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

Wednesday, September 18, 2002

Pitchers for the Hall of Merit

Let’s start discussing the pitchers here. I don’t have any adjusted numbers to post yet, but there’s no reason we can get the discussion cranking.

I take that back. I went through season by season a ways back and came up with pythagorean W-L records for each pitcher, based on his ERA vs. park adjusted league (season by season), adjusting for an average number of decisions in each season (based on the pitcher’s career IP/dec ratio for his career). Those numbers will be in the extended text.



John Clarkson 327-178 (361 Fibonacci wins)
Tim Keefe 351-215 (353)
Old Hoss Radbourn 304-200 (288)
Amos Rusie 261-158 (267)
Al Spalding 216-107 (254)
Tony Mullane 293-211 (252) .464 as a hitter too.
Pud Galvin 362-306 (252)
Jim McCormick 279-200 (241)
Mickey Welch 294-226 (234)
Will White 238-157 (225)
Sliver King 217-139 (212)
Jack Stivetts 202-134 (190)
Bob Caruthers 193-124 (186) .668 as a hitter.
Charlie Buffinton 220-178 (179)
Larry Corcoran 166-100 (170)
Guy Hecker 187-132 (164) .562 as a hitter.
Tommy Bond 220-178 (164)
Sadie McMahon 177-123 (158)
Candy Cummings 150-89 (156)
Bill Hutchinson 195-152 (153)
Ed Morris 171-122 (149)
Monte Ward 158-108 (144) .594 as a hitter.
Jim Whitney 210-185 (136) .548 as a hitter.
Dave Foutz 131-82 (130) .542 as a hitter.

Joey Numbaz (Scruff) Posted: September 18, 2002 at 05:16 PM | 571 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 6 of 6 pages  < 1 2 3 4 5 6
   501. Paul Wendt Posted: February 07, 2009 at 01:06 AM (#3069872)
One evolution that didn't happen.

According to a Base Ball Note in one distant newspaper:
Early in the 1887 season, with run scoring up thanks to strike four and some restriction on the pitching motion, St Louis AA owner Chris Von der Ahe suggested that teams must soon use pitchers in both left and right fields in order to prepare for all the substitute pitching that would be necessary.

In fact, Von der Ahe's team was on the fringe in using one pitcher regularly in right field. They were merely in the middle of the pack with 7 incomplete pitching games; every one of eight AA teams finished with 6 to 9 incomplete games. Eight NL teams finished with 2 to 9 incomplete games (Washington and Pittsburgh two each).

I don't know that captain/manager Charlie Comiskey felt any necessity to make even one pitching change simply because his pitchers were giving up more runs. I do know that Dave Foutz was worn out during spring training, alias another world's championship series with the Chicago White Stockings. Bob Caruthers did not practice at all; Nat Hudson held out (another holdout?); they scrambled and came up with Silver King.

I have read some newspaper references to ridiculous called strikes and to umpirical refusal to call obvious illegal deliveries. So I suspect that the executive branch softened the impact of the legislation --not necessarily by any presidential authority, maybe umpirical judgment about what's right.
   502. Paul Wendt Posted: February 07, 2009 at 01:22 AM (#3069880)
Perhaps they also saw the results the Cubs and Pirates were getting by using each pitcher less, and moved to copy someone else's successful model.

Yes, that's possible. Generally the dramatic change 1904-1909 that I mentioned may well be a time of change especially by closing the differences among teams. I am looking closely at aggregate statistics for 1904 to 1960 because it is a long period with stability in numbers of teams and team games. The period may open with extraordinary great differences among teams --maybe greater differences than immediately before 1904, too, because 1904 presented a problem to the teams. The two leagues expanded their schedules to 154 games (NL for the third time). It appears that Griffith and McGraw gave the extra games to their top pitchers. 154 games! Will 14 more make you Happy, Jack?

The dramatic 1904-1909 net change may have happened suddenly on many or most teams. Here is a sudden one.

pitcher starts distribution, Boston NL 1904-1909
43 38 36 21 15 2
42 41 38 27 4 2 1 1
41 37 35 32 3 3 1

32 31 28 25 16 12 3 2 2 1 (1907, Fred Tenney again, but they traded him to the Giants anyway)
31 30 28 20 11 11 8 7 5 3 2
32 30 19 17 13 10 9 7 6 4 3 3 2
   503. Paul Wendt Posted: February 07, 2009 at 01:27 AM (#3069883)
Why isn't #502 displayed here?
Reloading displays nothing new but the marginal notes say "502 - 7:22 pm"
   504. Paul Wendt Posted: February 07, 2009 at 01:30 AM (#3069886)
Hey, I meant to note

41 37 35 32 3 3 1 (1906, "Without Vic Willis we can finally go to town!")
   505. Paul Wendt Posted: February 10, 2009 at 03:42 PM (#3073053)
Let's see how this looks in everyone's favorite forum.

These listings may help understand the leading pitchers or the new statistics published in DT cards at BaseballProspectus. The background sets of pitchers are the top hundred and top fifty by career innings, with thresholds at 3800 and 3210 innings. The featured pitchers are everyone in those subsets who debuted before 1893 (18 and 11 in the two background subsets) and everyone who debuted during the 1980s (10 and 4 of them).

* marks Hall of Merit members who are on this week's agenda. Four of six (not Caruthers and Spalding) are among the hundred and all are among the fifty too.
o marks other Hall of Merit members

------ ------ represents DH=0 and
= represents a multiple of 50 hits above or below zero.

For example Bobby Mathews scores above 50, more hits than expected, and Young scores below -200.

>>
Delta-H (DH)
The number of hits above or below average for this pitcher, based on his own number of balls in play and his team's rate of hits (minus home runs) per ball in play; (H-HR) - BIP * (team (H-HR)/BIP). Essentially, the Voros McCracken number. For a team, Delta-H should be zero. Positive numbers signify more hits allowed than expected ("bad luck," if you believe pitchers have nothing to do with the outcome of a BIP), negative numbers mean fewer hits than expected ("good luck").
<< http://www.baseballprospectus.com/glossary/index.php?mode=viewstat&stat=45

The medians among 100 and 50 pitchers are DH = -65 and -46 hits.

ranks among 100 and 50 innings leaders: early pitchers

=
10 -5 Mathews
=
------ ------ (DH>0 for 21 of 100 pitchers; 7 of 50)
30 -- Terry
38 -- Bond
42 -- Griffith o
43 -- White
45 15 Welch
50 -- Buffinton
51 18 Radbourn *
=
53 20 Clarkson *
55 22 Weyhing
56 -- Whitney
70 29 McCormick
76 33 Nichols o
=
77 34 Galvin *
78 35 Keefe *
86 40 Mullane
=
90 -- Rusie o
=
97 47 Young o

One of the intermediate statistics, Balls In Play (BIP) is derived by subtracting homeruns, strikeouts, and free passes from the number of batters faced. That homeruns and strikeouts are not in play may be taken for granted today but does not fit early baseball very well. Between the two, strikeouts are more important. Strike three is a ball in play unless that would give the defense opportunity for a double play (compare the infield fly). When did the relatively routine play become one that might reasonably be taken for granted at the ballpark (we may have no alternative)? And how many "comebacks" to Al Spalding were routine plays?

Here is the same presentation for the pitchers with 1980s debuts, who are 10 of 100 and 4 of 50.

DH
ranks among 100 and 50 innings leaders: recent pitchers

=
=
15 -- Brown
17 -- Schilling
18 -- Wells
------ ------ (DH>0 for 21 of 100 pitchers; 7 of 50)
29 -9 Johnson
34 -- Smoltz
=
55 21 Glavine
64 -- Rogers
65 26 Maddux
74 -- Moyer
=
87 41 Clemens


By the way, although there are no extremes among 1980s debuts, or Clemens alone, the leaders and trailers are full of 1960s debuts, four of the top ten and five of the bottom ten among the hundred.
   506. Paul Wendt Posted: February 10, 2009 at 04:16 PM (#3073092)
[continued]

* marks Hall of Merit members who are on this week's agenda. Four of six (not Caruthers and Spalding) are among the hundred and all are among the fifty too.
o marks other Hall of Merit members

------ ------ represents zero and
------ represents a multiple of 50 runs above or below zero.

For example Jim Whitney scores above 50, more runs than expected (DR=98), and McCormick scores below -150.

>>
Delta-R (DR)
The number of runs, more or less, that a pitcher allowed, compared to his statistics. The pitcher's statistics (such as hits, walks, home runs) are run through a modified version of the equivalent runs formula to get estimated runs. Again, positive is "bad luck," negative is "good luck."
<< http://www.baseballprospectus.com/glossary/index.php?mode=viewstat&stat=46

The medians among 100 and 50 pitchers are DR = -33 and -37 runs.

ranks among 100 and 50 innings leaders: early pitchers

------
-2 -- Whitney
------
-7 -- Terry
11 -- Rusie o
12 -8 Keefe *
------ ------ (DR>0 for 33 of 100 pitchers; 13 of 50)
43 18 Clarkson *
53 24 Young o
63 29 Mathews
65 31 Nichols o
------
69 -- Griffith o
78 36 Weyhing
90 -- Buffinton
------
91 -- Bond
93 44 Radbourn *
94 45 Mullane
95 46 Galvin *
97 -- White
------
98 48 Welch
00 50 McCormick

The six pitchers with DR>50 are relatively recent: Ryan, the exception Whitney, Sutton, R.Johnson, Maddux, Hough. On the other hand seven of the ten pitchers below -100 runs are displayed here (debuts before 1893). The three others are Plank #99, Grove, and Willis.

Here is the same presentation for the pitchers with 1980s debuts, who are 10 of 100 and 4 of 50.

DR
ranks among 100 and 50 innings leaders: recent pitchers

------
_4 _3 R.Johnson
_5 _4 Maddux
------
18 -- Smoltz
22 11 Clemens
27 -- Schilling
28 -- Wells
------ ------ (DR>0 for 33 of 100 pitchers; 13 of 50)
36 -- K.Brown
41 -- K.Rogers
61 -- Moyer
------
86 43 Glavine

The first twelve all debuted during the 1880s (four) or the 1960s-80s. Grove is the youngest of the bottom ten and early rather than late 20th century pitchers have most of the scores below -50 runs.
   507. Paul Wendt Posted: February 10, 2009 at 04:47 PM (#3073128)
[continued]

* marks Hall of Merit members who are on this week's agenda. Four of six (not Caruthers and Spalding) are among the hundred and all are among the fifty too.
o marks other Hall of Merit members

------ ------ represents zero (located between zero and minus one)
------ represents a multiple of 5 wins above or below zero.

For example, Clark Griffith scores above 10, more wins than expected, and Tim Keefe scores below -20.

>>
Delta-W (DW)
The number of wins, more or less, that a pitcher won, compared to estimated wins. Estimated wins are derived from the pitcher's actual runs allowed and team average run scoring. Here, a positive number is "good luck," negative is "bad luck."
<< http://www.baseballprospectus.com/glossary/index.php?mode=viewstat&stat=47

The medians among 100 and 50 pitchers are DW = -2 and -1 wins.

ranks among 100 and 50 innings leaders: early pitchers

_5 -- Griffith o
------
10 _4 Mathews
------
24 _9 Galvin *
24 10 Welch
------ ------ (DW>=0 for 39 of 100 pitchers; 20 of 50)
44 25 McCormick
44 25 Clarkson *
44 -- Buffinton
50 -- Whitney
52 27 Mullane
54 30 Radbourn *
------
62 -- Bond
70 -- White
79 37 Weyhing
------
83 -- Rusie o
85 -- Terry
86 43 Young o
------
91 45 Nichols o
------
97 47 Keefe *

Here is the same presentation for the pitchers with 1980s debuts, who are 10 of 100 and 4 of 50.

DW
ranks among 100 and 50 innings leaders: recent pitchers

_1 -- K.Rogers
------
_3 _2 Glavine
_5 _3 Johnson
_5 -- Wells
_8 -- Moyer
------
38 19 Clemens
38 20 Maddux
------ ------ (DW>=0 for 39 of 100 pitchers; 20 of 50)
53 -- K.Brown
------
76 -- Smoltz
------
81 -- Schilling

The eight pitchers with 10 unexpected wins (DW>=10) are recent: five debuted in the 1980s, along with Joe Niekro, Griffith and Grimes. Eleven pitchers who score -15 wins or below include none so recent, Blyleven #2 and Eckersley alone with 1970s or 1960s debuts. Keefe #4 and Nichols debuted before 1893, which is proportional representation of that epoch. Here are all the "trailers".

DW =< -15 wins (eleven of the 100 innings leaders)
Spahn W, -16 wins
Nichols K
Friend B
Luque D
Powell J
Pierce B
Eckersley D
Keefe T
Ruffing R
Blyleven B
Johnson W, -26 wins
   508. Paul Wendt Posted: February 10, 2009 at 04:57 PM (#3073144)
The top twenty, down to DW=5 or 5 unexpected wins, include five 1960s and two 1970s debuts along with the five 1980s pitchers who score 10 wins.

number who score +5 wins by debut decade (top hundred innings leaders)

1870s: 1 0 2 ; 1 2 2 0 0 ; 0 5 2 5 0

career innings leaders by decade

1870s: 5 10 8 ; 7 10 8 4 4 ; 8 17 8 10 1 : top hundred
1870s: 3 6 4 ;; 4 5 3 2 2 ;; 2 10 5 4 0 :: top fifty
   509. Paul Wendt Posted: February 12, 2009 at 10:47 PM (#3075930)
#505 (DH)
>>
Delta-H (DH)
The number of hits above or below average for this pitcher, based on his own number of balls in play and his team's rate of hits (minus home runs) per ball in play; (H-HR) - BIP * (team (H-HR)/BIP). Essentially, the Voros McCracken number. For a team, Delta-H should be zero. Positive numbers signify more hits allowed than expected ("bad luck," if you believe pitchers have nothing to do with the outcome of a BIP), negative numbers mean fewer hits than expected ("good luck").
<< http://www.baseballprospectus.com/glossary/index.php?mode=viewstat&stat=45

The medians among 100 and 50 pitchers are DH = -65 and -46 hits.

vice versa.
Among the 50 career innings leaders, the median is DH = -65 hits. Among 100 innings leaders the median is -46 hits. Among 637 pitchers in my desktop table of Davenport statistics the median is -21 hits. (Those are not the top 637 pitchers by career innings but they include more than the top 150 and almost all are among the top 1000.)

The glossary description suggests that DH sums to zero at the team-season level. If so then generally negative scores (fewer hits than expected) for career innings leaders are balanced by their teammates somewhat positive scores.

Has anyone analysed DH to know whether the value is systematically different from zero for pitchers with very short careers? --or very short seasons?


#506 (DR)
>>
Delta-R (DR)
The number of runs, more or less, that a pitcher allowed, compared to his statistics. The pitcher's statistics (such as hits, walks, home runs) are run through a modified version of the equivalent runs formula to get estimated runs. Again, positive is "bad luck," negative is "good luck."
<< http://www.baseballprospectus.com/glossary/index.php?mode=viewstat&stat=46

Grove is the youngest of the bottom ten and early rather than late 20th century pitchers have most of the scores below -50 runs.

Among the 100 innings leaders early pitchers have most of the big negative scores that represent "good luck" in yielding runs. The most recent pitchers, especially the best of them, tend to have positive scores.

Evidently this measure sums to zero only by construction of the underlying "extrapolated runs" measure; that is, only for any population of pitchers where "extrapolated runs" is unbiased. That may be only the population of all major league pitchers.

Note. If there is no bias in extrapolated runs by date, then DR does measure something relative to a pitcher's contemporaries. Then there is a puzzle: why is the experience of early career innings leaders relative to their contemporaries so different from the experience of recent career innings leaders relative to their contemporaries?


#508 (DW)

>>
Delta-W (DW)
The number of wins, more or less, that a pitcher won, compared to estimated wins. Estimated wins are derived from the pitcher's actual runs allowed and team average run scoring. Here, a positive number is "good luck," negative is "bad luck."
<< http://www.baseballprospectus.com/glossary/index.php?mode=viewstat&stat=47

The very strong performance (more wins than expected) by recent pitchers among the 100 innings leaders suggest to me some association with the recent prevalence of relief pitching. Among other things, (a) a starting pitcher and a closer tend to score a win and a save when they both perform well; (b) closers and some other late-inning relief pitchers probably get more opportunities to lose games than to win them.

The early career innings leaders seem to have more than their share of low rankings, which mean fewer wins than expected. The rank-order data presented here is too crude for me to judge whether that is true, beyond what may be attributed to the opposite effect for the recent innings leaders.
   510. Paul Wendt Posted: March 07, 2009 at 05:16 PM (#3095324)
Moments ago I wrote this at Pitchers Ballot 1871-1892.
>>
Keefe also played for weaker teams than Radbourn did.

More important, if true, Providence for Radbourn (and Ward) was a better fielding team than was New York for Keefe (and Welch) --strikingly so according to the mid-aughts edition of Clay Davenport's analysis. From Dan Greenia I know that the DT cards are recently out in their late-aughts edition. I'm loathe to take a peek.
<<

The differences are radical, including relative differences in the revisions for Keefe and Radbourn.

XIP RA DH DR DW NRA RAA PRAA PRAR DERA IP
5093 4.40 -105 40 -22 3.67 470 436 957 3.73 5049 Keefe
2099 4.40 -105 40 -22 3.67 190 241 474 3.47 5049 Keefe, revised

XIP RA DH DR DW NRA RAA PRAA PRAR DERA IP
4525 4.51 -47 -111 -4 3.74 384 178 587 4.15 4535 Radbourn
1754 4.51 -47 -111 -4 3.74 135 247 442 3.23 4535 Radbourn, revised
   511. Paul Wendt Posted: March 07, 2009 at 05:29 PM (#3095341)
I have not looked for explanation of analytical changes at BaseballProspectus.com nor reread the glossary entries for the crucial statistics (XIP, RAA, PRAA, PRAR, DERA).

Previously I have noted that the estimates show almost all good career pitchers enjoyed the support of team fielding better than average, which was my interpretation of RAA > PRAA, commonly RAA >> PRAA.

At the moment all six pitchers on our agenda show RAA < PRAA, commonly RAA << PRAA. Worse yet, for stable interpretation, it appears that Radbourn and even Spalding have been re-evaluated relative to the others in this regard. Rather than most helped by good team fielding, they appear to be most hurt by bad team fielding (with Galvin).

For Keefe and Radbourn the crucial estimates repeated:

XIP RAA PRAA
5093 470 436 Keefe
2099 190 241 Keefe, revised

4525 384 178 Radbourn
1754 135 247 Radbourn, revised
   512. Paul Wendt Posted: March 16, 2009 at 04:50 AM (#3104316)
In "2010 Ballot Discussion" Chris Cobb reported the magnitudes of change in PRAA associated with the recent revision in the WARP rating system. (Those changes may be difficult to interpret as they are accompanied by large changes in XIP and RAA as well. See the examples just above.)

In #178 he concluded for now, in part,
... they have certainly made some significant adjustments in how they either a) assess fielding value or b) determine the split in responsibility between pitchers and fielders on a team by team basis.

On this latter point, the changes appear to me to move the system in the right direction. ...


Something seems quite wrong with the pair of revised statistics RAA and PRAA. Last week I illustrated my concern with Tim Keefe and Charley Radbourn. Note that Al Spalding, as much as Radbourn, now stands out as a pitcher who suffered with very poor team fielding support "Adjusted for Season". It's too much for me.
   513. Chris Cobb Posted: March 16, 2009 at 03:50 PM (#3104621)
Something seems quite wrong with the pair of revised statistics RAA and PRAA. Last week I illustrated my concern with Tim Keefe and Charley Radbourn. Note that Al Spalding, as much as Radbourn, now stands out as a pitcher who suffered with very poor team fielding support "Adjusted for Season". It's too much for me.

I would agree with Paul's concerns with respect, at least, to the pre-1893 pitchers. I don't have any confidence that these numbers are as accurate as the older WARP1 numbers, let alone more accurate. It's possible that there are still undetected bugs in the system that are distorting these results, or that the system simply produces bad results when applied to the conditions of the early game.

My comment that the changes appear to move in the right direction applies only to post-1920 pitchers. I don't see evidence that the 1893-1919 pitching era is being distorted, but I don't see changes that give me more confidence in the results, which I do see for more recent pitchers, where the changes reverse previous pitching evaluations that seemed significantly at variance with the "rawer" statistical record of the pitchers in question.
   514. Paul Wendt Posted: March 22, 2009 at 11:25 PM (#3110768)
1.
Chris Cobb, "Ranking the Hall of Merit Pitchers (1893-1923) - Discussion" #49
Then the WARP numbers changed drastically, including some systemic changes whose meaning I don’t yet understand and of whose validity I am dubious. So for now, I have decided to build my rankings around the average of my old home-grown system, my hybrid using using the old WARP numbers, and my hybrid system using the new WARP numbers.

I believe that's the way to go, if I understand correctly. That is, your third alternative or third component is "easy" to do by data entry and substitution, rather than modeling, design, or whatever you call the major intellectual work. There is too much unknown about the new WARP.

2.
Somewhere else Chris Cobb recently hoped that all would be explained in the print preseason guide from Baseball Prospectus. On my way out of town for three days, just returned, I looked at that volume standing in a bookstore. (That slowed my way out of town. At the time I planned to write this note when I reached my sister's home three nights ago, but I forgot about it, so this is a little more prone to error.)

Pages 618-19 by Clay Davenport provide some explanation of this major revision in the WARP system. I do not believe the explanation is adequate for HOM purposes but I may be wrong; those two pages are challenging, even if you sit down or you don't suffer from a heel bruise.

One focus of pages 618-19 is the definition of replacement-level fielding. There Davenport concedes a major error in his past approach (all previous editions, I infer), an error demonstrated for him by one inconvenient stylized fact; namely, the difference between average and replacement-level fielding is uniformly about 20 runs per team-season. Twenty runs in the 1890s or 1990s, at shortstop or firstbase. On the contrary, his past work (all WARP renditions?) implied or presumed larger margins between average and replacement in the 1890s, also larger margins at shortstop, with narrower margins in the 1990s or at firstbase. I didn't catch and retain whether that uniformity obtains only under his old definition of replacement level or whether it persists following revision but justifies revision on another point.

--
Somewhere they editors or Davenport alone say that they now provide WARP for active players in the main section of the volume, which is organized by team and provides some tabular data for all or most players. I looked quickly and found WARP only for batter-fielders. Maybe I missed it for pitchers.

Previously and concerning our batch of early pitchers (chiefly Keefe, Radbourn, and Spalding) I expressed my doubts and disbeliefs concerning some details of the revision. Regarding assessment as a system covering the whole historical range of major league pitching, 1871 to date, I feel that my concern is supported by several points in this note.
- The occasion to explain and justify revision is coverage of active players in the annual guide.
- That volume seems to use the new system for batter-fielders rather than pitchers. It may be that the impact on pitchers has languished as the residual effect of work whose focus is elsewhere.
- The cited scope of inconvenient facts, which force revision, is merely the 1890s to 1990s.

(I don't say the latter is wrong or inappropriate, only that is limitation complements my concerns based on early examples. It may be that all-time coverage is worse under the new edition yet abandoning that criterion is reasonable. If we aggregate the 1890s coarsely, everyone was using fielding gloves of some kind. The biggest steps in baseball organization had been taken. The changes from 1860s to 1890s were gross, to be followed by refinements.

Another caveat. After all, Davenport may have focused just as much on the 1870s as 1890s. The timespan 1890s to date roughly marks the epoch of reasonably uniform long schedules, which gives the "20 runs" gap somewhat more stable meaning than would any longer timespan.
   515. Paul Wendt Posted: April 17, 2009 at 04:06 PM (#3142248)
Here are DERA ratings for the major league pitchers in the first three chronological groups. 'DERA' is the 2008 edition and 'newnew' is the current edition. They are ordered by the current rating (column two).

1871-1892
In group one, Keefe is a big loser by 2009 revisions and Galvin is a big winner, as their ratings move up and down more than 20 points on the DERA scale. For the group the effect is strong convergence; they are now about equal by DERA.

DERA newnew
3.79 3.89 Clarkson J
4.20 3.96 Galvin J
3.73 3.96 Keefe T

4.15 4.02 Radbourn C
3.95 4.03 Caruthers B
4.13 4.06 Spalding A

Radbourn is no longer up there with Caruthers, Nichols, and Brown (ntm Spalding) as beneficiaries of all-time great team fielding. Indeed, his good score there now trails Clarkson and Keefe.
(More on Team Fielding ratings will follow.)

1893-1923
In group two there are five significant losers by the 2009 revisions: Waddell, Young, Plank, Nichols, and Brown in the range 0.14 to 0.10 (runs/9, namely PRAA per 9 XIP). Rusie is a huge winner at 0.35 which is the only bold revision.

DERA newnew
3.26 3.19 Johnson W
3.4 3.32 Alexander P
3.6 3.48 Walsh E
3.37 3.50 Young C
3.63 3.52 Coveleski S
3.88 3.53 Rusie A
3.59 3.63 Mathewson C
3.54 3.68 Waddell R
3.65 3.78 Plank E
3.95 3.84 McGinnity J
3.92 3.85 Griffith C
3.76 3.88 Nichols K
3.88 3.98 Brown Morde
3.97 4.00 Faber R
4.07 4.06 Rixey E

Nichols and Brown continue to be rated beneficiaries of all-time great team fielding. (more to follow)

1924-1958
In group three Ford is the significant loser at 0.13 runs and there seem to be four significant winners: Hubbell, Roberts, Spahn, and Lyons at 0.09 to 0.08. None is a bold revision.

DERA newnew
3.29 3.29 Grove L
3.50 3.49 Vance D
3.55 3.60 Newhouser H
3.84 3.75 Hubbell C
3.92 3.84 Roberts R
3.94 3.91 Ferrell W
4.00 3.92 Spahn W
4.02 3.94 Lyons T
3.93 3.97 Feller B
3.86 3.99 Ford W
4.01 4.07 Pierce B
4.28 4.23 Ruffing R
4.23 4.24 Lemon B
4.32 4.33 Wynn E

1959-19??
In group four Blyleven is a big loser at 0.18 runs with Gibson and Carlton also significant losers. Niekro is a huge winner like Rusie with Seaver and Jenkins the most significant others at 0.20 and 0.17 runs.

DERA newnew
3.65 3.45 Seaver T
3.53 3.49 Saberhagen B
3.55 3.50 Koufax S
3.73 3.61 Eckersley D
3.83 3.66 Jenkins F
4.02 3.68 Niekro P
3.72 3.85 Gibson B
4.01 3.89 Marichal J
3.91 3.91 Drysdale D
3.93 3.91 Stieb D
4.01 3.92 Perry G
4.03 3.93 Palmer J
3.79 3.97 Blyleven B
3.97 4.01 Bunning J
3.98 4.08 Carlton S
4.15 4.12 Ryan N
4.30 4.31 Sutton D

Palmer no longer rates with Nichols, Brown, & co. for all-time great team fielding. His good rating now trails Gibson in this group and it is not better than Sutton and Stieb. Niekro now rates in the opposite group, with team fielding estimated to be worse than league average by the margin 0.40 runs per game! The estimate for Jenkins may be extraordinary too. (I haven't seen enough ratings to judge. More to follow.)

--
Now I will let this rest a few days, both hoping to hear from Clay Davenport and planning to revisit the pages for some of the biggest winners and losers by revision, also some of the extreme revised values. See the Uberstats thread.
   516. Paul Wendt Posted: May 10, 2009 at 11:35 PM (#3173397)
Moments ago I posted a few comments and a little data on career OPS+ to the 1924-1958 Ballot thread. This version is revised only so that it does not depend on Howie Menckel's ballot remarks.

>>
[based on #60]
At career OPS+ 10, Dazzy Vance was merely a poor batter, not extremely bad, certainly no outlier.

Early Wynn was a good batter at OPS+ 54. That mark would have been good in Vance's time or Matty's; in the 1940s-50s it was excellent although Wynn too was no outlier.

At career OPS+ 19, Pierce was an average batter.

61. Paul Wendt Posted: May 10, 2009 at 07:16 PM (#3173382)
My desktop table of career pitcher data now includes 635 pitchers with at least 100 plate appearances. (That is not complete.) For 36 of them with debuts in the 1940s, median career OPS+ is 21 and Pierce is next.

For debuts in the 1910s (Vance) and 1920s (maybe a closer match to Vance's pitching career) the medians are OPS+ 39 and 23.

Among the 1920s debuts, 35 of 40 pitchers have career OPS+ in the range -3 to 62. Here are the five outliers.

100 Wes Ferrell
084 Red Lucas
081 Red Ruffing
-22 Si Johnson
-22 J Petty

Combining the 1910s and 1920s debuts provides 94 pitchers in my incomplete set. Median is OPS+ 28 {Rommel, Root, Ehmke}. The trio at OPS+ 10 {Vance, Faber, Toney} rank 75 to 77 which represents something like 18 to 21 percentiles. Among the group of pitchers at hand two rank slightly further down the list, Coveleski at 9 and Grove at 6.


Combining the 1930s and 1940s debuts provides 78 pitchers and a workable reference set for Pierce and Wynn (I hope). Median is OPS+ 21 {Parnell, Stobbs, Garcia}. The pair at OPS+ 19 {Pierce, Bridges} rank 43-44 and Wynn at OPS+ 54 ranks 8. Those ranks represent percentiles about 45 and 90.

Among the 1930s-1940s debuts, 72 of 78 pitchers have OPS+ in the range -13 (Maglie) to 69 (Walters). Perhaps the other six may reasonably be called outliers.

87 Schoolboy Rowe
85 Don Newcombe
82 Jim Tobin
82 Bob Lemon

-24 Preacher Roe
-41 Al Benton
   517. Paul Wendt Posted: May 10, 2009 at 11:37 PM (#3173399)
This version is revised only so that it does not depend on Howie Menckel's ballot remarks.

Excuse me. The closing paragraphs "Combining the 1930s and 1940s debuts" are new.
There should be a close quotation "<<" just above that point.
   518. Paul Wendt Posted: May 12, 2009 at 02:00 AM (#3175289)
Median OPS+ for pitchers by debut (double decade)
at least 100 plate appearances

debut number; OPS+(median)
1870/80s 88 ; 60 ; quartiles <39, 60, 92>
1880/90s 127; 56
1890/00s 105; 43
1900/10s 103; 36
1910/20s 94 ; 28 ; quartiles <16, 28, 53>
1920/30s 82 ; 23
1930/40s 78 ; 21
1940/50s 91 ; 15
1950/60s 128; +3 ; quartiles <-12, 3, 28>
1960/70s 121; +2 ; quartiles <-11, 2, 18>
1970/80s 99 ; -1 ; quartiles <-13, -1, 16>
1980/90s 85 ; -5 ; quartiles <-16, -5, 15>

For example consider the last line (bold).
I have 85 pitchers with 1980/90s debuts and at least 100 career plate appearances. Among them OPS+ = -5 is the median or 50th percentile. The 25th and 75th percentiles are -16 and +15.

The change from 1950s/60s debuts (also bold) is modest.


Median OPS+ for pitchers by debut double-decade
column 1: at least 500 plate appearances
column 2: at least 100 PA
column 3: linear model

62 60 50 : 1870s/80s
56 56 45
44 43 40
38 36 35
31 28 30 : 1910s/20s
24 23 25
25 21 20
21 15 15
+8 +3 10 : 1950/60s ; n=96 ; quartiles <-9, 8, 33>
+6 +2 +5
+5 -1 +0
+5 -5 -5

In this linear approximation (column three), career OPS+ declines by 5 points per decade from 50 at debut 1880 to 0 at debut 1980; hence -10 at debut 2000. That is generous chiefly to the most recent pitchers with many plate appearances (underline).
   519. Paul Wendt Posted: May 12, 2009 at 02:04 AM (#3175294)
I have 85 pitchers with 1980/90s debuts and at least 100 career plate appearances.

This data table is not complete down to 100 plate appearances in any decade. It happens to be the set of pitchers for whom I have DERA, newDERA, newnewDERA.
   520. Paul Wendt Posted: May 12, 2009 at 08:52 PM (#3176324)
I wrote three weeks ago #515
Now I will let this rest a few days, both hoping to hear from Clay Davenport and planning to revisit the pages for some of the biggest winners and losers by revision, also some of the extreme revised values. See the Uberstats thread.

For the 637 pitchers whom I have DERA and other Davenstats edition 2008, I have revisited the ten biggest winners and ten biggest losers, according to the 2009 revision in DERA; that is, the extreme values of the difference newnewDERA - DERA. There are no changes in three weeks so I guess that all of #515 stands.

Note that the tables in #515 are sorted by newnewDERA, the second column.

Here is the difference, the size of the revision on the DERA scale.
DERA difference (spring 2009 minus 2008)
difDERA nameFirst nameLast
-0.34 Phil Niekro
-0.20 Tom Seaver
-0.17 Fergie Jenkins
-0.12 Dennis Eckersley
-0.12 Juan Marichal
-0.10 Jim Palmer
-0.09 Gaylord Perry
-0.05 Sandy Koufax
-0.04 Bret Saberhagen
-0.03 Nolan Ryan
-0.03 Rich Gossage
-0.02 Dave Stieb
=0.00 Don Drysdale
+0.01 Don Sutton
+0.04 Jim Bunning
+0.07 Rollie Fingers
+0.10 Steve Carlton
+0.13 Bob Gibson
+0.18 Bert Blyleven

(Can someone point me toward deft management of the precision and the use of scientific notation, for purposes of copy and paste from MS Access query output?
For example, after naive copy and paste I have manually edited the entry for Don Sutton to +0.01 from
9.99999999999979E-03 Don Sutton.)
   521. Joey Numbaz (Scruff) Posted: May 13, 2009 at 12:57 PM (#3177355)
Paul my brain isn't working this early, I think you are saying Niekro's defenses just got a whole lot better and Blyleven's a whole lot worse? Those are some major changes on the fringe's and would definitely impact ranking of the pitchers.
   522. Paul Wendt Posted: May 14, 2009 at 12:30 AM (#3178534)
Yes, Joe, and #520 merely shows that change in estimates of team fielding support that underly the before and after measures of DERA in #515 --where I highlighted Niekro and Blyleven as big movers.

Clay Davenport evidently reacted to my report of a problem with the winter revision (the Advanced Pitching Stats for Cy Young were corrected next day). But he did not respond to my email inquiry about the new edition. Last night I returned to the big winners and big losers (by DERA, 2008 to the spring newnew edition), found no change, and decided that this edition is stable enough for me.

The winter new edition was ridiculous (#510-514). I should have emphasized last night that #515 features the 2008 and the spring newnew editions, that the latter is a big improvement over the winter farce, and that it seems stable. So #515 stands.

I regret the #515 layout with newnewDERA in column two.
   523. Paul Wendt Posted: May 14, 2009 at 12:44 AM (#3178541)
Aargh! Has anyone noticed that we have no edit feature here?
;-)
#520 merely shows that change in estimates of team fielding support that underly the before and after measures of DERA in #515 --where I highlighted Niekro and Blyleven as big movers.

#520 shows the change in DERA estimates, the differences between 2008 DERA and newnewDERA, or columns one and two in #515.

Here are the underlying estimates of runs saved "above average" for Phil Niekro only.

RAA PRAA inf [inferred estimate of team fielding contribution]
250 276 -026 Niekro, edition 2008
235 456 -221 Niekro, spring 2009 (newnew)

The 2008 estimates were 250 runs saved, allocated 276 runs to Niekro and -26 to team fielding. The newnew estimates are 235 runs saved, allocated 456 runs to Niekro and -221 runs to team fielding.

All right, here is Jim Palmer

RAA PRAA inf
336 198 +138 Palmer, 2008
299 217 +082 Palmer, spring 2009

For Palmer the revision is chiefly in column one, gross runs saved. The reallocation of runs to Palmer's pitching in particular, up from 198 to 217 runs, merely drops his DERA from 4.03 to 3.93. Compare the 34-point drop for Niekro from 4.02 to 3.68.
   524. Paul Wendt Posted: May 14, 2009 at 01:06 AM (#3178571)
Regarding the old knuckleball theme:

Among all 306 pitchers with mlb debuts after 1909 and at least 2000 career innings,
Phil Niekro, Wilbur Wood, and Emil Leonard rank 3-5-13 from the bottom by team fielding support, measured by (RAA - PRAA)/XIP in the current edition.

Charlie Hough, Tom Candiotti, and Tim Wakefield rank 46-70-79 from the bottom.
Hoyt Wilhelm and Joe Niekro rank 115 and 172 from the bottom, about percentiles 40 and 55.


By the way, here are the ranks and scores of the same eight pitchers by Delta-H ("DH", see #505), within the same set of 306 pitchers.

DH - name, rank among all 306 pitchers with 2000 innings, debut 1910 to date
-296 Hough, rank #1
-192 Wakefield, #4
-170 Phil Niekro, #8
-102 Wilhelm, #27
-083 Wilbur Wood, #44
-066 Candiotti, #60
-064 Emil Leonard, #64
-053 Joe Niekro, #83
   525. Paul Wendt Posted: May 14, 2009 at 01:10 AM (#3178577)
team fielding support, runs per 9 innings (RAA-PRAA)/XIP
(rank among 306 pitchers)
-0.40 Phil Niekro, #3
-0.36 Wilbur Wood, #5
-0.28 Emil Leonard, #13
-0.16 Charlie Hough, #46
-0.11 Candiotti, #70
-0.10 Wakefield, #79
-0.04 Hoyt Wilhelm, #115
+0.04 Joe Niekro, #172
   526. OCF Posted: May 14, 2009 at 01:25 AM (#3178590)
That does not look like a random effect. There are two possibilities.

1. Teams that are already poor defensively are more likely to employ knuckleballers. That's vaguely plausible, as a consequence of the idea that weaker teams in general tend to employ knuckleballers, since they've always been out of favor and unconventional, hence more likely to be used by a team with less to lose.

2. Knuckleballers by themselves, just by the way they pitch, cause measurable differences in evaluated defense. Do they give up more line drives but fewer HR? Are the PB leaking into the defensive stats in some way?

If #2 is true at all, then Davenport is doing something wrong here. An effect which is caused by the pitcher should be credited or charged to the pitcher, not laid off as part of his "support."

Here's another question: what happens with a pitcher's own fielding skills? For correct attribution, that should be part of who a pitcher is, and not part of his defensive support. The data on the other thread hinting that Gibson might have been better-supported on defense than Palmer (which I don't believe) is what gets me wondering about this - since Gibson was himself a very good defender.
   527. Paul Wendt Posted: May 14, 2009 at 01:27 AM (#3178593)
Among the leaders for team fielding support, among the same 306 pitchers (complete).

Dennis Leonard, Paul Splittorff, and Larry Gura rank #2-7-8. Were they teammates?

The top eight all have mlb debuts 1962-1974; Whitey Ford and Aaron Sele complete the top ten.

Someone asked about Tigers.
Milt Wilcox #11
Dan Petry #15
Jack Morris #24

Orioles,
McNally - Flanagan - Cuellar - Palmer, #6-22-37-40

Reds,
Billingham(Tigers), Gullett, Nolan, Grimsley(Orioles), Carroll
--without the career innings threshold (so the dataset is incomplete), all five of them are in the top 12 among 509 pitchers with mlb debuts 1910 to date. Jack Billingham, #1 at +0.53 runs, is the only one of the five Reds with 2000 innings.

The top 16 among 509 also include the three Royals;
Pirates Nelson Briles(Cardinals), Dock Ellis, and Bob Veale;
WWII Cardinals Mort Cooper and Max Lanier;
McNally
Johnny Murphy
Whitey Ford
   528. Crispix reaches boiling point with lackluster play Posted: May 14, 2009 at 01:30 AM (#3178595)
Very interesting, this issue of defensive support.
   529. Paul Wendt Posted: May 14, 2009 at 01:42 AM (#3178601)
Hi, OCF
Davenport ranks Nelson Briles #4 at +0.47 runs. --a Gibson teammate 1965-70.

something wrong?
add the fact that the eight leaders in my 306-pitcher group all share 1962-1974 debuts.

My dataset is complete for 2000 career innings back to 1871. There are 407 pitchers. Here are the 20 leaders by team fielding support, according to current Davenstats. There are eight with debuts 1962-74, Whitey Ford, and eleven with debuts before 1910.

tField
0.762951334379906 Spalding A
0.712230215827338 McBride D
0.533715925394548 Billingham J
0.524503311258278 Nichols K
0.479237909135320 Leonard Den
0.477379095163807 Briles N
0.463917525773196 Brown Morde
0.457473578258681 Ellis D
0.439868204283361 Grimsley R
0.434679334916865 Caruthers B
0.431221020092736 Phillippe D
0.431137724550898 Cuppy N
0.427685950413223 Cummings C
0.409594095940959 McNally D
0.391859914813062 Reulbach E
0.380549682875264 Splittorff P
0.372056514913658 Gura L
0.369663648124191 Ford W
0.364259028642590 King S
0.362720403022670 Stivetts J

The measure is (RAA-PRAA)/XIP*9.
(It's per nine innings but I sometimes forget to write "*9".)

Any qualms about XIP should extend to DERA.
This is the measure of team fielding runs per 9 that conforms with DERA.
   530. Paul Wendt Posted: May 14, 2009 at 02:30 AM (#3178626)
thoughts upon arrival home by bus:

I believe that what I call team fielding support includes the estimated contributions of nine fielders and the ballpark, insofar as the ballpark effects fielding.

Astroturf may be influential.
   531. Paul Wendt Posted: May 14, 2009 at 10:38 PM (#3179588)
We have sufficient evidence that there is a strong knuckleball effect in the new Davenstats.

Yet I checked some other prominent pitchers from a list of Knuckleball Pitchers that seems to rely heavily on Rob Neyer.

Mickey Haefner, Johnny Niggeling, and Roger Wolff, three teammates of Emil Leonard on the 1945 Washington Senators, all finished with very poor measured team fielding support, comparable to Leonard's.

Bob Purkey finished with average team fielding support, Eddie Cicotte and Eddie Fisher very good (close to +0.20 runs/9xip), and Gene Bearden super good (close to +0.40). Bearden worked 400 innings for Cleveland, then 400 for other teams, with all 34 of 34 career team fielding runs racked up in those first three seasons, then precisely average team fielding support. Bearden's entry in the Neyers/James Guide says that he learned the pitch at the beginning of this mlb career (1947).
   532. OCF Posted: May 14, 2009 at 11:07 PM (#3179634)
I believe that what I call team fielding support includes the estimated contributions of nine fielders and the ballpark, insofar as the ballpark effects fielding.

But if we want to use fielding support as an ingredient in making fair evaluations of pitchers, we want that to be the estimated contributions of eight fielders, not nine. If Greg Maddux was a significantly better fielder than Randy Johnson, then we don't want to discount that difference away from Maddux - it's part of who he was. It's probably true that on average left-handed pitchers are worse fielders than right-handed pitchers. I don't want to equalize that away. There are a whole array of reasons lefties have advantages over righties, or righties have advantages over lefties. And simply being right-handed or left-handed is part of a player's talent; it counts.
   533. Blackadder Posted: May 14, 2009 at 11:18 PM (#3179643)
This is really just accounting, but wouldn't it be best to determine a pitcher's defensive independent pitching, then attempt to independently determine his fielding contribution?
   534. Paul Wendt Posted: May 14, 2009 at 11:36 PM (#3179681)
I agree with OCF in #526 and #532 and I believe the differences among pitchers in fielding skills are big enough to be important here, especially in ranking the HOM pitchers.


I believe that what I call team fielding support
[--that is, (RAA-PRAA)/XIP] calculated from Davenstats--
includes the estimated contributions of nine fielders and the ballpark, insofar as the ballpark effects fielding.


That is, I think "PRAA" does mean pitching runs above average, not pitcher runs above average.

The BP glossary is ambiguous but generally supports this interpretation, I believe. The distinction between pitching and fielding seems deliberately drawn where the first word "pitcher-only" may be casual.

>>from the BP glossary
PRAA

Pitcher-only runs above average. The difference between this and RAA is that RAA is really a total defense statistic, and PRAA tries to isolate the pitching component from the fielding portion. It relies on the pitching/fielding breakdown being run for the team, league, and individual. The individual pitching + defense total is compared to a league average pitcher + team average defense, and the difference is win-adjusted.
<<
   535. Paul Wendt Posted: May 23, 2009 at 01:29 AM (#3190105)
Concerning team fielding support for pitchers in this group, Kevin Johnson wrote in 1959-1986 Discussion #54
I'm thinking DERA is not park-adjusted? If not, I'm guessing quite a bit of this is actually park effects instead of defense, as all of those first 4 guys played in pitcher's parks, while the bottom two played in extreme hitter's parks.

I replied in #56,
Kevin,
I don't believe so. This team fielding statistic is (RAA - PRAA) / XIP and DERA is calculated from the latter, PRAA/XIP. I believe your suggestion is really that RAA (runs allowed relative to average) is not park-adjusted; so that the difference RAA - PRAA includes estimates of both ballpark/schedule and team fielding effects. RAA and PRAA are one or two troubling black boxes; the key may be a surprise but I don't believe you have it here.


--
Rather than repeat that team fielding statistic for the pitchers on the ballot, here are the leaders and trailers among all 42 pitchers with 1950s/60s debuts and at least 2500 career innings.

tField name
0.41 McNally D (Cuellar and Palmer rank 8 and 9)
0.37 Ford W
0.26 Kaat J (Jim Perry ranks 10)
0.25 Burdette L
0.24 Hunter J
0.23 Gibson B
...
-0.19 Stottlemyre M
-0.24 Jenkins F
-0.27 Pascual C
-0.34 Coleman J
-0.36 Wood W
-0.40 Niekro P

And the same for threshold 1500 innings. Well, well, it seems to be Pirates and Reds at the head of the class.

tField name
0.53 Billingham J
0.53 Blass S
0.48 Briles N
0.46 Ellis D
0.45 Nolan G
0.43 Veale B
...
-0.36 Wood W
-0.38 Morton C
-0.40 Niekro P
-0.43 Johnson K (and which was your home ballpark, KJOK?)
-0.53 Farrell T
-0.56 Ramos P

Soon I will post a whole table again.
   536. Paul Wendt Posted: May 23, 2009 at 04:21 AM (#3190217)
Instead of fussing, here is a "whole" table of career data for 848 pitchers, available at YouSendIt for seven days.
http://www.yousendit.com/download/MnFnek9tRStTRTZGa1E9PQ

This includes the current edition of ten Davenstats in ten consecutive columns {newXIP, ... newDERA}.

There is some documentation at baseball-fever (see #26, 27). It concerns the preceding edition but the intensive scope is almost identical.
   537. Paul Wendt Posted: May 24, 2009 at 01:36 AM (#3190838)
In that table there is one new variable which identifies the Hall of Merit members in four special election groups {m1, m2, m3, m}, the next sixty major league pitchers selected and analysed by Chris Cobb in 2009 Ballot Discussion and above {cc}, and about a dozen more pitchers selected by me {c}.
   538. Paul Wendt Posted: May 24, 2009 at 05:17 PM (#3191240)
Joe Dimino, "Ranking ... 1959-1986" #60
I thought the difference between DERA and NRA was a team stat. Comparing Niekro to teammates that's definitely not the case.

http://www.baseballprospectus.com/dt//1981ATL-N.php

Paul, did this change with the new iteration? I'm almost positive when I was using this (before the updates this past winter) the difference was the same for teammates - which was why I was using it.



You may be right that this is one aspect of the change. During our discussion of meaning and "junk" status of these estimates, Dan Rosenheck reported thus (#307 above).
307. David Concepcion de la Desviacion Estandar (Dan R) Posted: October 20, 2008 at 04:06 PM (#2990308)
Hmm, I think you're right. Just looking at the 2002 Boston Red Sox, there's the same 8-run gap between Pedro and Lowe's RAA and PRAA, despite the fact that Pedro's K rate was a zillion times better. I stand corrected. Forgive me.

(Follow the link for context.)

Today I see this for the 2002 Boston Red Sox.
ADJUSTED FOR SEASON
 NRA  RAA PRAA PRAR  DERA  pitcher      ip    xip
2.62   55   50   87  2.80  Lowe      219.7  265.7
2.74   45   37   69  3.06  Martinez  199.3  229.6 


That shows better team fielding support for Martinez than for Lowe, same as you see for Perry and Niekro on the 1981 Atlanta Braves.
   539. Paul Wendt Posted: May 24, 2009 at 05:39 PM (#3191275)
Is that sufficient for you Joe?
I have career data only which is severely limiting.

There has been no revision in NRA. Here is the change in career DERA for some teammate pitchers.

difDERA
+0.10 Brown
+0.06 Reulbach

+0.13 Plank
+0.17 Bender

+0.13 Phillippe (48 -> 61)
+0.07 Leever (57 -> 64)

+0.00 Drysdale (54 -> 54)
-0.05 Koufax (43 -> 38)

Parenthetically those are before and after differences (DERA - RAA).
   540. Paul Wendt Posted: May 27, 2009 at 04:29 PM (#3195526)
Continuing re (RAA - PRAA)
Probably in "Pitchers" or "2009 Ballot Discussion", Chris Cobb reported for an earlier edition of Davenstats that the difference between RAA and PRAA --those runs above average evidently attributed to fielding rather than pitching, summed for all pitchers on a team-- is not equal to the sum of fielding runs above average for every player on a team. (We were discussing the relations among XIP, RAA, PRAA, and DERA, and which if any should be called "junk stats". I recall that Chris concluded with less than his previous confidence in DERA.)

Currently I find the following.

1906 Chicago Cubs
+150 = RAA - PRAA, sum for all pitchers
+193 = FRAA, sum for all fielders

1995 Atlanta Braves
-79 = RAA - PRAA, sum for all pitchers
-67 = FRAA, sum for all players

2001 Seattle Mariners
+238 = RAA - PRAA, sum for all pitchers
+205 = FRAA, sum for all players
   541. Paul Wendt Posted: September 22, 2009 at 02:57 PM (#3328855)
Regarding knuckleball pitchers and team fielding support estimated by Clay Davenport, hence DERA and WARP at BaseballProsectus (see especially #523-526, 531)

Tiboreau in his Combined Pitchers Ballot (Posted: September 22, 2009 at 02:53 AM (#3328689)
18. Phil Niekro—The beginning of a group of seven pitchers, both peak & career candidates, packed together quite tightly in a race to reach the top 20. Actually, at least at the time of the 1959 - Present pitchers ballot, BP appeared to absolutely love the knuckleballer; re-evaluation of BP's love affair to levels more similar to Joe Dimino's evaluation was, IMO, a more reasonable fit for Knucksie in the HoM Pitching puzzle.

Tiboreau,
Is this an allusion to some mid-2009 revision at BP? Or your revision?
   542. DL from MN Posted: September 22, 2009 at 09:40 PM (#3329292)
In 2009 BP WARP moved closer to the consensus replacement value.
   543. Joey Numbaz (Scruff) Posted: September 23, 2009 at 04:14 PM (#3329920)
All numbers I've posted on this site regarding pitchers are pre-2009 revisions.

I've always used runs allowed as my starting point, not earned runs. Then I've adjusted for team defense using DERA-NRA, which was always pretty much the same for every pitcher on a team. Now it appears that is no longer the case, so I'm not sure of what I can use as a proxy. I guess I could keep the old numbers?
   544. Tiboreau Posted: September 25, 2009 at 01:26 AM (#3331733)
Is this an allusion to some mid-2009 revision at BP? Or your revision?

The latter--including BP's numbers during the 1959 to Present pitchers vote meant that Niekro would be above Carlton, I think. So, essentially, I ignored BP's WARP & focused solely on Joe Dimino's data. The revision I refer to is simply dWAR appropriately doubled to account for the missing BP data to better compare him to the rest of the pitchers.
   545. Paul Wendt Posted: November 17, 2009 at 09:33 PM (#3389766)
This is the only one of the 2002 "Positional Threads" (so listed at "Important Links") that has remained active. So far, page 6 is nearly filled with notes by yours truly and chiefly concerned with the 2009 editions of "Advanced Pitching Statistics" by Clay Davenport at DT cards, BaseballProspectus. The key revised statistics are RAA, PRAA, and DERA (which are not independent), perhaps also XIP and PRAR. The last three pages represent just over one year's general and overall notes on pitchers by several of the people who have been active since the three-week cycles 1898-2007 have been succeed by annual elections and special projects.

CONTENTS
#1-268, 2002 to 2005
#269-300, January 2007
#303-5xx, October 2008 to date (pages 4-6)
   546. Paul Wendt Posted: November 17, 2009 at 09:44 PM (#3389781)
Newcomer 'epoc' says in "2010 Ballot Discussion" #286,
I'm a peak/prime voter. I use Rally's WAR for position players, and for pitchers I balance his WAR with WAR calculations of my own based on FIP vs. league average and pythagenpat.
...
13. Bob Friend - Another pitcher that DIPS really likes. Pitched really well in a tough era in the NL.


OCF close some remarks in #288: "OK, I get it - you're going by DIPS and not by ERA (or RA, which is what I use). I'm not sure I really trust that to justify lifting someone like Friend out of the group of his peers."

Ranking Bob Friend #13 is one way to get attention for a new rating system. I can't imagine that ranking is "right" yet I hope to learn more about that WAR based on FIP and/or DIPS.
   547. Paul Wendt Posted: November 20, 2009 at 08:48 PM (#3392774)
Probably I agree with DL from MN is probably right that the Hall of Merit is skimpy on pitchers. See "2010 Ballot Discussion" #314-316 (perhaps continuing there). See also the runups to previous annual elections above.

On the other hand, no specification such as 30% pitchers is innately plausible. Cooperstown honors more than 31% pitchers, the Hall of Merit more than 26%. But Cooperstown also honors a greater share of major league players from the last hundred years, when teams have employed more pitchers. The Hall of Merit honors more players from earlier times when fewer pitchers carried heavier workloads and teams employed fewer pitchers. There has been some change during those hundred seasons too.

For illustration only, suppose that 20% pitchers is about the right composition from pre-1893; about 25% from the "deadball era" and the Negro Leagues; about 30% for 1924-1958 or whatever; 35% for 1959-1984; 40% since then. What tolerance is appropriate for each period? What does all that imply for the aggregate composition of the Hall of Merit?

(My illustration puts all of the Negro Leagues with the deadball era because it does seem to me that NeL teams relied on a relatively small number of regular pitchers. The HOM has elected only Bret Saberhagen from the last period. The 2009 pitchers ranking project put him in group 4.)
   548. DL from MN Posted: November 20, 2009 at 09:05 PM (#3392811)
That's going to be hard to accomplish because we have a perpetually open ballot. It requires people to value pitchers based on their era much more than just using IP and ERA as a guide. That makes cross-era comparisons even tougher. It has been helpful that there has been a sliding replacement level across eras. I also believe that in some eras, pitchers were more "valuable" to the team than in others. Therefore it is possible that based on roster slots, you would expect 25% pitchers but there were a couple more or less guys in that group that were exceptional. Talent isn't distributed evenly. There is a dry spell in the 1980s and booms in the 70s and 90s for example.
   549. Paul Wendt Posted: December 01, 2009 at 12:27 AM (#3399578)
#547
For illustration only, suppose that 20% pitchers is about the right composition from pre-1893; about 25% from the "deadball era" and the Negro Leagues; about 30% for 1924-1958 or whatever; 35% for 1959-1984; 40% since then. What tolerance is appropriate for each period? What does all that imply for the aggregate composition of the Hall of Merit?

Because some tolerance is appropriate for each period, DL's criticism #548 is off the mark. Tolerance aside, however, it's only by reference to appropriate pitcher shares that vary in time --vary between time periods in my illustration-- that one may reasonably justify an appropriate aggregate pitcher share.

It is not reasonable to justify any aggregate share, such as 30%, by reference to modern roster composition only. DL makes an argument of that kind, in my reading. Although I agree with the conclusion that the Hall of Merit is "light" on pitchers, I am trying here #547 and elsewhere to warn against any argument from modern rosters as a point of reference.
   550. DL from MN Posted: December 01, 2009 at 12:47 AM (#3399593)
I don't even understand my own post fully. I think we're agreeing, sort of. I think # of starters in rotation is a better proxy than # of roster slots. Add a bump for ace relievers.
   551. Paul Wendt Posted: December 01, 2009 at 12:58 AM (#3399602)
Recently I have examined the current edition of Davenport's "Advanced Pitching Stats" for 100 major league pitchers. Those that vary from edition to edition now seem to be XIP, NRA, RAA, PRAA, PRAR, and DERA. I am interested primarily in RAA and PRAA (adjusted runs saved in aggregate and those attributed to pitching; the difference is attributed to fielding) and in DERA (the rate derived from PRAA). Tonight my remarks are limited to DERA>

Call this the fall 2009 edition. It is the third that I have examined this year. Evidently DL and Chris Cobb have examined it, too.

So far I have examined only 100 pitchers, down from 600-odd for the winter edition and 848 for the spring edition. Those 100 pitchers include 83 whose study pertains to this annual cycle, in my opinion.

- 60 featured by Chris Cobb in reexamination of post-1892 major league pitchers during the run-up to last year's election
- 12 whom I privately added to the Cobb Sixty at that time. I recommended three of them for consideration --Al Orth, Claude Passeau, and Ned Garver. A fourth may be important to some peakists, Harry Brecheen.
- 6 pre-1893 pitchers who have been named on preliminary or final ballots this year.
- 5 more knuckleball pitchers. The preceding edition or two clearly overrated knuckleball pitchers in general, because it underrated their team fielding support (ie, attributed too many of aggregate runs saved to the pitching). With three primary knuckleballists in the Cobb Sixty --Cicotte, Leonard, and Wood-- there are eight in my recent examination set.

There are 17 others whom I have set aside, essentially, after gathering their revised data.
- 9 upcoming candidates, Kevin Brown (eligible 2011) to Johnson, Smoltz, and Martinez (active 2009).
- 5 miscellaneous
- 3 relief pitchers: Bruce Sutter and Lee Smith, who have been named on preliminary or final ballots, and John Hiller.

The set of 100 does include all of the major league pitchers who have been named on any final ballot 2008 to 2010 (the three annual elections) or on preliminary 2010 ballots.
   552. Paul Wendt Posted: December 01, 2009 at 01:28 AM (#3399627)
examination of fall 2009 Davenstats, continued

Who gets a boost or takes a hit?

I have two general observations based on checking (career DERA* above baseline) x (career Innings). First, the six <u>pre-1893 pitchers have been downgraded enormously</u>. That's relevant to HOM consideration by anyone who has supported them partly in reliance on their winter or spring 2009 Davenstats --probably no one. Second, the eight <u>primary knuckleball pitchers have been downgraded in general</u>.

Let me make the point in terms of a rough code where { -, --, ---, ---- } represent small, medium, large, and x-l downgrades, same for upgrades. For the eight knucklists,

+ Cicotte, Candiotti
= Joe Niekro
- Wood
-- Wakefield
--- Leonard, Phil Niekro
---- Wiilhelm

This is very encouraging for two reasons
: there are many small, medium, and large upgrades for my 83 relevant pitchers
: the knucklists were badly overrated in the early 2009 editions.

Based on these two observations, I judge that the fall 2009 edition is a notable improvement. I think it's as good as the last edition before winter 2009.

Let me qualify that. I know that there has been another shift of explanatory burden between XIP, or translated innings, and the statistics denominated in runs, or RAA, PRAA, PRAR. I don't know how much this improves the set for Pennants Added estimation by Joe Dimino.
(my guess: it doesn't restore the perceived suitability of the last edition before winter 2009)
   553. Chris Cobb Posted: December 01, 2009 at 01:56 AM (#3399642)
Having found the spring 09 update providing perplexing data, I was very pleasantly surprised to find different data when I visited BP's site last week to do a spot check of my pitching rankings. My off-the-cuff assessment matches Paul's more considered one--it appears to be a notable improvement, with many of the bugs that crept in to the sweeping update in spring 2009 eliminated, but with the higher replacement level of WARP 2009 retained, I think. I was further delighted to find that runs above replacement position has been restored (after an absence of some years) to the batting stats--that's a huge convenience when dealing with pitchers, as I no longer have to cross-calculate with other systems to assess pitcher batting appropriately. I haven't taken the time to see how WARP uses that number in getting from BRAR/P, FRAR, and PRAR to WARP1. I hope they are using it instead of BRAR--that would be a big systemic upgrade!--but at least they make that number available to interested parties.
   554. Paul Wendt Posted: December 01, 2009 at 01:57 AM (#3399643)
Let me continue with the rough scale introduced above, representing the magnitude or downgrade or upgrade from spring 2009 to fall 2009 edition.

See also the knucklists, above.

Who gets a boost?

+++++ Jim Kaat (ie, he is no longer damned)
+++ Bob Friend
+++ Jim Vaughn
++ Tommy John
+ Nap Rucker
+ Wilbur Cooper

neutral:

= Phillippe, Adams, Quinn, Trout
= Reuschel, Morris, Gooden, Appier, Finley

Who takes a hit?

- Leever, Joss, Mays, Grimes, Newcombe, Tiant, Hershiser, Cone
-- Vic Willis, Jack Powell, Dolf Luque, Bucky Walters, Virgil Trucks, Frank Tanana
--- Noodles Hahn(*), Al Orth, Urban Shocker(**), Dizzy Dean, Lefty Gomez,
---- Eddie Rommel, Tommy Bridges, Ned Garver(**)


NOTES
A small upgrade or downgrade is small!

**
The spring 2009 edition was keen on Shocker and Garver(**), enough so that even their large and x-large downgrades leave them among the strong candidates by this simple derived rating.
*
The same is true for Hahn(*) by reference to league-average baseline (DERA = 4.50; DERA+ or DERA* = 100).
   555. Paul Wendt Posted: December 01, 2009 at 02:00 AM (#3399653)
#553.
I was further delighted to find that runs above replacement position has been restored (after an absence of some years) to the batting stats--that's a huge convenience when dealing with pitchers, as I no longer have to cross-calculate with other systems to assess pitcher batting appropriately.

Yes, that's good, too.
I haven't even glanced at any of the batting or fielding in the new edition.
   556. Joey Numbaz (Scruff) Posted: December 01, 2009 at 02:41 AM (#3399680)
Paul/Chris,

I was using NRA - DERA to approximate team defensive quality. Before the 2008 update, this number was basically the same for everyone on a team for one.

Is there something now that will work as a proxy for team defense? It looks like I can use the 'adjusted for all-time' version of this, but not they 'adjusted for season' version - at least according to the 1970 Braves (wanted Niekro in there).

Does this make sense? And if so, why does this 'work' for the all-time but not for season. Seems oddly inconsistent.

http://www.baseballprospectus.com/dt//1970ATL-N.php
   557. Joey Numbaz (Scruff) Posted: December 02, 2009 at 02:44 PM (#3400992)
Just bumping this . . . I'm going to start going through everyone again and updating them just want to make sure I get it right . . . any thoughts on my question in #556?
   558. Paul Wendt Posted: December 02, 2009 at 07:10 PM (#3401286)
Joe,
no substantial thoughts yet

IIUC you need these measures for every player-season stint, line by line from "Advanced Pitching Statistics". Maybe you can buy a copy from Davenport or BP. --for less than prohibitive price, I speculate, because a selection of six pitching fields would be inadequate to reproduce much of the rating system.
   559. Paul Wendt Posted: December 02, 2009 at 10:29 PM (#3401561)
Using Tim Keefe and Charley Radbourn for illustration,
#510 features the winter 2009 edition of "Advanced Pitching Statistics" (Davenstats), primarily {XIP, RAA, PRAA, DERA}. The next articles pursue the point and cite some discussion in other threads that refer to or depend on that revised edition.

Using all Hall of Merit pitchers and DERA for illustration,
#515 features the spring 2009 edition in contrast to the 2008; that is, incorporating two rounds of revisions. Much of #515-540 concerns that edition of the Davenstats.

This week's discussion
#551ff concerns the fall 2009 edition. (Call if "fall 2009" although none of us may know when between May and November that revision, or more than one revision, may have occurred.)


Joe Dimino's work on pitchers, whose product is Pennants Added, continues to rely on a 2008 or earlier edition of some Davenstats (at the season stint rather than career level, I understand). See #542-44, 556-57. That is conservative and it seems to be a good thing in this instance.
   560. Paul Wendt Posted: December 02, 2009 at 11:06 PM (#3401599)
Thanks to #510-11, I have the career lines for Keefe and Radbourn in the 2008 (or earlier?) and all three known 2009 editions. There has been no revision in {RA, DH, DR, DW} afaik. Those are now presented in the first four fields following IP (career innings, repeated from "Actual Pitching Statistics").

Six estimates that have been revised. They are now presented in the next six columns under the heading "adjusted for season". Here are their values in all four editions.

The differences are radical, including relative differences in the revisions for Keefe and Radbourn.


DT Cards, BaseballProspectus
Advanced Pitching Statistics
Adjusted for Season, four editions

Tim Keefe (5049 IP, 1880-1893)

XIP NRA RAA tfaa PRAA PRAR DERA
5093 3.67 470 034 436 957 3.73 Keefe, 2008(?)
2099 3.67 190 -51 241 474 3.47 Keefe, winter 2009
2099 ---- 190 063 127 418 3.96 Keefe, spring 2009
5093 3.67 410 264 146 712 4.24 Keefe, fall 2009

2796 3.97 165 134 031 342 4.40 Keefe, fall 2009 "adjusted for all time"

Charley Radbourn (4535 IP, 1880-1891)

XIP NRA RAA tfaa PRAA PRAR DERA
4525 3.74 384 206 178 587 4.15 Radbourn, 2008(?)
1754 3.74 135-112 247 442 3.23 Radbourn, winter 2009
1754 ---- 135 042 093 337 4.02 Radbourn, spring 2009
4525 3.74 323 209 114 616 4.27 Radbourn, fall 2009

2384 3.83 178 113 065 330 4.26 Radbourn, fall 2009 "adjusted for all time"


I have inserted tfaa, or team fielding (runs) above average, the numerical difference between RAA and PRAA. The allocation of RAA represented by the pair <tfaa, PRAA> is the allocation of credit and blame to fielding and pitching. I have retained the order of presentation from the preceding three editions on one point, PRAA before PRAR.

These tables now include NRA but show no change in the values for Keefe and Radbourn. For my 100+ pitchers there has been some modest revision, 0 to +0.03 in magnitude as follows.
+0.03, one (Joh Hiller)
+0.02, 7
+0.01, 23
none, 72
   561. Paul Wendt Posted: December 02, 2009 at 11:17 PM (#3401606)
Note that the current edition restores the 2008 measure of translated innings XIP adjusted for season, which is not much different from actual innings IP. Thus the recent revision shifts the burden of explanation massively back to the estimates denominated in runs.

Note that the current edition for Radbourn shows DERA=4.26 "adjusted for all time", which barely differs from the bottom-line rate DERA "adjusted for season". Translated innings XIP=2384 for all time, down about 50% from "adjusted for season"; the estimates denominated in runs are reduced to that scale in order of magnitude but not proportionally.

Not shown here, the current edition rates Al Spalding's pitching much below average, adjusted for season (DERA 4.93) and about the same, adjusted for all time (DERA 4.97). On the familiar index scale, roughly DERA+ = 91 either way.

That's all the analysis for now.
   562. Paul Wendt Posted: December 02, 2009 at 11:41 PM (#3401627)
For three of the four editions, I have uploaded the data for 105 pitchers in .csv format.
Here is the link for your file, which will be available for 7 Days.
http://www.yousendit.com/download/MVNkeFVYcVhCTW1Ga1E9PQ (Advanced Pitching Statistics, three 2008/2009 editions)

<u>The featured Davenstats appear in 21 columns</u> following some preliminaries.
RA, DH, DR, DW -- all three editions 2008, spring 2009, and fall 2009
nnXIP, nnNRA, nnRAA, nnPRAR, nnPRAA, nnDERA -- fall 2009
newXIP, newRAA, newPRAR, newPRAA, newDERA -- spring 2009; NRA completes that edition
XIP, NRA, RAA, PRAA, PRAR, DERA -- 2008 or earlier (PRAR is incomplete)

These are career measures only, one record for each pitcher.
   563. Paul Wendt Posted: December 03, 2009 at 12:06 AM (#3401642)
For 2008 and spring 2009 editions I have these career data for 637 pitchers (2008 PRAR is incomplete).
For the spring 2009 edition only I have the same for 211 more pitchers, 848 total.
Available upon request.

I wonder what others have. Generally I don't know what others mean when they write 'WARP' (except Dan Rosenheck who means rosenWARP or WARPenheck). DL, what is it that you will make available?
   564. DL from MN Posted: December 03, 2009 at 02:10 AM (#3401684)
I have Rosenheck's pitcher WARP which he appears to have lost. It's at work, I'll send it back to him tomorrow. I'm really tired of WARP moving around by hundreds of runs. If they're making changes that radical, perhaps they should call them WARP4, WARP5, WARP6.
   565. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 03, 2009 at 04:36 AM (#3401778)
Yes, the spreadsheet I used to calculate them went byebye when a hotel lost my bag with my laptop in it. I will have to start from scratch.
   566. Paul Wendt Posted: December 03, 2009 at 05:19 PM (#3402120)
Break up WARP3!

I'm really tired of WARP moving around by hundreds of runs. If they're making changes that radical, perhaps they should call them WARP4, WARP5, WARP6.

Yes indeed. At the same time, let's break up WARP3, which should be distinguished from WARP2 by one comprehensive adjustment --eg, adjustment for the number of team games played/scheduled.

How about 48 editions determined by selection from menus with 2, 2, 3, and 4 alternatives? Unfortunately we make too small and too impoverished a market.


That's all today. I need to shop for another flash drive, not wait for mail delivery either.
;-)
   567. Joey Numbaz (Scruff) Posted: December 04, 2009 at 08:10 AM (#3403110)
Here's a link to my pitcher pennants added spreadsheet. Not the most user friendly, but I don't think it's half bad.

The Rosenheck DB will be up shortly (FTP client says 17 minutes). It's the older version though, but I'm not sure when I'll get to update it (some time this weekend or possibly not until early next week) and didn't want to wait.
   568. Paul Wendt Posted: December 04, 2009 at 10:01 PM (#3403672)
"2011 Ballot Discussion" #57-83 includes Joe Dimino and Dan Rosenheck on relief pitcher replacement-level (slightly) and value above replacement (mainly). DanR argues for measurement that takes league-average bullpen staff as a reference point. Actually that seems to be the league-average "N best" pitchers on a team where N=6 for the last five seasons, the time period that illustrates the method. Here is the heart of it.

70. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 02, 2009 at 11:21 PM (#3401800)
... I'm using WPA rather than raw runs allowed because it's what I have handy, but the principles are the same. Here is what the average bullpen looked like from 2004-08:

Rank = rank on team in leverage index
LI = leverage index
WPA = win probability added
WPA/60 = win probability added per 60 IP
WPA/LI/60 = WPA/60 divided by leverage index, or
the WPA the reliever would have if he pitched 60 innings of an average impact on the game's outcome.

Rank   IP   LI   WPA WPA/60 WPA/LI/60
1    64.6 1.81  1.53   1.43      0.79
2    63.2 1.42  0.77   0.73      0.51
3    59.6 1.16  0.49   0.50      0.43
4    54.7 0.97  0.17   0.19      0.20
5    53.4 0.80 
-0.03  -0.03     -0.04
6
+   60.2 0.64 -0.05  -0.05     -0.07
TOT 355.7 1.15  2.88   0.49      0.42 
...
Do you disagree with any of this methodology?



In the relevant thought experiments we add a given pitcher to the average staff and bump the lesser pitchers down one rank each, or we substitute him for the pitcher in the most appropriate rank on the average staff, or we do that and then remove him.

Within a staff the pitchers are ranked by single-season WPA/LI/60 (or X/LI/60 using another count of achievement X).

In the model, if I understand correctly,
- X/LI/60 is the constant quality of a pitcher. He takes it with him from one role to another.
- Management knows the qualities of its pitchers and allocates roles according to quality.

I don't understand how the model management allocates innings. For example, how would it add to the average staff a pitcher with 80 innings and WPA/LI/60 = 0.333?
   569. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 05, 2009 at 01:54 AM (#3403959)
I don't have every last detail worked out, Paul Wendt, but typically I slot in a P where his LI says he fits in, and then have him take away/give back innings to the guy below him in LI, adjusting the LI of the guy below him to keep total bullpen LI constant. The Rivera example on that thread shows how I did this with the setup man.
   570. Paul Wendt Posted: October 09, 2010 at 04:09 PM (#3659651)
There is some discussion of pitching in this month's flurry of activity, 2011 Ballot Discussion #210-2??.
DL from MN seems to be working toward something new on playing time which I have selected as the point of entry for that link.

Pitcher Kevin Brown is one of the best new candidates.
   571. Bleed the Freak Posted: January 03, 2011 at 12:31 AM (#3722031)
Joe Dimino - to answer your question in the 2011 ballot discussion post 327

And if you can find me a good new proxy for team defense or a way to get at the old BPro cards (since the BPro cards no longer have what I need), I'm all ears!

A hat tip to Chris Jaffe for mentioning the old-school DT Cards in an article he wrote about Omar Vizquel:

http://www.hardballtimes.com/main/article/when-do-we-start-taking-omar-vizquels-cooperstown-case-seriously/

Under the references and resources section, he lists the direct link to Vizquel, and mentions that, to query for other players, using the baseball-reference (Lahman database) abbreviation will net the correct result.

http://www.baseballprospectus.com/dt/vizquom01.shtml

I hope this is helpful.
Page 6 of 6 pages  < 1 2 3 4 5 6

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Tuque
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.1913 seconds
49 querie(s) executed