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

Saturday, September 29, 2007

2006 Ballot Discussion

2006 (Oct 22)—elect 3
WS W3 Rookie Name-Pos (Died)

331 104.2 1986 Will Clark-1B
243 85.5 1989 Albert Belle-LF
210 87.0 1984 Orel Hershiser-P
187 79.5 1984 Dwight Gooden-P
147 67.0 1985 Rick Aguilera-RP
146 65.8 1987 Doug Jones-RP
148 57.4 1985 Ozzie Guillen-SS
162 47.0 1989 Gregg Jefferies-1B/LF
155 48.7 1988 Lance Johnson-CF
127 58.5 1989 John Wetteland-RP
132 52.5 1988 Tim Belcher-P
145 46.0 1987 Mike Stanley-C
137 44.3 1989 Roberto Kelly-CF
115 53.5 1988 Todd Stottlemyre-P*
110 55.1 1990 Alex Fernandez-P
123 43.0 1988 Walt Weiss-SS
119 40.9 1991 Mickey Morandini-2B
124 38.1 1987 Luis Polonia-LF
110 42.7 1989 Ramon Martinez-P*

Players Passing Away in 2005

HoMers
Age Elected

None

Candidates
Age Eligible

103 1952 Ted Radcliffe-C
97 1953 Al Lopez-C
89 1957 Mickey Owen-C
87——Bob Broeg-Sportswriter
84 1963 Bob Kennedy-RF/3B
83——Chuck Thompson-Broadcaster
79 1963 Gene Mauch-2B/Mgr
78 1971 Vic Power-1B
77 1965 Chico Carrasquel-SS
77——Harry Dalton-GM
73 1972 Don Blasingame-2B
70 1976 Earl Wilson-P
70 1978 Donn Clendenon-1B
67 1975 Dick Radatz-RP
61 1984 Nelson Briles-P
61 1987 Pat Kelly-RF
51 1997 Rick Mahler-P

Thanks, Dan!

John (You Can Call Me Grandma) Murphy Posted: September 29, 2007 at 10:30 PM | 296 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 3 of 3 pages  < 1 2 3
   201. DavidFoss Posted: October 15, 2007 at 03:08 PM (#2577022)
bump
   202. sunnyday2 Posted: October 15, 2007 at 03:28 PM (#2577048)
Don't miss #199 and #200, folks. They're now marooned back on page 2.

Vote for Bucky Walters. Don't vote for Bob Johnson. ;-)
   203. TomH Posted: October 15, 2007 at 03:41 PM (#2577066)
nah, vote for both of them. Johnson hit as well as Belle. And fielded better.
   204. sunnyday2 Posted: October 15, 2007 at 03:51 PM (#2577081)
See #200.

Though I haven't done Belle yet. But Puckett, Murphy and Reggie2 were better relative to their leagues than Bob.

Bucky was off the charts.
   205. DL from MN Posted: October 15, 2007 at 03:55 PM (#2577088)
> Bucky was off the charts

because you're using Win Shares.
   206. sunnyday2 Posted: October 15, 2007 at 06:01 PM (#2577243)
I prefer to think that Bucky was off the charts AND I'm using WS ;-)
   207. KJOK Posted: October 15, 2007 at 08:55 PM (#2577474)
Note that this question is not an endorsement of the win shares metric. I do, however, use win shares as one element in my system, along with WARP1 and WAR, and I would like to keep my system consistent after passing the 2001 line.


I wrote STATS, Inc. after 2002 to see if there was going to be a supplement to the electronic Win Shares issued that could be purchased, and they sent me a free file that contained Win Shares by Team, but with NO breakout for batting/fielding/pitching. After 2002, I don't think they even offered any supplemental file.
   208. sunnyday2 Posted: October 15, 2007 at 10:56 PM (#2577562)
Update

Vs. the median position regular in the respective league

1. Albert Belle +11 for 10 years
2. Kirby Puckett +8.5 for 11 years
3. Dale Murphy +8.5 for 10 years
4. Will Clark +7.5 for 13 years
5. Reggie Smith +6.5 for 11 years
6. Bob Johnson +6 for 10 years
7. Gavy Cravath +5 for 7 years

Versus the #1 (or #2) player at the position

1. Cravath +3
2. Belle -1.5
3. Puckett -5
4. Murphy -5
5. Smith -5.5
6. Will Clark -6
7. Bob Johnson -7.5

Belle's medians are Brunansky, Jorge Bell, Shane Mack, R. Henderson or Juan Gonzalez, R. Greer, Pride or Goodwin, Jose Cruz/Roberto Kelly sharing the position, BJ Surhoff, O'Neill or Dye and Buhner. The position leaders include Rickey, Molitor (DH), Juan Gone, Justice, Manny twice and Belle 4 times.

Clark's medians are Garvey (last year), G. Davis or Durham, Magadan or Von Hayes, Magadan or Galarraga, the same two or Bream, H. Morris, Merced, E. Murray, Joyner, JT Snow, John Jaha, Delgado, T. Clark or Segui. The position leaders include Keith Hernandez, Eddie Murray, Bags, Frank Thomas, Raffie and Mo, and Clark 4 times of course.

These are pretty competitive cohorts. There was some FAT manning these positions too, however. Nick Esasky 3 years at 8, 11 and 10 WS; F. Stubbs at 9 and 6; Cianfrocco and Vanderwal in successive seasons for Mtl (at 4 and 6 WS); rookies like Snow (3), Tino (7), Jaha (4); Sorrento at 9 and 11; Stahoviak at 6, 12 and 4, and that's just Clark. For Belle, add in Henry Cotto, Darnell Coles, Lee Stevens, Rich Amaral (2 years at 6 and 7); Rob Ducey, Bubba Trammell; more Sorrento; and Goodwin at 19, 9 and 5. Actually that 9 for Goodwin was his median year. That was 1996, a really terrible year for AL LF. Hammonds with 3 for Baltimore, Greenwell 7, Garrett Anderson 6, Tim Raines 7, Amaral 7, and these were all the #1 LF, though obviously most of them didn't play 162 games or anything close.

Belle may have had a short career but he was a very very good player just as long as any of the backloggers were. You've got to be rewarding hanging around pretty generously not to have Don't Call Me Joey somewhere around your ballot, I mean top 25-30 anyway.
   209. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 15, 2007 at 11:20 PM (#2577603)
Ya think? It's 4 very big years ('94-'96 and '98), two other All-Star seasons ('93 and '99), three seasons just above positional average ('91, '92, 97), and that's it, no? If all you look at is the top 6 seasons, Belle is definitely one of the top candidates, but a lot of backloggers with 80% of the peak had a good 5-6 more seasons at an All-Star level than Belle did. Unless you write off seasons 7-12 or so entirely--which are not "hanging around" but All-Star years for many candidates--it doesn't seem surprising to me that Belle isn't getting that much traction. He's only a viable option for the "pure" peak or short-prime voter.
   210. jimd Posted: October 15, 2007 at 11:22 PM (#2577612)
Are those win shares using the same exact formulas as Bill James win shares? My recollection is that they handle pitcher hitting differently.

THT used to document what their changes were, but that link is now broken. Instead it now points to an article announcing the arrival of the 2004 Win Shares.

IIRC, the three most significant differences were negative BWS, negative PWS, and a modification to how saves affected relievers.

Negative BWS penalizes pitchers that can't hit (also hitters that can't hit, but there aren't many of those). Bottom line result is it converts PWS into BWS. This could be a good idea if BJ Win Shares overrated pitchers and underrated hitters; that's not the case so its net effect is to accentuate one of WS's most basic underlying flaws. This mostly affects NL Win Shares.

Negative PWS penalizes pitchers that can't pitch anymore. If a pitching staff has a starter that gets lit up repeatedly in hopes he can turn it around, then he might be penalized enough Win Shares to inflate the other guys on that staff.

IIRC, the saves modification reduced the number of Win Shares that the closer received.
   211. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 15, 2007 at 11:43 PM (#2577687)
Some fodder for the pure peak voters: best top 5 non-consecutive seasons by eligible candidates, measured in WARP2.


1. Chuck Klein, 36.3
2. Albert Belle, 35.5
3. John McGraw, 35.2
4. Dale Murphy, 34.4
5. Bill Bradley, 34.0 (!!)--was he ever considered a candidate?
6. Jim Fregosi, 33.3
7. Dolph Camilli, 32.5
8. George Burns, 31.9
9. David Concepción, 31.9
10. Rico Petrocelli, 31.8
   212. DavidFoss Posted: October 16, 2007 at 01:09 AM (#2578081)
5. Bill Bradley, 34.0 (!!)--was he ever considered a candidate?

Yeah, he ran in 2000 but didn't get the nomination. (Notice that Gore is in the HOM!)
   213. Howie Menckel Posted: October 16, 2007 at 01:11 AM (#2578088)
And (Donie) Bush got some votes as well, yes?
   214. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 16, 2007 at 02:25 AM (#2578452)
hahah, funny, guys. Seriously, though, if Bradley hadn't had a career-ruining injury in the middle of 1906, he'd be a lock. He was *badass* from 1901-06.
   215. sunnyday2 Posted: October 16, 2007 at 03:58 AM (#2578681)
>If all you look at is the top 6 seasons

Ah, no. I looked through year 10 for all of them, though what I did was look at their primes defined as batting title eligible years ? 100 OPS+. If that isn't the heart of any resume, I don't know what is. I think it said that somewhere, it musta been hidden down there in the small type. (Except Cravath who only had 7 real years. I use his full MLE for balloting.)

>Unless you write off seasons 7-12 or so entirely--which are not "hanging around" but All-Star years for many candidates

Would some of seasons 7-12 fall within 10 seasons? I'm confused now.

But who had 5-6 more years than the 10 that are covered for all of them above, anyway?

OK, Reggie, I get Reggie. Belle is +4.5 WS per year vs. Reggie, who actually only has two extra "prime" years, that is, ? 100 OPS+ than Albert, the rest were < 100 OPS+. I like Reggie BTW, he is PHoM. But seriously, Belle has about 7100 PA, Reggie has 7900. That's not 5 years, that's a year-and-a-half in dog years. (See below.)

Bob Johnson, no not 5-6 more than the 10 that I already covered. And Belle was 5 WS per year better than Bob compared to the position median. Per year. For 6...NO, for 10 years. 10 years. 5 WS per year for 10 years vs. the position medians of their day. 10 years. +2.5 every year vs. Dale Murphy. +3.5 vs. Will Clark though the Thrill has 3 extra prime years and so he is #1 on my ballot, with Belle #2.

OPS+

Joey 144/191-75-75-58-47-43-34-24-17-10
Reggie 136/168-64-58-48-42-42-37-33-28-26-25-16

Win Shares (adj to 162)

Belle 257/37-34-34-31-27-24-18-16-15-15
Reggie 325/29-29-26-25-25-25-24-24-23-20-19-17-16-15-14

>If all you look at is the top 6 seasons

Like I said. Huh?
   216. rawagman Posted: October 16, 2007 at 12:32 PM (#2579108)
Sunny - You construct an interesting argument pro-Belle and anti-Johnson. I think one sufficient line of refute might be that a) WS underscores big differences in defensive ability, so much so that a good defender like Johnson will not benefit and a poor defender like Belle will not suffer for it and b) that WS system limits WSs in an extreme way for guys who played for big losers, as Johnson did and gives big benefits for guys who consistently played on winning teams, like IIRC, Belle did. I'm sure Dan R will comment on R. Smith, so I'll stick out of that one.
BTW - I have all three between 11-21 on my extended ballot, so I'm not really opposed to any one of them as such, all differences being micro.
   217. DL from MN Posted: October 16, 2007 at 02:36 PM (#2579279)
Crossposting from ballot thread
------------------
18. favre Posted: October 15, 2007 at 05:45 PM (#2577516)

DL,

The MLE's from Redding's thread see him as comparable to Hershiser. I'm interested to know why you have Willis as a comp.

19. DL from MN Posted: October 15, 2007 at 06:11 PM (#2577536)

Dick Redding MLE
Support Neutral MLE projected record: 230-177
ERA+ 114 (DERA 3.95)

Vic Willis
249-205 record
ERA+ 118

The main similarity is how they both packed a lot of innings and a small number of peak seasons. Outside of the peak both ate innings but had a lower level of effectiveness. It's hard to compare with Hershiser who never had a 3 year peak like either Willis or Redding. Another comparable for Dick Redding would be if Dwight Gooden's 1985 performance lasted 3 seasons instead of one-plus.

20. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 15, 2007 at 06:18 PM (#2577541)

DL from MN, you can't be serious. If you have three seasons of Doc Gooden 1985, you're Sandy Koufax.

29. DL from MN Posted: October 16, 2007 at 10:11 AM (#2579219)

Gooden had 33 Win Shares in 1985, correct?

Here's Dick Redding's 3 year consecutive peak in the i9 translations:

1915 40
1916 33
1917 27

Then he saw combat duty in WWI.

The buzz on Dick Redding at his peak is every bit as compelling as the anectdotes about Gooden in 1985. He's credited anywhere from 7 to 30 Negro League no-hitters depending on the source including a 17 strikeout perfect game. He had streaks of 17 (1912) and 20 (1915) consecutive wins. He was an absolute workhorse at his peak, working both games of a doubleheader at times.

30. John (You Can Call Me Grandma) Murphy Posted: October 16, 2007 at 10:18 AM (#2579235)

Are those numbers for Redding converted downward due to the less innings he would have been pitching during the Eighties, DL?

31. DL from MN Posted: October 16, 2007 at 10:26 AM (#2579250)

I don't know enough about the I9s to determine the answer to that question.
------------------------

I looked up the IP projections and they have Redding in the top 10 in IP 4 times with 320-250 IP. Gooden himself pitched 275 innings in 1985 to earn those 33 Win Shares so it can't be too much of a discount. I think Redding's 1915 is at least comparable to Gooden's 1985 and Redding has a rookie season in 1911 that seems more impressive than Gooden's 1984.
   218. karlmagnus Posted: October 16, 2007 at 02:46 PM (#2579290)
I thought we agreed thge I9s were at least 5-10% high.
   219. DL from MN Posted: October 16, 2007 at 02:47 PM (#2579294)
You said the I9s were high, Chris said he didn't think so.
   220. Dr. Chaleeko Posted: October 16, 2007 at 03:35 PM (#2579371)
DL,

But 33 WS in 1985 is a hell of a lot more impressive than 33 WS in the 1910s, right? Perhaps more impressive than even 40 WS in the 1910s.

I'd posit that 33 WS in 1985 is probably also more impressive than 33 WS in the 1960s, though that ground might be a bit shakier.
   221. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 16, 2007 at 04:02 PM (#2579410)
Of course 33 WS in 1985 is more impressive than 33 WS in the 1960's--average IP per starter were much lower. How could it not be?
   222. DL from MN Posted: October 16, 2007 at 04:10 PM (#2579423)
There was a bit of hyberbole in the statement that Redding's 1985 lasted 3 years but there is a nugget of truth - Redding was able to maintain his dominance longer than Gooden was.

What are the best 3 season peaks of the teens pitchers? Is Walter Johnson's 1912-1913 equivalent to Gooden's 1985 in terms of dominance or was Johnson even better?
   223. Chris Cobb Posted: October 16, 2007 at 04:48 PM (#2579487)
On the i9s numbers:

The major-league projections created by the i9 project attempt (as do my MLEs) to construct the player's record as if they had played in the major leagues of their own day instead of the Negro Leagues. So they place Redding in a 1911-2x context.

I generally thought the i9s estimates were too high, though it did not appear that their projections were entirely consistent, so some players seemed more exaggerated than others. I think my difference with Karl may have been that I suggested a 5% discount and he preferred 10%. I certainly concur with him that the i9s numbers should not be accepted at face value. On Redding in particular, I think that the innings pitched projections are, let's say, optimistic. Redding by their projections would have been rather more durable, on a seasonal basis during his prime in the late teens and early twenties, than any major-league contemporary.

As to _effectiveness_, that is harder to judge with pitchers. One has to make a lot of assumptions about a pitcher's fielding and hitting support to interpret the fragmentary records of pre-1920 play by African-Americans. Without having Redding's data in front of me from i9s and elsewhere, I wouldn't want to speculate about the accuracy of their projections in this respect.

I would not have said that any of Dick Redding's seasons were as good, in context, as Dwight Gooden's 1985. I would say that Walter Johnson's 1912-13 were as good as or better than Gooden's 1985. If you are denominating in win shares. There are quite a few 40-win share seasons by pitchers during the teens. There aren't any during the 1980s.
   224. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 16, 2007 at 04:49 PM (#2579489)
By rate, BP has Johnson with a DERA of 2.57 in 1912 and a 2.29 in 1913, which translate to 5.0 and 5.8 wins above average per 200 innings, while Gooden had a 2.32 in 1985, or 5.7 wins above average per 200 innings. The projected standard deviation of wins above average per 200 innings for pitchers was around 2.15 in the 1912-13 AL, and 1.91 in the 1985 NL, so Johnson's 1912-13 would have been worth 5.0 and 5.8 *1.91/2.15 = 4.4 and 5.2 wins above average per 200 innings in the 1985 NL. BP gives Johnson 300 translated innings in those years, to 280 for Doc's 1985, so that makes Johnson's 1912-13 6.6 and 7.8 wins above average in the 1985 NL, and Gooden 8.0 wins above average in the 1985 NL.

Now, replacement level. Tangotiger finds that an average non-closer reliever will have a .500 winning percentage as a reliever and a .410 winning percentage as a starter, and that a replacement pitcher will have a .470 winning percentage as a reliever. I will add an additional step and hypothesize that the average team's mop-up man has a .480 winning percentage as a reliever. So we then have:

Cost of replacing starter with average non-closer (long) reliever: .090 of winning percentage at a leverage of 1
Cost of replacing average non-closer (long) reliever with mop-up man: .020 of winning percentage at a leverage of 0.6
Cost of replacing mop-up man with replacement pitcher: .010 of winning percentage at a leverage of 0.2

Total cost of a replacement pitcher: .090*1 + .020*.6 + .010*.2 = .104. So in the modern game, a replacement starter has an effective winning percentage of .396, or -2.3 wins per 200 innings. Standard deviations for pitchers in the modern game, much like the 1912-13 AL, are also around 2.15, so a replacement pitcher would be -2.3*1.91/2.15 = -2.0 wins per 200 innings in 1985. Thus, the difference between an average and replacement starter in the 1985 NL would be 3 wins in Johnson's innings, and 2.8 wins in Gooden's innings. Add that on to the value above average, and we get Johnson with 6.6 + 3.0 = 9.6 1985 NL wins above replacement in 1912, 7.8 + 3.0 = 10.8 1985 NL wins above replacement in 1913, and Gooden with 8.0 + 2.8 = 10.8 1985 NL wins above replacement in 1985.

This is actually not *that* much trouble to do; it's just aggravatingly reliant on BP's assessments of team defense and innings translation system. But I could try to compile these numbers for the leading candidates before the next election if that would be helpful.
   225. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 16, 2007 at 04:53 PM (#2579492)
I forgot to mention that hitting seems to break even--both Johnson in 1912-13 and Gooden in 1985 were outstanding hitters for pitchers compared to their contemporaries.
   226. mulder & scully Posted: October 16, 2007 at 05:20 PM (#2579521)
According to SABR's Deadball Stars of the American League, Walter Johnson led the Senators in home runs from 1901 to 1919. That is either amazing for Johnson or really sad for the rest of the team.
   227. DL from MN Posted: October 16, 2007 at 05:25 PM (#2579529)
Your assessment of Gooden's 1985: 10.8
WARP1: 14.2

Your Johnson 1912: 9.6
WARP1: 17.2

Your Johnson 1913: 10.8
WARP1: 18.8

I would find a spreadsheet with your pitcher WAR useful. Even without the standard deviation adjustments it is a step up on WARP.
   228. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 16, 2007 at 05:40 PM (#2579554)
DL from MN, I'm doing these by hand! I don't have a spreadsheet with every pitcher ever or anything close to it. But I will try to put them together for the main candidates this week.

The biggest reason for the discrepancy is that I am calculating using BP's translated innings rather than actual innings. I've been doing a lot of work on innings translation and hope to have my own system working soon.
   229. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 16, 2007 at 05:42 PM (#2579557)
Johnson definitely had the greatest hitting season ever by a pitcher in 1925: 97 at-bats, .433/.455/.577, 163 OPS+.
   230. DL from MN Posted: October 16, 2007 at 05:56 PM (#2579574)
> I'm doing these by hand

Understood, I am just stating that getting to proper replacement value is a valuable first step.
   231. OCF Posted: October 16, 2007 at 06:39 PM (#2579623)
By RA+ Pythpat, I have Johnson 1912 and 1913 as 32-9 and 32-6, respectively. Converting that into single-season FWP gives us 47 and 53, and these are the two highest such years of the 20th century, at least among pitchers I've looked at. (Going back to 1993 picks up Rusie 1894 at 36-14, or 48 FWP.)

Here are the top post-WWII years, at least that I have worked up so far:

Gibson 1968 27-7 [41]
Gooden 1985 25-6 [39]
Carlton 1972 28-10 [38]
Clemens 1997 24-6 [37]
Perry 1972 27-11 [37]
Martinez 2000 21-3 [37]
Koufax 1966 26-10 [36]
Guidry 1977 24-7 [35]
Seaver 1971 24-8 [34]
Chance 1964 23-7 [34]
Maddux 1995 20-3 [34]

Of course, that's just a starting point, and I'm not doing any of the things Dan R. is doing.

I'm not adjusting for defensive support. I'm not adjusting for pitcher's own offense (which would boost Johnson and Gooden and lower Koufax and Chance). This is RA, not DERA; in particular, Gibson's 1968 has a little bit of a whiff of a BABIP fluke season.

I'm not adjusting for the expectations for IP between eras; this is clearly a large part of what puts Johnson out in front. Deadball pitchers, having fewer strikeouts and not fearing the HR, relied more on the defense behind them and thus contributed less to the value of that RA than did more recent pitchers. (Of course, in the context of his own time, Johnson was a high-K pitcher.) The high IP possible in the mini-deadball years of the mid-60's and early 70's contributes to several of the names on the list. FWP is a gimmick - it might not be fair in adjusting extreme high-RA+, low-IP cases (like Martinez 2000 or Maddux 1995) compared to high-IP cases.
   232. Mike Green Posted: October 16, 2007 at 06:40 PM (#2579628)
DanR, the Gooden/Johnson comparison is nicely done.

The innings translations and DERA that BP does for Johnson particularly are a bit difficult. It was easier for starting pitchers to accumulate innings in the deadball era due to the environment ("saving one's best for the clinches"). This naturally applied downward pressure on strikeout rates.

In Johnson's case, he struck out more relative to league averages in 1912 (7.39/4.25) than Gooden did in 1985 (8.72/5.50). The IP leaders in the league in 1912 were widely separated. Johnson threw 369, below Walsh's 394. Wood was next at 344, then Groom at 316 and Ford at 292. Gooden led the NL of 1985 with 276 innings, but there were 5 other pitchers with more than 260. I have no idea how BP arrives at a translated innings figure of 300 innings for Johnson, but I am pretty sure that it would be quite controversial.
   233. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 16, 2007 at 07:02 PM (#2579657)
To be clear, BP's DERA does NOT adjust for a pitcher's BABIP fluctuation, ONLY for the quality of his defense. DERA gives full credit for BABIP flukes, both good and bad.

Using my own (still very preliminary) translation system, I get Johnson's 1912 at 262 2005 innings, his 1913 at 245 2005 innings, and Gooden's 1985 at 248 2005 innings. Repeating the same procedure, we get:

2005 wins above average (same stdev as 1912-13)

Johnson 1912: 5.0 2005 wins per 200 innings * 262/200 = 6.6
Johnson 1913: 5.8 2005 wins per 200 innings * 245/200 = 7.1
Gooden 1985: 5.7 * 2.15/1.91 stdev correction = 6.4 2005 wins per 200 innings * 248/200 = 7.9

Replacement level

Johnson 1912: 2.3 2005 wins per 200 innings * 262/200 = 3.0 wins
Johnson 1913: 2.3 2005 wins per 200 innings * 245/200 = 2.8 wins
Gooden 1985: 2.3 2005 wins per 200 innings * 248/200 = 2.9 wins

Total

Johnson 1912: 6.6 + 3.0 = 9.6 2005 wins above replacement
Johnson 1913: 7.1 + 2.8 = 9.9 2005 wins above replacement
Gooden 1985: 7.9 + 2.9 = 10.8 2005 wins above replacement

This produces different results because BP translates '13 Johnson to having 30 more innings than Gooden, whereas I translate him to having 3 less.
   234. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 16, 2007 at 07:04 PM (#2579662)
Oh, and DERA makes no correction for strikeout rate or era either. DERA is simply the RA the pitcher would have had if he had pitched in front of an average defense in a league scoring 4.5 runs per game with a Pythagorean exponent of 2. The only things it corrects for are run environment and fielders.
   235. Tiboreau Posted: October 16, 2007 at 07:14 PM (#2579681)
From the ballot thread:

John (You Can Call Me Grandma) Murphy: Are those numbers for Redding converted downward due to the less innings he would have been pitching during the Eighties, DL?

DL from MN: I don't know enough about the I9s to determine the answer to that question.

It isn't the I9s, it's Win Shares. Dwight Gooden's WS in 1985 (like everything else) was historic compared to his peers. Cannon Ball Dick's WS look even better, but there were other pitchers in the deadball era who were closer to Dick's peak than there were '80s pitchers to Dwight Gooden's (of course, the same is true for Vic Willis). Redding's WS are adjusted into the league situation of his time, like all Negro League ballplayers. And Win Shares, like traditional numbers, doesn't correct for the decreasing innings and for how important defense was in Willis's or Redding's era, so their peak looks very impressive 80 years later as we compare them to backloggers who didn't play when they did.

Also, season-by-season data for NgL pitchers is even more iffy than it is for NgL position players, particularly for those who played before 1920. There is less to go off of, and in Redding's case there is no regression to correct for Short Sample Size issues like is the case for Oms, or Willard Brown, or Biz Mackey, etc. Someone else could speak to how much of an effect that'd have, but it seems to me that in situations where we've gone from non-regression to regression statistics (Dobie Moore) more subtracted from their peak and added to their outside peak years than I, at least, would add back to their regressed stats (typically 1, maybe 2 WS added to peak, subtracted from non-peak w\ regressed stats).

DanR: Tiboreau, why Cey over Nettles? Same era, same position, same peak and prime, but Nettles has meaningfully more career.

Win Shares agrees w\ you, but if I used just WS neither would be on my ballot and Sal Bando would. ;-)

The difference for me is their WARP. According to BP's numbers, Cey & Nettles' career numbers are fairly close: only 5 WARP1 advantage for Nettles, and a 10 WARP3 advantage (after adjusting for DH, I suppose). However, WARP appears to have Ron Cey's peak/prime solidly above Nettles: both gentleman have 1 year over 10 WARP1, but Cey has 5 years of 9 WARP1 while Nettles has only 2 years over 8 WARP1. Cey's advantage shrinks w\ WARP3, but as a peak/prime voter it's enough for the Penguin to make the last slot on my ballot (which he's done the past couple years) while Nettles sits 10 spots behind him.
   236. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 16, 2007 at 07:18 PM (#2579691)
Tiboreau, are you adjusting WARP1 for the DH? You need to add on 0.6 wins per 162 games to WARP1 for all players in DH seasons.
   237. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 16, 2007 at 07:41 PM (#2579748)
Sorry, DERA does correct for park as well. It is a very good stat, one of BP's best--the methodology is replicable and its results are accurate. You have to trust FRAA, but at the team level FRAA does correlate extremely well (.8-.9) to team defensive efficiency, at least in my spot checks.
   238. OCF Posted: October 16, 2007 at 07:44 PM (#2579755)
Hey, I have Willis 1899 as 28-10 [38] if you don't adjust for defense - right up there with Carlton 1972. In Willis's case, I did attempt to adjust for defense, which shaves it down to 24-14, and those are 1899 IP.
   239. sunnyday2 Posted: October 16, 2007 at 09:10 PM (#2579851)
>WS system limits WSs in an extreme way for guys who played for big losers, as Johnson did and gives big benefits for guys who consistently played on winning teams, like IIRC, Belle did.

That's a theory and I don't know that it's wrong. Bill James says it's wrong however. It seems that we the HoM voters have accepted WS unfavorable verdict on Chuck Klein though the same argument might be made for Klein as for Johnson. Why Johnson and not Klein? WARP seems to prefer Klein.
   240. sunnyday2 Posted: October 16, 2007 at 09:17 PM (#2579854)
Among the #1 starters on each team through the period of his own career, Willis was on average about a median #1 starter (4th or 5th best in the NL in any given year, on average). There are certainly candidates out there who were better than the median. I like Willis, I've had him on my ballot. Ditto Joss, who I now know to have been well below a median #1 starter because of his low workload. But neither were among the best starters of their day. Dean, Newcombe and Saberhagen were all substantially better compared to their contemporaries.
   241. jimd Posted: October 17, 2007 at 12:35 AM (#2580130)
WS system limits WSs in an extreme way for guys who played for big losers, as Johnson did and gives big benefits for guys who consistently played on winning teams, like IIRC, Belle did.

It's a strength of schedule effect within a league, and really doesn't apply to modern players because the leagues are twice as big. In the old 8-team leagues, it would be noticeable for teams that won about 60 games or less in a season (or lost 60 games or less), and gets stronger the further away from .500 the team is. In a modern 14 team league, the same size effect would occur at about 52 wins (or losses) or less; not enough of those teams around to worry about.
   242. jimd Posted: October 17, 2007 at 12:41 AM (#2580153)
IOW, in the 8 team leagues you'd notice the distortion at around .400/.600.
In modern leagues it's at around .300/.700.

At that level, the distortion is large enough to affect half the regulars on a team by one Win Share. Which is this should be considered a career affect for players on long-time cellar-dwellers (such as the late 30's A's or Browns or Phillies) or long-time dominant teams (such as the Yankees or Dodgers or Giants).
   243. Devin has a deep burning passion for fuzzy socks Posted: October 17, 2007 at 04:55 AM (#2580826)
There was a discussion in IRC about what was the best defensive team in history (after Curtis Granderson claimed it was the Rockies). I suggested the mid-80s Cardinals, but not really based on anything. I was wondering if Dan's defensive ratings had an answer to the question -if you can get one in a simple way, don't go to a lot of trouble.
   244. mulder & scully Posted: October 17, 2007 at 05:52 AM (#2580847)
The 1940 Reds would have to be on the list.

C - Ernie Lombardi was 6th among regulars
1b - Frank McCormick was 1st
2b - Lonny Frey was 1st
3b - Billy Werber was 1st
SS - Billy Myers was 4th
CF - Harry Craft was 1st
LF/RF - Their top 3 finished 1st, 3rd, and 10th

That is pretty damn impressive.
Actually, the Reds from 1939 to 1944 were very impressive.

Looking through my not-complete list of Dan R.'s Gold Glove winners I find the following teams.

1893 Pittsburgh: Jake Beckley 1st at 1st, Elmer Smith 2nd in LF, and Patsy Donovan 1st in RF.
The 1894-96, 98 Cleveland team all had 3 GG or very close
The 1898 and 1900 Boston had 4 GG
The mid to late 1890s Baltimore Orioles were defensively dominant as well.

There are many teams with 3 winners/very near winners throughout the 1901-15 era.
The 1904 Chicago White Sox, Boston Red Sox, and Cincinnati Reds each had 4.

Then you get the 1906 Cubs.
C - Johnny Kling - 1st
1b - Hofman and Chance were 2nd and 3rd
ss - Tinker was 2nd
LF, CF, and RF were all 1st place at their positions.

Then you have the 1908 Indians.
1b - 1st with George Stovall
2b - 1st with Nap Lajoie
LF and CF were 1st also.
RF was 2nd.

1909 Pirates had 4 - Gibson, Wagner, Clarke, and Wilson

1910 Athletics had 4 firsts and 2 seconds. Their 2 catchers tied for first, Eddie Collins and Frank Baker were GG, and their LF and RF were both 2nd.

From 1893 - 1915, the 1898 and 1900 Boston Braves, 1904 Red Sox, White Sox, and Reds, 1906 Cubs, 1908 Naps, 1909 Pirates, 1910 Athletics all had 4 GG winners. The Cubs and A's each had 2 seconds. But no one matches the 1940 Reds with 5.
   245. Chris Fluit Posted: October 17, 2007 at 02:15 PM (#2581015)
Of more recent vintage, I think of the late '90s Mets with Alfonzo-Ordonez-Alomar and Olerud, but it's possible that their outfield dragged them down.
   246. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 17, 2007 at 02:59 PM (#2581115)
Going by FWAA2, I have:

1. 1906 Chicago Cubs, 9.0. Guess Tinker to Evers to Chance were famous for a reason.
2. 1906 Cleveland Indians, 8.9. I've discussed them in previous threads. They led the league in defensive efficiency by a large margin and in double plays by 25. Nap Lajoie and Terry Turner were probably the greatest single-season DP combination ever, and Flick was a stud as well. I suspect Michael Humphreys would question this ranking, since he often says non-PBP systems overemphasize double plays, but there's no way they're not in the top 10. This further deflates Addie Joss's claim to Merit.
3. 1940 Cincinnati Reds, 8.5, as commented above. Take that, Bucky Walters.
4. 1973 Baltimore Orioles, 8.3. No surprise here, as Joe Dimino regularly reminds fans of Jim Palmer.
5. 1899 Boston Beaneaters, 8.2. Jimmy Collins's greatest year, plus terrific seasons from Duffy and Stahl in the outfield. Their .715 defensive efficiency led the next-best team by 23 points.
6. 1898 Boston Beaneaters, 8.1. Same $hit, different day.
7. 1969 Baltimore Orioles, 7.9. Same $hit, different day, part deux.
8. 1986 St. Louis Cardinals, 7.7, as commented above.
8. 1907 Chicago Cubs, 7.7. Same $hit, different day, part trois.
10. 1895 Baltimore Orioles, 7.6. The Hughie Jennings Show, plus fine work from Keeler, Kelley, McGraw, and the forgotten Steve Brodie.
11. 1919 Cincinnati Reds, 7.4. Morrie Rath, Roush, and Groh were the defensive stars of this tainted champion.
11. 1924 Pittsburgh Pirates, 7.4. Carey, Cuyler, Maranville, Traynor, and Glenn Wright.
11. 1897 Boston Beaneaters, 7.4. Getting repetitive.
14. 1943 St. Louis Cardinals, 7.3. In wartime, Stan Musial had his best defensive season and Marty Marion was Marty Marion.
15. 2002 Anaheim Angels, 7.1. Statheads may hate Darin Erstad, but he got them to the playoffs with his glove.
15. 1897 Baltimore Orioles, 7.1. Heard this before?
15. 1905 Chicago White Sox, 7.1. Lee Tannehill, the brother of deft hurler Jesse, was the original good-field, no-hit infielder, and he played third base as well as anyone has before or since for a while. Fielder Jones also excelled for the precursor to the Hitless Wonders.
18. 1999 Cincinnati Reds, 7.0. Regularly recognized as one of the top fielding teams of the modern era, thanks to Pokey and Larkin. Even 45-HR man Greg Vaughn caught the ball.
18. 1916 Brooklyn Robins, 7.0. Rube Marquard had George Cutshaw, Mike Mowrey, and Zack Wheat to thank for his success on this pennant winner.
18. 1919 New York Yankees, 7.0. The immediate pre-Ruth Yankees could pick it all over the field. Ping Bodie, Chick Fewster, Duffy Lewis, Roger Peckinpaugh, and Del Pratt were all better than a win above average.
18. 1910 Philadelphia Athletics, 7.0. For a cool $100,000.
18. 1931 Philadelphia Athletics, 7.0. The second A's dynasty wasn't known for their defense, but defense they played. Bishop, Simmons, Jimmy Dykes, and Bing Miller were the main contributors. Not that Lefty Grove needed the help.
18. 1944 St. Louis Cardinals, 7.0. Another wartime Marion job.
24. 1900 Boston Beaneaters, 6.9. Appears to be the most prolonged defensive dynasty.
25. 1896 Cincinnati Reds, 6.8. The carcasses of Buck Ewing and Bid McPhee were all this team had for star power, but everyone could field, most notably third baseman Charlie Irwin. A beneficiary of their defense, pitcher Billy Rhines, is a favorite of mine in WhatIfSports SimLeagues.
25. 1908 St. Louis Browns, 6.8. Bobby Wallace, Hobe Ferris, and Danny Hoffman starred.
   247. Devin has a deep burning passion for fuzzy socks Posted: October 17, 2007 at 03:09 PM (#2581136)
Cool, thanks Dan. I guess if you were going to put a timeline factor in, you'd come up with the 69-73 Orioles on top, unless you went with a really steep timeline and then have the Angels or Reds.
   248. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 17, 2007 at 03:15 PM (#2581145)
I haven't studied this, but the number of pre-1920 teams on the list doesn't surprise me at all--you would think that the standard deviation of defense would steadily decline over time as the percentage of Three True Outcomes increased...
   249. DavidFoss Posted: October 17, 2007 at 05:27 PM (#2581339)
That is pretty damn impressive.
Actually, the Reds from 1939 to 1944 were very impressive.


Bill James wrote about these teams in his book on baseball managers. In 1938, McKechnie came in and retooled the team to be glove-centric.
   250. mulder & scully Posted: October 17, 2007 at 06:05 PM (#2581390)
James' Manager Book is one my favorites of his. McKechnie came to the Reds and used 3 SS in his infield. Lonny Frey came up as a SS, but played 2b, Bill Werber played some SS when he started, but played 3b, and Billy Myers was the regular SS.
   251. Best Dressed Chicken in Town Posted: October 17, 2007 at 06:34 PM (#2581431)
According to SABR's Deadball Stars of the American League, Walter Johnson led the Senators in home runs from 1901 to 1919. That is either amazing for Johnson or really sad for the rest of the team.

Wow, what a great stat. I checked and it's true. Johnson was 12th in PA for Washington during that time, and had the most HR, 16 to Danny Moeller's 15.
   252. Joey Numbaz (Scruff) Posted: October 17, 2007 at 08:30 PM (#2581689)
My translations get Johnson 1912-13 at 334 and 349.7 IP.

Gooden 1985 is at 278.

I use the middle half of Nth league leaders, where N = teams in league. So in an 8 team league, I take #3-#6 to set the standard. In a 12-team league it's #4-#9.

I take the average of those pitchers, and set them to 258.3 IP. This puts them exactly on the same scale as Prospectus tIP. But mine aren't as prone to flukey league leaders that are way above everyone else throwing off the curve - those guys actually get credit for their insane innings totals.

Prospectus uses the top 5 IP in the league and sets them equal to 275 IP. They use the top 5 whether it's an 8-team or 16-team league, which I believe is another flaw.

Historically, setting my sub-group to 258.3 is the same as setting the top 5 to 275, so that's why I use that number.
   253. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 17, 2007 at 08:54 PM (#2581730)
I have been tearing my hair out over innings translation for a few months now with no success. The problem I have with straight multiplication, as Joe and BP do, is that it takes no account of the standard deviation of innings pitched. Baseball history has seen periods where innings per starter are closely clustered together, and periods when they are widely spread out. If you use a straight multiplier, you over-reward the workhorses and over-penalize the laggards of high-stdev eras, and do the opposite for low-stdev eras. But if you normalize the distributions, you wind up translating guys who threw 130 innings in the 1890s *upwards* to 160 or so for the modern era, because fifth starters throw more innings now than they did back then. I've been trying to use some sort of exponential distribution, but I can't get a good fit unless I include relievers--in which case I wind up translating guys who threw 130 innings in the 1890s down to 40 innings for the modern era, because the 30th percentile of all pitchers were fourth starters back then and are relievers now. So I'm stuck. Any ideas?
   254. Joey Numbaz (Scruff) Posted: October 17, 2007 at 09:08 PM (#2581753)
Dan, I don't think standard deviation should be an issue here.

I think they should be compared to what the top guys that season did. I remove the top few to avoid having the flukes throw off the curve, of course.

But innings is mostly a function of 'how much' not 'how good' - though of course you have to at least be good enough to get the innings. So I don't think 'easy to dominate' should be an issue in what is mostly a case of endurance.

If a guy is throwing significantly more innings than his peers in a season, I think that needs to be reflected.
   255. Joey Numbaz (Scruff) Posted: October 17, 2007 at 09:28 PM (#2581776)
I would add that I could be convinced otherwise, but haven't been yet.
   256. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 17, 2007 at 10:56 PM (#2581829)
Joe, how can you translate Johnson's 1913 innings *upwards*? Starting pitchers obviously worked more innings in 1913 than they did in the liveball era...
   257. Joey Numbaz (Scruff) Posted: October 17, 2007 at 11:32 PM (#2581848)
Johnson's starter IP in 1913 are scaled down from 311 to 282.

However, Johnson also threw 35 innings in relief at an estimated leverage index of 1.93. In those 35 innings he was 7-0 with 2 saves (he also had a 0.26 ERA, though that doesn't weigh into the leverage calc).
   258. Joey Numbaz (Scruff) Posted: October 17, 2007 at 11:37 PM (#2581851)
If we are going to give credit to relief aces for leveraged innings, we also have to give them to starters when they relieve.

One thing though - I think that I need to tweak my 'leader' base to only use starter innings in the calc and not all innings. I was just using the top X in innings (depending on the number of teams in the league). But I probably should tweak that to be top X in starter innings. Although that would be agonizingly painful to calculate throughout history.
   259. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 18, 2007 at 12:44 AM (#2581874)
I strongly disagree with simply multiplying innings by leverage if you are not going to also account for "chaining" (see http://www.hardballtimes.com/main/article/jonathan-papelbon-and-replacement-level). No manager would ever put a replacement player in as his closer, he simply promotes his setup man. You are vastly overrating relievers--by over 50%, according to this article--if you do one but not the other.
   260. Joey Numbaz (Scruff) Posted: October 18, 2007 at 12:45 AM (#2581878)
I find that very hard to believe (that I'm overrating relievers), but I will read the article.
   261. Joey Numbaz (Scruff) Posted: October 18, 2007 at 12:52 AM (#2581883)
Interesting article. I don't agree that the back end of a bullpen is a 100 ERA+, for one - especially not throughout history. So I think his conclusions are overstated. But it is interesting.
   262. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 18, 2007 at 01:00 AM (#2581887)
Tangotiger finds that replacement relievers, used as relievers, have a .470 winning percentage, so 94 RA+. That's certainly what I'd use.
   263. 'zop sympathizes with the wrong ####### people Posted: October 18, 2007 at 01:24 AM (#2581902)
I strongly disagree with using standard deviation of starting pitcher IP, because once you get below the top X number of pitchers (X probably being something like the number of teams in the league), innings pitched is controlled by the quality of the pitcher as well as any global contextual starting pitcher usage.

Or, to put it another way, pitchers who suck dont pitch as much as pitchers who are good.

This complication doesn't exist with position players, as a much higher percentage of regular players play full-season (exceptions being platoon guys, and the diff in PA b/w the top and bottom of the order). But anyways, the point is that IP correlates to pitcher quality, and that ##### everything up.
   264. Dr. Chaleeko Posted: October 18, 2007 at 01:52 AM (#2581918)
But anyways, the point is that IP correlates to pitcher quality, and that ##### everything up.

And I wonder if IP also varies with style of pitching as well. Some styles are less efficient, force a pitcher to throw more pitches, and may cause him to lose innings as a result: high K/high walk/flyball pitchers who can't field their position and don't hold runners well for instance.

But more efficient pitchers (sinkerballers who don't walk or K anyone, field their position well, and have a great move) may well have an advantage over their less crafty brethren, despite not necessarily being as generally effective.

Ryan and Clemens are freaks of nature, but are TJ and Kaat, or are they simply efficient pitchers?

So quality mostly, perhaps some style as well? I also wonder if home park might have a noticable effect. If you pitch in Safeco or Petco, are you likely to toss more innings than an identical hurler who does his work in Arlington or in Cincy? Once you get into the late 20th/21st century, and the unbalanced schedule, I wonder if your division may even have a very slight impact. If you pitch in the AL East, the land of the big offenses, you may hurl fewer innings. But if you pitch for Minnesota or Cleveland in the Central, facing KC and those awful Tiger teams 19 times a year, you're catching a lot of bad offensive teams. And if you pitch in the AL West, you get two good pitchers parks 20 times a year (and Arlington...).

Anyway, just interesting to think about.
   265. Joey Numbaz (Scruff) Posted: October 18, 2007 at 02:21 AM (#2581941)
Dan, there is about zero chance that a 94 ERA+ (equivalent) replacement level for relievers applies throughout history.
   266. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 18, 2007 at 02:25 AM (#2581945)
OK, Joe, but the point holds that simply multiplying innings by leverage and then just comparing to a replacement reliever is w-r-o-n-g. No true replacement player in history has ever pitched in a situation with a LI over 2. The correct approach--I feel VERY strongly about this--is:

1. Create a theoretical average bullpen, much like the theoretical lineup we use to calculate runs/wins above average for hitters, with estimated leverage for each role.
2. For the pitcher in question, slot him into the role appropriate for his leverage, and calculate the leveraged winning percentage of the Average Bullpen Plus Player.
3. Remove the player in question, and promote everyone else up one role. Insert a .470 WPCT pitcher into the mopup role, and calculate the leveraged winning percentage of the Average Bullpen Plus Replacement Player.
4. Subtract 3 from 2, and multiply the difference by the fraction of the team's innings pitched by the bullpen and then by 162 to get wins above replacement.

Here's an example. I am just guessing these numbers. It is very important to get the assumptions right, and also to adjust them as you go backwards in time, if anyone is inspired to do the research.

Average bullpen

PitcherIP LevI LIP WPCT
Closer
-- 70 2.00 140 .600
Setup
--- 80 1.60 128 .570
LOOGY
--- 45 1.35 061 .520
Middle1
60 1.15 069 .495
Middle2
60 1.00 060 .485
Middle3
50 0.85 043 .480
Long
---- 80 0.65 052 .475
Mopup
--- 55 0.30 017 .470
TOTAL
-- 500 1.14 569 .536 


OK, now let's put a super-duper closer on the team. If we increase the closer winning percentage from .600 to .700 and leave everything else the same, the leveraged winning percentage jumps to .560.

On to the replacement. If we replace the closer and move everyone up a role (except the LOOGY who is untouched), we get:

PitcherIP LevI LIP WPCT
Closer
-- 70 2.00 140 .570
Setup
--- 80 1.60 128 .503
LOOGY
--- 45 1.35 061 .520
Middle1
60 1.15 069 .484
Middle2
60 1.00 060 .479
Middle3
50 0.85 043 .475
Long
---- 80 0.65 052 .471
Mopup
--- 55 0.30 017 .470
TOTAL
-- 500 1.14 569 .511 


So, our super-duper closer is actually worth (.560-.511)*162*500/1458 = 2.7 wins above replacement. If you just went by winning pct minus .470 times leveraged, you'd get (.700-.470)*162*140/1458 = 3.6 wins above replacement. That seems about the right size correction to me--25%.

Note that the size of the correction gets smaller as you move down the leverage chart. If I repeat the same exercise with the Middle 2 guy, the correction is only 3.4%.
   267. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 18, 2007 at 02:28 AM (#2581947)
As I said, Joe, this would DEFINITELY need to be adjusted for era to reflect bullpen usage. And I have no idea how you'd do that before leverage statistics are available, even if you had the time. But the point holds--if you multiply for leverage and don't factor in chaining, you are *substantially* overrating ace relievers.
   268. rawagman Posted: October 18, 2007 at 02:42 AM (#2581960)
Well - this gets right to the heart of what exactly is entailed in a pitcher's workload. Not all Ip's are created equal. Far from it. An IP should only be seen as a measurement of how many outs a particualr pitcher was on the mound for. As far as workload is concerned, it makes much more sense to look at pitch counts. I don't think a straight pitch count will tell the whole picture, but I also don't think there is any one way to measure it in terms of abuse as a catch all for pitchers. Each pitcher tires according to his own context.
Since returning to Toronto in June, I managed to watch the vast majority of the Jays' second half games. And that reminded me - there's only so much we can learn from the numbers - sometimes you just have to watch the games.
The Jays this year had a pretty young rotation, all but Burnett were trained in the system. The four true Jays (Halladay, McGowan, Marcum and Litsch) were all trained to pitch to contact. Burnett is a strikeout artist.
Halladay - 5.55 K/9. 1.91 BB/9. ERA+ 125/225.3IP
McGowan - 7.73 K/9 3.27 BB/9. ERA+ 114/169.7IP
Marcum - 6.91 K/9 2.77 BB/9. ERA+ 113/159IP
Litsch - 4.05 K/9 2.92 BB/9. ERA+ 122/111IP

Burnett - 9.56 K/9 3.59 BB/9. ERA+ 124/165.7IP

That's a pretty effective rotation by any measure. Other than Litsch, all are legitimate starting pitchers, and all pitched alot of the time against the highly touted offenses of the AL (ie. Red Sox and Yankees). IP is not a good indicator for any of them because, 1) Halladay missed 3 starts with appendicitis. 2) McGowan, Marcum and Litsch did not begin the season in the rotation 3) Burnett was/is/will remain being injury prone.
I'll focus on Halladay - Roy Halladay has finished in the top 4 in the AL in IP 4 times in the last 6 years for a few reasons:
1) He pitches to contact. His K numbers have dropped in that time and this year's 125 was his lowest ERA+ in a full year in his career. But they haven't dropped too much. He does not rely on striking people out. He has been lucky enough to have a solid infield behind him most of the time and he trusts them to do their job. I'd love his pitches per inning count as compared to league averages to prove this hypothesis.
2) He is in excellent physical shape. He has a long distance running background. He does not tire late in games. Ever. In memory, he has ptiched two 10 inning complete games. How many other pitchers have done that in the last 10 years? When he doesn't pitch more than 7 innings, it is due to effectiveness, not tiredness. The two years he didn't finish in the top 4 in IP, he was injured. Once was a broken leg (took a line drive off the shin) and he had shoulder problems in 2004.
3) He has very good location. 2.09 UIBB/9 for his career.
4) Main pitch is a sinker which forces alot of groundouts. Keeps him away from big innings.

So I think this helps to identify heavy factors in pitcher types/characteristics that would be conducive to heavy workloads with minimal injury risk.
Be in top shape. Keep the ball down and in the strike zone. Trust your defense.
   269. Howie Menckel Posted: October 18, 2007 at 04:04 AM (#2581989)
nice work, rawagaman

I have a ton of interest in K rate in fantasy ball, which has in general been a quality indicator of future success.

I consider K rate to be of minimal value in this retrospective project. Zero to start, but plus avoiding errors that contribute to runs, more or less.

I'll take any effective pitcher, don't care how he gets there (ok, factor in his defensive help, yes)
   270. Paul Wendt Posted: October 18, 2007 at 01:01 PM (#2582119)
241. jimd Posted: October 16, 2007 at 08:35 PM (#2580130)
> WS system limits WSs in an extreme way for guys who played for big losers, as Johnson did and gives big benefits for guys who consistently played on winning teams, like IIRC, Belle did.

It's a strength of schedule effect within a league,


In this forum TomH focused on another effect or two, I think.
Part may be that runs do not produce wins at a constant rate but at a low rate for extreme teams and a "high" rate for average teams. "High" is highlighted because it may be equivalent to what we consider the normal rate, thanks to the overwhelming numbers of average teams (eg all 30 teams in 2007) and the nonlinearity of the effect (order of magnitude smaller at .400 than at .300).
   271. Mike Green Posted: October 18, 2007 at 01:59 PM (#2582181)
The innings translation is just a mess no matter which way you slice it. Take the 1876 NL season. In an eight team league, Jim Devlin threw 600+ innings leading the league by a fair shot. The game was so different from the game that we know (K/HR/E rates) that saying he threw the equivalent of 280, 300, 320, 340, 360 or 380 1985-equivalent innings is pretty much meaningless as far as I can tell. It is not quite the same for Walter Johnson's 1912/13 seasons, but the conditions were sufficiently different from the post-deadball era that the range of equally plausible innings translations is quite large.
   272. Mike Webber Posted: October 18, 2007 at 02:14 PM (#2582205)
Hi,

I think I am going to meet Jimmy Wynn next Thursday, have we sent/gave him "The Santo Letter" yet?

Relive the Rivalry at NLBM - Royals and Yankees in the 1970's
   273. Joey Numbaz (Scruff) Posted: October 18, 2007 at 02:18 PM (#2582210)
We haven't Mike - do you know somewhere we could send it.

The only one we ever sent was to Santo, and I never heard anything back from him.
   274. sunnyday2 Posted: October 18, 2007 at 02:39 PM (#2582238)
Maybe Mike could carry it in and give it to him by hand.
   275. Dr. Chaleeko Posted: October 18, 2007 at 02:41 PM (#2582240)
Joe,

We should send it to Mike, so he can give hand deliver it to Wynn. Not a big ceremony, just "here's something a few guys wanted to pass along to you with their highest regards." Hell, here's a draft to get us started if need be:

Dear Mr. Wynn,

The Baseball Hall of Merit is an online organziation of approximately 60 highly informed baseball fans. Our mission is to recognize baseball's greatest players, as we see them, and in so doing to provide a more rigorous alternative to the Baseball Hall of Fame in Cooperstown.

We wish to congratulate you on your election to the Baseball Hall of Merit. Our elections began with players who retired in 1892 or earlier, moving chronologically forward in the same fashion as the Cooperstown Hall of Fame. You were elected in 198?, your Xth year of eligibility. Our voters were keenly impressed with your combination of speed, power, and plate discipline, as well as with your numerous All-Star and MVP-type seasons. We have created an online plaque to memorialize your election, and it is located at www.baseballthinkfactory.org/files/hall_of_merit/xxxxxxxxxxxx.

In electing you to the Hall of Merit, we recognize you as one of the greatest 230 retired players to ever wear a Major League Baseball uniform, one of approximately the best 1% of the 16,000 men who have suited up. And while our voice is a small one, we hope that our deliberations may create a ripple effect and help your case with Cooperstown's electors; our elections suggest your omission is one they should strongly reconsider.

Congratulations once again from all of us at the Baseball Hall of Merit.

Yours sincerely,

XXXXXXXX
   276. sunnyday2 Posted: October 18, 2007 at 02:59 PM (#2582261)
Do it!
   277. Joey Numbaz (Scruff) Posted: October 18, 2007 at 03:35 PM (#2582306)
That's a great letter Dr. C. I'm definitely OK with giving that to Mike. Mike, if you can, drop me your address off line, so I can sign it and mail to you . . .
   278. Mike Webber Posted: October 18, 2007 at 03:43 PM (#2582315)
Sent it to you via BBTF email.
I thought it would be cool to drop it off in person.
   279. 'zop sympathizes with the wrong ####### people Posted: October 18, 2007 at 04:34 PM (#2582403)
I'd love his pitches per inning count as compared to league averages to prove this hypothesis.

Isn't that just [(P/PA)*BFP]/IP?

That's easy to get (P/PA is on THT and ESPN, the rest you can get anywhere)
   280. rawagman Posted: October 18, 2007 at 04:50 PM (#2582421)
'zop - thanks. I didn't know that.
Just looking at Halladay's numbers, since 2002, he averaged between 3.39 ('03) and 3.66 ('04) P/PA. Between 13.5 ('05) and 15.5 ('04 again) P/IP. 2004, his high water mark year for both, was also his worst year since proving himself as a regular.
Which all brought me to the realization of a single stat that probably should be used when trying to determine pitcher effectiveness. IP, on its own, is not very informative. Batters faced per inning, OTOH, is. That lets us know how effective said pitcher was. I suppose you could remove IBB, and maybe errors, but if you could combine this with the number of outs made (or IP) you would have a great way of measuring effective within the context of durability.
   281. jimd Posted: October 18, 2007 at 09:48 PM (#2582913)
You were elected in 198?, your Xth year of eligibility.

Wynn was elected in 1996, his 14th year of eligibility.
   282. John (You Can Call Me Grandma) Murphy Posted: October 18, 2007 at 11:01 PM (#2582984)
That's a great letter Dr. C. I'm definitely OK with giving that to Mike. Mike, if you can, drop me your address off line, so I can sign it and mail to you . . .


Hopefully he will respond to it, too.
   283. KJOK Posted: October 18, 2007 at 11:43 PM (#2583023)
The innings translation is just a mess no matter which way you slice it. Take the 1876 NL season. In an eight team league, Jim Devlin threw 600+ innings leading the league by a fair shot. The game was so different from the game that we know (K/HR/E rates) that saying he threw the equivalent of 280, 300, 320, 340, 360 or 380 1985-equivalent innings is pretty much meaningless as far as I can tell. It is not quite the same for Walter Johnson's 1912/13 seasons, but the conditions were sufficiently different from the post-deadball era that the range of equally plausible innings translations is quite large.


I agree. When comparing pitchers, I leave innings alone, and make all the adjustments at the 'value' level at the end - giving modern pitchers credit for around 70% of run prevention, and 1876 pitchers something far less.
   284. Joey Numbaz (Scruff) Posted: October 19, 2007 at 04:53 AM (#2583995)
I'm probably going to leave out the 14th year of eligibility and just say he was elected in 1996. Any objections?
   285. Joey Numbaz (Scruff) Posted: October 19, 2007 at 04:54 AM (#2584000)
Interesting approach KJOK - I can definitely see the positives of that approach. How to do figure out what the % of run prevention is in a particular season?
   286. KJOK Posted: October 19, 2007 at 05:07 AM (#2584010)
Interesting approach KJOK - I can definitely see the positives of that approach. How to do figure out what the % of run prevention is in a particular season?


Well, what I do mainly is 'educated guess', doing what would resemble a bit of timelining from current back to 1876, only it's not really timelining, it's just trying to estimate pitcher vs. fielder credit. What I would hope is someone who has a really good pitcher evaluation methodology, like you or Dan R., would be able to figure out something a little more solid, probably based on balls in play for each team/season, that would better apportion run prevention for each team pitching staff, or maybe even each individual pitcher, in history?!?
   287. Joey Numbaz (Scruff) Posted: October 19, 2007 at 05:17 AM (#2584021)
Heh - while I appreciate the compliment, my system is based on runs, I don't even track peripheral data.

I'm not a huge fan of tying it to BIP. While I see the general trend of more BIP more help needed by fielders - I still think pitchers have a much higher influence on the result of BIP than most statheads . . . I'm from the Emeigh school when it comes to DiPs; which generally says there's a lot more to it, due to natural selection (pitchers that give up a lot of hits on BIP don't last very long).
   288. Joey Numbaz (Scruff) Posted: October 19, 2007 at 05:18 AM (#2584022)
Don't get me wrong there, I still think fielders can have a huge influence on what looks like a pitcher's record (see Palmer, Jim) but I'm still not sure about tying everything to BIP.
   289. Dr. Chaleeko Posted: October 19, 2007 at 01:23 PM (#2584156)
I'm probably going to leave out the 14th year of eligibility and just say he was elected in 1996. Any objections?

Looks good to me. Here's the text of the letter with this and the year correctly filled in:

Dear Mr. Wynn,

The Baseball Hall of Merit is an online organziation of approximately 60 highly informed baseball fans. Our mission is to recognize baseball's greatest players, as we see them, and in so doing to provide a more rigorous alternative to the Baseball Hall of Fame in Cooperstown.

We wish to congratulate you on your election to the Baseball Hall of Merit. Our elections began with players who retired in 1892 or earlier, moving chronologically forward in the same fashion as the Cooperstown Hall of Fame. You were elected in 1996. Our voters were keenly impressed with your combination of speed, power, and plate discipline, as well as with your numerous All-Star and MVP-type seasons. We have created an online plaque to memorialize your election, and it is located at http://www.baseballthinkfactory.org/files/hall_of_merit/discussion/the_hall_of_fame_plaque_room_s_z.

In electing you to the Hall of Merit, we recognize you as one of the greatest 230 retired players to ever wear a Major League Baseball uniform, one of approximately the best 1% of the 16,000 men who have suited up. And while our voice is a small one, we hope that our deliberations may create a ripple effect and help your case with Cooperstown's electors; our elections suggest your omission is one they should strongly reconsider.

Congratulations once again from all of us at the Baseball Hall of Merit.

Yours sincerely,


XXXXXXXXXXXXXx
   290. Dr. Chaleeko Posted: October 19, 2007 at 01:27 PM (#2584163)
Joe,

I just sent you a copy of the letter (with organization spelled correctly), via email as a Word .doc file. Feel free to play around with it as you need. That long URL won't fit on one line, you might want to ensmallen the font to get it onto one line.

Anyway, it's ready to go. Just sign it, and mail it to Mike. Or if you have your signature scanned and on file, plop it in, and email it to Mike to print and hand off!
   291. Joey Numbaz (Scruff) Posted: October 19, 2007 at 01:53 PM (#2584193)
I don't have my signature scanned . . . good idea though.

I will print it on nice resume paper when I get home tonight and it will be in the mail this evening. Thanks a bunch!

I suppose we could start sending more of these if we can get addresses . . .
   292. Cblau Posted: October 20, 2007 at 01:40 AM (#2584797)
The problem with that letter to Wynn is it doesn't explain that the election actually took place this year, not 1996. Probably should say it was the election of players who retired in or before 1990.
   293. Joey Numbaz (Scruff) Posted: October 20, 2007 at 03:56 PM (#2584975)
Good point cblau . . . I already mailed the letter - Mike, do you think that's something you'll get a chance to explain - I could mail another letter also . . . changing it to something like:

"You were elected in our '1996' election (which actually took place in March of this year)."

Does that work?
   294. Paul Wendt Posted: October 21, 2007 at 03:43 AM (#2586058)
Hopefully he will respond to it, too.

Depending on the nature of the meeting and how it goes, I think he is likely to respond to Mike W. in person.
   295. Dr. Chaleeko Posted: October 21, 2007 at 01:19 PM (#2586158)
I doubt he'd say much other than "Thank You." I mean what's he going to say? 60 seamheads in "room" isn't nearly so flattering to him as the HOF's various election sqauds.

But that said, it beats not getting any letter.

Anyway, agreed with Joe that we ought to send all living HOMers letters, and I'll add that we should also send them a complete list of all the players we've elected to give them some context (particularly the HOM-not-HOF guys). We might want to designate on this list when each guy played and whether he was an MLB guy or a NgL guy. I'd bet that plenty of players don't have a lot of historical background, and a little scaffolding never hurt anyone.

I think we should do it after we wrap up our final current-cycle election (be that 2008 or 2012).

Depending on whether we have an interest in promoting this project, we might look to see if they are represented by someone and send the letter to that person also.
   296. sunnyday2 Posted: October 21, 2007 at 05:16 PM (#2586285)
I was looking at my voting history and just for the hell of it I compiled my top streaks.

1. Hughie Jennings--voted for him for 53 straight years (from 1908 until he was elected in 1960)
2. Rube Wadell--49 (1938-elected 1986)
2. Pete Browning--49 (1957-elected 2005)
4. Tommy Bond--48 (1927-1974)
5. Addie Joss 45 (1956-2000)
6. George Sisler 44 (1936-elected 1979)
6. Dobie Moore 44 (1948-elected 1991)
8. Ed Williamson 39 (1913-1951)
9. Jose Mendez 34 (1952-elected 1985)
10. Larry Doyle 32 (1975-still counting in 2006)

11. Charley Jones 30 (1913-1942)
12. Sam Thompson 28 (1902-elected 1929)
13. Ralph Kiner 27 (1961-elected 1987)
14. Dickey Pearce 25 (1907-elected 1931)
15. Charlie Bennett 23 (1899-elected 1921)
15. Edd Roush 23 (1975-elected 1997)
17. Harry Wright 21 (1906-1926)
18. Cupid Childs 20 (1920-1939)
19. Lip Pike 19 (1898-1916)
19. Willard Brown 19 (1958-elected 1976)

3 streaks--Williamson 39 (1913-51), 15 (1956-70), 16 (1991-2006 and counting), total 70 years

2 streaks--Pike 19 (1898-1916), 16 (1925-40), total 35
Bond 11 (1907-17) and 48 (1927-74), total 59
Roush 12 (1945-56) and 23 (1975-1997), total 35

Voted Most Times

1. Williamson 70
2. Bond 59
3. Jennings 53
4. Waddell and Browning 49
Page 3 of 3 pages  < 1 2 3

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Sheer Tim Foli
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.1060 seconds
49 querie(s) executed