Baseball for the Thinking Fan

Login | Register | Feedback

You are here > Home > Hall of Merit > Discussion
Hall of Merit
— A Look at Baseball's All-Time Best

Monday, February 09, 2004

1920 Ballot Discussion

It’s time to start discussing the 1920 ballot . . . only one gets in this year . . .

BOBBY WALLACE (tough call, but I think it is fair to make him eligible now, he was 40 in 1914)

Joey Numbaz (Scruff) Posted: February 09, 2004 at 09:46 AM | 167 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

Page 2 of 2 pages  < 1 2
   101. Marc Posted: February 13, 2004 at 07:03 PM (#521680)
Carl, "short career in the 19th century" is just code. Translated it stands for "timeline."
   102. OCF Posted: February 13, 2004 at 07:58 PM (#521681)
Carryover ballot from last year:

1 & 2: elected
   103. Al Peterson Posted: February 13, 2004 at 08:25 PM (#521682)
1920 prelim here. Finding the right someone to vote for is becoming a challenge. I'd say I'm sold on the top 5 guys, everyone else still has some warts. Spots 13-22 are so tight; I think at one time or another they all have made a ballot from me.

1. Charlie Bennett (1). Didn't play everyday but the position at the time didn't allow it. Fine hitter, excellent receiver, positional adjustment = HOM member.

2. Jimmy Collins (4). Jimmy Collins was named on too many All-time All-Star teams from the 1st half of the 20th century to not be an extraordinary player. A+ Defensive Win Shares Grade from Bill James.

3. Ed Walsh (-). Peak-heavy pitcher who outdid the other short timers like Waddell and Joss. His workloads were impressive, impacted pennant races in the 1900s, whats not to like.

4. Rube Waddell (5). Yes, I like pitchers. His strikeout numbers were well ahead of his time.

5. Sam Thompson (6). Slugging RF got off to a late career start. Ended early also due to injuries.

6. Jimmy Sheckard (7). Some of his accumulated stats are hard to ignore. Looks to be a well rounded player.

7. Bobby Wallace (-). Part of my analysis on him says put him in quickly, then I look at other SS we have in the HOM and I wish to hold back. Not a super peak but did play many years. Didn't do too well on the managing side of things.

8. Joe McGinnity (8). Start, relieve, whatever to help the team. Sidearming wonder who had (non-ML) longevity.

9. Frank Grant (9). Sticking with Frank as the man to bear the early Negro League torch.

10. Cupid Childs (10). Allowing a shorter career length for infielders in the rough and tumble 1890s helps Cupid. Still hit with the best of them some years, regardless of position. If McPhee was the fielding 2B, Childs was the hitting 2B. And saying Cupid couldn't field is a disservice.

11. George Van Haltren (12). The OF glut is still around. Broad skills certainly didn't hurt his team.

12. Clark Griffith (11). The many pitching metrics presented show he's in the mix as HOM worthy. Most people voting are saying the same thing: Nice career, here's a low ballot spot, thanks for playing.

13. Jimmy Ryan (13). A thumbs up for career length, thumbs down for not doing more with that time.

14. Fielder Jones (14). Very steady production for 13 years for this centerfielder. Outstanding with glove, a leader of overachieving teams in Chicago.

15. Hughie Jennings (-). Return of the peak monster. Hard to argue his value over a five-year span, hard to ignore his other years. Another Bill James A+ fielder.

Who is below that you ask?

16. Bob Caruthers. Another peak guy, this time doing it pitching and hitting. Hurt (slightly) by doing work in AA and by my cap on value added per season. IOW one player can only make so much difference. Most likely back on the ballot in a couple of years for Freedom Bob.
   104. karlmagnus Posted: February 13, 2004 at 09:33 PM (#521683)
Cap on value per season should surely be twice as high for those leading their team in both pitching and hitting, like Caruthers at his best and the early Ruth (do you cap Ruth 1918 or 1920?). It appears to have shortened his career, so he should be given full credit for it while he did it and not arbitrarily marked down. Otherwise it?s just an ad hominem ripoff of Caruthers, of which there seem to be plenty round here.
   105. ronw Posted: February 13, 2004 at 10:29 PM (#521684)
Tentative ballot with new ranking system.

Don't worry, Howie, I still have critical thinking.

1. Bennett (1,1,1) - Scores much higher than any other catcher in my system. I want to see what happens when Bresnahan gets added in, and I just might get my wish.
   106. Rick A. Posted: February 13, 2004 at 10:42 PM (#521685)

This was posted in the important links thread

Also, if you're going to try to be witty about something and use someone's handle, please use something like "not the real Rick", even though Rick's number 3 was stupid, he shouldn't have his handle stolen, even if it was obvious that it wasn't the real Rick . . .

I didn't get an opportunity to look at any threads yeaterday. (busy day at work) What is this in reference to? Was there a post on the thread that was impersonating me?
   107. Rick A. Posted: February 13, 2004 at 10:49 PM (#521686)

Never mind. I just saw it in the WARP3 thread. None of those posts were made by me. Not even the ones made by 'the real Rick'. If there are any other Ricks that want to legitimatly contribute to the discussion and join this group, that's great. But please DO NOT post under other peoples names and pretend to be them. As I said, I have not posted anything in the last couple of days, so please don't attribute those messages to me.
   108. jimd Posted: February 13, 2004 at 10:59 PM (#521687)
I can speak only for myself, but I doubt anyone confused "Rick" (the troll) with you, Rick A. His posts didn't seem to be quite your style ;-)
   109. Howie Menckel Posted: February 13, 2004 at 11:29 PM (#521688)
Touche, to Ron Wargo.
   110. Marc Posted: February 14, 2004 at 01:23 AM (#521690)
Rick A., it was quite obviously not you.
   111. Jim Sp Posted: February 14, 2004 at 01:43 AM (#521691)
Ok, I'm in a contentious mood. I'll probably regret this but I'll bite.

Caruthers has an ERA+ of 123 in 2828 IP, with a high of 158. 3 years with an ERA+ over 137.

Walsh has an ERA+ of 145 in 2964 IP, with a high of 189 and six straight years at 144 or higher. Big difference, there.

League strength: not even close. Many pitchers with less than HoM credentials racked up gaudy years in the AA, it's just not an impressive achievement. Look up Tony Mullane, Silver King, Guy Hecker, Will White, Dave Foutz, Ed Morris...

Team support: Walsh went 18-20 with an ERA of 1.27 in 1910. His team was known as the "hitless wonders". Caruthers played for the strongest team in a weak league.

Timeline: why bother, it's not necessary. or explain to me why Tony Mullane, Silver King, and Will White aren't on your ballot. they all had more impressive pitching careers than Caruthers. More IP, all of them. Even Silver King had north of 10% more IP than Caruthers, same ERA+, and his peak year is way more impressive: 199 ERA+, 585 IP.

So if you have Caruthers high on your ballot, it must be because of his special hitting value.

If his hitting is only 25% of his value, then he doesn't make the HoM by a long shot.

Otherwise how can Tony Mullane be completely off the ballot? Mullane pitched 1600 more innings at almost the same quality as Caruthers, hit quite well himself, and he proved himself in the 1892 single league NL.

Oh, and look, Mullane's peak year was 1883 in the AA for...St. Louis. Is this a coincidence? Probably not.

I've had Caruthers on my ballot in past years, but now I'm not buying it.

I do have a stiff AA discount (20%), any consistent analysis either has a stiff AA discount or puts more AA pitchers on the ballot.

So...explain to me why Caruthers hitting should be more than 25% of his total value and I'll vote for him. Or put Mullane on your ballot.
   112. karlmagnus Posted: February 14, 2004 at 02:26 AM (#521692)
Caruthers' ERA+ (a pretty suspect stat anyway) is deflated by his last 2 year when his arm was falling off. His W/L record is far better than anyone else on the ballot, and for a CAREER (not a season) wins and W/L% are the most important single pitcher stat. Most of Caruthers' teams were the best because he was on them; only the 1886 St Louis team was really strong. Mullane is just off my ballot, but his best years were in 1882-84, when general consensus (as distinct from randomly invented 20% discounters) is that AA was not as strong as it later became -- certainly a 20% discount is ludicrous for 1886, when St Louis won the World Series. Caruthers wasn't just a good hitter for a pitcher, he was in the Stovey class, and only a little behind Browning and Thompson, the best hitters of his era. He was a unique talent, the best player in the AA (and paid as such) for several years. Only bending the stats like a pretzel gives you an excuse for keeping him out of the HOM -- for example, the idea from fielding "win shares" (Jim Spencer post 90) that Caruthers was a uniquely bad fielder is laughably implausible -- he was 5 foot 7, 135 pounds and in his early 20s.
   113. Marc Posted: February 14, 2004 at 02:45 AM (#521693)
Karl is right on. A 20% discount is OK for '82, pretty high for '83-'84 and totally unjustified thereafter until '91. An AA peak in '82-'84 is a completely different kettle of fish than an AA peak in '85-'87.

Then there's the little fact of Caruthers' peak value. If you're not a peak voter, fine. But some are and those would be the FOBC.

Jim, where were you when McPhee and Stovey waltzed in?
   114. jimd Posted: February 14, 2004 at 03:07 AM (#521694)
If his hitting is only 25% of his value, then he doesn't make the HoM by a long shot.

James' unadjusted Win Shares has Bob's hitting at 27% of his value. The rest is 71% pitching and 2% non-pitcher fielding.

WARP-1 has Bob's hitting at 30% of his value, 59% pitching and 11% fielding. According to BP his OF fielding is nothing special, somewhat below average, a 96 rating in RF and a 94 rating in LF. OTOH, his fielding as a pitcher appears to have been excellent to spectacular, 111 for his career, including a peak season at 131. His Range Factor (as a pitcher) was 2.5 compared to league average of 2.2. His overall error rate (as a pitcher) was 7% compared to league average of 11%. In both 1887 and 1889, his error rate was 3% (and if you've looked at 19th century error rates, that is spectacular). He appears to be a near-certain Gold Glove (as a pitcher) in 1887, and might be in the mix for other seasons (I haven't checked other AA pitchers to determine who his rivals might be). If you toss out the fielding as irrelevant, then his hitting is 34% of his value, not James' 27% (because pre-1893 pitching value in Win Shares is allegedly grossly out-of-whack, deflating his pitching Win Shares increases the proportion of Bob's value that is hitting.)

The other pitchers. Well, I have Silver King sitting just off of my ballot. If the next few years have a dearth of qualified candidates, he may make a return appearance. Mullane? As far as I'm concerned, the AA is of very uneven quality. I think your 20% discount is appropriate only from 1882-4 and in 1890; at Caruthers peak I think the league imbalance is more like the AL/NL discrepancy of the late 50's. Unfortunately for Tony, his best seasons are in the weak AA. To make my ballot, Caruthers argument is a peak argument; Tony has no such peak, that his career adds up to similar totals is like the Pappas/Drysdale argument.
   115. Chris Cobb Posted: February 14, 2004 at 04:30 AM (#521695)
1920 Preliminary Ballot

Lots of changes from 1919. New arrivals Bill Monroe, Bobby Wallace, and Ed Walsh find ballot spots, and I do quite a bit of shuffling. The new arrivals are the vanguard for very strong group of candidates who'll be moving onto the ballot over a four-year period. This year and next year are the weakest ballots we'll see until the late 1920s, and every one of the top nine players on my ballot has a good argument for being the best eligible player. In placing them, I've given priority to players at underrepresented positions. Giving priority to players from underrepresented periods would also be justifiable.

Following Chris J.'s lead, I've been reading _Cool Papas and Double Duties_. Anyone who wants a comprehensive sense of the experts' views would do well to read it, though the format is not the most straightforward, and if McNeil had gotten all of the experts to follow our "best practices" ballot construction, it would have been more helpful. Looking foward to getting Holway's book in the mail soon . . .
   116. Jim Sp Posted: February 14, 2004 at 04:46 AM (#521696)
As some of you probably suspect, Post #90 is not a serious comment. The Rocky Colavito mention could have been a tip off for the attentive reader ;-).

Hmmm...I am missing an annual correction for league strength. In Caruthers case this is the crucial point as his peak exactly corresponds to the period in question.

I guess I can accept that the 1885-1889 AA is a different beast and I missed by not having an annual discount. What's exactly is the evidence for that, though? I've seen posts on this subject but it's not clear to me what the methodology was.

If I reduce the discount to 10% and say 1/3 of his value was his hitting, plus fielding as a non-pitcher, then he's on the ballot. Still, I don't see him competing with Ed Walsh. Ed's 464 IP in 1908 led the league by 140(with a 1.42 ERA, nonetheless). Caruthers 482 IP in 1885 was third in the league...

So if the discount for Caruthers should be 10%, I'll put him about #8. If 0%, then I'd have him #2, but still behind Walsh.

My problem with the AA is that the young "stars" very predictably fade out quickly after the collapse of the AA. Or so it seemed to me as I've been spending far too much time poking around But if others have studied it, I'll accept 10% for 1885-1889.

If so Caruthers will be at #8, but to be fair to King and Mullane with a little bump in those years they'll have to move onto the ballot as well. Everyone will love my ballot with more pitchers on it...though I suppose Welch will drop off this year anyway. Sigh.

I am not going to apologize for using ERA+...getting outs and preventing earned runs seem far from irrelevant to me. Win Shares and WARP3 at some level have to be taken on faith, I certainly can't run those on a spreadsheet and play around with the league factors for example. Sure, my ballot has an "ERA+ bias", others have an "extended prime bias" or a "adjusted WARP3 bias", whatever. Sounds like most of us are having fun creating our systems. I don't mind letting the consensus even it all out.

And no, Stovey wasn't my favorite pick but that's water under the bridge now.
   117. Dag Nabbit is a cornucopia of errors Posted: February 14, 2004 at 04:52 AM (#521697)
Caruthers' ERA+ (a pretty suspect stat anyway)
   118. Paul Wendt Posted: February 14, 2004 at 04:17 PM (#521699)
OCF #130:
   119. Paul Wendt Posted: February 14, 2004 at 04:20 PM (#521700)
   120. Paul Wendt Posted: February 14, 2004 at 05:07 PM (#521701)
DanG #140
   121. Chris Cobb Posted: February 14, 2004 at 05:15 PM (#521702)
Chris J. wrote:
   122. Paul Wendt Posted: February 14, 2004 at 05:43 PM (#521703)
Credit Clay Davenport with relative league values that are grounded in systematic research. That isn't the only criterion but it is an important one.

karlmagnus #155
   123. karlmagnus Posted: February 14, 2004 at 05:54 PM (#521704)
I posted this a month or so ago, but when Caruthers was sold to the Brooklyn Bridegrooms in winter 1887-8, he was paid $5000, the highest wage in the AA. Some of the NL stars (Anson?) may have been making more, but not many.
   124. Paul Wendt Posted: February 14, 2004 at 06:04 PM (#521705)
Chris Cobb #158:
   125. John (You Can Call Me Grandma) Murphy Posted: February 14, 2004 at 06:11 PM (#521706)
The best CF is . . . Roy Thomas?

He's my pick, Paul. Fielder Jones for the best AL centerfielder?
   126. John (You Can Call Me Grandma) Murphy Posted: February 14, 2004 at 06:22 PM (#521707)
BTW, who is Pete Hill?
   127. Marc Posted: February 14, 2004 at 06:25 PM (#521708)
>hit his best in high-offense contexts when it was less valuable.

I'll join Paul in piling on Chris. Not sure I see this statement as accurate. Certainly 100 RBI is not as valuable in '95 as in '85 or '05. But leading the league in RBI (being the best at the time) seems to be approximately equally valuable in any given year. It's not like he led the league with 100 RBI, his totals were appropriate to the environement and best of breed. (Using RBI purely e.g. assuming all else--BA, OB, SA, etc. etc.--also vary accordingly.)
   128. Chris Cobb Posted: February 14, 2004 at 06:39 PM (#521709)
BTW, who is Pete Hill?

My pick at present for best centerfielder of the aughts. Played in Negro Leagues from 1899-1926. Here are excerpts from his bio in _Cool Papas and Double Duties_:

"Pete Hill, a powerfully built left handed hitter who career covered the first quarter of the twentieth century, was called the most consistent hitter of his time by Cum Posey . , . He broke into professional baseball with the Pittsburgh Keystones in 1899. He joined Rube Foster on the Philadelphia Giants team in 1903 and followed Foster to the Chicago Union Giants in 1907. When Foster formed the Chicago American Giants four years later, Pete Hill was his first recruit.

"The Chicago American Giants were the most dominant team in Negro baseball during the teens, a team built on pitching, speed, and defense. . . The defense up the middle (in 1917) was unmatchable, with Bruce Petway, Bingo DeMoss, John Henry Lloyd, and Pete Hill. . . . The offense was led by Hill and Lloyd.

"Pete Hill stood 6'1" tall, weighed a muscular 215 pounds, and had power to all fields. The left handed slugger was a line-drive hitter who hit the ball where it was pitched, and hit it with authority. John Holway credits Hill with a .300 batting average for 1555 at bats, but those stats are from the Negro National League only, from 1920 to 1926, when Hill was in his 40s, and are not truly representative of his hitting ability. According to Jim Riley, Hill hit .423 in 1910, .400 in 1911, .357 in 1912, and .302 in 1914, a .371 clip for four years. . . .

"The big outfielder was a five-point player who excelled in all phases of the game. He was a center fielder with exceptional range and a strong, accurate throwing arm. On the bases, he was fast and agressive. Jim Riley said, "He was a nervy baserunner who upset pitchers and infielders like Jacke Robinson was to do a quarter [century] later. He was described as a 'restless type, always in motin, jumping back and forth, trying to draw a throw from the pitcher.'"
   129. John (You Can Call Me Grandma) Murphy Posted: February 14, 2004 at 06:50 PM (#521710)
Re: post #171

I agree with both of you, Marc. I agree with you that the best of an era is always valuable in any given year (at least for HoM consideration, that is), while Chris is correct that we have to place high-offense stats in their proper place. A 160 OPS+ in 1908 is more impressive to me than one in 1911.
   130. John (You Can Call Me Grandma) Murphy Posted: February 14, 2004 at 07:17 PM (#521711)
Chris C.:

For some reason, he evaded my radar (probably because he hasn't retired yet). At the very least, he was a very, very good player.

Thanks, Chris!
   131. Chris Cobb Posted: February 14, 2004 at 07:21 PM (#521712)
I wrote: hit his best in high-offense contexts when it was less valuable.

Marc wrote: I'll join Paul in piling on Chris. Not sure I see this statement as accurate. Certainly 100 RBI is not as valuable in '95 as in '85 or '05. But leading the league in RBI (being the best at the time) seems to be approximately equally valuable in any given year. It's not like he led the league with 100 RBI, his totals were appropriate to the environement and best of breed. (Using RBI purely e.g. assuming all else--BA, OB, SA, etc. etc.--also vary accordingly.)

As I see it, Thompson's accompishments turn on the argument that he was a truly extraordinary hitter. Given his short career, indifferent fielding value, and injury history, he has to be a very great hitter to be a worthy candidate for the HoM. It's my view that Thompson's performances in very favorable hitting environments makes his career offensive accomplishments look better than they were. This argument has two parts:

1) Eric C, I think, pointed out that Thompson's three great years all came during three of the best offensive environments in baseball history, 1887, 1894, 1895. It appears to be the case that it is easier for very good hitters to exceed league-average performance in favorable offensive-context seasons. So I am less impressed by Thompson's OPS+ and RC scores in those seasons.

2) Yes, leading the league in RBIs (or some other more meaningful stat) is more or less equally valuable from season to season. If you judge players by black-ink scores it is indeed equally valuable. But in actuality, each league-leading RBI season is _not_ equally valuable, some are better than others, depending upon how far it exceeds that of other players in the league, and how many runs it takes to win a game in a given season. It looks to me like Thomspon a) was not all that much better than other hitters in his league when he was putting up his big numberrs in the mid-90s (I accept 1887 as a great season) and that, if you scale his runs created according the runs per win in his environment, his totals are not exceptional in comparison to league leaders in other seasons.

If you put points 1 and 2 together, it looks to me like Thompson put up his best numbers at a time when it was easier to exceed league averages and at a time when creating many runs does not lead as directly to creating many wins as it usually does. This conduces to Thompson being overrated. He's overrated by some voters because they're impressed by the raw numbers. He's overrated by other voters because WARP, which isn't linked back to actual wins as WS is, doesn't take account of run contexts as fully as WS does (see Ed Williamson 1884 as an example).

This case is inferential, based on Eric's work and my understanding of how WS works: I haven't gone through and re-calculated offensive WS for Thompson season by season to see if that's the reason why he looks pretty ordinary, or if it has more to do with less defensible factors like WS's low zero point in relation to actual replacement level.

Given that I've found WS to be a trustworthy offensive metric in most cases and that by its measures Thompson is not close to being ballotworthy, I haven't taken the time to study him intensively to see if _maybe_ there's some horrible mistake. Since my view is not the consensus view, I acknowledge that the consensus might have it right by indicating that if my _slightly_ casual dismissal of his value might begin to make a difference in whether Thompson or not is actually elected, I will then make absolutely sure of my position (I hope those who have Charlie Bennett about where I have Sam Thompson will be equally scrupulous). Of course, anyone else who wants to promote Thompson's case with me and other doubters is _welcome_ to do this analysis and post the results. I'd be interested to see them. But I'm not interested enough to do that work myself, when I'm still busy refining my analysis of pitchers and learning about the great Negro-League players 1900-1930. Those seem to me to be more important projects.

<i>Charge: Incenting votes for Big Sam Thompson.
   132. Paul Wendt Posted: February 14, 2004 at 10:20 PM (#521713)
   133. Yoenis Cespedes, Baseball Savant Posted: February 15, 2004 at 12:20 AM (#521714)
When is Home Run Johnson ballot-eligible? He'd probably make my top 5-7.
   134. OCF Posted: February 15, 2004 at 01:51 AM (#521715)
My two cents on the Sam Thompson discussion that Chris Cobb got into:

Here's an idea. Maybe it's an attempt to reinvent the wheel, but it is another approach to the relationship of high-scoring and low-scoring times. The source is a Stats Inc. Major League Handbook, which records Runs Created, Runs Created/27 outs, league runs/game, and park factor. By dividing RC by RC/27 outs, you get the number of outs used. Proceeding from that:

In 1887, Thompson is credited with 152 RC at a rate of 11.46 RC/G. That's an equivalent of 13.3 games worth of outs. His enrvironment was a 6.08 R/G league and a 100 park factor. So an average player would have created 81 runs in those outs, leaving Thompson as creating 71 runs above average (or 87 runs above 80% of average if you would prefer that).

In 1894, that was 137 RC at a 13.05 rate in a 6.94 environment: 64 RC above average.

In 1895, that was 154 RC at an 11.96 rate in a 6.56 environment: 70 RC above average.

For contrast, let's cherry-pick three years from Frank Chance's career:

In 1903, 116 RC at a 9.96 rate in a 4.49 environment: 64 RC above average.

In 1905, 105 RC at a 9.54 rate in a 4.11 envoronment: 60 RC above average.

In 1906, 112 RC at an 8.34 rate in a 3.61 environment: 64 RC above average.

So far, ranked against average, Thompson's three years are 71, 64, and 70, while Chance's three years are 64, 60, and 64. But now I propose equalizing the run environments to something arbitrary, say 5.00 runs per game. Take the figures above and multiply the three Thompson years by 5.00/6.08, 5.00/6.94, and 5.00/6.56, and multiply the three Chance years by 5.00/4.49, 5.00/4.11, and 5.00/3.61.

Do that, and we get: Thompson 59, 46, 53; Chance 71, 73, 88. If the comparison is to 80% of average instead of average, the result is: Thompson 72, 57, 66; Chance 82, 84, 102.

This is being unfair to Thompson in several ways, the first of which is that Chance had longer seasons available to him than Thompson did. I also think that it is an overcorrection in favor of the low-scoring times over the high-scoring times. But is it that much of an overcorrection? By using James's RC instead of OPS, we are being friendlier to Chance's OBP-heavy offense than we are to Thomspson's SLG-heavy stats, and we are giving Chance quite a bit of credit for his base stealing.

But what of it? Who really had the best three offensive years - Thomspson or Chance?
   135. OCF Posted: February 15, 2004 at 02:09 AM (#521716)
My pick at present for best centerfielder of the aughts....John Holway credits [Pete] Hill with a .300 batting average for 1555 at bats, but those stats are from the Negro National League only, from 1920 to 1926, when Hill was in his 40s, and are not truly representative of his hitting ability.

Too late now, but this is another case in which in might have been nice, when this project was set up in the first place, to have eligibility keyed to a certain fixed age (like 45) rather than five years after retirement. We are right now discussing outfielders of the aughts. We've elected Flick, we're debating Sheckard, and Magee will be on the ballot soon. Hill's career started in the same year as Crawford, and Crawford will be eligible before Hill is. These are Hill's contemporaries.
   136. Chris Cobb Posted: February 15, 2004 at 02:44 AM (#521717)
Thanks, OCF, for that analysis of Thompson!

Too late now, but this is another case in which in might have been nice, when this project was set up in the first place, to have eligibility keyed to a certain fixed age (like 45) rather than five years after retirement. We are right now discussing outfielders of the aughts. We've elected Flick, we're debating Sheckard, and Magee will be on the ballot soon. Hill's career started in the same year as Crawford, and Crawford will be eligible before Hill is. These are Hill's contemporaries.

Our method of determining eligibility certainly creates some placement trouble for Negro-League players because they played so long. It strikes me that we should be making an effort not only to ascertain when Negro-League stars are eligible, but to identify when they are reaching the end of their primes. If we do that, we'll be better positioned to assess them and their white contemporaries.
   137. Jim Sp Posted: February 16, 2004 at 02:39 AM (#521719)
In my system a discount is from a very high replacement value, not 0.

For example, a 20% discount is not taking a 140 ERA+ and making it 112. It's taking a 140 ERA+ and making it 130 (20% of the difference from 90).

So when people say a 20% discount is too much for the 1885-1889 AA, is this what they mean?

ERA+ (a pretty suspect stat anyway)...W/L% are the most important single pitcher stat.

gotta disagree with you there. the problems with W/L% are greater than the biases in ERA+. to find a few examples: Jesse Tannehill, Chief Bender, Jack Stivetts, Art Nehf, Charlie Buffinton, Jack Chesbro, Herb Pennock...lots of guys rack up impressive W/L without impressive ERAs, that's exactly where you find the questionable HoF selections. Playing for good teams is not going to put them on my ballot anytime soon.
   138. EricC Posted: February 16, 2004 at 03:02 AM (#521720)
Which standard deviations have you calculated?

Paul- (1) Standard deviations of Win Shares per plate appearance for postion players (all positions). (2) S.D. of WS/PA for each individual position. (3) S.D. of ERA+ for pitchers.

By the way, I fit the performances to the extreme right tail of a normal distribution ("bell curve") for the entire North American population, not to a normal distribution for the population of baseball players. This way of looking at things was inspired by a Bill James essay in one of the old Abstracts about how talent is not normally distributed in baseball. I'm not sure if it has been incorporated in this manner into a rating system before.

Given one measure of "established ability ...", the question may reduce to how you have defined position and era.

I fit the performances for each year to their own bell curve.
   139. Max Parkinson Posted: February 16, 2004 at 03:08 AM (#521721)
Before I get into my prelim ballot, which will probably change before its final incarnation, the issues that I've been grappling with this week are:

Ed Walsh
   140. Max Parkinson Posted: February 16, 2004 at 07:30 AM (#521723)
Please scratch my inadvertant slagging of Addie Joss' defensive ability. According to Clay Davenport, he was an excellent fielder, probably third to Walsh and Nig Cuppy. I meant McGinnity. And change the end of that sentence to read he might fall as far as 4th.

Sorry for the insult, Mr. Joss.
   141. Marc Posted: February 16, 2004 at 03:09 PM (#521724)
Primey for #184 (Max Parkinson). Great analysis. Pike and Thompson have been sorely underrated. Collins and Sheckard will have their time. But all four were "the best of their time." No matter what the parallel discussion of RC/27 says, Chance never was. He was overshadowed by a dozen or so guys (including Bobby Wallace) who played in the same environment as Peerless.
   142. Chris Cobb Posted: February 16, 2004 at 03:43 PM (#521725)
Max, thanks for a most searching analysis!

I'm left with a big question about Wallace, though. How does his profile compare to other current eligibles? I'm perfectly convinced that he doesn't match up on this prime measure, to most of the candidates we've elected so far, but no one on the ballot at present matches up particularly favorably to most of the candidates we've elected so far. That's why they're still on the ballot :-) . I'd find it helpful to see how his profile compares to Van Haltren, Duffy, Jennings, Williamson, Long, and Childs, especially, if you'd have a chance to post it.

Also, given that WARP2 is the primary measure used in constructing these profiles, I'm not convinced that Thompson's profile is accurate in comparison to his peers. In adjusting defensive statistics for all time, WARP advantages right fielders over left fielders (see thread on WARP3 for confirmation of this), and I believe (although I can't cite hard evidence) that in all-time adjustments WARP also privileges power statistics over non-power contributions to offense. In both cases, WARP2 privileges Thompson over most of his peers when converting them to "all-time" standards.
   143. Dag Nabbit is a cornucopia of errors Posted: February 16, 2004 at 04:46 PM (#521726)
One thing I went looking up the other day - the different levels of defensive support pitchers in contention (for my ballot at least - Walsh, McGinnity, Griffith, Caruthers, Joss, Waddell, Willis) received throughout their careers. My idea: take a pitcher's individual IP in a season, divide by the team's total IP, & multiply by the team's defensive win shares for the season. Then add those altogether for every season the pitcher pitched for, & figure how many Def. Win Shares the pitcher received per 1000 innings in his career. Makes sense to me. The results, from worst to best defensive support:

1. Addie Joss 30.6/1000IP
   144. Howie Menckel Posted: February 16, 2004 at 05:29 PM (#521727)
   145. Chris Cobb Posted: February 16, 2004 at 05:33 PM (#521728)
Chris J., in deciding how to act on these findings you might consider two things:

1) While fielding support tends to even out over the course of a pitcher's career, it is much less consistent season by season, so that the shape of a pitcher's peak is much more influenced by defensive support than a pitcher's career. Caruthers enjoyed fabulous defensive support during his peak years in St. Louis and suffered atrocious defensive support later in his career, for example. Willis likewise enjoyed fabulous support during his early years in Boston. I don't know how much you consider peak in rating pitchers, but since many people do consider peak more heavily than they do career, the seasonal variability of defensive support is worth considering.

2) The win shares normalization of the pitching/fielding ratio of defensive win shares is probably at work here, esp. with respect to Caruthers relative to the rest of the group. I would trust WS to show that the fielding support, for their careers, of Joss, Willis, Waddell, McGinnity, and Walsh was not all that great, but I wouldn't trust WS to compare the fielding support of these pitchers to that of 1880s pitchers or 1930s pitchers.
   146. Dag Nabbit is a cornucopia of errors Posted: February 16, 2004 at 05:49 PM (#521729)
Chris Cobb - both occurred to me, especially the latter. Not too concerned about the 2nd one, though, as I mainly tryto use this method to look at a pitcher's peers & most of these guys played at about the same time. As for the first comment, good point, but for now I'm not really sure how to deal with that in a single way I like. Shouldn't be that difficult, but for now my mind's a little hazy on it.
   147. Max Parkinson Posted: February 16, 2004 at 06:08 PM (#521730)
Chris Cobb,

I'm terrible at HTML formatting, so bear with me if this looks awkward. The first number is the player's peak position, highest ranking attained 1-10. The second number is the number of years on the "Top 10" list. Every player's years on the leaderboard were consecutive, save for King Kelly, who appeared on for one year, off for one, and then back on for 10 more.

   148. Dag Nabbit is a cornucopia of errors Posted: February 16, 2004 at 09:18 PM (#521731)
Didn't take too long - found a season-by-season version I liked. Figure the total Def WS in a league each year, divide by total league IP, multiply by a pitcher's own IP & see how many Def WS he should've had with league average D behind him & compare with the Def WS he actually had him. (Note: I made an error in post #189. Walsh's D wasn't that good - should've been 32.0WS/1000IP) Results (make your own timeline adjustments here). Plus is good, negative means bad D:

Bob Caruthers:
   149. Marc Posted: February 16, 2004 at 10:07 PM (#521732)
This kinda plays into DIPS, doesn't it, Chris J? Kinda makes you wonder if the whole history of baseball is rigged for and against certain pitchers. And makes you wonder if we shouldn't maybe be takin' another look at guys like Weyhing and Breitenstein. If they basically got screwed in life and got on the wrong team, then there'd be nothin' they could do to put up decent numbers.

On the other hand, I could look at Chris' post and conclude that a tax cut would solve the problem. Either that or knock Caruthers down a couple more notches.
   150. Paul Wendt Posted: February 17, 2004 at 02:28 AM (#521734)
Max Parkinson #184
   151. Chris Cobb Posted: February 17, 2004 at 02:51 AM (#521735)
TomH asked for batting and fielding WS, season by season, for Thompson, Sheckard, Ryan, Van Haltren, and Duffy. Here they are, with additional info on when the totals are among top 5 in the league.

BWS ? FWS (in top 5)

Sam Thompson, 85-96
   152. Chris Cobb Posted: February 18, 2004 at 03:24 PM (#521737)

Thanks for the additional information on how current candidates rank over ten-year stretches! I'm still digesting the implications, but it's very helpful data.

As to your question re: WARP2, I was under the impression that the main culprit was FRAR in the Davenport system as opposed to FRAA. That is, he balances players around the average of their positions very well, but the relative difficulty of those positions seem off for early baseball. Based on Tango's input on this board, I have been using FRAA in these "all-up" calculations.

You're right, FRAR is the main culprit. I had neglected to note that you were using FRAA. I'm not sure if RF FRAA are adjusted differently for all time than LF FRAA are.
   153. Chris Cobb Posted: February 18, 2004 at 03:42 PM (#521738)
However, I should add that simply using FRAA doesn't solve the problem. Since it compares players only to other players at the same position, its numbers do not offer a valid basis for comparison between players at different positions. Using them in that fashion will tend to overrate good fielders at less important defensive positions.
   154. Al Peterson Posted: February 19, 2004 at 04:12 PM (#521740)
Question for the group. I'm trying to evaluate 1880's vs 1890's to try and decide on the best candidates not in the HOM yet. Wanted to look at team composition so does anyone know roster sizes at this time?

For an example year like 1887, my guess would be 13 - 8 position players, 1 backup C, 1 utility player, 3 pitchers. By 1897 more like 8 position players, 1 backup C, 1 backup IF, 1 backup OF, 4 pitchers for a total of 15.

Do those numbers sound reasonable? This helps with me finding a population size of ML players when comparing 1 league vs. 2 league strength.
   155. jimd Posted: February 19, 2004 at 07:20 PM (#521741)
I don't know the actual roster limits throughout MLB history. I suspect that during the 19th century, the only limits were financial (how many players were you willing to pay to travel), but I could be wrong. Anyone know when formal roster limits were imposed?

I've mentioned before my definition of a regular (played in more than half the maximum games-played, or played defensively in more than half the games-played of the league leader, with pitchers also qualifying at more than half of the games-started or innings-pitched of the league leaders). I have a definition of an "irregular" which is similar but uses one quarter instead of one half.

The table below takes the number of players that I define as either an "irregular" or regular and divides it by the number of teams to get an "effective roster size". This is the number of players that a team used fairly regularly during the season. (Modern teams have been getting around the 25 player limit by roster juggling since the 1980's, to the extent that an average team now uses 28 players per year at the level of "irregular" or more.)

Effective roster size:

1871-80 10 09 10 10 10 10 10 10 11 11
   156. Al Peterson Posted: February 19, 2004 at 09:16 PM (#521742)
Thanks jimd. That's just the sort of thing that will be useful.
   157. Dag Nabbit is a cornucopia of errors Posted: February 20, 2004 at 06:13 PM (#521743)
Re: My comment in post #194 where I tried to figure out how much defensive support a pitcher got relative to league average in terms of defensive win shares.

Went through & figured that for a total 29 pitchers. Note: this doesn't include any NA experience as I have no idea what the defensive win shares were for those leagues. Also, some pitchers' stats for 1889 NL may be messed up (long boring story). For anyone curious, here are the results (make your own era adjustments):

1. Kid Nichols +36.5
   158. KJOK Posted: February 20, 2004 at 07:23 PM (#521744)
I think we've discussed this before, but is there any reason to think that win shares over-rates Hugh Duffy?

A few I had been using:

1. Outfield Defense in general over-valued by fielding metrics for 19th century.

2. More variability in W/L vs. Pythagoras for 19th century (probably due to shorter schedules?), and Duffy's teams seemed to have been often on the "high" side, which means there were more "unassigned" win shares to assign to players.

3. 445 ft to LC field, which might mean he was able to get more left field putouts than his contemporaries, which would also drive his fielding metrics up.
   159. Paul Wendt Posted: February 22, 2004 at 01:43 AM (#521745)
Chris J #190
   160. Chris Cobb Posted: February 22, 2004 at 03:17 PM (#521746)
Is this a tautology in the Win Shares system?

Chris J., the analysis you provided is fascinating as usual. In thinking about Paul's question, I see two possible explanations. One is that the WS system is so set to normalize the ratio between fielding and pitching that the most terrible defensive teams, ones that are a good deal below any others, will necessarily appear worse in both pitching and fielding. It's hard for me to believe that a quirk of the system isn't contributing to this effect. James notes that bad offense has a lower bound -- 0 -- but bad defense does not. If either the fielding or the pitching is outrageously bad, James' system may not be sensitive enough to detect which part is truly responsible for dragging the defense down.

However, a second possibility is that, since pitching and fielding are interdependent, pitching that is suffiently bad _produces_ bad fielding, and vice versa. It's easy enough to see how pitchers who don't have confidence in their fielders could pitch worse: they'll nibble at the corners, trying to make the perfect pitch, get behind in the count, and give up more walks and home runs, which will appear as their responsibility, even though the situation that influenced their pitching was set up by bad defense. Similarly, if the pitchers are being hit hard all the time, defensive efficiency will go down, and that will reflect poorly on the fielders. They may also press, try to do too much, have to make more high-risk, error-prone plays like throwing out runners on the basepaths, etc. They look bad for responding to difficult situations created by bad pitching. And at what point does the team start to give up and no longer play its best?
   161. Marc Posted: February 23, 2004 at 01:05 AM (#521747)
>However, a second possibility is that, since pitching and fielding are interdependent, pitching that is
   162. Chris Cobb Posted: February 23, 2004 at 02:26 AM (#521748)
OTOH, wouldn't DIPS say that it is (almost) always versa? (Bad fielding produces bad pitching?)

DIPS would say that bad fielding produces bad performance for pitchers in traditional pitching stats like ERA and h/9. It doesn't assert (as far as I understand the DIPS view) that bad fielding leads to worse performance by pitchers even in their ostensibly defense-independent statistics: walks, home runs, and strikeouts.

Chris J.'s findings in WS suggest that bad fielding may negatively influence pitcher performance even in these "defense-independent" areas, at least in a small way.

Studies of hits allowed on balls in play (generally used as a measure of fielding quality) have shown that bad pitchers as a group (identified as those pitchers who don't last in the majors) allow more hits on balls in play that good pitchers, so we know that the influence works in the other direction as well.
   163. Dag Nabbit is a cornucopia of errors Posted: February 23, 2004 at 04:27 AM (#521749)
Studies of hits allowed on balls in play (generally used as a measure of fielding quality) have shown that bad pitchers as a group (identified as those pitchers who don't last in the majors) allow more hits on balls in play that good pitchers, so we know that the influence works in the other direction as well.

It would have to work in the other direction as well because it's a zero sum game. If the bottom 10% pitchers are averaging a -X, then the top 90% would have to average +X.

I'm also reminded that I looked up the Hit Delta (the Voros number) for all pitchers in James's top 100 pitcher list on their prospectus playing cards - about a dozen had given up more hits than expected - & half of them were under +10. The total sum of all the plus pitchers could be neutralized by the combination of Walter Johnson & Luis Tiant (& Tiant didn't have a particularly high H=Delta, it's just that his plus Johnson's brought the others to zero).

One thought has popped into my head: in an era when pitchers are only throwing it with everything they have 15% of the time or so, pitchers are going to end up more reliant on their defenses & thus pitchers of this era in general are going to need
   164. Howie Menckel Posted: February 23, 2004 at 12:14 PM (#521751)
Joe, 637 IP is about a 25 pct difference in not only their careers, but those of most pitchers of the era.
   165. Jeff M Posted: February 23, 2004 at 01:16 PM (#521752)
How can anyone rate Ed Walsh highly and not Addie Joss?

They were basically the same quality, and Walsh threw 600 more innings. 1 2/3 seasons is the difference between finishing at or near the top, and finishing 23rd or something? That blows me away.

Some of it is probably attributable to how close the top 23 players are to each other. Then there's the domino effect. Walsh throwing more innings also means Walsh compiled some additional kudos (more wins, more complete games, more shutouts, for instance).

And forgive me for a HoF Digression, but there's this: the average HoF pitcher (excluding the ridiculously good, e.g., Johnson, Mathewson, Alexander and Young; and the ridiculously bad, e.g., Bender, Chesbro, Haines and Marquard), has the following 3-year peak WARP1, 5-year consecutive WARP1, 7-year WARP1, WARP1 per 1000 IP and career WARP1 (all season-length adjusted, except the per 1000 IP):


33.4 (3 year)
   166. OCF Posted: February 23, 2004 at 05:10 PM (#521753)
I'm one of the ones targeted by Joe's #213, as I had Walsh first and Joss not on the ballot. One key thing: I don't see them as having essentially the same quality per inning; I see Walsh as better.

The following tables list RA+ and RA+ turned into Pythagorean equivalent record for Walsh and Joss:

   167. jimd Posted: February 23, 2004 at 06:11 PM (#521754)
It's the IP, Joe. Walsh was a horse; Joss was not. If you've got two position players putting up the same OPS+, but one plays 154 games and the other 120 almost every year during their peaks, who's more valuable to his team? Like Collins vs McGraw, it's not close.
Page 2 of 2 pages  < 1 2

You must be Registered and Logged In to post comments.



<< Back to main

BBTF Partner

Support BBTF


Thanks to
Martin Hemner
for his generous support.


You must be logged in to view your Bookmarks.

Hot Topics

Manny Ramirez
(15 - 7:31pm, Jan 28)
Last: theboyqueen

Toby Harrah
(15 - 2:32am, Jan 28)
Last: Ardo

Tommy John
(17 - 2:20am, Jan 28)
Last: Ardo

2016 Hall of Merit Ballot Discussion
(65 - 1:50am, Jan 28)
Last: Ardo

Wally Schang
(12 - 1:40am, Jan 28)
Last: Ardo

Ernie Lombardi
(53 - 4:18pm, Jan 27)
Last: The District Attorney

Most Meritorious Player: 1901 Ballot
(9 - 11:17am, Jan 27)
Last: DL from MN

Luke Easter
(108 - 10:51pm, Jan 26)
Last: Kiko Sakata

Most Meritorious Player: 1902 Discussion
(11 - 9:41am, Jan 26)
Last: DL from MN

Vic Willis & Sam Leever
(58 - 2:28pm, Jan 25)
Last: Kiko Sakata

Ernie Banks
(71 - 8:32am, Jan 24)
Last: djrelays

Trevor Hoffman and Billy Wagner
(77 - 2:22am, Jan 23)
Last: Rob_Wood

Rube Waddell
(30 - 5:45pm, Jan 21)
Last: W. G. Braund

Most Meritorious Player: 1901 Discussion
(39 - 3:36pm, Jan 18)
Last: TomH

Ben Taylor
(87 - 11:44am, Jan 17)
Last: Chris Cobb


Page rendered in 0.5525 seconds
43 querie(s) executed