Do three backloggers go in this "year?"

Looks like a safe bet.

I just want to say that I am appalled by the number of voters who are not giving war credit due to "uncertainty". This applies particularly to Keller who may or may not be on the 1996 ballot, but also to other players who had their primes interrupted by the war.

What really riles me up are the people who deny peak/prime war credit on grounds of "uncertainty" or "I don't like giving extra credit", but who then vote for NgL candidates or guys from weak leagues in the pre-1900 era.

What hypocracy! Guys like Keller and Rizzuto were off fighting a war! It's a damned shame that mandatory war credit wasn't included in the HoM constitution, but at the least the voting population should get together and reach some agreement on how to fairly judge wartime players without penalizing them for their decade of birth.

PS: This also goes for the potentially underrepresented NgL players of the wartime/integration era.

Keller is going in in '96, so it's a moot point for him, 'zop.

But was it a just war?



:)

Rick Reuschel's a personal favorite pitcher of mine.

In part it's because I'm a Cubs fan, but it's more. I remember as a kid, this fat ol' pitcher seemingly resurrecting from nowherere to start the All Star Game at age 40, just three years after nearly leading the league in losses.

He was the most out of place member on the 1984 Cubs. Former hero during the dark days hanging on the end of the bench with a bunch of new guys, none of whom had been there when he was in his prime.

He's one of only two men to give up home runs to Hank Aaron and Barry Bonds.

In 1977, he had the worst run support of any 20-game-winner since Nap Rucker. He also just had lousy run support throughout his career.

The most runs his team ever scored in a game was 16. They lost that game. Windy day at Wrigley.

According to David Gassko's (questionable) methodology in the 2007 THT, he's the 27th best pitcher in baseball history, ahead of Eddie Plank, Jim Bunning, Carl Hubbell (!!), Bob Feller (!), Hal Newhouser, Frank Tanana, Don Drysdale, Whitey Ford, Juan Marichal, John Clarkson, Dave Stieb, Pud Galvin, Billy Pierce, Tim Keefe, Dazzy Vance, Ted Lyons, Early Wynn, Red Faber, Eppa Rixey, Sandy Koufax, Ed Walsh, Tommy Bridges, Bob Newsom, Red Ruffing, Joe McGinnity, Al Spalding, Rube Waddll, Danny Darwin, Mordecai Brown, Old Hoss Radbourn, and Babe Adams. . . Hey, I said it was questionable results.

I think Keith Hernandez has a good shot at election. He's a little difficult to compare to Keller because his career shape is so different, but his other top competition is Jimmy Wynn and Nellie Fox.

Offensively he and Wynn are even, and defensively it's a question of how possibly the greatest defensive first baseman of all time compares to a below-average CF and corner outfielder. Hernandez also has about one season of good production on Wynn.

Hernandez is, of course, way ahead of Fox offensively, and Fox is way ahead of him in defensive value, but again evaluation depends on how much weight one places on the positional difference.

Anyway, I think Hernandez has to be considered a serious candidate for election when he is compared to top of the backlog.

What hypocracy! Guys like Keller and Rizzuto were off fighting a war! It's a damned shame that mandatory war credit wasn't included in the HoM constitution, but at the least the voting population should get together and reach some agreement on how to fairly judge wartime players without penalizing them for their decade of birth.

As one of the hypocrites I'd just like to say this is how I see Keller,
If I give him 25 win shares for 1944, and 15 more for 1945 - ALL-STAR Level play but not MVP level play, he has 268 career win shares with 4 MVP type seasons.

That makes him not much different from Augie Galan (263 career WS, 3 MVP type season - though during the War years), Bobby Veach (265 WS, 3 MVP type seasons), Larry Doby's MLB career only (268, 3 MVP type seasons), Bobby Murcer (277, 2 MVP type seasons).

Even guys as ridiculously good as Frank Robinson had seasons in the middle of their career where they were "only" all-stars.

If you say Keller would have definitely posted MVP numbers in the seasons he missed, I think you can make a strong Hughie Jennings/Dobie Moore argument for his induction.

But I also think you are lying to yourself to say you absolutely know that they would have been MVP seasons.

>I think Keith Hernandez has a good shot at election.

I think Keith is under-rated though he is still bouncing around #16-18 for me, preliminarily speaking.

But the way new borderliners are flying' in, I don't doubt that he does.

Also, can someone explain to me how Darrell Evans (119 career OPS+, arguable 3B defense, post-1980 at 1B/DH) flies in on the first ballot, but Greg Nettles (110 career OPS+, definately superb defense) is mired in 27th place?

Anyway, I think Hernandez has to be considered a serious candidate for election when he is compared to top of the backlog.

What role does his drug problems play in the HoM? One manager dumped him for pennies on the dollar, and went to two more World Series anyway. The team that got him went to one, but had its two youngest/brightest stars develop coke habits.

Hit for the cycle in the Rick Camp game.

Hernandez's case will really be a test of how the electorate values prime, I think. He doesn't have the handful of really big years that Jimmy Wynn or Charlie Keller does, but he was an outstanding player for more seasons than they were. He doesn't have the career value that contributed to Darrell Evans' election, but he played at an all-star level much more consistently than Evans did.

I know that he'll be somewhere on the top half of my ballot, but I don't see clearly what the electorate will do with his candidacy.

Will he be competing with Jimmy Wynn for #1, or will he be competing with Bobby Bonds for #34? Given how little difference there is between Wynn and Bonds, the tightness of the ballot boggles my mind. Of course, I think Bonds is being a little bit underrated, partly because his case is mostly prime. Of course, Hernandez doesn't have all those strikeouts. . . It's going to be interesting.

As one of the hypocrites I'd just like to say this is how I see Keller,
If I give him 25 win shares for 1944, and 15 more for 1945 - ALL-STAR Level play but not MVP level play, he has 268 career win shares with 4 MVP type seasons.

That makes him not much different from Augie Galan (263 career WS, 3 MVP type season - though during the War years), Bobby Veach (265 WS, 3 MVP type seasons), Larry Doby's MLB career only (268, 3 MVP type seasons), Bobby Murcer (277, 2 MVP type seasons).

Even guys as ridiculously good as Frank Robinson had seasons in the middle of their career where they were "only" all-stars.

If you say Keller would have definitely posted MVP numbers in the seasons he missed, I think you can make a strong Hughie Jennings/Dobie Moore argument for his induction.

Why would you give credit less Win Shares than he earned in seasons before and after the war? If it's because of uncertainty, don't you need to ding NgL players too? And we don't do that.

We're supposed to make the best estimate of a players' value during seasons in which he was forced not to play by acts of god, war, and segregation. I can't see how in good faith you can argue that 25WS is your "best estimate" for Keller's wartime peak. You really believe that?

Given how little difference there is between Wynn and Bonds, the tightness of the ballot boggles my mind.

Listening to people talk about RPI and who is on the bubble for the NCAA basketball tournament today reminded me a lot of the process of figuring out how to rank the backloggers on the HOM ballot.

There are more factors than anyone can possible sort out without some type of system, but when the RPI spits out an answer somone doesn't like they claim it doesn't work. Of course it WORKS! You give it data, tell it how to rank the data and then wait for the output. If it says Southern Illinois is a top 10 team and Nellie Fox is in the HOM you can't then say, well it place too much weight on conference wins and second base defense.

What is worth more, road wins or career win share totals? Which do you think is better, wins against ranked opponents or MVP type seasons? Do you give War Credit and how do deal with a team that started the season hot but has played .500 ball down the stretch?

Hernandez will be second on my ballot.

Dave, 9 points of career OPS+ is a pretty substantial difference. Evans had the Big Year (1973) that Nettles never did. Plus, clearly the electorate *was* persuaded by WS/WARP's take on Evans's fielding as Gold Glove level. That said, I vote for Nettles.

I can understand crediting Keller at a 25-WS level rather than a 30-WS level for his war time. Think of it this way--if you did a PECOTA or ZIPS projection for Keller after 1943, would it really come out to 30 WS? Don't you have to factor in the risk that he gets hurt, or drops off, etc? Those systems only project 30-WS type seasons for the Pujolses of the world; all the other stars come out in the 25 range, because no one besides Pujols is really a better-than-50% bet to accumulate over 30 WS, regardless of whether they've strung together three straight 30-WS seasons.

I don't know that the electorate was persuaded by the WS/WARP defense - I think if they had, he would have had Ted Simmons-type vote totals. I think they were convinced he was at least a good third baseman, if not Schmidt-level. And let's not forget, the guy did have over 400 home runs. It's not like he didn't bring anything to the table.

From the quick lines of new eligible players:

363 116.1 1971 Darrell Evans-3B/1B
321 105.2 1969 Graig Nettles-3B

Evans has a 42 point lead in WS and an 10.9 lead in WARP3. Considering the number of voters who use one or the other of those tools, I'm not surprised that Evans was elected before Nettles. It's certainly not the travesty you're making it out to be.

Charlie Keller – a quick and dirty study, batters only in the study below.

Keller had 102 Win Shares from age 24-26.

Only 12 men have had more, and all of them are in the HOM except Will Clark.

35 men have had more than 90 in those three seasons.

These players had totals within 10 of Keller.



Throwing Keller out, at age 27 they averaged 29 win shares, throwing out Keller and DiMaggio they averaged 26 at age 28.

I guess you could say I am under rating him, and you could point out Bonds. I could point out Cronin and say I’m not. But I feel pretty comfortable saying Mattingly (hey, a bad back guy, I’m sure it’s a coincidence).

Another chart Win Share from the Age 30 season to the end of their career


That’s Bonds through age 36, old spreadsheet I guess.

A thread for Dan Quisenberry? In the early '80s, James made some persuasive arguments that he was the best pitcher in the AL in 1982 and 1983, really he and Stieb. How did Pete Vukovich win the CY in 1982?

Dave, 9 points of career OPS+ is a pretty substantial difference.

Aren't you the guy who's always going on about positional adjustments?

I would like to take this opportunity to urge each and everyone of us to take some time in the next two weeks to give proper reevaluations to each of the top 20 guys (or more) in our backlog. Very likely that 3 will go in in 3 weeks time. Let's do our best to make sure that they are the 3 most worthy:
last year name last year's vote total
4 Charlie Keller 386
5 Jimmy Wynn 370
6 Nellie Fox 348
7 Edd Roush 336
8 Pete Browning 313
9 Jake Beckley 297
10 Rollie Fingers 292
11 Charley Jones 269
12 Tony Perez 263
13 Bob Johnson 257
14 Bucky Walters 253
15 Cannonball Dick Redding 242
16 Hugh Duffy 236
17 Gavvy Cravath 227
18 Alejandro Oms 219
19 Roger Bresnahan 171
20 George Van Haltren 167
21 Tommy Leach 161
22 Dizzy Dean 147
23 Burleigh Grimes 140

DCW3--um, not for two guys who play the same position....

how Darrell Evans (119 career OPS+, arguable 3B defense, post-1980 at 1B/DH) flies in on the first ballot, but Greg Nettles (110 career OPS+, definately superb defense) is mired in 27th place
Because OPS+ understates the large difference in offense between them.

TomH, how so? Neither was a great baserunner, both had tons of K...I guess Evans was a bit more OBP-heavy, but that's worth 1-2 points at most...

I don't vote (I don't feel confident rating players pre-WWII, and I'm really leery of rating players across leagues), but I read alot of the discussion in the threads. I wonder about war adjustments, but I really question "length of season" adjustments.

For war adjustment, I often think of a guy like Ted Williams as the perfect example of my questions: WWII, obviously, took significant time from the prime of his career. I think it obvious you should make an adjustment. Korea, however, took years from his mid-30's. Isn't it possible/probable that such a player is just delaying the end of his career (and I use Williams because he was awfully productive to an "old" age)?

But that isn't even quite the issue that "length of season" adjustments have in my mind. Much is made of (for instance) Beckley's stat adjustments to a 162 game schedule; is that appropriate? I'm guessing that playing 20-30 fewer games a season would add a couple of years of career length - if anything, it might enhance his counting stats, not suppress them.

As a non-voter, I don't expect a response, but I'd appreciate if someone would point out the error(s) in my thinking.

relative OBP and SLG to league/park avg
Nettles OBP +0.6% SLG +9.4%
Evans. OBP +9.1% SLG +9.9%

If you adjust Evans' OBP by subtracting out 60+ fewer GIDPs that he hit into (relative to Nettles)--
Evans OBP +11.0% SLG +9.9%

I weight an increase in OBP about 50% more than SLG; so
Nettles .6*1.5 + 9.4 = 10.3
Evans 11*1.5 + 9.9 = 26.4

Which is a lot bigger than the OPS+ diff of 119 to 110.

SB-CS ##s would not change it much at all.

Prelim

PHoM: Tony Perez, Keith Hernandez, Cupid Childs

1. Tony Perez
2. Jake Beckley
3. Rusty Staub
4. George Van Haltren
5. Edd Roush
6. Jimmy Wynn
7. Tommy Leach
8. Nellie Fox
9. Mickey Welch
10. Graig Nettles
11. Lou Brock
12. Keith Hernandez
13. Norm Cash
14. Hugh Duffy
15. Reggie Smith

16-20. Bonds, Singleton, Cepeda, Johnson, Browning
21-25. Willis, Redding, S.Rice, Streeter, Grimes
26-30. McCormick, Strong, W.Davis, Doyle, Greene

Hernandez will be second on my ballot.

Hernandez will be forty-third on my ballot.

I second the call for a Quisenberry thread. Short career, but not too many relievers have put up a 6 year run like he did from 1980-1985.

A thread for Dan Quisenberry?

I'm going to be posting more in a week. I didn't want to knock all preexisting threads off of the Hot Topics section.

I do want to point out here that in the vase of Keller, I highly doubt that his career was extended because of the war. Many have argued (did Mike Webber just now?) that Keller's back would have acted up sooner if he had been playing ball in 1944 and 1945. I would counter by saying that ball players (with the exception of Joe D who was sick or hurt most of teh time) were either a) fighting a war like Feller or Williams, or b) playing baseball year round and all over the globe for troop morale. Both of those things are at lest as taxing as playing baseball and if his back was ready to go out at any time it would have.

Plus in 1945, Keller had 9 WS in in 44 G, a tiny bit more than a quarter of a season. (all of my season totals have been adjusted to 162 games). That would equal 33 or 34 WS, so not only did he play at an MVP (30 WS) level in 1942, 43, and 46, but he also did so upon his return in 1945. I dont' see much reason (at lesat for me) not to give him 30-31 WS during those two seasons.

I can understand crediting Keller at a 25-WS level rather than a 30-WS level for his war time. Think of it this way--if you did a PECOTA or ZIPS projection for Keller after 1943, would it really come out to 30 WS? Don't you have to factor in the risk that he gets hurt, or drops off, etc? Those systems only project 30-WS type seasons for the Pujolses of the world; all the other stars come out in the 25 range, because no one besides Pujols is really a better-than-50% bet to accumulate over 30 WS, regardless of whether they've strung together three straight 30-WS seasons.


You're absolutely right, Dan. If all the information we had about Keller was his pre-war seasons, then I concur that, projecting one season foward we would have to assign a 25 WS value.

But that's not what we're presented with here. We have data from before and after the seasons we need to project, and both the before AND the after suggest a 30WS/162G rate for the missing time. To give Keller less than his established level of play unfairly penalizes him, because while he might have had an off-year and only rung up 25 WS, he might have had a career year and rung up 36. War credit should be given at the established level of play; value should not be subtracted for "uncertainty".

1996 (elect 3)

In all its glory, all the way down to Dick McAuliffe at #150. Hernandez debuts at #17, Quiz at 22, Lynn at 34, Frank White at 149. Reuschel and Downing were close to the top 150 (in the 150s) but its not like they were gonna work their way up to ballot consideration or anything so they're not listed.

I’m comfortable with my top 5 but in a backlog year it gets dicey as early as #6. But I’ve made those choices long ago (PHoM process), though I’m uncertain of Ellie Howard or Dick Redding at #15.

Willie Keeler, Darrell Evans and Ezra Sutton (after all these years) are prelim PHoMers but I’m gonna fiddle with that a bit. Hernandez, Perez, Leach, Newcombe and Trouppe are also in the running.

Oh, and now that I’m PHoMing all of these HoMies, I hope you guys will reciprocate and HoM some of my PHoMies. Actually all but Wynn and Beckley among the top 10 would be good though Roush would be my first preference, Keller second, and then I don't really feel strongly about Fingers or Browning (I mean I like 'em both but the preference between the two is slight).

1. Edd Roush (2-2-4, PHoM 1976)—nice peak of 38*-33-30 (*short WWI season adjusted to 154), very well-rounded skills.

2. Charlie Keller (3-4-6, PHoM 1985)—the ultimate all-peak candidate.

3. Rollie Fingers (4-6-7, PHoM 1991)—there’s no uber-stat that says Fingers is ballot-worthy, but I go back to Chris Cobb’s old test—who do you want in the HoM? And on that simple basis, subjective as it is, I want the #3 reliever of all-time (as of 1995) in mine.

4. Pete Browning (6-15-10, PHoM 1961)—do you believe in numbers?

5. Addie Joss (5-8-9, PHoM 1967)—not a workhorse, to be sure, but damn effective and not a “small sample” of IP.

6. Orlando Cepeda (8-9-16, PHoM 1987)—pretty interchangeable with F. Howard, Cravath and (now) Tony Perez, but the best of the group.

7. Reggie Smith (9-13-17, PHoM 1988)—cannot quite see how he’s not better than Jim Wynn.

8. Charley Jones (10-17-30, PHoM 1921)—even with hefty AA discounts.

9. Phil Rizzuto (7-10-22, PHoM 1995)—306 career WS with war credit. His tenth best season is 21 WS, compared to Nellie Fox at 19, Darrell Evans at 18, e.g.

10. Larry Doyle (11-5-15, PHoM 1975)—same OPS+ as Ed Roush.

11. Ed Williamson (12-14-13, PHoM 1924)—more peak and more glove than Darrell Evans.

12. Frank Howard (14-23-27, PHoM 1987)—PHoM opening forced me to re-evaluate a lot of guys, and his numbers are just too good.

13. Nellie Fox (13-12-8, PHoM 1971)—I’ve decided I like Rizzuto a bit better, though they are two very very comparable players. I think Phil’s peak was a bit more valuable.

(13a. Darrell Evans [18-new], PHoM 1996—not quite the peak I usually like to see, plenty of career, high borderline)

14. Gavvy Cravath (21-21-14, PHoM 1995)—if I’m in a generous mood with MLEs, which I am today.

15. Elston Howard (15-7-12, PHoM 1994)—never really thought of him as a HoM or HoF or PHoM type of player, but I now see him as one of those few players whose opportunities were least commensurate to his ability. Still prefer him to Trouppe.

Very Very Close to Ballot

16. Dick Redding (16-11-26, PHoM 1971)—next best arm, great peak.

(16a. Willie Keeler [24a-30a-22a], PHoM 1996—after all these years)

(16b. Ezra Sutton [18a-39a-32a], PHoM 1996—after all these years, double. Thank Darrell Evans for having reviewed his case.)

Close—i.e. right around in/out line, as I think we will elect another 10 or so backloggers before we’re done

17. Keith Hernandez (new)—under-rated, but likely to get over-rated here

18. Tony Perez (19-19-18)
19. Don Newcombe (20-22-20)
20. Tommy Leach (21-16-29)
(20a. Quincy Trouppe [22-50-54])

21. Norm Cash (21-43-31)
22. Dan Quisenberry (new)—the closest thing to Gossage’s peak, clearly if narrowly better than Sutter
23. Al Rosen (24-18-37)
24. Hack Wilson (25-28-39)
25. Burleigh Grimes (26-38-70)

HoVVG

26. Roger Bresnahan (27-31-36)
27. Jim Wynn (28-52-52)
28. Sal Bando (29-29-49)
29. Johnny Pesky (30-80-76)
30. Eddie Cicotte (31-24-29)
(30a. Jim Bunning [31a-30b-28a])

31. Ken Singleton (32-71-72)
32. Bruce Sutter (33-44-new)
(32a. Joe Sewell [33a-30a-28b])
33. Alejandro Oms (34-40-53)
34. Fred Lynn (new)—never fulfilled his promise, but how many turn out to be the second coming of Joe DiMaggio; what he did was pretty damn impressive
(34a. Ken Boyer [33b-22a-22a])
35. Bob Elliott (35-57-61)
(35a. Don Sutton [35a-20-new])
36. Lefty Gomez (36-32-46)
37. Luis Tiant (37-48-85)
38. Thurman Munson (38-37-19)
39. Chuck Klein (39-30-24)
40. Luke Easter (40-25-60)

41. Vern Stephens (41-33-33)
42. Bill Monroe (42-46-63)
43. Bobby Bonds (43-45-51)
(43a. Jimmy Sheckard [43a-45a-54a])
44. Bus Clarkson (44-82-66)
45. Cesar Cedeno (45-49-32)
46. Wally Berger (46-47-40)
47. Tommy Bridges (47-58-93)
48. Bucky Walters (48-27-25)
49. Fred Dunlap (49-59-58)
(49a. Cool Papa Bell [49a-49a-76a])
50. Graig Nettles (50-53-new)

51. Pie Traynor (51-39-45)
52. Bob Johnson (52-60-64)
53. Luis Tiant (53-new)
(53a. Wes Ferrell [53a-51a-54b])
54. Hilton Smith (54-51-43)
55. Dick Lundy (55-26-23)
56. Dizzy Dean (56-33-14)
57. Dave Bancroft (57-35-38)
58. Tommy Bond (58-34-11, PHoM 1929)

HoVG

59. Jake Beckley (59-54-48)
60. Vic Willis (60-63-56)

61. Hugh Duffy (61-62-42)
62. Gene Tenace (62-75-73)
63. Jim Rice (63-new)
64. Ben Taylor (64-73-75)
(64a. Biz Mackey [64a-40a-65b])
65. John McGraw (65-64-77)
66. Frank Chance (66-69-69)
67. Bobby Estalella (67-42-47)
68. Tony Oliva (68-76-50)
(68a. Joe Kelley [68a-64a-65a])
(68b. Pete Hill [68b-81a-87a])
69. Lou Brock (69-65-57)
70. Rocky Colavito (70-56-55)

71. Mickey Welch (71-70-35)
72. Vada Pinson (72-78-81)
73. Wilbur Cooper (73-67-NR)
74. Marvin Williams (74-68-44)
75. Jim McCormick (75-83-34)
(75a. Billy Pierce (75a-70b-76b)
76. Dave Concepcion (76-41-new)
(76a. Early Wynn [76a-70a-60a)
77. George Van Haltren (77-74-65)
78. Rusty Staub (78-71-63)
79. Jimmy Ryan (79-79-90)
80. Jim Fregosi (80-84-88)

81. Tommy John (81-new)
82. George Burns (82-81-76)
83. Tony Mullane (83-72-59)
84. John Clapp (84-86-62)
85. Ron Cey (85-61-82)
86. Ernie Lombardi (86-66-89)
87. Herman Long (87-NR)
88. Urban Shocker (88-85-96)
89. Bert Campaneris (89-90-83)
90. Boog Powell (90-NR-NR)

91. Red Schoendienst (91-92-HM)
92. Dolf Luque (92-HM-84)
93. Luis Aparicio (93-55-41)
94. Tony Lazzeri (94-88-HM)
95. Artie Wilson (95-94-97)
96. Bobby Veach (96-89-80)
97. Wally Schang (97-97-NR)
98. Maury Wills (98-77-NR)
99. Mike Tiernan (99-HM-74)
100. Buddy Bell (100-new)

Honorable Mention

101. Kiki Cuyler (101-79-72)
102. Carl Mays (102-92-89)
103. Jim Kaat (103-98-99)
104. Ellis Kinder (104-HM-NR)
105. Steve Garvey (105-HM-new)
107. Rabbit Maranville (107-86-81)
108. Joe Tinker (108-HM-67)
109. Johnny Evers (109-HM-68)
110. Bobby Murcer (110-HM-new)

111. Gil Hodges (111-HM-97)
112. Spot Poles (112-HM-HM)
113. Al Oliver (113-HM-87)
114. Cecil Travis (114-NR-NR)
115. Billy Nash (115-NR-NR)
116. Mickey Vernon (116-HM-95)
117. Andy Cooper (117-HM-100)
116. Bill Byrd (116-HM-HM)
119. Sol White (119-HM-HM)
120. Pancho Coimbre (120-HM-HM)

121. Bobby Avila (121-NR-NR)
122. Catfish Hunter (122-HM-HM)
123. Silvio Garcia (123-HM-99)
124. Hippo Vaughan (124-HM-NR)
125. Jake Fournier (125-HM-HM)
126. Lon Warneke (126-HM-HM)
127. Amos Otis (127-HM-92)
128. Roy White (128-HM-NR)
129. Lave Cross (129-HM-NR)
130. Jose Cruz (130-HM-new)

131. George Scales (131-HM-91)
132. Sam Rice (132-HM-NR)
133. Denny Lyons (133-HM-HM)
134. Vida Blue (134-HM-NR)
(134a. Pud Galvin [134-NR-NR])
135. Silver King (135-HM-HM)
136. Fielder Jones (136-HM-NR)
137. Davey Lopes (137-HM-NR)
138. Jim Whitney (138-HM-HM)
139. Virgil Trucks (139-HM-HM)
140. Dave Orr (140-HM-HM)

141. Roger Maris (141-HM-NR)
142. Sparky Lyle (142-HM-HM)
143. Ray Dandridge 143-(HM-NR)
144. Darrell Porter (144-new)
145. Jim Creighton (145-NR-NR)
146. Mickey Lolich (146-HM-NR)
147. Ron Guidry (148-new)
148. Leroy Matlock (150-NR-NR)
149. Frank White (new)
150. Dick McAuliffe (147-NR-NR)

You're absolutely right, Dan. If all the information we had about Keller was his pre-war seasons, then I concur that, projecting one season foward we would have to assign a 25 WS value.

But that's not what we're presented with here. We have data from before and after the seasons we need to project, and both the before AND the after suggest a 30WS/162G rate for the missing time. To give Keller less than his established level of play unfairly penalizes him, because while he might have had an off-year and only rung up 25 WS, he might have had a career year and rung up 36. War credit should be given at the established level of play; value should not be subtracted for "uncertainty".

So ballplayers always play at their established level of play? Frank Robinson in 1961,62, and 64 put up 34,41,33 Win Shares. If he missed 1963 where do you put him at? Is it the 23 Win Shares he actually earned?

I'm still going to curb my enthusiasm on Keller some when giving credit.

1996 Prelim Ballot

1) Keith Hernandez - I've been wanting to go over this one for weeks. I've gone over the numbers in detail and the spreadsheet keeps throwing him up on the top of the heap. I'm puzzled by this because I never once thought while following Keith Hernandez' career that he was a clear 1st ballot top of the heap HoM player. Heck, I never voted for him for all-star balloting. The bat isn't quite as good as Norm Cash but I believe Keith's defensive value is much higher and I have Cash #4. As I look down the ballot I can see that Hernandez is clearly worth voting for considering the other players I have on there. But still, #1?!
2) Luis Tiant - may benefit from my indecision on Hernandez and take 1st. I'm certain he was about as good as Drysdale, Pierce, Bunning and Marichal. Why doesn't anyone else think so?
3) Bob Johnson
4) Norm Cash - I wonder if stdevs puts him closer to Hodges
5) Tommy Bridges - where are all the pitchers we need to elect going to come from? Ryan, Blyleven, Gossage and Eckersley are the only sure things in the next 10 years. That's 13% pitching from here on out. We're already down to 28% pitching in our hall and that would drop us down to 26% pitching. There are only 5 pitchers in the top 20 backlog - Fingers, Redding, Walters, Dean and Grimes and I'm not considering anyone but Redding at this time. Tiant blows the backlog away and I like Bridges better also. I think Redding needs to resurface in the top 10.
6) Tony Perez
7) Jake Beckley
8) Reggie Smith
9) Bus Clarkson - We owe Clarkson and Easter our due diligence. There should be a fairly intensive research project for each of them. I am convinced that Clarkson is the best glove man available. If he was a Polanco/Larkin level fielder he should be #1 on the ballot.
10) Rick Reuschel - a better version of Dutch Leonard. Decent hitter for a pitcher and a very good fielder. The spreadsheet says 10th and it ignores his fielding.
11) Gavy Cravath
12) Rusty Staub
13) Virgil Trucks
14) Jim Wynn
15) Edd Roush - I'm glad they both fit on. I felt bad giving Wynn a 5 point boost for making the ballot when there was only one slot between them
16-20) Dutch Leonard, Orlando Cepeda, Bob Elliott, Ron Cey, Charlie Keller
21) Cannonball Dick Redding - could rise if I give peak performance a little more boost to get more pitching more representation in my pHoM. I don't know how you could vote for Vic Willis and NOT his contemporary Dick Redding.
22-25) Jack Quinn, Vic Willis, Urban Shocker, Johnny Evers
26-30) Tommy John, Dizzy Trout, Luke Easter, Dave Bancroft, Hilton Smith
31-35) Frank Howard, Alejandro Oms, Bobby Bonds, Charley Jones, Pete Browning
Pete and Charley move up as I re-do their credits and demerits. Oms and Bonds look like copies of each other.
36-40) Tony Lazzeri, Roger Bresnahan, Jerry Koosman, Jim McCormick, Rollie Fingers

So ballplayers always play at their established level of play? Frank Robinson in 1961,62, and 64 put up 34,41,33 Win Shares. If he missed 1963 where do you put him at? Is it the 23 Win Shares he actually earned?

I'm still going to curb my enthusiasm on Keller some when giving credit.

Don't be obtuse. The point is that the mean expected value of the missing season is X, which is his established level of play. The possible outcomes for that season are (roughly) normally distributed around that mean. If a players established level of play is 30 WS, then the most probable outcome for the missing season is 30WS, and it is, more or less, equally likely that he would have earned 35 WS or 25 WS.

Don't you see this? You can cherrypick Robinson's 1963, but you could also pick years where Robinson exceeded his established level of play. If you only give credit for the worst reasonable possible outcome (say, 1 stdev below the expected value), then you aren't giving a player full credit for the time he missed! If you can justiify giving Keller 25WS for his missing time, then you can equally justify giving him 35.

*(actually, seasonal value is probably non-normal at these ultra high levels of play, but it's close enough that assuming a normal distribution gives you the most accurate result.

Maybe I'm in the minorty, but where I come from, there is no such thing as "established level of play". There definitely is no established anything for a kid who's only spent 4 full seasons in the majors.
Credit, yes. I go with career norms. But nothing is established. If baseball had established things, then the highest payroll every year would always win the World Series and I would find other way to kill my time.
Any mention of established level of play is absurd at best. What we can have is a probable level of play.
BTW - I do not give minor league credit for show me seasons either. Much too slippery a slope.

'zop, that might be true for rate, but not for playing time...since you can never play more than every game, the most likely outcome had he played would not have been 154-game seasons but probably I dunno 140-145. I certainly think you have to deduct for the chance of an injury, even if someone has a great health record otherwise...

If Keller earned 30 wins shares in 1944 and 1945 he would have had six straight seasons with more than 29 Wins Shares. How many players have ever had 6 straight seasons of play at that level?

Only 304 Batters in MLB history have had more than 180 Win Shares in their career and at least one season with 29 Win Shares.

Honus Wagner 13 straight – Best Season - 59
Lou Gehrig – 12 straight – Best Season - 44
Hank Aaron – 11 - 41
Willie Mays – 10 - 43
Babe Ruth – 9 and 8 straight - 55
Mickey Mantle – 9 - 51
Eddie Collins – 9 - 43
Stan Musial – 8 - 46
Ty Cobb – 7 - 48
Joe Morgan – 7 - 44
Mel Ott – 7 - 38
Tris Speaker – 6 - 51
Joe DiMaggio – 6 – 41

So 13 players have done it, and it’s the inner circle group.

Keller’s best season was 36, which is behind everyone on this list, and Mel Ott is the only other player without a 40+ Win Share season.

Keller probably would have had a 40 against the weakened wartime competition.

Keller probably would have had a 40 against the weakened wartime competition.

Well, everyone should give that as much weight as they do when they evaluate Dizzy Trout's 42 Win Share season in 1944.
Actually only Appling and Trout had 40 Win Share seasons between 1943 and 1945.

Don't be obtuse. The point is that the mean expected value of the missing season is X, which is his established level of play. The possible outcomes for that season are (roughly) normally distributed around that mean. If a players established level of play is 30 WS, then the most probable outcome for the missing season is 30WS, and it is, more or less, equally likely that he would have earned 35 WS or 25 WS.

Don't you see this? You can cherrypick Robinson's 1963, but you could also pick years where Robinson exceeded his established level of play. If you only give credit for the worst reasonable possible outcome (say, 1 stdev below the expected value), then you aren't giving a player full credit for the time he missed! If you can justiify giving Keller 25WS for his missing time, then you can equally justify giving him 35.

*(actually, seasonal value is probably non-normal at these ultra high levels of play, but it's close enough that assuming a normal distribution gives you the most accurate result.

Yes I cherrypicked Robinson's 1963. Just to show a guy, age 27, who when everything should have gone right, had a down year. Happens but you're right, my bad on selective data choice.

I went back and looked at what I do for Keller in 1944, 1945. 28 WS each. I'm good with that...

I am an ardent advocate of giving credit for seasons missed due to war. But the amount of credit I give depends upon how established the level of play is (i.e., number of seasons before and after the war the established level of play is based upon). In general, I credit at around 90% total value for the reasons that have been mentioned above -- missed games due to injury, etc., just having a down year, etc.

> I credit at around 90% total value for the reasons that have been mentioned above --
> missed games due to injury, etc., just having a down year, etc

How does this work for people who miss multiple seasons? Do you assume 3 down years consecutively?

No thread for Chet Lemon?

When comparing players from different positions, we all attempt to determine how much different a shortstop
hits, for example, than a right fielder. In so doing, should we use the positional 'average' or some type
of 'replacement'?

RCAP uses average. Advantages: easy to calculate, & it's an accepted mathematical measure.
Disadvantage: a few stars really skew the average, and you can't often replace your player's value with aWhen comparing players from different positions, we all attempt to determine how much different a shortstop
hits, for example, than a When comparing players from different positions, we all attempt to determine how much different a shortstop
hits, for example, than a right fielder. In so doing, should we use the positional 'average' or some type
of 'replacement'?

RCAP uses average. Advantages: easy to calculate, & it's an accepted mathematical measure.
Disadvantage: a few stars really skew the average, and you can't often replace your player's value with a
magically average player.

WS and WARP measure value above 'replacement, but thye really implicity use a Historical Average; 2B over
some time period hit XX worse than 1B, ergo they get a positional bonus relative to 1Bmen.
ADV: not susceptible to random short-term fluctuations
DISADV: insensitive to TRUE fluctuations which actually do skew the market

Dan R's methdos use the worst few MLB regulars at their positions to find 'replacement'.
ADV: Uses actual data comparing the player to value the team likely would have had if the player had been
lost to them
DISADV: Ignores the other players in the positional "pool"

Okay, Tom, what do you mean by that last statement?

The key is this: when determining value above replacement, how much do the other players at his position
matter? Are other stars who play that spot irrelevant? How about the average guys?

Consider this thought exercise:

You are drating a 'team' that consists of a total of 4 'players', each who play one specific position.
You are drafting against 5 other 'owners'.

The players have known values, which are described on a 0 to 10 scale.

Player values are as follows

Position A: 9 5 4 3 2 1 avg = 4 ; bottom = 1 ; avg of last 5 = 3.0 ; avg of last 3 = 2
Position B:10 9 8 6 2 1 avg = 6 ; bottom = 1 ; avg of last 5 = 5.2 ; avg of last 3 = 3
Position C: 9 9 6 4 2 0 avg = 5 ; bottom = 0 ; avg of last 5 = 4.2 ; avg of last 3 = 2
Position D: 8 4 3 1 1 1 avg = 3 ; bottom = 1 ; avg of last 5 = 2.0 ; avg of last 3 = 1

This kind of situation occurs regularly in the sterile environment of simulation games, such as ABPA,
Diamond Mind, Strat-o-Matic, and Scoresheet. It loosely mimics real life.

So, who is your first pick?

By RCAP (Value above average at position), player "D8" is 5 pts above the overall avg, and 6 above the
remaining D players. Same with A9.

If you fear getting stuck with the worst at each spot, B10 and C9 are 9 pts above their worst counterparts.

If you use the bottom 3 as a proxy for replacement, A9 B10 C9 and D8 are all equally 7 pts above replacement.

None of the three above metrics are foolproof, "correct", or "best".

--
Going back to real life, if shortstops happen to be (at a given moment in time) a combination of studs
and yuks, how do we value a stud SS? Value over avg? That is corrupted by other studs. Over yuk? But if
I lose my stud, I can through trades or free agency acquire some other stud or even avg SS, so you cannot
completely discount that they exist.

Hence, I conclude that meausres like RCAP and WS/WARP general replacement level and Dan R's method
of position-specific replacement level ALL contribute to general knowledge. In my pracitce at these sim
games, I've found that what's left at the bottom is somewhat MORE important than the other studs,
but the overall average of play at a position IS a relevant piece of data that we ought not to discard.right fielder. In so doing, should we use the positional 'average' or some type
of 'replacement'?

RCAP uses average. Advantages: easy to calculate, & it's an accepted mathematical measure.
Disadvantage: a few stars really skew the average, and you can't often replace your player's value with a
magically average player.

WS and WARP measure value above 'replacement, but thye really implicity use a Historical Average; 2B over
some time period hit XX worse than 1B, ergo they get a positional bonus relative to 1Bmen.
ADV: not susceptible to random short-term fluctuations
DISADV: insensitive to TRUE fluctuations which actually do skew the market

Dan R's methdos use the worst few MLB regulars at their positions to find 'replacement'.
ADV: Uses actual data comparing the player to value the team likely would have had if the player had been
lost to them
DISADV: Ignores the other players in the positional "pool"

Okay, Tom, what do you mean by that last statement?

The key is this: when determining value above replacement, how much do the other players at his position
matter? Are other stars who play that spot irrelevant? How about the average guys?

Consider this thought exercise:

You are drating a 'team' that consists of a total of 4 'players', each who play one specific position.
You are drafting against 5 other 'owners'.

The players have known values, which are described on a 0 to 10 scale.

Player values are as follows

Position A: 9 5 4 3 2 1 avg = 4 ; bottom = 1 ; avg of last 5 = 3.0 ; avg of last 3 = 2
Position B:10 9 8 6 2 1 avg = 6 ; bottom = 1 ; avg of last 5 = 5.2 ; avg of last 3 = 3
Position C: 9 9 6 4 2 0 avg = 5 ; bottom = 0 ; avg of last 5 = 4.2 ; avg of last 3 = 2
Position D: 8 4 3 1 1 1 avg = 3 ; bottom = 1 ; avg of last 5 = 2.0 ; avg of last 3 = 1

This kind of situation occurs regularly in the sterile environment of simulation games, such as ABPA,
Diamond Mind, Strat-o-Matic, and Scoresheet. It loosely mimics real life.

So, who is your first pick?

By RCAP (Value above average at position), player "D8" is 5 pts above the overall avg, and 6 above the
remaining D players. Same with A9.

If you fear getting stuck with the worst at each spot, B10 and C9 are 9 pts above their worst counterparts.

If you use the bottom 3 as a proxy for replacement, A9 B10 C9 and D8 are all equally 7 pts above replacement.

None of the three above metrics are foolproof, "correct", or "best".

--
Going back to real life, if shortstops happen to be (at a given moment in time) a combination of studs
and yuks, how do we value a stud SS? Value over avg? That is corrupted by other studs. Over yuk? But if
I lose my stud, I can through trades or free agency acquire some other stud or even avg SS, so you cannot
completely discount that they exist.

Hence, I conclude that meausres like RCAP and WS/WARP general replacement level and Dan R's method
of position-specific replacement level ALL contribute to general knowledge. In my pracitce at these sim
games, I've found that what's left at the bottom is somewhat MORE important than the other studs,
but the overall average of play at a position IS a relevant piece of data that we ought not to discard.
magically average player.

WS and WARP measure value above 'replacement', but they really implicity use a Historical Average; 2B over some time period hit XX worse than 1B, ergo they get a positional bonus relative to 1Bmen.
ADV: not susceptible to random short-term fluctuations
DISADV: insensitive to TRUE fluctuations which actually do skew the market

Dan R's methdos use the worst few MLB regulars at their positions to find 'replacement'.
ADV: Uses actual data comparing the player to value the team likely would have had if the player had been lost to them
DISADV: Ignores the other players in the positional "pool"

Okay, Tom, what do you mean by that last statement?

The key is this: when determining value above replacement, how much do the other players at his position
matter? Are other stars who play that spot irrelevant? How about the average guys?

Consider this thought exercise:

You are drating a 'team' that consists of a total of 4 'players', each who play one specific position.
You are drafting against 5 other 'owners'.

The players have known values, which are described on a 0 to 10 scale.

Player values are as follows

Position A: 9 5 4 3 2 1 avg = 4 ; bottom = 1 ; avg of last 5 = 3.0 ; avg of last 3 = 2
Position B:10 9 8 6 2 1 avg = 6 ; bottom = 1 ; avg of last 5 = 5.2 ; avg of last 3 = 3
Position C: 9 9 6 4 2 0 avg = 5 ; bottom = 0 ; avg of last 5 = 4.2 ; avg of last 3 = 2
Position D: 8 4 3 1 1 1 avg = 3 ; bottom = 1 ; avg of last 5 = 2.0 ; avg of last 3 = 1

This kind of situation occurs regularly in the sterile environment of simulation games, such as ABPA,
Diamond Mind, Strat-o-Matic, and Scoresheet. It loosely mimics real life.

So, who is your first pick?

By RCAP (Value above average at position), player "D8" is 5 pts above the overall avg, and 6 above the
remaining D players. Same with A9.

If you fear getting stuck with the worst at each spot, B10 and C9 are 9 pts above their worst counterparts.

If you use the bottom 3 as a proxy for replacement, A9 B10 C9 and D8 are all equally 7 pts above replacement.

None of the three above metrics are foolproof, "correct", or "best".

--
Going back to real life, if shortstops happen to be (at a given moment in time) a combination of studs
and yuks, how do we value a stud SS? Value over avg? That is corrupted by other studs. Over yuk? But if
I lose my stud, I can through trades or free agency acquire some other stud or even avg SS, so you cannot
completely discount that they exist.

Hence, I conclude that meausres like RCAP and WS/WARP general replacement level and Dan R's method
of position-specific replacement level ALL contribute to general knowledge. In my (successful!) practice at these sim
games, I've found that what's left at the bottom is somewhat MORE important than the other studs,
but the overall average of play at a position IS a relevant piece of data that we ought not to discard.

John Murphy mentioned that he was going to start some other newbie threads later this week or next. He specifically mentioned Dan Quisenberry but I would hope that Chet Lemon is also a part of that second batch.

What role does his drug problems play in the HoM?

None for me, thanks.

For war adjustment, I often think of a guy like Ted Williams as the perfect example of my questions: WWII, obviously, took significant time from the prime of his career. I think it obvious you should make an adjustment. Korea, however, took years from his mid-30's. Isn't it possible/probable that such a player is just delaying the end of his career (and I use Williams because he was awfully productive to an "old" age)?

At the HOM we love our lurkers!

As noted elsewhere, Williams, for instance, was in combat, which I'd assume takes more out of your body than baseball.

But that isn't even quite the issue that "length of season" adjustments have in my mind. Much is made of (for instance) Beckley's stat adjustments to a 162 game schedule; is that appropriate? I'm guessing that playing 20-30 fewer games a season would add a couple of years of career length - if anything, it might enhance his counting stats, not suppress them.

OK, IMHO, you got to split this into two questions:
1) Is it appropriate? I think it is because our consstitution says we got to be fair to all eras, which means we must compare them endlessly. Without proration, we end up with real comparitive problems, whether we're working in countings stats or uberstats. Whether proration adds two or three years is somewhat irrelevant since if you don't then the latter-day guys have that two-three year advantage anyway.

2) How do you do? I suspect most folks just say, for example, 162/154 * WS because there's little to be gained from using a slightly more accurate formulation. But others have suggested that regression for very short seasons might be a better idea. I don't know if it is or not, but it's out there.

3) In general, if you have a robust system that measure value in numerous ways, this whole thing might not make much differences. If you are incorporating the answers to questions like "How often was he the best ...," then you'll be doing context-neutralization in many instances anyway.

As a non-voter, I don't expect a response...
I hope in so saying that you aren't implying that we aren't responsive to lurkers. If so, we'll try harder, if not,...PHEW! Lurkers are important to us, and they are frequent sources of insight, information, good cheer, support, and funnies. Go lurkers!

Arrgh!!! Stoopid server!!!! took me 4 tries to get through, and it came thru gobbedly gooked up 4 times. world record for longest post that is 80% oops.

1860s - 1 (Pearce) (SS)

1870s - 9 (Anson, Barnes, McVey, Pike, Spalding, Start, Sutton, White, Wright) (P, C-2, 1B-2, 2B, 3B, SS, CF)

1880s - 17 (Bennett, Brouthers, Caruthers, Clarkson, Connor, Ewing, Galvin, Glasscock, Gore, Hines, Keefe, Kelly, O'Rourke, Radbourn, Richardson, Stovey, Ward) (P-5, C-2, 1B-2, 2B, SS-2, LF-2, CF-2, RF)
{Candidates – Browning, C Jones, Welch, Williamson}

1890s - 16 (Burkett, Childs, Dahlen, Davis, Delahanty, Grant, Griffith, Hamilton, Jennings, Keeler, Kelley, McPhee, Nichols, Rusie, Thompson, Young) (P-4, 2B-3, SS-3, LF-3, CF, RF-2)
{Candidates – Beckley, Duffy, Van Haltren, McGraw}

1900s - 17 (M Brown, Clarke, J Collins, Crawford, Flick, R Foster, Hill, G Johnson, Lajoie, Mathewson, McGinnity, Plank, Sheckard, Waddell, Wagner, Wallace, Walsh) (P-6, 2B, 3B, SS-3, LF-2, CF, RF-2)
{Candidates – Bresnahan, Leach, Willis, Joss}

1910s - 16 (Alexander, Baker, Carey, Cobb, E Collins, Groh, Jackson, W Johnson, Lloyd, Magee, Mendez, Santop, Speaker, Torriente, Wheat, Williams) (P-4, C, 2B, 3B-2, SS, LF-2, CF-4, RF)
{Candidates – Roush, Redding, Cravath, Doyle, Taylor}

1920s - 18 (Beckwith, Charleston, Coveleski, Faber, W Foster, Frisch, Goslin, Heilmann, Hornsby, Mackey, Moore, Rixey, Rogan, Ruth, Sewell, Sisler, Vance, Wilson) (P-6, C, 1B, 2B-2, 3B-2, SS-2, LF, CF, RF-2)
{Candidates – Oms, Grimes, Mays, Schang, Traynor}

1930s - 29 (Averill, Bell, R Brown, Cochrane, Cronin, Dickey, Dihigo, Ferrell, Foxx, Gehrig, Gehringer, J Gibson, Greenberg, Grove, Hartnett, Herman, Hubbell, Lyons, Medwick, Ott, Paige, Ruffing, Simmons, Stearnes, Suttles, Terry, Vaughan, Waner, Wells) (P-8, C-4, 1B-5, 2B-2, SS-3, LF-2, CF-3, RF-2)
{Candidates – B Johnson, Dean, Bridges}

1940s – 17 (Appling, Boudreau, W Brown, DiMaggio, Doerr, Feller, Gordon, Hack, Irvin, Leonard, Mize, Musial, Newhouser, Reese, Slaughter, Trouppe, Williams) (P-2, C, 1B-2, 2B-2, 3B, SS-3, LF-2, CF-3, RF)
{Candidates –Keller, Walters, Elliott, Rizzuto, Stephens, Pesky}.

1950s – 17 (Ashburn, Banks, Berra, Campanella, Doby, Ford, Kiner, Lemon, Mantle, Mathews, Minoso, Pierce, Roberts, J Robinson, Snider, Spahn, E Wynn) (P-6, C-2, 2B, 3B, SS, LF-2, CF-4)
{Candidates – Fox}

1960s – 20 (Aaron, Boyer, Bunning, Clemente, Drysdale, Freehan, B Gibson, Kaline, Killebrew, Koufax, Marichal, Mays, McCovey, B Robinson, F Robinson, Santo, Torre, Wilhelm, B Williams, Yastrzemski) (P-6, C-2, 1B-2, 3B-3, LF-2, CF, RF-4)
{Candidates – Cepeda, Cash, Brock, E Howard}

1970s – 18 (Allen, Bench, Carew, Carlton, DEvans, Grich, RJackson, Jenkins, Morgan, Niekro, Palmer, Perry, Rose, Seaver, Schmidt, Simmons, Stargell, Sutton) (P-7, C-2, 1B-2, 2B-3, 3B-2, RF-2)
{Candidates- J Wynn, Fingers, Perez, Staub, Tiant, Singleton, Bonds, Bando, RSmith, Munson}

Before we get too excited about Keith Hernandez, Dave Parker or the other Evans, we should remember that Jimmy Wynn is about to go in and we still have Blyleven, Carter and Gossage to go for the '70s. We're going to hit 22 easily for that decade. We might want to be careful about pushing that total up to 25 or higher.

But isn't the total number of electees based on the number of teams? In that case, isn't the number of electees for the 1970s supposed to be higher?

DCW3--um, not for two guys who play the same position....

Again, Evans was only a third baseman for slightly over half of his career, while Nettles played third base in close to 90% of his games. I think that eats up most of the offensive difference between them on a career level.

Is it appropriate? I think it is because our consstitution says we got to be fair to all eras, which means we must compare them endlessly. Without proration, we end up with real comparitive problems, whether we're working in countings stats or uberstats. Whether proration adds two or three years is somewhat irrelevant since if you don't then the latter-day guys have that two-three year advantage anyway.

Because he'll likely get elected soon, let's use Beckley as an example. He's 44th in career PA, so let's ask two questions:

1. With 665 PA/yr (which is what Craig Biggio has averaged, excluding '94-95) instead of the 571 he averaged in the meat of his career, he adds 1598 PA to his total. Is it fair to assume someone would have played the equivalent of almost 3 seasons (on top of a historically long career, mind you) "only if"? Adding those PAs put him over 12000 for his career, ahead of Palmeiro; it would put him in 9th all-time in "non-DH" ABs (and 6 of those ahead of him played in the "greenie" era). Again, is this fair to assume?

2. Since he isn't playing as much every year, isn't fair to assume he is less likely to play tired, and thus be able to perform closer to his peak ability more of the time?

Please be clear (especially those of you who think Beckley's been shafted so far) I'm using him only as an example - I don't know enough about him, or his era, to argue for or against him.

As a non-voter, I don't expect a response...
I hope in so saying that you aren't implying that we aren't responsive to lurkers.

Not at all (I'm not 100% sure who are and aren't the voters at just a glance); I just figured someone would say "We answered this 80 years ago; go to that thread". I didn't assume rudeness, but I'm not part of your "group" in here, so I expected to just be pointed in the direction of the answers while you guys discussed the players.

Before we get too excited about Keith Hernandez, Dave Parker or the other Evans, we should remember that Jimmy Wynn is about to go in and we still have Blyleven, Carter and Gossage to go for the '70s. We're going to hit 22 easily for that decade. We might want to be careful about pushing that total up to 25 or higher.

First, I'm not sure how this list was generated, but I have some disagreements here. First, other than 1972, Dick Allen is in the HOM based on what he did in the 60's, not 70's. Second, both Hernandez and the other Evans had the majority of the best years in the 80's, not the 70s.

Finally, I always thought the reason for expanding the number of elected positions in the last few drafts was to account for the expaning talent pool (and expanding number of teams) in the 70's and beyond. IMO, the number of elected players per team should be relatively consistent across time.

Whale of a post there, TomH. :) Slight clarifications:

1. Win Shares uses a replacement level? News to me...

2. BP WARP's replacement level, which is the 1899 Cleveland Spiders, introduces two further distortions. First, by setting rep level far below its actual empirical value, it grossly distorts the proper ratio of players' performances (eg if Guy A generates 100 runs, Guy B 80, and rep level is 60, Guy A is twice as valuable as Guy B. But BP, by setting rep level at 30, will say that Guy A is just 40% more valuable than Guy B). Second, by seeing how many wins a player would add to the 1899 Cleveland Spiders rather than to an average team, it introduces OBP and Pythagorean effects that cause some teams to wind up with more WARP than their actual win total (I believe, or damn close to it at any rate).

I'd certainly be interested to hear why average (either vs. position or league) as opposed to replacement (however defined) matters for valuation purposes...is it something you can express mathematically?

Re # of electees by decade: 23-25 for the 1970s wouldn't be too many, because the pool is larger. Hernandez and Carter had more value in the 1980s than the 1970s, anyway.

2. Since he isn't playing as much every year, isn't fair to assume he is less likely to play tired, and thus be able to perform closer to his peak ability more of the time?

Before the advent of the 154-game season, teams tended to play a number of non-league, "exhibition" games during and after the regular season, so players weren't actually playing fewer games: there wre just fewer regular-season games.

I doubt the 8-game difference between the 154-game schedule and the 162 game schedule makes a notable difference in terms of wear-and-tear, and the improvements in medical care, training, and nutrition surely more than offset the longer schedules in the modern game in terms of enabling players to be at peak performance level on a game-by-game basis. Great modern players are more likely to have longer careers than great players of the nineteenth century.

John Murphy mentioned that he was going to start some other newbie threads later this week or next. He specifically mentioned Dan Quisenberry but I would hope that Chet Lemon is also a part of that second batch.

Lemon, Quiz, Buckner and Boone to be exact, Chris.

Do we really think Big B's going to make it in 2005? Wonderful if so, but having had him top of my ballot most years since 1930 I remain skeptical.

The extra games vs conditioning question can also be looked at the other way, as a question of how many standard deviations above the mean was his career length. Although Beckley played fewer games than Palmeiro and only just beats him in hits when you scale up (by a bit more than 162/154 because they still weren't playing 154 early in his career) his career length is exceptional for the time -- second most hits after Cap Anson when he retired. If we're looking at standard deviation on OPS+, we should surely look at it on career length also. 20 years was a LOT back then.

Judging by what we know of players' habits back then, you also have to scale up for elimination of hangovers (except against a minority of 1980s players.) Even the wilder modern players are normally seeing 1 ball not 2 when they swing!

Before we get too excited about Keith Hernandez, Dave Parker or the other Evans, we should remember that Jimmy Wynn is about to go in and we still have Blyleven, Carter and Gossage to go for the '70s. We're going to hit 22 easily for that decade. We might want to be careful about pushing that total up to 25 or higher.

First, I'm not sure how this list was generated, but I have some disagreements here. First, other than 1972, Dick Allen is in the HOM based on what he did in the 60's, not 70's. Second, both Hernandez and the other Evans had the majority of the best years in the 80's, not the 70s.

I agree with that. I think I also have Wynn as a 60s guy, although that's very close, and I have Schmidt as the (first!) 80s guy. But on the other hand, you left out Nolan Ryan, and possibly Carlton Fisk.

Finally, I always thought the reason for expanding the number of elected positions in the last few drafts was to account for the expaning talent pool (and expanding number of teams) in the 70's and beyond. IMO, the number of elected players per team should be relatively consistent across time.

I think that's true to an extent, but I wouldn't want to go overboard with it. If the 70s come out a little bigger than the 60s, it's not a big deal, but I'd be surprised if it's a lot more than that.

First, I'm not sure how this list was generated, but I have some disagreements here

The list was originally created and maintained by former voter Brent. I've been updating it for the last year or so since Brent stopped voting on a regular basis. And it's a pretty simple, sometimes blunt, tool. I look at a player's statistics and assign him to one decade or another. It's certainly not precise, but it's not supposed to be. It's meant to be a helpful shorthand. And as such, I appreciate corrections.

First, other than 1972, Dick Allen is in the HOM based on what he did in the 60's, not 70's.

You're right. I don't know why I assigned Allen to the '70s when I did (maybe it was that MVP year in '72), but it's clear that more of the meat of his career was in the '60s. I'll move him over before I post the list again. Actually, Allen's case shows a weakness in the decade listing. He's more of a '65-'75 player than either a '60s or a '70s one. In fact, that might be an interesting chart to look at.

Second, both Hernandez and the other Evans had the majority of the best years in the 80's, not the 70s.

You're right again. I was looking at the big Brett/Ryan/Yount class in '99 as the first of the '80s players but after reading your post, I took a better look at Hernandez and Dwight. They would be the first of the '80s players.

Finally, I always thought the reason for expanding the number of elected positions in the last few drafts was to account for the expaning talent pool (and expanding number of teams) in the 70's and beyond. IMO, the number of elected players per team should be relatively consistent across time.

I agree with you. I said something similar to that when I posted this list in the '95 discussion. I don't think it's a problem that we've stepped up from 16-18 players in most decades to 20 in the '60s and 22 in the '70s. The leagues were expanding. Some would argue that the talent pool was also expanding (although our elections of Negro League players offset both of those to some degree). With more teams and more players to choose from, we shouldn't be surprised if we have more honorees as well. We should even expect it. But I still think that we should be careful that our jumps aren't too big or disproportionate. And that's all that my warning was about.

Hernandez and Carter had more value in the 1980s than the 1970s, anyway.

I had looked at Carter and Fisk previously and placed one in the '70s and one in the '80s. I was working somewhat from memory when I made my earlier comment so it's likely I had them reversed.

I'd certainly be interested to hear why average (either vs. position or league) as opposed to replacement (however defined) matters for valuation purposes...is it something you can express mathematically?
As much as I'd like to, DanR, it seems to me to be one of those "you can prove anything depending on your assumptions" items; you know, how some assessments are 'robust', in that they hold true regardless of how underlign distributions change, etc. Unfortunately, you can Easily cook the ##s differently on this 'avg vs replacement' discussion. But I'll try sometime this weekend to put something numerical together.

As Bruce keiden used to say, loose-leaf thoughts from a tight-leaf notebook.

Keith Hernandez-I have a real hard time with the Mattingly/Hernandez defense question. I mean my gut tells me that Hernandez was special in ways that don't come along very often. Mattingly was my boyhood hero, and he could really turn the 3-6-3 like nobodies business. It's a shame that Mattinly didn't play in the NL where we could have seen his arm used more often in different situations. That said, both guys were also presences on the field, the sort that exuded confidence and didn't let you down. I feel blessed having watched both of them growing up.

Chet Lemon-The Mike Cameron of his day.

Fred Lynn-The Jim Edmonds of his day, except for the part about ending up as near-HOM/HOFer. Instead just an All-Star and more fragile than Brittany's ego. That one's for you, yest... ; )

Rick Reuschel-Who would join him on the Not-So-Athletic-Looking All-Star team? Thurman Munson? Stanky? Hack Wilson? John Olerud? Greg Luzinski? Cecil Fielder, naturally. Fernando. Charlie Hough?

Frank White-If he could only either walked or hit .300, he'd have a damn-close to HOMable career. Gee, I guess you could say that about almost anyone.

Bill Buckner-If only Mookie or Stanley were also due up this year.... An amazingly long career for a guy who didn't walk and couldn't hit for power.

Bob Boone-I want to write silly stuff about how the guy thought he was a genius manager, but that's not really fair. In the second half of his career, he kept himself on a vicious workout regimen and stayed in the game far longer than anyone would have figured. At the time of his retirement, he was talked about as though he were a kind of second manager, an on field leader whose veteran presence was trusted. Why wouldn't he think that would translated well to managing? Well, he wasn't a very good manager, but his legacy would seem much less awful if the Royals had turned it around by now. Then he'd be a faded memory or be remembered as the last guy before they righted the ship. Instead, he reminds everyone that the Royals have been down a long time, and he's one in a disastrous litany of lunacy that's pushed the team further and further down in the standings. Each of their managers has put his own nutty stamp on things, and Boone's a one.

Claudell Washington-I really like him when I was 11, growing up in NY. I remember when the Angels dealt him to the Yanks for an aging Winfield, and someone said, "Great deal if he's the Caludell we used to have." He wasn't, but Winfield was still the Winfield of old. Yuck. I had his baseball card in 1982, and I had no idea how to pronounce his name.

Dan Quisenberry-RIP.

Johnny Ray-Here's the thing about Johnny Ray: terrible percentage player. Hit 30 points over the league, but his OBP was league average because he didn't walk. His SLG was also league average, meaning he had no power. But a 101 OPS+ career somehow. Meanwhile, he steals 80 bases, but at a 62% clip. Ack!

John Tudor-Most feared lefty in the NL of 1985? I think so. I remember how he dominated the league that year. In 1990, he went 12-4 with a 2.40 ERA and hung it. Was he injured? Either way, this has rank among the best adieu seasons of any pitcher.

Bob Knepper-three words: Garbage Pail Kids.

Dave Collins-He was a bad idea in 1983.

Phil Bradley-Did he have a reputation for surliness or uncoachableness? I seem to have a hazy recollection of it, but I don't really know for sure.

Gary Ward-Here's the perfect guy to talk about boxscores. I never think of this guy by his actual name, I think of him as GWard (guh-ward), due to my having seen his name thusly abbreviated a hundred times in box scores. That's why I also call Roberto Alomar "RAlomar." Or Brady Anderson "ByAnderson." Or one of the Evanses "DEvans" (pronounced DeeEvans).

Greg Minton-Retrobermanism: Franklin Minton.

Ron Oester-He makes really great appliances, much better than Ham Beach.

Jeff Leonard-One flap down, of course.

Mike Scott-I put my Mike Scott 1982 Topps card into my elementary school's time capsule...He looks like someone famous, maybe Philip Seymour Hoffman or something....Anyway, it's pretty close between him, Gooden, and Hershiser about who the best NL pitcher of the mid-late 1980s is. We've talked some already about Who is the pitcher of the 1980s? Well, Scott's a muddling influence in this question. He's the Dave Stewart of the NL in some ways, and in general, the 1980s seemed to be filled with reclamation projects (Scott and Stewart plus Reuschel and Tanana), late bloomers (Tudor), fizzles (Gooden, Danny Jackson, Soto, Hurst), misunderstood stars (Stieb, Blyleven), and long-career innings eaters with occasional flashes of brilliance (Morris, Tanana, Sutcliffe, Candelaria, Darwin). Not that every decade doesn't have these types of players, but this decade had a lot of muddle, but without the benefit of having a group of stand-outs who were obviously the tops in their generation and who in particular who spanned the entire period. Essentially Stieb and Blyleven are it if you're looking for long-spanning, high quality hurlers. And even Blyleven's credentials there are a little mixed due to injury. See what a mess selective endpoints can make!!???!!! ; )

The list was originally created and maintained by former voter Brent. I've been updating it for the last year or so since Brent stopped voting on a regular basis.

As Mark Twain supposedly said, "The reports of my death are greatly exaggerated."

(I've been waiting for an opportunity to say that ;-)

I've continued to vote regularly since the 1930s, though judging from the groundswell of support for my best friend, Phil Rizzuto, it appears that no one actually reads my ballots. To give credit where it's due, it was actually Devin McCullen who first put together such a list of HoMers by decade. I liked the idea and updated it a few times, though (like Chris Fluit) every time I posted it there would be a bunch of criticism. So I quit updating it about a year ago and Chris picked it up. I still like it -- it's nothing sophisticated, but it does give a quick snapshot of how we're doing on being fair to all eras. For example, for a long time we were quite concerned about the 1890s being underrepresented, but with the inductions of Childs, Griffith, and Jennings I see that it's now in line with the surrounding decades.

I hope you keep it up, Chris.

Who is the pitcher of the 1980s? Well, Scott's a muddling influence in this question.

Don't forget that at the peak of his career, Scott was pitching in a park with a park factor in the 95-98 range. Here are my RA+ PythPat equivalent year-by-year records for the six-year span 1985-1990: 13-12, 22-9, 17-11, 14-10, 14-11, 10-13. Sure, there's one monster year in there, but to be honest, you've seen better peaks. Everything outside this six-year span is essentially worthless. And even that one monster year, the equivalent 22-9? If you move it back a year, to 1985, it matches Tudor, not Gooden. Gooden's 1985 is on another planet: an equivalent 25-6.

Here's another pitcher's RA+ equivalent record over the exact same 6 years I counted for Scott: 7-4, 20-8, 22-9, 19-10, 17-11, 19-6. Think about that for a moment and you might identify who I'm talking about.

Here's another pitcher's RA+ equivalent record over the exact same 6 years I counted for Scott: 7-4, 20-8, 22-9, 19-10, 17-11, 19-6. Think about that for a moment and you might identify who I'm talking about.

I'm going with Clemens. How scary is it that you could make a legitimate claim that the 80s were his third-best decade?

Correctly identified. That was Clemens 1985-1990. (He was an equivalent 8-7 as a rookie in 1984, and he didn't exactly stop having years like that after 1990.)

Okay, since Jimmy Wynn seems likely to get elected soon, I have a question: why is Wynn going to go into the Hall of Merit, while Cesar Cedeno fails to show up on a single ballot? They look very, very similar to me.

John Tudor-Most feared lefty in the NL of 1985? I think so. I remember how he dominated the league that year. In 1990, he went 12-4 with a 2.40 ERA and hung it. Was he injured? Either way, this has rank among the best adieu seasons of any pitcher.

In 1985, Tudor pitched 10 shutouts. Ten. And he had an 11th game in which he pitched more than 9 innings without ever allowing a run.

Yeah, he was injured; the usual (arm injury). It cost him nearly the entire 1989 season, and pretty much nullified what the Dodgers got out of trading Pedro Guerrero for him. Even though he was very effective in 1990 (I have it as an equivalent 11-5), I think he was still in pain.

In Tudor's case, those Crash Davis to Nuke LaLoosh lessons never took: Tudor never learned how to speak cliche. I don't really know whether he was too smart, or whether he just couldn't master the tone of voice. As a fundamentally shy guy, I think he regarded the public spotlight and the questions of reporters as an ordeal - and that he wasn't that upset about walking away.

Tudor goes on the all-royalty pitching staff along with Silver King, Ray King, Mel Queen, and Pedro Borbon. The A's have a guy whom I saw pitch in NCAA ball and who might come up this year: Jason Windsor.

And, for that matter, Wynn doesn't seem notably better than Fred Lynn, and Lynn probably won't get in either.

Tudor goes on the all-royalty pitching staff along with Silver King, Ray King, Mel Queen, and Pedro Borbon. The A's have a guy whom I saw pitch in NCAA ball and who might come up this year: Jason Windsor.

That's awesome! But where's that Hohenzollern reliever I seem to remember from my youth????

By the by, could Ray King be thought of as King King? I don't know that Raymond comes from the same root as Rex or Rey, but it would be kind of funny if so.

I like Wynn's peak a godo bit better than both of them. Just because their end of career stats look the same doesn't mean they were nearly equal players.

Wynn: 36,32,32,31,28,28,27
Lynn: 34,33,27,22,22,21,19
Cedeno: 33,30,30,27,26,23,20

I know that WS has its flaws and that it isn't the end all be all, but this is a prety drastic seperation here. The only time that Wynn loses to either Lynn or Cedeno is when his second best year is compared to Lynn's second best year. He beats Lynn in yeras 1-3 then destroys him in year 4-7. Out of the 21 seasons here Wynn has the 1st, 4th, 5th, 6th, 9th, and tied for 10th best seasons here.

On Wynn vs. Cedeno and Lynn: BP's WARP shows much the same thing. Wynn's career value and his career rates are not all that much ahead of Cedeno's and Lynn's, but on a season-by-season basis, he was consistently a more valuable player. On those occasions when Cedeno and Lynn put togher full seasons in full health, yes, they were highly comparable to Wynn. But they put together full seasons in full health at their best quality notably fewer times than Wynn did.

Sorry, Eric - I couldn't find any Hohenzollern. Or Hapsburg, or Romanov (although there is a Romano or two). Any more royal houses to look for?

1. Win Shares uses a replacement level? News to me...

Dan, I thought you had been around for all of those discussions?!

A team of zero win share players (TRUE zero win shares, NOT negative win shares zeroed out to zero) would not win zero games, but would win around 32 games in a 162 game season. This has to do with the way James calculates Win Shares based on MARGINAL Wins. There have been some good explanations of this in various places on the internet.

but would win around 32 games in a 162 game season

Or, curiously, more wins than the Spiders would have had in a 162 sked.

OCF,

How about Ming? Han, Qin, Quing, Jin, Song, Yuan, Sui, Zhou? ; )

OCF Posted: March 07, 2007 at 12:08 PM (#2308185)
Sorry, Eric - I couldn't find any Hohenzollern. Or Hapsburg, or Romanov (although there is a Romano or two). Any more royal houses to look for?

Stuarts. Stu(art) Miller, (Dave Stewart)

Bush?

Honesly, though wasn't there a Korean pitcher named Seung Song? He was mentioned in Moneyball I believe.

How about Jason Windsor? He pitched 4 games for Oakland last year.

On Wynn vs. Cedeno and Lynn: BP's WARP shows much the same thing. Wynn's career value and his career rates are not all that much ahead of Cedeno's and Lynn's, but on a season-by-season basis, he was consistently a more valuable player.

Yes, WARP does show that...sort of. But <U>why</U> does it show that? Here are the best seasons for each player by WARP, seasons that WARP sees as essentially identical: each has a WARP3 between 11.0 and 11.1, and a WARP3 between 11.0 and 11.5:

Wynn, 1965: 663 PA, .275/.371/.470, 143 OPS+, 43 SB/4 CS, 6 FRAA
Cedeno, 1972: 625 PA, .320/.385/.537, 162 OPS+, 55 SB/21 CS, 6 FRAA
Lynn, 1979: 622 PA, .333/.423/.637, 179 OPS+, 2 SB/2 CS, 6 FRAA

I guess Wynn's baserunning must be giving him a big boost. But those numbers don't even match up to others on BPro's own site. Since all three guys played the same position, and were equivalent fielders by FRAA, we can look at VORP:

Wynn: 53.5
Cedeno: 64.5
Lynn: 82.8

Even after factoring the different offensive levels of those three seasons, that would work out (by my calculations) to around 5.8 WARP for Wynn, 7.2 for Cedeno (in a 153-game season), and 7.7 for Lynn. Is anyone else troubled by this?

and a WARP3 between 11.0 and 11.5:

That should say a <U>WARP1</U> between 11.0 and 11.5...

A question about BP FRAA--is one BP FRAA always 1/9 of a win (as one BP BRAA is, due to the 4.5 R/G run environment)? If so, wouldn't the "raw" FRAA totals be higher in the 1890s and lower in the 1910's?

elect Van Haltren before Roush!

For those of you somewhat newer to our group, you may not know that George Van Haltren was once thought of as a sure fire HoM-to-be, while Roush lagged in the backlog. Lately they have swapped places.

I think this is a mistake.

Quick review:
Career - this is easy, ain't it? In WS, GVH leads 344 to 314. In WARP3, GVH leads 78.9 to 74.0. You could give a little credit to Roush for time unfairly missed, but you'd also have to lengthen Van Haltren's shorter league schedule.

Prime/most of career - ages 22 to 36, both of their best 15 yrs, GVH leads both metrics again. 325 to 303, 76.8 to 72.6.

Peak - By WS (unadjsuted for schedule length), Roush has the best 4 seasons. GVH beats him in all yrs 5 to forever. By 3 times as much. Is peak worth more than 3 times the 'rest of career'?? WARP shows less of a peak advantage for Roush.

League strength? That won't help Roush's case any. GVH played in the one-league MLB for quite a while.

It's easy to look at GVH's career and be unimpressed with the numbers, partly because he pitched for a few years, curbing his career batting totals. But he got on base a lot, was durable, stole a bunch of bags which was awfully important in the game of 1890, and thus he scored a pile of runs; at least 109 runs every year 11 years in a row when he wasn't pitching. Roush did cool things like not strike out and hit .350. And he scored 95 runs in a season. Once. (Yes, different game, but check the leaderboards)

Van Haltren didn't gain more career stats because he played LONGER. He entered the majors a year earlier than George, and exited a year later.

Either WS and WARP are BOTH wrong. Or we're headed down the wrong path.

Doh.
Van Haltren didn't gain more career stats because he played LONGER. <u>ROUSH</u> entered the majors a year earlier than George, and exited a year later.

DCW3, I have Wynn's 1965 and Lynn's 1979 of virtually equal offensive vlaue as well. Here's the math:

1965 Wynn: 108.8 runs with a 93 PF = 117.0 runs in 411 outs. The league scored .155 runs per out, and the average team had 4,238 batting outs, so an average team with Wynn would score 117 + (.155*(4238-411)) = 593 for a total of 710 runs. The average team scored 656 runs, and 710 RS/656 RA in a 162 game season is 86.8 wins.

1979 Lynn: 132.9 runs with a 106 PF = 125.4 runs in 358 outs. The league scored .183 runs per out, and the average team had 4,116 batting outs, so an average team with Lynn would score 125 + (.183*(4116-358) = 688 for a total of 813 runs. The average team scored 752 runs, and 813 RS/752 RA in a 162 game season is 86.9 wins.

Not sure why OPS+ is getting this one so wrong, although playing time (Wynn had 40 more PA), stolen bases and park factors (are the same ones being used?) all play a part.

Are you taking into account the fact that Wynn's league had pitchers hitting? If you plugged him into an average lineup, he wouldn't be taking away the outs made by pitchers.

And how does Cedeno's '72 compare?

DCW3, good point. The way I actually deal with that issue in my system is just by comparing them to replacement, since a replacement player will have worse performance vs. average in the Al than in the NL as well. Note that BP WARP1 does NOT account for the DH.

To do this properly, then (without counting fielding, and using the long-term average for replacement level rather than the specific year's level):

1965 Wynn: 5.8 offensive wins above average in 658 PA. Replacement NL CF: 1.4 wins below average per 676 PA. 5.8 + 1.4*(658/676) = 7.2 WARP.
1972 Cedeño: 6.0 offensive wins above average in 624 PA (straight-line-adjusted for season length). Replacement NL CF: 1.4 wins below average per 648 PA. 6.0 + 1.4*(624/648) = 7.3 WARP.
1979 Lynn: 5.9 offensive wins above average in 622 PA. Replacement AL CF: 1.8 wins below average per 676 PA. 5.9 + 1.8*(622/676) = 7.5 WARP.

OK, so Lynn's the best of the three, but all of them are well within the margin of error of any reasonable system. They were equally valuable seasons.

1965 Wynn: 108.8 runs with a 93 PF = 117.0 runs in 411 outs.
1979 Lynn: 132.9 runs with a 106 PF = 125.4 runs in 358 outs.

Dan, where are you getting your numbers from? In 1965, Wynn had 98 runs created, and was responsible for 435 outs. In 1979, Lynn had 143 runs created, and used 370 outs. I see where your out totals from, but where did you get your runs totals from?

Now you *really* want to get into the nitty-gritty of it, don't you? You never give me an inch of breathing room, Dizzypaco. (That's a good thing).

I use eXtrapolated Runs, except that I allow the out value to float so that each league-season's XR matches its actual runs scored. I don't know what RC formula baseball-reference is using, but I'll bet you my life savings it's not as accurate as this is.

NetDP is the number of double plays the player hit into above or below league average given his opportunities. EqBR is James Click's Equivalent Non-SB Baserunning runs, which are available for 1972 to the present.

1965 Wynn: 96 singles*.5 + 30 doubles*.72 + 7 triples*1.04 + 22 home runs*1.44 + 81 UIBB*.33 + 5 HBP*.33 + 3 IBB*.25 + 43 SB*.18 - 4 CS*.32 + 5 SF*.37 +6.79 NetDP*.37 - .098*409 AB-H = 108.8 XR.

1979 Lynn: 95 singles*.5 + 42 doubles*.72 + 1 triple*1.04 + 39 home runs*1.44 + 78 UIBB*.33 + 4 HBP*.33 + 4 IBB*.25 + 2 SB*.18 - 2 CS*.32 + 5 SF*.37 + 2.81 NetDP*.37 + 2.5 EqBR - .100*354 AB-H = 132.9 XR.

I don't know what RC formula baseball-reference is using, but I'll bet you my life savings it's not as accurate as this is.

Its good that you are are convinced of the accuracy of the system you use, but it is not at all settled that it is more accurate than runs created, the formula devised by Bill James that is used in win share calculations.

And its not nit-picking. In this case, the difference is huge. Using runs created, Lynn's unadjusted advantage over Wynn is huge, easily enough to overcome differences in park and league.

Thanks for presenting your numbers, Dan. But I still think that people who use stats like WARP to evaluate Wynn should realize just how different his scores on those numbers in his best seasons are from measures like OPS+ or even semi-uberstats like VORP before taking any of those stats at face value. Wynn was a great basestealer in 1965, but it's not like he was Maury Wills or Vince Coleman out there. Lynn's '79 beats Wynn's '65 by 33 points of OPS+ and 219 points of raw OPS. That's a ridiculous error range. Cedeno's '72 is even more comparable, since he played in the same ballpark as Wynn, in a very similar offensive environment. He was significantly better offensively than Wynn in every way except basestealing--and it's not like Cedeno was a zero on the basepaths himself.

How about Bobby Murcer's 1971? In 624 PAs, he hit .331/.427/.543 for a 181 OPS+--the highest OPS+ any CF has had in the last forty years. He stole 14 bases and was caught 8 times. His AL leading VORP was 70.4, which works out to about 7.9 WARP. Yet Wynn's 1965 season still rates as significantly better than Murcer's '71 on WARP1 (11.4 to 10.7) and far better on WARP3 (11.0 to 9.6). Now, FRAA does rank Wynn as a much better defender (+6 to -12), but you'd think that that would, at best, push Murcer down to Wynn's level, not put Wynn well ahead.

Well, I'll do it with BaseRuns too, then. BaseRuns gives 108.4 for Wynn, and 133.4 for Lynn (counting the EqBR). I'm happy to send you the spreadsheet if you want.

And how about BP's EQR estimator, which is also terrifically accurate? It has 109 for Wynn, and 132.5 for Lynn (counting the EqBR). (check their BP pages)

So we have XR, BaseRuns, and EQR all in agreement to within literally one run. By contrast, we have baseball-reference's RC--which, according to http://www.baseball-reference.com/about/bat_glossary.shtml, is the most basic version, just(H+BB)*TB/PA--which is on another planet.

I won't vote for Lynn or Cedeño, and Wynn missed my ballot this year, so I don't know why I'm spending so much time on this, but attempting to cast doubt on the most accurate run estimation systems we have--all of which are in perfect agreement--just muddies the waters unnecessarily. With all the substantive issues we have left to debate--replacement levels, fielding stats, quality of play, career vs. peak, you name it--the *last* thing we need to do is make run estimation seem less precise than it is. It's sort of similar to the people who "caution" that evolution is "just an unproven theory." Sure, but it's a theory that meets the most rigorous standards science has to offer. So do XR, EQR, and BaseRuns--and Simple RC (the one you see on the baseball-reference pages) doesn't.

I'm not going to quibble with the accuracy of those run estimators--but it seems like the fact that Lynn created about 25 more runs while using about 55 fewer outs ought to easily overwhelm any park/era differences. (I plugged the numbers for each year into one of Tango Tiger's Linear Weights formulas--I don't know how much stock to put in it, since I don't know how accurate it is for different eras, but Wynn comes out at 39 runs above average [excluding pitchers] and Lynn at +58. Both numbers are park-adjusted.)

It's funny - there was a time when I felt like I was one of the guys arguing against George Van Haltren, and now he may be in my top 5. I have Roush in my top 15 as well, but definitely behind GVH. But I wanted to remind everybody that Van Haltren gets an excessive boost in his WS for his pitching, which was no more than league average at best.

Here are my numbers, which come from a different raw source - the RC calculations of a Stats Handbook.

Wynn, 1965: 104 RC, at a rate of 6.51 per 27 outs in a 4.03 league, but with a 90 park factor (!) Just dividing the RC by the rate, 104/6.51 = 15.98 or about 431 outs. Taking into account the park factor, that's about 46 RCAA. In that scoring environment, a win is worth about 8.2 runs; dividing 46 by 8.2 give 5.6, or 56 as I report it in my tables.

Lynn, 1979: 134 RC. at a rate of 9.71 per 27 outs, in a 4.67 league with a 104 park factor. 134/9.71 = 13.80 or about 373 outs. That gives me about 67 RCAA. In that scoring environment, a win is worth about 10.1 runs; dividing 67 by 10.1 gives 6.6, or 66 as I report it in my tables.

But that's by far Lynn's best offensive year, at 66. I have him at 55 in 1975, 35 in 1978, 33 in 1984, 30 in 1976, and a bunch of years in the 20's.

For Wynn, I have 1965 as his 4th best year in this system. His top year is his 76 in 1969, followed by 58 in 1974, 57 in 1968, and then that 56 in 1965. There's also a 52 in 1972, a 39 in 1967 and a 34 in 1975.

DCW3--Let's be clear. Are you suggesting that Lynn's offense in '79 was, in fact, meaningfully more valuable (more than 0.5 wins better) than Wynn's in '65? If so, how? Is there an error in the math in posts # 85/88/90? The run estimation is clearly right, since EQR and BaseRuns confirm it precisely. I certainly can't see any mistakes in the theoretical team or Pythagorean calculations. Do you have different park factors than 93 and 106? If so, can you tell me why they're better than baseball-reference's? I'm happy to be convinced.

Otherwise, why does it matter what Lynn's raw OPS, or OPS+, or VORP was? OPS+ and VORP (which is based on the MlVR formula, which uses BA/OBP/SLG) are simply not as accurate measures of offensive value as the best run estimators available to us, as measured by RMSE or any other such check. If the best tools we have say those years were offensively equal, I don't see why we'd cast unnecessary doubt on them by comparing them to inferior ones--it's like saying there must be something wrong with UZR when it disagrees with fielding percentage (albeit much less extreme). I don't mean to sound condescending, nor to toot my own horn; I didn't come up with the run estimators. But unless you're saying that what they're telling us is wrong and OPS or OPS+ or VORP are right--and if so, please do tell me how!--I'm not sure what the point in bringing them into the discussion is.

There are TONS of problems with BP WARP, which I frequently criticize at great length, but their raw BRAA numbers are pretty damn accurate. Again, let's focus on what we DO need to correct for with BP--first and foremost its replacement level, and subsequently the black-boxness of FRAA and the inscrutable adjustments of WARP2 and WARP3--before undermining the one thing it does well (offensive value).

As for Murcer in '71:

116.5 XR in 357 outs. 94 park factor brings him to 124.0. An average 1971 AL team scored 623 runs in 4,145 outs for .15 XR/Out, so a theoretical team with Murcer would score 124 + ((4145-357)*.15) = 692 runs. 692 RS and 623 RA in 162 games is 88.9 wins. Replacement at CF is 1.4 wins below average per 667 PA, Murcer had 623 PA, so 88.9 - 81 + (1.4*623/667) = 9.2 WARP before counting fielding, which is an absolutely tremendous season. At 8.68 runs a win in the 1971 AL, Murcer would have had to be 16-17 fielding runs worse than Wynn in 1965 for their seasons to be equally valuable.

I use eXtrapolated Runs, except that I allow the out value to float so that each league-season's XR matches its actual runs scored. I don't know what RC formula baseball-reference is using, but I'll bet you my life savings it's not as accurate as this is.

NetDP is the number of double plays the player hit into above or below league average given his opportunities. EqBR is James Click's Equivalent Non-SB Baserunning runs, which are available for 1972 to the present.

1965 Wynn: 96 singles*.5 + 30 doubles*.72 + 7 triples*1.04 + 22 home runs*1.44 + 81 UIBB*.33 + 5 HBP*.33 + 3 IBB*.25 + 43 SB*.18 - 4 CS*.32 + 5 SF*.37 +6.79 NetDP*.37 - .098*409 AB-H = 108.8 XR.

1979 Lynn: 95 singles*.5 + 42 doubles*.72 + 1 triple*1.04 + 39 home runs*1.44 + 78 UIBB*.33 + 4 HBP*.33 + 4 IBB*.25 + 2 SB*.18 - 2 CS*.32 + 5 SF*.37 + 2.81 NetDP*.37 + 2.5 EqBR - .100*354 AB-H = 132.9 XR.

Okay, this is the last work I'll be doing on this, but here's what I tried: according to Baseball-Reference, the park-adjusted league-average line for Houston in 1965 (excluding pitchers) was .250/.312/.378. For Boston in 1979, it was .277/.342/.419. Using these numbers, I estimated that an average hitter playing for the Astros in 1965 and using up 409 outs (just defined as AB-H) would have the following line:

95 1B, 25 2B, 3 3B, 13 HR, 49 BB+HBP, 7 SB, 4 CS

While the average hitter playing for the Red Sox in 1979 and using up 354 outs would have this line:

96 1B, 23 2B, 4 3B, 13 HR, 48 BB+HBP, 9 SB, 5 CS

I plugged these numbers into the same XR formula Dan used, with a few minor simplifications--I ignored DP, SF and EqBR, and treated IBB and UIBB the same. Wynn comes out with 104.3 runs compared to 63.4 for the 1965 league-average batter. Lynn comes out with 122.6 runs compared to 62.4 for the 1979 league-average batter. So Wynn is at 40.9 runs above average, while Lynn is at 60.2 (quite close to those Linear Weights numbers). That would work out to 4.5 wins above average for Wynn, versus 5.67 for Lynn.

And yes, that's compared to average, rather than replacement level, but Wynn had only 41 more PAs than Lynn. The difference between average and replacement level over such a small sample would be on the order of one run.

DCW3--well, the era difference is just accounted for by putting in the league average RS per team and using the Pythagorean theorem as you can see; unless you've got a better method to convert runs into wins than the Pythagenport equation (and again, if you do, I'm all ears!), I'm afraid the numbers speak for themselves. As for park, that just depends on what park factors you use! I'm using baseball-reference's, which have a 14% difference between the Astrodome and Fenway--that's an extremely large gap. The combination of era and those park factors, along with the fact that Wynn had more PA, is just enough to make the two seasons roughly even, mathematically speaking (Lynn is still 0.3 wins ahead). If you have reason to believe that the gap between the Astrodome and Fenway was smaller than 14%, that would increase Lynn's advantage; I'd like to hear why.

OCF--I have Lynn as 0.3 wins ahead of Wynn. You have Wynn at 116 park-adjusted RC (I have him at 117), and Lynn at 129 (I have him at 125), so there's another 0.5 wins. Not sure where the last 0.2 are coming from.

Why do you use "theoretical" rather than real outs totals? Your outs numbers are about 20 too high. And what accuracy advantages, if any, do the Stats Handbook's RC numbers have over XR, EQR, or BaseRuns?