidentally overwrote my excel file I was using as a database. However, I'm choosing to look at this as a positive; I was growing distisfied with my specific methodology. Mainly, while I am a career voter, I do try to distinguish between guys who pile up an extra X WS or Y WARP3 just by hanging on and guys who put up the same career totals without so much in the way hang-on tallies. I view this as an opportunity to pump out a new method (one I suspect, though I don't know, will favor the shorter career/big peak types more than I was), so look for major changes on my ballot this year.

I've just finished doing my own WARP calculations for this ballot. Maybe all this should be put on another thread, as I'm going to be unloading a lot of material on the group here, so if someone wants to move my next bunch of posts to their own thread, please do.

METHODOLOGY
Position players:
Batting:
I used eXtrapolated Runs/XR (1B = .5 runs, 2B = .72, 3B = 1.04, HR = 1.44, BB/HBP = .33, SB = .18, CS = -.32, SO = -.098, outs on balls in play = -.09) to calculate batting runs and batting runs per 27 outs for a player, and batting runs per 27 outs for the league as a whole. I worked backwards from positional equivalent average data (obtained from Max Parkinson) to get a positional XR/27 rate. Using games played at position data, I calculated what fraction of the season a player played at each position and used a weighted average to get the league average XR/27 for his position (e.g., if a player played 75% of the season at catcher and 25% at first base, and a league average catcher had 3 XR/27 and a league average first baseman had 5 XR/27, his positional average XR/27 would be 3.5 XR/27). I then subtracted out the player’s own contribution to this positional average by using Positional XR/27*# of teams in the league, minus the player’s XR/27 times the fraction of the season he played, divided by the number of teams in the league minus one plus the fraction of the season he didn’t play. This gives me the average XR/27 rate for players at his (weighted) position, excluding himself (PXR27’). A .300 winning percentage replacement team will score .1515*League Average Runs Scored – 50 runs fewer than the league average, and allow .1515*League Average Runs Scored + 50 runs more than the league average, so I used .8485*PXR27’ – (50/162) as the replacement positional XR/27. Then I calculated how many XR/27 the rest of the team would produce by using ((LgXR/27*9)-(PXR/27))/8, since the rest of the team doesn’t include another player at his position.


Fielding: A .300 replacement team will allow .1515*LgXR/27 + 50 runs above average. Using an arbitrary 60/40 pitching/fielding split, I credited a league-average team’s fielders with .0061*LgXR/27 + 20 fielding runs above replacement. Using the Baseball Prospectus all-time position weights (catcher 39 runs out of a total 219, etc), I divided these FRAR among the positions and credited a player with fielding runs equal to his Baseball Prospectus FRAA plus his position’s fraction of these FRAR.


I then imagined a team that was league-average except for the player in question, and compared it to a team that was league-average except for a replacement player with the same number of outs. I determined how many runs each team would score and allow using the league average XR/27, the player’s XR, and the XR/27 rate of a replacement player at his position. I used the Pythagenport formula to derive winning percentages for these two teams, and from this I obtained WARP.

Example: Pete Browning, 1882 American Association produced 8.59 XR/27. His league generated 2.95 XR/27 (291 XR+). He played 57.5% of his games at second base, 17.8% at third base, and 24.7% at shortstop. The average AA second baseman in 1882 produced 2.49 XR/27, the average third baseman 3.44, and the average shortstop 2.85, so his positional average was (.575*2.49)+(.178*3.44)+(.247*2.85) = 2.75 XR/27. He played in 69 of his team’s 84 games, or 82.1%. There were six teams in the AA that year. Thus, the positional average excluding him is ((2.75*6)-(8.59*.821))/(5+.179) = 1.82 XR/27 (a big drop!). A .300 WPCT replacement player at his position would generate 1.82-(.1515*1.82)-.309 = 1.23 XR/27. The rest of the team would score ((2.95*9)-2.75)/8 = 2.98 XR/27.
Browning made 179 outs, with his teammates making the other (84*27)-179 = 2089 outs. The rest of the team would thus score 2.98*(2089/27) = 231 runs. Browning produced 57 XR. A replacement player at his position(s) would have produced 1.23*(179/27) = 8 runs. Thus, a team with Browning would score 231+57=288 runs, while a team with a replacement player at his position would have scored 231+8=239 runs. It was an 84 game season, so in order to make comparisons across season lengths, I normalize to 162 games: Average Team Plus Browning scores 288*162/84=555 runs, Average Team Plus Replacement Player At Browning’s Position scores 239*162/84=461 runs.
Now on to the runs allowed. An 1882 AA league-average team playing 162 games would allow 2.95*162 = 478 runs. This league-average team plus Browning would allow 478 minus his BP Fielding Runs Above Average times a season-length adjustment (in this case, 162/84). In fact, Browning had zero FRAA that year, so the team with Browning scores 555 runs and allows 478. (If he had had, say, 4 FRAA, the team with Browning would allow 478-(4*162/84) = 470 runs).
A team with replacement pitchers and fielders would allow (.1515*478) + 50 = 122 runs more than the average team. Using a 60/40 pitching/fielding split, I allot 122*.4 = 49 runs to the fielders. Baseball Prospectus has second basemen accounting for 15.5% of league average fielding runs above replacement, third basemen 11.9%, and shortstops 17.4%. .575*.155 + .178*.119 + .247*.174 = .153, so a league-average fielder at Browning’s position(s) would be .153*49=7 runs better than a replacement fielder per 162 games (how BP manages to credit average SS with 39 R above replacement per 162 eludes me). Thus, the league average team with a replacement fielder at Browning’s position would allow 478 + 7 = 485 runs, scoring scoring 461.
The last step is to convert this to WARP. The league had 2.95 XR/G, so its Pythagorean exponent is .45 + 1.5*log10(2.95*2), which is 1.61. The Browning team scores 555 runs and allows 478, for a winning percentage of .560. The team with a replacement player in Browning’s stead scores 461 runs and allows 485, for a winning percentage of .480. Over 162 games, a .560 WPCT is 90.7 wins, and a .480 WPCT is 77.8 wins, so Browning is 12.9 wins above replacement.

Pitchers:
Pitchers’ hitting and fielding are the same as batters, although I am considering a replacement-fielding pitcher to be a league-average fielder (FRAA-FRAR = 0).
To derive an XR/27 allowed for pitchers (ERA), I used a modified DIPS principle. I took pitchers’ K, BB, HR, and HBP totals without adjustment. To calculate their BABIP, I took the ratio of their BABIP to that of their teammates (to remove the influence of defense) and reduced the random wide fluctuations in this ratio due to luck by weighting the ratio in a given year ¼, the ratio over the three-year period surrounding it ¼, the ratio over a five-year period surrounding it ¼, and the career ratio ¼. Then I multiplied this smoothed ratio by the league BABIP to get a pitcher’s defense-independent BABIP and thus the number of non-HR hits and outs. Then I just used the XR equation to get XR allowed per 27 (a DIPS ERA, effectively). Again using the 60/40 pitching/fielding split, I defined a replacement pitcher’s XR allowed/27 as League XR/27 + .6*(.1515*LgXR/27 + .309), which simplifies to 1.0909*LgXR/27 + .185. From here I proceeded the same as for batters.

To show how all of this works in tandem, let’s look at the combo man himself, Parisian Bob Caruthers, in the 1887 American Association, where the league XR/27 was 5.19. Here’s his stat line:


Batting:

 AB   H  2B 3B HR SB CS BB HBP SO
364 130  23 11  8 49  9 66   6 30

Fielding:
50 games in RF, 39 at P, 7 at 1B, 3 in LF, 2 in CF. 15 Fielding Runs Above Average.


Pitching (TBIPA is his team’s batting average allowed in balls in play when he was not pitching)

 BFP   H HR BB SO HBP TBIPA
1442 337  6 61 74  16  .280

First, let’s look at his hitting. The stat line translates into 10.97 XR/27 (211 XR+), in 234 outs in a 135 game season. Caruthers played 49.5% of his games in RF, 38.6% at P, 6.9% at 1B, 3.0% at LF, and 2.0% at CF. 1887 American Association right fielders generated 5.71 XR/27, pitchers 3.68, first basemen 6.20, left fielders 7.08, and center fielders 6.60, so Caruthers’ positional average XR/27 is (.495*5.71)+(.386*3.68)+(.069*6.2)+(.03*7.08)+(.02*6.6), which is 5.02. Caruthers played in 98 of his team’s 135 games that year, 72.6%, so the positional average excluding him is ((5.02*8)-(.726*10.97))/(7+.274) = 4.43 XR/27. A .300 WPCT replacement player at this hybrid position would generate 4.43-(.1515*4.43)-.309 = 3.44 XR/27, and the rest of the team will produce ((9*5.19)-5.02)/8 = 5.21 XR/27.
Caruthers made 234 outs, with his teammates making the other (135*27)-234 = 3411 outs. The rest of the team would thus score 5.21*(3411/27) = 658 runs. Caruthers produced 95 XR. A replacement player at his position(s) would have produced 3.44*(234/27) = 30 runs. Thus, a team with Caruthers would score 658+95 = 753 runs, while a team with a replacement player at his position would have scored 658+30 = 688 runs, which after season length adjustment comes to 904 runs scored for Average Team Plus Caruthers, and 825 runs scored for Average Team Plus Replacement Player.

Now, the fielding. An 1887 AA league-average team playing 162 games would allow 5.19*162 = 841 runs. This league-average team plus Caruthers the fielder would allow 841 minus his BP Fielding Runs Above Average (15) times a season-length adjustment (in this case, 162/135), giving us 823 runs allowed for the Caruthers team.

A team with replacement pitchers and fielders would allow (.1515*841) + 50 = 177 runs more than the average team. Using a 60/40 pitching/fielding split, I allot 177*.4 = 71 runs to the fielders. Baseball Prospectus has right and left fielders each counting for 9.1% of league average fielding runs above replacement, pitchers 0%, first basemen 5.5%, and center fielders 13.7%, so Caruthers accounts for (.495*.091)+(.386*0)+(.069*.055)+(.03*.091)+(.02*.137) = .054, or 5.4% of those fielding runs above replacement. A league-average fielder at Caruthers’ position(s) would be .054*71 = 4 runs better than a replacement fielder per 162. Thus, the league average team with a replacement fielder at Caruthers’ position would allow 841 + 4 = 845 runs.

Ah yes, but the man could pitch, too! Let’s get his pitching into the equation. Caruthers faced 1,442 batters, striking out 74, walking 61, hitting 16 batters, and surrendering 6 home runs. None of those events involved his fielders, so we’ll give him full credit for those. The remaining 1,285 batters put the ball in play, thus bringing defense into the equation. Caruthers allowed a .258 batting average on balls in play (BABIP), while the other pitchers on his team gave up .280, a ratio of (.258/.280) = .919. However, in the 19th century and deadball eras as well as now, there were great year-to-year fluctuations in BABIP that resulted overwhelmingly from luck, so I’ve taken some of the noise out of the data by looking at seasons in context. For his career, Caruthers’ allowed BABIP was 96% of his teammates’. In 1885 and 1886 it was 103% and 92% of his teammates’, and in 1888 and 1889 it was 101% and 94%. Weighting for the number of batters he faced each year, and then weighting 1887 ¼, the 1886-88 average ¼, the 1885-1889 average ¼, and his career average ¼, I got a revised, “smoothed” ratio of Caruthers BABIP = 94.9% of his teammates’ BABIP. Then taking out the effects of defense and park, I multiplied the league BABIP of .295 by .949 to get Caruthers’ defense-independent BABIP of .280 (coincidentally exactly what his teammates allowed, but this is entirely random). 1,285 balls in play times a .280 BABIP makes 360 hits on balls in play. I then plug the revised stat line into XR, and get 147 XR allowed in 333 innings, for a dERA of 3.97 (131 dERA+).


Now let’s put it all together. Caruthers pitched 28.5% of his team’s innings, so let’s take our league average team giving up 5.19 runs a game. Plug in Caruthers for 28.5% of the innings, and you get (.715*5.19)+(.285*3.97) = 4.84 runs allowed per game, times 162 games makes 784 runs. Now remember that Caruthers the Fielder saved another (841-823) = 18 runs and subtract that, and you get 766 runs allowed for the Caruthers Team (along with 904 runs scored).
Using the 60/40 pitching/fielding split again, 60% of the runs a league average team saves versus a replacement team are due to pitching. 60% of 177 is 106 runs. Caruthers pitched 28.5% of the team’s innings, so a replacement pitcher would have allowed (106*.285) = 30 runs more than an average pitcher during Caruthers’ innings. Add those 30 runs to the 845 established in the fielding section and you get the replacement team allowing 875 runs.
Finally, we’re ready to calculate WARP. The Average Team Plus Caruthers scores 904 runs and allows 766, while the Average Team Plus Replacement Player scores 825 runs and allows 875. The league XR/27 was 5.19, so the Pythagorean exponent is .45 +(1.5*(log10(5.19*2))) = 1.97. The Caruthers team has a winning percentage of (904^1.97)/((904^1.97)+(766^1.97)), or .581, while the replacement team has a winning percentage of (825^1.97)/((825^1.97)+(875^1.97)), or .471. .581*162 is 94.1 wins for the Average Team Plus Caruthers, while .471*162 is 76.3 wins for the Average Team Plus Replacement Player, making Caruthers an eyepopping 17.8 wins above replacement.

Below are the WARP and dERA+ or XR+ scores for everyone with available ML stats who got at least 100 points in the 1929 election, plus a newcomer. These numbers are not adjusted for park or for league quality.

Bob Caruthers
Year Lg WARP dERA+ XR+
1884 AA  2.9  120  134
1885 AA 10.9  110   95
1886 AA 16.1  125  251
1887 AA 18.0  131  211
1888 AA 13.3  137  125
1889 AA 11.5  119  139
1890 NL  7.5  114  135
1891 NL  5.5  107  130
1892 NL  2.3   86  145
1893 NL  0.5  N/A  149

TOTAL   88.5
TOP 5   69.7


Clark Griffith
Year Lg WARP dERA+
1891 AA  1.8    97
1893 NL  0.3   120
1894 NL  6.3   120
1895 NL  6.1   108
1896 NL  8.3   127
1897 NL  8.9   129
1898 NL 10.5   150
1899 NL 11.3   150
1900 NL  8.1   144
1901 AL  9.6   140
1902 AL  2.5   104
1903 AL  4.3   120
1904 AL  0.8   107
1905 AL  2.6   152
1906 AL  0.9   112
TOTAL   82.3
TOP 5   48.6

Jimmy Sheckard
Year Lg  WARP XR+
1897 NL  0.1  160
1898 NL  0.8  119
1899 NL  7.5  137
1900 NL  2.9  167
1901 NL  8.0  203
1902 NL  6.1  137
1903 NL 11.4  195
1904 NL  1.1  105
1905 NL  5.1  153
1906 NL  5.3  136
1907 NL  4.1  146
1908 NL  1.6  124
1909 NL  2.8  123
1910 NL  4.1  127
1911 NL  7.9  162
1912 NL  3.6  119
1913 NL  0.6  100
TOTAL   73.2
TOP 5   40.9

Pete Browning
Year Lg WARP XR+
1882 AA 12.9 291
1883 AA  6.8 201
1884 AA  5.5 201
1885 AA 11.9 233
1886 AA  5.0 181
1887 AA 10.2 223
1888 AA  4.5 185
1889 AA -0.4 101
1890 PL  7.0 186
1891 NL  4.1 153
1892 NL  2.9 143
1893 NL  1.5 167
1894 NL  0.0  88
TOTAL   72.0 
TOP 5   48.8

Vic Willis
Year Lg WARP dERA+
1898 NL  3.0   106
1899 NL  6.5   123
1900 NL  1.2    98
1901 NL  6.2   123
1902 NL  8.6   127
1903 NL  5.6   126
1904 NL  5.8   116
1905 NL  4.6   110
1906 NL  6.6   130
1907 NL  6.6   135
1908 NL  6.8   138
1909 NL  5.4   127
1910 NL  3.8   122
TOTAL   70.7
TOP 5   35.1

Cupid Childs
Year Lg WARP XR+
1890 AA 10.8 201
1891 NL  5.1 147
1892 NL  9.9 192
1893 NL  7.1 163
1894 NL  5.9 145
1895 NL  4.1 108
1896 NL 11.0 167
1897 NL  8.5 148
1898 NL  3.5 131
1899 NL  0.7 108
1900 NL  1.5  79
1901 NL  2.4 106
TOTAL   70.4
TOP 5   47.4

Addie Joss
Year Lg WARP dERA+
1902 AL  8.6   154
1903 AL  9.3   161
1904 AL  4.4   156
1905 AL  7.7   149
1906 AL  8.1   156
1907 AL  9.1   156
1908 AL 12.7   205
1909 AL  6.6   175
1910 AL  3.0   173
TOTAL   69.4   
        47.8

Charley Jones
Year Lg  WARP XR+
1875 NA  -0.1 136
1876 NL   0.7 147
1877 NL   6.1 175
1878 NL   6.1 165
1879 NL  17.3 224
1880 NL   4.1 175
1883 AA   9.8 173
1884 AA  10.5 217
1885 AA  10.9 188
1886 AA   3.9 140
1887 AA   0.1 104
1888 AA  -0.4  40
TOTAL    69.1
TOP 5    54.7

Jimmy Ryan
Year Lg WARP XR+
1885 NL  0.1 343
1886 NL  4.0 137
1887 NL  4.9 132
1888 NL 11.4 222
1889 NL  9.0 162
1890 PL  6.0 150
1891 NL  4.8 139
1892 NL  5.6 161
1893 NL  1.8 130
1894 NL  3.7 125
1895 NL  2.8 123
1896 NL  1.9 114
1897 NL  3.4 123
1898 NL  4.1 165
1899 NL  1.2 112
1900 NL -0.6 105
1902 AL  3.1 143
1903 AL  0.9 103
TOTAL   68.2
TOP 5   36.9

Eddie Cicotte
Year Lg WARP dERA+
1905 AL  0.0    85
1908 AL  2.5   108
1909 AL  2.1   112
1910 AL  2.8   106
1911 AL  3.2   115
1912 AL  3.6   119
1913 AL  6.4   134
1914 AL  4.7   119
1915 AL  4.0   123
1916 AL  3.8   132
1917 AL 10.1   167
1918 AL  8.8   153
1919 AL  9.6   165
1920 AL  5.7   125
TOTAL   67.2   130
TOP 5   40.6

Rube Waddell
Year Lg WARP dERA+
1897 NL  0.3   118
1899 NL  1.8   135
1900 NL  5.9   149
1901 NL  6.1   127
1902 AL  9.8   161
1903 AL  8.6   154
1904 AL  8.4   146
1905 AL  8.1   153
1906 AL  5.3   129
1907 AL  4.4   132
1908 AL  5.0   135
1909 AL  2.7   128
1910 AL  0.2   113
TOTAL   66.6   141
TOP 5   41.1

Tommy Leach
Year Lg WARP XR+
1898 NL -0.3 -21
1899 NL  2.0 111
1900 NL  0.1 72
1901 NL  5.9 135
1902 NL  8.6 150
1903 NL  3.4 134
1904 NL  6.7 112
1905 NL  2.8 103
1906 NL  3.0 125
1907 NL  6.4 161
1908 NL  6.7 138
1909 NL  3.9 132
1910 NL  1.6 102
1911 NL  1.9  92
1912 NL  2.2 115
1913 NL  4.8 151
1914 NL  4.5 129
1915 NL  2.0 103
1918 NL  0.2 124
TOTAL   66.4
TOP 5   34.2

Hughie Jennings
Year Lg  WARP XR+
1891 AA   2.3 107
1892 NL   0.9  70
1893 NL  -1.9  33
1894 NL   8.2 124
1895 NL  14.6 169
1896 NL  14.0 193
1897 NL  10.7 185
1898 NL  10.8 190
1899 NL   2.3 151
1900 NL   1.4 107
1901 NL   0.3 114
1902 NL   1.2 112
TOTAL    64.9
TOP 5    58.4

Jake Beckley
Year Lg  WARP XR+
1888 NL   0.7 172
1889 NL   0.9 119
1890 PL   2.7 144
1891 NL   5.2 129
1892 NL   2.7 103
1893 NL   5.0 127
1894 NL   3.8 127
1895 NL   4.6 125
1896 NL   2.1 107
1897 NL   2.7 137
1898 NL   2.3 123
1899 NL   6.7 147
1900 NL   4.4 141
1901 NL   4.5 129
1902 NL   5.3 167
1903 NL   4.4 152
1904 NL   5.2 160
1905 NL   0.7 117
1906 NL  -0.4  90
1907 NL  -0.7  41
TOTAL    62.6
TOP 5    26.6

George Van Haltren
Year Lg WARP XR+
1887 NL -0.2  64
1888 NL  4.1 158
1889 NL  5.4 152
1890 PL  4.3 147
1891 AA  6.0 168
1892 NL  4.5 158
1893 NL  5.0 144
1894 NL  2.5 113
1895 NL  3.6 144
1896 NL  6.3 149
1897 NL  5.4 124
1898 NL  3.7 141
1899 NL  3.3 118
1900 NL  3.4 129
1901 NL  4.9 157
1902 NL  1.0 147
1903 NL -0.9  89
TOTAL   62.4
TOP 5   28.1

John McGraw
Year Lg WARP XR+
1891 AA -0.3 118
1892 NL  1.5 123
1893 NL  5.4 160
1894 NL  6.2 143
1895 NL  8.3 172
1896 NL  1.0 150
1897 NL  5.0 167
1898 NL 10.0 205
1899 NL 11.3 270
1900 NL  6.2 215
1901 AL  4.8 240
1902     1.7 159
1903 NL  0.2 170
1904 NL  0.3 201
TOTAL   61.6
TOP 5   41.9

Frank Chance
Year Lg WARP XR+
1898 NL  0.6 113
1899 NL  1.5 106
1900 NL  1.9 145
1901 NL  1.6 145
1902 NL  2.8 177
1903 NL  8.5 198
1904 NL  6.9 184
1905 NL  7.2 217
1906 NL  9.1 216
1907 NL  6.2 176
1908 NL  5.0 143
1909 NL  2.0 134
1910 NL  2.8 151
1911 NL  0.8 170
1912 NL  0.1 180
1913 AL  0.3 119
TOTAL   57.4
TOP 5   37.9

Hugh Duffy
Year Lg WARP XR+
1888 NL  2.8 136
1889 NL  2.0 127
1890 PL  8.3 145
1891 AA  8.0 181
1892 NL  4.1 152
1893 NL  5.1 151
1894 NL  8.6 216
1895 NL  3.9 150
1896 NL  1.2 110
1897 NL  5.3 149
1898 NL  4.5 129
1899 NL  0.3  99
1900 NL  0.9 126
1901 AL  1.7 122
1904 NL  0.4 195
1905 NL  0.1 113
TOTAL   57.2
TOP 5   35.3

Del Pratt
Year Lg WARP XR+
1912 AL  5.1 126
1913 AL  3.4 128
1914 AL  5.6 135
1915 AL  5.3 116
1916 AL  6.4 112
1917 AL  2.3 100
1918 AL  4.6 112
1919 AL  7.0 116
1920 AL  5.6 120
1921 AL  3.4 118
1922 AL  3.2 110
1923 AL  1.7 117
1924 AL  0.9  96
TOTAL   54.4
TOP 5   29.8

Roger Bresnahan
Year Lg  WARP  XR+
1897 NL   0.2  121
1901 AL   0.8   98
1902      1.9  127
1903 NL   6.0  204
1904 NL   4.5  166
1905 NL   4.7  163
1906 NL   6.4  182
1907 NL   4.3  160
1908 NL   6.2  177
1909 NL   2.6  126
1910 NL   2.7  158
1911 NL   2.9  154
1912 NL   2.0  160
1913 NL   0.8   92
1914 NL   3.3  154
1915 NL   1.4   94
TOTAL    50.6
TOP 5    27.8

Lip Pike
Year Lg WARP XR+
1871 NA 10.9 199
1872 NA  4.0 133
1873 NA  7.2 141
1874 NA  8.9 200
1875 NA 10.0 222
1876 NL  8.4 186
1877 NL  0.5 142
1878 NL  0.1 146
TOTAL   50.0 
TOP 5   45.3

Mike Tiernan
Year Lg WARP XR+
1887 NL  2.3 125
1888 NL  6.9 186
1889 NL  7.3 193
1890 NL  5.6 173
1891 NL  5.8 178
1892 NL  2.5 146
1893 NL  3.8 142
1894 NL -0.1  93
1895 NL  4.0 160
1896 NL  6.3 178
1897 NL  3.6 140
1898 NL  1.3 131
1899 NL -0.8  75
TOTAL   48.4
TOP 5   31.9

Gavvy Cravath
Year Lg  WARP XR+
1908 AL   1.6 158
1909 AL   0.4 133
1912 NL   3.2 130
1913 NL   7.5 194
1914 NL   6.7 190
1915 NL  11.6 209
1916 NL   5.7 172
1917 NL   5.6 173
1918 NL   0.5 117
1919 NL   4.3 263
1920 NL   0.6 179
TOTAL    47.6
TOP 5    37.0

Mickey Welch
Year Lg WARP dERA+
1880 NL  8.7   103
1881 NL  1.3   101
1882 NL -2.0    85
1883 NL  2.8    98
1884 NL  7.3   105
1885 NL  5.3   104
1886 NL  4.0    99
1887 NL  7.3   123
1888 NL  4.0   107
1889 NL  3.7   104
1890 NL  4.2   111
1891 NL -1.2    85
1892 NL -0.1    64
TOTAL   45.4   103
TOP 5   32.9

Larry Doyle
Year Lg WARP XR+
1907 NL -1.2  96
1908 NL  2.0 161
1909 NL  3.9 160
1910 NL  3.8 145
1911 NL  4.9 176
1912 NL  4.4 153
1913 NL  2.7 134
1914 NL  1.2 120
1915 NL  4.3 158
1916 NL  5.9 132
1917 NL  4.0 115
1918 NL  2.4 134
1919 NL  4.4 148
1920 NL  0.6 119
TOTAL   43.3
TOP 5   24.0

After unloading all that, here's my preliminary 1930 ballot. I will figure out how to weigh my own data against BP WARP, WS, and Ink by 1931, but since I put all the time into this I think I'm just going to go by these. I may be setting a record here for change in one ballot to the next.

Tier One (75+ career WARP)

1. Bob Caruthers (88.5 career WARP, 69.7 peak)
Further analysis leads me to conclude that the fact that Caruthers has not yet made it in is a travesty. Not only does he have by far the best peak value on the ballot (at the Ruth, Bonds, Wagner level), he has the best career value on the ballot. 88.5 wins above replacement for his career puts him below Plank (96.1) but ahead of Three-Finger Brown (84.8). An unprecedented and unparalelled dominator. Adjusting for season length, he's also tops on the ballot in CAREER Win Shares (just nudging out GVH) and obviously blows everyone away in peak Win Shares. The closer I look, the more Caruthers looks to me like an *inner-circle* HoM'er. If he doesn't get in during this backlog I'm gonna vomit.

2. Clark Griffith (82.3 career WARP, 48.6 peak)
This is why I spend the time on this stuff--Griffith didn't even make my ballot when I started, and now he's in an elect-me spot. ERA+ makes it seem that Griffith had one dominant year in 1898 and was just above average elsewhere. In fact, he was just as good in 1899 (look at K, BB, HR, and BABIP/Teammates' BABIP), was a reliable workhorse, and pitched at an All-Star level for a decade. You can't see his greatness on the surface, but look deeper into the numbers and from 1896-1901 he was a genuine superstar. His 82.3 career WARP are a big jump ahead of anyone else (Sheckard's next at 73.2). After Parisian Bob gets elected, I'll be the new Best FOCG. Expect to hear a lot more from me about Griffith in the years to come.

Tier Two (65-75 career WARP, 45+ peak)

3. Cupid Childs
Once again, someone I scorned gets massively reevaluated. Being "the best 2B of the 1890's" is not enough, but he was an offensive juggernaut at a scarce position with often excellent leather for eight years. A bona fide superstar in '90, '92, and '96, and a strong All-Star in '93 and '97. Didn't play forever but so good that he accumulated comparable career value to Sheckard, more than the "career guys" like GVH/Beckley. Since we don't have anyone from his era at his position, and because he played in a stronger league than the other guys on the ballot with similar WARP totals for career and peak, he gets my no. 3 vote.

4. Charley Jones (69.1 career WARP, 54.7 peak)
Well, I'm a little dubious on BP's crediting Jones with 27 FRAA in the 85-game 1879 season. But even if he wasn't quite as good as that would suggest--17.3 WARP that year is the second-best season I've calculated, better by a good bit than Ty Cobb's best (13.8 in 1917)--he was still a force, making mincemeat of the 1883-85 AA along with Browning. Everyone in this group is comparable, I think, so he gets a boost for having one of the greatest seasons in major league history in '79 and for his blacklist years.

5. Addie Joss (69.4 career WARP, 47.8 peak)
What a boomerang! I went from being a good FOAJ, to dropping him from my ballot altogether, and now moving him back up to 5th! Joss had a remarkable ability to prevent hits on balls in play, allowing a BABIP 31 points lower than his teammates' for his career (.238/.269). He had six seasons where he was absolutely one of the best in the biz, including 1908 which was particularly standout. I now believe he was so good, in fact, during his brief career that he too had more career value than Ryan, GVH, Beckley etc. The AA discount for Browning gives him the edge here.

6. Pete Browning (72 career WARP, 48.8 peak)
This guy really needs more support. 1890 showed us he was for real, so his knock-em-dead years in '82, '85 and '87 have to be taken seriously. Like everyone else in this group, more career value than the career guys, and a true dominator for three seasons. Hopefully I can drum up some support for him; he really deserves it.

7. Lip Pike (50 career WARP, 45.3 peak)
Well, clearly he was an upper-echelon superstar from 1871-76. The pre-NA data suggesting he was one of the best players in the second half of the 1860's and the fact that he clearly was at his peak in '71 leads me to credit him with 8 WARP for his best two pre-NA years and 5 WARP for the rest, putting him up to 71 with a 46ish peak, which is squarely in Tier II. Also, gotta support the tribe of Israel here. Go Lipman Go--Hank Greenberg, Sandy Koufax, and Shawn Green smile upon you. So does Doron Sheffer.

8. Hughie Jennings (64.9 career WARP, 58.4 peak)
The last of the high-peak guys. His standing here is very much dependent on FRAA's extraordinarily positive evaluation of his defense (averaging 47 FRAA/130 G in 95-96 is pretty nuts). If there is reason to be skeptical of his defensive brilliance, he'll fall. But if not, he is a Caruthers type--so good for five years that he was more valuable than guys who played for three times as long. 4 years of 170-190 XR+ at the SS position, plus that defense, makes him HoM-worthy to me.

Wow Dan. That's a lot of work. Will take a while to digest. Thanks for posting the numbers.

Stepping back from the data for a second, how does your WARP differ from BP's WARP on a conceptual level? In other words, since your WARP numbers are quite a bit different than BPs WARP numbers (which is why you did the exercise, of course), what factors are you considering or not considering (or weighting differently) that produce the different result?

Also, are your XR numbers park adjusted?

Tier Three (60+ career WARP, 40+ peak)

9. Jimmy Sheckard (73.2 career WARP, 40.9 peak)
A strange career. Much of his value in defense and walks, including an outlandish 147 in 1911, when he hadn't been particularly great in eight years. Add it all up and you get a mid-ballot candidate as far as I'm concerned, just squeaking into the HoM.

10. Rube Waddell (66.6 career WARP, 41.1 peak)
Rube still just makes it past my in/out line, but he's taken a big hit with my reevaluation. I *love* the K's, but now that I can see that deadball pitchers really could prevent hits on balls in play, he stands out less than he did before. It's worth nothing that his 1903 season was just as good as his much more highly regarded '04--almost as many innings, same BB/K/HR rates, similar propensity to giving up line drives (BABIP 5% higher than teammates' in '03, 6% in '04). '02 was really his best season though. One of the best pitchers in baseball from '02-'05, but not an otherworldly dominator and not enough career to push him further up the ballot.
------------------
In-Out Line
------------------
11. Eddie Cicotte (67.2 career WARP, 40.6 peak)
He really was a premier, superstar pitcher from 1917-19, and was serviceable in 1913 and 1920. A slightly above league average pitcher for the rest of his career.

12.John McGraw (61.6 career WARP, 41.9 peak)
He didn't play long enough to make the HoM, and rarely played full seasons even when he did. But man, was he good--an on-base machine the likes of which the game has rarely seen since. Despite the partial seasons, he still was worth over 40 wins above a replacement player in his 5 best years and breaks 60 for his career, which merits a lower-ballot placement.

Tier Four<b> (over 60 WARP, under 40 peak)
13. <b>Jimmy Ryan (68.2 WARP, 36.9 peak)
He really was great for two years (1888-9), and played forever.
14. Vic Willis (70.7 WARP, 35.1 peak)
Just kept churning out those innings at an above-average level. The Beckley of pitchers, but a more valuable career than Beckley and at least a genuine All-Star once or twice.
15. Rube Foster (n/a)
Because I'd rather vote for the best Negro League pitcher of his generation than an ML player I really can't support.

Top 10 Returnees Left Off:

Dickey Pearce: First of all, I don't think he was playing baseball, and secondly, I am very wary of subjective accounts of defensive brilliance. People said he was a good fielding shortstop. People say the same of Derek Jeter. I strongly don't think he deserves to get in. Please ask yourselves how reliable the reports are of his excellence before elevating him ahead of players with real stats. I know he was old, but he was a pretty awful hitter in the NA.

Jake Beckley: He wasn't an All-Star for 15 years. He was a slightly above average player for a long time, accumulating less career value than guys who played half as long. And of course he had no peak. Unless you think that replacement level in his era was so low that just playing in the majors meant a lot to helping a team win, he's really got no case. His best year (1899) was worse than Del Pratt in 1919. Also, yes I really do think he was barely better than a replacement player in 1889, when he hit .301. The ABC boys meant that the positional average XR/27 was 6.43, a notch ahead of his 5.79. Add in his zero defensive value (-3 FRAA) and you get one win above a replacement 1B.

George Van Haltren: People compare him to Sheckard? I see him as much more comparable to Beckley: identical, less-than-stunning career value (62.5ish WARP) and no peak whatsoever.

Jeff M--For hitters, I don't think they are that conceptually different. The big factor is that BP uses a ludicrously low replacement level for fielders. Does anyone think that a league-average hitting and pitching team with AAA fielders would go 61-101? Cause that's what they're suggesting. Also, I don't think they subtract out a player's own contribution to a league positional average, which makes a big difference in an 8-team league (i.e., the positional EqA for 1908 NL catchers would be 10 points lower if you take out Bresnahan). For pitchers, they do try to separate pitching and defense, but not by the method that I've employed (comparison to teammates).
No, they're not park adjusted--how do you do that?

By contrast, my system holds that an average pitching and hitting team with replacement fielders would go 75-87.

Yeah... great work Dan! Maybe you could post them to the yahoo group at some point so that we can browse them more easily.

The XR+ numbers all look high. That might not be bad, because they are all higher and its all relative. But it will take a while to digest the difference between XR+ and OPS+. Did league average hitters have an XR+ of 100?

I can post the spreadsheets there. I'll do that now. You have to download all of them in the same directory for them to work, howver.
Yes, league average hitters have an XR+ of 100--XR+ is just Player XR/27 times 100 divided by League XR/27. Take Browning in 1882, say--he produced 8.59 XR/27 against the league's 2.95. 8.59*100/2.95 = 291.

I'd really pay more attention to the WARP than the XR+, which I calculated just as an afterthought. XR+ doesn't take into account position, or fielding, or how much someone played, and a 200 XR+ in one run scoring environment is very different from a 200 XR+ in another. The WARP number factors in all of this.

No, they're not park adjusted--how do you do that?

I'm not sure where you'd put it in the XR formula, but it may simply be that you multiply the XR for the hitters times the inverse of the Batting Park Factor and multiplying the pitcher's XR times the Pitching Park Factor.

But maybe you'd have to apply the factors to the positional reference point, to put the replacement player in the same ballpark as the player you are evaluating. Otherwise, it seems like a hitter in a run-friendly park will likely show as even better against a replacement player in a neutral park.

The baseball-reference factors are based on Total Baseball's method of calculation. Bill James uses a simpler method, but smooths it out over five years (and there's no publication showing year by year park factors by this method...that I know of). The Lehman database also has park factors, but I can't remember from whence they are derived.

Dan, fantastic stuff, not least because it nails down rigorously and to my eye convincingly the analysis I had done in a sloppy "back of an envelope" way which was that for a brief period, but still several years, Parisian Bob was a very special player indeed, as in Ruth/Cobb/Wagner special.

I particularly like the way your system downgrades the salience of fielding somewhat -- that looks right to me, given what we know about fielding's contribution in the modern game. For example, the HOF elected Rabbit Maranville, and fielding-oriented systems give him high marks, whereas I don't think I'll have him on my ballot.

I wonder why your system doesn't like Beckley, though. To me, Beckley has about the same quality or slightly better as Duffy/Ryan/Van H (who I bracket close together) but then went on for about 20% longer, meaning that with any kind of season length adjustment he was well over 3000 hits. Plus he was a 1B, more valuable than LF and RF in those days, and possibly than CF. Is there a distortion from possibly the 3 greatest hitters of his early career all being 1B? (wasn't true later on, though.)

Do you end up a FO Mickey Welch or an EO Mickey Welch? If the latter, why, and is it the result of a distortion from Keefe being on the same team? (like Drysdale being only 1 win ahead of team for his career, but his team was a pitching juggernaut and included Koufax)

secondly, I am very wary of subjective accounts of defensive brilliance. People said he was a good fielding shortstop. People say the same of Derek Jeter. I strongly don't think he deserves to get in. Please ask yourselves how reliable the reports are of his excellence

If you're going to convince me of the unreliability of subjective accounts of defense, you're going to need to do better than bring out one example. I read this & I'm reminded of the George Sisler comment on p441 of the NBJHA:

"Also, while articles of Sisler invariably claim that he was a fine fielder, the Win Shares system is unable to see that this is true. The Win Shares system does document the defensive excellence of Keith Hernandez, Wes Parker, Bill White, Vic Power, Gil Hodges, Bill Terry, Frank McCormick, Charlie Grimm, and almost all other top defensive first baseman."

The lesson I take from that is that while subjective accounts can be well off, they tend to be pretty good.

Also, from what I remember of John Murphy's campaigning, Pearce wasn't just a good glove, he's the guy who largely invented it, turning what had been an afterthought of a defensive position into the most critical defensive position.

Also, yes I really do think he was barely better than a replacement player in 1889, when he hit .301. The ABC boys meant that the positional average XR/27 was 6.43

Someone can (hopefully) probably verbalize it better than I can, but there seems to be something off about that to me. Didn't Bill James used to figure replacement value based on the performance of the worst players at a position?

Someone can (hopefully) probably verbalize it better than I can, but there seems to be something off about that to me. Didn't Bill James used to figure replacement value based on the performance of the worst players at a position?

Hypothetically, its supposed to be guys that are readily available and you can pick up for almost nothing. AAA players... but not the blue-chip AAA guys who might be MLB ready already but in AAA because of a blockage or a conservative GM... the AAA veterans... the ones left exposed for drafts and the like.

You are right, that the presence of superstars in the league should have little effect on the replacement level, but its so tough to estimate replacemnt for a number of reasons. A few teams may employ below-replacement players as regulars out of loyalty, faith-that-its-just-a-slump, or perhaps extreme youth.

In modern times its not too bad to estimate replacement level with all of the MLE data out there and the plethora of major league and minor league teams... (though I imagine you'll get quite a bit of disagreement between people who calculate it). In the 1800's with the small leagues... its really tough. Finding a historical relationship between average and replacement and applying it en masse to all has flaws, but is often convenient.

In The Politics of Glory, Bill James looked at the percentage of major league at-bats accounted for by HoFers over the years. He found that the historical norm was about 10 percent, and that the 1924-1934 period was overrepresented because more than 20 percent of the at-bats were by HoFer's in each of these years.

As a way of looking at pitching representation in the HoM, I've computed the percent of innings pitched that were by HoMers by year and by league in the table below.

If Mickey Welch were elected, the percentage of NL innings pitched by HoMers from 1880 through 1892 would jump to 29-32-33-29-26-29-24-25-21-24-20-26-15.

Percentage of innings pitched by HoMers (through 1929 inductees)    

Year NA NL AA UA PL AL FL ML
1871 11                   11
1872 12                   12
1873 14                   14
1874 15                   15
1875 11                   11
1876    13                13
1877     3                 3
1878    10                10
1879    21                21
1880    19                19
1881    26                26
1882    28  0             18
1883    23  9             16
1884    19  4  0           7
1885    23  1             12
1886    18  4             11
1887    21  0             10
1888    16  0              8
1889    20  0             10
1890    17  0     9        9
1891    25  0             12
1892    14                14
1893    13                13
1894    11                11
1895    10                10
1896     7                 7
1897     7                 7
1898     7                 7
1899     7                 7
1900     9                 9
1901     7          11     9
1902     4           9     7
1903    10           7     9
1904    12           8    10
1905    10           7     9
1906     8           7     8
1907     8          10     9
1908     8           9     9
1909     6           7     6
1910     6           7     6
1911     6           6     6
1912     4           6     5
1913     4           3     4
1914     3           2  2  2
1915     2           0  5  2
1916     1           2     2
1917     0           1     1
1918+ -nil-

Eric, about that sag with all the 7's lined up from 1896 through 1899 and again in 1902 - how much difference would either Clark Griffith or Vic Willis make?

One thing to remember about 1904 being a high point is that Kid Nichols pitched > 300 innings in 1904, with < 200 the year after and none in the year before.

Because there were fewer pitchers per team and hence fewer pitchers when the game was young, a greater fraction of innings were pitched by stars. It's not that unreasonable for the HOM percentage to drop some at around 1893. How much it should drop is less clear.

about that sag with all the 7's lined up from 1896 through 1899 and again in 1902 - how much difference would either Clark Griffith or Vic Willis make?

Adding Griffith would make the '96-'02 percentages
9-10-9-9-12-10-8;
adding Willis would make them
7-7-9-9-13-10-9.

CAREER
1896-1902


INNINGS PITCHED                 IP     
1    Cy Young                 2573.2   
2    Red Donahue              2039     
3    Clark Griffith           2034     
4    Kid Nichols              2024     
5    Chick Fraser             1929.2   
6    Jack Powell              1894.1   
7    Win Mercer               1822     
8    Pink Hawley              1782.1   
9    Al Orth                  1740.2   
10   Brickyard Kennedy        1681.2   
11   Vic Willis               1605     
12   Nixey Callahan           1541.1   
13   Bill Dinneen             1510.2   
14   Jesse Tannehill          1499     
15   Joe McGinnity            1443     
16   Ted Lewis                1405     
17   Ted Breitenstein         1393.2   
18   Doc McJames              1344.1   
19   Bill Carrick             1324.2   
20   Noodles Hahn             1316.2   
21   Willie Sudhoff           1306.2   
22   Frank Kitson             1251.2   
23   Jack Taylor              1242     
24   Frank Killen             1241     
25   Jouett Meekin            1227.1   
26   Jack Taylor              1218.1   
27   Wiley Piatt              1209.1   
28   Bert Cunningham          1182.2   
29   Deacon Phillippe         1168     
30   Ed Doheny                1144.1   

Griffith and Willis would help a bit, but the fact is that there just weren't a lot of HOM candidates eating up IP during this time.

In the Neyer/James guide to pitchers they talk about the rule change in 1893 (moving the mound back). They talk about how it had to happen at some point because the hard-throwers were dominating at the closer distance.

After these rule changes, pitchers have to experiment more with trying to get batters out... it took a while for the pitchers to catch up. Similar things happened in 1920 when they eliminated the doctored ball.

Big ballot changes this year. First, my thoughts on the newcomers.

Pratt - Just squeezes onto my ballot. Nice hitter, nice defender, quality full career with a touch of peak. Won't stay on my ballot for long, though.

Daubert - Near the very bottom of my consideration set. No peak, little career, nothing to see here.

Gardner - About 10 slots ahead of Daubert, which is to say nowhere close to the ballot. More career, still no peak.

Donaldson - I'll cop to not being completely informed on him, but he seems like a nice pitcher who has no real case for being better than the many other pitchers who aren't near my ballot.

----------------------------------------

Alright, had to rebuild my rankings from scratch after accidentally overwriting my old spreadsheet (Note to self: Backup). My consideration set at this moment consists of everyone who got a vote in the last election plus a couple people I remember that were decently ranked in my old rankings (Griffin, Thomas), and everyone from 1921 forward with a pulse. If there's anyone I'm likely overlooking, let me know, I don't intend to be exclusive with it.

I was dissatisfied with my methods of evening out "hang around value" with my career voter leanings. I'm much happier with my new system. Instead of attempting to punish those guys, I reward guys with star or superstar seasons (my definition of those totally subjective). This causes a near complete reshuffling of my ballot in a lot of places, though in others it is remarkably similar.

1. Foster
2. Jennings
3. Caruthers
4. Sheckard
5. Ryan
6. Childs
7. Van Haltren
8. Poles
9. F. Jones
10. Griffith
11. Monroe
12. Williams
13. Cross
14. Pearce
15. Pratt

16-20. Long, Duffy, Beckley, Griffin, Pike
21-25. McGraw, Dunlap, Thomas, Waddell, Bresnahan
26-30. Bond, King, McCormick, Huggins, Konetchy

Anecdtaol only, but re Beckley, baseballlibrary.com

"Beckley also won fame with his glove. His 25,000 chances and 23,696 putouts rank first all-time among first basemen. However, he had a notorious scatter-gun arm, and runners always took the extra base on him. He also had a unique hidden-ball trick: He hid the ball under one corner of the base."

He also had a unique hidden-ball trick: He hid the ball under one corner of the base.

That would be a tough one to pull off today.

I'd be interested to see what the at-bat numbers look like, including the pitchers, if there's a major difference from James' numbers.

Does anyone know anything about Spotswood Poles' defense? Of course, you assume he was a good fielder because he played CF, but was he being compared to Tris Speaker, or just good enough to get by?

Now for the topic that's really bugging me these days.

I have Rube Foster down in the 20s, and I'm starting to wonder if there's a flaw in my thinking or not. (Anyone with opinions either way is welcome to comment.) From what I understand about his career, it seems there is a six-year span of greatness, surrounded by periods where he just wasn't pitching that much. Now, that isn't a fatal flaw in his candidacy, it's really not that disimillar to Walsh or Joss or McGinnity as a career path.

But for me to put someone like that in the HoM, I have to know that in those peak years, he really was great, one of the top few pitchers in the game. Walsh met that standard, Joss really didn't, and the Iron Man is right on the border. But for me to support him, I have to be certain that he was really that good. With Foster, given the limited stats available and the uncertainty of the competition (and a ridiculous-sounding W/L record doesn't help, because it makes me think the competition stunk), I really don't feel like I can be certain that he was at that top level.

Now that's where I wonder if I'm being unfair. Am I setting the standard so high for a Negro League pitcher in that era that it's unreachable? The position players I've supported all had long careers as a support for their argument, but can a pitcher really have that in this era? (How many white HoM candidates from this time had long, consistent careers? Young, Mathewson, Plank and probably Nichols are about it.) Also, when I read through the summations of Jose Mendez' career, I was kind of underwhelmed for someone being presented as a top 2 candidate. So I am starting to wonder if I'm being too tough on Negro League pitchers.

A couple of related points on the subject. First, several people have mentioned strong summations of Foster's career, and I don't really recall seeing that, so I don't know if I missed something or I disagree with the interpretation. Can anyone point me to a specific thread or post? Second, I'm not really convinced by the argument that the contemporaries would have found taking Hill over Foster inconceivable. How would the contemporaries have felt about taking Sheckard over T-E-C? I know it's a different situation, but I just don't think it's a truly relevant point. Third, I do wonder if some people, including me, are overcompensating, You know you have to mark down Foster's reputation because there's so much off-field stuff involved, which we're not judging here. But that's an inherently negative judgement, and I worry it may lead to knocking him down too much.

OK, sorry I rambled on so long without much of a point. But half of me feels like I'm doing something wrong, and the other half is afraid that I'm trying to talk myself into going along with the crowd. Any thoughts on the subject would be appreciated.

Karlmagnus, Chris J.--A *strong* EOMW. He came out the second-worst of anyone I looked at--103 lifetime dERA+, 45.4 career WARP. Yes, some of that has to be due to pitching alongside Keefe who was tossing so many innings, but I think a lot of it was also that he pitched on Giants teams full of Hall of Meriters who presumably played great defense. I was very surprised at how poorly Mickey fared. As for Beckley, I really think he sucks, but I also do think that the ABC triumvirate unfairly raises his replacement level--if he were, say, 4 WARP (his prime average)in 1889, he'd pull ahead of GVH-- but I'm really not sure how to calculate replacement level any other way--I'd be inclined to take it as the second-worst starter at a position in a league, or something, but that would take forEVER to calculate for every position in every league in every year, and I can't spend ALL my time playing around with baseball stats (only 90%).

Does a batting park factor of 120 mean that you should discount runs in that park by 20% or 10% (since teams only play half their games at home)

Does a batting park factor of 120 mean that you should discount runs in that park by 20% or 10% (since teams only play half their games at home)

I think it depends on which park factors you use. I think -- but I'm not sure -- that the park factors from Total Baseball (which appear on baseball-reference.com) already take that into account and don't need to be further adjusted. The Hidden Game of Baseball and Total Baseball apply the factors "as is" (except when applying it to averages like OBA), which means the adjustment is either built in or they've totally overlooked the issue. Also, the Total Baseball park factors take into account that a team's hitters don't face its own pitchers and vice versa, which a simple ((RS_Home+RS_Home)/HomeGames)/((RS_Away+RA_Away)/Away Games) does not.

Baseball-reference.com uses the Total Baseball method, smoothed over three years.

Bill James uses the simple formula and smooths it over five years. There's an article on page 86 of the Win Shares book about this. First, he establishes the smoothed park factor for the '27 Yanks (.910166).

James says you could halve the discount, but "The problem with that is it assumes that everybody's 'road' parks are the same, which they aren't. To adjust for that, we figure that the [1927] Yankees, in an eight-team league, play 14 sets of games, 7 at home and 7 on the road. The seven at home are represented by 7 times the park factor [6.37116]....The seven on the road are represented by 8.00 (the total of eight average parks) minus the Yankee park factor [7.08983]...We add these two together and divide by 14, creating a Park Run Adjustment, for Yankee Stadium in 1927, of .961500. If they were in a 14 team league, we'd be dividing by 26."

SABR's Baseball Biography project recently completed its biography of Del Pratt. It appears at

Pratt Bio on SABR

but I think a lot of it was also that he pitched on Giants teams full of Hall of Meriters who presumably played great defense.

Don't see it. I've got my own win shares-based method for figuring defensive support & by the results Welch's support was nothing spectacular. And I already know his hitting support was only a little above average. How'd he win all those games & have such a good winning percentage while doing it?

Lots of good work and discussion in recent days. Thanks to Dan Rosenheck for his detailed analysis and sharing of conclusions, which have renewed discussions.

While being appreciative of that work, I am not ready to place much weight upon it yet, for several reasons that could be accounted for in improving the system.

1) Park factors can have quite significant effects that must be accounted for before any fine distinctions between great players can be trusted. Clearly, this is a matter than can be addressed, and that Dan is working on.

2) Comprehensive metrics need to be tested by comprehensive results at the team and league level, so that they can be checked readily against actual results. Without some kind of cross-check, I'm not prepared to trust the results too much.

3) There are several "all-time" estimates of batting, fielding, and pitching value built into Dan's system that may distort its representation of the early game. Are the run values for various kinds of batting events appropriate pre-1900? Is the pitching/fielding split? Are the defensive values assigned to each position? In working with win shares, I've found it necessary to correct for its use of static formulas for all-time, and I see Dan's system as subject to similar problems. These might be addressed.

4) Now that we are comparing players from 1858-1924, figuring out how to account for differences in competition quality from era to era is becoming a central problem for ranking. Dan's system, like win shares, doesn't provide any guidance on this matter. I don't agree with WARP's findings, but they provide one cross-period assessment. Dan's system is probably right to put three stars of the 1880s AA near the top of his rankings in terms of their in-context achievements, but these players are still in the pool of eligibles because the electorate has been discounting their achievements because they occurred in the weaker league in perhaps the weakest competitive era since the founding of the National League (that's WARPs view, at least). Dan's analysis doesn't provide any evidence for addressing the question of how to discount performance against weaker competition. No fault of Dan's work, as this is not a factor that can be addressed from inside the system, but we shouldn't take it as demonstrating something it doesn't demonstrate.

Chris Cobb--thanks for your feedback.
I have now figured out how to park-adjust these numbers, and have updated spreadsheets including park adjustments which I will post to the Yahoo group once I sort a few more things out.
Not sure what you mean about "comprehensive metrics," but I certainly don't have the time or resources to calculate WARP for every player in every season ever, which I imagine would be required to do the kind of testing you are discussing.
It would definitely be an improvement to change positional weights, pitching/fielding splits, and XR values year-by-year, if anyone has any idea how to do that.
In my own balloting, I've just arbitrarily decided to discount the early (1882-84) and late (1888-91) AA by 5-20%, depending on the year, and 15% for the Federal League. This could be wildly off--Browning's 1882 season is perhaps the most challenging example of this, as he put up an extraordinarily valuable season against a cast of nobodies.

As for putting stock in the results, my best argument in support of what I've found is that I've explained my methodology and posted the spreadsheets that show exactly how I got from raw stats to WARP. I'd encourage everyone else to play around with them and try to improve them (and post what you've done!)

After park adjustments and weighting my own system against WS and BP WARP, I've changed my preliminary ballot posted on this thread. I'll repost a new one now.

1930 ballot

See Ballot Discussion Thread for WARP methodology—if you’re skeptical about something, the spreadsheets are available in the Yahoo group. WARP totals are park-adjusted. I have discounted the 1882 AA 20%, 1883 and 1891 AA and 1914-15 FL 15%, 1884, 1889, and 1890 AA 10%, and 1888 AA 5%. I am referring to a five-year peak throughout this ballot.

Tier One (75+ career WARP)

1. Bob Caruthers (86.4 career WARP, 67.9 peak)
Further analysis leads me to conclude that the fact that Caruthers has not yet made it in is a travesty. Not only does he have by far the best peak value on the ballot (at the Ruth, Bonds, Wagner level), he has the best career value on the ballot. 86.4 wins above replacement for his career puts him below Plank (96.1) but ahead of Three-Finger Brown (84.8). An unprecedented and unparalelled dominator. Adjusting for season length, he's also second on the ballot to Welch in career season length-adjusted Win Shares (just nudging out Van Haltren) and obviously blows everyone away in peak season length-adjusted Win Shares. The closer I look, the more Caruthers looks to me like an inner-circle, no-doubt-about-it HoM'er. If he doesn't get in during this backlog I'm gonna vomit.

2. Charley Jones (79.3 career WARP, 51.8 peak)
Yes, I’m serious. Conservatively crediting him with 5 WARP for each of his blacklist years, he blows past everyone besides Caruthers on the ballot. I think the short season lengths he played in have caused people to sell him short… I'm a little dubious on BP's crediting Jones with 27 FRAA in the 85-game 1879 season. But even if he wasn't quite as good as that would suggest--17.3 WARP that year is the second-best season I've calculated, better by a good bit than Ty Cobb's best (13.8 in 1917)--he was still a force, making mincemeat of the 1883-85 AA along with Browning. Third-best peak on the ballot (after Caruthers and Jennings), third-best career total (Caruthers and Griffith). I would want to give him a little extra credit for having one of the greatest seasons in major league history in ’79, but he doesn’t need it. And it’s not just me—once you adjust BP’s WARP1 for season length, Jones finishes third on the ballot in peak. If you conservatively credit him about 80% of his (less than stellar) 1883 performance for his blacklist years, he’s first in career season-length-adjusted WARP1; if not, he’s still third or fourth. I would strongly encourage all of you to reevaluate him.

3. Clark Griffith (82.1 career WARP, 48.6 peak)
This is why I spend the time on this stuff--Griffith didn't even make my ballot when I started, and now he's rocking the top of my ballot. ERA+ makes it seem that Griffith had one dominant year in 1898 and was just above average elsewhere. In fact, he was just as good in 1899 (look at K, BB, HR, and BABIP/Teammates' BABIP), was a reliable workhorse, and pitched at an All-Star level for a decade. You can't see his greatness on the surface, but look deeper into the numbers and from 1896-1901 he was a genuine superstar. His 82.1 career WARP are a big jump ahead of anyone else. I'm the new Best FOCG—expect to hear a lot more from me about Griffith in the years to come.

Tier Two (65-75 career WARP, 40+ peak)

4. Lip Pike (72.1 career WARP, 46.8 peak)
I’ve given Pike a total of 19.2 WARP for his pre-ML 1866-70 career, an average of under 4 WARP a season, which I think is pretty conservative. There’s a glut of candidates with 45-48 peak WARP and 67-72 career, so Pike gets the edge because a) he has the most career WARP of the group and b) BP season-length-adjusted WARP1, when given the same pre-ML credit ratio as my WARP, likes him a little better than his peers. Obviously a truly dominant player in the NA and 1876 NL Also, gotta support the tribe of Israel here. Go Lipman Go--Hank Greenberg, Sandy Koufax, and Shawn Green smile upon you. So does Doron Sheffer.

5. Pete Browning (70.7 career WARP, 47.5 peak)
Almost indistinguishable by season-length and league quality-adjusted WARP, BP WARP, and Win Shares from Childs, so his much better Ink scores get him the nudge. This guy really needs more support. 1890 showed us he was for real, so his knock-em-dead years in '82, '85 and '87 have to be taken seriously. More career value than the “career” guys GVH/Beckley by my measure, and a true dominator for three or four seasons. Hopefully I can drum up some support for him; he really deserves it.

6. Cupid Childs (67.2 career WARP, 45.3 peak)
Once again, someone I scorned gets massively reevaluated. Being "the best 2B of the 1890's" is not enough, but he was an offensive juggernaut at a scarce position with often excellent leather for eight years. A bona fide superstar in '90, '92, and '96, and a strong All-Star in '93 and '97. Didn't play forever but so good that, like everyone else in this group, he accumulated more career value than the "career guys." Since we don't have anyone from his era at his position, and because he played in a stronger league than the other guys on the ballot with similar WARP totals for career and peak, he gets my no. 6 vote.

7. Addie Joss (69.4 career WARP, 47.8 peak)
What a boomerang! I went from being a good FOAJ, to dropping him from my ballot altogether, and now moving him back up to 5th! Joss had a remarkable ability to prevent hits on balls in play, allowing a BABIP 31 points lower than his teammates' for his career (.238/.269). He had six seasons where he was absolutely one of the best in the biz, including 1908 which was particularly standout. His rate stats were so good that even despite his innings problem, he still comes out mid-ballot on both career and peak. Take a look inside the numbers and see for yourself.

8. Hughie Jennings (63.4 career WARP, 57.2 peak)
Just falls short of the career cutoff for Tier Two, but if you believe FRAA (averaging 47 FRAA/130 G in 95-96 is pretty nuts), the peak is so high you can’t ignore it. If there is reason to be skeptical of his defensive brilliance, he'll fall. But if not, he is a Caruthers type--so good for five years that he was more valuable than guys who played for three times as long. 4 years of 170-190 XR+ at the SS position, plus that defense, makes him HoM-worthy to me.

Tier Three (60+ career WARP, 40+ peak)
9. Jimmy Sheckard (71.6 career WARP, 40.2 peak)
A strange career. Much of his value in defense and walks, including an outlandish 147 in 1911, when he hadn't been particularly great in eight years. But it all adds up to a mid-ballot candidate as far as I'm concerned, just squeaking into the HoM. I do think there are much stronger candidates than election-favorite Sheckard on this ballot—Pike seems to have enough support that he could make a preferable alternative, IMO.
--------------------
In-Out Line
--------------------
10. Rube Waddell (66.6 career WARP, 40.9 peak)
Rube’s taken a big hit with my reevaluation. I *love* the K's, but now that I can see that deadball pitchers really could prevent hits on balls in play, he stands out less than he did before. It's worth nothing that his 1903 season was just as good as his much more highly regarded '04--almost as many innings, same BB/K/HR rates, similar propensity to giving up line drives (BABIP 5% higher than teammates' in '03, 6% in '04). '02 was really his best season though. One of the best pitchers in baseball from '02-'05, but not an otherworldly dominator and not enough career to push him further up the ballot or into my revised PHoM.
11. Eddie Cicotte (67.2 career WARP, 40.6 peak)
He really was a premier, superstar pitcher from 1917-19, and was serviceable in 1913 and 1920. A slightly above league average pitcher for the rest of his career.

The Rest
12. Jimmy Ryan (62.1 career WARP, 34.3 peak)
He doesn’t fare that well in my system, but I do have to give respect to his near-ballot-topping career season-length adjusted BP WARP1 and top 4 career WS, and he did at least have two great years in 1888 and 89.
13.John McGraw (59 career WARP, 40.3 peak)
Last guy on the ballot with 40 peak WARP, and almost as many career WARP as Ryan. He didn't play long enough to make the HoM, and rarely played full seasons even when he did. But man, was he good--an on-base machine the likes of which the game has rarely seen since.
14.Vic Willis (70.7 career WARP, 35.1 peak)
Just kept churning out those innings at an above-average level. The Beckley of pitchers, but a more valuable career than Beckley and at least a genuine All-Star once or twice. Can’t keep 70+ WARP off my ballot altogether.
15.Rube Foster (n/a)
Because I'd rather vote for the best Negro League pitcher of his generation than an ML player I really can't support. I'll nod to the consensus here that Foster > Monroe.

...because they occurred in the weaker league in perhaps the weakest competitive era since the founding of the National League (that's WARPs view, at least).

This comment made me wonder something. Of course we all discounted the AA to one degree or another vis-a-vis the NL. But because the NL was watered down by the AA, PL and UA (sort of), the NL from 1882-1891 was far weaker than the NL prior to and after that.

To what extent are the voters taking that into account? E.g., Welch having a 113 ERA+ primarily in that era doesn't seem as impressive to me as a pitcher having a 113 ERA+ from 1892-1900.

I'm not suggesting timelining or even a year-by-year league quality adjustment (for instance, measuring the NL vs. the AL). I'm talking about a period where the talent was spread very thin among the various leagues.

To what extent are the voters taking that into account? E.g., Welch having a 113 ERA+ primarily in that era doesn't seem as impressive to me as a pitcher having a 113 ERA+ from 1892-1900.

My view is that, taken collectively, the electorate has not taken this into account sufficiently, leading to slight over-representation of the 1880s and under-representation of the 1890s so far. It's not a major problem, but I think it is a problem, and it could lead to what I called "serious error," in the next 20 years if we keep putting 1880s guys in.

Of course, I think we made a mistake in not electing Welch, so I don't advocate a moratorium on 1880s players.

On the subject of Welch, here's my reasoning for seeing WARP's evaluation and Dan Rosenheck's evaluation of him as inaccurate.

Chris J., as he mentioned has used win shares to calculate the fielding support above average received by pitchers. He calculates that Welch gained 5.4 wins from good fielding in his career.

Using a study of team defensive efficiency to calculate hits saved above average at the team level, I have calculated that Welch gained 8.1 wins from good fielding in his career.

I have just used Chris J's RSI numbers and Dan's XERA+ numbers to calculate Welch's record independent of defensive support, according to Dan's modified DIPS analysis. I haven't used pythaganport exponents, so there may be some small inaccuracies in my calculations on that basis, but these numbers suggest that, independent of defensive help (that is above average defense), Welch would have won 35.6 fewer games than he did. That's 30.2 wins more than Chris J. calculates, and 27.5 wins more than I calculate. If you add that difference to Welch's WARP (each additional win being a full win above average that must be to his credit, since it is not a product of offensive support or fielding support), Welch moves from being an off-ballot candidate to a strong candidate.

I think there's a missing win problem here. It seems to me much more likely that calculations of Welch's value derived from his component stats are missing his value than that calculations derived from team fielding statistics are missing the value of the fielders, taken as a group.

I further doubt, that if Welch was basically an average pitcher of durability just a bit above average, that he would have lasted to throw nearly 5000 innings.

I'm open to persuasion, certainly, but here's what someone needs to explain to me:

In 1884, Welch's XERA+ and RSI produce a record of 33.3 - 26.7. Welch's actual record was 39-21. Defensive efficiency shows that New York had below average defensive efficiency that season (not by much, I calculate 8 runs), which should lead to Welch having a record _worse_ than his run-support and components allow. Instead, he beat that expected record by 5.7 wins. That's a lot of luck, and it appears to me to have continued throughout his career. If someone can show me how those 5.7 (or, for the career, 27.5) extra wins really belong to the defense rather than to Welch, I'll assent to dropping him out of the serious candidate picture. But now, I see the most reasonable interpretation of the evidence as pointing to a failure on the part of the metrics to register Welch's skill properly.

I further doubt, that if Welch was basically an average pitcher of durability just a bit above average, that he would have lasted to throw nearly 5000 innings.

Bobby Mathews did. Tony Mullane threw only 300 innings less (not even a full season). I don't think they are HoMers. How durable was Welch? I don't really know. He threw a lot of innings, but it wasn't too tough to pitch back then.

I see a lot of conflicting stuff about Welch. We've got unspectacular defensive support and run support, which seems to indicate he "earned" all those wins. He's also got fantastic Wins Above Team. Win Shares likes him, though Win Shares overstates pitching contributions during this period -- I shift about 49 of Welch's pitching WS to the fielders.

On the other hand, his biggest year (1885) was the year he got significant help from his defense. Then we have the fact he wasn't dominant. He didn't lead the league in much of anything. WARP doesn't like him. Palmer's linear weights doesn't like him. He pitched when talent was spread thin, yet he has a modest (by HoM standards) ERA+ of 113. If he didn't have spectacularly good or bad defensive support, then his ERA ought to be fairly reflective of his achievements. Presumably the league ERA against which we calculate his ERA+ has average defense and run support (by definition). How did he win all those games with an ERA+ of 113?

In the end, I tend to reserve the best ballot spots for players who do well across multiple systems of measurement.

On the other hand, his biggest year (1885) was the year he got significant help from his defense.

Yes, he did. The team defense saved, for the season, 76 runs above average by my calculations, which gained Welch 3.7 wins.

Welch's record, derived from RSI and XERA+, should have been 34.4 - 20.6. He beat that projection by 9.6 wins. Even in 1885, it looks to me like the systems that rate Welch as nothing special don't have any obvious way of accounting for how the team won the games that it did when Welch was pitching.

It's possible that my system for calculating fielders' contributions is faulty. Here's one probable shortcoming. Looking at 1885, I see that in addition to being excellent at hit prevention (130 hits saved above average), the defense was also much less error-prone, committing 113 fewer errors than league average. Since I don't have an average run value for an average error, I have not included errors in my defensive efficiency calculations. Errors not treated as terribly significant in the modern game, since the range of variation among teams relative to the total number of defensive chances is very small, so I left them out. With so many more errors being committed, obviously the range in errors from team to team _is_ a significant component of team defensive performance. Still, the average run value of an error would have to be greater than the average run value of a hit for the errors to make up the full difference between predicted and actual wins.

In the end, I tend to reserve the best ballot spots for players who do well across multiple systems of measurement.

I'd say that's generally a reasonable approach to take, but if a significant flaw in a system's evaluation of a player is identified, it seems unreasonable to continue to place weight on that system's evaluation. I'd like to see us push on the systems we have until they yield results in which we have high confidence, if we can.

It would be useful to see Dan Rosenheck calculate his WARP and XERA+ numbers for Tim Keefe, to see if the same gap arises between his projected and actual wins as it does in the case of Welch, or if the Keefe projection turns out to be more accurate. That would help us test the various measures.

where can I find season by season leaders for all defensive stats?

OK, I figured out the % total for at-bats per year. When there's more than 1 number, the first is the percentage for the top 2 leagues (NL & PL for 1890), followed by the NL, AA or AL, and any 3rd league. For pre-1892 seasons, I'll add the number of teams that year in parenthesis. I'm not entirely sure what conclusions to draw from this, so I won't draw any.

1871  10.1  (9)
1872  11.6  (11) 
1873  15.8  (9) 
1874  15.9  (8) 
1875  13.5  (13)
1876  14.8  (8)
1877  18.1  (6)
1878  19.5  (6)
1879  21.2  (8)
1880  19.7  (8)
1881  25.7  (8) 
1882  16.5  26.5   1.8  (8/6)
1883  13.5  23.2   3.7  (8/8) 
1884  10.4  21.8   2.4   0.6  (8/13/13) 
1885  13.3  23.8   3.0  (8/8) 
1886  12.1  22.4   3.2  (8/8)
1887  11.3  20.7   2.7  (8/8)
1888  10.5  18.2   2.9  (8/8)
1889  12.0  20.0   4.2  (8/8)
1890  13.5  14.3   0.0  12.8  (8/9/8)
1891  12.4  21.4   3.3  (8/9) 
1892  16.3
1893  15.4
1894  13.8
1895  13.6
1896  13.4
1897  12.4
1898  12.9
1899  11.9
1900  18.6
1901  10.5  16.8   4.0
1902   8.9   7.8  10.1
1903   8.3   5.7  10.9
1904   9.0   6.3  11.7
1905   9.0   7.1  10.8
1906   8.2   6.9   9.4
1907   7.8   5.8   9.7
1908   7.4   6.7   8.1
1909   6.1   4.9   7.4
1910   5.4   4.2   6.5

That seems about as far as it's useful to go.

Sunnyday2: Back on the 1927 ballot discussion thread, you presented the results of a study of the 1860s stars using a rough index of value, that was helpful in showing that Pearce was clearly more valuable than Harry Wright, the only player in the comparison group who was his true contemporary.

Would you be willing to expand that study to include Joe Start, Al Reach, Dick McBride, and Charley Smith, now that DavidFoss has kindly posted numbers for all of them? Since there have been questions raised about how one can tell whether Pearce really stood out among the best players of his day, I think your analysis would be particularly helpful in answering it.

I'm working on a similar study myself, but I think your approach was a bit more sophisticated than what I'm doing.

where can I find season by season leaders for all defensive stats?

yest, baseball-reference.com has season-by-season listings by position of all traditional fielding stats. You can't see the leaders at a glance, but you should be able to find the leaders you're looking for fairly quickly there. Go to the league pages, and click on the link for fielding.

Would you be willing to expand that study to include Joe Start, Al Reach, Dick McBride, and Charley Smith, now that DavidFoss has kindly posted numbers for all of them? Since there have been questions raised about how one can tell whether Pearce really stood out among the best players of his day, I think your analysis would be particularly helpful in answering it.


In case people didn't get the email. These posts are now available in the "Files" section of the HOM yahoo group. The 1927 discussion was getting awkward to browse.