Baseball for the Thinking Fan

Login | Register | Feedback

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

Monday, May 24, 2004

1927 Ballot Discussion

The top new candidates (thanks DanG and Chris Cobb!):

WS W3 Rookie Name-Pos (Died)
287 73.1 1907 Ed Konetchy-1b (1947)
227 52.6 1908 Dode Paskert-CF (1959)
205 50.7 1910 Hippo Vaughn-P (1966)
189 48.4 1908 Slim Sallee-P (1950)
180 42.2 1910 Duffy Lewis-LF (1979)
161 44.9 1911 Ray Caldwell-P (1967)
155 38.0 1909 Dots Miller-1b/2b (1923)
143 37.9 1911 Lefty Tyler-P (1953)
106 28.8 1914 Braggo Roth-RF (1936)

Negro Leaguers
HF%—from expert voting in _Cool Papas and Double Duties_
BJ—Bill James positonal ranking in _NBJHBA_
MVP—Sum of Bill James’ best player and best pitcher awards and John Holway’s MVP and best pitcher awards
All-Star—number of times player designated as seasonal all-star by John Holway

HF% Career  Name-pos (born)          BJ    MVP All-Star
68% 1901-25 Pete Hill-CF/LF (1880)  #4 lf   2     5*
08% 1909-21 Frank Wickware-P (1888)         2     1*
00% 1914-21 Horace Jenkins-OF       (??)    1     2*

Players Passing Away in 1926

HoMers
Age Elected
75 1914 Cal McVey-C/1B
50 1924 Eddie Plank-P

Candidates
Age Eligible
66 1901 Bill Hutchison-P
64 1899 George Pinkney-3B
63 1900 Danny Richardson-2B
60 1903 Lou Bierbauer-2B

Joey Numbaz (Scruff) Posted: May 24, 2004 at 10:12 AM | 387 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 4 pages  < 1 2 3 4 > 
   101. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 25, 2004 at 07:16 PM (#643676)
My approach to the 1927 ballot will be radically different from 1925 and 1926.

First of all, I bought the Win Shares digital update so I am now including them in my calculations.

Second, I've decided to try to rank players statistically rather than by gut-feeling comparisons of some basic numbers. I will try to evaluate 30 players per election. For each of them, I consider eight categories. I take the player's five best years in three stats: OPS+/ERA+, WARP2 straight-line adjusted for season (adding them together to get 5-year-peak WARP), and WS adjusted for season length (added to get 5-year-peak WS), all adjusted for league difficulty (using an 11% AA discount). These make the three peak categories. Next, I take the player's career season/league adjusted WARP2 and WS, and career league-adjusted OPS+/ERA+ minus 80 (replacement level) times the number of seasons played. These are the three career categories. Finally, I modified the Ink tests for my personal preferences (3 points for BA/OBP/SLG and ERA/WHIP/K, 2 points for HR/R/RBI/SB and IP/W/SV/CG, 1 point for H/BB/TB and W-L%/GS/SHO). In each of these categories, I calculate the mean and standard deviation for the 30 players I'm considering and then get z-scores in each category for each player. Finally, I add these z-scores to get a final score. I'd love to hear thoughts/criticisms from the statisticians among us on this method so as to improve it.

Third, I'm trying to give the same mathematical consideration to Negro Leaguers by turning KJOK's MLE's into these usable numbers. I am finding this a confusing and challenging task, and will be posting again and again until I find a way to do it that makes sense.

I will post a preliminary major league ballot right after this. Hill, Foster, and Monroe will not be on it until I figure out what to do about them, although they will be given full consideration for my final ballot.
   102. sunnyday2 Posted: May 25, 2004 at 07:18 PM (#643682)
Zap, let me ask the question another way, but first I'll just lay my bias out there. I don't like rate stats because they're too confusing. i.e. to compare Joss and McGinnity by virtue of their rate and then make adjustments from there. It's too hard.

What makes more sense is to start with something like WS and then factor in rates and whatever as fine-tuning. It just makes more sense to go that way.

So my question, well, I guess I already said it as a statment. Why not start with WS or WARPies and work from there?
   103. sunnyday2 Posted: May 25, 2004 at 07:21 PM (#643689)
andrew: 1. you want the kids at the library to be able to read the HoM don't you? and 2. if you ever run for public office, the opposition researchers will be googling you all over the place, so the censorship is for your own good. I mean you don't really want them to know about your vocabulary!? (insert smiley face just in case this isn't funny, then you'll know that it was really meant to be, but just wasn't a competent joke) ;-)
   104. andrew siegel Posted: May 25, 2004 at 07:27 PM (#643704)
I'll answer for Zap:

because for pitchers, WS and WARP are simply rate stats adjusted for quantity and an inappropriate replacement value.

One could, I suppose start with WS or WARP and adjust the replacement level within the system, but why bother? Why not start with the ERA or RA data and adjust that for workload and whatever replacement level you find appropriate.
   105. John (You Can Call Me Grandma) Murphy Posted: May 25, 2004 at 07:29 PM (#643707)
Like a broken record, I will continue to ask how an offensive stud who was an above-average defensive 2B and dominated his position for a decade during the hardest era ever for middle infielders doesn't register on 2/3 of the ballots?

I'm just happy that Doyle didn't receive more votes because he had the advantage of playing during an easier time for infielders (which helped create more WS for the Laughing One).

4) The Manager/GM thing is John Murphy's idea that McGinnity's ERA would have been lower if he'd pitched fewer innings, like Joss did. In other words, having John McGraw as a manger made McGinnity's ERA go up. I'm not convinced.

All you have to do is go through baseball history to see that I'm right (not that this is my brainchild). But regardless, McGinnity was pitching quality innings when a non-suitable replacement was filling in for Joss most of the time.
   106. Max Parkinson Posted: May 25, 2004 at 07:56 PM (#643752)
Unfortunately, I was away last week, accounting for much of Hughie Jennings’ drop in the polls.

My prelim 1927 Ballot, with congrats going out to Pete Hill and Ed Cicotte…

1. Hughie Jennings – still the only eligible player to be the best player in the game. The arguments against giving him this credit rest on Ed Delahanty being a better hitter than Jennings, ignoring that Jennings was the best defensive SS in the world, and Delahanty was an adequate corner outfielder. Craig B used to post about the Orioles fantastic runs against numbers year in and year out during the ‘90s, and their decidedly average pitching. Infield defense had to play a pretty large role, no?

2. Bob Caruthers – flipping back and forth with Thompson at 2/3. I think that discussion here has passed the point where I can be useful…

3. Sam Thompson – The best hitter on the board. Someone please convince me that Joe Jackson was a better player than Big Sam, because I just don’t see it.

4. Jimmy Sheckard – The good bat, great glove man. Lots of bags in the aughts, when that was oh so important, and could still belt the ball enough to finish on power leaderboards. To be fair to Pete Hill, he’s just slightly behind Sheckard on my list, although the tightness of the ballot makes that 4 spots.

5. Bobby Wallace – Poised for a decade-long run as the best shortstop in the game, and then Pittsburgh converted their right fielder to the middle…

6. Dickey Pearce – He made the MP HoM last year, and this position is my best compromise between my two prevailing opinions on him: 1. He was one of the best players in the world for damn near a decade. 2. He still wasn’t as good a player as most of the others on this ballot.

7. Lip Pike – We’ve elected everyone better than him from the NA (and a couple who weren’t as good). Are we done with ‘70s players?

8. Pete Hill – based on everything I’ve read, I’ve put him just a little behind Sheckard. That gets him in this year as far as my HoM is concerned.

9. Ed Cicotte – Above Joe Jackson, you ask? Well, yes. Cicotte had a longer career, their primes were similar, but Cicotte’s was a few years longer, and Eddie gets a lot more MVP points than Joe. (2nd in 1917 and 3rd in 1919, as opposed to Jackson’s 4th place in 1911.)

10. Joe Jackson – If you only play for 10 years (regardless of the reason), you better have a couple of MVP-type seasons if you want to get up the ballot.

11. Fielder Jones – The best defensive outfielder that the game saw until Speaker, and a pretty good hitter in his own right. Note that no credit is given for his “blacklist” years.

12. Jim McCormick – Again, I’ll take the player who has the legitimate claim as Best in the Game ahead of the player that doesn’t, in this case McGinnity and Griffith.

13. George Van Haltren – A long and good, if rarely great, career in centre, with some pretty good pitching to boot.

14. Clark Griffith – The fourth best pitcher of the ‘90s just ekes out…

15. Joe McGinnity – the fourth best pitcher of the ‘00s. The other 5 (Young counts twice) are all automatics, and Griffith has already made my personal HoM. McG has a pretty good shot to get in before 1933.

16-20. Bond, Beckley, Monroe, Nash, Ryan
21-25. Williamson, Whitney, Cross, Foster, Buffinton
26-30. Konetchy, McGraw, Waddell, King, Seymour
31-35. Long, Force, Jimmy Williams, Childs, Duffy
36-40. Willis, Tannehill, Tenney, Doc White, Griffin
   107. Zapatero Posted: May 25, 2004 at 08:45 PM (#643835)
OK, I've normalized for league, using two methods. In each method, the gap between the two pitchers narrows. In each case, I still believe Joss is more valuable to the team.

Method one: I took the player's baseball-reference ERA+ score (which is league and park adjusted) and assumed league average was 3.00. This led to a 2.33 8-year ERA for McGinnity and a 2.02 score for Joss. (The formula is 3.00/ERA+) Once again assuming Joss pitches 277.3 IP and McGinnity pitches 368 IP, Joss now gives up 62 runs while McGinnity gives up 95 runs. That's a 33 run difference, but of course, we also have to account for the 90.6 IP. If Joss's partner pitches to a 3.28 ERA, then Joss plus the 3rd guy equals McGinnity. By definition, 3.28 ERA is a 91 ERA+. So Joss + average replacement is clearly superior to McGinnity, Joss + replacement-level replacement is worse than McGinnity, and Joss + somewhat below average, but not terrible is McGinnity's equal.

Method two: I took the player's FIP score (the hardball times formula) and substituted 2.00 for the constant. This would work just as well with any constant, but this way the numbers look right. McGinnity's FIP-derived ERA is 2.29, Joss's is 1.78. McGinnity gives up 94 runs in 368 IP, Joss gives up 55 in 277.3. The difference is 39. Our third pitcher needs to meet only a 3.85 ERA in this set-up for Joss + third guy to match McGinnity. A player with a FIP score like that would have a 5.05 ERA in today's run-scoring environment. Not quite replacement, but close.

The pitchers are much closer now than on my previous example from #62. I'd still take Joss, and the reason is that a pennant-contending team should be able to find a 91 ERA+ or better pitcher to pitch the extra 90.6 innings. (I'll point out that 91 ERA+ is the highest result I've gotten -- every other method has allowed the third pitcher to be even worse than that and still Joss + 3rd guy is better than McGinnity).

McGinnity helps a bad team more than he helps a good team. Joss helps a good team more than he helps a bad team. Since they're so close to one another in terms of runs prevented, I'd rather have the guy who would help me win a championship than the guy who merely gets me up from bad to respectible. It takes real greatness to help a very good team get better. I'd still take Joss over McGinnity.

That said, Joss is likely dropping in my ballot after this. Accounting for the league, I still think he was truly great, but there are other players of higher merit than Joss.
   108. PhillyBooster Posted: May 25, 2004 at 09:05 PM (#643871)
So Joss + average replacement is clearly superior to McGinnity, Joss + replacement-level replacement is worse than McGinnity, and Joss + somewhat below average, but not terrible is McGinnity's equal.

It looks like we do not disagree. My point is simply that when looking at same-year replacement (rather than post-retirement replacement), then "replacement-level replacement" is the only appropriate point of comparison.

Also, I give McGinnity extra credit for several "Cy Young award" seasons that Joss does not earn, but that is a more subjective bonus that I can't really quantify.
   109. Zapatero Posted: May 25, 2004 at 09:33 PM (#643925)
I don't think the replacement-level replacement is the only appropriate point of comparison. We're talking about pennant-contending teams here. Pennant contenders aren't full of replacement-level players. If they were, the Texas Rangers would have won World Series after World Series in the past decade, given their stars (and the Mariners before them).

Pennant contenders and World Series winners carry average players to complement their stars. Addie Joss pitching 277.3 innings and an average guy pitching 90.6 will allow fewer runs than Joe McGinnity pitching 368. I think that's an appropriate point of comparison.

As I've already said, I think a guy who makes a good team great is tougher to find than a guy who makes a lousy team decent. That scarcity makes him more valuable.
   110. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 25, 2004 at 09:56 PM (#643949)
First half of preliminary 1927 ballot, excluding Negro Leaguers until I figure out how to assess them.

1. Joe Jackson (7.75). Plus ca change. Jackson is above average for this ballot in every category except Black Ink, and again, it’s not his fault that Speaker and Cobb played in the AL. His 183 five-year peak OPS+ dominates the ballot (z=2.21) and he leads in my Gray Ink (z=1.77), while he’s second in lifetime OPS+*years (z=1.54) and third in peak WS (z=1.05). I do see him as a no-doubt-about it HoM kinda guy.

2. Bob Caruthers (7.35). Well, he jumps a few light years up my ballot. I know WS may overrate 1880’s pitchers, but after applying the AA discount he tops out at 40 pitching WS for a season, which seems reasonable. As a two-way player matched only by the transitional Babe Ruth, Caruthers’ 1885-89 peak just dominates everyone else. My 287 calculated Win Shares over those five years are almost silly (z=4.03), but I think that shows just how much of an outlier this guy was—he was the best hitter and the best pitcher in his league in ’86, and the best pitcher and second-best hitter in ’87. WARP loves his peak too (z=2.40), and he actually tops the ballot in career WS (z=1.60) while placing well above average in the Ink categories. The only stat dragging him down is Peak OPS+/ERA+, and that’s because I’m not giving him credit for his hitting there (if I did, he’d probably zip by Jackson). The more I look at him, he was simply an unparalleled player, and we would be amiss not to recognize this level of dominance.
3. Sam Thompson (5.45). Drops a spot due to my massive reevaluation of Caruthers, but emerges from my new rating system none the worse for wear. Big Sam doesn’t dominate any category, but I am more convinced than ever he is the second-best hitter on the ballot: 1.34 SD above average in my Gray Ink, 1.14 in peak OPS+, 1.09 in career WARP. Only WS doesn’t really like him, and even then he comes out just below average for this ballot in both career and peak. Let’s face it, guys: The man could hit.
4. Iron Joe McGinnity (3.88). A drop-off from a clear top three for me, but McGinnity has fared well in my reevaluation, leapfrogging many players I was sure were better than him. The guy just racked up Black Ink like there was no tomorrow (z=2.48), and due to all those innings had the second-best peak Win Shares (z=1.08) of anyone on the ballot. He’s a bit unimpressive in peak ERA+ and career WARP, but man, did he fill up those leaderboards. He’s not my kind of player, but I’ve gotta give him credit.
5. Lip Pike (3.52). First of all, I found out he was a Jew (ahem, Rosenheck, ahem). He doesn’t get any extra points from me for paving the way for Hank Greenberg, Sandy Koufax, and (of course!) Shawn Green, but it doesn’t hurt either. What he does get extra points for is his pre-1871 play, which I have arbitrarily estimated at about 5.5 WARP per season. This elevates him above his very close competitors Browning and Waddell. His 177 peak OPS+ earns major brownie points (second to Jackson with z=1.81), and he’s average or above everywhere else. He really did own the NA. I don’t see what makes McVey, who got in over a decade ago, so much stronger than the Lip.
6. Pete Browning (3.42). Rated higher before I applied the AA discount, but was still the fifth-best hitter of his era (Brouthers, Connor, Anson, Kelly). He tops the ballot in OPS+*Years at z=2.00, while registering well-above-average scores in Peak OPS+ and both Ink categories and holding his own in peak WARP and WS and career WS. Only career WARP is holding him back from the no. 4 spot, where I would be otherwise inclined to put him. Just an extraordinary and underappreciated hitter.
7. Rube Waddell (3.30). Oh, how my favorites have fallen. As a DIPS devotee, I have a soft spot for Waddell, who from 1902-08 struck out 1/3 more batters per 9 than the league's second-place finishers in K’s per 9, including 1902 when he fanned hitters 82% more often than next-best finisher Jack Powell. Rube clearly got vastly more outs through his own pitching than any of his contemporaries, which is reflected in his strong showings in Black and Gray Ink (1.81, .83). He also fares well in lifetime ERA+*Years (z=.91) and peak ERA+ (z=.80). But WARP and WS just don’t see it. His peak registered just average on this ballot by both measures, and his career totals were over half a SD below the mean in each category. He’s one of my favorites, and I have no doubt he is a HoM’er, but he really only pitched for ten years and was only great in four of them.
   111. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 25, 2004 at 09:57 PM (#643951)
Second half of preliminary 1927 ballot, again excluding Negro Leaguers for now:

8. Jimmy Ryan (2.75). This is the second dropoff point on my ballot, and where my PHoM ends (Ryan wouldn’t quite make it). Very strong career totals, as he’s second on the ballot in career WARP (z=1.73) and fourth in total Win Shares (z=1.43). But it’s just not enough peak for me, as he’s right at the mean in all three peak categories and half a SD below in both the Inks. While an average peak for a 30-player HoM ballot is certainly impressive, I’m one of those guys who thinks you can never get great by just being good.

9. Jimmy Sheckard (2.65). Surprise surprise, Ryan and Sheckard back to back. This Jimmy sure amassed some nice career WARP and WS (z= 1.38 and 1.24), and his candidacy is strengthened by a high peak WARP score (z=.56). But he’s merely average everywhere else, and couldn’t put together a particularly compelling OPS+ in his peak (144 makes a z of -.41). A lot to like, but not enough to love for me.

10. Vic Willis (2.17). I wouldn’t claim for a second that Willis was a special player. But he always found his place among the league leaders—he’s second on the ballot in Gray Ink (z=1.74). Combined with basically ballot-average scores everywhere else, he fills out the top ten.

11. Gavvy Cravath (1.41). Oh how I love thee, Cactus Gavvy. When all those weenie deadball slap hitters were jackrabbiting around the bases, you were playing baseball as it was meant to be played, drawing walks and cranking longballs. Your superb power-hitting peak shows up in a dominant Black Ink score (z=2.00), and good finishes in the OPS+ categories. But even crediting you as a 13.5 WS player from 1909-1911, you just can’t get it done in the counting stats, especially WARP, which found your glovework (aside from 1915) quite unimpressive. Since you’re around a half SD below the mean in peak WARP and career WARP and WS, you can’t make the top ten, or my PHoM. How the mighty have fallen.

12. Eddie Cicotte (1.40). Good peak OPS+ (z=1.07) and Gray Ink (z=.67), average most elsewhere.

13. Mike Tiernan (.99). Average everywhere (within 0.15 SD of the mean in six categories), slightly above in OPS+ (z=.53 peak, .45 career).

14. Charley Jones (.42). As a peak voter, I’ve gotta give Charley his props, as he scored a Z of over .8 on Peak OPS+ and Gray Ink. But he gets killed on career WARP and WS. Man was he good in 1879, though.

15. George Van Haltren (.30). I thought he was indistinguishable from Ryan and Sheckard, but while he is third overall in career WS (z=1.58), he really never had a peak worth writing home about, while the Jimmys both had their share of really strong years.
   112. Chris Cobb Posted: May 25, 2004 at 11:31 PM (#644105)
Zapatero,

You comparison of Joss and McGinnity has become much more reasonable, and we may be down now to the "philosophical differences" level, but I'll mention in McGinnity's favor that

1) if you do the analysis using Runs Allowed rather than Earned Runs, McGinnity will look better yet. He allowed slightly fewer than average totals runs per earned run; Joss appears to have allowed slightly more. Some are inclined to attribute this sort of difference to luck or to defense; I see the pitcher as having a share in almost all scoring against him.

2) Your argument about the sort of player that puts a team over the top a) doesn't fit the careers of the two players and b) overlooks the way that McGinnity's extra innings would likely help _any_ team that doesn't have an entire staff composed of outstanding pitchers. When he throws 400 innings at 169 era+ or 434 innings at 137 era+, he gives much better value than an average pitcher would in those innings, and it's a _rare_ team that has a pitching staff composed entirely of above average pitchers. These two seasons are tremendously valuable, more valuable than any of Joss's seasons, as my metrics see it, because McGinnity combined high performance with unmatched durability. He paid for it later, of course, with weak seasons, but if you are concerned about how much the players actually helped their teams win pennants, rather than with how much a player of Joss's or McGinnity's average caliber would theoretically help a team, you need to give McGinnity credit for his peak value, and a career-oriented analysis does not do that.
   113. OCF Posted: May 26, 2004 at 01:01 AM (#644321)
A bunch of numbers, with an explanation: RC and RC/27 outs for each hitter from a Stats Inc. Handbook. Outs obtained from those numbers. Park-adjusted league runs times the outs subtracted from RC to get a version of RCAA. This RCAA then scaled to a 5.00 R/G environment. Then, for each of these players, the normalized RCAA per year sorted in decreasing order. There's no season-length adjustment here, but Frank Chance fares very well, and he was playing his own personal version of a short season. Also nothing about position or defense. No league-strength adjustments, something to bear in mind when looking at Kauff or Browning. Also, no conclusions: make your own inferences.

Baker        90 76 71 62 45 37 30 22 19 14  3  1  0    
Bresnahan    62 49 46 42 28 26 21 20 19 10  7  6  1  0 -2 -6 -6
Browning     57 56 53 52 46 41 34 27 27 20 19  0 -6    
C. Jones     70 42 42 32 30 22 19 18  6  4 -2      
Chance       88 73 71 58 47 33 30 28 25 13  9  8  4  2  0  0 
Cravath      80 62 58 54 51 42 26 16  6  4  3      
Donlin       83 67 51 40 31 16 11 10 10  8  4  0     
Doyle        68 52 47 42 37 35 29 29 22 21 18 17 13 -5   
Flick        78 76 72 61 61 56 54 46 38 21  4  3 -1    
Hartsel      62 57 54 43 40 33 30 28 17  7  6  4  1 -2   
Jackson      94 91 87 76 74 66 62 50 34 16 10 -2 -3    
Kauff        83 82 59 51 34 28 13  1         
Konetchy     50 46 45 37 34 30 26 23 20 20  9  7  7 -5 -9  
Magee        93 80 60 58 51 50 34 32 29 25 24 20 19 17 13  0 
McGraw       72 68 48 36 34 32 29 25 11  5  3  2  1  0   
Schulte      63 39 29 22 19 17 15 13 11  7  6  5 -1 -1 -8  
Sheckard     70 58 55 35 31 30 27 26 22 21 15 14 14  5  3 -2 -5
Sisler       71 70 59 48 47 41 40 15 11 10  5 -3 -7 -9 -10  
Stone       105 63 57 39 16  8  0          
Thomas       56 55 51 48 42 37 37 34 24 23  6 -2 -5    
Thompson     59 53 46 42 34 31 26 22 21 19 18  8  4  0 -2  
Tiernan      60 56 55 50 49 43 35 26 24 20  8 -3 -5    
Titus        62 37 37 34 31 28 25 23 15 12  7      
   114. Chris Cobb Posted: May 26, 2004 at 02:20 AM (#644641)
Re DavidFoss's post of Dickey Pearce data:

David, what does Wright indicate about the completeness of his records? Are the games he reports all of the games played, or are these just the ones of which some details have survived? I'm working on the data now, and while I can see treating the years back to 1864 as seasons (they are comparable to many NA seasons in length), I have a hard time treating 5 or 8 games as a season. I'm not very knowledgeable about the era, but my understanding is that a lot of the playing took place between different teams "within the club." Is that true? If so, how do you (or others) treat such competition? Does Wright provide any records from intraclub play?
   115. Chris Cobb Posted: May 26, 2004 at 02:24 AM (#644659)
OCF,

Very informative post! Would it be too much to ask for you to run these numbers for Beckley, Childs, Duffy, Jennings, Ryan, and Van Haltren? It woudl be really useful to have a broader view of the 1890s players in comparison to the stars of the early 20th century.
   116. OCF Posted: May 26, 2004 at 02:37 AM (#644708)
Chris, give me till tomorrow. I already have 5 of your 6, but need time to format it. The only one I don't have yet is Jennings - I was focusing on "offense-first" candidates, which isn't exactly what Jennings is.

I also have Lajoie, Wagner, Collins, Cobb, Speaker, Crawford. I left them off the post above so you wouldn't focus on them. Of course they blow away these mere mortals. Besides, Cobb's 7 years above 100 would have screwed up the column spacing.
   117. DavidFoss Posted: May 26, 2004 at 04:28 AM (#644879)
Re: Dickey Pearce Data

No records are kept for intraclub play. (All games listed are against other clubs)

Wright has this to say about completeness:

"In compiling this information, it became evident how frustratingly incomplete it was. For some Association clubs complete information was available -- for others only tantalizing fragments from spotty press coverage. For the games, sometimes the playing date remains unknown, sometimes a score as well. For the players, in several cases, box scores were not extant thus making their record incomplete. In other instances, players were given stats in more games than thier team played because of their participation in 'picked nine' (all-star) games."

That last part explains Pearce's 10 games in a seven game season.

1861-63 appear to be setback years in the growth of the NABBP. The war took its toll on rosters and teams cut back on their schedules as well.

There are a *lot* of team some of these years. Documentation is best for the better teams, though. You have to go back to 1862 to start seeing missing box scores from the top teams (Eckfords 14-2 record that year had runs/outs data for only 8 of the games). Before that missing box scores are not unexpected.

Top teams are documented well, some of the lesser teams are missing... I think a couple of Harry Wright's seasons might be missing... though I do recall that he played Cricket instead of baseball a couple of years.

Its only Games, Runs and Outs up until 1867. Hits and/or Total Bases are added for some teams in 1868. Pro-only games differentiated starting in 1869 (though the stats were still for all games). Pro-only stats came for some teams in 1870 (though notably only all-game stats available for Cincy). Chicago kept track of At Bats that last year.

His sources include 'The New York Clipper', 'American Pasttimes' by C. Peverelly and 'Beadle's Guides'.

He does say the following:

"From the beginning of recorded baseball history, statistics have played a central role."
   118. TomH Posted: May 26, 2004 at 05:00 AM (#644897)
testing my ability to post.....

yahoo! I'm "in"!
   119. TomH Posted: May 26, 2004 at 05:06 AM (#644902)
Negro star Pete Hill

I can’t see myself putting Hill high on my ballot. His offensive stats don’t seem to match up with Grant Johnson. I’m not sure the would be much better than Monroe’s or Grant’s either, both of whom played middle infield. Who was it that posted a week or 2 ago research showing that his own teammates hit as well as Pete? That sure ain’t impressive, if it’s anywhere near true.

As far as his place among Negro League outfielders, Spotswood Poles comes along in a bit, and he looks better, as well as some true no-brainer OF choices. I can’t see Hill placing among the top NegLeg OF stars; so I encourage initial caution on Hill this ‘year’ unless more information is forthcoming.
   120. TomH Posted: May 26, 2004 at 05:15 AM (#644908)
I made a comment re: Dickey Pearce last week that stated Joe Start was much better than Pearce for Pearce's NA years (age 35 plus). This conclusion was questioned, given Pearce's ability to play shortstop. Well, the difference between SS (was it the at or near the lowest ranked hitting positon in 1873?) and 1B isn't nearly going to make up for this:
Start EqA age 35-42 is over .300
Pearce EqA age 35-41 is .240

The best argument for Pearce is that he Might have been the best player before the late 1860s. To me, this is the equivalent of saying that I will hold a USA Tiddlywinks tournament in 2005, with a prize of $8000 to the winner. The winner, will, I assume, be a very good Tiddlywinker. But if I then advertise a Tournament in 2008 with a prize of $200,000, people will come out of the woodwork, even giving up their day jobs to practice. It's a tough call whether the top 2 players of the 1860s trump the best 10 of the 1890s.
Having said all of that, Pearce will probably finally make the bottom of my ballot next week.
   121. DavidFoss Posted: May 26, 2004 at 05:16 AM (#644909)
Just for yucks... Here are the pre-NA numbers for GEORGE WRIGHT. I picked him because he didn't play for Atlantic and because he played extremely well. Note that he jumped from Gotham to Union in mid-1866

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

GEORGE WRIGHT

1864 -- New York Gotham (3-7-1)
-- Pos -- C
-- Competition: NYC+/Newark
-- Hitting : T1st on team with 19 Runs (Gibney 19, Cohen 18, HWright 14)
1865 -- DNP ? (Not on NYGotham roster-- but Harry did)
1866-1 -- New York Gotham (4-4)
-- Pos -- C (5 games)
-- Competition: NYC+/NJ/Wash
-- Hitting -- 1st on team with 4.20 R/G (Connell 4.17, Shreves 3.4) (FYI -- HWright-- DNP)
1866-2 -- Union Morrisania (25-3)
-- Pos -- SS-C (only 9 games)
-- Competition: NYC+/Newark/CT/Philly/Wash/Albany
-- Hitting -- 4.67 R/G (best regulars -- GSmith 4.93, Hannegan 4.36, AMartin 4.21)
1867 -- Washington National (29-7)
-- Pos -- 2B-SS-P
-- Competition : Wash/ Bal/ NYC+/NJ/Richmond/Rockford/Indy/Cincy/Columbus/Lou/StL/Chi
-- Hitting -- 1st on team with 6.28 R/G (ESmith 6.16, Studley 6.09, Norton 6.0)
1868 -- Union Morrisania (37-6)
-- Pos -- SS-2B
-- Competition -- NYC+/CT/NJ/Alb/Det/Cle/Mil/Chi/Rock/Stl/Ind/Lou/Cincy/Philly/Syr
-- Hitting -- 1st on team with 4.53 R/G (Goldie 4.34, GSmith 4.0)
1869 -- Cincinnati (57-0) (Pro: 19-0)
-- Pos -- SS-P
-- Competition -- Cincy/Ind/Buf/Cle/Roch/Alb/Bos/NYC+/Wash/Rock/Mil/Syr/Pit/Stl/SF/Phi/Lou
-- Hitting (all games)
---- 1st on team with 5.95 R/G (Waterman 5.14, McVey 4.60, Gould 4.53)
---- 1st on team with 5.33 H/G (Waterman 4.00, HWright 3.88, Sweasy 3.84)
---- 1st on team with 10.77 TB/G (Sweasy 7.40, Waterman 6.61, Gould 6.34)
1870 -- Cincinnati (67-6-1) (Pro: 27-6-1)
-- Pos -- SS (58 games)
-- Competition -- Lou/NO/Mem/Lex/Cle/Bos/NYC+/NH/Phi/Bal/Wash/Ind/StL/CHi/Rock
-- Hitting (all games)
---- 1st on team with 4.27 H/G (Waterman 3.86, Leonard 3.63, McVey 3.63)
---- 1st on team with 7.08 TB/G (Waterman 5.75, Leonard 5.70, Gould 5.66)
   122. TomH Posted: May 26, 2004 at 05:43 AM (#644925)
I am proud to be the best friend of Clark Griffith.
From 1894 to 1902 with Chicago, he went 190-110. And Chicago was a below .500 team without him. They finally won a pennant in 1901, 4 games over Boston, with Griffith going 24-7.
Add to that, the 1890s were a tough one-league MLB era.
His ERAs are deceptive: it is commonly agreed to, I believe, that he 'pitched to the score'. This is borne out by the fact that his W-L record, given his run support, is much better than would be expected, as has been documented previously.
Yes, Clark gets from me a coveted #2 ballot spot next week.
   123. KJOK Posted: May 26, 2004 at 06:35 AM (#644957)
I've seen MLE projections for RC at 1800 (comp. Jesse Burkett) and 1441 (comp. George Davis). The same projections show him hitting .308. His
OPS projection of 868 comps Bill Dickey (#83 all-time) and his OPS+ of 140 ties Burkett, Wally
Berger, Tip O'Neill and Reggie Jackson for #56 all-time. His comp of 2,985 hits is pretty
damn good, comping Sam Rice. His R and RBI comp Jimmie Foxx and Tony Perez at #17 and #18 all-time.

In other words, these comps cannot be trusted.


Just a couple of words on my own MLE's for Pete Hill.

First, his OPS+ is around 122, not 140! That's a big difference. I think you may not be comparing to the 1955-2003 timeperiod?

Second, since the methodology places players into the 1955-2003 NL, when translating deadball players it will make "power hitters" look a little more valuable than they actually were and "singles hitters" a little less valuable. I'm working on revised MLE's that keep the value contstant across eras.

Third, the methodology does not currently take season for season and convert it. Because of the lack of stats, the methodology takes career numbers and then "spreads" them back based on the year-by-year plate appearances on the "I9" website.
   124. KJOK Posted: May 26, 2004 at 06:38 AM (#644961)
oops,continuing...

Finally, as was mentioned above, Hill for most of his career played in a pitcher's park. I have not yet adjusted my MLE's for park impacts as I'm still trying to get enough data to adjust for Negro League Parks. However, so far I can say that Chicago generally = Dodger Stadium, St. Louis generally = Coors Field....
   125. KJOK Posted: May 26, 2004 at 06:42 AM (#644963)
For now, one more comment. Using my MLE's, these are the guys I have as having the most similar hitting value to Hill:

Andre Dawson
Jim Rice
Dale Murphy
Enos Slaughter
Jimmy Sheckard
Rusty Staub
Harold Baines
Dave Parker
Al Oliver
Chili Davis
Jose Cruz
Don Baylor
Brian Downing
   126. Kelly in SD Posted: May 26, 2004 at 09:57 AM (#644991)
Hi. I am a first-time poster on the HoM. I've lurked for a long time while trying to accumulate enough knowledge about the first 60 years of organized baseball to vote in an election. I'm close, but not there yet.

I came upon some interesting facts that I haven't seen posted (at least recently). If this is old info, just ignore.

In comparing Welch to his frequent teammate Keefe while they are on the same team, they are remarkably similar:

Player: Keefe, Welch
Wins: 211, 228
Losses: 140, 154
W %age: .601, .597
ERA: 2.62, 2.71
Runs All/9: 4.40, 4.79
%R earned: .60, .56
WinShares: 239, 261
GStarted: 368, 391
CG: 350, 384
IP: 3146, 3521
H: 2795, 3349
H/9: 8.00, 8.56
BB: 759, 959
BB/9: 2.17, 2.45
K: 1564, 1263
K/9: 4.47, 3.23
HR allow: 49, 78
WHIP: 1.13, 1.22
BlkInk: 45, 3
GryInk: 139, 176

Again, these numbers are just while they are on the same team. The %R earned is the percentage of runs allowed that were earned. The league average for the years was about .59, while the team's average overall was .56.
In combination with the Retrosheet information about RSI and the quality of Welch's opponents and the above info, I am finding that I cannot make a ballot without including Welch on it.
I understand his Black Ink total is verrrry low, but his Grey Ink is verrrry high and he generally finished behind other HoM'ers.

Any comments
   127. Kelly in SD Posted: May 26, 2004 at 09:59 AM (#644992)
Arrggh. I'm sure it has been asked before, but is there a way I should have formatted this info so the columns would line up correctly? Thank you much.
   128. Kelly in SD Posted: May 26, 2004 at 10:38 AM (#644996)
Re: post 112

McGinnity and Joss did have a slight difference in the number of runs allowed compared to their earned runs allowed. I think this is what Chris Cobb meant.

Over his career, McGinnity allowed 3.76 r/9 and had an era of 2.66. He allowed 1.41 runs per earned run or 71% of his runs allowed were earned. His teams over this period (except for 1902 - I don't know how to balance the 2 teams in different leagues) allowed 3.95 r/g with 2.80 era. This works to 1.41 runs to earned runs or 71% of his team's runs were earned runs.

In comparison, Joss allowed 2.82 r/g and had an era of 1.89. He allowed 1.49 runs per earned run or .67 of his runs allowed were earned. His teams over this period allowed 3.66 r/g with an era of 2.51. This works to 1.46 runs to earned runs or 69% of his team's runs were earned runs.

With the individual player's totals matching their team's so closely, is the difference between the two more a function of the team?

I hope this information is helpful to the discussion and that I did not misinterpret your post, Chris. If I did, I apologize.
   129. Kelly in SD Posted: May 26, 2004 at 10:40 AM (#644997)
Oh, none of the numbers in the above post were park adjusted so the league era numbers may not match the numbers from baseballreference.com.
   130. karlmagnus Posted: May 26, 2004 at 01:20 PM (#645043)
John, sorry but third best hitter on the 1860 Brooklyn Atlantics, playing a 16 game season from teams as far away as New Brunswick doesn't do it for me as far as Pearce is concerned. Of course he's a great pioneer, as were Cartwright, Creighton and Harry Wright (equivalent data on H. Wright would be very interesting indeed.) But he wasn't a stand out player in his early years, even on the Church softball level competition they were playing. Joe Start has enough of an ML career to indicate he might have been really good, and late 60s figures indicate he was. If Pearce had been even 5 years younger, maybe he'd have done the same thing. But unfortunately he wasn't, and he didn't. It's like those Negro Leagues figures for 1911-14 -- they prove nothing, but they make one look more positively on Poles and Lloyd, and with some caution on Hill.
   131. Chris Cobb Posted: May 26, 2004 at 01:37 PM (#645064)
Kelly,

Thanks, you've explained what I meant to say exactly. I hadn't checked the team numbers; it's probably mostly a function of the team here, then. This is a small issue in comparing McGinnity and Joss, since their RA/ERA ratios are quite close and their teams', are pretty close as well, but if this discussion explands to include Rube Waddell, it would become important, since we found a few years back that Rube consistently allowed significantly more unearned runs, in proportion to his earned runs, than his teammates did.
   132. PhillyBooster Posted: May 26, 2004 at 02:11 PM (#645115)
I am very conflicted on Welch. On the one hand, he certainly has the "results" (wins, win%, IP) compared to Keefe.

On the other hand, he did not match Keefe very well at all on the route there (RA, H/9, BB/9, K/9, HR).

The answers seems to be either Welch was very lucky, or Welch was a "clutch" pitcher, or some combination of the two.

I'm not sure if it's possible to figure out which one of the two it is.
   133. Chris Cobb Posted: May 26, 2004 at 02:41 PM (#645153)
KJOK, thanks for your expanded explanation of your MLEs!

Lots of interesting discussion here on Pete Hill. Let me respond to TomH and then try to sum up my impression of KJOK's several posts.

TomH wrote (good to see you back on line, btw):

I can’t see myself putting Hill high on my ballot. His offensive stats don’t seem to match up with Grant Johnson. I’m not sure the would be much better than Monroe’s or Grant’s either, both of whom played middle infield. Who was it that posted a week or 2 ago research showing that his own teammates hit as well as Pete? That sure ain’t impressive, if it’s anywhere near true.

I would agree that Hill wasn't as good a hitter as Grant Johnson. He was definitely a better hitter than Monroe: I guess I'd better repost Monroe's numbers from Holway. A few weeks back Chris J. posted data _from Hill's bad years, 1914-1918_, drawn from the old MacMillan encyclopedia, comparing him to his teammates to see to what extent his drop in numbers was attributable to playing in an extreme pitchers' park. He found that _during this decline period_ Hill had teammates who were indeed hitting better than he was. The records for Hill from 1903-1913 are quite different. He was almost certainly the best hitter in black baseball during his prime, better during those years even than Home Run Johnson. Overall, I think Johnson was a better hitter because he looks to have been a) a tremendous hitter in his very lightly documented early career and b) appears to have remained a significantly above average hitter into his late 30s, which Hill did not. However, basing an assessment of Hill's career off of data focused on the worst years of his career would lead to a deeply erroneous conclusion.

As far as his place among Negro League outfielders, Spotswood Poles comes along in a bit, and he looks better, as well as some true no-brainer OF choices. I can’t see Hill placing among the top NegLeg OF stars; so I encourage initial caution on Hill this ‘year’ unless more information is forthcoming.

The information is there. The experts think very highly of Hill, higher than they do of Poles, mostly, and the data from Hill's prime supports it. He was the best position player in black baseball during this period (again, his older contemporary Johnson was better overall, but he peaked earlier), and careful study of the batting data (most importantly KJOK's, though my less scientific work reaches similar conclusions) shows that he compares well to the top long-career outfield candidates now on the ballot: Jimmy Sheckard, George Van Haltren, Jimmy Ryan.
   134. Chris Cobb Posted: May 26, 2004 at 02:42 PM (#645155)
Now let me turn to KJOK's data. The comparison set he provides is really useful in showing that Hill was not a _tremendous_ hitter; in the majors, he would have been among the top 3-5 hitters during his best seasons, but otherwise not (Andre Dawson, Jim Rice, Dave Parker, Dale Murphy seem like reasonable modern comps for imagining Hill's impact as a hitter).

The hitters on KJOK's list are all look like they'll be borderline candidates for the HoM, so one may wonder how Hill could possibly merit a #1 or other high-position ballot spot this year. First, we have a slate of borderline candidates. Jimmy Sheckard, Hill's best contemporary comp as a hitter, placed 5th in 1926. Second, if Hill truly was most comparable to these players as a hitter, then he had a better career.

Why?

First, he played the second half of his career in an extreme pitchers' park, but the numbers KJOK uses have not been park-adjusted, while he's being compared to park-adjusted figures for these other players.

Second, in each case, if his hitting was equal, he had either a longer career or more defensive value.

Andre Dawson -- Without the bad knees, would he be borderline?
Jim Rice -- much less def. value
Dale Murphy -- Best modern comp, declined as a hitter about when Hill did, at his peak was dominant. Hill did it without the launching pad, though.
Enos Slaughter -- lasted longer, depending on how you credit Hill's comeback. Less documented peak (lost to war years), less defensive value (RF to CF) Only HoFer on the list, and the best player on it.
Jimmy Sheckard -- closest contemporary comp, less defensive value, left field to center field
Rusty Staub -- Much less defensive value
Harold Baines -- much less defensive value
Dave Parker -- much less defensive value (Shares the trough and recovery career shape with Hill, but tanked in his early 30s)
Al Oliver -- lacks Hill's peak, not a strong defensive outfielder
Chili Davis -- much less defensive value
Jose Cruz -- becomes a borderline candidate _when_ you adjust for the Astrodome, not before
Don Baylor -- much less defensive value
Brian Downing -- much less defensive value

Maybe there are other comps who would show as great fielders, but the impression I get from this list is that these are mostly players who were stars, who had great careers, because they were hitters, and they are borderline candidates because a) they weren't the top hitters of their generation (no 170 OPS+ seasons here) and b) they were not great defensive players. When you find a hitter like this who is a great defensive player, he becomes a good candidate. Jimmy Sheckard is a good example; Andre Dawson and Dave Parker "might have been" examples.

I'm not saying Hill is the best negro-league candidate we've seen so far -- Johnson was better. I'm not saying he's necessarily the best candidate available in 1927. Depending on what you value, there are several players whom one might prefer. I am saying, though, that he's the second-best negro-league candidate we've seen (we could argue about Frank Grant), and I am saying that he's better than Jimmy Sheckard. Not everybody likes Jimmy Sheckard, but that's the key comparison. And the key years to study in Hill's career are 1903-1913. Study his prime.

We've been right to try to understand his decline years to see how evaluate his career length, but if your image of the candidate is based on those years (and from the way some write I think that it is case), you are missing out on a great ballplayer.
   135. PhillyBooster Posted: May 26, 2004 at 03:17 PM (#645218)
Thank you to everyone who voted for Grant and Johnson. My ballot no longer looks like I’m an undercover agent from the Negro League Hall of Fame.

1. Caruthers (2) – Climbs the ladder up to #1
2. Jackson (3) – He hit like Caruthers pitched, but didn’t pitch like Caruthers hit.
3. Pete Hill (n/e) – adjusting for the pitchers’ park, he’s conservatively here.
4. Beckley (5) – inducted any first basemen lately? Didn’t think so. Konetchy is no comparison.
5. Bresnahan (6) – no candidate dwarfs the #2 candidate at his position by as much.
6. McGinnity (7) – Only a little bit better than Addie Joss, if Joss is allowed to bring a friend along with him.
7. Cravath (4) – bringing him down a little closer to Thompson, who I think is a good comp.
8. Pike (8) – best centerfielder – an underrated position so far.
9. Foster (10) – top black pitcher
10. Williamson (11) – We spent a long time talking about how third base was different then, but we didn’t change our voting as a result.
11. Thompson (12) – Cravath with more majors, but less overall.
12. Pearce (13) – Top shortstop
13. Childs (14) – Top second baseman with Grant now in.
14. van Haltren (15) – top post-NA centerfielder, and he could pitch too.
15T: Welch (16) – 300 wins is 300 wins.
15T: Wallace (17) – Normally, Welch would have just slid up, but with Wallace the top runner up this year, I wanted to look at him again. They won’t end up tied, but I didn’t want to leave off the potential winner without taking some more time.
   136. OCF Posted: May 26, 2004 at 03:24 PM (#645223)
An extension of my post #113, in response to Chris Cobb's #114. Here are the players Chris suggested I include. I also added back in both Kelley and Crawford as part of the context for dealing with Jackson. Note how large the factor of equalizing to the same run environment is - large enough to make Ryan's 1888 more impressive than Duffy's 1894.

Beckley     40 40 33 32 30 30 25 20 20 19 18 18 16 14 13 10  9  4 -9-13
Childs      61 56 55 34 32 28 20 15  9 -1 -1 -5-20       
Crawford    80 67 64 60 59 53 52 46 46 45 43 42 33 31 26 19 11  2 -9 
Duffy       62 58 42 35 34 33 33 22 21  7  7  7  1  0  0  0 -1   
Jennings    59 56 46 39 14  5  4  3  1  1  0  0 -4 -5-15-21    
Kelley      70 67 58 51 39 35 31 25 24 22 21 18 10  6  1 -2 -17   
Ryan        65 52 47 39 29 26 21 17 17 11 10 10  9  7  6  3  2 -7  
Van Haltren 51 44 44 38 36 35 34 24 23 23 20 17 14 10  5 -4-12
   137. OCF Posted: May 26, 2004 at 03:33 PM (#645242)
There's been quite a bit of debate on this thread about McGinnity and Joss. The following calcuations are based on RA+, converted into a winning percentage by the variable-exponent PythPat system previously discussed (I can't find which thread, but it's there somewhere). This winning percentage is multiplied by IP/9 to give a record. Some other people here have based this kind of thing on decisions; I'm basing it on IP. On each line, there is a year, the RA+, a W-L record, and the (single season) equivalent FWP based on this record.

Joss
1902 113 17-13 12
1903 122 18-13 16
1904 151 14-07 17
1905 132 20-12 19
1906 146 20-11 23
1907 128 23-15 21
1908 167 25-11 32
1909 132 17-10 16
1910 127 12-07 07

McGinnity
1899 144 27-13 32
1900 123 23-15 21
1901 104 22-20 13
1902 109 12-10 08 (AL)
1902 134 11-06 11 (NL) combined 23-17 19
1903 150 33-16 39
1904 179 33-12 45
1905 111 19-16 14
1906 107 20-18 12
1907 100 17-17 09
1908 096 10-11 04

Griffith
1894 126 18-11 18
1895 122 24-16 22
1896 126 22-14 22
1897 102 20-19 11
1898 181 27-09 38
1899 115 20-15 16
1900 110 15-13 11
1901 136 19-11 20
1902 102 12-12 07
1903 139 15-08 17
1904 167 08-03 11
1905 209 09-02 14
1906 142 04-02 05

Totals:
Joss 161-98
McGinnity 227-155
Griffith 213-134

I think you should all take another look at Griffith. I also urge caution for all of these pitchers. Some other records for pitchers not yet eligible but coming soon:

Covaleski 209-134
Rixie 275-224
Faber 255-199
Vance 201-120

Note: I found an error on my spreadsheet. I like these numbers better than the last set I posted.
   138. Jeff M Posted: May 26, 2004 at 04:07 PM (#645319)
Thank you to everyone who voted for Grant and Johnson. My ballot no longer looks like I’m an undercover agent from the Negro League Hall of Fame.

PhillyBooster, where's Monroe? He was at least as good as Grant, and maybe better.
   139. OCF Posted: May 26, 2004 at 04:44 PM (#645376)
The time has come to deal with the Black Sox. I expressed my revulsion with a one-year boycott, including not even mentioning their names during that year, but now I must consider them as players.

First: do recall that we are considering them early. Jackson is only 37 years old now. He's younger than Cobb, Speaker, Collins, Johnson, Alexander, and Baker. He's only 6 years younger than Ruth, 4 years younger than Sisler. He's being considered ahead of his true contemporaries. Had he played out a full career, he would have had a better record - but he'd also be caught up in the 33-34 logjam of candidates. Cicotte, whose career got off to a late start, is 42. Even that's a little bit young for HoM consideration.

First, Cicotte. My RA+-equivalent record for him is 209-149. Very good, but behind Waddell and Willis (even with a severe defensive adjustment to Willis.) I do have him ahead of Powell, Shocker, Reulbach, Leever, and Joss. I had Waddell 9th and Willis 14th on my last ballot, but I may have been overrating Waddell and Willis (and McGinnity). And see that list above: Covaleski, Rixey, Faber. They had careers that overlap with Cicotte, and I'll probably be taking them ahead of Cicotte as well. Sure, he might have gone on and on had he not been banned from baseball. But in the real world his 1921 value as a major league baseball player was the same as Addie Joss's value in 1911. I'm not cutting him any slack for it. Cicotte might make my top 25 but not my top 15.
   140. OCF Posted: May 26, 2004 at 04:44 PM (#645377)
Now, Jackson. We should be very cautious about him because of his youth.

There's a romantic legend that encrusts Jackson. The essential features of this legend include: (1) He was a true giant of the game, a player on the same plane as Cobb and Speaker. (2) As a stupid hick, he didn't have the capacity to understand the implications of the fix. (3) He didn't personally throw the World Series, as his BA attests.

I don't buy any of that. Illiterate doesn't necessarily mean stupid. Although he wasn't a ringleader, I hold him as guilty as the rest. And he wasn't really a giant of the game. In the early teens, in Jackson's best years, they did have a voted award (the Chalmers award). They were working their way through the true giants: Cobb in 1911, Speaker in 1912, Johnson in 1913, Collins in 1914. True, Jackson did finish 5th in 1911 and 2nd in 1913. Had there been a voted MVP in 1919, I think Jackson would have won (to the dismay of the paleo-sabermetricians) - and there would have been a 1920-21 writers' campaign to rescind the award. Jackson was one of those players, like Mel Ott or Cesar Cedeno, who blew into the league at a very young age and were great nearly right away but never really got any better than that.

To be honest, I expected to find enough evidence to pronounce Jackson severely overrated as a player so I could take him down. That evidence isn't there. See posts #113 and 136. The only people I have ahead of Jackson for career total of that version of calibrated RCAA are the true giants (Cobb, Wagner, Speaker, Collins, Lajoie), and Crawford. If I make it calibrated RC above 80% of average instead of above average, Clarke and Magee also go ahead of Jackson, but Jackson has a more impressive peak than any of Crawford, Magee, or Clarke.

I'm probably going to find some excuses for not having Jackson at #1 or #2 - but they'll involve such things as discounting either his 1919 or 1920 seasons, or creating some factor involving his age. He's awfully hard to ignore.
   141. Chris Cobb Posted: May 26, 2004 at 04:57 PM (#645403)
Following up on my comments on Hill vs. Monroe above, I've posted the full data set on Monroe on the new Bill Monroe thread. Those who are arguing that Monroe is better than Hill or Grant, please take a look.
   142. PhillyBooster Posted: May 26, 2004 at 05:51 PM (#645498)
PhillyBooster, where's Monroe? He was at least as good as Grant, and maybe better.

I give "extra credit" for being the best of one's era at a position. Grant certainly was, based on available evidence. HR Johnson was, too. Pete Hill is clearly my #1 centerfielder. Rube Foster is my #3 pitcher, after Caruthers and McGinnity, and that warrants a ballot spot, too.

I just don't see it for Monroe. I rank by position first, and then compare the best at each position to each other to come up with rankings.

My rankings of second basemen goes (1) Childs, (2) Evers, (3) Doyle, (4) Monroe, (5) Dunlap.

There is certainly evidence of Monroe's greatness, but is there any indication he was greater than the three I see as better? I'm open to being convinced, but when I look at him against the others, taking into account the numbers and the subjective data, he just seems to fall short, looking at both contemporary Negro Leaguers and available second basemen.
   143. Kelly in SD Posted: May 26, 2004 at 06:09 PM (#645547)
Re: Rube Waddell's Earned Runs and Runs Allowed:
Chris (and other interested people), I did figure Waddell's numbers also, both for his career and for his time with the Athletics only.

Career numbers: Waddell had an ERA of 2.16 and Runs Allowed (RA) of 3.23. That works out to Waddell allowing 1.49 runs per earned run for his career or 67% of his RA were earned.
Waddell's teams' numbers (99 Lou, 00 Pit, 01 ChiN, 02-07 PhA, 08-10 StA (I dropped the 2 games he pitched for Pit in 1901)): Team allowed 4.04 runs a game of which 2.80 were earned. That works to 1.44 runs per earned run or 69% of his team's Runs Allowed were earned.

Philadelphia numbers ONLY: Waddell had an ERA of 1.97 and a RA of 2.88. That works out to him allowing 1.46 runs per Earned Run or 68% of his RA were Earned Runs.
His team had an ERA of 2.61 and RA of 3.68. That works out to the team allowing 1.41 Runs per Earned Run or that 71% of its Runs Allowed were Earned Runs.

Also, the League averages for his career were 69% of Runs Allowed were Earned and for his time in Philadelphia, the League average was 71%.

So, throughout Waddell's career, he allowed Earned Runs at about 94 to ninety_seven percent of his team or league
   144. sunnyday2 Posted: May 26, 2004 at 06:16 PM (#645567)
Great charts, OCF. Aside from the obvious HoMers, these charts really put to rest the idea that Pete Browning had a short career.

I mean, whenever a new guy comes on the ballot we ask how he ranks against his contemporaries and we say, well, Pete Browning had a short career compared to Gore and Hines and Stovey and...whomever. But then, after we've elected Gore and Hines and Stovey and all the other long-career guys, the tag still sticks. Short career.

Now, in a backlog year, his career's not that short! Of course, those who like a long career have a legit. alt. in Jimmy Sheckard or VH or Ryan. But there are also lots of guys getting votes, and about whom we do not hear "short career," but whose careers are no longer. Even an Ed Konetchy did not have a "longer" "productive" career.
   145. OCF Posted: May 26, 2004 at 06:18 PM (#645573)
Since Kelly in SD brought it up, here's Waddell in the same style as post #137. I've always used RA, not ERA.

Waddell
1897 154 01-00 01
1899 130 05-03 06
1900 130 14-09 15
1901 098 14-14 06
1902 170 22-09 30
1903 141 23-13 26
1904 141 27-15 29
1905 169 26-11 34
1906 119 17-13 14
1907 106 17-15 16
1908 120 18-13 16
1909 104 13-12 07
1910 070 01-02 -1
   146. Howie Menckel Posted: May 26, 2004 at 06:42 PM (#645605)
Here goes:

Most HOM teammates
O'Rourke 19
Kelly 18
Brouthers 16
Hines 15
Ward 14
Clarkson 14

Most "seasons with" HOM teammates
O'Rourke 78
White 58
Ward 51
Connor 49
Ewing 47

Most years as HOM teammates (10 G minimum)
Wagner-Clarke 15
Young-Burkett 11
Lajoie-Flick 11

HOMers by year (10 G minimum)
1871: 8
1872-77: 10
1878: 11
1879: 16
1880: 17
1881: 20
1882: 21
1883: 20
1884-87: 21
1888-89: 23
1890: 27
1891: 28
1892: 27
1893: 24
1895: 20
1896-97: 19
1898-99: 18
1900: 17
1901: 19
1902-03: 17
1904-05: 19
1906: 17
1907-08: 16
1909: 15
1910: 13
1911: 11
1912-13: 9
1914-16: 7
1917: 4
1918-19: 1

HOMers by year, shorter form
1871-77: 8 to 11
1879-80: 16 to 17
1881-87: 20 to 21
1888-89: 23
1890-92: 27 to 28
1893: 24
1895-1906: 17 to 20
1907-09: 15 to 16
1910-13: 9 to 13

I see six teams with six HOMers (counting 1887-89 NY Giants as one team with same six HOMers each yr). I see 10 more with six HOMers, and a dozen teams with four HOMers - all of those teams playing before 1896. Could be a long while before we see another.
   147. OCF Posted: May 26, 2004 at 06:46 PM (#645612)
I should have added up the columns for Waddell: 200-129, 191.

An odd thought: what if Walter Johnson had gotten sick in the middle of the 1915 season and never pitched again, so that his career had been cut as short as Joss's? Then I'd have his equivalent record as 198-92, and I would have had no problem voting for his induction several elections ago ahead of Brown or Walsh. The Big Train is the real peak monster.
   148. Al Peterson Posted: May 26, 2004 at 07:14 PM (#645667)
General question for the group:

Should I try and take into account run scoring strategies when analyzing player value? Specifically I'm talking about Sheckard for the Cubs during the deadball era.

Seems he had many sacrifice hits during a time where he still was a productive player who probably should have been hitting away. But with Chance calling the shots Jimmy was laying those bunts down with great frequency. Do these thrown away plate appearances hurt his value? Was he adept at bunting - deemed valuable during low run-scoring eras - and that fact isn't accounted for? Or do WARP, WS, etc. account for this?

I'm just not sure if the numbers available are a good picture of value since the game was played differently that before 1900. Deadball era baseball seems like a game of sacrificing an individuals' stats for the team and some were asked to play this game more than others.
   149. Dag Nabbit is part of the zombie horde Posted: May 26, 2004 at 07:39 PM (#645719)
Re: #143,

I looked this up way back when I started here - maybe 10 elections ago (& odds are, like almost every lengthy post I made here, it got gutted in the transfer to the new site).

I compared, for each season, what percent Rube Waddell's RA were earned & what % of the team's RA were earned & then determined how many ER Waddell gained/lost each season. Then I added those year-by-year totals up & determined that over the course of his career, that given the number of RA he allowed, & the percentages of ER/RA his teams had, Rube Waddell should've allowed (IIRC, & in fact I think I do remember correctly) 50 more earned runs than he did. That would make his ERA a 2.31 instead of a 2.16 & OCF (I'm 95% sure it was OCF) informed me an ERA of 2.31 would net him an ERA+ of around 125ish.

Considering what an ERA+centric candidate he is, that's pretty damning info. He fell off my ballot & has been dropping ever since.
   150. DanG Posted: May 26, 2004 at 08:25 PM (#645791)
Maybe someone can help me here. Does anyone know how I can get the Baseball Prospectus cards for Lefty O'Doul and Bob O'Farrell? I was getting the WARP3 numbers off of BPro but couldn't get to the cards for names beginning with O'.
   151. DavidFoss Posted: May 26, 2004 at 09:16 PM (#645858)
http://www.baseballprospectus.com/dt/o=doule01.shtml

http://www.baseballprospectus.com/dt/o=farbo01.shtml

Their search engine appears to be broken.

A slow-loading index is here:

http://www.baseballprospectus.com/dt/
   152. EricC Posted: May 26, 2004 at 10:56 PM (#645966)
I compared, for each season, what percent Rube Waddell's RA were earned & what % of the team's RA were earned & then determined how many ER Waddell gained/lost each season.

Chris- This assumes that runs are strictly proportional to earned runs. I picked 1906 as an example, and found that the team ERA-RA data for that year is best fit by the formula RA = 0.53 + 1.16xERA.
   153. Jim Sp Posted: May 26, 2004 at 11:08 PM (#645992)
Cicotte #18, Konetchy #19, Hippo #30.
1)Jackson—1911-3 are all top 100 seasons by OPS+. .356 average with power and walks, if he’s eligible then he’s got the peak to get in. Really should have to wait for his contemporaries to be eligible, but in a weak field I don’t see anyone else to compete for the #1 spot.
2)Doyle—Compare to contemporary George Cutshaw, who was a regular 2B for 11 years with an OPS+ of 86. Doyle’s 126 OPS+ at 2B is only exceeded by Hornsby, Lajoie, Collins, Morgan, Robinson, Richardson, and Dunlap. #19 all time in innings at 2B, how bad a fielder could he have been? Regularly in the 2B defensive Win Shares leaders, WS Gold Glove in 1917. Top 10 in Win Shares 1909-12, 1915.
3)Beckley—I’m no longer Beckley’s best friend, but close. Keeler’s election convinced me to stop downgrading Beckley. Beckley is the better fielder, about the same as a hitter for his career, and at an underrepresented position that with more defensive value. Behind the big 3, much better than any other dead-ball 1B. Win Shares best fielder at 1B in 1893, 1895, 1899, and 1900. Add in 2930 hits, with power and walks. No peak but a lot of consistent production.
4)Pete Hill—Enters my HoM this year. Could belong anywhere between #2 and #8, this is where he is today.
5)Waddell—Waddell has a run of 7 years (1902-1908) in which he was blowing people away, striking out people at rate that is extremely high for the era. Each year allowing at least 20% fewer runs than an average pitcher, in three of those years with an ERA+ over 165. 134 ERA+ in 3000 IP is worthy, his W/L record isn’t impressive because his run support wasn’t impressive. A seven year peak for a pitcher is much more rare than a seven year peak for a hitter, I give the short peak pitchers a lot more credit than the short peak hitters.
6)Wallace—long career, good hitter, played shortstop well, and gets a boost for his pitching. A shortstop with a long career who can hit belongs in the HoM.
7)Cravath—I’m tempted to put him above Beckley, but I’ll be a little less aggressive with him in his first year.
8)Bill Monroe—Riley’s Biographical Encylopedia likes him a lot.
9)Bresnahan--Best hitting year was as a CF, not a C, so that hurts him a bit.
10)Griffith—Comp is Marichal, plus he could hit.
11)McGinnity—Win Shares NL best pitcher in 1900, 1903, and 1904. Terrible hitter.
12)Lave Cross—great fielder. Caught some too. Only hit well in weak leagues, but still that’s a lot of career value…2645 career hits with a lot of defensive value. All time leader in Win Shares / 1000 innings at 3B.
13)Joss—Comp is Koufax…a terrible hitter.
14)Pearce—placement is quite subjective, putting him above Childs and McGraw feels right.
15)Welch— Better than Galvin. His 1885 season (44-11, 1.66 ERA, 492 IP) is a great peak year, he had 3 other great years (1884, 1888, 1889) plus another 6 good seasons. Welch played every year in the toughest league. He could hit a little (68 OPS+). Career 307-210…he deserves some of the credit for that.
   154. EricC Posted: May 26, 2004 at 11:54 PM (#646124)
1927 prelim, in under 1000 characters.

1. Joe Jackson
2. Eddie Cicotte 3223 IP; ERA+ over 150 4 times.
3. George "Rube" Waddell
4. Roger Bresnahan In top 5 of NL OPS+ 1903-1912 decade.
5. Bobby Wallace
6. Pete Hill I see his career arc as a blend between Joe Kelley's and Sherry Magee's. I also see his value between that of Magee (who I was bullish on), and Kelley (who only ever made the lower half of my ballot).
7. Jake Beckely
8. Lip Pike
9. Dickey Pearce
10. Addie Joss
11. Hughie Jennings
12. Frank Chance
13. Gavvy Cravath
14. John McGraw
15. Jimmy Ryan He's back.
Van Haltren and Sheckard are close. Nothing new to say about Thompson, McGinnity, and Caruthers. Ed Konetchy and Hippo Vaughn would be strong candidates if the Hall were twice as big.
   155. Dag Nabbit is part of the zombie horde Posted: May 27, 2004 at 01:01 AM (#646482)
Chris- This assumes that runs are strictly proportional to earned runs. I picked 1906 as an example, and found that the team ERA-RA data for that year is best fit by the formula RA = 0.53 + 1.16xERA.

I have no idea what this means.

In 1900 Rube Waddell allowed 55 ER out of 96 RA - 57.3 were earned. The Pirates as a team 417 ER in 612 RA - 68.1% were earned. If 68.1% of his 96 RA were earned he'd have 65 ER. I see no reason why this shouldn't be done proportionately. Throughout his career he gave up 50 fewer ER than he actually did, so his ERA & ERA+ are suspect.
   156. Dolf Lucky Posted: May 27, 2004 at 01:37 AM (#646674)
Just wanted to point out that our most recent electee, Sherry Magee had a lifetime WARP2 score of 75.4 in 2087 games.

(WARP2 is the one that doesn't adjust for season length, but does adjust for league strength)

As a matter of comparison, Sam Thompson had a lifetime WARP2 score of 76.3 in 1407 games.

Magee probably went in too soon, but it's time to put Big Sam in.

That is all.
   157. DavidFoss Posted: May 27, 2004 at 03:51 AM (#647218)
There was a request up there somewhere for HARRY WRIGHT's pre-NA numbers. Unfortunately, his pre-Cincy play is not well documented. I heard he took time off to play cricket, but I won't guess when that was.

==================================

HARRY WRIGHT

1858 -- New York Knickerbocker (0-3-1)
-- Pos -- P-C
-- Competition -- NYC+
-- Hitting -- 6 Runs in 2 Games (3 games have boxes, fwiw -- 3.0 R/G leads team)
1859 -- New York Knickerbocker (1-3)
-- Pos -- SS-C
-- Competition -- NYC+
-- Hitting -- 1st on team with 3.5 R/G (Davis 3.0, DeBost 3.0)
1860 -- New York Knickerbocker (0-1)
-- Pos -- C
-- Competition -- Bkn (Excelsior)
-- Hitting -- 1 run scored in 32-9 loss (Welling only player to score 2)
1861-2 -- ?? Knickerbocker games not documented (not on Gotham roster)
1863 -- New York Gotham (3-4)
-- Pos -- ?
-- Competition -- NYC+
-- Hitting -- 1st on team with 2.0 R/G (Stokem 1.75, 1.72)
1864 -- New York Gotham (3-7-1)
-- Pos -- 1B-SS-3B
-- Competition -- NYC+/Newark
-- Hitting -- 4th on team with 2.0 R/G (GWright 2.38, Stokem 2.16, Gibney 2.11)
1865 -- New York Gotham (5-7)
-- Pos -- OF-SS-P (7 games)
-- Competition -- NYC+/Philly
-- Hitting -- 1st on team with 3.14 R/G (Dockney 2.89, Hatfield 2.43, Beadle 2.3)
1867 -- Cincinnati (16-1)
-- Pos -- P
-- Competition -- Cincy/Lou/Ind/Wash
-- Hitting -- 2nd on team with 6.59 R/G (Schwarz 6.67 (in 12G), Howe 6.4, Storer 6.11)
1868 -- Cincinnati (36-7)
-- Pos -- P-2B
-- Competition -- Cincy/Col/NYC+/Philly/Alb/Buf/Cle/Chi/StL/Pit/Wash
-- Hitting
---- 6th on team with 3.65 R/G (Hatfield 4.81, Waterman 4.40, Howe 4.18)
---- 4th on team with 3.28 H/G (Hatfield 4.14, Waterman 3.75, Gould 3.60)
1869 -- Cincinnati (57-0) (Pro: 19-0)
-- Pos -- OF-P
-- Competition -- Cincy/Ind/Buf/Cle/Roch/Alb/Bos/NYC+/Wash/Rock/Mil/Syr/Pit/Stl/SF/Phi/Lou
-- Hitting (all games)
---- 9th on team with 4.07 R/G (GWright 5.95, Waterman 5.14, McVey 4.60, Gould 4.53)
---- 3rd on team with 3.88 H/G (GWright 5.33, Waterman 4.00, Sweasy 3.84)
---- 8th on team with 5.82 TB/G (GWright 10.77 Sweasy 7.40, Waterman 6.61, Gould 6.34)
-- Pitching -- 1.23 R/IP in 118 IP, (Brainard 1.20 - 338, GWright 1.29 - 14)
1870 -- Cincinnati (67-6-1) (Pro: 27-6-1)
-- Pos -- OF-P
-- Competition -- Lou/NO/Mem/Lex/Cle/Bos/NYC+/NH/Phi/Bal/Wash/Ind/StL/CHi/Rock
-- Hitting (all games)
---- 6th on team with 3.47 H/G (GWright 4.27, Waterman 3.86, Leonard 3.63, McVey 3.63)
---- 7th on team with 4.94 TB/G (GWright 7.08, Waterman 5.75, Leonard 5.70, Gould 5.66)
-- Pitching -- 1.14 R/IP in 108 IP (Brainard 1.00 - 440, Atwater 0.89 - 93)
   158. Kelly in SD Posted: May 27, 2004 at 04:27 AM (#647252)
This is more Rube Waddell ER vs Runs Allowed Stuff:

I broke down season-by-season (except for 1897 for some reason) the total runs and earned runs Waddell allowed and the total runs and earned runs his teams gave up when Waddell's numbers are removed.
The verdict: In four years, Waddell allowed more earned runs than would be expected or his earned runs were a higher proportion of his runs allowed than the rest of his team.
In 1899, he allowed 38 runs and 27 were earned. If he had the same proportion as the rest of his team, it would have been 25.
In 1903, actual was 109 runs, 88 earned, if he matched the rest of his team - 83.
1906, actual 89 r, 67 earned, if matched team then 62.
1910, actual 19 r, 13 earned, if matched then 12.
In these seasons, he is 13 runs "unlucky" allowing 13 more earned runs than would be expected.

On the negative side:
1900: 96 total, 55 earned. If same ratio as team, then 68. 13 runs "lucky".
1901: 135 total, 84 earned. If same ratio as team, then 90. 6 runs "lucky" 19 total.
1902: even
1904: 109 total, 69 earned. If same ratio as team, then 80. 11 runs "lucky", 30 total.
1905: 86 total, 54 earned. If same ratio as team, then 60. 6 runs "lucky", 36 total.
1907: 115 total, 68 earned. If same ratio as team, then 83. 15 runs "lucky", 51 total.
1908: 93 total, 60 earned. If same ratio as team, then 65. 5 runs "lucky", 56 total.
1909: even.

In sum, he is 56-13 for 43 runs lucky. If you want to add these runs to his earned runs total it would make his career era 2.29 instead of 2.16.

Does this make sense? What problems do you see?
Any critiques welcome.
   159. Jeff M Posted: May 27, 2004 at 04:45 AM (#647263)
As a matter of comparison, Sam Thompson had a lifetime WARP2 score of 76.3 in 1407 games.

Yes, but can you explain WARP2? Is it right? How would we know?
   160. Jeff M Posted: May 27, 2004 at 04:48 AM (#647265)
Should I try and take into account run scoring strategies when analyzing player value?

Al, this is a noble idea, but if you do it for Sheckard, you'll have to do it for everyone. First, that would be a TON of work and might not be fruitful. Second, is it even possible to do it for everyone? Sheckard played for Chance for quite a while, and we know quite a bit about Chance's managing style because he was high profile. It would be more straightforward with Sheckard.

But what about players who had multiple managers (some of whom we know very little about)? I don't have any idea if Pants Rowland liked the sacrifice or not. I just like his name. :)
   161. DavidFoss Posted: May 27, 2004 at 05:19 AM (#647271)
strong
em
u
strike
stats
.(JavaScript must be enabled to view this email address)
   162. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 27, 2004 at 05:53 AM (#647283)
Preliminary ballot part III: Missing the Cut

16. Clark Griffith (.28). There's a real glut with my 14-17, thankfully none of them will make it because of Negro Leaguers so I don't have to split hairs. Griffith is within half a SD of the mean for this ballot in every single category. Was great in 1898, but not enough elsewhere to bump him up.

17. Hugh Duffy (.25). Fares much worse on my Black Ink test (Z=.57) than Bill James's. It was tough to stand out in the 1894 offensive craziness, but .440/.502/.694 certainly qualifies, and he racked up the career WS (z=.97). But he only had one other year over 130 OPS+, leading to a -.68 Z on Peak OPS+ that sinks him off my ballot.

18. Jake Beckley (-.1). Unsurprisingly, gets his ass kicked in the peak categories (combined Z=-2.73), more than makes up for it in the career categories (combined Z=4.00), but his score of Z E R O on my Black Ink test (z=-1.25) does him in. He seems to comp to Rafael Palmeiro for me, but didn't have One Great Year like Raffy's 1999.

19. Addie Joss (-2.25). Well, this is the real cutoff line, the biggest gap on my scoring system. I used to adore Joss for his shiny career ERA+. But I now think he was so high in part because his career was so short (coddled, no decline phase, etc.). I am certainly a peak voter, but Joss' best five years don't really stand out--he's over .5 SD below the mean for Peak WS and WARP. My system likes McGinnity's, Waddell's (especially), Cicotte's and Griffith's peaks over Addie's. Joss was very, very good, but his only real knock-'em-dead year was '08. And needless to say, he's near the bottom in career WARP and WS. I've really reevaluated Addie, and have joined the consensus that he just doesn't have what it takes.

20. Hughie Jennings (-2.34). WARP just goes gaga about Jennings' peak, preferring it even to Caruthers' best 5 (!) for a Z of 2.83. WS is less enthusiastic although still impressed with Hughie's best 5 (z=.66). However, lackluster OPS+ marks, career WS, and Ink totals are all well below the ballot averages and drop him this low.

21. Cupid Childs (-2.78). WARP is a fan, with z=.72 for his best 5 and .55 for his career. Comes out below average everywhere else, though. The title of "best 2B of the 1890's" only gets you so far.

22. Bobby Wallace (-3.39). WARP goes nuts for this guy's defense, and he dominates the ballot in career WARP (z=2.43) while coming out very strong in career Win Shares (z=1.46) as well. But he's just putrid in the OPS+ categories, peak Win Shares, and both Inks. I just haven't seen enough to believe Wallace was really the Ozzie Smith of the deadball era.

23. Hippo Vaughn (-3.65). Vaughn actually put up some nice Ink marks, once leading his league in ERA, WHIP, and Wins, and twice in K's and IP (Black ink z=.76). But none of his stats match up with the competition here.

24. Ed Konetchy (-3.89). The career voters should give him a long look, as all three of his career categories (OPS+*Years, WS, and WARP) are about average for this ballot. But he just didn't peak like the other guys here.

25. Larry Doyle (-4.25). Not too much to recommend--career WS and Gray Ink are solid, I suppose. Like Childs, the fact that he was the best at his position in his league just reflects that there weren't any really HoM-worthy 2B in the NL in the teens.

26. Frank Chance (-5.07). Peak OPS+, peak WS, and Gray Ink are near ballot average; nothing else is.

27. Roger Bresnahan (-8.92). Yes, he was the best catcher of his era. No, he wasn't a HoM player, and I'm not convinced he was any more than good. Fares atrociously in every single category. Can't we accept that catcher was a weak position and get on with it?

28. Slim Sallee (-10.67). I thought he was worth considering. I was wrong. He doesn't measure up anywhere.
   163. Kelly in SD Posted: May 27, 2004 at 07:15 AM (#647299)
Some Addie Joss info and does anyone know where I can find similar info pre-sorted so I don't have to go through Retrosheet game by game for other pitching candidates such as McGinnity, Griffith, Caruthers, Waddell, and Welch?

The totals won't match his stats exactly because Retrosheet only lists starting pitchers, but since Joss completed and/or got decisions on almost all his starts, the numbers should be interesting.
Joss started 135 games against teams over .500 and the team went 71-64.
Joss started 120 games against teams under .500 and the team went 90-30.
The cumulative record of all the pitchers he faced was .528.
The cumulative record of the pitchers he defeated was .498.
The cumulative record of the pitchers he lost to was .577.
The above 3 figures were calculated by seeing who Joss started against each year, that person's record and totalling the figures. A pitcher's record was counted as many times as he faced Joss. While he benefited from facing Walsh 4 times when he won 40 games he also faced a pitcher named Winter 4 times the same year who finished 5-19.

Does anybody have any similar information to compare other pitchers with the above numbers?
Does the average opponent winning %age seem high or low? Based on Joss' 29 decisions a year, he went 18-11 while facing a 15-14 pitcher every time on the mound.

I don't know if I will vote for Joss. I am putting the numbers out here to get feedback.
Thank you.
   164. DanG Posted: May 27, 2004 at 02:11 PM (#647365)
This is too good to save for my ballot. Here are all the firstbasemen with 950 or more RBI thru the 1926 season:

2076 Anson
1575 Beckley
1322 Connor
1296 Brouthers
1060 McInnis
992 Konetchy
968 J. Doyle
952 H. Davis

Beckley's total is still the 8th best all-time among firstbasemen. Besides Anson (#1 all-time) there's Gehrig, Foxx, Murray, Palmeiro, Perez and Killebrew ahead of Beckley. McCovey and McGriff round out the top ten. Bagwell and Thomas figure to join this group next season.
   165. robc Posted: May 27, 2004 at 02:21 PM (#647372)
Made some adjustments this year. Didnt affect much at the top. Pitchers got a small boost, Jackson got a big boost, but still doesnt make my ballot (31->17). Top 11 I consider to be HoMers, with 7 thru 11 being marginal. I dont see any reason for expanding the ballot, I am already voting for more guys than I want to.

1. Wallace
2. Sheckard
3. Thompson
4. VanHaltren
5. Cross
6. Ryan
7. Beckley
8. Childs
9. Jennings
10. Caruthers
11. Hill
12. Long
13. Jones, F
14. McGinnity
15. Williams

16. Nash
17. Jackson
18. McCormick
19. Konetchy
20. Tiernan
21. Griffin
22. Waddell
23. Pearce
24. McGraw
25. Pike
26. Bresnahan
27. Griffith
28. Mullane
29. Willis
30. Selbach
   166. Chris Cobb Posted: May 27, 2004 at 02:22 PM (#647373)
Dave Rosenheck,

Just finished reading through your preliminary ballot -- lots of food for thought. Thanks!

Three notes/questions:

In your comment on Lip Pike, you wondered what distinguished Cal McVey from Pike -- the main differentiating factor was that McVey played half his career at catcher, the most valuable and demanding defensive position. In short, he hit about like Pike did, but had more defensive value.

That leads me to my first question: perhaps I missed your full description of your new system, but how are you assessing defensive value or defensive excellence in it? Defense will be part of WARP and WS, of course, but are you giving it any separate treatment? I ask because it seems like you are multiplying the impact of good hitting by looking at OPS+ and black ink, but not the impact of good defense. I think it's proper, at the present moment, to have a ballot that is heavy on outfielders, but the fact that you have 10.5 outfielders and no infielders suggests strongly to me that your system isn't giving fielding its due.

Second question. In your comments on Bresnahan, you indicate that he doesn't rate well and all, and you suggest that catcher was simply a weak position. Have you tested your system on a catcher or two who are widely acknowledged to be great -- possibilities who are now active and beginning to make quite a mark are Mickey Cochrane and Gabby Hartnett. If your system shows _them_ to be good candidates, then it's probably a fair system. But if _they_ don't get close to your ballot, then you might reconsider how your system treats catchers.
   167. Al Peterson Posted: May 27, 2004 at 02:23 PM (#647376)
1927 prelim ballot. Joining the ballgame is the Shoeless One and other new eligibles.

1. Joe Jackson (-). No he isn't Cobb or Speaker. I don't think that is a problem since those guys are pretty fine players. Excelled in the short career he compiled before some integrity issues dragged him down.

2. Sam Thompson (1). Patiently waiting his turn. During 10 year period (1886-1895) was top 10 in league in total bases 9 times. .308 EQA, .684 OWP, man could hit a little.

3. Joe McGinnity (2). Eight straight years 20+ Wins, 300+ Innings. Iron Joe is quite accurate.

4. Pete Hill (-). On the opinion side, most sources like Bill James and the Pittsburgh Courier poll place him as an upper echelon Negro League performer. Then the numbers calculated show a solid offense at a defensive position we're looking for candidates. Not much to dislike.

5. Jimmy Ryan (3). Giving credit of CF being more valuable than LF in terms of defensive spectrum. Being very good for long periods of time gets some points from me. Interesting note: was tried at SS for Chicago in 1889 after Ed Williamson flamed out.

6. Jimmy Sheckard (9). Another well-rounded player, successful for the great Cub teams. The comparisons to Magee show them on the same plane - I have a harder time believing Sheckard was so awesome defensively, Magee so wretched.

7. Rube Waddell (7). Won six straight strikeout titles - dominance you don't get everyday. His 1902 season: 12-8 out in LA to start season, brought back to Philly by Connie Mack in June and went 24-7. That's a lot of pitching...

8. George Van Haltren (4). Similar arguments to Ryan, just a little less to them.

9. Pete Browning (8). If we're looking for an 1880s ballplayer this could be your man. Even with league discounts he swung some mean lumber. Career OPS+ = 162 which puts him in company with names like Foxx, McGwire, Frank Thomas. Discount it because of AA play? At the OPS+ = 147 level you're talking Heilmann, McCovey, and Schmidt. That's some pretty lofty company. Batting Average Placement within league 1882-1891: 1, 2, 3, 1, 2, 2, 3, -, 1, 3. Had a career .341 average in a league environment of .257 . Offensive Winning Percentage - .745. His guy was a WOW when hitting. He might not be the ideal multi-dimensional player but when you are this much an outlier at part of the game its going to get a bonus from me.

10. John McGraw (10). Limited playing time but what he did with it is nonetheless outstanding. Positional bump as well. Cons include just not playing enough but was on base all the time when participating.

11. Bob Caruthers (11). My tug-of-war with his value continues. I'm more sure now that if push came to shove Freedom Bob deserves in before the rest below him. Affected pennant hopes greatly per individual year with his play yet due to short career couldn't affect a large number of pennants.

12. Jake Beckley (12). Tougher and tougher to ignore with dearth of 1B for a number of years; career totals eventually add up to quite the player despite lack of peak. Not being one of the ABC trio at 1st base doesn't mean you were bad.

13. Hugh Duffy (13). Couple of great spikes to go with other uneven performances. Gets bump based on contemporary opinion as being one heck of a ballplayer. Win Shares love the D.

14. Dickey Pearce (14). Impressed by the fact he was a regular in the NA at an age that was very old compared to most players. Still unsure as to career arc - was he a Cal Ripken, Alan Trammell, Dave Concepcion, none of the above? The reviews from early baseball are glowing - the issue of competition level at an unorganized, evolving time is frustrating.

15. Gavvy Cravath (15). Get's a lift from noteworthy performance in minors between stints in the majors to go along with peak achievement at the major league level at an advanced age. Ten years too early to reap benefits of the lively ball.
   168. Al Peterson Posted: May 27, 2004 at 02:25 PM (#647377)
1927 ballot also rans:

16. Tommy Leach. 3B and CF; what a weird combination.

17. Cupid Childs. OK, the claim is best 2B of the 1890's. So then why did we just elect Frank Grant?

18. Bobby Wallace. Career is long, not enough peak to make the top candidate. We've have plenty of SS to consider (and elect) and I don't like him as much as some of the others above.

19. Tony Mullane. Helped most by my reexamination of Caruthers. Another pitcher who was good with the stick. Pitched many innings with good rate stats. He won 284 games playing with some teams that weren't nearly the strength of Freedom Bob. Remember he's missing a year (1885) when in a contract dispute or his numbers would have been better. If people have Caruthers high on their ballot I'd expect seeing Mullane - lower but in close proximity. Great nicknames as well - Count, Apollo of the Box.

20. Roger Bresnahan. See the pros as solid hitting catcher with some versatility. Cons are OK, not great, in terms of durability, fielding is not Bennett level. Fielding numbers seem to be hurt in 1902-04 when he was all over the place with different positions. I'll wait for better catchers before force feeding an induction on the Duke.

21. Hughie Jennings. Were taking on SS's quickly - has short term excellence on his side.

22. Clark Griffith. 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.

23. Larry Doyle. Between Childs and Dunlap, more like the lower end probably.

24. Rube Foster. Fine pitcher. Was probably used against only the best opponents so his Strength of Schedule, to borrow from football, was high.

25. Ed Ciccotte (-). Knuckles is like a lot of pitchers down below the top 15 on the ballot. They had nice careers but HOM worthy?

26. Charley Jones. Having a hard time ignoring the eye popping production. Blackballed years are recognized but not used to jump over many candidates.

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

28. Lip Pike. For the time of baseball before 1885 I'd prefer Pearce or Charley Jones over the Lipster. Maybe even Harry Wright if push came to shove.

29. Vic Willis. I have a feeling he wasn't as good as the win totals show. Just a feeling I guess.

30. Mickey Welch. Just win baby. Pitched a ton of innings - probably deserves more credit than I'm giving. Pitcher reevaluation is coming up and we'll see what's in store for this longtime ballot filler.

Konetchy is in the mid-30s. Other newbies are not too attractive.
   169. karlmagnus Posted: May 27, 2004 at 02:28 PM (#647379)
Thank you very much David, that was extremely helpful indeed. It tells me, from the Cincy data, that I must slide Wright down towards Pearce, and therfore off my ballot -- to be a true HOMer, in those years before the NA/NL, he should have been top 3 in Cincy in 1868-70, not 6th-9th, even though that was an All Star team. Wrigght and Pearce should of course go in Pioneer Wing of the HOM, or Wright in a Managers Wing.

Do you have equivelent data on Pike and Meyerle (neither of them, IMHO, good candidates for Pioneer Wing.) Everyone has Pike higher than Meyerle, based on his pre'71 play, because from '71 on, Meyerle was better.

Dan Rosenheck, welcome to the Parisian Bob club. I know this is supposed to be scientific and dispassionate, but after 12 "years" with Bob #1, I can't help cheering when someone newly comes to share my view of him, more or less.
   170. DanG Posted: May 27, 2004 at 02:28 PM (#647380)
This question is for the Friends of Sam Thompson: Why not Mike Tiernan?

When Tiernan came on the ballot in 1905, there were innumerable comments about his close similarity to Thompson. Almost as inseparable as Van Haltren/Ryan.

The voters took this to heart and voted Tiernan exactly three places behind Thompson in each election 1905-08. However, Tiernan's support was concentrated at the bottom (12-15), while Thompson was more just below mid-ballot. So it's not surprising when Tiernan lost many votes as the ballot got more crowded.

Fast forward to 1926. The ballot has thinned out and Thompson has just received 16 top-4 votes. That should indicate a sturdy recovery for Tiernan. But he got merely two votes.

At this point I'm reminded of Bill James writing that as time goes by, a player's hall of fame cases rests more and more on his statistics. Thompson has the superficial numbers generated by a great lineup and favorable home parks. That's the only basis I can see for the huge chasm now separating him and Tiernan.

Either ST is being vastly overrated or MT is being drastically short-changed.
   171. OCF Posted: May 27, 2004 at 03:28 PM (#647456)
Oops - I've got to correct a serious error of mine. In calculating Clark Griffith's RA+, I failed to enter the actual league runs scored for the entire AL part of his career, and the default number that wound up there was a > 5 R/G turn-of-the-century number. As a result, his AL RA+ was grossly overstated. Here are the correct numbers - fixing the mistake costs him 10 wins and adds 10 losses to his record, and adding in his late ineffective token appearances adds another 2 losses.

Griffith
1894 126 18-11 18
1895 122 24-16 22
1896 126 22-14 22
1897 102 20-19 11
1898 181 27-09 38
1899 115 20-15 16
1900 110 15-13 11
1901 126 18-12 17
1902 089 10-13 02
1903 108 13-11 08
1904 111 06-05 04
1905 145 07-04 08
1906 098 03-03 01
1907--14 00-02 -2

The overall equivalent record is 203-146. That's certainly a good pitcher, and he did have one great year in 1898, but it lands him in a territory occupied by too many other pitchers. As a result, I am withdrawing my support for Griffith. Like Cicotte, he might make my top 25, but probably not my top 15.
   172. andrew siegel Posted: May 27, 2004 at 03:53 PM (#647495)
Based on all the pitcher discussion, I went back and looked at the PRAR numbers for all the eligible pitchers. Since one of my purposes was to see if I could get a handle on how good a pitcher Bob Caruthers was, I kept a list of the seven best PRAR seasons for all the eligible pitchers (as Caruthers had a great seven year run and nothing else). This is not corrected for season length or league quality and obviously there is a selection bias because I used the length of Caruthers's career to set the parameters. For what it's worth, here are the leaders among all pitchers eligble by 1927:

Rusie 854
Young 798
Mathewson 740
Walsh 730
Nichols 684
Clarkson 681
Keefe 660
Radbourn 607
Galvin 607
McGinnity 598
Buffinton 574
King 572
Plank 550
Welch 549
Waddell 548
Hutchinson 548
Cicotte 542
Breitenstein 533
Willis 529
Vaughn 521
Caruthers 519
Mullane 517
Griffith 515
Morris 504

Three-Finger Brown is not included because his card is not accessible. Joss is at 478. a bunch of guys soon to join the ballot would be high on the list (among them: Johnson 887, Alexander 774, Colveski 617, Shocker 547, Mays 522, Cooper 507, Shawkey 498).

Make of it what you will.
   173. OCF Posted: May 27, 2004 at 04:26 PM (#647540)
Thanks, Andrew. For my own sake, I don't like to make lists that combine pitchers from before and after 1893, because so many things were different. But from before 1893, besides the elected ones and besides Welch, that's Buffinton, King, Hutchison, and Breitenstein all ahead of Caruthers. Of course, Caruthers was a better hitter than those four.

One of the issues within the 1933-35 new candidate flood will keeping alive the candidacies of those who come in with that crowd but not on top of it. Some people will need advocates, and I've already picked out the guy I'm going to follow: Stan Covaleski.
   174. Chris Cobb Posted: May 27, 2004 at 04:32 PM (#647546)
This question is for the Friends of Sam Thompson: Why not Mike Tiernan?

I'm no FoST, but I can give you one of the answers: WARP, which supports a high ranking for Thompson while leaving Tiernan off the ballot. You've got the other answer in black ink and the raw stats behind it.

It was WS that first aligned Thompson and Tiernan. Those voters who still trust WS on this one (such as myself) still have Thompson and Tiernan close together (and well off the ballot). But with two other credible sources that uphold Thompson's reputation, the majority of voters have firmly separated the two.

In my view, OCF's analysis of runs-created in context corroborates and shows the foundation of Thompson's unspectacular win share totals. I'd like to see a supporter of Thompson explain why that evidence is unpersuasive.

I'd also like to see a reprise of discussions of the extent to which OPS+ is an accurate measure of a hitter's value, esp. in the early game. It's cited more than it was ten-fifteen elections ago, and it plays a significant role in Thompson's candidacy, the continuing support for Browning, and the high placement of Joe Jackson. I don't have the mathematical skills to do this one myself, but if those who do wouldn't mind, I'd like a refresher on the pros and cons of OPS+.
   175. DavidFoss Posted: May 27, 2004 at 04:42 PM (#647560)
Thank you very much David, that was extremely helpful indeed. It tells me, from the Cincy data, that I must slide Wright down towards Pearce, and therfore off my ballot -- to be a true HOMer, in those years before the NA/NL, he should have been top 3 in Cincy in 1868-70, not 6th-9th, even though that was an All Star team. Wrigght and Pearce should of course go in Pioneer Wing of the HOM, or Wright in a Managers Wing.

Do you have equivelent data on Pike and Meyerle (neither of them, IMHO, good candidates for Pioneer Wing.) Everyone has Pike higher than Meyerle, based on his pre'71 play, because from '71 on, Meyerle was better.


Switching from fact-dumping mode to having-an-opinion mode....

Pearce is certainly a better candidate than Wright IMO. It all depends on how much you timeline his early 60's play (if at all). He certainly was a great player from 59-64 for a top team.

HWright's pre-Cincinnati days are simply not documented well by the book I have. He appeared on all star games in the late 50's, but the data that Marshall Wright has (Runs/Outs in documented inter-club games) does not confirm his stardom. I certainly don't want to say that he was not a star -- its just not confirmed by this one book I have.

I have data for Meyerle & Pike... Pike's last two years are up in the Pearce stats as are most of Start's. Start also played for Bkn-Enterprise in 1860-61. I have to wait until I get home though.
   176. Jeff M Posted: May 27, 2004 at 04:48 PM (#647569)
Have you tested your system on a catcher or two who are widely acknowledged to be great -- possibilities who are now active and beginning to make quite a mark are Mickey Cochrane and Gabby Hartnett. If your system shows _them_ to be good candidates, then it's probably a fair system. But if _they_ don't get close to your ballot, then you might reconsider how your system treats catchers.

An excellent comment and we should all be testing our systems to see if good catchers fare well. Catchers are a historically weak hitting position, but I don't think that we can characterize it as a weak position generally when we know it is the toughest defensive assignment on the field.
   177. John (You Can Call Me Grandma) Murphy Posted: May 27, 2004 at 04:57 PM (#647579)
HWright's pre-Cincinnati days are simply not documented well by the book I have. He appeared on all star games in the late 50's, but the data that Marshall Wright has (Runs/Outs in documented inter-club games) does not confirm his stardom. I certainly don't want to say that he was not a star -- its just not confirmed by this one book I have.

I'm probably going to remove Wright with this information at least temporarily. Thanks, David.

Have you tested your system on a catcher or two who are widely acknowledged to be great -- possibilities who are now active and beginning to make quite a mark are Mickey Cochrane and Gabby Hartnett. If your system shows _them_ to be good candidates, then it's probably a fair system. But if _they_ don't get close to your ballot, then you might reconsider how your system treats catchers.

My system treats them both kindly as top-tier players. Dickey and Schang also do very well. Ernie Lombardi doesn't, however (though he still may make my ballot when we get to the '50s).
   178. andrew siegel Posted: May 27, 2004 at 05:26 PM (#647620)
I think you need to break Caruthers's offensive contribution into two parts: additional value on the days he pitched and contributions on days he didn't pitch. My rough calcuations say that the effect of Caruthers's bat on the days he pitched was sufficient to move his value on his pitching days up about 100 runs over his top 7 seasons, moving those seasons into rough parity with the top 7 seasons of McGinnity, Coveleski, Radbourn, and Galvin.

In evaluating the careers of these players against each other you neeed to take into account:

(1) Their contributions in the remaining years of their careers (in Caruthers's case including those games he played other positions during the seven seasons).
(2) Season length.
and (3) League quality.

Based on those criteria, I think Galvin remains signficiantly ahead of Caruthers based on his longevity and Radbourn beats him based on league quality and season length adjustments. Leaving Spalding out of it for the minute, I think Caruthers trails every pitcher we've thus far elected.

As for those still on the ballot, I think McGinnity's three additional good pitching seasons put him a little behind Caruthers's additional work (a partial pitching season at a very high quality, one good hittting season as a full-time OF, and the equivalent of about two other superb hitting seasons for the games he played in the OF while still pitching). I think McGinnity certainly deserves points for pitching against tougher competition (at least during his early days in the one-league NL) and for the huge IP gap between him and his contemporaries but his seasons were longer and he put up some of his best seasons during the transition to a two-league system. It's very close, but I now have Caruthers a tad ahead of McGinnity, with both barely on the bubble of deservingness.

Based on these numbers, Caruthers deserves to rank ahead of Buffinton, King, Welch, Cicotte, Waddell, and Joss, all of whom trail him on best 7 pitching seasons once you make an adjustment for his hitting and trail him or come in roughly even with him on other achievements (remembering that we are crediting Caruthers with two additional excellent seasons for the games he played in the OF while still pitching).

The only pitchers who present an issue are Mullane, Willis, and Griffith, all of whom trail Caruthers's 7 seasons by between 75-12f runs and had significantly longer careers than Caruthers. For these guys, it depends on your balance of career, prime, and peak value. Given that I am primarilly a prime voter, I'm inclined to keep Caruthers (and some of the others) ahead of these guys.

In conclusion, I think I signficiantly underrated Caruthers, who justifiably ranks with McGinnity as the best pitcher on the ballot. On the other hand, looking at the slew of pitchers just one or two notches below those guys, I don't think any of the pitchers on the ballot are banging on the door of the HoM.

My revised prelim:
(1) Jackson
(2) Hill
(3) Sheckard
(4) Childs
(5) Van Haltren
(6) Jennings
(7) Wallace
(8) Jones
(9) Ryan
(10) Pike
(11) Caruthers
(12) McGinnity
(13) Duffy
(14) Williamson
(15) Monroe
   179. OCF Posted: May 27, 2004 at 05:45 PM (#647645)
From Andrew Siegel (#172):

a bunch of guys soon to join the ballot would be high on the list (among them: Johnson 887...)

As I said in #147, "The Big Train is the real peak monster."

He's still active, and he's been eating innings reasonably well. I've got his RA+-equivalent record over the last 6 years as 114-77, but that's nowhere near what he could do at his best. We don't yet know that this coming year will be the end of the line for him. What we think of him in 1927 must resemble what the informed baseball fan of 2004 thinks of Roger Clemens. His place among the greats is secure, and even though he's not retired yet, we're already talking about whether he's the greatest ever.
   180. Dag Nabbit is part of the zombie horde Posted: May 27, 2004 at 06:13 PM (#647711)
Does this make sense? What problems do you see?
Any critiques welcome.


Works for me. Looks like nice work. And it keeps Waddell low on my ballot. He's a pitcher with a low total of IP who needs a lot of quality/IP to get on my ballot & the impact of UER knocks him off.

Does anybody have any similar information to compare other pitchers with the above numbers?
Does the average opponent winning %age seem high or low? Based on Joss' 29 decisions a year, he went 18-11 while facing a 15-14 pitcher every time on the mound.


Great info on Joss. I did something similar (only looking at WPct of opposing teams instead of opposing pitchers) with Brown, Griffith, Welch, Caruthers & McGinnity - & then I just looked at median WPct - only sometimes checked on W/L percentage versus them. Found out that Brown & Griffith faced the toughest competition - median opponenet WPct of a little over .520. Figure that teams tended to match up their best pitchers against each other & their opposing pitcher percentage would likely be a little higher. Welch & McGinnity were right around .500. Caruthers was at .480ish.
   181. andrew siegel Posted: May 27, 2004 at 06:32 PM (#647746)
How's this for a totally underrated guy:

I just figured a bunch of other guys using the same methodology as in post 172 and Nap Rucker comes in at 587, good for 11th place, behind 10 HoMers and McGinnity. I'm going to give him another look.
   182. Al Peterson Posted: May 27, 2004 at 06:54 PM (#647782)
This question is for the Friends of Sam Thompson: Why not Mike Tiernan?

Similar hitters but to me dissimilar fielders. Tough to find a fielding metric that likes Silent Mike, even a little. It might be right field we're talking about but if were trying to split hairs on a ballot then this would be a minus in his column.

Also how many New York Giants do we want n the HOM? He played with Ward, Glasscock, George Davis, Ewing, Connor, Gore, Keefe, Rusie. Also considered are Welch and VanHaltren.

I might have him a little low combined with Sam a little high. All fixable so I'll give another glance.
   183. OCF Posted: May 27, 2004 at 07:39 PM (#647856)
How's this for a totally underrated guy: ... Nap Rucker

Andrew - having just confessed to an arithmetical error myself, I'd suggest you re-examine that calculation to see if you've got some formula wrong. For post-1893 pitchers, your list in #172 is generally fairly consistent with mine. You've got this order: Walsh, Covaleski, McGinnity, Plank, Waddell, Shocker, Cicotte, Willis, Vaughn, Mays, Griffith, Cooper. I've got this order: Plank, Walsh, McGinnity, Covaleski, Willis, Waddell, Cicotte, Cooper, Griffith, Shocker, The differences between those lists mostly have to do with peak versus career (you don't do Plank any favors by only looking at 7 years.) There's pretty broad agreement.

But Rucker? I have his RA+-equivalent career record as 156-108. Sure, that's much better than his actual 134-134 record, but as a career accomplishment, that puts him in the same neigborhood as Happy Jack Chesbro, Hippo Vaughn, and Jesse Tannehill. Rucker has the advantage in a 7-year system that nearly 90% of his career innings fall into his first 7 years, but still I don't see how that should make him as good as McGinnity or Waddell.
   184. Jim Sp Posted: May 27, 2004 at 07:45 PM (#647866)
I'd also like to see a reprise of discussions of the extent to which OPS+ is an accurate measure of a hitter's value

OPS+ has really terrible marketing. OPS+ is a pretty good normalized estimator of runs created, not OPS.

You'll hear people say "OPS+ underestimates his contribution because he had high on-base percentage". This is a confusion caused by bad marketing. OPS has this problem, OPS+ does not (well, at least it does not as a first order effect). OPS+ is a variation on the PRO stat from Total Baseball, unfortunately quite a bit of information on it could be found on the baseball-reference.com site, but I'm not finding it right now. Hopefully someone can find the link, it has all the details.

A quick derivation without having the link to refer to:

Assume runs are proportional to OBP*SLG. This is very similar to the basic Bill James runs created formula, but he sets the proportionality constant to 1 and garbles the denominator.

Normalize OBP and SLG for league and park. The difference between PRO and OPS+ as I understand it is a techncial difference in how you do the normalization.

So if a players OBP is 1.03*average for the league and park, then we'll say his OBP' is 1.03.
Similarly for SLG, we'll get SLG'.

His runs created compared to league/park average:

RC' = OBP' * SLG'

For values of OBP' and SLG' close to one, we'll have:

RC' = (1 + delta O') * (1 + delta S')

where delta O' is OBP' - 1, S' is SLG' - 1. delta O' is .03 in my example above.

which for delta O' and/or delta S' close to 0 approximates to

RC' = 1 + delta O' + delta S'

so

RC' = 1 + (OBP' - 1) + (SLG' - 1)

RC' = OBP' + SLG' - 1.

OPS+ is just RC' * 100. A 110 OPS+ is an approximation that a player created runs at a 10% greater rate than average.

It's main limitations (for good hitters, bad hitters aren't too interesting for our purposes):

1) based on OBP*SLG, so ignores a number of factors such as basestealing and GIDP.

2) dropping (delta O' * delta S') means that it understates the run creation extremely good players such as Babe Ruth.

3) dropping (delta O' * delta S') will tend to underrate players who are good at both OBP and SLG, compared to specialists.
   185. andrew siegel Posted: May 27, 2004 at 08:02 PM (#647898)
OCF:

Just to be clear, that's not my list of the player's ranking, just a list of how their first seven years rank according to WARP's PRAR calculation, without adjusting for season length or league quality or pitcher hitting.

As for Rucker, I've added the number again and they're right. But, having examined the raw data and every other system's assessment of Rucker, I'm having a hard time understanding how baseballprosepctus justifies them. If I can't trust the source of the numbers, then maybe the whole project was a waste. Any thoughts?
   186. John (You Can Call Me Grandma) Murphy Posted: May 27, 2004 at 08:03 PM (#647903)
Also how many New York Giants do we want n the HOM? He played with Ward, Glasscock, George Davis, Ewing, Connor, Gore, Keefe, Rusie. Also considered are Welch and VanHaltren.

Glasscock and Gore really don't belong on your list since they were more valuable on other teams (Cleveland and Chicago, respectively).
   187. andrew siegel Posted: May 27, 2004 at 08:04 PM (#647904)
OCF--

P.S.

It's not like the other guys you mention do that badly in the 7-year PRAR calculation-- Vaughn is at 521 and Chesboro and Tannehill are well into the 400's.
   188. Rick A. Posted: May 27, 2004 at 08:20 PM (#647931)
I'm looking to tweak my system a little. (Nothing major, just want to make it a little more automated, and a little less subjective) Does anyone here know where I can find defensive innings played for position players? I've looked on Baseball-Reference, Baseballprospectus, and lahman's database.

Thanks.
   189. Howie Menckel Posted: May 27, 2004 at 08:32 PM (#647950)
Re 19th century NY Giants, I have it this way (10 G minimum, includes 1890 PL, which kept this core):

1885-86 NY NL O'Rourke Ward Keefe Ewing Connor
1887-89 NY NL Gore O'Rourke Ward Keefe Ewing Connor
1890 NY PL Gore O'Rourke Keefe Ewing Connor
1891 NY NL Gore O'Rourke Ewing Connor Glasscock Rusie
1892 NY NL Gore O'Rourke Ewing Rusie Richardson Keeler
1893 NY NL Kelly Ward Connor Rusie Davis
1894 NY NL Ward Connor Rusie Davis
   190. Michael Bass Posted: May 27, 2004 at 09:24 PM (#648015)
JimSP -

I'm intrigued by your post, but I'd like to hear more.

If I read what you're saying correctly, these two players:

Player A - 3% better OBP than league average, 6% better SLG than league average

Player B - 6 % better OBP than league average, exactly league average SLG

...have exactly the same value offensively (other statistics such as SB, GIDP, whatever held constant)?

That's awfully counter intuitive.

And, yes, I know that if you assume that OBP * SLG is a perfect predictor, that the above comparison is true, but you're going to have to point me to something that convinces me that OBP * SLG doesn't underuse OBP.
   191. Michael Bass Posted: May 27, 2004 at 09:27 PM (#648025)
Extra point onto that last post (I gotta get out of the habit of realizing things after posting)

Actually, I also realize that the two players wouldn't be the same in OPS+ because of unrelated point 3) of your flaws list. Leaving that aside, let's change the example to

Player A - 6% above league OBP, league average SLG
Player B - League average OBP, 6% above league SLG
   192. Jim Sp Posted: May 27, 2004 at 09:38 PM (#648051)
Michael,
In your example player A is better than B, I think all runs estimators such as Bill James Runs Created will agree with OPS+ on that. Your example seems to argue for player B--that a 3% difference in OBP is worth more than a 6% difference in slugging, is that what you intended?

By OPS+:

A: 1.03*1.06 is a 1.0918, approximated as 1.09, an OPS+ of 109.

B: 1.06*1 is 1.06, an OPS+ of 106.

Note that player A lost .18 of an OPS+ point as a result of the OPS+ approximation, because he was above average in both categories.

Limitations 2 and 3 refer to the (normally) small inaccuracies such as losing .18 of an OPS+ point above.
   193. Jim Sp Posted: May 27, 2004 at 09:50 PM (#648065)
Never mind, just saw the second post.

A and B are pretty darn close, if A gets the edge it's not by much.

That's what I mean by "at least it does not [bias] as a first order effect [like OPS does]".

OPS has a large bias towards SLG, OPS+ is only biased to the extent that OBP*SLG is biased. I don't have any data on the size of that bias, but someone here probably does.

OPS+ is not normalized OPS, that would be bad. It is normalized OBP*SLG, which is pretty good.
   194. Jeff M Posted: May 27, 2004 at 11:27 PM (#648165)
Does anyone here know where I can find defensive innings played for position players?

A player's total defensive innings at a particular position are in the Win Shares book (although I think they are estimated). It isn't done on a year by year basis though.

I don't think actual defensive innings have been compiled prior to STATS coming on the scene and keeping track.

On p.155 of the Win Shares book, there's an essay about estimating defensive innings. James makes the point that it is a really bad idea to rely too heavily on estimated defensive innings. He says if you are going to do it:

"The best thing to do is just to assume that each player's defensive innings at a position are proportional to his plays made at the position, unless:

....a) this would imply that he was playing more than 9 innings per game, or

....b) this would imply that he was playing less than 3 innings per game."

This apparently works a lot better when a guy played lots of innings at a position.
   195. sunnyday2 Posted: May 28, 2004 at 12:11 AM (#648257)
Great stuff re. Pearce and the Wright brothers, wish we had had it many years ago.

karl wrote that with this info he will move Harry down toward Pearce (and off his ballot), and he should. Harry may have been just as good an athlete and baseball player as Pearce, we certainly do not know today. Henry Chadwick praised Wright more lavishly than Pearce. But Pearce clearly had more value.

Wright was born in 1835 in England, and came to the U.S. as a boy because his father came here to play cricket. Harry was an outstanding cricketer but also played baseball at an early age. The Wrights were gentlemen, however, so Harry played baseball with the Knickerbockers. The Knicks virtually invented the game but, frankly, were never that good. By 1857, working class teams were being formed and they were better than the Knicks from the git-go.

Pearce started with the Atlantics in 1857 at age 21 and over the next 14 years the Atlantics on 254 games and lost 49. Dickey was their #3-6-1-3-1-1-2-2-5-8-2 hitter by runs/game, then bounced around from 2nd to 6th from '68 to '70 as we have more categories of stats to look at. His '68 may provide some insight: He was second in hits and runs but only 6th in TB. You might infer that he was a lead-off or #2 type hitter, at least he seems to have functioned that way. All the while playing mainly SS, then C, then SS again.

Meanwhile, Harry played cricket and baseball, but playing baseball more for fun apparently--that is, with a lousy team though he himself, again, was recognized as one of the better players. He finally moved to the Gothams in '63 but they weren't much good either. He was their best hitter (in R/G, which of course may or may not be really indicative of his ranking) 2 out of three years, and in the 3rd year (1864) he was out-scored by brother G. 2.38 R/G to 2. (He was scoring about 2 R/G throughout that period, while Pearce and his top teammates were scoring 3-4.)

Harry quit baseball in '66 and moved to Cincy to play cricket. He was immediately drawn back into baseball, however, and over 4 years his team (which he organized as well as played for) went 176-14 (46-6) vs. pro teams. Meanwhile, during these same years, the Atlantics were in decline, going 147-35 but only 35-22 against professional teams. But of course they beat Wright's team to end its 100+ game win streak. Pearce continued scoring 3-4 R/G, but now Harry was doing the same. And while Pearce was still playing SS, Harry was pitching and playing 2B as well as the OF. Harry may have been the 6-7-8-9 hitter for the Red Stockings but he was racking up stats similar to those of Pearce and Start. Of course, what we don't know is how many outs any of them used up. But if you want to give Pearce the benefit of the doubt (Harry had better teammates and probably more PA), well, then maybe you have to give Harry the same benefit for his Knick and Gotham years.

Both were in decline by the time of the NA, though Pearce, only one year younger, clearly gets the better of it. He is approx. as good on offense and plays a more important defensive position.

In sum, from '57-'58 through '66 Pearce blows Wright away for *value* though we don't know what their relative skills are. Chadwick called Wright one of the best hitters of his time, while Pearce's defense received much comment.

'67-'70, given that Harry is the principal pitcher in '67 and backup P the rest of the time (pitching about 20% of the team's innings and very very nearly as effectively as Brainard), it is hard to say Pearce was better or more valuable. If using WS, well, the Red Stockings earned a lot more, so even the #7-8 hitter and #2 pitcher could have racked up a lot of them.

After '70, it's mostly Pearce again.

So anyway, I never ever thought Harry was more valuable than Pearce. But for those who have Pearce in your top 5-6 (you know who you are), I don't see Harry as a different caliber of player. Just one who played a little more cricket.
   196. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 28, 2004 at 12:25 AM (#648290)
Chris Cobb--in my new system (counting positional leads in WS as Ink), Johnny Bench comes out in the top three, right where he should be. But Cochrane and Hartnett still don't make it. I guess what I want to know is, is it possible that catcher is just a historically weak position? Cochrane amassed 45.2 WARP3 and 142 WS in his 5 best years, both just around or under my 30-player ballot average (comparable to Van Haltren's 5 best by WARP and Doyle and Chance's 5 best by WS). His career is no more impressive--83.9 WARP3 and 275 WS are also close to the ballot averages, matching up neatly with Cupid Childs' career (if season lengths are straight line adjusted). Does a Van Haltren/Doyle/Chance peak and a Childs career make a sure thing first ballot HoM'er, as I suspect people will consider Cochrane? If so, why?
   197. sunnyday2 Posted: May 28, 2004 at 12:38 AM (#648331)
To be totally blunt, I give catchers a bonus, up to 15% of any uber-stat (like WS) if he played 100% of the time at C. For guys like Bresnahan it is .15 x .xx, with .xx being the percent of career games at C.

I did that because of a comparison of all HoFers at each position and a finding that man-for-man, #1 vs. #1, #2 vs. #2, etc., C have about 15% less value. That AND that there is an obvious explanation for it in terms of wear and tear, as opposed to attributing it to chance.

Obviously there are not a lot of 3B in the HoF, but I found that at the top of the heap, the best 3Bs were nearly as valuable as the best at other positions, and besides I could not satisfy myself that the lack of 3Bs was not chance.

I also give relief pitchers a bonus, without which none would be in the HoF. Count me among those who think some RPs should be in the HoF--at least, Hoyt Wilhelm and Goose Gossage and Eck. Without a boost, no go.

But it is even more obvious that we should have Cs in the HoF (and HoM). C is the one position where the numerical values and the subjective value (as I see it, at least) is most out of whack. Cochrane clearly stands out among Cs, no way would I want Van Haltren in the HoM and not Mickey Cochrane. But then, I want Cupid Childs in my PHoM, too.
   198. Jeff M Posted: May 28, 2004 at 12:41 AM (#648339)
I guess what I want to know is, is it possible that catcher is just a historically weak position?

Catchers are a historically weak hitting position -- no doubt about it -- but I don't think that we can characterize it as a weak position generally when we know it is the toughest defensive assignment on the field.

Two things about that: (1) the wear and tear of catching adversely affects hitting and (2) catchers are not scouted primarily for hitting ability. They are an anomaly.

The WARP and WS numbers you give put Cochrane at the top of the catching position historically.

When (if) the Catcher positional thread comes back up, you may want to scroll through some of the arguments there.
   199. David Concepcion de la Desviacion Estandar (Dan R) Posted: May 28, 2004 at 01:11 AM (#648429)
If they are so defensively valuable, why don't they accumulate 15+ fielding WS/FRAR?
   200. DavidFoss Posted: May 28, 2004 at 01:14 AM (#648442)
I'm not satisfied with the way either WS or WARP handles positional adjustments.

Not really a complaint, but I just don't believe players who play different positions yet accumulate the same WS or WARP are necessarily equal caliber players (or HOM candidates).
Page 2 of 4 pages  < 1 2 3 4 > 

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Francis
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.0312 seconds
49 querie(s) executed