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, May 04, 2005

Relief Pitchers

Should we maybe start a thread for general discussion of relief pitchers at some point?

- that “definitely immoral” cat :-)

John (You Can Call Me Grandma) Murphy Posted: May 04, 2005 at 05:31 PM | 460 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

Page 3 of 5 pages  < 1 2 3 4 5 > 
   201. JC in DC Posted: June 02, 2006 at 05:20 PM (#2048384)
One can consider it to be relevant or not. I think it's inescapably relevant, as one of the ways in which we try to understand what relief pitching is, as a role in comparison with other roles: how the athletes who perform it have been selected, and what that selection process tells us about the relative scarcity of relief pitching talent, and what that selection process and talent scarcity tells us about how teams typically value the contributions of starters vs. relievers. It's an interesting piece of the puzzle, I think.


This is well put, but I agree w/Daryn (and I suspect bunyon). The problem with this analysis, to my mind, is a static and undefended conception of skill. There is no reason to assume that the skills involved in being a good or great reliever are the same as those involved in being a good or great starter, or even a starter at all. To my mind a helpful analogy might be to look at offensive v. defensive linemen. OTOH, they are superficially similar: the tasks require strength, size, and agility. OTOH, the tasks are very different, and a great offensive lineman is NOT interchangeable w/a great defensive lineman; nor even are great offensive linemen interchangeable w/each other (say from guard to tackle or guard to center). The point is, that while it's clear very often a starting pitcher w/limited success will become a relief pitcher, it's much less often the case that a starting pitcher will become a GREAT, or even very good (of the Bob Wickman type) relief pitcher. There's no reason to conclude Rivera wouldn't have been at least a #3 or 4 starter, given the opportunity. But he's a lights-out relief pitcher, and as Farnsworth showed last night, that's nothing to undervalue.

Additionally, I think Daryn has done a good case arguing for the extraordinarily small margin for error granted to relief pitchers. If a relief pitcher had started as slowly as Robinson Cano or Andy Phillips, or Cabrera, they would have been shipped out much more quickly than are those everyday players whose failures ARE more easily hidden, or more forgivable.

If you want to argue for normative managerial patterns, then, then you have to notice, as bunyon has, the changing attitude towards relievers which we can all assume (defensibly, I think) is the consequence of experience: relief pitchers are extreme commodities possessing rare skills (an "out" pitch or two, quick recovery and durability, composure), are very fragile, burn out very quickly, and are thus NOT interchangeable.
   202. Steve Treder Posted: June 02, 2006 at 05:34 PM (#2048399)
Clemens wasn't a smashing success right out of the gate either.

As a rookie, Clemens was 9-4 with a 96 ERA+ in 20 starts at the age of 21. Rivera as a rookie was 3-3 in his 10 starts with a 77 ERA+ at the age of 25.

Nor was Maddux or Johnson.

Maddux did indeed struggle mightily, but he was 20 and 21 at the time. By the age of 22 he was an 18-game winner.

Johnson's struggles are more akin to those of Rivera: he wasn't that young, and it took him several years to find the plate. Indeed it wouldn't have been at all surprising for someone to have converted Johnson into a closer; he certainly had the requisite attributes. But Johnson's late-blooming dramatic success is quite unusual among great starters.

Even superlative pitchers require a season or 2 to acclimate themselves to the faster level of competition and to begin to understand the weaknesses in the hitters they wish to exploit.

Of course, every young player develops over a period of years. But it's simply a fact that the great majority of the best starters, the types of starters the HOM is honoring, were established as successful major league starters in their early 20s.
   203. Daryn Posted: June 02, 2006 at 05:36 PM (#2048404)
I was looking at that chart and saw a string of 6s and 7s under the column R and I thought kevin was being sarcastic. Then I woke up.
   204. Steve Treder Posted: June 02, 2006 at 05:40 PM (#2048408)
I don't see enough there that would discourage me from giving him another shot at starting.

Well, maybe not me either, depending of course on what my other options were. But the fact is that it did in fact discourage the Yankees from giving him another shot at starting. Whether they made a wise choice or not, that's the choice they made, and a similar sort of story can be found in nearly every other relief star's background. I see no reason to reject or ignore this recurrent historical pattern in the consideration of how best to establish a value assessment for relievers for HOM purposes.
   205. Daryn Posted: June 02, 2006 at 05:51 PM (#2048418)
Well, maybe not me either, depending of course on what my other options were. But the fact is that it did in fact discourage the Yankees from giving him another shot at starting. Whether they made a wise choice or not, that's the choice they made, and a similar sort of story can be found in nearly every other relief star's background. I see no reason to reject or ignore this recurrent historical pattern in the consideration of how best to establish a value assessment for relievers for HOM purposes.

I think you are drawing wrong conclusions from agreed facts. At my workplace, we have lots of young lawyers -- we evaluate their skills and often steer them to be solicitors (research, document drafting, analysis based, non-litigation lawyers) rather than barristers (litigators). The litigators are the starting pitchers of the law world and make more money and get more respect. But someone who showed the promise to be a superlative solicitor (or someone who likes that role better) sometimes doesn't get the opportunity to show that they could have been a great barrister.

And as BL has said several times (I'm not sure if you are reading him Steve :), so maybe this is why you haven't responded to the following valid point) none of that matters when you are evaluating value or merit -- why he is in the role shouldn't influence the value of the role.

JC's linemen examples are also terrific iterations on the points we have been discussing. Similar roles are not all that similar when you are at the major league level.

Closers, like secondbasemen, are immensely skilled athletes doing their part for the team. The closers we elect will have the same value flaws as the bottom half of the secondbasemen we elect.
   206. Steve Treder Posted: June 02, 2006 at 06:04 PM (#2048426)
why he is in the role shouldn't influence the value of the role

Why he is in the role doesn't influence the value of the role. But our entire challenge is to determine the value of the role, and that determination is enhanced by an understanding of how teams have actually, historically valued the role, and their actual, historical choices in which athletes they select for the role, as well as how readily available they sense replacement talent to be.

Of course it's all about value. The issue is that one shouldn't assume that the value of the relief pitcher role is equal to the value of other roles without testing the assumption. Every value methodology we're familiar with (WARP, Win Shares, etc.) doesn't conclude that relief pitching is equal in value to other roles. I asked you this before, and you haven't responded: do you disagree with that assessment?

Closers, like secondbasemen, are immensely skilled athletes doing their part for the team.

Yes, but so what? So are backup catchers and utility infielders. The relevant question is what value "their part" brings to the team.

The closers we elect will have the same value flaws as the bottom half of the secondbasemen we elect.

Sure, but that isn't the issue. The issue is should the HOM elect the same number of relief pitchers as it does second basemen.
   207. Daryn Posted: June 02, 2006 at 06:11 PM (#2048434)
The issue is should the HOM elect the same number of relief pitchers as it does second basemen.

That's never been the issue. We are trust trying to evaluate value. There are no quotas here. When people make reference to the number of closers they think ought to be in the HoM it is just shorthand for guessing the number that will have the value to make the top 225. And noone has suggested that more than 9 closers ought to make the HoM even using that shorthand, while more than 20 2b will make it.
   208. Steve Treder Posted: June 02, 2006 at 06:22 PM (#2048440)
I thinka more apt analogy would be "Do we evaluate leadoff hitters the same way we evaluate cleanup hitters". Leadoff hitters don't have the power the cleanup guys do but they still are, nevertheless, critical components to the offense and have to be evaluated on what they do, rather than what they don't do. If you use RBI's as a yardstick, then they all suck, from Ricky Henderson on down.

No, that isn't an apt analogy, because every position player has the opportunity to be a leadoff hitter or a cleanup hitter. So position players aren't evaluated on the basis of whether they're leadoff hitters or cleanup hitters; instead they're evaluated on the basis of their total offensive contribution (whatever form it took) within the context of their position-specific defensive contribution.

And in fact, I'm sure you'll find that the HOM has a whole lot more cleanup hitters represented in it than it does leadoff hitters, especially as first-ballot types.
   209. Steve Treder Posted: June 02, 2006 at 06:31 PM (#2048447)
That's never been the issue.

That's the sum total of the issue. That's the issue in a nutshell. The entire question is whether it should be nobody, or just Wilhelm, or just Wilhelm and a few others, or Wilhelm and 20 others. That's precisely the point of this discussion.

We are trust trying to evaluate value.

Of course, because in coming up with an agreed-upon general understanding of the value of the best relief pitchers, a determination can be made in comparing that value to the value delivered by the best players at other positions.

There are no quotas here. When people make reference to the number of closers they think ought to be in the HoM it is just shorthand for guessing the number that will have the value to make the top 225.

Let me see ... if it walks like a quota, and talks like a quota ...

And noone has suggested that more than 9 closers ought to make the HoM even using that shorthand, while more than 20 2b will make it.

Substitute the term "relief pitcher" for the term "closer," and you have just demonstrated that you acknowledge a greater value contribution from the second base position than from the relief pitcher position.
   210. Daryn Posted: June 02, 2006 at 06:37 PM (#2048453)
Substitute the term "relief pitcher" for the term "closer," and you have just demonstrated that you acknowledge a greater value contribution from the second base position than from the relief pitcher position.

Of course I do, at least in the sense of is the best secondbaseman more valuable than the best closer. We are not trying to win (or lose) arguments here. We don't play "gotcha". We share our views and information and learn from each other. As sappy as that may sound to you, we have enjoyed it for 80 years now.
   211. John (You Can Call Me Grandma) Murphy Posted: June 02, 2006 at 06:38 PM (#2048454)
There are no quotas here. When people make reference to the number of closers they think ought to be in the HoM it is just shorthand for guessing the number that will have the value to make the top 225.

Let me see ... if it walks like a quota, and talks like a quota ...


It's still not the same thing, Steve. We did the same thing with the Negro Leaguers trying to see what was a good number for them as a benchmark, but in the end, the only thing that mattered was if and where they belonged on our ballots. No NeLers were getting a boost so as to meet that benchmark either. The same thing will also apply to firemen/closers.
   212. Steve Treder Posted: June 02, 2006 at 06:40 PM (#2048457)
Of course I do

Then we completely agree.
   213. Steve Treder Posted: June 02, 2006 at 06:46 PM (#2048465)
It's still not the same thing, Steve. We did the same thing with the Negro Leaguers trying to see what was a good number for them as a benchmark, but in the end, the only thing that mattered was if and where they belonged on our ballots. No NeLers were getting a boost so as to meet that benchmark either. The same thing will also apply to firemen/closers.

No, the Negro League analogy doesn't apply. The question there is simply assessing and electing the best players on each ballot, regardless of what league(s) they played in.

The issue with relievers is different. Here the question of properly assessing the value generated by relief pitching as a role, in comparison with the value generated by other positional roles, is inescapable. This must be done in conjunction with considering how good each reliever is in rank order amongst themselves, or else there is no way of knowing whether the 10th-best reliever is worth considering as HOM-worthy or not.
   214. Dr. Chaleeko Posted: June 02, 2006 at 06:52 PM (#2048472)
I think as I'm trying to visualize HOMable value, one analogy is a clear-glass pot of water. The meniscus of the water is the value level of the worst HOMer. As you add heat (read greatness), the water will bubble. But the bubbles are not uniform. Some bubbles (Babe Ruth) are big, lid-jarring gobs of plasma-filled energy that gather strength at the bottom then come roaring to the top. Some bubbles (Bill Terry) barely make a ripple on the surface as the gently make their way up the side of the pot.

The innate value of the positions and of starting pitcher will consistently produce bubbles that break above the surface of the water. But relief pitchers, due to the intrinsic nature of their roles are limited, no matter how great. They don't accrue enough value to make big bubbles, and even the greatest among them appears to only make a Bill Terry sized bubble.

So it comes down to this:
-either the omnibus measures, which do, apparently account for leverage, consistently devalue relief pitching and chronically underrate the role

or

-the role isn't that valuable to begin with.

That's the whole enchilada. We either measure wrong, or we're measuring something that almost never measures up because it's intrinsically impossible for it to be able to.
   215. Steve Treder Posted: June 02, 2006 at 06:57 PM (#2048475)
So it comes down to this:
-either the omnibus measures, which do, apparently account for leverage, consistently devalue relief pitching and chronically underrate the role

or

-the role isn't that valuable to begin with.

That's the whole enchilada. We either measure wrong, or we're measuring something that almost never measures up because it's intrinsically impossible for it to be able to.


Exquisitely put.

One of the most common criticisms of Win Shares, for example, is that it doesn't properly evaluate relievers; some complain that it values them too highly, others that it doesn't value them highly enough. The two things everyone should be able to agree upon are:

1) That comparative valuation question between relievers and everyone else is critically important

and

2) It's awfully damn difficult to do
   216. Daryn Posted: June 02, 2006 at 07:20 PM (#2048490)
1) That comparative valuation question between relievers and everyone else is critically important

and

2) It's awfully damn difficult to do


When these two factors exist, common sense also plays a role.
   217. Howie Menckel Posted: June 02, 2006 at 07:25 PM (#2048493)
My problem with the analogies of the offensive linemen or even second basemen is that we know that a great QB couldn't play OT, and vice versa.

My thinking is whether the great SP could be a great closer, and if the great closer could be a great SP.
There isn't really enough data to give an ironclad answer, but at this point it's my feeling that the great SPs could indeed be great closers - but not the other way around.
Maybe someone will talk me out of that part. If not, I may tend to give the benefit of the doubt to the borderline SP over the borderline closer.

Take wins vs saves, as another example.
A closer can and often does pitch only one inning with a two or three-run lead. Sometimes he gives up a run or two, but he gets the 'money stat' anyway - a save.

Sure, a SP can have a mediocre to poor game and still get a win. But he always has to throw 5 full innings, at least 75 pitches or so, and many times he pitches quite well and doesn't get the win.
But a closer whose team is losing gets the day off, so his team waits until the optimal moment arrives, then puts him in the game to perform a task that isn't all that difficult at least half the time (get 3 outs before you give up 2-3 runs).

I acknowledge that even into the 1990s this isn't exactly the way closers were used, and the more challenging the task that was asked of a player, the more credit I will give him (Gossage being a good example).
   218. sunnyday2 Posted: June 02, 2006 at 07:38 PM (#2048504)
>Charlie Bennett couldn't play a full schedule because the physical demands were too great.

Well actually we elected Bennett in part because he did play damn near a full schedule.

His experience shows that the platoon catcher was basically "the way we do things nowadays."

As long as humans pull the triggers, "the way we do things" is part of it. Our job is to determine how that fits in with everything else we know.
   219. Dr. Chaleeko Posted: June 02, 2006 at 07:39 PM (#2048506)
Commonsense then is telling me that I've overvalued HOM slots for relievers in the past because I've used a purely mathematical way to look at their representation. I used positional balance as the guiding theory. That told me 5-7 of them would be proporitional to the representation at other positions.

Now I'm thinking 0-3 when I've considered what our value systems say and why...as well as what good old baseball argumentation and logic say.

But I'm still very open to being convinced otherwise. What would convince me? An explication of WS and WARP that shows their leverage systems err substantially and undervalue relief work by something like 20%. I don't know the exact number, but that's the one I'll pull out for now.

If after such a process we find instead that WS and WARP undervalue relieving by 3% or 7%, that's not really going to change my mind much because it still won't push anyone but Mo and Goose over the top.

On other other hand, the win-expectancy approach has some real potential, and I think my next task is going to be to figure out what WXRLs look like and whether they reveal anything useful about RP vs SP/Pos Players.
   220. sunnyday2 Posted: June 02, 2006 at 07:45 PM (#2048510)
I refuse to believe that relievers as a category categorically cannot produce a HoMer.

But I never said 5-7 sounded right. I think 1-3 (not 0-3) is gonna be closer.
   221. bunyon Posted: June 02, 2006 at 07:52 PM (#2048516)
Sure, a SP can have a mediocre to poor game and still get a win. But he always has to throw 5 full innings, at least 75 pitches or so, and many times he pitches quite well and doesn't get the win.

I don't disagree with your overall assessment, but SP can and do record wins with poor outings. Someone the other day fell behind 7-0 in the first and got a win. A reliever who has a horrific outing has no shot at a save.

Wins and saves, though are equivalent in being too team dependent. And, as I said, I think you're probably right in that a good SP could be a good to great closer while the converse is less likely.
   222. JC in DC Posted: June 02, 2006 at 07:59 PM (#2048525)
My problem with the analogies of the offensive linemen or even second basemen is that we know that a great QB couldn't play OT, and vice versa.


Whatever, but that's not how the analogy worked. It was from offensive line to defensive line, which is more like from SP to RP than from CF to SP or RP.

And, like Steve, you're begging the question in asserting that b/c often RPs are "failed SPs" therefore RPs are less valuable and/or more interchangeable.

In fact, AFAICT, Steve has committed himself to a conclusion he doesn't allow the facts to challenge. This is why I wrote that he's committed to a static conception of skill. Every team obviously values RPs. RPs increasingly come from a variety of sources (they're targeted in HS and pitch relief through college, they're converted OFs or SSs, or SPs). The fact that they once did something else and can then move to relief is no knock on their relative value, it only speaks to the dire need for quality relievers; I just don't see how that argument works, especially given their extraordinary scarcity, despite Steve's assertion to the contrary.
   223. Dr. Chaleeko Posted: June 02, 2006 at 08:01 PM (#2048528)
OK, so I'm looking at WXRLs on BP. Using 1968 as an example, the WXRL tables say

Hoyt Wilhelm 2.348

and that's supposed to tell me that Hoyt Wilhelm added 2.348 wins to his team above what a replacement reliever would have given the quality of batting opposition Hoyt faced.

OK. So how do we integrate this? BP's player cards offer no help, they don't include any of the usual reliever-report stats nor this one. As Chris mentions they include XIP, but that appears to be a blunt instrument.

Wilhelm's pitching in 1968 by WARP1 is determined to be worth 4.7 wins above replacement. Does this mean that WARP1 is overstating Hoyt's performance (4.7 WARP vs. 2.3 WXRL)? Or does it mean something else? Or are these two unintegratable in any meaningful way?
   224. Steve Treder Posted: June 02, 2006 at 08:28 PM (#2048547)
And, like Steve, you're begging the question in asserting that b/c often RPs are "failed SPs" therefore RPs are less valuable and/or more interchangeable.

It doesn't automatically mean that they're less valuable than SPs, although it very obviously raises it as a distinct possibility. But it does clearly mean that they're more interchangeable, given that every team has a ready supply of candidates for its bullpen (its starters) that the starting rotation doesn't conversely have.

Every team obviously values RPs. RPs increasingly come from a variety of sources (they're targeted in HS and pitch relief through college, they're converted OFs or SSs, or SPs). The fact that they once did something else and can then move to relief is no knock on their relative value

It isn't directly a knock, of course. But the fact that SPs are converted to relief at many, many times the rate that RPs are converted to starter forces one to carefully consider the value question in relation to SPs, at the very least. And the feeder pool of replacement talent (again, SPs) is relevant to the value question too; value in any realm is partially a function of resource supply.

Why is it, JC, that metrics such as Win Shares and WARP rate relief pitchers as significantly less valuable than starters or position players? Are they wrong in that assessment?

it only speaks to the dire need for quality relievers; I just don't see how that argument works, especially given their extraordinary scarcity, despite Steve's assertion to the contrary.

It works because there is no extraordinary scarcity. Starting pitchers are a far more scarce, difficult to develop and replace commodity than relievers.
   225. Daryn Posted: June 02, 2006 at 08:43 PM (#2048555)
Commonsense then is telling me that I've overvalued HOM slots for relievers in the past because I've used a purely mathematical way to look at their representation. I used positional balance as the guiding theory. That told me 5-7 of them would be proporitional to the representation at other positions.

Now wait a minute Doc -- I only want you to use common sense if it will cause you to agree with my pre-conceived notion that 5-7 relievers is the right number. Go back to your mathematical proportionality thingy if doing otherwise will lower your support for relievers. :)
   226. Daryn Posted: June 02, 2006 at 08:54 PM (#2048570)
It works because there is no extraordinary scarcity. Starting pitchers are a far more scarce, difficult to develop and replace commodity than relievers.

I think you must mean #1 starting pitchers as compared to closers. Because if you mean #5s compared to closers or run of the mill relievers, that's just wrong. I disagree with you on both points, but the latter isn't even tenable. Of the 30 closers, 3 to 8 are replaced each year. Of the 30 #5s, 15 to 30 are replaced each year. Of the 90 run of the mill relievers, 30 to 60 are replaced each year. It is the 5th starter that is most fungible in MLB, and in any event no less fungible than a run of the mill reliever.

This again raises what I view as an error in your fungibility argument. You say starters are scarcer and closers are easy to replace because there is always/often a virtually as good set up man waiting to replace him. But the better comparison is closer/set up guy versus #1 starter/#2 starter. In this comparison you will see a real analog -- many teams have only a small drop off and some teams have a big drop off and that applies to both categories.
   227. DL from MN Posted: June 02, 2006 at 09:07 PM (#2048580)
Just looking at how many 50-60 WARP relievers there are gives me a sense that it isn't a terribly difficult thing to find a 50-60 WARP reliever. I believe the scarcity (or lack of scarcity) argument. I also am starting to think the leverage is partly cancelled out by the higher replacement value. If so, it comes down to how well you pitched and in how many innings.

I think there are only a very few pure relievers worth considering: Wilhelm, Rivera, Gossage, Lee Smith, Trevor Hoffman, John Franco and Rollie Fingers. These guys are all above 75 WARP. Eckersley wouldn't make it without his days as a starting pitcher. Lee Smith is borderline for my PHoM so I guess I'm saying 3-7 is the right number and I think WARP is close enough. I don't believe there is such a thing as an "inner-circle" relief pitcher.
   228. Steve Treder Posted: June 02, 2006 at 09:08 PM (#2048581)
I think you must mean #1 starting pitchers as compared to closers.

I do, and I also mean #2 starters compared to #1 setup men, and #3 starters compared to #2 setup men. The top of the rotation is far harder to fill than the top of the bullpen.

I agree that 5th starters are nearly as fungible as back-end-of the bullpenners. That's one of many reasons why a strict 5-man rotation is a suboptimal approach.

This again raises what I view as an error in your fungibility argument.

Well, let's remember that it really isn't "my" fungibility "argument." We got into this tangent because of the acknowledgement of the empirical fact that the best closers tend to remain in the role for far shorter periods than the best starters, and relievers generally have shorter careers overall. These are facts; the question becomes how to explain them.
   229. Daryn Posted: June 02, 2006 at 09:20 PM (#2048586)
I don't believe there is such a thing as an "inner-circle" relief pitcher.

Yesterday I disagreed with this, but I think you are right if inner circle means something like top 50 players of all time. Someone said that Bill James rated the best reliever ever at 27 all time among pitchers. That sounds about right to me, and makes the top handful or perhaps more clearly HoM quality in my view.
   230. Howie Menckel Posted: June 02, 2006 at 09:23 PM (#2048587)
JC in DC:
"Of the 30 closers, 3 to 8 are replaced each year."

Huh? The number is WAY higher than that within every given season.
Do you mean from September til April? That stills seems low, but I'm not sure whether that would be relevant.


To modify the analogy as you request, I would say that if I believed most offensive linemen could be equally successful as defensive linemen, but that the converse was not true, yes, I would tend to give a borderline OL the benefit of the doubt.


In what way is there an "extraordinary scarcity" of good RPs compared to good SPs?
Not only is it not true, imo, that's something that could easily be corrected by a given team - good SPs, I say, will mostly be good RPs (I won't even bring up the Ryan Dempsters and such who suddenly seem to gain new life in the pen).

But if you have a bunch of good RPs yet are short of SPs, well, you've got a problem.
   231. Mark Donelson Posted: June 02, 2006 at 09:25 PM (#2048591)
If the fact that your system misses Wilhelm and looks like it will miss relief pitchers in general is troubling to you, then broaden your system!

Well, that's precisely what I'm trying to do! But I need to find a real reason to do so, not say "Mariano Rivera feels like he should be an HOMer, so I'm going to multiply his WS by X until he looks like one."

As for Wilhelm, my system doesn't entirely "miss" him: I agree with you that he drops in around the level of Bunning and Drysdale. Bunning was 11th on my ballot last year; Drysdale was 15th the year he went in. I was more interested in those who didn't adore Bunning/Drysdale who have Wilhelm slotted for first (like Marc, who has explained why since).

I still think I'll have Wilhelm on-ballot, but low. And I still have no sense of what I'll do with Gossage/Rivera/anyone else, but I find myself agreeing an awful lot with Dr. Chaleeko's recent posts.
   232. Steve Treder Posted: June 02, 2006 at 09:27 PM (#2048593)
JC in DC:
"Of the 30 closers, 3 to 8 are replaced each year."


It was Daryn who said that.
   233. Howie Menckel Posted: June 02, 2006 at 09:44 PM (#2048601)
Gotcha.
JC on "extraordinary scarcity," but daryn on "3 to 8 closers changed."

thanks.
   234. Backlasher Posted: June 02, 2006 at 10:02 PM (#2048611)
But it does clearly mean that they're more interchangeable, given that every team has a ready supply of candidates for its bullpen (its starters) that the starting rotation doesn't conversely have.


No it doesn't and even if it did its moot to this discussion. I don't care if in your opinion SP Billy Grabinass from 1927 could have gotten 95 saves a year if he pitched in the modern environment. He didn't get those saves. He didn't contribute that value. He also didn't "fail" in the closer's role because he may have finished a couple of games in relief but the team decided to use him as a starting pitcher.

All that matters is the actual value contributed. Not what you have to jump through ten logically constrained hoops to figure what someone else could have done. You aren't electing persons that "had the best true talent level"; you are electing the persons whose performance gives them a distinguishment of merit.

Substitute the term "relief pitcher" for the term "closer," and you have just demonstrated that you acknowledge a greater value contribution from the second base position than from the relief pitcher position.


That is assinine, substitute any term you want and all that you will do is expect that of the people that have performed that role, here are the numbers that distinguish themselves. It has nothing to do with a quota; it has nothing to do with anything else.

If you want to say that over the course of history, the class of starting pitchers have provided more value than the class of relief pitchers, that could be correct. Here is not the place to argue it. You are dealing with individuals, not a caste system that you want to impose.

Yes, but so what? So are backup catchers and utility infielders. The relevant question is what value "their part" brings to the team.


No, the relevant question is how much value did this individual bring to the team. If you have a backup catcher that adds enough value to be inducted, then he should get inducted. It doesn't matter that you put a label on him of "backup catcher".

But its highly unlikely a backup catcher will be able to provide that much value. Throughout most of history, it would have been highly unlikely a relief pitcher could have provided that much value. That is primarily because the utilization models didn't provide that opportunity.

But Hoyt Wilhelm did provide that value, so he deserves consideration. No that the utilization model is more optimized, there are going to be more relief pitchers capable of providing enough value to be considered. That class of pitchers includes closers, and includes at least one slam dunk member Mariano Rivera. It includes others that are worthy of consideration.

But branding someone a "failure" or a "loser" or a "backup such and such" means nothing. Its a means of caste segregation. An I am not going to consider such-and-such because he was in the "Negro Leagues" or was a "relief pitcher". You find the value the people contributed, and then you decide if its worthy of honor. It has nothing to do with woulda/coulda/shoulda for anybody else. Replacement is moot, in fact it inures to their favor because it shows the difficulty in maintaining longevity, the exact same way it does for people playing catcher or people playing second base.
   235. Backlasher Posted: June 02, 2006 at 10:12 PM (#2048617)
My thinking is whether the great SP could be a great closer, and if the great closer could be a great SP.
There isn't really enough data to give an ironclad answer, but at this point it's my feeling that the great SPs could indeed be great closers - but not the other way around.
Maybe someone will talk me out of that part. If not, I may tend to give the benefit of the doubt to the borderline SP over the borderline closer.

Take wins vs saves, as another example.
A closer can and often does pitch only one inning with a two or three-run lead. Sometimes he gives up a run or two, but he gets the 'money stat' anyway - a save.

Sure, a SP can have a mediocre to poor game and still get a win. But he always has to throw 5 full innings, at least 75 pitches or so, and many times he pitches quite well and doesn't get the win.
But a closer whose team is losing gets the day off, so his team waits until the optimal moment arrives, then puts him in the game to perform a task that isn't all that difficult at least half the time (get 3 outs before you give up 2-3 runs).

I acknowledge that even into the 1990s this isn't exactly the way closers were used, and the more challenging the task that was asked of a player, the more credit I will give him (Gossage being a good example).


There is nothing wrong with creating such tiebreakers, and it doesn't matter too much whether any person could have done a different role, they did what they did. I wouldn't try talking you out of any of those things; I also would not think any of those things would go into your evaulation of performance.

But I will offer you this on saves, because I see that repeated very often. A closer has one duty that he either succeeds or fails in. He must finish the game with his team winning. If he does, he is successful, if he does not his team fails.

A closer needs to be successful about 88% of the time to keep his job.

Those are his performance metrics. If he gives up 2 runs but gets the save he has still done his job. At the end of the day, you shouldn't really care much more about that then if Matheson decides to cost in a 7-0 game and give up 3 runs but still get the W.

However, what you do care about is not only the success of the person, but also the value that person provides. That is going to be measured by his LI to some degree. And LI will include any deductions, bonuses, etc. that you want to make for value. IOW, if somebody is always saving 1 run games, he is providing more value than someone saving 3 run games, because of the context of the situation. I cannot imagine that you would not want to look at both things rather than anyone alone.

And if you don't like LI, then I think that is burden in front of you merit boys. You need to figure out how to measure that value. Hopefully, you want need a protracted conversation to recognize that value differs.

But that is the test. And its also one of the reasons why you may find that some of the people that you think were really, really great actually don't perform as well as people that you may think are pedestrian. For instance if Wagner stays healthy and even declines a bit, he is likely to provide far more value than THE GOOSE.
   236. Paul Wendt Posted: June 02, 2006 at 10:18 PM (#2048622)
S.T.
Every value methodology we're familiar with (WARP, Win Shares, etc.) doesn't conclude that relief pitching is equal in value to other roles.

methodology? conclude?
Do you mean that the authors of rating systems demonstrate, show differences rather than postulate, presume them?


That's never been the issue. We are trust trying to evaluate value. There are no quotas here. When people make reference to the number of closers they think ought to be in the HoM it is just shorthand for guessing the number that will have the value to make the top 225.

Some express that ideal and a few may strive for it.
   237. Backlasher Posted: June 02, 2006 at 10:19 PM (#2048623)
No, that isn't an apt analogy, because every position player has the opportunity to be a leadoff hitter or a cleanup hitter. So position players aren't evaluated on the basis of whether they're leadoff hitters or cleanup hitters; instead they're evaluated on the basis of their total offensive contribution (whatever form it took) within the context of their position-specific defensive contribution.


And what you propose would be exactly what Kevin describes. Any person on the roster can play at any position. Rickey Henderson is not used as a cleanup hitter; Mariano Rivera is not used as a starting pitcher. Either could do those things. They may or may not exhibit better performance than the people doing those things. But they did what they did, were used how they were used, and provided the value they provided.

You can attach a label to Rickey that is just as valid and true as the label you attach to Rivera. The issue here is that you are using the label you attach to Rivera as derisive, and trying to drag up history that is not even relevant to promote why you want to segregate.

And yes, there are more cleanup hitters than leadoff hitters, and there will be more Starting pitchers than relief pitchers. That doesn't mean that a leadoff hitter should be downgraded because he is a leadoff hitter or a relief pitcher should be downgraded because they are a relief pitcher. It doesn't matter how many times leadoff hitters are replaced. It doesn't matter if Reggie Jackson woulda/coulda/shoulda been a great leadoff hitter.

All that matters is what the people did in their role. And if you find that because of the batting order that player provided extra value (which is unlikely to be true) then he should be given credit for the value that he actually provided.
   238. Backlasher Posted: June 02, 2006 at 10:25 PM (#2048628)
We got into this tangent because of the acknowledgement of the empirical fact that the best closers tend to remain in the role for far shorter periods than the best starters, and relievers generally have shorter careers overall. These are facts; the question becomes how to explain them.

No, you got into this tangent because you kept bringing things up. And rather than listening, you are playing dog without a bone. You have maybe three posts where you introduce something new to the conversation. Nothing new is relevant to assessing value. Everything else is just rhetoric hoping that you catch someone in focusing on one set of data and overvaluing it, rather than looking at all the data together and properly valuing it.

And if it "begs the question"; it doesn't beg the question in a Hall of Merit thread. Because all they need to do is assess value. Not decide what Billy Grabinass should have done; not follow a weak logical chain that ends in 3 Man rotations, ACE RELIEVERS, and throwing 19 year old kids for 300 innings.

These are people whose careers are over. They need to assess their value.
   239. Steve Treder Posted: June 02, 2006 at 10:29 PM (#2048629)
methodology? conclude?
Do you mean that the authors of rating systems demonstrate, show differences rather than postulate, presume them?


I project no motives onto the authors of rating systems other than what they say they're attempting to do: measure the actual value of player performance.

Assuming that that's actually what they're trying to do, then how are we to interpret the fact that multiple, different systems developed by different authors all present a consistent finding: that the best relief pitcher seasons aren't as valuable as the best seasons by starting pitchers or position players, and as well that the typical relief pitcher season is less valuable than the typical season by a starting pitcher or position player.

That is the situation. As Dr. Chaleeko points out, this can mean only two things:

a) All these systems are wrong in this regard, and thus the question becomes why are they all wrong, and how should they be corrected?

or

b) Relief pitchers actually aren't as valuable as starting pitchers or position players.

Which do you think it is?
   240. Backlasher Posted: June 02, 2006 at 10:32 PM (#2048632)
Why is it, JC, that metrics such as Win Shares and WARP rate relief pitchers as significantly less valuable than starters or position players? Are they wrong in that assessment?


That's simple. They don't do that. The rating systems rate each individual based on the contribution they provide. They don't discount a pitcher for pitching in relief, any more than they discount a player for being Dominican.

Certain players because of the role they play may have different opportunities to acquire different levels of value.

And more important, nothing is dealt with at a class level, its dealt with at an indvidual level.

That statement is like saying how come RBIs always show that cleanup hitters are better than leadoff hitters.
   241. dlf Posted: June 02, 2006 at 10:47 PM (#2048641)
All that matters is the actual value contributed. Not what you have to jump through ten logically constrained hoops to figure what someone else could have done. You aren't electing persons that "had the best true talent level"; you are electing the persons whose performance gives them a distinguishment of merit.


I'm curious whether the HoM voters are really trying to measure pure value. I haven't voted here and only read some of the threads. But it seems to me that there is more than a little "what this guy would have done in other circumstances" discussion about ability. Examples are giving minor league or National Association credit for some players. War credit for others. Some of the statistics being thrown about in both position player and pitching discussion attempt to take the player out of the context specific to him and place him in a neutral environment. That is not a value-neutral choice. I think that a pure value question is interesting, particularly in a discussion of relievers vis a vis starters. But, at least to this interested observer, I don't think a retrospective review of value is all that the voters here are striving for.
   242. Backlasher Posted: June 02, 2006 at 10:48 PM (#2048643)
Assuming that that's actually what they're trying to do, then how are we to interpret the fact that multiple, different systems developed by different authors all present a consistent finding: that the best relief pitcher seasons aren't as valuable as the best seasons by starting pitchers or position players, and as well that the typical relief pitcher season is less valuable than the typical season by a starting pitcher or position player.


You don't need to interpret at all. You aren't judging classes when inducting people into the HoM. You are inducting persons for the value that they provided.

That is the situation. As Dr. Chaleeko points out, this can mean only two things:

a) All these systems are wrong in this regard, and thus the question becomes why are they all wrong, and how should they be corrected?

or

b) Relief pitchers actually aren't as valuable as starting pitchers or position players.

Which do you think it is?


Please stop that. It does not mean only one of two things. In fact, it could very well mean both of those things at the same time.

As Daryn suggested, discuss, don't try to play a rhetorical game.

This is about individuals performance. How do you explain the fact that these systems constantly show that "clean up hitters perform better than leadoff hitters"

Its a meaningless statement.

But if you want an answer to that limited question, what it likely means, since those stats COUNT VALUE, is that the class of starting pitchers have more opportunity then the class of relief pitchers.

I could just as easily say, how come things like LI, ERA+, ERA and just about any rate stat constantly show that the class of CLOSERS are far superior to the class of starters.

That is the exact same statement that you made where I cherrypick a type of stat. Now when someone tells you the problems with that all you are going to do is keep repeating the same thing.

But if I throw that right back at you, your reaction is very predictable. You will now insist that:

(1) Somebody is arguing that CLOSERS ARE BETTER THAN #1 STARTERS.
(2) Change your position to say, All I ever said was that closers aren't better than #1 starters.
(3) Now do what everybody has done to you, show the problem in relying on only one measure.

But nobody should be dumb enough to do that. Because me using ERA+ to show that CLOSERS as a class are better than starters is:

(1) Irrelevant to the topic under discussion;
(2) Something that I don't necessarily believe;
(3) Has logical problems.

In other words, it is sophistry. And that is exactly the same thing you are doing with this argument. The HoM is not about class warfare, its not about creating some weird building block for a pet theory in current utilization. Its not a soapbox to test your ideas. And it should be insulated from historical meandering about Billy Grabinass because the thread is either about one person or about a value system for FUTURE POINTS IN TIME.

There is no reason to do any of this here at all. You have plenty of threads in Primer to do it.

But you are doing to the HoM exactly the same thing you did with Primer. You are taking a forum that has knowledgable to semi-knowledgable people, infecting with weird positions, then dominating the conversation. You may dupe a few of them with the incomplete logic, but the harm is far greater. Because eventually, a lurker like me is going to be tired of seeing that happen, and specifically point out the various rhetorical tricks and logical problems.

And for all the "That's all I'm saying" you aren't going to do anything except keep plugging with more rhetoric.

If this is just about winning and losing to you, its over. You lost, a long time ago. If its about having "an intelligent discussion" you would never in a thousuand lifetimes offer up such a sophistic argument on an inrrelevant point.
   243. Paul Wendt Posted: June 02, 2006 at 10:56 PM (#2048651)
There's too much for me to read here.

Four score and seven messages ago . . .
no, longer than that, somewhere on page one

Daryn
There are hundreds or examples of great players having down years, really bad years, and recovering. Closers are not allowed to do this. That makes the ten year closer equivalent to the 20 year hitter, in my view.

equally scarce, not equally valuable (nor equally meritorious, in my opinion)

Daryn,
Broadly, this is quota thinking rather than value thinking, thus broadly contrary to your position on page three. (I commented on that earlier this hour, noting that this project isn't purely one of measuring value, a very different point.)
   244. Backlasher Posted: June 02, 2006 at 10:57 PM (#2048652)
But, at least to this interested observer, I don't think a retrospective review of value is all that the voters here are striving for.


Really, so you think that if they determined that Dwight Gooden could have won 350 games if not for cocaine, he goes into the HoM? You think that if they determine absent a Beane ball, Dickie Thon would have gotten 3000 hits he should go in the HoM.

I have read a few of these threads and I don't know Grasscock from Grabinass, so I'm not sure what they have always done. I've seen them not penalize someone or setup arbitrary threshholds because of lost opportunity. IOW, value rate higher, or in their venacular "peak performance"

I do see them context-neutral adjust way too much for my taste, but I laid out of the discussion. It would seem the intent in doing that is to reach a value proposition rather than a woulda/coulda/shoulda proposition. Sometimes I think several of them miss badly, but I stay out of that discussion too.

I've had that argument many times, the proponents will just walk in circles claiming what they are doing is determining value when you are expressly showing them why they have not done that, but have created an artificial environment. If they truly believe they are calculating value, then it takes years to show them why they are not.

But I haven't seen anyone, anywhere go to the extreme of class degredation. Treder is pounding over and over and over and over again that you should discount the actual value because they are inherently inferior to somebody else because how a third party has decided to use them (even though that third party is likely optimizing value).
   245. Paul Wendt Posted: June 02, 2006 at 11:08 PM (#2048656)
Repeating #249 with references.
Four score and seven messages ago . . .
no, longer than that, somewhere on page one

Daryn #088
There are hundreds or examples of great players having down years, really bad years, and recovering. Closers are not allowed to do this. That makes the ten year closer equivalent to the 20 year hitter, in my view.

equally scarce, not equally valuable (nor equally meritorious, in my opinion)

Broadly, this is quota thinking rather than value thinking, thus broadly contrary to your position on page three #213. (I commented on that earlier this hour, noting that this project isn't purely one of measuring value, a very different point #242.)

--
I don't know what everyone has learned in the last eight score and fourteen articles. Since I don't have time to find out so I'll now pass on this one.

Eric Chalek: maybe you need a website for posting tables such as those back near #088.
   246. dlf Posted: June 02, 2006 at 11:17 PM (#2048661)
Really, so you think that if they determined that Dwight Gooden could have won 350 games if not for cocaine, he goes into the HoM? You think that if they determine absent a Beane ball, Dickie Thon would have gotten 3000 hits he should go in the HoM.

No, but I've seen them extend backwards the career of Lip Pike or Dickie Pearce into the pre-NA days. In this "years" discussion of Clemente, one poster mentions a one year bonus for an early death. Likewise, in the discussion of Wilhelm, someone mentions a bonus for WWII service. There have been others that I'm sure I could find that are similar and - more importantly - on borderline candidates rather than the apparently overwhelming Roberto and Hoyt. It seems that the starting point is value, but it has been adjusted by some voters to add extra credit for things that did not happen between the lines and enter somewhat into the woulda-coulda.

I really have no interest in another Backlasher / Treder argument and don't feel a need to agree or disagree with the position either proponent is offering with regard to relief pitching. Nor am I suggesting it is right or wrong for the HOM to have moved away from pure value. For good or bad, I do think that they have abandoned the pure count-up-the-value-actually-added approach that you suggest they use in the first part I quoted in the last post.
   247. John (You Can Call Me Grandma) Murphy Posted: June 02, 2006 at 11:41 PM (#2048677)
No, but I've seen them extend backwards the career of Lip Pike or Dickie Pearce into the pre-NA days.

Extend backwards? We gave them credit for what they did as professionals. Sounds like crediting value to me.

In this "years" discussion of Clemente, one poster mentions a one year bonus for an early death.

I don't know who the poster was, but that's actually frowned upon by our group.

Likewise, in the discussion of Wilhelm, someone mentions a bonus for WWII service.

Okay, now you have a case. But as Bill James pointed out in the NBJHA, DiMaggio, Gordon, Willams, and Greenberg were still great players during WWII. However, they were denied the opportunity to play, so why should they be penalized for that? Besides, the WWII generation will be shortchanged without proper credit compared to other eras..
   248. Backlasher Posted: June 02, 2006 at 11:53 PM (#2048682)
I really have no interest in another Backlasher / Treder argument and don't feel a need to agree or disagree with the position either proponent is offering with regard to relief pitching.

Nor am I at all interested in discussions about how people are above the fray, etc. etc.

I am interested in the specific points you are discussing namely

No, but I've seen them extend backwards the career of Lip Pike or Dickie Pearce into the pre-NA days. In this "years" discussion of Clemente, one poster mentions a one year bonus for an early death. Likewise, in the discussion of Wilhelm, someone mentions a bonus for WWII service. There have been others that I'm sure I could find that are similar and - more importantly - on borderline candidates rather than the apparently overwhelming Roberto and Hoyt.

and assertations that you attribute to me such as:

For good or bad, I do think that they have abandoned the pure count-up-the-value-actually-added approach that you suggest they use in the first part I quoted in the last post.

On the former, yes they do that. They also use DER as a substitute for figuring out pitching talent. I agree with neither of these things. I don't think you bonus up people for woulda/coulda/shoulda, and I don't think you use an artificial environment that discounts value when making a value based judgment.

But its reaching heavily to say that I assert "count-up-the-value" I have said its about value, and it is about value. Their constitution expressly says

"Voters are strongly encouraged to consider only a player’s on-field accomplishments and other factors which had an impact on the outcomes of the player’s baseball games. "

They can interpret however they want, and have done so without my inteference. I just don't want to see someone who is going to introduce irrationality disgused as rationality to infect that process.

I'm much more interested in how and why they arrive at their decisions, then influencing their decision. I don't want them to be duped either.

There is no practical way to "count up the value" It just doesn't exist. IF they reached consensus that XYZ properly accounts for value, then game over. They just perform a calculation and take the top of that number.

Determining value requires a multiple part process.

If they want to compare rate of output for someone who had a reasonable basis for not accumulating total value; I'm just going to watch. I think they call this "peak" versus "career".

How they do that I will agree and disagree probably without much input.

But what is being advocated is an assessment methodology and a point of view that devalues a class of players. You strip away the rhetoric, and the desire to be right, this is what they are being encouraged to do:

(1) Reduce or ignore rate stats like ERA+
(2) Ignore or reduce the amount of leveraged value that is being supplied
(3) Give a bonus just because of opportunity (e.g. IP) without regard to performance in that opportunity.
(4) Look upon an entire class of players with disdain, because of (insert lots of irrelevant and irrational arguments).

That's a much different class of woulda/coulda/shoulda.

I personally don't think there should be a lot of relief pitchers that would receive an honorific. I opine that if the method is objective, the amount will grow with time because the value provided by certain relief pitchers has grown with time.

There isn't any reason for them to have a quota by any stretch of the imagination. But no matter how good somebody may think Kent Tekulve was, its extremely unlikely he put up enough value for even the borderline case. And I seriously doubt that THE GOOSE put up enough value either, even if everyone remembers that he was so impressive. But I'd bet he probably does make it in.

I can't see any plausible argument that is going to keep Rivera and Wilhelm out.

But the interesting thing is going to be for a lot of the post 1990 guys. There are players still active that are severly underrated in terms of value. There are a few they will decide upon, including one that started this conversation, that just don't deserve to be short-shrifted.

Most of the merit boys came to this conclusion. But I imagine most everyone lost site of where this argument started (and around which player).

If these guys want to do woulda/coulda/shoulda that is there perogative. I don't vote.

But its going to be pretty weird if they start diminishing an all-time great season because they get duped into ignoring ERA+ and provide a class based demerit on a player because of the way in which they provided value.
   249. Daryn Posted: June 03, 2006 at 12:02 AM (#2048687)
equally scarce, not equally valuable (nor equally meritorious, in my opinion)

That's a good and fair point Paul. I should say that scarcity is a possible/probable indicator of value, and clearly not equal to value. It is like Sim scores -- most (perhaps all) players whose most similar player scores under 800 are superstars, but it doesn't have to be the case.
   250. Harold can be a fun sponge Posted: June 03, 2006 at 12:15 AM (#2048703)
duped into ignoring ERA+

Just because someone downplays or ignores ERA+ doesn't mean that they're being duped. While ERA+ is great, it still has plenty of flaws, and the magic 606 represents a bunch of those flaws (as well as a truly awesome pitching performance). I'm not going to re-start the ERA+ argument; I just bristle at the characterization of people who understand these flaws as rubes who have been conned.
   251. Daryn Posted: June 03, 2006 at 12:20 AM (#2048710)
Nobody can diminish Eck's 606 season -- more saves than baserunners. 3 walks was it? That season was insane.
   252. Backlasher Posted: June 03, 2006 at 12:26 AM (#2048719)
I just bristle at the characterization of people who understand these flaws as rubes who have been conned.

If you understand the flaws and account for them you are neither a rube nor have you been conned.

If you simulteanously understand the virtues, then you would seem to be a wise man.
   253. JC in DC Posted: June 03, 2006 at 01:59 AM (#2048891)
Apparently the Yankees didn't get the memo about the interchangeability of RPs, and the ease w/which they can be replaced. Erickson and Farnsworth are blowing another late lead for NY.
   254. Steve Treder Posted: June 03, 2006 at 02:10 AM (#2048916)
JC, what's your response to this question?

Why is it, JC, that metrics such as Win Shares and WARP rate relief pitchers as significantly less valuable than starters or position players? Are they wrong in that assessment?
   255. JC in DC Posted: June 03, 2006 at 02:33 AM (#2048960)
I don't really have an answer, Steve. I don't know enough to make a judgment, and I wouldn't rule out that they are wrong, either.
   256. Steve Treder Posted: June 03, 2006 at 03:07 AM (#2048998)
Fair enough, JC. Please understand that it is a fact that the cumulative judgement of those who have devoted countless years to studying such things with painstaking precision is that relief pitchers are significantly less valuable than starters or position players. This isn't my opinion or anyone else's; it is a fact. Of course they might be wrong, but the mere statement of the obvious truth that they might be wrong doesn't begin to amount to an argument that they are wrong, much less in what manner they are.
   257. Chris Dial Posted: June 03, 2006 at 04:01 AM (#2049063)
I think LI is too high. As BL notes, saves are converted at a really high rate, and even blown saves are wins about half the time.

I think that most closers are over-valued. Rivera has been exceptional, but he pushes the limit for any other reliever.

The difference between a great Mo season and Braden Looper is about 13 runs, and that's with 10-15 games that aren't save situations.
   258. Dr. Chaleeko Posted: June 03, 2006 at 04:48 AM (#2049084)
Although I feel like BL and ST are ping-ponging between different sides of this question as I read, what I'm drawing from them is not mutually exclusive.

Gentlemen, I apologize for being reductive, but...

ST says...

1. that relievers have, by all the omnibus total value stats we all use as a shorthand for value, shown that relievers post less value per annum and per career than do other players.

2. that unless leverage is shown to be poorly accounted for among the omnistats, we have to address the group expectation for what a HOM reliever looks like so that we can set reasonable standards for RP induction.

Meanwhile BL says ...

1. that doesn't relievers are unfit for consideration as a group.

2. that value is value is value.

I think these are all four reasonable points. My main concern is with ST's point two: that we make sure our measuring sticks are correctly calibrated for the job, and that we use them and wisely. I do think we recognize BL's points as a group already, and in fact, I think that the WS/WARP uberstats go a long way in helping us address these points. In general, I think we've made an effort to ensure that both his major points have become important parts in our election process. (Or I could be acting self-congratulatory toward the HOM; ain't the first time!)
   259. Dr. Chaleeko Posted: June 03, 2006 at 04:55 AM (#2049090)
To clarify a really unclear past-midnight post..., what i mean is that the non-discriminatory/non-classicist themes of BL's posts and the value is value is value theme (or a pennant is a pennant is a pennant approach) have been integral in our election process. And that WARP/WS actually contribute to this process by offering proxies for value that we can all loosely agree on, then make our own individualized adjustments to, in order to avoid the classist/biased judgements that can occur when you use a number in isolation of context.
   260. Backlasher Posted: June 03, 2006 at 05:01 AM (#2049094)
I think these are all four reasonable points. My main concern is with ST's point two: that we make sure our measuring sticks are correctly calibrated for the job, and that we use them and wisely. I do think we recognize BL's points as a group already, and in fact, I think that the WS/WARP uberstats go a long way in helping us address these points. In general, I think we've made an effort to ensure that both his major points have become important parts in our election process. (Or I could be acting self-congratulatory toward the HOM; ain't the first time!)

FWIW, I do think you are addressing the concerns. It's likely I will not totally agree where you land, but I am not and have not criticized the process.

IMHO, I think you guys have some work ahead. I think the BBWAA is struggling with this same question contemporaneously.

And when I watch you it seems that you discuss the issue thoroughly, e.g. IF the metrics under-represent the class are the metrics correct? Should rates be valued as high, since rates should be increased in that role? Should leverage play a role, and do we have our hands around the leverage? etc.

I watch your threads and I see that discourse, and it seems like you are pretty balanced on discussing the problem. I have no desire to force you to my train of thought on the issue. I actually trust that, even though I may later disagree with the final result, most of the folks that have been active in these threads are reasonably fair. (I had some skepticism that would be the case when I really started reading these threads).
   261. Joey Numbaz (Scruff) Posted: June 03, 2006 at 05:21 AM (#2049100)
"They also use DER as a substitute for figuring out pitching talent."

Just as a point of information - DERA is NOT DIPS. Nothing to do with DIPS. It's simply an adjustment to a pitcher's ERA for the quality of the defense that played behind him. It doesn't give bonuses for strikeouts, etc.. It's just an adjustment to accounts for the difference between having Ozzie Smith at SS as opposed to a bad SS.
   262. Joey Numbaz (Scruff) Posted: June 03, 2006 at 05:24 AM (#2049101)
"Erickson and Farnsworth are blowing another late lead for NY."

Huh? Farnsworth retired everyone he faced. I just got back from the game.

Erickson on the other hand is a complete disaster. What the hell are they thinking there.

I can't believe I'm longing for the days of Tanyon Sturtze . . .
   263. Joey Numbaz (Scruff) Posted: June 03, 2006 at 05:32 AM (#2049103)
"Why is it, JC, that metrics such as Win Shares and WARP rate relief pitchers as significantly less valuable than starters or position players? Are they wrong in that assessment? "

For one, if WARPs leverage adjustment for Wilhelm is only 1.14, that's a big reason. His leverage had to be higher than that.

Give Wilhelm a 1.3 LI and I get him equal to Bunning and Drysdale, who just waltzed into the Hall of Merit. Give him anything higher than that and he's ahead of the game.

In the NHBA the "Wilhelm reliever" model (there were 5 - Clint Brown, Elroy Face, Hoyt Wilhelm, Bruce Sutter, Robb Nen) had more leverage than any other type - almost double the leverage of a typical inning (just 4.47 runs needed to flip a full game, as opposed to the typical 9-10). But a LI system that is based on saves only is going to miss this.
   264. Joey Numbaz (Scruff) Posted: June 03, 2006 at 05:46 AM (#2049105)
OK - I was basing that 1.14 in post 269 off what someone had said earlier in this thread - in reading the Wilhelm thread it appears they rate his leverage higher than that. I'm not sure if that rating makes its way into WARP though.

It looks like it might, because they get Wilhelm right in the Drysdale/Bunning ballpark with WARP1 . . . and with my system he needs about a 25-30% boost to get there - and it's a system that's similar to WARP (I make many adjustments, and have a higher replacement level). But again, I think that's on the low side of what his leverage actually was.
   265. Dr. Chaleeko Posted: June 03, 2006 at 06:32 AM (#2049110)
Here's something from a BP Mailbag article that explains their WXRL stat for relievers.

http://www.baseballprospectus.com/article.php?articleid=3968

Woolner writes:
Win Expectancy is inherently a context-sensitive measurement (indeed, that's what separates it from context-independent measures like VORP or EqA). It would be possible to measure something like "percentage of possible increase in win probability attained", and in some circumstances that might be an interesting measure to look at.

I don't quite understand what the second means because that's what I thought we were getting with WXRL, but what it means to me is that WXRL and WARP are not very compatible. One's about context, one is not. So then the question becomes which one to use? I think we'd need a big cross-reliever comparison of WARP and WXRL to establish which stat measures what and how before we answer that.
   266. Dr. Chaleeko Posted: June 03, 2006 at 06:36 AM (#2049111)
Also Woolner says this:
I chose to focus more on the pitching side, because of the close relationship with BP stats like Support Neutral Win/Loss records (SNWL).

Chris or Jimd, is SNWL any part of the WARP equation? If so, this might be the spot where WARP/WXRL meet.

I'll read up on it in te 2005 annual this sesaon and tell you what I can.
   267. Joey Numbaz (Scruff) Posted: June 03, 2006 at 06:41 AM (#2049113)
I'm not a big fan of Win Expectancy. I kind of hate it really.

I mean if you stretch it to Pennant Expectancy, is Bucky Dent's 1978 the greatest season of all time?

I think it's pretty much useless for hitters - they can't be leveraged. ARod's 3rd inning HR is every bit as important as Ortiz's 9th inning HR. You can't leverage a hitter, unless he's a PH.

But for relievers it's somewhat different because they can be leveraged, so managers do get added value out of it. But I'm still not a huge fan of Win Expectancy.
   268. Dr. Chaleeko Posted: June 03, 2006 at 06:50 AM (#2049114)
Uh, it only took me like two minutes to look up win expectancy in the 2005 and 2006 bp annuals. Turns out I haven't got the foggiest. Woolner explains the process of arriving at Win Expectancy (as you'd expect), but beyond describing leverage, he doesn't offer a lot of context. There's a reliever table with best/wort seasons since 1972, but I can't for the life figure out where the derived stats come from.

He does do a bit with SNWL for starters (as noted above). The 2006 book is all hitting and running.
   269. Joey Numbaz (Scruff) Posted: June 03, 2006 at 06:59 AM (#2049117)
One thing I highly recommend is reading the article in the NHBA called Valuing Relievers, which starts on page 232.

One nugget . . .

"A hundred runs saved by a relief ace have substantially more impact than a hundred runs saved by a starting pitcher in all cases, even when the reliever's job is just to come in and pitch when the starter is knocked out."
   270. Joey Numbaz (Scruff) Posted: June 03, 2006 at 07:05 AM (#2049119)
By the way, the usage for the "Clint Brown" reliever, the one referred to at the end of post 275 was his workload was determined randomly based on the starter's lack of effectiveness and the reliever's fatigue. He averaged 58 games per season and 106 IP, 10 saves. If the starter was knocked out early, he might pitch 5 or 6 innings that day, then be out for 3 or 4 days, etc..

Even in that pattern, one win was gained for every 5.88 runs he saved (starters turned one win for every 8 to 9 runs).
   271. Joey Numbaz (Scruff) Posted: June 03, 2006 at 07:08 AM (#2049121)
He also concluded that a run saved had anywhere from 36%-97% more win impact, depending on the comparison, and that the short answer would be a run saved by a relief ace has about 70% more win impact than a run saved by a starting pitcher.
   272. John (You Can Call Me Grandma) Murphy Posted: June 03, 2006 at 11:57 AM (#2049137)
Just as a point of information - DERA is NOT DIPS. Nothing to do with DIPS. It's simply an adjustment to a pitcher's ERA for the quality of the defense that played behind him. It doesn't give bonuses for strikeouts, etc.. It's just an adjustment to accounts for the difference between having Ozzie Smith at SS as opposed to a bad SS.

It's one of the few things from the BP gang that I actually look at.
   273. sunnyday2 Posted: June 03, 2006 at 12:44 PM (#2049140)
Somebody has probably already answered this question and it really goes to basics.

But is it universally agreed that there is leverage? That leverage is a valid concept?

Does leverage deny that an inning is an inning, an out is an out? I'm pretty sure it posits that an out with RISP in a one-run game in the 9th inning is more valuable than the first out in the 1st, right?

Does everyone agree that this is so? Or are there hold-outs who say otherwise?
   274. John (You Can Call Me Grandma) Murphy Posted: June 03, 2006 at 01:13 PM (#2049144)
But is it universally agreed that there is leverage?

I'm pretty sure Archimedes cleared that up a while back, Marc.
   275. Howie Menckel Posted: June 03, 2006 at 01:50 PM (#2049152)
Actually, last night's "blown save" by Farnsworth was illustrative of the problems with save pct.
Farnsworth inherited a bases-loaded, one-out situation with a 5-4 lead in the bottom of the 8th.
- He induced a grounder to Cairo at 3rd that should have been an inning-ending DP. Cairo bobbled it, and settled for getting the guy at 1st as the tying run scored.
- Farnsworth then got the 3rd out, and punched out the Orioles in the 9th for a Yankees win.

This is an outstanding performance, yet it goes in as a "blown save." That would be true even if he had merely allowed an outright error with the bases loaded and NO one out - if the tying run scores, 'you blew it.'

Interestingly, 21st-century closers rarely are put in this spot, which is part of the reason their save percentages are so high. Clearly Farnsworth "saved the game" in the colloquial sense - he put put out the fire in the 8th with manageable damage (none of it his fault), and he shut down the Orioles in the 9th for a Yankees win.
This is FAR more valuable than a scoreless 9th in a 5-2 game, obviously.

As for the rest of this thread, the whole "they" thing I keep reading seems a little odd. I'd understand it better if it were a separate BTF thread that happened to refer to the Hall of Merit; as it is, um, we're right here.
And paternal concerns aside, no, we aren't going to be dazzled by anyone's bright and shiny new idea, like kids at the toy store.
But we will take into account fresh opinions. They're quite helpful, especially with a challenge such as this one.
No offense intended.
   276. Mefisto Posted: June 03, 2006 at 01:57 PM (#2049155)
Somebody has probably already answered this question and it really goes to basics.

But is it universally agreed that there is leverage? That leverage is a valid concept?

Does leverage deny that an inning is an inning, an out is an out? I'm pretty sure it posits that an out with RISP in a one-run game in the 9th inning is more valuable than the first out in the 1st, right?

Does everyone agree that this is so? Or are there hold-outs who say otherwise?

I don't know if I count as a holdout for HOM purposes, but I'm far from sold on Win Expectancy as a measure of value.
   277. Daryn Posted: June 03, 2006 at 02:22 PM (#2049161)
As for the rest of this thread, the whole "they" thing I keep reading seems a little odd. I'd understand it better if it were a separate BTF thread that happened to refer to the Hall of Merit; as it is, um, we're right here.

The other thing about the "they" references is that there is no "they" here. There is no consistency of thought or process among the 50 voters, which I see as a good thing. There are 50 different ways of evaluating value, that at best could be separated into 6 or 7 general groups. I proudly belong to the quasi-maligned traditonal stats group whose most public members are yest and karlmagnus. You can identify us because our posts reference things called "hits" and "wins".

It reminds me of a Far Side knock off cartoon about the ubiquitous "they", I think it was Bizarro, where a guy is sitting in a room with a phone lablelled THEY and answering questions about the weather: "Yes, it is going to rain".
   278. Steve Treder Posted: June 03, 2006 at 02:30 PM (#2049165)
Give Wilhelm a 1.3 LI and I get him equal to Bunning and Drysdale, who just waltzed into the Hall of Merit. Give him anything higher than that and he's ahead of the game.

Joe D.,

I completely agree with you that James's "Valuing Relievers" article is extraordinarily good, the most comprehensive and straightforward look at this issue I've yet encountered. And it is highly relevant to this discussion that his analysis concludes that the "Wilhelm Model" is the most highly-leveraged, more highly-leveraged than the modern closer usage pattern, the "Nen Model."

Yet James's player performance rating system itself, Win Shares, doesn't see relievers as delivering nearly the value of players at other positions. In 1964 and 1965, Wilhelm's most monster great years, his best reliever years according to Win Shares, the system gives him scores of 21 and 19, nothing close to what the best at other positions were generating:

1964:
Allen 41
Mays 38
Santo 36
Mantle 34
Aaron 33
F Robinson 33
B Robinson 33
E Howard 32
Chance (starting pitcher) 32
Clemente 30
Hansen 30
.
.
.
Drysdale 26
Larry Jackson 25
Marichal 25
Koufax 24
Gibson 24
.
.
.
Bunning 22
Wilhelm 21 (the highest of any reliever in 1964)

1965:
Mays 43
Allen 33
Koufax 33
B Wiliams 33
Oliva 33
Versalles 32
Santo 32
Aaron 31
Wynn 31
Buford 30
Morgan 30
Marichal 30
.
.
.
Drysdale 27
Bunning 27
Gibson 26
McDowell 25
Short 24
Maloney 23
Stottlemyre 23
.
.
.
Stu Miller 22 (the highest of any reliever in 1965)

In career totals, Win Shares has Wilhelm at 256, which is indeed almost identical to Drysdale (258) and Bunning (256), but that is of course entirely a function of Wilhelm's far greater longevity, as well as the fact that he spent a couple of years as a starter.

So here's the thing: James is the guy who writes, "A hundred runs saved by a relief ace have substantially more impact than a hundred runs saved by a starting pitcher in all cases, even when the reliever's job is just to come in and pitch when the starter is knocked out." Yet even his system has the greatest of all relievers valued far, far lower than starting pitchers and position players in his very best seasons.

If you're correct, Joe, and Win Shares under-measures leverage, it is the case that it under-measures it by a very great deal, and its creator has clearly stated how hugely he sees relievers garnering leverage.

Is Win Shares truly that far off the mark?
   279. Mark Shirk (jsch) Posted: June 03, 2006 at 02:44 PM (#2049169)
I would expect Wilhelm's LI to be less than Rivera's. When you pitch three innings at a time you are bound not to have the LI of someone who pitches 1 carefully selected inning.

That said, I am not sure we have to worry about adjusting Wilhelm to the level of Drysdale/Bunning (which sounds about right to me anyway) as Wilhelm will also ease into the HOM. It is the guys after him that are the impetus for the conversation.
   280. Steve Treder Posted: June 03, 2006 at 02:51 PM (#2049171)
That said, I am not sure we have to worry about adjusting Wilhelm to the level of Drysdale/Bunning (which sounds about right to me anyway) as Wilhelm will also ease into the HOM. It is the guys after him that are the impetus for the conversation.

Sure, but the conceptual issue of assessing the proper relative value of a reliever against that of a starter or a position player is going to be crucial there, and Wilhelm, being that he's generally agreed to be the very best of all relievers, is highly instructive as a test case in this regard.
   281. Joey Numbaz (Scruff) Posted: June 03, 2006 at 03:40 PM (#2049182)
"I don't know if I count as a holdout for HOM purposes, but I'm far from sold on Win Expectancy as a measure of value. "

I'm with you Mefisto.

But I still think leverage is important when measuring relievers. The reason WE is absolutely useless to me for everyone else is because they aren't being leveraged. They come up once every time through the order, etc..

But relief aces (note, I didn't say relief pitchers, the non aces and non setup men are used in low leverage actually) are specifically used in high leverage situations. They pitch in close games where every at-bat is magnified (you don't have 7 innings to regain the lead) and it's not a random thing - that's that difference.

If the idea that relief aces pitch in higher leverage innings than starters hasn't been universally accepted, it should be. The only question is "how much leverage is there?"; not, "is there leverage?"

****************

Steve - Win Shares makes a very weak attempt at crediting leverage. You get what are called "Save Equivalent Innings". SEI are calculated as SV*3 + Holds, to a maximum of 1.9 * IP.

You also get credit in one of the claim point areas for (3*W) - L + SV.

I've built a spreadsheet that calculates WS. What's great about it, is that I could just pull out something like Saves and see how the numbers change.

I did this for a Diamond Mind league that we played based on the 1924 season.

Take Rube Ehrhardt, who was used as a closer in our league that year. He pitched 86 innings, went 8-1 with 20 saves and a 2.08 ERA. He ended up with 16.45 WS. It was a 154 game season.

Removing his saves and holds (it was a DM league, so we have holds too) from the equation and the saves of everyone on the team and he drops to 10.13 WS. So I guess you could say WS gives him a 1.62 LI for that year. That's just one example, I wish I had more, but I've never entered a modern season into the sheet.

Here's the rub though - you can't compare pitchers straight up against hitters using WS. Read my posts on the 101-200 page of the 1977 ballot thread for more detail - no time to rehash that now, but essentially replacement level for position players in WS is much higher than for pitchers. For a 154-game season, without the DH replacement level for position player in 162 games should be 11.7 WS. For a pitcher in 220 IP, replacement level is only 6 WS.

Also, since relievers pitch fewer innings than starters, they lose less when you subtract off for replacement level.

Both of these built in biases explain why it looks like WS doesn't value relief aces as high as it actually does. This also explains part of the reason why everyone always says WS don't credit pitchers enough - they do if you subtract the right amount from each group of players.
   282. Backlasher Posted: June 03, 2006 at 03:41 PM (#2049183)
I would expect Wilhelm's LI to be less than Rivera's.

His LI is lower than Rivera's. That discussion shows many of my fears.

(1) James has walked away from that article.
(2) That text you dealt with was about the results of a simulation , not actual results of the player.
(3) There are many flaws with the simulation for that purpose. It doesn't account for injury, ineffectiveness from use, durability etc. In fact, the criteria that was used was only if the pitcher had not been used in the previous day, and often that pitcher was used for three inning stints. James statements on use of relievers is incorrect.
(4) Its been articulated in such a way that it has been made to look like:
(a) a fact that a pattern of usage is more optimal.
(b) a fact that Wilhelm was more leveraged than Rivera.

The other thing about the "they" references is that there is no "they" here.

"They" reference the group of regular contributors to the Hall of Merit, especially the "voters".

There is no consistency of thought or process among the 50 voters, which I see as a good thing. There are 50 different ways of evaluating value, that at best could be separated into 6 or 7 general groups.

Yes, I have said that over and over. And the problem that I am worrying about is already starting to happen. There is no way that a debate about utilization strategy goes in this forum. But you already see the real objective as it is being segued into. And its already being done so with misdirection. Moreover, its already being played with rhetorically. James old study is being labeled as "objective", which is clearly not the case. James wanted to believe such things, and later moved away such things, particularly after his experience in Boston. Its being labled as "comprehensive". Its a couple of pages and all it did was run pitchers through a simulator with artificial criteria. In just about any thread on the subject, more study has been made, more has been written on the intracy of the subject, and it has been done more objectively.

That said, I am not sure we have to worry about adjusting Wilhelm to the level of Drysdale/Bunning (which sounds about right to me anyway) as Wilhelm will also ease into the HOM.

You don't have to adjust. Even if you use your existing uberstat, he is already there.

That said, I am not sure we have to worry about adjusting Wilhelm to the level of Drysdale/Bunning (which sounds about right to me anyway) as Wilhelm will also ease into the HOM.

I haven't seen anyone over here as a holdout. If you go to the mainboard you will see holdouts.
   283. Joey Numbaz (Scruff) Posted: June 03, 2006 at 03:42 PM (#2049184)
"Read my posts on the 101-200 page of the 1977 ballot thread for more detail - no time to rehash that now"

Sorry that's the 1977 ballot discussion thread.
   284. Joey Numbaz (Scruff) Posted: June 03, 2006 at 03:50 PM (#2049187)
Backlasher I agree it was a simulation - a simulation with very specific parameters.

It was looking at 'effectiveness' equal, what usage pattern gets the most leverage.

That's important for us - as you've said numerous times in this thread - we care about value. So his study shows us, for a given effectiveness (ERA+, DERA, ERC whatever your preference), what pattern gets the most out of relief aces? Ballpark, what's the leverage of a relief ace, given a certain effectiveness.

All it proves is that there is leverage, given a certain usage pattern of a pitcher - relief ace.

I'm not talking about the part where he advocates a certain usage pattern. What's relevant to us from that article is 1) Leverage clearly exists 2) He gives us an idea of just how much leverage there is.
   285. Steve Treder Posted: June 03, 2006 at 04:24 PM (#2049195)
If the idea that relief aces pitch in higher leverage innings than starters hasn't been universally accepted, it should be. The only question is "how much leverage is there?"; not, "is there leverage?"

Absolutely.

Win Shares makes a very weak attempt at crediting leverage.

I would call it a clumsy attempt. I'm not nearly as convinced as you are that its bottom-line effect is to significantly under-weight leverage.

you can't compare pitchers straight up against hitters using WS

Well, you can. James does so, freely and eagerly. He would argue, I'm sure, that the differences in replacement level he's established between pitchers and hitters are based on differences in the actual scarcity of talent between pitchers and position players, as even starting pitchers are single-skill specialists to a degree that position players aren't.

But assume that WS is messed up in this regard, and you genuinely can't compare pitchers straight up against hitters. You can still compare relief pitchers straight up against starting pitchers, and despite the fact that:

since relievers pitch fewer innings than starters, they lose less when you subtract off for replacement level

relief aces still score significantly lower than starters. In 2001, for example, in the heart of the closer era, Rivera had the most WS of any reliever, with 19. Seven starters had more WS that season, led by Randy Johnson with 26.

Both of these built in biases explain why it looks like WS doesn't value relief aces as high as it actually does.

I wouldn't put it that way. Both of these biases explain why WS doesn't value relief aces as high as it values starters and position players. One can say this is a flaw of Win Shares, but it assigns the scores it does from a finite starting point of team wins. It actually does value relief pitchers with the WS totals we see; if we were to tweak the formula and assign more WS to relief aces than they have, those incremental WS are going to have to come from someone else's existing total.

This also explains part of the reason why everyone always says WS don't credit pitchers enough - they do if you subtract the right amount from each group of players.

Fine, but if you subtract the right amount from each group of players (or modify the results in any other way), you've come up with a different metric than the existing WS system.

I would agree with you (or perhaps, I think you might agree with me) that the biggest weakness of the WS system is its handling of pitchers, and particularly relief aces. Like you, I'm not at all willing to take its scores as anything close to the last word on these guys. But I think I'm a lot less willing than you are to confidently assert that it's simply a matter of WS undervaluing leverage, for two reasons: the consistent finding of WS and WARP regarding their relative valuation of relief aces vs. other players (that consistent finding makes me hesitant to assume that both happen to be wrong about the same thing), and the general trickiness of quantitatively measuring leverage. I'm not confident that these metrics have it right, but I'm also not confident to what degree and in which direction they don't have it right.
   286. rawagman Posted: June 03, 2006 at 04:30 PM (#2049200)
Lots of conflicting ideas being bandied about here.

Relievers, as we know and value them, are a fairly new invention.
MLB managers and scouts and GMs and organizations in general are still working on the best ways to use all of their pitchers so as to maximize their chances of winning.
You see teams spend freely on good arms for their pen (Cubs), teams who rely on kids (A's), teams who ignore the bullpen and get roasted (Red Sox and Blue Jays in recent seasons - not yet this year).
Today, a good closer can make big money - ask B.J. Ryan.
In the task the Hall of Merit has taken on, there is no real historical viewpoint for relief pitchers. Why?
There is no history for relief pitchers. Rather, it is small enough and still evolving quickly enough that we cannot gain the proper consesus on it.
At this point, intuition is the best we can truly offer.
Most will agree about Wilhelm (within 3 ballot spaces) and Rivera. Then we may either come up with more broadly acceptable systems to rate the rest, or wing it, one reliever at a time.
   287. Backlasher Posted: June 03, 2006 at 04:31 PM (#2049201)
Goodness, it looks like this little tangent isn't going to end until it has been disproven factually.

Ok, there is a lot of Loosey-Gooseyness on how one would define a person, a starter, a reliever, or a closer. There are also persons that use variations on the Win Share formula.

However, if you rely on the Hardhat times last year and their categorization of starter and reliever, and you also pick out the players with the most saves and designate them the most frequent closer then.

Median Starting Pitcher - 6 win shares
Median Reliever - 2 win shares
Median Closer - 11 win shares

There is nothing super-duper class warfare going on.

If you are trying to measure "talent" in some way, and you gave Mo Rivera 60 additional garbage innings, kept the other appearances constant and he produced even at a slightly better league average rate, he woulda, coulda been at near the top of the league in Win Shares.

But that is pretty irrelevant because:

(1) Then you would not have likely been able to use him in the leveraged situations where you did.
(2) He would be likely to get injured, and if he didn't, you are measuring him being a freak of nature, not a performer.
(3) He didn't do it anyway, so that value is not there as it might be slightly reflected in Win Shares.

But none of this matters. As you can see this was just a bridge to try to get people to look through a warped prism to conclude that James simplistic, untrue, and outdated statement about a "Wilhelm" model is superior to a simplistic, untrue and outdated statement about a "Nen Model".

If for some reason, you decide "the best releif pitcher(s) of each era should be inducted" then knock yourself out. That is going to have some strange inductees like Quisenberry who really didn't add that much value.

If instead you are trying to induct players based on their performance and the effect on outcome of games, it may feel weird, because ultimately you will leave out the Gossages and induct the Wagners and Hoffmans. But that is the reality of the value provided by these players.

Most likely, you will do like you do with every other position and do some combination of both things at the opportune time.

If your trying to uber-number the situation, Win Shares will be really crude, as is already discussed. You should probably derive your own uber-number. If instead the uber-number is part of the equation, then you can look at these numbers and just understand the flaws.

For Rivera and Wilhelm it doesn't matter too much. But in now way whatsoever, should you ever worry about class warfare. You might worry about performance inside a class.

If you follow the adolescent argument presented above, you can take it to a lot of levels. Let's just cherry pick a year, like last year for instance:

2005 Pujols A STL NL 1B 36.3 0.0 2.0 19 .998 25 38 180
2005 Rodriguez A NYA AL 3B 33.3 0.0 3.3 19 .989 24 37 318
2005 Lee D CHN NL 1B 34.3 0.0 3.0 19 .993 24 37 151
2005 Giles B SD NL OF 32.3 0.0 3.1 18 .974 23 35 234
2005 Bay J PIT NL OF 30.4 0.0 3.3 19 .889 20 34 57
2005 Ramirez M BOS AL OF 30.9 0.0 2.9 17 1.018 22 34 310
2005 Sheffield G NYA AL OF 30.5 0.0 2.2 17 .978 21 33 401
2005 Teixeira M TEX AL 1B 29.1 0.0 3.3 18 .884 20 32 69
2005 Delgado C FLA NL 1B 30.2 0.0 1.1 17 .947 20 31 243
2005 Ortiz D BOS AL 1B 31.4 0.0 0.2 14 1.149 22 31 108

There is your top 10 in WS. Maybe we should just conclude that everybody is inferior to corner players. That no pitchers at all should ever be in, or any SS, 2B, CF, or C.

I highly imagine that the corner OF and 1b are more represented in your hall, but so what. You still elected Biz Markey and Grasscock for reasons that you thought took into account their value. And if you use WS, remember that does have its crude accounting for defense.
   288. Backlasher Posted: June 03, 2006 at 04:43 PM (#2049210)
It was looking at 'effectiveness' equal, what usage pattern gets the most leverage.

That's important for us - as you've said numerous times in this thread - we care about value. So his study shows us, for a given effectiveness (ERA+, DERA, ERC whatever your preference), what pattern gets the most out of relief aces? Ballpark, what's the leverage of a relief ace, given a certain effectiveness.

All it proves is that there is leverage, given a certain usage pattern of a pitcher - relief ace.

I'm not talking about the part where he advocates a certain usage pattern. What's relevant to us from that article is 1) Leverage clearly exists 2) He gives us an idea of just how much leverage there is.


Domino,

FWIW, I agree with your implicit assertion, that once pitchers were selected for leverage, it becomes important to measure leverage. But I'm not trying to advocate that position.

Also, I am not responsive to your statement which I think has value. I just think certain people were waiting for it so they could make their claim about most effective utilization, which is definately not a subject matter for HoM.

But we have had 1000s of posts on utilization model situations. Someone just wants a new audience to confuse. He has been decimated and left with no arguments whatsoever on that front.

But, if the dynamic of this thread leads to a repeat of that discussion, the people that are here go visit those discussions, or later participate in what will likely be that discussion again, I ask one big favor.

PLEASE, PLEASE, PLEASE, read post #84 carefully, memorize it, bookmark it.

Because this is what is going to happen. We will destroy every silly little argument about ACE RELIEVER, and as we slowly start to get people to see why that is the case, Treder will then assert, "all I ever said is that the model will change"

He clearly and unambiguously asserts something about the "Wilhelm model" There is no doubt about that whatsoever.

I have tried very hard to keep those needless conversations out of your Hall of Merit. But he is not going to stop. He is going to keep saying the same thing over and over and over again to get you to devalue relievers without adding anything new. And it will be no new information, just rhetoric, and as you saw above "gotchas".

Right now, he is working hard to get you to "yes" To try to get you to admit something that is not true, e.g. systems devalue relievers in hopes that you will fall like dominos afterward. its not even good rhetoric, its what lawyers call "a jailhouse con."
   289. Backlasher Posted: June 03, 2006 at 04:55 PM (#2049216)
I'm not confident that these metrics have it right, but I'm also not confident to what degree and in which direction they don't have it right.

Good, then if you aren't confident that it is right, the persistent argument of something you know is logically incorrect is called sophistry. You can stop posting the nonsense.

It doesn't matter, because by that logic, WS and WARP fact constantly:

Value position players higher than any pitcher
Value 1b higher than SS
Value 1b higher than C
Value 1b higher than 2b
Value 1b higher than CF
Value RF higher than SS
....

It is meaningless.

They have four things at their current disposal:

(1) Rate stats
(2) Counting stats
(3) Value stats
(4) Comparisons against peers

They may develop a new one of any of these things. But just like with any player before, they should use ALL of these things.

200 posts ago you were telling them they should discount the rate stats b/c of the inherent advantage RPs have against SPs. No you are telling them they should weigh the value stats pretty heavily because the SPs have an inherent advantage over RPs. And then you are using just those value stats to claim that you shouldn't compare against peers because these people are an "inherently inferior class of players" You are waiting for someone to talk about the counting stats because then you can take back the mindshare ground and use your arsenal of "Well, anybody who knows anything about baseball" In fact, you hope that I do it and then really hope somebody else makes a statement first, then you can "Well said" it make your bombastic insult, and still pretend like you are ignoring me.

If you were interested in discussion, you certainly wouldn't have asked those "gotcha" questions over and over. That is nothing but "getting to yes", and old sales technique to seal the deal before people realize that they just bought so much of the blue sky.


You can keep claiming taht "all you are saying is", but that isn't the case. They have four traditional categories they can look at. YOu are asking them to devalue anything that makes the relievers look good, and overcredit anything that makes them look bad. And as a second step, ignore comparitives because they are just those inferior relief pitchers.
   290. Dr. Chaleeko Posted: June 03, 2006 at 04:58 PM (#2049217)
SEI are calculated as SV*3 + Holds, to a maximum of 1.9 * IP.

A leverage of 1.9 would be about right for a slightly above-average closer if I remember Tango's studies right. Which means that it's possible WS is measuring leverage pretty reasonably for closers since this is a proxy for leverage.
   291. Chris Cobb Posted: June 03, 2006 at 05:22 PM (#2049226)
SEI are calculated as SV*3 + Holds, to a maximum of 1.9 * IP.

A leverage of 1.9 would be about right for a slightly above-average closer if I remember Tango's studies right. Which means that it's possible WS is measuring leverage pretty reasonably for closers since this is a proxy for leverage.


Win Shares and WARP, by basing their leverage estimate on saves, holds, and decisions will probably get leverage about right for the post-1985 "closer," because most of the time in which he appears in a single high-leverage inning, it gets marked as high-leverage by counting as a save, hold, or a decision. I think they are probably pretty close to right about the value of Rivera, Wagner, and Hoffman and their contemporary closers.

Any time that relievers are frequently throwing two or more innings or are throwing innings in non-save, non-decision situations that are nevertheless higher than average in leverage, the win-share and WARP calculations of leverage will be increasingly inaccurate. Non-closer relievers and pre-1985 relievers are both likely to be somewhat underrated because their leverage will be undercounted.

If we can get a little bit better understanding of what exactly the leverage index numbers on BP mean and how to apply them appropriately to IP, I think we'll have the information that we need to know in order to evaluate the value of the top relief pitchers post-1960 with reasonable accuracy, and a little tweaking for dealing with their handling of inherited baserunners in assessment of their effectiveness would help, too.
   292. Daryn Posted: June 03, 2006 at 07:29 PM (#2049360)
On the relievers as failed starters issue, which we abandoned yesterday for its limited relevance to begin with, Kelvim Escobar is an interesting case study.

He starts as a starter and is terrible. He is converted to a closer and is average. He then is converted back to a starter and becomes very good. To me this career path lends credence to the theory that the relievers who are given a very short chance as a starter and perform poorly are mostly performing poorly because they are young and adjusting to the big leagues and the success they have as closers might well have also been achieved had they been given the chance to mature as starters.
   293. Daryn Posted: June 03, 2006 at 07:46 PM (#2049384)
Well, you can. James does so, freely and eagerly.

Steve, who is this James idol of whom you so frequently speak?

I ask this because my idols at the tabernacle of baseball analysis are named Cobb and Chalek and Dimino and a handful of others here who have created their own analysis of the issues we are discussing as well as piggybacking upon the existing analysis. Unlike James, who studiously ignores others' research, these researchers studiously study others' research to help improve their own without any sense of proprietary interest in the analysis. It leads to some damn fine information and analysis.

Which is a long way of saying that citing that James believes something to be so is no more persuasive than you saying you or I believe it to be so. Now if Chris Cobb believes it to be so, for better or worse, that will effect this electorate...
   294. BDC Posted: June 03, 2006 at 08:41 PM (#2049499)
This has been a fascinating discussion to read through on a rainy Saturday in New York, and I will try to compose a few stray thoughts and post them after the page break in hopes that they'll get read there :)
   295. BDC Posted: June 03, 2006 at 08:56 PM (#2049552)
It seems the the primary inciting point in the discussion is that relief pitchers simply do not pitch very many innings. Aside from the occasional Mike Marshall, they don't qualify for ERA titles; full-time starters all do. That means that the burden of the argument is all on the position that relievers, particularly closers, do things that are much more valuable per inning than things starters do.

Much of the discussion has swirled around the issue of how relievers get to be relievers, how hard relieving is, &c. This is a very knotty point. Closing could be both very easy to do and very valuable, for instance. So maybe Trevor Hoffman, e.g., is not much of a pitcher but has lucked into doing a very valuable job for many years. He's still helped the Padres win a bunch of games, which is a value, right? no matter how he got the job or how bad he would have been as a starter.

You know what's really hard? being an effective pinch hitter season after season. Even someone like Elmer Valo fluctuated from great to terrible. So did Gates Brown. Manny Mota had some rotten years, and he was about the most automatic pinch hitter I ever saw. And yet, all pinch hitters are failed regulars, right? And yet, pinch hitting is of pretty marginal value: important, occasionally tremendously clutch in the context of a single game, but nobody used to pick the Dodgers to win the pennant because they had Manny Mota.

All this tends to lead me to believe that being a closer is something a lot of pitchers could do well, simply because I observe a lot of them able to do it for a year or two and a fair number able to do it for ten. But the key HOM questions, as people keep coming back to, are whether Mariano Rivera's 80 innings a year are really as valuable to his club as Roger Clemens's 240 innings, and then if they are, whether Mo's ten seasons are really as valuable as Roger's 20. Maybe this means that ounce for ounce Rivera has to be six times as good as Roger. I dunno.
   296. Steve Treder Posted: June 03, 2006 at 10:06 PM (#2049795)
Steve, who is this James idol of whom you so frequently speak?

I ask this because my idols at the tabernacle of baseball analysis are named Cobb and Chalek and Dimino and a handful of others here who have created their own analysis of the issues we are discussing as well as piggybacking upon the existing analysis. Unlike James, who studiously ignores others' research, these researchers studiously study others' research to help improve their own without any sense of proprietary interest in the analysis. It leads to some damn fine information and analysis.

Which is a long way of saying that citing that James believes something to be so is no more persuasive than you saying you or I believe it to be so. Now if Chris Cobb believes it to be so, for better or worse, that will effect this electorate...


Well, gosh, Daryn, I beg your wise forgiveness for being so crudely sacreligious.

In my feeble defense may I simply point out that it was your idol Dimino who invoked James's relief pitcher analysis up there in #275. But of course I will endeavor in the future to remain more faithful to your analytic orthodoxy. :-)

By the way: do YOU believe Win Shares significantly undervalues relief aces? If so, wny?
   297. Joey Numbaz (Scruff) Posted: June 03, 2006 at 10:35 PM (#2049864)
"But the key HOM questions, as people keep coming back to, are whether Mariano Rivera's 80 innings a year are really as valuable to his club as Roger Clemens's 240 innings, and then if they are, whether Mo's ten seasons are really as valuable as Roger's 20. Maybe this means that ounce for ounce Rivera has to be six times as good as Roger. I dunno."

No one would argue Rivera is as good as Clemens. That's not debatable, IMO.

The issue is, is he as good as Clark Griffith? Stan Coveleski? Bob Lemon? Jim Bunning?
   298. Steve Treder Posted: June 03, 2006 at 10:43 PM (#2049881)
The issue is, is he as good as Clark Griffith? Stan Coveleski? Bob Lemon? Jim Bunning?

That's precisely the issue, and coming up with a generally agreed-upon answer to it will go a long way to answering the follow-up issue, which is: where in the list of the Trevor Hoffmans, Lee Smiths, John Francos, Rollie Fingerses etc. does it become clear that HOM consideration is no longer reasonable?
   299. Backlasher Posted: June 03, 2006 at 10:44 PM (#2049882)
Well, gosh, Daryn, I beg your wise forgiveness for being so crudely sacreligious.


Well, golly, then instead of quoting others, why not present some actual analysis.

No one is going to consider you an expert because you memorized an erroneous conclusion from a long time ago.

You have googled up some stats. None of which are real probative.

And lets not lose site of what you have asserted.

-Mariano Rivera is a failed starter. He is inferior to a starting pitcher.

-You have to adjust for ERA+ and other rate stats.

- You should knock relievers just because they were relievers.

Quit trying to get people to "yes". Its not a gotcha. It doesn't matter how that question is answered, it doesn't support your bluster.

If you want to play some gotcha games, here is a question for you: Do you think Kelvim Escobar is a failed closer?

But just let it go. Let these guys figure out how to analyze relief pitchers and quit trying to sell the snake oil to a new crowd.

Win shares offers a decent, but not perfect measure of value. Everybody has already said it. Now let them figure out how they are going to combine: value, rate, peak, milestone, and comparison without all the silliness.

Because by the simplistic, cherrypicked stuff you offer, guys like Roger Clemens and Greg Maddux aren't going to get consideration. Rivera outperformed Halladay on your little stat last year, and even if they look at nothing but win shares, there aren't too many people that are going to compare with the elite closers.

Todd Jones might not, but that is not someone they are going to elect.
   300. Howie Menckel Posted: June 03, 2006 at 10:48 PM (#2049888)
Well, I have a lot of problems with Win Shares anyway, so this debate doesn't change my impression of them.

Watching the Mets game today, I see Dave Righetti, the SF pitching coach.
Reminds me: He had an excellent 105 IP in the strike year of 1981 (175 ERA+), then also is a pretty good SP in 1982-83 at age 22 and 23 (400 total IP with ERA+s of 106 and 114).
Becomes a closer in 1984, and is among the best in baseball the next 3 years, with a total of more than 300 IP in those years. Really not that overwhelming the 5 years after that (for example, can't get his ERA+ over 130), yet he keeps his job. Basically done by age 33.
I think it will be of some use annually to notice the incoming retired Ps - will we see more examples of successful SPs who easily handle closing? Will we see some failed closers become successful starters? Certainly there have been attempts on both ends. My mind's open on this.


this was interesting by daryn...
"The other thing about the "they" references is that there is no "they" here. There is no consistency of thought or process among the 50 voters, which I see as a good thing."
Page 3 of 5 pages  < 1 2 3 4 5 > 

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
greenback calls it soccer
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.8571 seconds
49 querie(s) executed