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

Tuesday, December 23, 2014

Trevor Hoffman and Billy Wagner

Eligible in 2016

DL from MN Posted: December 23, 2014 at 11:07 AM | 85 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.

   1. Dr. Chaleeko Posted: December 24, 2014 at 08:16 PM (#4868197)
This seems like a place to run my reliever system by youse guys to make sure I'm not doing something incredibly stupid.

BBREF WAR seems to undervalue RPs because although it does use leverage, relievers are deployed as situation-specific, tactical weapons especially in order to close down games (and end-of-game relievers are the ones we really care about here). WPA does a nice job of capturing that value where WAR struggles with it. I want to meld the two because I want to capture both performance and its impact for this type of player.

In other words: PERFORMANCE + IMPACT = RELIEF VALUE or WAR + WPA = RP VALUE

Of course, you can't just add WPA and WAR together. They are both denominated in wins, and there's repetition between them as a result of their both assessing the same performance in a different way. But not complete repetition because WPA is set against average and therefore analogous to WAA, so we want to conserve the value between replacement and average that WAR provides. So what I really want to do is...

(PERFORMANCE - CONTEXT) + (IMPACT - PERFORMANCE) = RP VALUE

If you know what I mean.

BBREF tells us that WAA + WAA_adj + REP = WAR. The context embedded in WAR is the WAA_adj column (which also has a small other function, but let's not get into a thicket). Therefore we might say now:

(WAR - WAA_adj) + (IMPACT - PERFORMANCE) = RP VALUE

On the other side of the plus sign, BBREF tells us that WPA/LI is WPA with the leverage taken out. Or, we might say that WPA/LI is the performance, which, I think, means that the rest of WPA is the impact. Therefore:

(WAR - WAA_adj) + (WPA - WPA/LI) = RP VALUE

Here's an example using John Smoltz since he's a hot topic of late. Let's take his 2002 to 2004 relief seasons.
2002: (1.2 WAR - 0.0 WAA_adj) + (5.0 WPA - 1.8 WPA/LI) = 4.4 Wins
2003: (3.3 WAR - 0.6 WAA_adj) + (3.1 WPA - 1.9 WPA/LI) = 3.9 Wins
2004: (2.2 WAR - 0.2 WAA_adj) + (5.1 WPA - 1.8 WPA/LI) = 5.2 Wins

One might wonder if this method would pump relievers up too high. The answer, i believe, is no. I did this with the most important relievers, and the result was that Goosage went from a fringy HOMer to a solid one, and Mariano went from solid to nearly automatic (when he arrives, of course). Eck picks up three wins that he didn't really need, but are nice to have. Billy Wagner picks up four or five wins, not enough to push him across. Trevor Hoffman picks up ten wins, but he's still streets behind the worst SPs in the HOM. Fingers only adds 1.5 Wins. Smoltz was leveraged unusually well and picks up about 2 to 2.5 wins a year, but all that does is make him look better when he is HOMward bound (for me) anyway.

An objection to this idea might be that now the wins won't add up across leagues/seasons. I don't worry too much about this because we're dealing with a very small subset of players.

At this time, this can only be used in seasons where we have PBP data and where the pitcher is really obviously a reliever because BBREF doesn't have WPA or WPA/LI splits, and it would be murder to go through game by game for the oldsters. I'd love to be able to calculate something like this for Miner Brown, Dizzy Dean, Eddie Rommel, Red Faber, Chief Bender, Firpo Marbury, and Lefty Grove.

Anyway, this is something I've been using but not entirely sure if it is doing something unwise with WAR, so I welcome feedback from the group. Thanks! And Merry Christmas!
   2. Win Big Stein's Money Posted: December 24, 2014 at 08:53 PM (#4868199)
They're generally regarded as the two best closers (non-android division) to play the game. If you're the best player at your position, I say you qualify.
   3. Chris Fluit Posted: December 26, 2014 at 07:17 PM (#4868585)
Thanks for posting this, Dr. Chaleeko. I agree that WAR underrates relievers but I've been having trouble figuring out what system to use instead. Win Shares? WPA? WAR per IP? Every other approach also has its problems.

I think we made a mistake when we elected Rollie Fingers (I was one of the voters at the time) as he was the first pure reliever to really merit attention. Yet, at the same time, I think that Hoffman is a credible candidate. I don't yet know if he's worthy of a ballot spot or should fall short but it would be good to a have a more extensive evaluation. Wagner might be an even better candidate than Hoffman but it's hard to support someone who didn't even reach 1000 IP for his career.
   4. karlmagnus Posted: December 26, 2014 at 08:00 PM (#4868594)
According to my formula of Pitcher Points = (ERA+-90)*(IP/1000) Hoffman is nowhere close (55) while Wagner is low borderline (88) -- close to Sam Leever. Fingers (49) and Gossage (65) were nowhere close, Rivera is a slam-dunk (148) as was Wilhelm (126, compared to Schilling's 122.)

Relievers can get in, but they'd better post an other-worldly ERA+.

I think I can live with the results of that formula; quick-and-dirty, but gets relievers vs starters about where I want them.
   5. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 27, 2014 at 08:21 PM (#4868848)
Dr. Chaleeko, I think your method double-counts leverage. RA-based wins above average times leverage index should correlate very closely to WPA--if it doesn't, something's wrong with how LI is being calculated. The reason why BB-ref WAR for relievers is coming in lower than you would "like" is because it regresses the LI halfway towards 1 to account for "chaining"--the fact that when a closer gets hurt, those maximum-leverage innings don't go to a replacement-level reliever, but rather to his setup man, who is also likely to be an above-average pitcher. To properly calculate a reliever's value, you have to look at his work in the context of a whole bullpen. The exact result of this calculation depends on your assumptions about what an average bullpen looks like and who would get which innings, but Tom Tango has written that this usually averages out to about (LI+1)/2, which is the amount that BB-ref uses. Your method is effectively valuing leverage at 1.5x, assigning credit both for its full impact (via WPA - WPA/LI), and then granting a further 50% from BB-ref WAR.

I'm quite confident that BB-ref WAR (which as you well know I very often take issue with!) is correct on overall reliever valuation. The only way modern closers, even Rivera, deserve to get into the HoM is a) with some quota system equivalent to the "catcher bonus" that so many of us use or b) with heavy postseason credit, which is merited since their usage in the playoffs is much greater than in the regular season. I would object to the reasoning behind A, because you can't field a baseball team without a catcher, but you certainly can field one without a Proven Closer, and teams did so for about 100 years.
   6. theboyqueen Posted: December 27, 2014 at 11:13 PM (#4868871)
They're generally regarded as the two best closers (non-android division) to play the game. If you're the best player at your position, I say you qualify.


The problem with your logic is that "closer" is a position like "pinch hitter" is a position.
   7. bjhanke Posted: December 28, 2014 at 12:00 AM (#4868884)
I got to thinking about this business of evaluating relievers by leverage. While I agree with the idea, it also occurred to me that any leverage a reliever receives is leftover from the previous pitcher. In other words, the leverage the first reliever in the game gets is whatever the starter left him, and so on down the relief chain. I think that, if you account for the leverage the reliever inherits and also the leverage the reliever (or starter) leaves for the reliever to inherit, you'd come up with much better evaluations. In particular, some starters are going to look better than others, because they generally don't leave the game until an inning is over. Other starters, who often have to be relieved in high-leverage situations, because they are collapsing out there, wouldn't score as high. I don't have any method for doing this, although most of the thinking is pretty obvious - Markov Chains galore. So, I'm just tossing the idea out there to see if anyone else thinks that would be useful. - Brock Hanke
   8. bobm Posted: December 28, 2014 at 12:08 AM (#4868887)
[6] The problem with your logic is that "closer" is a position like "pinch hitter" is a position.

For cumulative seasons, From 1914 to 2014, in Sv Situ (within Game Outcome for Pitcher), sorted by greatest Runs Created for this split

                                                                                     
Rk   I              Player      Split From   To   G    RC  W  L  ERA  GF  SV       IP
1                Lee Smith in Sv Situ 1980 1997 615 296.1 21 40 2.94 526 478 745.0001
2              John Franco in Sv Situ 1984 2005 605 272.6 24 39 2.80 478 424 655.3335
3             Jeff Reardon in Sv Situ 1979 1994 498 260.9 18 35 3.38 412 367    617.1
4           Rollie Fingers in Sv Situ 1969 1985 486 251.7 28 48 2.69 412 341    790.0
5        Roberto Hernandez in Sv Situ 1992 2007 501 243.5 23 38 3.49 370 326    519.0
6               Todd Jones in Sv Situ 1993 2008 495 240.4 12 34 3.78 347 319 509.6668
7             Bruce Sutter in Sv Situ 1976 1988 412 230.7 18 35 2.86 344 300    651.2
9        Francisco Cordero in Sv Situ 1999 2012 458 218.5 12 32 3.61 353 329    458.2
10            Rich Gossage in Sv Situ 1972 1994 455 216.9 27 38 2.41 363 310    714.0
11          Trevor Hoffman in Sv Situ 1993 2010 696 215.4 12 31 2.71 634 601    698.1
12         Jeff Montgomery in Sv Situ 1987 1999 413 214.3 16 30 3.16 342 304 498.9999
13              Doug Jones in Sv Situ 1986 2000 412 213.4 15 31 2.94 344 303    520.2
14          Mariano Rivera in Sv Situ 1995 2013 763 212.9 21 25 1.95 692 652 836.6665
17               Jose Mesa in Sv Situ 1987 2007 459 202.3 11 31 3.26 346 321 455.3332
21           Rick Aguilera in Sv Situ 1986 2000 402 196.2 20 34 3.37 363 318 453.6665
22         Dan Quisenberry in Sv Situ 1979 1990 321 191.9 17 21 2.49 281 244    555.2
23     Francisco Rodriguez in Sv Situ 2003 2014 499 191.0 15 17 2.76 371 348    495.1
30          John Wetteland in Sv Situ 1989 2000 394 179.7 17 25 3.08 366 330    432.2
31             Randy Myers in Sv Situ 1987 1998 445 178.1 12 27 2.74 377 347    495.2
33         Armando Benitez in Sv Situ 1995 2008 419 172.7  9 23 2.98 311 289 417.3331
34        Dennis Eckersley in Sv Situ 1975 1998 480 172.2 15 26 2.83 422 390    528.0
40      Jason Isringhausen in Sv Situ 1999 2012 393 162.0 12 26 3.49 322 300    395.0
42           Troy Percival in Sv Situ 1995 2009 455 159.6  9 25 3.22 380 358    453.0
44            Billy Wagner in Sv Situ 1996 2010 504 153.1 11 22 2.43 447 422 514.6668
45                Robb Nen in Sv Situ 1994 2002 369 152.7 13 21 2.89 345 314    388.2
50               Tom Henke in Sv Situ 1982 1995 384 147.4 15 27 2.45 347 311    458.1
72              Joe Nathan in Sv Situ 1999 2014 450 127.9  9 15 2.40 390 376    438.1
79       Jonathan Papelbon in Sv Situ 2005 2014 376 125.5  8 17 2.36 346 325 382.0001
92           Jose Valverde in Sv Situ 2003 2014 350 116.8  7 16 2.52 308 288    350.2


Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 12/28/2014.

For cumulative seasons, From 1914 to 2014, as PH (within Defensive Positions), sorted by greatest Runs Created for this split

                                                                       
Rk   I          Player Split From   To   G   RC  PA   BA  OBP  SLG  OPS
1         Mark Sweeney as PH 1995 2008 799 99.9 799 .258 .353 .380 .733
2         Lenny Harris as PH 1989 2005 883 82.3 883 .264 .317 .337 .654
3    I   Smoky Burgess as PH 1949 1967 582 82.0 582 .285 .378 .437 .815
4          Dave Hansen as PH 1990 2005 703 78.5 703 .234 .348 .358 .705
5      John Vander Wal as PH 1991 2004 624 78.0 624 .236 .342 .409 .751
6          Matt Stairs as PH 1992 2011 490 74.8 490 .252 .357 .476 .833
7    I      Manny Mota as PH 1962 1982 592 71.7 592 .300 .373 .368 .741
8           Greg Gross as PH 1973 1989 733 68.9 733 .240 .362 .282 .644
9          Steve Braun as PH 1971 1985 482 66.9 482 .281 .386 .389 .775
10         Gates Brown as PH 1963 1975 499 66.6 499 .251 .357 .422 .779


Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 12/27/2014.
   9. bobm Posted: December 28, 2014 at 12:34 AM (#4868894)
[7] the leverage the first reliever in the game gets is whatever the starter left him

And whatever leverage existed when the manager elected to relieve the starter, etc. The number of relief innings started "clean," ie leading off an inning, seems to be increasing. In that situation, without baserunners or outs, the leverage is a function of the inning, the runs surrendered by the prior pitchers and the runs scored by the offense.

I guess one could look at some amalgamation of starters' IP/GS and BQR/GS (bequeathed, ie left to be inherited, runners per start), plus things like quality start % and run support.
   10. Howie Menckel Posted: December 28, 2014 at 12:48 AM (#4868896)

"I think we made a mistake when we elected Rollie Fingers"

I believe HOM founder Joe Dimino noted an otherwordly strand rate by Fingers entering in mid-inning. Modern guys have it easy.
   11. theboyqueen Posted: December 28, 2014 at 12:54 AM (#4868897)
I'm not sure what the list in [8] is meant to illustrate. It is not news to me that the best closers have been much more valuable than the best pinch hitters, but the question remains: What makes "closer" a position separate from "pitcher"? How far down the list of pitchers in history would you have to go to get to Lee Smith? I'm asking this genuinely as I didn't even know runs created was a thing for pitchers. My guess is that a guy like Jered Weaver has already had a more valuable career than any of Lee Smith/Billy Wagner/Trevor Hoffman, but I'm willing to be shown how I am wrong.

Who is the best middle reliever of all time?
   12. alexking Posted: December 28, 2014 at 02:40 AM (#4868907)
One rough cross-check on the WAR figures is to check salaries: who are similarly-paid contemporaries of Hoffman/Wagner/Rivera? I looked for players with career earnings within $20 million and who debuted within +/- 2 years of each player: (debut years were determined somewhat arbitrarily and on the fly but shouldn't make too much of a difference)

Hoffman debuted in 1993 and made $80 million. These are his comparables in salary:

Player debut WAR salary
Hampton 1995 29 95
Gonzalez 1991 39 87
Kent 1992 55 86
Edmonds 1994 60 86
Alou 1992 40 81
Anderson 1995 26 76
Vaughn 1992 27 70
Leiter 1994 43 68
Durham 1995 34 67
Gonzalez 1991 52 66
Mondesi 1994 29 66
Radke 1995 46 62
Lofton 1992 68.2 61
Salmon 1993 40 61
Lopez 1995 30 61
Belle 1991 40 60
Klesko 1995 27 60

That's a group with average bbref WAR 40.3 and salary $71 million. Of these 17 players only 3 to 6 are even HOM candidates (Edmonds, Kent, Lofton, maybe Salmon, Belle, and Radke).

Wagner debuted in 1996 and made $93 million:

Player debut WAR salary
Edmonds 1994 60 86
Green 1995 35 101
Hampton 1995 29 95
Anderson 1995 26 76
Damon 1996 56 112
Renteria 1996 32 85
Kendall 1996 42 80
Jones 1997 63 107
Giles 1997 51 81
Garciaparra 1997 44 78
Schmidt 1997 32 77
Cameron 1997 47 76
Lowe 1998 33 112
Lee 1998 34 91
Millwood 1998 30 90
Carpenter 1998 36 88
Colon 1998 45 85
Wood 1998 27 73

These 18 players averaged 40.1 WAR and made $88.5 million, and 5-6 of them are serious candidates (Edmonds, Jones, Giles, Garciaparra, Damon, maybe Cameron).

Meanwhile Rivera's salary comparables paint a different picture (I stretched the boundaries a little due to a dearth of comparable players):

Player debut WAR salary
Thome 1994 72.9 143
Pettitte 1995 61 140
Delgado 1996 44 146
Helton 1998 62 161
Jones 1995 85 168
Pedro 1993 86 146

These guys have an average of 68.5 WAR and made $151 million (Rivera made $169 million).

So overall, although this is a very rough way of evaluating players, the salary data seems to support Rivera as a clear HOMer and Hoffman/Wagner somewhat below the cut.
   13. DL from MN Posted: December 28, 2014 at 10:27 AM (#4868956)
Who is the best middle reliever of all time?


Kent Tekulve was pretty good.

I agree that if you use leverage you need to account for the higher replacement level (chaining). Managers aren't bringing anyone but their top 3 bullpen guys in a save situation if they can help it.

How much difference is there between these two approaches?

PWAR + leverage - chaining
PWAR

In other words, does the chaining cancel out the leverage?
   14. cardsfanboy Posted: December 28, 2014 at 12:06 PM (#4868996)
I believe HOM founder Joe Dimino noted an otherwordly strand rate by Fingers entering in mid-inning. Modern guys have it easy.


Hoffman is probably the best in recent history at stranding inherited runners(20.2%) most modern day closers are around 27% (Rivera had 367 and allowed 29.16% to score) . But he inherited 346 in comparison to Fingers 782(28.7%).
   15. Dennis Eclairskey, closer Posted: December 28, 2014 at 01:16 PM (#4869030)
Not that it matters as much in HOM voting as it does in HOF voting but didn't Wagner get run out Philly and also seemed to have some high profile blown saves as a member of the Mets. I know Hoffman had the big blown save against Colorado & limped to 600 svs but seemed to be more appreciated than Wagner
   16. DL from MN Posted: December 28, 2014 at 01:43 PM (#4869035)
The really important question. "Enter Sandman" or "Hell's Bells"?
   17. Howie Menckel Posted: December 28, 2014 at 01:46 PM (#4869037)
I'm pretty confident that Fingers not only had far more "fires" to put out, they also blazed more brightly.

That is, Hoffman would not be used with a 4-0 lead, but if the leadoff man walks, the setup man gets two outs, and then there is a single, now the tying run is on deck. Save situation, so here comes Hoffman. All he has to do is get 1 out before he allows 1 hit, basically, to boost his strand rate (and he can allow a 3-HR and a double off the wall plus a couple of walks, if he likes, before getting 1 out and getting credit for a silly SV).

The Fingers-era usage was not only to not wait until the 9th inning, but to feel free to use a RP in a much tougher spot. So if a 4-0 lead in the 6th or 7th becomes a 4-2 lead with 2nd and 3rd and none out, that's a spot for a "real" RP fireman. The MGR didn't care if Fingers et al got stuck with a blown SV even if he did an acceptable job in ending the inning tied. The focus on a W eclipsed the preciousness of a closer's SV pct, back then.

The game is now as rigged as it can possibly be for the better RPs. A surface-level comparison to Fingers or Gossage is fairly useless. Even strand pct is now rigged for the current closers, as noted.
   18. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 28, 2014 at 02:51 PM (#4869046)
DL from MN, what do you mean by PWAR? Just the gap between runs allowed and those that would have been allowed by a 94 ERA+ mopup man? In that case, chaining negates about half the leverage at the back of the bullpen. (For the mopup man himself, there is no chaining effect, as his replacement is the true replacement reliever.) If by PWAR you mean BB-ref WAR, well, that already Is RAR + leverage - chaining converted to wins.
   19. Dr. Chaleeko Posted: December 28, 2014 at 03:27 PM (#4869054)
Dan, thanks for your help. DL, your comment about chaining equalling replacement level really helped me understand the concept better.. Much obliged.
   20. DL from MN Posted: December 28, 2014 at 04:53 PM (#4869075)
Just the gap between runs allowed and those that would have been allowed by a 94 ERA+ mopup man?


That's what I meant. Thanks for the answer.

For Hoffman it's the 13.7 WAA that is going to keep him from ever seeing my ballot (roughly 15 if you zero out his last season). He's roughly equivalent to Lee Smith and I don't vote for him either.

Mariano Rivera's 32.7 WAA is much more impressive and not far behind pitchers like Reuschel, Saberhagen, Tiant and Stieb. He's a step ahead of Hershiser and John. Tim Hudson is the rough equivalent contemporary starting pitcher.
   21. bobm Posted: December 28, 2014 at 05:24 PM (#4869084)
I'm pretty confident that Fingers not only had far more "fires" to put out, they also blazed more brightly.

Based on BR PI Event Finder data:

59% of Hoffman's IR were RISP
56% of Fingers'
56% of Lee Smith's
52% of Rivera's
52% of Wagner's
   22. Joyful Calculus Instructor Posted: December 29, 2014 at 12:19 AM (#4869173)
Wait, Fingers made the Hall of Merit? In Bill James' Historical Abstract, he makes a scathing review of Fingers and why he did not list Fingers as a top 100 pitcher all-time:

In my opinion the BBWAA did something dumb when they gave Fingers an MVP award and then compounded the dumbitutde by using that as a reason to put him in the Hall of Fame.
   23. Bring Me the Head of Alfredo Griffin (Vlad) Posted: December 29, 2014 at 10:26 AM (#4869242)
Kent Tekulve was pretty good.


He was, but he also had 184 career saves. That feels like a lot for a middle reliever.
   24. Misirlou doesn't live in the restaurant Posted: December 29, 2014 at 10:55 AM (#4869263)
Tekulve finished 638 of 1050 games. that's not a middle reliever.
   25. Howie Menckel Posted: December 29, 2014 at 12:18 PM (#4869326)

Thanks, bobm, although we also need how many outs when the RP arrived on the scene. Ideally, margin of error as well. I think we know that Hoffman is not coming into a 3-2 game in the 9th with a man on 3rd and nobody out (an extreme, of course, but it would be cool if we could figure out actually degree of difficulty. 21 is an interesting start).

   26. Rob_Wood Posted: December 29, 2014 at 01:32 PM (#4869392)
Here is something I was able to do with BB-Ref's pitching split finder. It is pretty basic and I will keep looking for more detailed info on relief pitchers.

Below is the IP, Saves, and IP/Save in Saves during their careers for any pitcher with at least 20 saves in any season, sorted by IP/Save:

Player IP in Saves Saves IP/Save
Bob Stanley 250 132 1.89
Bill Campbell 238 126 1.89
Dick Radatz 224 120 1.87
Mike Marshall 347 188 1.85
Hoyt Wilhelm 420 228 1.84
Gene Garber 392 218 1.80
John Hiller 221 125 1.77
Dan Quisenberry 422 244 1.73
Stu Miller 264 153 1.73
Terry Forster 216 127 1.70
Clay Carroll 242 143 1.69
Jack Aker 210 124 1.69
Greg Minton 252 150 1.68
Darold Knowles 238 143 1.66
Tug McGraw 299 180 1.66
Ron Perranoski 295 178 1.66
Lindy McDaniel 288 174 1.66
Roger McDowell 261 159 1.64
Gary Lavelle 219 136 1.61
Rollie Fingers 547 341 1.60
Bruce Sutter 473 300 1.58
Rich Gossage 487 310 1.57
Dave Giusti 227 145 1.57
Sparky Lyle 371 238 1.56
Steve Bedrosian 282 184 1.53
Dave LaRoche 193 126 1.53
Willie Hernandez 225 147 1.53
Don McMahon 230 152 1.51
Duane Ward 183 121 1.51
Ted Abernathy 224 149 1.50
Kent Tekulve 275 184 1.50
Ron Davis 194 130 1.49
Jim Brewer 197 133 1.48
Jesse Orosco 205 144 1.43
Jay Howell 218 155 1.41
Mel Rojas 170 126 1.35
Roy Face 251 191 1.31
Dave Righetti 327 252 1.30
Dave Smith 280 216 1.30
Jeff Reardon 470 367 1.28
Mike Henneman 244 193 1.26
Bobby Thigpen 254 201 1.26
Doug Jones 380 303 1.25
Jeff Brantley 214 172 1.24
Lee Smith 593 478 1.24
Bryan Harvey 211 177 1.19
Danny Graves 214 182 1.18
Keith Foulke 224 191 1.17
Tom Henke 364 311 1.17
Michael Jackson 166 142 1.17
Dan Plesac 183 158 1.16
Mitch Williams 222 192 1.16
Jeff Montgomery 350 304 1.15
Steve Farr 149 132 1.13
Randy Myers 390 347 1.12
Todd Worrell 286 256 1.12
John Smoltz 172 154 1.12
John Franco 471 424 1.11
Jeff Shaw 225 203 1.11
Jeff Russell 206 186 1.11
Dennis Eckersley 432 390 1.11
Gregg Olson 235 217 1.08
John Wetteland 356 330 1.08
Mike Timlin 152 141 1.08
Rick Aguilera 342 318 1.08
Billy Koch 174 163 1.07
Roberto Hernandez 347 326 1.06
Mariano Rivera 688 652 1.06
Tom Gordon 166 158 1.05
Eric Gagne 196 187 1.05
Joakim Soria 186 178 1.05
Robb Nen 327 314 1.04
Armando Benitez 301 289 1.04
Ugueth Urbina 246 237 1.04
Bob Wickman 275 267 1.03
Jason Isringhausen 309 300 1.03
Todd Jones 328 319 1.03
Jim Johnson 127 124 1.03
Chad Cordero 131 128 1.02
Antonio Alfonseca 132 129 1.02
Jonathan Papelbon 332 325 1.02
Billy Wagner 431 422 1.02
J.J. Putz 193 189 1.02
Rafael Soriano 211 207 1.02
Huston Street 280 275 1.02
Jose Valverde 293 288 1.02
Brian Wilson 175 172 1.02
Kevin Gregg 179 177 1.01
Rod Beck 289 286 1.01
Francisco Cordero 332 329 1.01
Mark Wohlers 120 119 1.01
Troy Percival 361 358 1.01
Kazuhiro Sasaki 130 129 1.01
Trevor Hoffman 602 601 1.00
Eddie Guardado 187 187 1.00
Fernando Rodney 220 220 1.00
Heath Bell 168 168 1.00
Brad Lidge 224 225 1.00
Chris Perez 132 133 0.99
Jose Mesa 319 321 0.99
Matt Capps 137 138 0.99
Francisco Rodriguez 345 348 0.99
Craig Kimbrel 184 186 0.99
Bobby Jenks 171 173 0.99
Joe Nathan 370 376 0.98
LaTroy Hawkins 121 124 0.98
Brian Fuentes 198 204 0.97
Mike Williams 139 144 0.97
Joe Borowski 126 131 0.96

P.S. Looks like I need a refresher course in formatting!
   27. Rally Posted: December 29, 2014 at 02:00 PM (#4869414)
For a starting pitcher WPA will fall way short of WAR, since the baseline is league average. An average pitcher in 275 innings (assuming he's not remarkable in situational pitching) will have the same WPA as one pitching zero innings.

But for relievers, the baseline for WAR is pretty close to league average, so WAR and WPA over the long term should be in the same ballpark. First guy I checked was Mo, and he comes out at 56.6 by either WAR or WPA. Some others:

pitcher, WAR, WPA

Wilhelm 50.1, 31.1
Goose 41.8, 32.5
L Smith 29.4, 21.3
Hoffman 28.0, 34.1
Wagner 27.7, 29.0
Nathan 26.1, 31.0
Fingers 25, 16.2
Quiz 24.9, 20.7
Sutter 24.5, 18.2
Franco 23.7, 19.1
Papelbon 22.2, 27.5
   28. Rally Posted: December 29, 2014 at 02:12 PM (#4869430)
What Dr. Chaleeko is doing in post #1 is not double counting of wins, that was my first impression but once I grokked the post I realized it's something different.

What that method does is give the reliever full credit for the extra leverage. In the WAR calculation we are giving the relief ace half of the credit for high leverage. If you don't know why, then look into BBref's documentation or search TangoTiger's site for "chaining". The quick explanation is that if a relief ace (Mariano) goes down, his high leverage innings are given to the next best reliever (Soriano, Robertson) and not the rookie they call up from Columbus to replace him on the roster.

So for Gossage, (WAR - WAAadj) + (WPA - WPA/li) = (41.8 - 7.7) + (32.5 - 15) = 51.6.

WAAadj is that half leverage credit. Double it to give him full leverage credit, and it's 41.8 + 7.7 = 49.5. Seems close enough.
   29. Harvest Posted: December 29, 2014 at 02:29 PM (#4869442)
I thought I would join the discussion by posting my own method of reliever value here. I use shutdowns and meltdowns (and therefore indirectly WPA) to measure the tactical benefits of using a reliever that WAR doesn't pick up. Specifically, I view a closer that comes into the game for the 9th and records a SD (or any other reliever in such a situation) as having essentially "won" that inning for their team, or equivalently 1/9 of that game. On the other side of the coin, a reliever that pitches an inning and records a MD has "lost" 1/9 of the game. I put "won" and "lost" in quotes because these are obviously not the same as wins in the Wins Above Replacement sense, and I don't view them as such.

This can be extended to full seasons and entire careers by looking at the piecewise "wins" and "losses" a reliever produced by recording SD and MD; this generalization makes sense as to record a SD or MD, one needs to affect WPA by plus or minus 0.06, so a generic closer (read: one-inning guy) with 9 SD has increased WPA by at least 0.54, enough to ensure a victory for their team in one game from the starting point assuming the contributions from hitters on their team are neutral.

The exact formula I use is (SD - MD)*(IP/(9*Gms)). The usage component of this should stabilize at slightly over 1/9 for the typical closer, meaning that they would gain points as they accrue shutdowns. I apply this after making my usual WAR-based calculations for evaluating a player; it brings Hoffman and Wagner up from being joke candidates to respectability, although they're both still well below my ballot. FWIW I have both of them above Rollie Fingers. It also moves Mariano from someone that I would support anyway to an absolute lock (his starter doppelgänger in terms of overall value in my system is Phil Niekro), and pushes the Goose into the Hall where previously he was nowhere near it.

Comments, suggestions, etc. are welcome.
   30. Sunday silence Posted: December 29, 2014 at 02:53 PM (#4869456)
In my opinion the BBWAA did something dumb when they gave Fingers an MVP award and then compounded the dumbitutde by using that as a reason to put him in the Hall of Fame...-Bill James


which is amusing and ironic because James spent a good portion of that Historical Baseball Abstract using just such criterion as one of his measuring sticks.
   31. bobm Posted: December 29, 2014 at 09:45 PM (#4869673)
[25] Thanks, bobm, although we also need how many outs when the RP arrived on the scene. Ideally, margin of error as well. I think we know that Hoffman is not coming into a 3-2 game in the 9th with a man on 3rd and nobody out (an extreme, of course, but it would be cool if we could figure out actually degree of difficulty. 21 is an interesting start).

For degree of difficulty, I left aside the inning and looked at lead and base-out state for what I perceived to be the hottest fires. That is, how many times did the RP enter into one of the following states:

Tied: 0 Outs or 1 Out, At least 1 runner on 2B or 3B
Lead +1: 0 outs, men on; 1 Out, At least 1 runner on 2B or 3B
Lead +2: 0 outs, 2B and 3B, or bases loaded
Lead +3: 0 outs, bases loaded

Fingers: 10%
Smith: 6.3%
Hoffman: 2.6%
Wagner: 1.3%
Rivera: 1.0%
   32. Mike Emeigh Posted: January 01, 2015 at 03:26 PM (#4871025)
Before 1975, relief pitchers almost always entered games either (a) with runners on base mid-inning or (b) as a result of the previous pitcher leaving for a pinch-hitter.

Since 1995, most relief pitchers enter games at the start of an inning.

Between 1975 and 1995, relief pitcher usage patterns changed drastically. This has affected how we perceive relievers whose career started during that time period (Lee Smith and Jeff Reardon, primarily) and makes it very difficult to compare relievers of this era with relievers of the prior era. IMO, it's about the same level of difficulty as comparing 19th century starting pitchers with today's starters.

-- MWE
   33. SoSHially Unacceptable Posted: January 01, 2015 at 03:35 PM (#4871038)
Tekulve finished 638 of 1050 games. that's not a middle reliever.



I think of Timlin before him, but even he finished close to 500, with several seasons with double-digit saves.

For a pure middle reliever, I always think of Scot Shields: 10 seasons, never more than 7 saves in a season, 114 GF out of 491 games, 139 ERA+. Also gets bonus points for being the rarest of birds, the middle reliever who spent his entire career with one franchise.
   34. Ziggy's screen name Posted: January 01, 2015 at 04:09 PM (#4871077)
I'm not a HOM voter, but I feel compelled to say something before lots of modern relief pitchers hit the ballot. The HOM is our (well, your) means of improving on the HOF. Voters for the HOF often conflate "a good story" with "actual contribution to the team". I would like to ask you not to do the same. LI is a good way to represent how exciting a particular game situation is, and high leverage situations produce the memorable moments that make watching baseball so much fun. But it is a terrible way to measure a player's contribution to his team. This has, of course, been addressed many times over many years on this site, but a run prevented in the first inning matters just as much as a run prevented in the ninth. Maybe it is possible for a relief pitcher to be productive enough to merit a place in the HOM, on that question many of you are more qualified to say than I am. But the leverage of the situations in which a player finds himself does not effect the value of his contribution, and shouldn't be used to measure it.
   35. Chris Cobb Posted: January 01, 2015 at 05:07 PM (#4871128)
Ziggy, the reason at least some of us are inclined to consider leverage index in assessing relief pitchers is that their usage in terms of the number of innings they throw is deliberately limited in order to have them available at maximum effectiveness in high-leverage situations. If they were employed in a fashion that enabled them to throw as many effective innings as they could healthily manage (that's the goal, at least, with starting pitchers), then they would throw quite a few more innings than they do, at a somewhat lower level of effectiveness. In order to compare relief pitchers to starting pitchers, some consideration of leverage is therefore appropriate in assessing their merit (which is not necessarily the same thing as value) in comparison to starting pitchers. How much to factor in leverage is very much an open question, but to rule out leverage altogether is essentially to conclude that no relief pitcher has as much value as HoM-worthy starting pitchers. That may be the position you are advocating for, but the electorate as a whole is not convinced of that point. It appears the electorate is also trying to get a handle on how much other situational factors of the game state in which the reliever enters should also be taken into consideration. As the HoM has elected only 3.5 relief pitchers so far--Wilhelm, Fingers, Gossage, and Eckersley, I don't think the history of the institution suggests a headlong rush to elect relievers. To the contrary, I think the trend now is to consider that we have elected too many, and to view the case of any non-Rivera reliever candidate as doubtful.
   36. Rob_Wood Posted: January 01, 2015 at 11:05 PM (#4871408)
Is a run scored in the first inning just as important as a run scored in the tenth inning?
   37. Howie Menckel Posted: January 01, 2015 at 11:13 PM (#4871415)

thanks again, bobm, and another crumb of extra credit for Hoffman.
   38. bobm Posted: January 01, 2015 at 11:51 PM (#4871457)
thanks again, bobm, and another crumb of extra credit for Hoffman.

You're welcome.

Before 1975, relief pitchers almost always entered games either (a) with runners on base mid-inning or (b) as a result of the previous pitcher leaving for a pinch-hitter.

Since 1995, most relief pitchers enter games at the start of an inning.


% of games entered in base-out state 0-on, 0-out during innings 2-end

Rollie Fingers 43%
     Lee Smith 67%
Mariano Rivera 77%
Trevor Hoffman 81%
  Billy Wagner 86%
   39. bobm Posted: January 02, 2015 at 12:21 AM (#4871493)
[38] Goose Gossage 43% of (relief) games entered in base-out state 0-on, 0-out during innings 2-end
   40. Squash Posted: January 02, 2015 at 03:29 AM (#4871586)
Is a run scored in the first inning just as important as a run scored in the tenth inning?

Clearly it's more important, since if you scored that run in the first inning you would never get to the tenth inning.
   41. jdennis Posted: January 02, 2015 at 10:01 PM (#4871960)
I rate pitchers using ERA+ and WHIP+, along with IP for a scalar value, and I account for league progression and unearned runs and the like. I did some work on R and found that ERA+ and WHIP+ should be close to equally weighted, at least when both numbers are above 100. I think this is good because it accounts for pitchers who pitch to misses/record Ks but maybe more BBs, and those who pitch to contact/strand runners, and does not favor one or the other. I believe this is fair for relievers because they post more extreme rate stats but in fewer innings, so the career values even out for HOF/HOM length careers. Interestingly, the model endorses a lot of the controversial reliever CYA. I am adding ancillary skills (hitting, fielding) and seeing how my rankings change. Obviously, this doesn't affect modern relievers nearly as much.

As I've mentioned before, I have Hoffman rated as a borderline case among all pitchers, and I believe he would be third among relievers currently in the hall, behind Eckersley and Wilhelm. He certainly had more career value than Sutter, Fingers, or Gossage. Not to mention Smith, whose saves record he extended by 123. He has a strong identity with one team, which I like honestly. I don't see him as an inner-circle guy, but I would confidently vote him in.

By comparison, Rivera I have all the way up at 12th, with his consistently obscene ERA and WHIP numbers. Rivera by my model is an inner-circle HOFer/HOMer, without a shadow of a doubt, and the best reliever of all time by a mile.

Unfortunately, I have not evaluated Wagner yet. I guess my main point is, I'd vote for Hoffman.
   42. Russ Posted: January 03, 2015 at 12:05 AM (#4872000)

Clearly it's more important, since if you scored that run in the first inning you would never get to the tenth inning.


This is something that is simply not always true. Baseball innings are not independent of one another... What happens in one inning depends on what happened in previous innings. I do not know what the causal effect of a run in the bottom of the first inning is, but I surely know what the causal effect is of a run in the bottom of the 10th. Therefore, from the perspective of value provided to winning a game, these are quite different things.
   43. Walt Davis Posted: January 04, 2015 at 11:47 PM (#4872983)
You know I'm bored at work when I wander into an HoM thread. :-) Nitpicks:

The quick explanation is that if a relief ace (Mariano) goes down, his high leverage innings are given to the next best reliever (Soriano, Robertson) and not the rookie they call up from Columbus to replace him on the roster.

Isn't this true for DH as well? When the DH goes down, the PA go to the best available hitter. Largely true for 1B, LF, RF. Definitely not the case for C and SS, probably somewhere in between for 2B, 3B and CF.

Hoffman debuted in 1993 and made $80 million.

Seems a bit odd to look at career earnings rather than AAV. Rivera topped out at about $15 M a year, basically the same as a top DH (Ortiz), substantially below a good starting OF (or even Soriano).

Also Hampton signed a rather infamous 8/$120 M contract that beats Hoffman's career earnings easily (if $80 M is accurate) ... so not sure what the source for career earnings is. If you're using b-r, be careful as they tend to miss signing bonuses, buyouts, etc. For example they show Hampton's contract years as $106 M. You've got Belle listed there as $60 M career earnings -- even b-r has his career earnings at about $100 and that's under-reporting his 5/$65 contract with Baltimore as $61 M.

The only news reports I could find on Hoffman contracts were for the last three seasons which look accurate at b-r, depending on incentives earned. But earlier values could be out of whack.

Not a big deal, the comps seem generally reasonable -- Shawn Green or a short-career SP like Wood. Wood had only 400 more career IP than Wagner and 300 more than Hoffman (at much lower ERA+ and leverage) so that sounds superficially correct. Also WAR/WAA puts it at 27/16 for Wood and 28/17 for Wagner and 28/14 for Hoffman. Of course he also never really got the chance to sign a big FA contract when a healthy starter, with better timing he probably earns well over $100 M.
   44. Ziggy's screen name Posted: January 09, 2015 at 10:28 AM (#4876701)
That they aren't independent doesn't mean that they aren't equal in value. Sure you'll play the game differently if you have a lead rather than if you're tied, and whether you get a chance to score that run in the 10th depends on whether you have lead or if you are tied. So that they aren't independent is correct. But the run in the first wins the game just as much as the run in the 10th does.
   45. David Concepcion de la Desviacion Estandar (Dan R) Posted: January 10, 2015 at 11:26 AM (#4877453)
Walt Davis, yes, and that's why replacement level for DH is basically a pitchers-excluded-league-average hitter. This is the source of positional value.

Ziggy, if the point is to win games, then you win more games by saving your limited-innings relievers for the highest-leverage situations than you do by using them in the first inning. A run in the 1st does *not* matter/count as much as one in the 10th, because there's no prize for best Pythagorean record, only for best won-lost record. In fact, for a backwards-looking valuation project like the HoM (rather than a forward-looking projection system), I think that philosophically we should really be using total Win Probability Added above replacement rather than batting runs above replacement (and also weight defensive plays by their leverage--a HR-saving catch in the 10th is worth more than one in the 1st). I suspect the only reasons we don't are 1. that we haven't had access to historical WPA for all that long and 2. over a 20-year career, WPA should track pretty closely to BRAA/10 for most players. The only place you'd expect a big gap over a full career is with relievers, since their usage pattern is designed to maximize their impact, so that's why it only tends to get invoked in their cases. Now, whether you want to go further into "Championship Probability Added" is another question, but there's certainly a strong case to be made that Joe Carter or Bill Mazeroski had more valuable careers than Ernie Banks did.
   46. Rob_Wood Posted: January 10, 2015 at 12:56 PM (#4877512)
Ziggy, I was trying to be funny by pointing out a silly extreme that (I thought) everyone would understand. I honestly thought that everyone would acknowledge the obvious fact that a run scored in the tenth inning contributes more to its team winning the game than a run scored in the first inning.

Consider a new-fangled NBA rule in which overtime is now "golden goal" (i.e., the first basket wins). Suppose the typical NBA final score is around 125-120 and many games are blow-outs. Then, I hope everyone would agree that a basket scored in overtime is worth more than the first basket scored in a game. To be clear, we are not just talking about that one game that went to overtime. We are talking about all the games in the season. So a basket scored in overtime *always* wins the game (by rule). But the first basket scored in some random game (that may or may not eventually go to overtime, and in which there are many many made baskets by both teams, and the game may well prove to be a blow-out (by either team)) obviously has limited "win value".

If we cannot agree on this, then there is no point in continuing this discussion.
   47. Chris Cobb Posted: January 10, 2015 at 01:17 PM (#4877527)
Dan R,

Agreed on the principle of leverage as meaning not all runs created are of equal value, but strong disagreement on the philosophical orientation of the whole HoM project to WPA above replacement. If we're going to take that kind of stance, what's the argument for minor-league credit when a player is blocked or excluded from the majors? The functionality of wins above replacement level is its estimation of value in a neutral win-probability context. It's conceivable that Joe Carter had a more valuable career than Ernie Banks, but is that due to any extraordinary merit of Joe Carter's or to his good fortune to be in a place where the contextual value of his skills was maximized?

I think it's good for us to be looking at RE24 and WPA and related stats, but I don't think we have a handle on how fully contextualized value is related to merit or of the implications of looking at this sort of game-context-based value for the project of being fair to all players from all eras.
   48. Kiko Sakata Posted: January 10, 2015 at 02:52 PM (#4877616)
In fact, for a backwards-looking valuation project like the HoM (rather than a forward-looking projection system), I think that philosophically we should really be using total Win Probability Added above replacement rather than batting runs above replacement (and also weight defensive plays by their leverage--a HR-saving catch in the 10th is worth more than one in the 1st). I suspect the only reasons we don't are 1. that we haven't had access to historical WPA for all that long and 2. over a 20-year career, WPA should track pretty closely to BRAA/10 for most players. The only place you'd expect a big gap over a full career is with relievers, since their usage pattern is designed to maximize their impact, so that's why it only tends to get invoked in their cases.


Dan R, I think there are issues with WPA that make it not perfectly apt for this - specifically, it overvalues close games vis-a-vis blowouts when, retrospectively, wins are wins. But I've tried to put together something that does exactly this - Player won-lost records. I have a really long article going into the gory details of what I'm doing, how, and why - including how my system differs from pure WPA - here. This is all done using Retrosheet play-by-play data. Retrosheet has done extraordinary work and (counting deduced games) they have full data going back to 1948 and have released at least partial data for every season back to 1934, but that does still leave a bit of a gap in terms of how one would incorporate players from the 1930's and earlier (and, of course, all non-white-MLB players/stats), which is a large reason why I haven't tried to jump into HOM voting.
   49. Rob_Wood Posted: January 10, 2015 at 06:21 PM (#4877728)

Kiko, I really like what you have done. Good job.

P.S. Please join in HOM discussions and voting!
   50. John DiFool2 Posted: January 10, 2015 at 06:47 PM (#4877748)
One more issue...

Note that the overwhelming consensus is that Rivera is the overwhelming consensus #1 reliever of all time.

At no other position do we see such a gap, except arguably in RF (The Babe vs. The Field).

The question is why. Yes, the surface assumption is that he was the greatest simply because he was, full stop. OK...

But maybe the position is more challenging, in terms of career value, than is popularly supposed. And certainly Fruitbat deserves full credit for remaining at or near his peak for close to 20 years. But there are plenty of closers who have burned out at relatively young ages and never came close to lasting that long, tho their peaks match his (thinking Papelbon here just off the top of my head for a recent example, ).

And things were worse in that regard during the heydays of Fingers and Goose, given that these guys typically threw more innings; the recent remodeling of the closer role was in part a reaction to this perceived heavy workload, to allow these pitchers to have longer careers at near peak levels. And it has apparently worked, with quite a few recent pitchers on the overall saves leaderboard.

Just saying that these long-term relievers perhaps deserve a bit more credit for being that durable.
   51. alilisd Posted: January 10, 2015 at 07:07 PM (#4877766)
Is there really any more of a gap between Rivera and whomever as between Ruth and Aaron?

But more to the point, I think the difficulty of the"position" is in repeating success year after year. Many pitchers have been great for a few years here and there, or a few years and then flamed out. It does appear to be exceedingly difficult to maintain success in the position for what would normally be considered the length of a HOF career, say 15 or so years.
   52. Kiko Sakata Posted: January 10, 2015 at 07:13 PM (#4877768)
Kiko, I really like what you have done. Good job.

P.S. Please join in HOM discussions and voting!


Thanks, Rob. I will try to do something for the 2016 election - now that I have about 10 months to do so. As I said, I just need to figure out how to incorporate what other systems say about the really old folks with my system that's missing them. I guess the first step is to figure out exactly which "old guys" I need to incorporate. I assume it would violate the spirit (if not the explicit rules) of the project to just throw up my hands and say "well, anybody pre-1930 who hasn't been elected yet probably wouldn't make my ballot anyway", right?
   53. DL from MN Posted: January 10, 2015 at 08:13 PM (#4877830)
I don't really think of players that way. I think more of choosing up sides. Which guy would I want on my team if I had him for his whole career. It certainly isn't Ernie Banks' fault he played for the Cubs.
   54. Bleed the Freak Posted: January 10, 2015 at 09:39 PM (#4877872)
The really important question. "Enter Sandman" or "Hell's Bells"?


While a great question, too bad neither is from the pinnacle of either artists career, either Bon Scott AC/DC or 80s era Metallica.

Orion and Live Wire are two of my recommendations.
   55. Bleed the Freak Posted: January 10, 2015 at 10:07 PM (#4877895)
Dan R, I think there are issues with WPA that make it not perfectly apt for this - specifically, it overvalues close games vis-a-vis blowouts when, retrospectively, wins are wins. But I've tried to put together something that does exactly this - Player won-lost records. I have a really long article going into the gory details of what I'm doing, how, and why - including how my system differs from pure WPA - here. This is all done using Retrosheet play-by-play data. Retrosheet has done extraordinary work and (counting deduced games) they have full data going back to 1948 and have released at least partial data for every season back to 1934, but that does still leave a bit of a gap in terms of how one would incorporate players from the 1930's and earlier (and, of course, all non-white-MLB players/stats), which is a large reason why I haven't tried to jump into HOM voting.


This is great stuff Kiko and I sure hope that you can vote in the 2016 election, share your insights from the mountain of Retrosheet information you have compiled, and maybe you have your own personal hall to share?

I find this table to be one of the more interesting tidbits:

http://baseball.tomthress.com/Leaders/UberLeaders.php?y1=1949&y2=2013&p=0&e=1&w=.051&a=1&r=.949&na=1&nr=1&c=1&b1=1&b2=1&b3=1&ss=1&lf=1&cf=1&rf=1&dh=1&ph=1&pr=1&o1=1&sp=1&rp=1&psw=0&psa=0&psr=0>=0&ga=0&n=1000

Would this key stat represent how you would vote...I plan to post the best eligibles in the 2016 ballot discussion thread.

Notable relievers (less than 60 GS) come out as follows:
31.5 (1) - Mariano Rivera
27.6 (2) - Hoyt Wilhelm
22.8 (3) - Rollie Fingers
19.7 (4) - Kent Tekulve
18.3 (5) - Dan Quisenberry
18.1 (6) - Trevor Hoffman
17.4 (7) - Goose Gossage
17.3 (8) - Billy Wagner
14.7 (9) - Doug Jones
14.7 (10) - Roy Face
14.7 - Tug McGraw
14.4 - Mike Marshall
14.2 - Joe Nathan
14.1 - John Franco
13.8 - Sparky Lyle
12.7 - Tom Henke
12.6 - Lee Smith
12.6 - Mark Eichhorn
12.4 - Terry Forster
12.2 - Jesse Orosco
11.1 - Francisco Rodriguez
11.0 - Keith Foulke
10.1 - Don McMahon
9.90 - Ron Perranoski
9.80 - Jonathan Papelbon
9.80 - Bruce Sutter
9.10 - Jeff Montgomery
8.90 - John Hiller
   56. Kiko Sakata Posted: January 10, 2015 at 10:13 PM (#4877900)
Would this key stat represent how you would vote...I plan to post the best eligibles in the 2016 ballot discussion thread.


Thanks, Bleed. I'm glad you like it. That table would certainly be my starting point.
   57. Bleed the Freak Posted: January 17, 2015 at 08:50 AM (#4882554)
Profiling Lee Smith but really a summation of relievers in general from Kiko's site, adding detail to the chart I posted previously:

http://baseball.tomthress.com/Articles/2013HOF/LeeSmith.php

Unlike the key stat data quote above, Hoffman fairs great in the Lee Smith link, a hair or notch ahead of all but Rivera while Wagner fades into the crowd.
   58. Kiko Sakata Posted: January 19, 2015 at 09:21 PM (#4884255)
Unlike the key stat data quote above, Hoffman fairs great in the Lee Smith link, a hair or notch ahead of all but Rivera while Wagner fades into the crowd.


The difference is that the key stat data you quote above (and elsewhere) uses only context-neutral (teammate-adjusted) data. But if you don't take account of context, then basically no relief pitchers end up worthy. If you compare the numbers you posted in #55 here with the numbers you posted in #28 of the ballot discussion thread, Mariano Rivera (31.5) ends up below Jim Rice (57.7) and Kirby Puckett (41.0), among others. The only way to get relief pitchers into the coversation is to incorporate context (as BB-Ref does w/ WAR, for example).

This sort of segues into an issue I'm wrestling with. Within context, including postseason, Mariano Rivera ranks among the top 25 players in pWins over positional average (pWOPA) among players for whom I've calculated Player won-lost records and 53rd measured against replacement level (pWORL). Trevor Hoffman, meanwhile, ranks 133rd in pWOPA, but that's the only thing he has going for him; switch to pWORL and he's outside the top 250 (and my data only go back to the 1930's, so outside the top 250 is pretty much outside the HOM).

One thing that I think was talked about upthread is that, generally speaking, relief pitchers have historically tended to have very short careers. So, one way that one could boost relief pitchers vis-a-vis other positions would be to give them a positional boost, as I think a lot of people do with catchers, for example. So, if, for example, you were to give a 50% positional adjustment for relief pitchers (which is arguably defensible in my system, that would boost Hoffman's pWORL up to ~40, which would put him around the top 90 or so in career pWORL in my first link of this comment, around guys like Jeff Kent, Jim Bunning, and Edgar Martinez. So, that would certainly be enough to put Hoffman (and a few other relievers) into the HOM conversation.

But, to be consistent, if you did the same with Mariano, who already has 48 pWORL - adding another 50% on top of that gets you up to 72 pWORL, which would put Rivera in the top 10 in pWORL in my first link, between Warren Spahn and Alex Rodriguez. And as great as Mariano Rivera was, that seems silly. He wasn't one of the top 10 (or 20) players in major-league baseball since the 1940's.

This is equally true with BB-Ref's WAR, for example. Trevor Hoffman has 28.0 career WAR. To get him up to the, say, 50-60 range that tends to be the lower bound for HOM consideration, you'd need to multiply that by 2. Mariano Rivera has 56.6 career WAR (regular season only). Double that and he's at 113 WAR which slots him between Alex Rodriguez (116.0) and Lou Gehrig (112.4). Again, Mariano Rivera was great, but Lou Gehrig great?

So, where I'm coming down is essentially that it can't really make sense to make an adjustment that says that Mariano Rivera was THAT great, but anything less than that is going to leave pretty much every other relief pitcher in major-league history outside of HOM consideration. In effect, Mariano Rivera proved that it IS possible to put together a long enough, excellent enough career as a relief pitcher to warrant HOM consideration, but nobody else has come particularly close to doing so except for Mariano.

Does that make sense and is it "fair" to "all positions" - i.e., is it fair to relief pitchers?
   59. Ziggy's screen name Posted: January 19, 2015 at 10:44 PM (#4884292)
Take a game in which the visiting team scores a bunch of runs in the first inning, and whose pitchers/defense don't totally melt down. Then there are no high leverage situations for the pitchers in this game. Let's let them score 75 runs in the first inning, and let the pitcher throw a no-hitter. Given the huge lead, and hence little leverage, if we used leverage to evaluate pitcher performance we would need to conclude that the no-hitter (with a 75 run lead) reflected less well on the pitcher than a much worse performance in a much closer game. But surely the pitcher's merit is independent of the performance of his offense (let it be an AL game, he didn't even bat). One of the problems with using leverage to evaluate pitchers is the same as one of the problems with using RBI to evaluate hitters, you end up giving credit to players (or relative demerits, since you're using this to compare players) for things that were done by their teammates.
   60. theboyqueen Posted: January 19, 2015 at 11:39 PM (#4884324)
One thing that I think was talked about upthread is that, generally speaking, relief pitchers have historically tended to have very short careers. So, one way that one could boost relief pitchers vis-a-vis other positions would be to give them a positional boost, as I think a lot of people do with catchers, for example. So, if, for example, you were to give a 50% positional adjustment for relief pitchers (which is arguably defensible in my system, that would boost Hoffman's pWORL up to ~40, which would put him around the top 90 or so in career pWORL in my first link of this comment, around guys like Jeff Kent, Jim Bunning, and Edgar Martinez. So, that would certainly be enough to put Hoffman (and a few other relievers) into the HOM conversation.


Am I missing something? The purpose of a positional boost generally is to reflect the scarcity of a particular (and necessary) skillset, in order to reflect the true value of a performance. We know that Larry Walker can't catch effectively, otherwise he would have been a catcher. We don't know what kind of relief pitcher Kevin Appier would have been but there is enough data to suggest he probably would have done quite alright. The idea of giving a relief pitchers a positional boost because they don't actually pitch much and tend to have short careers seems like it's begging the question. In fact, the more defensible position would be to give them a penalty for not being able to start.

I don't understand the need to "boost" relief pitchers vis-a-vis other positions for any reason other than the purely psychological. What is so wrong with the idea that Mariano Rivera is the only historically great relief pitcher, that we need to make up fudge-factors to correct this "problem"?
   61. Kiko Sakata Posted: January 19, 2015 at 11:48 PM (#4884328)
I don't understand the need to "boost" relief pitchers vis-a-vis other positions for any reason other than the purely psychological. What is so wrong with the idea that Mariano Rivera is the only historically great relief pitcher, that we need to make up ways to correct this problem?


I guess the only thing "wrong" with that is that Hoyt Wilhelm, Goose Gossage, and Rollie Fingers are already in the Hall of Merit and short of giving relief pitchers "extra credit", I can't see any way to get around viewing that as a huge mistake. And I don't necessarily want to jump into a project that's been going on for 100+ "years" and start by laying out the mistakes that the electorate has made.
   62. theboyqueen Posted: January 20, 2015 at 01:23 AM (#4884356)
Well then what you are doing is something more akin to HOF/HOM monitor than a valuation tool, no?
   63. Chris Cobb Posted: January 20, 2015 at 08:30 AM (#4884386)
No one who has been involved with the HoM project for a long time thinks it is free of mistakes, and each of us could probably quickly produce our personal list of the biggest ones! A new voter does not need to accept earlier selections as precedent. However, if you reach the conclusion that the HoM's mandate to be fair to all positions and eras means that the best relief pitchers ought to be represented, then it's perfectly appropriate to consider adjustments--indeed, analogous to the ones catchers get--to put them on a level footing with candidates whose playing time was not limited by the nature of the position. Whether relief pitcher is a position and whether, as a position, it deserves to have comparable stature to starting positions and not the stature of "fourth outfielder" or "utility infielder" is a key question that won't be answered purely on quantitative grounds.

Rivera's performance raises questions about the appropriateness of reliever adjustments that could not have been framed a decade ago, and it will probably take another three decades of observing relief pitchers to discover how much of an outlier he really is. We do our best with the information we have to work with.

To get a sense of the range of agreement/disagreement among the electorate over the election of relief pitchers, one could look back at the election results and ballot discussion threads from the years they were candidates. That might be quite interesting.
   64. DL from MN Posted: January 20, 2015 at 10:10 AM (#4884441)
Wilhelm pitched 2200 innings and has a WAR of 50 with a WAA of 27. That doesn't look like a mistake to me. Gossage is marginal but I'd probably not have him on ballot at the moment. With hindsight Fingers looks like a big mistake. I'd be fine with Rivera and Wilhelm as the only two relievers deemed worthy.
   65. Kiko Sakata Posted: January 20, 2015 at 10:16 AM (#4884444)
Thanks for the responses. I'm leaning toward not necessarily thinking of "relief pitcher" as a position distinct from "pitcher", which would probably end up with Mariano Rivera being the only relief pitcher in my personal Hall of Fame (I think Hoyt Wilhelm tends to end up around the borderline of a "personal Hall" around the size of the Hall of Merit and I would have to go back and re-create past ballots to see how I might have voted for him).
   66. Tom T Posted: January 20, 2015 at 11:20 AM (#4884488)
I honestly thought that everyone would acknowledge the obvious fact that a run scored in the tenth inning contributes more to its team winning the game than a run scored in the first inning.


I think the problem here is conflating the event of *scoring* a run with the *run* itself. Semantics, here, I acknowledge ("scoring a run" vs. "a run scored"), but tghis can be very important when trying to convey leverage. I know this phrasing confused the hell out of me for a while until I figured out that folks were being imprecise. (Not that I can explain it clearly, but I haven't had to teach it in front of a classroom yet....)

So, to Ziggy's point....If a team wins 1-0, it does not matter --- a posteriori --- *when* the run scored. The run contributes precisely the same amount to the outcome (a win) regardless of the inning in which the event occurred.

And, given that a team does not get "extra credit" for runs that score in the 10th as opposed to those scored in the 1st inning, these runs (at the end of the game) have the exact same value in determining the team's overall record, playoff opportunities, etc.


That said, I don't believe the Ziggy would disagree with the argument that the context in which the run scores in the 10th leads to a substantially greater reduction in the variance of the outcome of the game than does a run scored in the 1st inning.

So, the underlying context is that it is clearly more vital to a team's desired outcome that they score a run if it is 0-0 in the 10th than in the 1st, but the contribution of the run (i.e., increasing the team's score by 1 run) is fundamentally not different. Thus, the leverage is capturing the context, and scoring a run leads to a more certain outcome, but the run (in and of itself) does not meaningfully contribute more to the win if it is N-(N-1) victory, independent of how many innings are played beyond the 5th.

Eh, that may all not be quite clear....

What I am trying to get at is that asking whether "a run scored in the 10th more important than a run scored in the 1st" is the wrong question --- they really aren't. BUT...*scoring* a run in the 10th IS (undoubtedly) more important than *scoring* a run in the 1st. Rob's question is fundamentally asking this, but the jargon isn't clear.

(I will now go back to demonstrating my ignorance through lurking, having -- I am sure I will be told -- filled my mouth quite well with my shoe.)
   67. Chris Cobb Posted: January 20, 2015 at 11:45 PM (#4885037)
Excellent post, Tom T!

On the subject of reviewing the history of the treatment of relief pitchers in the HoM and considering whether the electorate has made mistakes, I'll note that I took a stroll down memory lane this evening and visited the 2000 Ballot Discussion thread (the year Gossage and Fingers were elected) from actual year 2007, the Rich Gossage thread (mainly 2007), and the Relief Pitchers thread, which built up several hundred posts in 2006, when Hoyt Wilhelm was on the ballot.

The tools we have available for evaluating relievers, and the more-or-less accepted Sabermetric wisdom have changed a lot in nine years. On the Relief Pitchers' thread, we are just discovering that we can get Leverage Index data from BP and discussing what exactly that means. We're working with Clay Davenport's WARP1 (BP's in those days) and Win Shares as our two comprehensive metrics. Both of those try to estimate leverage from holds and saves--a far cry from what we have now from the play-by-play data. Neither system includes the ERA adjustment for relievers that is now a (largely) accepted part of Baseball-Reference's WAR. If you go over to Clay Davenport's site and check out the views of WARP, with its XIP adjustment and no ERA adjustment, you'll see that WARP1 gives Rich Gossage and Dave Stieb equivalent career values, where B-R's WAR now shows Stieb 10+ WAR ahead of Gossage for their careers. On that thread, I noted that WARP1 valued John Smoltz's top year as a reliever (2003) about the same as his first year back as a starter (2005), and it still does, but none of the second- (third?-) generation WAR systems support that view at all.

Howie Menckel may be right to say that the electorate is relying too much on B-R's WAR, but there's no doubt that it now includes adjustments as a matter of course that we couldn't possibly make then and that were being hammered out on the cutting edge of sabermetric data analysis around that time.

As an aside, the 2000 election results were really interesting in light of the way later elections went--there were large and rapid shifts in the evaluation of a lot of players in the course of the next 10 elections. (Maybe it was around that point that Dan R came back strongly into the HoM project?)
   68. Rob_Wood Posted: January 21, 2015 at 06:48 AM (#4885070)

Some inning by inning versus win pcts data is available over at Baseball-Reference. As far as I can tell, in 2014 major league teams had a little lower than a .700 win pct in games in which they scored in the first inning. And major league teams had slightly higher than a .900 win pct in games in which they scored in the tenth inning (or any other extra inning). If anyone is a database wizard and has access to the raw data, it would be interesting to look at these win pcts by inning.

Since I am a San Francisco Giants fan, I tallied them for the 2014 Giants. The tally was done by hand so there may be a few minor errors. But I get the win pcts by scoring inning as follows:

1 = .720
2 = .739
3 = .643
4 = .775
5 = .615
6 = .775
7 = .745
8 = .833
9 = .560 (largely due to the fact that home teams don't bat in the ninth inning if leading)
-------
10 = 1.000
11 = .500
12 = 1.000
13 = 1.000
14 = 1.000


Based upon the posts in this thread, I honestly don't know if people are surprised by the disparity in win pcts between scoring in the first inning vs. scoring in the tenth inning. Since my hypotheticals have gone over like the proverbial lead balloon, here's another. Imagine a hypothetical world in which you have a fantastic pinch-hitter, say a slugger who can hit a home run every time up, but can only bat once a week. When would you use him? Leading off a game? Of course not. I don't think I would ever use him in the first inning of any game all season. To get the most bang for your buck, you want to save him until a key moment in a game arises in which a home run can tilt the game from a very likely loss into a very likely win. If I was his manager, he'd probably never bat before the 8th inning of any game.

Note: this is not the same issue where it can reasonably argued that a game in April is worth a game in September. For each game does indeed count the same (modulo opponent) towards winning the pennant. Such is not the case for runs scored by inning affecting the team's win pct.
   69. Rally Posted: January 21, 2015 at 11:18 AM (#4885209)
Kiko, I like your site. Good job.
   70. Kiko Sakata Posted: January 21, 2015 at 11:25 AM (#4885215)
Thanks, AROM. I love your work on WAR, so that means a lot.
   71. David Concepcion de la Desviacion Estandar (Dan R) Posted: January 22, 2015 at 08:54 AM (#4885751)
I believe I returned around the 1993-94 votes.
   72. Tom T Posted: January 22, 2015 at 11:01 AM (#4885847)
Okay, so while avoiding writing progress reports and proposals my mind has been ruminating on this.

I have to believe that someone out there has applied either some economic or behavioral modeling to address the concept that the "event" and the "outcome" can be categorized differently, providing a more obvious framework for explaining leverage? If so, can someone please point me to it?

Also, just because I figured out how to get around a flaw in my previous argument, an example to refute the idea that "a run" (vs. the event of "scoring") is more valuable in the 10th than the 1st would be to consider a game in which the home team hits a 3-run home run to win by 2 (e.g., 7-5 after training 5-4). Clearly the 5th and 6th runs contribute significantly to reducing the (game) outcome variance by inducing transitions within the "game-state" space, but the 7th run contributes no marginal value and is, therefore, well behind any runs scored in innings 1-8 in terms of producing the desired (binary) outcome for the game. Ergo, not all runs scored in the 10th are of greater value/contribution than runs scored in the 1st. However, the (singular) event of scoring that produced a transition in the game-state space (the 3-run homer in the 10th) was clearly a far greater contributor to reducing the variance than the same event of scoring (a 3-run homer) would be in the 1st. I suspect one could prove that all such events in the 10th ARE of greater value than any such event occurring in the 1st.

Like I said...this MUST have been formalized by someone along the way. As an engineer who works with random processes, there is not always a meaningful distinction between the "event" and "outcome" (e.g., "scoring a run" vs. "the number of runs scoring in a game" is clearly a different dichotomy than "a service call came in" vs. "the total number of calls to the center in an hour") so I'm not equipped with the right tools to think about this much further, and would greatly appreciate a nudge toward source material of interest.
   73. Tom T Posted: January 22, 2015 at 11:08 AM (#4885859)
Rob --- Cool data. Thanks for taking the time to compile it! I am struck that the reduction in variance of outcome across innings 1-7 appears to be essentially null (i.e., each inning is equally predictive of outcome) and that the dominant effect only occurs once a team can (in effect) push the game into a terminal state by its final at-bat (8th/9th and beyond).

How memoryless is the run scoring during a game? What are the conditional probabilities of scoring 1+ runs in the 8th if you have scored {1,2,...K} runs across innings 1-7 (individually or as composite events)? Again, I assume this information must be out there...presumably these are incorporated into the WPA tables? or are they simply based on the score differential?

Dang...have to somehow justify a sabbatical at some stage so I can blow off 6-months enjoying this material, instead. (Also, I doubt Manfred is as likely as Goodell to surreptitiously attempt to get me to wear concrete shoes.)
   74. Rob_Wood Posted: January 22, 2015 at 05:37 PM (#4886240)

Since it was easy to do, I wrote a program to simulate baseball games using the 2014 major league distribution of scoring by inning (the same distribution was used for every half inning). In the first simulation, one team was given a bonus of one extra run in the first inning in every eighth game of the season. In the second simulation, one team was given a bonus of one extra run in a late inning (ninth or later) if the game was tied with the proviso that they could not use the bonus run in more than 1/8 of their games.

The winning pct of the first team was .511 and the winning pct of the second team was .541. Clearly these are extremes as there exists no player who can manufacture a run on command. Or, from a relief pitching perspective, make as much of a difference as a single run in each appearance over the course of a season.

Anyway, I guess this proves the obvious that close/late runs are more valuable in securing wins than the same number of runs sprinkled randomly throughout a season.
   75. bobm Posted: January 23, 2015 at 01:35 AM (#4886412)
[68] Since I am a San Francisco Giants fan, I tallied them for the 2014 Giants. The tally was done by hand so there may be a few minor errors. But I get the win pcts by scoring inning as follows:

1 = .720
2 = .739
3 = .643
4 = .775
5 = .615
6 = .775
7 = .745
8 = .833
9 = .560 (largely due to the fact that home teams don't bat in the ninth inning if leading)
-------
10 = 1.000
11 = .500
12 = 1.000
13 = 1.000
14 = 1.000


I got slightly different numbers using retrosheet gamelog data:

   1     2     3     4     5    6     7     8     9     10    11     12      13      14
.720  .739  .643  .775  .615 .775  .745  .833  .560  1.000  .500  1.000   1.000   1.000 
.740  .739  .667  .775  .615 .775  .766  .833  .560  1.000  .500  1.000   1.000   1.000 


The average MLB rates were lower and, as far as I could tell, had a narrower range over innings 1-8:

   1     2     3     4    5    6    7    8    9    10    11    12    13   14   15     16     19
.629  .652  .662  .633 .657 .641 .647 .648 .531  .917  .790  .880  .870 .900 .800   .000  1.000 



   76. bobm Posted: January 23, 2015 at 01:45 AM (#4886413)
I tried to program the calculation for 2014 win pct by scoring inning for any selected team.

Here's my stab at the Royals, for comparison:

   1    2    3    4    5    6    7    8    9   10   11    12     13     14
.703 .650 .657 .698 .620 .630 .698 .743 .676 .500 .667  .000  1.000  1.000 


The Angels:

   1    2    3    4    5    6    7    8    9   10    11     12      13      14     16     19
.679 .762 .804 .696 .767 .826 .778 .680 .594 .800 1.000  0.500   1.000   1.000  1.000  1.000 



The D-Backs:

   1    2    3    4    5    6    7    8    9    10    11     12     14      15
.543 .533 .525 .529 .462 .467 .625 .575 .417 1.000 1.000  1.000  1.000   1.000 
   77. Rob_Wood Posted: January 23, 2015 at 02:22 AM (#4886419)
Thanks bobm for doing that. I think there is a typo in your 16th inning result (four teams scored in the 16th inning in 2014 and all four teams won their games [in the 16th inning, of course], so win pct should be 1.000).

From these results, I find that the 2014 major league per-inning cumulative win pct of a team that scores in an inning 1-8 was .646 (9th inning is special so we leave it aside), and the per-inning cumulative win pct of a team that scores in an extra inning was .876.

P.S. I just saw your Royals posting -- they may have hated the 12th inning in the regular season, but they loved it in the postseason!
   78. bbmck Posted: March 09, 2015 at 12:34 PM (#4909382)
Rk               Player ERAOPS+  WHIP   SO9 SO/W From   To   Age    G  GF     IP  FIP
1        Mariano Rivera  223   42 0.972  8.30 4.38 1996 2013 26
-43 1096 950 1216.2 2.63
2          Billy Wagner  187   49 0.998 11.92 3.99 1995 2010 23
-38  853 703  903.0 2.73
3     Jonathan Papelbon  185   51 0.996 10.42 4.84 2006 2014 25
-33  576 500  593.1 2.58 


Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 3/9/2015.

Wagner's entire career, Rivera and Papelbon without their first season during which they started some games.

Rk           Player ERAOPS+  WHIP   SO9 SO/W Year Age  Tm  G GF   IP  FIP
1    Trevor Hoffman  265   30 0.849 10.60 4.10 1998  30 SDP 66 61 73.0 2.04
2    Trevor Hoffman  226   32 0.907  8.00 3.43 2009  41 MIL 55 46 54.0 2.63
3    Trevor Hoffman  206   93 1.111 11.00 3.67 2003  35 SDP  9  7  9.0 3.03
4    Trevor Hoffman  198   47 0.936  9.76 4.87 1999  31 SDP 64 54 67.1 2.60
5    Trevor Hoffman  190   55 0.968  7.14 3.85 2006  38 SDP 65 50 63.0 3.46
6    Trevor Hoffman  177   39 0.920 11.35 3.58 1996  28 SDP 70 62 88.0 2.66
7    Trevor Hoffman  170   66 0.915  8.73 6.63 2004  36 SDP 55 51 54.2 2.74
8    Trevor Hoffman  162   62 1.054 10.93 3.40 1994  26 SDP 47 41 56.0 2.70
9    Trevor Hoffman  148   66 1.020 12.28 4.63 1997  29 SDP 70 59 81.1 2.70
10   Trevor Hoffman  144   61 0.995 10.58 7.73 2000  32 SDP 70 59 72.1 2.50
11   Trevor Hoffman  138   72 1.180 10.47 3.83 2002  34 SDP 61 52 59.1 2.04
12   Trevor Hoffman  135   74 1.116  6.91 2.93 2007  39 SDP 61 50 57.1 2.94
13   Trevor Hoffman  131   72 1.110  8.43 4.50 2005  37 SDP 60 54 57.2 2.50
14   Trevor Hoffman  116   81 1.144  9.40 3.00 2001  33 SDP 62 55 60.1 4.21
15   Trevor Hoffman  108   90 1.322  7.90 2.03 1993  25 TOT 67 26 90.0 4.01
16   Trevor Hoffman  105   93 1.163  8.78 3.71 1995  27 SDP 55 51 53.1 4.38
17   Trevor Hoffman  101   85 1.037  9.13 5.11 2008  40 SDP 48 42 45.1 3.99
18   Trevor Hoffman   69  127 1.437  5.70 1.58 2010  42 MIL 50 36 47.1 5.21 


Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 3/9/2015.

I just don't get how that guy is even in the discussion with Wagner/Papelbon and of course Rivera stands alone.


   79. bbmck Posted: March 09, 2015 at 01:13 PM (#4909399)
Player           ERAOPS+  WHIP   SO9 SO/W From   To   Age   G GS  GF    IP  FIP
Joe Nathan        104   95 1.378  6.76 1.41 1999 2003 24
-28 121 29  14 266.1 5.05
Trevor Hoffman    118   83 1.204  8.98 2.73 1993 1995 25
-27 169  0 118 199.1 3.74
Joe Nathan        237   39 0.934 11.14 4.32 2004 2009 29
-34 412  0 361 418.2 2.40
Trevor Hoffman    176   49 0.945 10.98 4.57 1996 2000 28
-32 340  0 295 382.0 2.51
Joe Nathan        126   70 1.153  9.63 3.18 2011 2014 36
-39 243  0 210 231.2 3.22 (did not play Age 35)
Trevor Hoffman    131   73 1.087  8.29 3.55 2001 2010 33-42 526  0 443 508.0 3.25 


Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 3/9/2015.

There is a guy to compare Hoffman to, more decent innings after peak vs worse peak.
   80. alilisd Posted: March 09, 2015 at 02:41 PM (#4909467)
Am I missing something? The purpose of a positional boost generally is to reflect the scarcity of a particular (and necessary) skillset, in order to reflect the true value of a performance. We know that Larry Walker can't catch effectively, otherwise he would have been a catcher.


Hm, I never thought of postional adjustment for catchers that way at all. I've always thought of them being necessary as catchers do not (and realistically cannot) play 150+ games at their position as other players do, and have shorter careers as well. It's an adjustment for playing time, both in season and career, not a scarcity adjustment.

Also, we absolutely do not know that about Walker. In fact, excellent offensive catchers are sometimes moved off the position to get their bats into the game more often. I don't see anything about the position which would preclude an otherwise skilled athlete capable of playing other positions in the field from playing catcher. Isn't it more often the slower, less athletic players who are put at the position due to their incapacity for other positions rather than the other way around?
   81. alilisd Posted: March 09, 2015 at 03:00 PM (#4909477)
I just don't get how that guy is even in the discussion with Wagner/Papelbon


I think it's pretty clearly narrative. Of course he's not being discussed/compared with Papelbon as he is still playing, but Wagner comes up short in the Saves, which is what it's all about, and he moved from Houston to Philly to NY to Boston to Atlanta. Hoffman has a strong association with San Diego, I'll bet many people forget he ever played for Miami/Florida, and tend to gloss over his last two seasons in Milwaukee.
   82. SoSHially Unacceptable Posted: March 09, 2015 at 03:11 PM (#4909483)
Also, we absolutely do not know that about Walker. In fact, excellent offensive catchers are sometimes moved off the position to get their bats into the game more often. I don't see anything about the position which would preclude an otherwise skilled athlete capable of playing other positions in the field from playing catcher. Isn't it more often the slower, less athletic players who are put at the position due to their incapacity for other positions rather than the other way around?


Bats get moved off catcher because the position is so physically demanding and career-shortening that you frequently don't want to risk losing an elite bat. But if your idea was true, then you'd see far more attempts to convert first or third sackers with bats that may not carry the position into backstops, and that almost never happens.

Guys can move along the defensive spectrum in positions 3-9, but almost no one transitions to catcher. It's just a different defensive beast.

   83. alilisd Posted: March 09, 2015 at 05:52 PM (#4909596)
Guys can move along the defensive spectrum in positions 3-9, but almost no one transitions to catcher. It's just a different defensive beast.


True, at the professional level you're not going to pick up catching, but if an athlete has the skills to succeed at MLB, such as a Larry Walker, he could surely have learned the position as an amature.
   84. bbmck Posted: March 10, 2015 at 09:40 AM (#4909808)
Player          ERAOPS+  WHIP   SO9 SO/W     IP From   To   Age  FIP
Greg Maddux      191   48 0.947  7.04 6.33  938.1 1995 1998 29
-32 2.57
Billy Wagner     187   49 0.998 11.92 3.99  903.0 1995 2010 23
-38 2.73
Randy Johnson    173   59 1.047 11.82 5.08 1118.0 2000 2004 36
-40 2.53 


Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 3/10/2015.

Player                ERAOPS+  WHIP   SO9 SO/W From   To   Age    IP  FIP
Trevor Hoffman         118   83 1.204  8.98 2.73 1993 1995 25
-27 199.1 3.74
Bud Black              118   83 1.192  3.56 1.69 1989 1989 32
-32 222.1 3.51
Trevor Hoffman         176   49 0.945 10.98 4.57 1996 2000 28
-32 382.0 2.51
Dwight Gooden          176   55 1.013  9.90 3.83 1984 1985 19
-20 494.2 1.93
Trevor Hoffman         131   73 1.087  8.29 3.55 2001 2010 33
-42 508.0 3.25
Fernando Valenzuela    132   70 1.099  7.18 2.65 1980 1982 19
-21 495.0 2.63 


Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 3/10/2015.
   85. SoSHially Unacceptable Posted: March 10, 2015 at 10:56 AM (#4909864)
True, at the professional level you're not going to pick up catching, but if an athlete has the skills to succeed at MLB, such as a Larry Walker, he could surely have learned the position as an amature.


I doubt it's that simple. It's possible Larry Walker could have made an OK catcher, but it's by no means guaranteed (and whether he was good enough to handle the spot at the big league level is another matter entirely). We see guys who grow up catching, the D'Backs O'Brien being the most recent example, who may not be able to handle the position defensively in MLB. I don't think simply having a decent arm and being a good athlete is necessarily enough.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Dynasty League Baseball

Support BBTF

donate

Thanks to
James Kannengieser
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 0.8049 seconds
59 querie(s) executed