Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Hall of Merit > Discussion
Hall of Merit
— A Look at Baseball's All-Time Best

Wednesday, September 18, 2002

Pitchers for the Hall of Merit

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

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



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

Joey Numbaz (Scruff) Posted: September 18, 2002 at 05:16 PM | 571 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

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

Page 5 of 6 pages  < 1 2 3 4 5 6 > 
   401. Paul Wendt Posted: December 16, 2008 at 03:26 AM (#3030033)
Joe,

RAA - PRAA per nine innings is equivalent to NRA - DERA.

We are using essentially the same statistical bases except for rounding errors and revisions in the DT cards, and our measures of playing time (I have XIP and IP available, you are using a translated innings), and our clerical errors.

--
On ERA+,

Joe #397
I definitely agree with your general premise that ERA+ is not the answer. It is a nice eyeball metric, for sure. But that is it. It really should not be used as the sole basis of any serious ranking of pitchers.

Howie #398
I refer to this stat a lot, but I totally agree with this.
Anyone has to factor in career length and in-season durability relative to league, for starters.


Those two qualifications pertain to the use of any career rate, such as for pitchers RA or NRA or DERA as well as ERA, or the versions of same that incorporate batting (eg, my today articles use DERA*).

While it may be fair to recommend those qualifications as "starters", it may also be fair to consider them "preliminaries" because they may have been covered elsewhere. Here in the Hall of Merit, in particular, one may suppose that they have been covered regarding the use of OPS+ as the basis for general assessment of players. You know it, Tom Hanrahan's slippery slope.
54 50 44 41 37 36 33 30 09 09 03

Beyond these points for starters or preliminaries, some intermediate points and some advanced points are general to many uses of career rates. In principle one may answer some of those general questions once for all applications. Many do so in Hall of Merit practice. Best 5-year rate and peak rate and so on are examples; they are answers with general application.
   402. Howie Menckel Posted: December 16, 2008 at 03:49 AM (#3030062)
Joe, that's true if Palmer is pitching THE SAME WAY as he would have ithout that defense.
He's trying to strike people out, a lot of them hit it somewhere, and lucky for him he has good fielders. Possible, I suppose.

But if you assume THAT, you always may have problems - especially with outliers like HOM pitchers.

Palmer never gave up a grand slam in his entire career.
Ok, maybe he got lucky and the wind kept blowing in when he gave up long flies to big hitters.
Or maybe not.

Anecdotally, I think I've read interviews of him where he realized what he had defensively.

I agree that team defense is an issue, but again I won't automatically demerit 100 pct of the difference against a pitcher - especially a great one.

We've mentioned Wang 2007 before - sample size issues, but it seemed as if he struck out guys when he had to (guy on 3rd, one out) and otherwise relied on getting grounders.

Great pitchers do interesting things - there's a Maddux anecdote about him once having a nice lead, and throwing an inside fastball that was a little dangerous. HR allowed, but he won anyway and he felt like he had the upper hand next time around.

Hell, we go all the way back to the weird Clark Griffith tale of him supposedly being superstitious against throwing shutouts.
Sounds silly, but he didn't have one in his first 1150 IP (high-scoring era, admittedly), and only had 4, oddly, in 36 CG in 1898 with a 1.88 ERA. Griffith had none while completing 35 of 38 starts in 1899 with a 2.79 ERA.
He seemed to shake the pattern beginning in 1900, but there seems to be something there.
   403. Joey Numbaz (Scruff) Posted: December 16, 2008 at 04:15 AM (#3030114)
Howie, DERA doesn't take into account whether or not the pitcher pitched to his defense. It just adjusts for the quality of defense behind the pitcher.

Do you really think Palmer would have pitched just as well (runs allowed wise) with a bad defense? If so why did Earl Weaver bother getting all of those good fielders. He should have just put a bunch of sluggers out there if the fielders didn't matter.
   404. Joey Numbaz (Scruff) Posted: December 16, 2008 at 04:19 AM (#3030120)
Greg Maddux only gave up 3 GS in his career. Mike Mussina 4. In a much more power friendly era.
   405. Howie Menckel Posted: December 16, 2008 at 04:33 AM (#3030133)
Oh, no, his defense helped him a lot, and it needs to be accounted for.

But great players do interesting things.
   406. Joey Numbaz (Scruff) Posted: December 16, 2008 at 04:42 AM (#3030142)
Absolutely. And Palmer was definitely a great player. Just not quite as great as most think, that's all.
   407. Howie Menckel Posted: December 16, 2008 at 04:45 AM (#3030145)
Well, Palmer's perceived placement is just behind the first set of all-time greats, and I think there's room in the next tier where surprisingly his benefit of better defense doesn't make a ton of difference.

But I could be wrong.
   408. Joey Numbaz (Scruff) Posted: December 16, 2008 at 08:22 AM (#3030251)
I think that's fair.

Before adjusting Palmer for the updated (better I assume) DERA numbers, I have him placed #37 among eligible post 1893 SP. Behind Reuschel, Bunning, Koufax, Marichal; ahead of Faber, Saberhagen, John, Ferrell, Hershiser, Coveleski.

His best years are 7.3, 7.0, 6.0, 6.0, 5.9, 5.8, 5.8. His DRA+ is 113.

If however, I assume he pitched in front of average defenses, he vaults to #21. Now he's in the group just behind Hubbell and Plank, just ahead of Sutton, Jenkins, Drysdale and Wynn.

His best years become 8.8, 7.8, 7.5, 6.9, 6.8, 6.5, 5.9. His DRA+ becomes 120.

I think it's reasonable to adjust that much for defense. It's not like the guys he's being compared to weren't also great pitchers who could pitch to their defenses if that had to also, etc..
   409. Paul Wendt Posted: December 16, 2008 at 05:23 PM (#3030486)
Joe,
This isn't a big deal but it's a medium-size one.

In these lists we have recently posted here, the positive magnitudes are greater than the negative ones. Elsewhere I have summarized by saying that the median for a good career pitcher is near +0.05. Maybe I said that here, too, re interpretation of Tommy John +0.01 and Kaat -0.04.

So Jim Palmer may slosh up and down the ranks a little less than you guess, when you toggle the adjustment for team fielding. (Ten-year teammate Dave McNally must slosh up and down less than you guess, because he is down in the ranks where the population is dense.)

Among the 5000-inning pitchers Kid Nichols benefat most from team fielding, +0.43 runs/9ip. Don Sutton +0.14 (a greater share of his runs saved than Palmer), Warren Spahn +0.13. Those numbers may fly under the radar, but they represent 79 and 76 runs each.

Among 4000-inn pitchers, Radbourn is there with Nichols. Mullane and McCormick bracket Palmer. . . . Clarkson +0.19. Then five-year teammate Dennis Martinez +0.18.

D.Martinez really pitched only 3999.7 inns, Palmer 3948 inns. In that narrow range we have Vic Willis, 3996 inns and +0.20 runs.

. . .
Suppose I cut my 550 pitchers at 1500 innings minimum as a quick way to get rid of most mainly relief pitchers. Of course that costs many of the starting pitchers with extreme records too. The reason is sample size, in a broad sense, not the sample of a pitchers work (where 1400 inns is plenty) but the sample from a team fielding history. "Many" starting pitchers with 1400 inns may "sample" a time of uniformly good or uniformly bad pitching in one team's history. But almost no one who plays twice as long enjoys or suffers almost uniformly good or bad team fielding.

Johnny Murphy and Dick Radatz did not work 1500 inns, so Carl Mays and Ned Garver represent the extreme 20th century slots bar none. Here is the complete list with bad team fielding worse than -0.20 runs/9ip.

. . .
-0.21 Claude Passeau
-0.22 Emil Leonard
-0.23 Wilbur Wood
-0.26 Don Wilson
-0.24 Larry Dierker
-0.29 Ned Garver

(So I have passed over 1890s debuts, Brewery Jack Taylor -0.21 and Win Mercer -0.54(!), five 1880s debuts, and Bobby Mathews with 4900 inns at -0.28.)

On the plus side, the pitchers who enjoyed good team fielding, 0.20 runs/9ip is relatively common, and +.30 corresponds roughly to -0.20.

+0.37 Carl Mays
+0.36 Morde Brown
+0.35 Ed Reulbach
+0.34 Dave McNally
+0.33 Jim Palmer
+0.32 Dan Petry (Jack Morris down at +0.26)
+0.30 Freddie Fitzsimmons

(So I have passed over 1890s debuts Kid Nichols +0.43 and Deacon Phillippe +0.30. Just missing, teammates Jack Stivetts +0.33 with 1889 debut and Sam Leever +0.29. Twelve others pitched mainly in the 1870s-1880s.)
   410. Joey Numbaz (Scruff) Posted: December 16, 2008 at 07:08 PM (#3030658)
Interesting Paul. The spreads are wider now, for one. The highs are higher and the lows are lower, at least from my eyeballing things. I'm guessing maybe they are more confident in the method now and are willing to make a bigger statement with it, so to speak.

Ideally, I want to convert my gigantic spreadsheet into an access database, and include every pitcher. It will take some work, but to have seasonal NRA/DERA in a table, for every pitcher season ever would be the ultimate goal. I'm a BPro Premium subscriber, I need to see if that info is in the exportable stats, but I don't think it is.

Heck, I'd be willing to pay a small fee for the data even.
   411. Joey Numbaz (Scruff) Posted: December 16, 2008 at 07:38 PM (#3030720)
Question for the Leverage Index chaining proponents (DanR, anyone else?):

How do I do this historically? The bullpen in 1928 or 1938 was much different than in 1998-2008. Specifically it wasn't nearly as deep. There were tons of doubleheaders, meaning nothing like a regular rotation really existed, good starters were leveraged against good opponents more (see Chris Jaffe's work); starters who were in-between starts were usually the relief aces.

Because of this, I feel like the impact of chaining was much smaller then.

Any thoughts? Now that we have PBP data into the 1950s, has anyone looked to see if the chaining impact in the 50's and 60's was smaller (or I guess possibly larger, though that'd be extremely unintuitive).

I want to apply this to my system, but I want to make sure I get it right, and I have to be comfortable I (we?) come up with something that works across all eras. Should I make this a separate thread?
   412. Joey Numbaz (Scruff) Posted: December 16, 2008 at 08:56 PM (#3030867)
Another question for Access experts . . .

Does Access have something similar to the Excel "solver" function? My system requires the use of the Solver plug-in, to determine what the correct RA/9 value is for the given predicted winning %. Basically I figure out what the WPct should be, then covert it to RA/9 using Solver.

I suppose I could make a table by hand for every .001 WPct increment, with a corresponding RA/9 value. It isn't linear, and I use a dynamic exponent (PythaganPat), which is why solver is needed.

Any thoughts would be appreciated.
   413. Paul Wendt Posted: January 13, 2009 at 10:18 PM (#3050958)
Occasionally in this forum we have discussed pitchers in terms of DERA or another new statistic published in player 'DT cards' at baseballprospectus. Occasionally I have posted analysis that relies on both DERA and OPS+ at once, for instance.

Now I have posted a data table that covers 637 pitchers and includes DERA, IP, ERA+, and more.
pitchers, edition 637 now available

Visit that page which is not itself the file but an explanation.

Intensive Scope
* playerID, nameFirst, nameLast, debut,
* lahmanID, time, name,
* XIP, RA, DH, DR, DW, NRA, RAA, PRAA, PRAR, DERA,
* IP, ERA+, PA, OPS+
* RpS (Reliefs per Start)

Essentially this extends a table I have posted twice before with intensive scope marked bold, or the same plus RpS. Essentially the extension is to cover many more pitchers. I say that because the first four fields are a minor convenience and the four numerical fields {RA, NRA, PRAR, PA} are incomplete, present mainly for the pitchers I have added this winter.
   414. Howie Menckel Posted: January 17, 2009 at 05:38 PM (#3054304)
Are we going to vote on all-time pitchers next?
My preference was splitting them into 3 groups, which would give us similar amounts of players to weigh compared to what we had with the other positions.
I guess by era (as opposed to by ERA)?
I'd be amenable after that to picking our top 20 - it would get a lot easier once we've digested 3 manageable groups to further cull from there.

But I'll labor through all 60 or so at once if need be.....
   415. John (You Can Call Me Grandma) Murphy Posted: January 27, 2009 at 01:00 AM (#3061035)
Spoke to Joe. He thinks we should break the pitchers up like this:

Pitcher's Box (1871-1892)
Deadball (1893-1920)
Pre/Post WWII (1921-1961)
Expansion (1962-1992)
Steroid (1993-present)
Relievers (Wilhelm, Fingers, Goose, Eckersley)

I'm in inclined to agree, even though I had originally wanted to rank all of them together, which was foolhardy of me in retrospect.

These groups would make it easier for us talliers, too. :-)

What do you guys think about them?
   416. HGM Posted: January 27, 2009 at 01:04 AM (#3061041)
Sounds perfectly reasonable to me.
   417. HGM Posted: January 27, 2009 at 01:41 AM (#3061092)
One question, how would we break up some players? Is it by year of debut? Majority of career in the time frame? Take Pete Alexander, for example. Would he be in the deadball era, or the Pre/Post WWII era?
   418. Howie Menckel Posted: January 27, 2009 at 02:27 AM (#3061120)
How many in each group?

I hope we'll consider a 'consolidation' pitching ranking after these are over.
We'd each vote for 15 out of the 60+, and it would be a lot easier after we've ranked big groups of them (no need at the end to consider about half of them in the final poll).

I'm agnostic on including the relievers in their eras or not.
   419. KJOK Posted: January 27, 2009 at 02:55 AM (#3061135)
We have somewhere between 60 - 66 pitchers to rank I think. 3 groups of 20 - 22 would seem just about perfect.
   420. KJOK Posted: January 27, 2009 at 02:55 AM (#3061136)
EDIT: I think we should pull the relievers out and rank them as their own, '4th' group.
   421. Howie Menckel Posted: January 27, 2009 at 03:12 AM (#3061140)
If we want 4, just combine expansion/steroid and leave the relievers in with their peers...
   422. Joey Numbaz (Scruff) Posted: January 27, 2009 at 05:39 PM (#3061503)
Well there really isn't anyone from the steroid era yet, so not really a problem.

The main reason for breaking out the steroid era was the pitcher workloads dropping in the late 80s - you could probably start the steroid era with 1987 for pitcher ranking purposes.

I would say the era where the pitcher has the most value should determine what group he's in.

Alexander would be deadball, IMO. That's where his best years were. Whitey Ford would fall into the pre-post war group.

Any others? We should settle those issues before we start voting.
   423. Joey Numbaz (Scruff) Posted: January 27, 2009 at 05:45 PM (#3061511)
I would think once we have these groups ranked, we could then have one poll where everyone votes for their top 20 and use that as an overall tally.
   424. Chris Cobb Posted: January 27, 2009 at 07:06 PM (#3061614)
I would think once we have these groups ranked, we could then have one poll where everyone votes for their top 20 and use that as an overall tally.

If we want to have a complete ranking, we will need more than one round of voting to combine the lists. There will be a number of pitchers who won't make anybody's top 20, and the ranking of pitchers who appear on just a few ballots would not be reliable.

We have somewhere between 60 - 66 pitchers to rank I think. 3 groups of 20 - 22 would seem just about perfect.

This is my view, which I argued for earlier when the format for breaking up pitchers was discussed. The issue upon which a proposal for three groups of approximately equal size foundered was Joe's insistence that the pre-1893 pitchers be ranked separately from the others, because the position was so different when the pitching distance was different.

I argued to the contrary that a) there was more difference between Al Spalding and John Clarkson than between Clarkson and, say, Ed Walsh and b) it would be harder to rank Rube Foster against Rube Waddell than to rank Charley Radbourn against Rube Waddell. If we could rank NeL pitchers against ML pitchers within period, we ought also to be able to rank pre-1893 pitchers against post-1893 pitchers on a single ballot.

I don't think the discussion went farther from there. If Joe answered those arguments directly, I have unfortunately forgotten.

I think three rankings groups would be optimal, but I think we would do better at this point to start the balloting soon rather than spending a lot of time debating how to set up the ballots.
   425. Joey Numbaz (Scruff) Posted: January 27, 2009 at 09:05 PM (#3061809)
I don't really see the value in equal groups . . . I think breaking up pitchers by era is far more valuable. What does 20-22 person groups accomplish, other than setting arbitrary cutoffs? I don't think the cutoffs I proposed are really arbitrary - there's a reason for each one, as the game changed pretty dramatically each time.

Also, Spalding is the only 1870s pitcher we elected - I don't really see what harm there is in throwing him in with Clarkson, Galvin, et al. I think there is definitely a major difference in what the guys like Clarkson were asked to do and what the guys like Nichols and Young were asked to do.
   426. Joey Numbaz (Scruff) Posted: January 27, 2009 at 09:09 PM (#3061814)
The only change I would think we should make to what I proposed (through John) in #415 is:

Pitcher's Box (1871-1892)
Deadball (1893-1920)
Pre/Post WWII (1921-1961)
Expansion (1962-1986)
Steroid (1987-present)
Relievers (Wilhelm, Fingers, Goose, Eckersley)

Do you all think 1987 or 1993 is a better cutoff date? I think the SP innings, CGs, etc., all started dropping off greatly in the late 80s, so that seems a little more logical than cutting it off at 1993. But it's a pretty minor point, as either way, I imagine we'll throw Saberhagen in with expansion, so maybe for now we do this:

Pitcher's Box (1871-1892)
Deadball (1893-1920)
Pre/Post WWII (1921-1961)
Expansion (1962-late 80s)
Relievers (Wilhelm, Fingers, Goose, Eckersley)

And worry about the steroid era guys in 10-15 years once we've elected some of them.
   427. John (You Can Call Me Grandma) Murphy Posted: January 27, 2009 at 09:19 PM (#3061824)
I think there is definitely a major difference in what the guys like Clarkson were asked to do and what the guys like Nichols and Young were asked to do.


Of course, the tradeoff was a relative lack of durability from the post-1893 group compared to the pre-1893 group.

With that said, all of these cutoffs are arbitrary. The main thing is, if possible, to have approximately the same number of pitchers within each group (with the exception if we went with a Relievers group, of course) and that the hurlers have some connection to each other.

If we get it down to no more than 15 per group, I could use Evan Reich's ballot counter, too. :-D
   428. Joey Numbaz (Scruff) Posted: January 27, 2009 at 10:21 PM (#3061876)
Here's what I get for each group:

Pitcher's Box (6): Caruthers, Clarkson, Galvin, Keefe, Radbourn, Spalding.

Deadball (16): Alexander, M.Brown, Faber, R.Foster, Griffith, Johnson, Mathewson, McGinnity, Mendez, Nichols, Plank, Rusie, Waddell, Walsh, Williams, Young.

Pre/Post WWII (21): R.Brown, Coveleski, Dihigo, Feller, Ferrell, Ford, W. Foster, Grove, Hubbell, Lemon, Lyons, Newhouser, Paige, Pierce, Rixey, Roberts, Rogan, Ruffing, Spahn, Vance, Wynn.

Expansion (16): Blyleven, Bunning, Carlton, Drysdale, Gibson, Jenkins, Koufax, Marichal, Niekro, Palmer, Perry, Ryan, Saberhagen, Seaver, Stieb, Sutton.

Relievers (4): Wilhelm, Fingers, Goose, Eckersley

??
Coveleski could go war or deadball
Ford could go war or expansion
Rixey could go war or deadball

Others are pretty easy to slot, right?
   429. HGM Posted: January 27, 2009 at 11:24 PM (#3061924)
I think Red Faber belongs in the Pre/Post WWII era. Other than that, I think your groupings are solid.
   430. jimd Posted: January 27, 2009 at 11:38 PM (#3061934)
Pitchers in chronological order by birtdate (groups of 15):

Spalding Radbourn Galvin Keefe Ward?
Clarkson Caruthers Young Nichols Griffith
McGinnity Rusie Plank Waddell MBrown

RFoster Mathewson Walsh JWilliams Alexander
Mendez WJohnson Faber Covelski Rogan
Vance Rixey Grove Lyons Hubbell

Ruffing BFoster Dihigo Paige Ferrell
RBrown Feller Wynn Lemon Spahn (Wilhelm)
Newhouser Roberts Pierce Ford Bunning

Gibson Koufax Drysdale Marichal Perry
Niekro Jenkins Seaver Carlton Sutton (Fingers)
Palmer Ryan Blyleven (Gossage) (Eckersley) Stieb Saberhagen

Relievers in (). Was Ward ranked with the SS's? (I forget)
   431. DL from MN Posted: January 28, 2009 at 12:00 AM (#3061947)
I'd throw the relievers in their respective groups. Eckersley has a lot of starting value. Wilhelm was in a separate era from the other 3 relievers. Voting on a group of 4 seems really tiny.

6/16/22/19 looks like a good split.
   432. OCF Posted: January 28, 2009 at 01:18 AM (#3061987)
Ward was not ranked with the shortstops, so he has to go here. Similarly, Dihigo has not been ranked anywhere else yet.

Our general principle has been that for each candidate we determine a position and place the candidate in a group. But once the voting for that group begins, we vote for each player based on total value as a baseball player. That means that we must consider Ward's value as a SS/2B and Dihigo's value as an OF/etc. in our ranking here. That's going to make placing Ward and Dihigo particularly hard.

I'm with DL from MN. Placing Wilhelm in a group that includes Ford, Pierce, and Bunning, or placing Gossage, Fingers and Eckersley on a list with Stieb and Saberhagen is an easier problem than placing Spalding, Ward, or Dihigo on their respective lists. I don't want to consider the relievers separately. (And if that results in the relievers crowding the bottom of the list - so be it. We did elect them.)
   433. HGM Posted: January 28, 2009 at 01:28 AM (#3061989)
According to the shortstop ranking results (http://www.baseballthinkfactory.org/files/hall_of_merit/discussion/election_results_wagnerlloyd_ripken_vaughan_davis_dahlen_and_yount_get_at_l/), Ward was ranked with them.
   434. OCF Posted: January 28, 2009 at 01:43 AM (#3061993)
How about this split?

Spalding Radbourn Galving Keefe Ward
Clarkson Caruthers

Young Nichols Griffith
McGinnity Rusie Plank Waddell MBrown
RFoster Mathewson Walsh JWilliams Alexander
Mendez WJohnson Faber Coveleski Rogan
Rixey

Vance Grove Lyons Hubbell
Ruffing BFoster Dihigo Paige Ferrell
RBrown Feller

Wynn Lemon Spahn Wilhelm
Newhouser Roberts Pierce Ford Bunning
Koufax

Gibson Drysdale Marichal Perry
Niekro Jenkins Seaver Carlton Sutton Fingers
Palmer Ryan Blyleven Gossage Eckersley
Stieb Saberhagen

Cone, if elected, would start the next group: Cone KBrown Clemens Maddux, etc.

Of course, splitting between Feller and Wynn/Lemon/Newhouser/Roberts may not be right, and putting Koufax with Pierce/Ford/Bunning rather than with Gibson/Drysdale/Marichal may not be right.
   435. OCF Posted: January 28, 2009 at 01:45 AM (#3061995)
Oops - sorry, HGM, you're right, Ward was ranked with the SS. I scanned my tally and somehow missed his name. My bad. So Ward should not be listed with the pitchers. (But Spalding, Dihigo, and to a lesser extent, Caruthers, do present ranking challenges.)
   436. John (You Can Call Me Grandma) Murphy Posted: January 28, 2009 at 01:53 PM (#3062134)
I can go with almost any group, but how about this?

19th century: Spalding Radbourn Galvin Keefe
Clarkson Caruthers Young Nichols Griffith Rusie

Deadball: McGinnity Plank Waddell MBrown
RFoster Mathewson Walsh JWilliams Alexander
Mendez WJohnson

'20s/'30s: Faber, Covelski, Rogan
Vance Rixey Grove Lyons Hubbell
Ruffing BFoster Dihigo Paige Ferrell
RBrown

'40s/'50s/'60s: Feller Wynn Lemon Spahn (Wilhelm)
Newhouser Roberts Pierce Ford
Bunning, Gibson Koufax Drysdale Marichal

Expansion: Perry
Niekro Jenkins Seaver Carlton Sutton (Fingers)
Palmer Ryan Blyleven (Gossage) (Eckersley) Stieb Saberhagen

Splitting it up this way would make it much easier for me to tally, instead of those 16+ groups that would force me not to use Evan's ballot counter.
   437. DL from MN Posted: January 28, 2009 at 03:41 PM (#3062214)
We could follow up the era rankings with a "top 15 all-time" ranking. I would like to keep Young, Nichols, Griffith and Rusie in the "Deadball" group. You can do that without busting the ballot counter.
   438. John (You Can Call Me Grandma) Murphy Posted: January 28, 2009 at 03:47 PM (#3062221)
I would like to keep Young, Nichols, Griffith and Rusie in the "Deadball" group. You can do that without busting the ballot counter.


True, but their value was concentrated more during the 19th century. Why throw them in with the Deadball guys?
   439. OCF Posted: January 28, 2009 at 04:04 PM (#3062241)
Year of last season of at least 10 equivalent FWP in RA+ system:

Rusie 1898
Griffith 1901
Nichols 1904
McGinnity 1906
Waddell 1908
Young 1909
Walsh 1912
Mathewson 1913
Brown 1915
Plank 1916

I'm reasonably comfortable with putting Rusie back with Clarkson, Keefe, et al., in part because a large fraction of Rusie's value lies in the pitcher's box era, but (and it's because he kept going forever) it would feel odd to have Young in a different group than McGinnity and Waddell.
   440. DL from MN Posted: January 28, 2009 at 04:04 PM (#3062242)
Because they pitched after the rubber moved.
   441. DL from MN Posted: January 28, 2009 at 04:07 PM (#3062247)
Yeah, I'm ambivalent about Rusie and I lean toward the later group with Griffith but understand how someone could argue the other direction. I want Young v. Johnson on top of the ballot though.
   442. karlmagnus Posted: January 28, 2009 at 04:17 PM (#3062260)
I like Grandma's groupings. I agree that about equal groups makes it easier, and that 4 relievers as a separate group makes no sense (we'd probably be nearly unanimous on their ranking, but so what?) Adding 1890s pitchers in with their predecessors gives you one group with rapid evolution, followed by 3 fairly stable ones. The reality is that pitching in 1896 was a much harder job than in 1908, so a 1900 divider makes sense. Late 1870s also differs substantially from late 1880s, even after Spalding's gone. I shall try to participate in this phase, as my lack of PhD level sabermetrics is less of a problem, and my #1 favorite Caruthers is in this group (I regretted not voting for my #2 favorite Beckley, but it would only have moved him from bottom to penultimate.
   443. John (You Can Call Me Grandma) Murphy Posted: January 28, 2009 at 04:18 PM (#3062262)
Because they pitched after the rubber moved.


I figured that, DL, but pitching during the '90s was much different than during the next decade, too. Looking at the stats, the 1880's hurlers look more like the guys from the 1890s than the 1890s/Deadball guys do.

but (and it's because he kept going forever) it would feel odd to have Young in a different group than McGinnity and Waddell.


Still, his best years were during the first half of his career. I understand the feelings that you have, OCF, but there is a reason why he has a Cleveland Spiders cap instead of a Boston one. ;-)

With that said, if we can keep the groups big enough that they can fit on my ballot counter without blowing it up, I'll go with the consensus opinion.
   444. Paul Wendt Posted: January 28, 2009 at 04:35 PM (#3062279)
baseball history is too long and baseball careers are too short

For maximum-size 15, either {Young Nichols Griffith Rusie} or {Faber Coveleski, Rogan, Rixey} should be removed from the previously recommended "deadball" ballot (19), but it is not necessary to remove both groups as JTM has done.

(Joe, I see an argument for putting Coveleski early, Faber late, but no argument vice versa.)


>> (my emphases)
425. Joe Dimino Posted: January 27, 2009 at 03:05 PM (#3061809)
I don't really see the value in equal groups . . . I think breaking up pitchers by era is far more valuable. What does 20-22 person groups accomplish, other than setting arbitrary cutoffs? I don't think the cutoffs I proposed are really arbitrary - there's a reason for each one, as the game changed pretty dramatically each time.

Also, Spalding is the only 1870s pitcher we elected - I don't really see what harm there is in throwing him in with Clarkson, Galvin, et al. I think there is definitely a major difference in what the guys like Clarkson were asked to do and what the guys like Nichols and Young were asked to do.


[from Joe Dimino 426.]
Do you all think 1987 or 1993 is a better cutoff date? I think the SP innings, CGs, etc., all started dropping off greatly in the late 80s, so that seems a little more logical than cutting it off at 1993.

427. John (You Can Call Me Grandma) Murphy Posted: January 27, 2009 at 03:19 PM (#3061824)
Of course, the tradeoff was a relative lack of durability from the post-1893 group compared to the pre-1893 group.

With that said, all of these cutoffs are arbitrary.


If these pitcher ballots make round one, a staging ground for round two that will rank all pitchers (not to say all on one stage two ballot), then it doesn't matter much where the lines are drawn or even whether everyone shares.

At the other extreme, if we will write some prose account of periodization with rationale such as JoeD provides in #425-426 then we should deliberate.

It seems plausible to me that we will one day do something in between, devote to each of these elections a webpage that includes a serious preface (as well as links to the Discussion, Ballot, and Results pages, which might be the main purpose of a webpage for each election).

If rule changes will be primary, then 1893-1920. If the theme will be "what guys were asked to do" then that doesn't make sense. Then there is a good argument for putting Coveleski, Faber, Rixey -- and Shocker or Cooper or Quinn if elected -- and Redding if elected, and Rogan -- all in an early group that runs to about 1923. The workloads measured by starts or by innings did not plummet. Grimes or Luque if elected, rather than Coveleski, would be the modern pitcher with a lot of deadball characteristics in early big seasons.

1910-1939 pitcher counts

40 starts for one team
1 2 2 0 5 3 2 3 0 1 ; 24 teams 1914-15; very short and short 1918-19
1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 2 0

36 starts for one team
_7 _9 10 _9 17 10 _9 13 _1 _3 ; 24 teams 1914-15; very short and short 1918-19
13 12 _6 14 _4 _4 _2 _4 _3 _3
_3 _2 _3 _3 _3 _2 _5 _0 _5 _3


308 innings pitched for one team
6 8 6 5 14 8 8 7 3 0 ; 24 teams 1914-15; very short and short 1918-19
6 6 3 6 3 0 1 1 2 0
0 0 1 2 3 2 2 0 1 1

"308 innings" adjusted 13 8
6 8 6 5 9 5 8 7 13 8 ; AL and NL only 1914-15; 252 inns 1918, 280 inns 1919 (2ip per game)
6 6 3 6 3 0 1 1 2 0
0 0 1 2 3 2 2 0 1 1
   445. karlmagnus Posted: January 28, 2009 at 04:36 PM (#3062280)
Young was only 3 years younger than Caruthers. We think of him as an aughts pitcher because there are about 10 times as many books about the oughts Red Sox than about the 90s Spiders -- the 90s books are almost all about the Orioles. Anyway, I don't think Young/Johnson is difficult; it was pretty clearly Johnson, who I would rank #1 in the final pitcher list, whereas Young has difficulty fighting off Alex, Grove, Clemens and Maddux.
   446. Paul Wendt Posted: January 28, 2009 at 04:52 PM (#3062291)
277+ innings for one team
12 13 16 12 30 23 16 17 _6 _8 ; 24 teams 1914-15; very short and short 1918-19
17 14 10 14 _3 _3 _4 _6 _7 _3
_4 _4 _5 _7 _6 _5 _7 _2 _5 _4



20 complete games for one team
24 20 26 16 31 30 16 23 14 20 ; 24 teams 1914-15; very short and short 1918-19
27 18 18 17 12 11 _9 10 13 _8
12 10 10 10 _9 _9 12 _8 _7 _7

15 complete games for one team
41 39 42 37 57 57 31 36 28 31 ; 24 teams 1914-15; very short and short 1918-19
45 34 37 35 29 35 24 27 33 27
27 31 25 27 28 25 22 21 21 15

10 complete games for one team
70 63 61 55 94 90 54 62 60 56 ; 24 teams 1914-15; very short and short 1918-19
65 60 54 53 55 57 52 59 49 53
48 54 50 53 50 44 45 49 53 37

By the three complete games measures 1920 is an outlier with the early 1920s roughly similar to the mid-teens (pre-war). The clearest one-year declines in complete games occur 1924, 1926, 1928 for pitchers with 20, 15, and 10 complete games.
   447. John (You Can Call Me Grandma) Murphy Posted: January 28, 2009 at 04:53 PM (#3062292)
We think of him as an aughts pitcher because there are about 10 times as many books about the oughts Red Sox than about the 90s Spiders -- the 90s books are almost all about the Orioles.


Not to mention the fact that it was hard to relegate the all-time leader in wins with the "dinosaurs" of the 19th century.
   448. John (You Can Call Me Grandma) Murphy Posted: January 28, 2009 at 05:00 PM (#3062293)
For maximum-size 15, either {Young Nichols Griffith Rusie} or {Faber Coveleski, Rogan, Rixey} should be removed from the previously recommended "deadball" ballot (19), but it is not necessary to remove both groups as JTM has done.


Based on MiL credit, Coveleski may belong with the Deadballers. It's possible that Faber belongs with WWI credit, also. As for Rixey and Rogan, however, I can't see them as part of a Deadball group (cutoff year: 1919) since they didn't pitch that long to qualify for it.
   449. John (You Can Call Me Grandma) Murphy Posted: January 28, 2009 at 05:04 PM (#3062297)
I would think once we have these groups ranked, we could then have one poll where everyone votes for their top 20 and use that as an overall tally.

If we want to have a complete ranking, we will need more than one round of voting to combine the lists. There will be a number of pitchers who won't make anybody's top 20, and the ranking of pitchers who appear on just a few ballots would not be reliable.

If multiple rounds of voting would be needed, then we should still go that route. Would be fun, too.
   450. Paul Wendt Posted: January 28, 2009 at 05:04 PM (#3062298)
As readers must know but glancers may miss,
The focus of all tables #444, 446 should be on the second row of data representing the 1920s. Boldface is merely a device for warning that the counts for two seasons cover more teams. Better I would have multiplied those counts by 2/3 or counted AL and NL pitchers only.

By the way Vance, Lyons, Ruffing, Foster, Dihigo, and Paige accomplished nothing or next to nothing in the early 1920s.
   451. John (You Can Call Me Grandma) Murphy Posted: January 28, 2009 at 05:33 PM (#3062332)
If rule changes will be primary, then 1893-1920. If the theme will be "what guys were asked to do" then that doesn't make sense. Then there is a good argument for putting Coveleski, Faber, Rixey -- and Shocker or Cooper or Quinn if elected -- and Redding if elected, and Rogan -- all in an early group that runs to about 1923. The workloads measured by starts or by innings did not plummet. Grimes or Luque if elected, rather than Coveleski, would be the modern pitcher with a lot of deadball characteristics in early big seasons.


If the cutoff year was 1923 rather than 1919, then I would agree with you, Paul.
   452. Joey Numbaz (Scruff) Posted: January 28, 2009 at 06:25 PM (#3062404)
True, but their value was concentrated more during the 19th century. Why throw them in with the Deadball guys?


Because they moved the mound back in 1893, not in 1901.

I definitely goofed on Faber, he's definitely in with the 20s group - I must have just been thinking Black Sox.

I have no problem splitting out pre-War/post-War, though I don't really see the difference. Is it that hard to tally more than 20 just once? With plenty of notice could the tabulation worksheet be updated to accommodate more than 20? Then we start getting into things like which group Feller and Newhouser belongs in too, for one.

I figured that, DL, but pitching during the '90s was much different than during the next decade, too.


I don't necessarily agree. Bill James touches on this in his comment on Kid Nichols in the NHBA.

Young was only 3 years younger than Caruthers.


I don't really see how that is relevant - the birthday. Career dates should be all that matter in something like this.

36 starts for one team
_7 _9 10 _9 17 10 _9 13 _1 _3 ; 24 teams 1914-15; very short and short 1918-19
13 12 _6 14 _4 _4 _2 _4 _3 _3
_3 _2 _3 _3 _3 _2 _5 _0 _5 _3


That's a pretty good case for treating 1893-1921 as a group right? The dropoff from 14 -> is 1921 -> 1922.

I mean if we want even smaller groups, we could go with 1871-1892; 1893-1909; 1910-21; 1922-1942; 1946-1962; 1963-1986; 1987-present.

But do we really want to break it down that small? I still think 1871-1892; 1893-1921; 1922-1962 (or 1922-1942 and 1946-62); and 1963-86 makes the most logical/practical sense . . .

If we do move the cutoff from 1920 to 1921, it does make more sense to move Coveleski back with the deadball guys, moves his split from 5-5 to 6-4 in terms of good seasons, including his two best.
   453. Joey Numbaz (Scruff) Posted: January 28, 2009 at 06:27 PM (#3062406)
Oh, and moving it back to 1921, along with giving credit for 1918 military services makes a fair argument for dropping Faber back with the deadball guys as well.
   454. Joey Numbaz (Scruff) Posted: January 28, 2009 at 06:34 PM (#3062421)
Whoops, I counted wrong - the dropoff is 1923 -> 1924, not 1921 -> 1922.

I can definitely see making the one group as 1893-1923. I think Paul has pretty solid evidence as justification for the dropoff being that year, and the game didn't just go from deadball to liveball overnight; I mean in 1923 Harry Heilmann finished 3rd in the AL with 18 HR and Jack Fournier finished 2nd in the NL with 22.

This should move Coveleski, Faber and Rixey all to the deadball group.
   455. Chris Cobb Posted: January 28, 2009 at 07:15 PM (#3062490)
An additional argument for putting Coveleski and Faber in the deadball group is that they were grandfathered spitballers, so if one is looking at rule changes rather than usage patterns as the marker for the divide, they pitched under the old rules, even after the rules were changed.

As to where we draw the early line, what would Paul's data show about changes in usage patterns, 1880-1910?
   456. John (You Can Call Me Grandma) Murphy Posted: January 28, 2009 at 07:26 PM (#3062511)
Because they moved the mound back in 1893, not in 1901.


Again, I don't think 1893 was any momentous than, for example, the fooling around with the strike zone during the '60s, Joe.

I don't necessarily agree. Bill James touches on this in his comment on Kid Nichols in the NHBA.


His argument, IIRC, was that Nichols should be considered one of the all-time greats. I totally agree, except that I also think Clarkson, Radbourn, etc. shouldn't be left out of the conversation because "the game was different." Unlike James, who doesn't think too much about 19th century baseball anyway, I just can't agree with that.
   457. John (You Can Call Me Grandma) Murphy Posted: January 28, 2009 at 07:29 PM (#3062520)
With that said, Nichols can be placed in the Expansion group for all I care, as long as I don't have any tallying nightmares. :-)
   458. Joey Numbaz (Scruff) Posted: January 28, 2009 at 09:05 PM (#3062652)
Again, I don't think 1893 was any momentous than, for example, the fooling around with the strike zone during the '60s, Joe.


We have a cutoff in 1962 - could easily move it to 1963. That was a very big change. I called the era 'expansion' but the main reason for the divide was just how much pitching changed in the early 60s.

On the Nichols thing James was arguing that Nichols should be one of the greats, in part because he lasted into the early 1900s and was still pitching well despite being in his 30s. I thought he also talked about how the game changed after the mound moved back, but I could be mistaken.

Either way, I still think 1893 is much closer to 1923 than to 1883.
   459. Paul Wendt Posted: January 29, 2009 at 04:33 PM (#3063242)
454. Joe Dimino Posted: January 28, 2009 at 12:34 PM (#3062421)
Whoops, I counted wrong - the dropoff is 1923 -> 1924, not 1921 -> 1922.

I have used many of these tables. Always the purpose of layout in multiple rows of ten is to read the time scale easily but without the clutter of explicit dates. A row begins when the third digit of one of our annual dates turns over, a date that ends in zero; a row therefore represents one "decade" in the sense 1910s, 1920s, 1930s.


455. Chris Cobb Posted: January 28, 2009 at 01:15 PM (#3062490)
An additional argument for putting Coveleski and Faber in the deadball group is that they were grandfathered spitballers, so if one is looking at rule changes rather than usage patterns as the marker for the divide, they pitched under the old rules, even after the rules were changed.

I did think of that regarding Burleigh Grimes as I wrote,
"Grimes or Luque if elected, rather than Coveleski, would be the modern pitcher with a lot of deadball characteristics in early big seasons."

<i>As to where we draw the early line, what would Paul's data show about changes in usage patterns, 1880-1910?


Today and tomorrow, at least, I cannot afford to do more than provide the longer timespans for a few of the pitcher counts which I prepared and saved in a microsoft notepad.

underline represents seasons shorter than 154 games, by schedule or curtailment of the season, not cancellation of odd games.
Notes in the right margin concern the number of teams.

40 starts for one team (1893-2008)
begin 1893
- - - 10 9 8 11 5 8 4 : twelve teams 1893-99
4;1 4 4 10 4 3 3 3 1 : eight teams 1900, then 16 except 1914-15
1 2 2 0;5 3;2 3 0 1 : 24 teams 1914-15, then 16 again
1 1 1 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 2 0
0 1 0 0 2 0 1 0 0 0
0 0 0 1 0 0 1 0 0 0
0;0
- -;2 4 1 5 4 1 2;7 : 1960s expansions to 18, 20, 24 teams
6 2 6 12 9 3 2;2 3 1 : 1977 expansion to 26
0 0 1 0 0 0 0 1 0 0
0 0 0;0 0 0 0;0 0 0 : 1990s expansions to 28 and 30
0 0 0 0 0 0 0 0 0


no time for any more markup or other editing

36 starts for one team (1893-2008)
1893-97: 11 to 16 pitchers on 12 teams, 132-game schedule
1898-99: 19 and 14 pitchers on 12 teams, 154-game schedule
1900 and after: . . . 8 teams 1900 (bold); 140 games 1900-03 (underline)
7 12 8 10 21 16 12 10 9 2
7 9 10 9 17 10 9 13 1 3 (very short and short 1918-19)
13 12 6 14 4 4 2 4 3 3

3 2 3 3 3 2 5 0 5 3
5 3 1 1 4 1 1 3 4 5
5 4 3 3 4 1 6 3 4 8
7 7
- - 12 12 13 20 14 16 13 23
23 28 19 31 33 20 16 17 16 14
14 -- 12 9 5 12 10 11 4 5 (none in short 1981, 1994, 1995)
2 4 2 5 -- -- 2 0 1 0
0 0 1 2 0 0 0 0 0 (Glavine 02, Maddux 03, Halladay 03)



308 innings for one team (1910-1979)
begin 1910
6 8 6 5 14 8 8 7 3 0
6 6 3 6 3 0 1 1 2 0
0 0 1 2 3 2 2 0 1 1
1 1 0 0 3 1 1 0 1 0
1 2 2 1 1 0 1 0 0 0
0 0
- - 1 3 1 2 2 0 3 6 : 1962-79, 308 innings
2 4 4 7 7 3 2 2 1 1

- - 0 0 0 1 0 0 2 2 : 1962-79, 324 innings (two per game scheduled)
1 3 4 4 2 1 0 1 1 1
   460. jimd Posted: January 30, 2009 at 12:21 AM (#3063709)
There is no mound to move in 1892/93. In 1892 there was a pitching box, the back of which was 55'6" from the plate, and the pitcher had to keep his foot on that back line while delivering the ball. In 1893, there was a pitching rubber, 60'6" from the plate, and the pitcher had to keep his foot on that rubber while delivering the ball. Distance moved back 5 feet, rubber replaced chalk line (making it easier for a 2-umpire team to detect irregularities).

The first mounds were apparently introduced in the late 1890's. Combined with the introduction of the foul strike rule, there's an argument for 1901 being a significant break in pitching conditions.

(Thanks to cblau, who corrected me on this issue back in the 1942 discussion thread.)
   461. jimd Posted: January 30, 2009 at 12:23 AM (#3063712)
Rereading, it strikes me that I'm not sure how the pitching box was delineated in 1892 (grass/dirt, chalk, ??).
   462. Howie Menckel Posted: January 30, 2009 at 02:42 AM (#3063799)
"I would think once we have these groups ranked, we could then have one poll where everyone votes for their top 20 and use that as an overall tally."

I like this. I'm not THAT worried about whether one pitcher should be ranked 42nd or 44th, anymore than I am disappointed about not ranking all 60 or so OFs at once. We're looking for the best 20 or so all-time at each position, in my mind, and we rank them in order at all previous positions and we'll to some extent do it with pitchers, too (granting that the last few in the top 20 aren't much more popular than the next several. But then again, the last couple of infielders at each slot sometimes aren't much more popular than the guys we haven't quite elected, either).

I'll be fine with whatever groupings we pick. I'm now more sold on the "add the relievers, too" side, though. Wilhelm in particular gets a chance there, perhaps.
   463. Joey Numbaz (Scruff) Posted: January 30, 2009 at 09:08 AM (#3063968)
You know, on the relievers, we could do both. Rank them within, and then separate. Wouldn't really hurt anything, win/win, right?

jimd - interesting, I hadn't heard that before. Why did offense shoot up in 1893 then? R/G went up 29%, from 5.10 to 6.57. Something changed, and the new pitching rules make the most sense, right?

I would think moving the distance back 5 feet (9%) has to be a big change, much bigger than pitching off a mound from the same distance, right?
   464. Paul Wendt Posted: January 30, 2009 at 06:30 PM (#3064343)
459.
Distance moved back 5 feet, rubber replaced chalk line (making it easier for a 2-umpire team to detect irregularities).

one umpire commonly behind the plate, in some seasons after 1892/93. (Maybe 1893 was one of the two-umpire seasons, but that will be a surprise to me, given the time of retrenchment.)

The first mounds were apparently introduced in the late 1890's. Combined with the introduction of the foul strike rule, there's an argument for 1901 being a significant break in pitching conditions.

1.
The NL introduced foul strikes in 1901 but how many minors and colleges went along? Almost none, I have inferred. (Adding the foul strike and dropping the hit batsman helped foment rebellion against the NL that winter and spring. They restored the hit batsman by telegraph vote on opening eve.)

So 1903 for the foul strike. Along with the "modern World Series" and the AL replacement of Baltimore by New York that clinches it ;-)

2.
Gradual introduction of mounds in the late nineties doesn't help support a break in 1901.

461. jimd Posted: January 29, 2009 at 06:23 PM (#3063712)
<i>Rereading, it strikes me that I'm not sure how the pitching box was delineated in 1892 (grass/dirt, chalk, ??).


don't know. Some of the time they used lines. (lime?)


463. Joe Dimino Posted: January 30, 2009 at 03:08 AM (#3063968)
jimd - interesting, I hadn't heard that before. Why did offense shoot up in 1893 then? R/G went up 29%, from 5.10 to 6.57. Something changed, and the new pitching rules make the most sense, right?

What is the question? A mound helps the pitcher. Gradual introduction of pitching mounds must be one cause of a decline of batting between the 1893 and 1900/1902 seasons.

A pitcher's rubber or metal or anything solid in contrast to a line on the ground also helps the pitcher, at least in practice where pitchers wear spiked shoes --not to mention practice where placement "flush to the ground" permits digging a trench in front of it.

Of course it takes some time for pitchers to adapt. Big league pitchers devote a couple weeks or months to preparation for the season but that cannot be sufficient.

I would think moving the distance back 5 feet (9%) has to be a big change, much bigger than pitching off a mound from the same distance, right?

Do you mean a change in the balance of power between pitcher and batter? Or a rule change or grounds change that directly causes change in pitching mechanics, with secondary effects on bodily strain and injury, and secondary effects on the balance of power?

Foul strike rules help a pitcher who cannot field bunts but simultaneously help a pitcher who can field bunts unusually well. Foul strike rule spares everyone some pitches and thereby some strain.

Rules changes governing the pitching motions demand greater adjustment (from some) and they cause many strains, some injuries, some simple loss of effectiveness. Consider regulations of shot, discus, and hammer field events, or butterfly and breast swimming races, or carrying the basketball (traveling, palming). Tweaking those regulations, either written rules or practical interpretations, may take someone's "game" away. I'm sure that that happened to some pitchers in the 19th century and I suppose that it happened to some in 1920 and 1921.
(This theme is practically beside the point here because changes were too scattered in time.)

The rule permitting tactical substitution, as well as injury substitution or courtesy substitution, makes a big difference in the pitchers' responsibilities (and the captains and managers, of course) at the game and roadtrip levels. In history baseball took a long time to increase the pitching staffs and especially the numbers of pitchers commonly used in one game. Now that teams carry twelve pitchers and everyone has adopted relief pitching, some further relaxations of substitution rules would have sudden impact, as would most restrictions of course.
   465. Paul Wendt Posted: January 30, 2009 at 07:45 PM (#3064454)
413. Paul Wendt Posted: January 13, 2009 at 04:18 PM (#3050958)
Occasionally in this forum we have discussed pitchers in terms of DERA or another new statistic published in player 'DT cards' at baseballprospectus. Occasionally I have posted analysis that relies on both DERA and OPS+ at once, for instance.

Now I have posted a data table that covers 637 pitchers and includes DERA, IP, ERA+, and more.

pitchers, edition 637 now available

Follow that link for explanation of the table and a link to the table.

I have revised it without changing the filename, 637pitchers.txt

Now it provides career PA for all 637 pitchers, PA=0 if no plate appearances, and career OPS+ for everyone with PA>0.

At one time I had entered OPS+ = -50 or -30 or 0 for many pitchers with small numbers of PA. A month or three ago I forgot about that and entered another batch or two of pitchers with their true OPS+, and only recently noticed the mixture. Oops.

(Now I have checked all of the {-50, -30, and 0} values for OPS+ and confirmed or corrected them.)
   466. Joey Numbaz (Scruff) Posted: January 30, 2009 at 09:12 PM (#3064541)
Given everything that's been said Paul - where would you draw the line?

I still think the biggest fundamental shift is 1892 -> 1893. But even if it's a gradual change over 1892 -> 1901, wouldn't it make sense to keep the 1892-1901 pitchers in with the latter group (1902-1923)?

It eliminates the tough call on Cy Young for one. I still think of Rusie, Nichols and Young in with the guys like Mathewson than Radbourn, for example.

In 1892 NL, 5 guys started 51 games or more, with the leader at 70. In 1893 the leader had 52, then 48, 46, 45, 44, 44. That seems like a drastic dropoff from one year.

Is there a strong consensus one way or the other that I'm not picking up?
   467. Joey Numbaz (Scruff) Posted: January 30, 2009 at 09:52 PM (#3064583)
Whoops. The schedule dropped from 154 to 131 games.

By 1897, the leaders were at 42, 42, 41, 40, 40, schedule 132 games.

Top 5 in innings 1893 - 383-482 IP. By 1897, 337-368. In 1901, with a 140 game schedule, we're at 336-375 IP.

Or looking at an average of IP by the league leaders, NL (adjusted for # of teams in a league), 1871-2007, maybe this will give us an obvious place to draw the lines?

NLa is my methodology, that averages the x-y ranking of IP leaders, where in an 8 team league x = 3 and y = 6, in a 12 team league x = 4 and y = 9, in a 16 team league, x = 5 and y = 12.

IP/tG adjusts for schedule.

NL      NLa      IP/tG
1871    238.6    
1872    376.1    
1873    429.0    
1874    495.4    
1875    509.7    
1876    433.7    6.67
1877    443.5    7.39
1878    425.8    6.94
1879    550.5    6.86
1880    554.6    6.52
1881    458.3    5.46
1882    421.8    5.16
1883    488.0    4.97
1884    519.4    4.65
1885    440.3    3.96
1886    470.3    3.58
1887    415.2    3.14
1888    425.3    3.12
1889    374.4    2.75
1890    451.4    3.35
1891    437.5    3.35
1892    440.2    2.87
1893    379.3    2.90
1894    379.3    2.87
1895    355.4    2.68
1896    359.3    2.72
1897    332.2    2.46
1898    355.8    2.32
1899    350.6    2.28
1900    321.6    2.26
1901    336.6    2.43
1902    312.3    2.24
1903    322.1    2.31
1904    351.2    2.25
1905    332.4    2.15
1906    312.8    2.04
1907    307.0    1.99
1908    329.8    2.12
1909    296.8    1.91
1910    298.3    1.91
1911    312.2    2.02
1912    301.4    1.96
1913    301.8    1.96
1914    322.5    2.05
1915    293.8    1.89
1916    304.0    1.95
1917    307.3    1.97
1918    266.0    2.09
1919    274.8    1.97
1920    301.4    1.95
1921    292.0    1.90
1922    271.3    1.75
1923    307.3    1.99
1924    260.0    1.69
1925    269.3    1.75
1926    268.0    1.74
1927    280.6    1.82
1928    270.8    1.76
1929    269.0    1.75
1930    256.7    1.66
1931    254.3    1.65
1932    273.3    1.77
1933    284.5    1.86
1934    293.5    1.92
1935    267.8    1.74
1936    278.8    1.80
1937    260.5    1.68
1938    264.0    1.73
1939    268.5    1.74
1940    272.5    1.76
1941    255.3    1.64
1942    267.5    1.75
1943    270.8    1.75
1944    271.8    1.75
1945    237.3    1.54
1946    235.5    1.52
1947    257.0    1.65
1948    249.5    1.61
1949    249.3    1.61
1950    278.3    1.80
1951    284.5    1.84
1952    253.8    1.64
1953    242.0    1.56
1954    258.0    1.67
1955    239.3    1.55
1956    267.8    1.73
1957    249.3    1.61
1958    261.0    1.69
1959    272.5    1.76
1960    271.3    1.76
1961    254.5    1.60
1962    270.6    1.67
1963    280.9    1.74
1964    271.1    1.67
1965    291.8    1.80
1966    276.5    1.71
1967    272.8    1.68
1968    276.9    1.70
1969    303.9    1.87
1970    276.5    1.71
1971    275.0    1.70
1972    264.1    1.70
1973    266.4    1.65
1974    271.2    1.67
1975    262.1    1.63
1976    261.0    1.62
1977    252.1    1.56
1978    256.4    1.59
1979    249.2    1.54
1980    241.7    1.49
1981    160.9    1.50
1982    255.8    1.58
1983    248.0    1.53
1984    232.1    1.43
1985    257.2    1.59
1986    246.4    1.52
1987    232.6    1.44
1988    244.6    1.51
1989    237.4    1.47
1990    229.6    1.42
1991    231.8    1.43
1992    232.4    1.43
1993    232.2    1.43
1994    166.1    1.45
1995    197.3    1.37
1996    230.5    1.42
1997    230.9    1.43
1998    229.9    1.42
1999    223.6    1.38
2000    226.0    1.40
2001    223.4    1.38
2002    218.1    1.35
2003    214.6    1.32
2004    216.0    1.33
2005    224.8    1.39
2006    219.3    1.35
2007    210.5    1.30 


I graphed these, I wish I could show it here. You should be able to cut and paste those numbers into Excel and make your own graphs, or drop me a note and I'll email it to you.

Using the IP/tG on a logarithmic chart (base 1) shows the differences better as 1871-95 kind of drowns out the other 112 years otherwise.

Anyone see obvious cutoff points? I'll refrain from making any observations for now.

What I'm going for with determining the group cutoffs are the best combination of (in no particular order):

1) Rules changes
2) pitcher usage trends
3) obvious cutoff points that make things neat and simple (WWII, leagues being added dropped, expansions, huge fluctuations in offensive levels, etc.)
   468. Joey Numbaz (Scruff) Posted: February 04, 2009 at 09:24 PM (#3067692)
Nothing on this? Long post lost everyone?

You all really want me to decide on this groupings entirely on my own? Trying to push for consensus . . .

:-)

Either way, we should have this settled and start discussion for the first group by the beginning of next week.
   469. DL from MN Posted: February 04, 2009 at 09:55 PM (#3067720)
There's a drop from 44 to 45. The 1921 cutoff could be 1923 at the latest.

1871-1892; 1893-1921; 1922-1945; 1946-present
   470. Joey Numbaz (Scruff) Posted: February 04, 2009 at 10:03 PM (#3067736)
I agree that 1923-44 is a good grouping - assuming we want to break it down that small.

1871-1892
1893-1923
1924-1944
1945-1962
1963-1986
1987-present

Does that work?

While I wanted to do relievers separately, except for Eck, they all fit in the 1963-86 group at this time. Maybe once we elect Rivera and whoever else from this era, we could combine them all in a ranking.
   471. DL from MN Posted: February 04, 2009 at 10:59 PM (#3067790)
Just throw 87-present with the previous group for now.
   472. Joey Numbaz (Scruff) Posted: February 04, 2009 at 11:08 PM (#3067800)
Yeah, agreed. Only guys that really affects are Saberhagen and Eckersley anyway. Although those two head to head is an interesting question in and of itself.
   473. Chris Cobb Posted: February 05, 2009 at 03:46 AM (#3067958)
By Joe's suggestion, these are the actual groupings I get

1871-1892: (6) Spalding Radbourn Galvin Keefe Clarkson Caruthers
1893-1923: (18) Young Nichols Griffith Rusie McGinnity Plank Waddell MBrown RFoster Mathewson Walsh JWilliams Alexander Mendez WJohnson Faber Coveleski Rixey
1924-1944: (12) Rogan Vance Rixey Grove Lyons Hubbell Ruffing BFoster Dihigo Paige Ferrell RBrown
1945-1962: (11) Feller Wynn Lemon Spahn (Wilhelm) Newhouser Roberts Pierce Ford Bunning Koufax Drysdale
1963-1986 (plus post-1986 for now): (16) Gibson Marichal Perry Niekro Jenkins Seaver Carlton Sutton (Fingers) Palmer Ryan Blyleven (Gossage) (Eckersley) Stieb Saberhagen

This looks workable. For greater evenness of numbers in each group and of years included in each grouping, we could tweak it to

1871-92: no change (6)
1893-1919: (15) Move Faber, Rixey, Coveleski to next group.
1920-1944: (15) Add Faber, Rixey, Coveleski
1945-1968: (13) Add Gibson and Marichal from next group.
1969-present (eventually with break c. 1990): (14) Move Gibson and Marichal to preceding group.

I think Faber, Rixey, and Coveleski can legitimately be in either group: they peaked on the cusp of the transition between dead ball and lively ball, so we might as well move them for evenness.

The rule-change division between 1968 and 1969 does not show up in pitcher usage in any significant way, but the break at 1969 seems to me to capture a meaningful generational transition: Marichal certainly did his best pitching when Koufax, Drysdale, and Bunning were doing theirs, and Gibson's peak was ending just as the peaks of Seaver, Jenkins, Carlton, et al, were getting started.

Just a couple of suggestions.
   474. HGM Posted: February 05, 2009 at 04:31 AM (#3067979)
I like Chris Cobb's second list.
   475. Joey Numbaz (Scruff) Posted: February 05, 2009 at 07:06 PM (#3068424)
i would put Bunning/Koufax/Drysdale in the 1963-86 group. That's when they had their best years. I also think that's definitely where Gibson and Marichal fit.

Feller is a tough one, I could legitimately see him in either group (1923-44, 1945-62).

I mean Koufax was "KoufaxTM" from 1962-65.
   476. Joey Numbaz (Scruff) Posted: February 05, 2009 at 07:07 PM (#3068425)
I think pitching from 1963-68 resembled 1969-76 a lot more than 1945-62, right?
   477. Chris Cobb Posted: February 05, 2009 at 09:54 PM (#3068667)
i would put Bunning/Koufax/Drysdale in the 1963-86 group. That's when they had their best years. I also think that's definitely where Gibson and Marichal fit.

Using BP's WARP1, this is not true of Drysdale. He definitely has more value pre-1963 than 1963 and after (143 PRAA/544 PRAR 1956-62 and 76 PRAA/477 PRAR 1963-69), and more of his best seasons were pre-1963 than 1963 and after.

Here are his seasons ordered by PRAR, with DERA for another reference point:
1964 108 (3.39)
1962 99 (3.75)
1960 98 (3.51)
1959 93 (3.77)
1957 87 (3.26)

1963 87 (4.11)
1967 82 (3.82)
1961 75 (3.66)
1968 72 (3.64)
1965 68 (4.61)
1958 60 (4.20)
1966 57 (4.54)
1956 32 (3.65)
1969 3 (5.97)

So pre-1963 gets 4 of the top 5 seasons by PRAR, and 5 of the 8 full seasons with DERA below 4.00.

Drysdale did throw slightly more IP 1963- on, 1802.3 vs. 1629.7, but looked at by overall value there's definitely more 1956-62. If one normalizes IP for era, that might close the gap. here, also.

Throw in batting and fielding and the result is the same: 53.7 WARP1 vs. 44.4 WARP1.

Bunning is split almost exactly down the middle, 48.2 WARP1 1955-62, 49.8 WARP1 1963-71. His biggest years clearly are 1965-67, but he was nearly worthless after that, and his next five best seasons are all before 1963. I don't think it would misrepresent his career significantly to put him in the earlier group.

I think pitching from 1963-68 resembled 1969-76 a lot more than 1945-62, right?

It did, but it may be important that folks like Bunning, Koufax, Drysdale, Marichal, and even Gibson started in the pre-1963 environment, then moved into the 1963-76 environment, whereas the next group started in the 1963-76 environment and moved into the 1977-92 environment.

It's no accident, I think, that the former group all had their biggest years 1963 and later, because they were positioned to take advantage of the new conditions, but all except Gibson had their arms fall off pretty early. Nobody from that generation even made it to 4000 IP, unlike Spahn and Wynn and Roberts among immediate post-war pitchers and unlike a whole raft of pitchers in the next generation, many of whom made it to 5000 IP: Ryan, Blyleven, Seaver, Sutton, Carlton, Niekro, Perry, Jenkins, looking at HoMers only.

I'd suggest that the sequence of conditions a pitcher moves through in the course of his career might be as important for determining the group in which it is appropriate to place him as is the particular playing condition in which he had his peak.
   478. Joey Numbaz (Scruff) Posted: February 06, 2009 at 12:20 AM (#3068855)
Very interesting Chris . . . my system with Drysdale:

WAR
1956-62: 2.2, 6.8, 3.6, 6.5, 7.6, 5.7, 7.0 (39.4)
1963-68: 5.4, 9.0, 5.4, 3.5, 6.0, 5.3 (34.5)

tIP/DRA
1956-62: 1607.7/3.45
1963-68: 1615.7/3.74

1962 is the year throws him over the top to either group for sure. It is much closer than I realized, and surprised it leans towards the earlier group, but it clearly does.

This caused me to go back and look at the year by year data I posted . . .

Does anyone else think we could/should make the cutoff 1958-59 instead of 1962-63?
   479. DL from MN Posted: February 06, 2009 at 12:27 AM (#3068860)
I think we're splitting hairs at this point. Make it easy for the ballot counters and go with Chris' group 2.
   480. OCF Posted: February 06, 2009 at 02:04 AM (#3068947)
There are problem cases with any cutoff. I'll side with what DL from MN just said: Chris's second groupings.
   481. OCF Posted: February 06, 2009 at 05:05 AM (#3069065)
A search through my RA+ Pythpat equivalent record files, looking for the top individual seasons in each grouping, ordered by equivalent FWP:

I.

I never worked up this group, except to extend Young, Nichols, and Rusie back to the beginnings of their careers. Young 1892 and Nichols 1890 both show up as big years, but I have no idea what the monsters of the 80's look like, although Radbourn 1884 is an obvious candidate year.

II.

Johnson 1913: 32-6 [53]
Rusie 1894: 36-14 [48]
Johnson 1912: 32-9 [47]
McGinnity 1904: 33-12 [45]
Chesbro 1904: 35-16 [42]

There's an obvious skew towards the earlier parts of this period; I don't think Johnson could have done that in 1918 or so even if he were still at his absolute peak.

III.

Luque 1923: 27-9 [38]
Grove 1931: 25-7 [37]
Hubbell 1937: 26-8 [37]
Hubbell 1933: 26-9 [36]
Grove 1932: 24-8 [34]

IV.

Gibson 1968: 27-7 [41]
Koufax 1966: 26-10 [36]
Chance 1964: 23-7 [34]
Koufax 1963: 25-10 [33]
Newhouser 1946: 24-9 [33]

The problem, of course, is that both Gibson and Newhouser are sitting on the doorways of the period, looking to the next period over. The best post-Newhouser, pre-expansion is Roberts 1953: 26-12 [32].

V.

Gooden 1985: 25-6 [39]
Carlton 1972: 28-10 [38]
Perry 1972: 27-11 [37]
Guidry 1978: 24-7 [35]
Seaver 1971: 24-8 [34]

VI.

Clemens 1997: 24-6 [37]
Martinez 2000: 21-3 [37]
Maddux 1995: 20-3 [34]
Martinez 1998: 21-5 [32]
Martinez 1999: 19-4 [31]

There are probably individual years that I've missed completely - it's not a thorough search.
   482. Howie Menckel Posted: February 06, 2009 at 02:43 PM (#3069218)
I'll accept that list, but seems a little weird that we've been voting on about 20 per position, then when we divide up pitchers, we divide up on significantly smaller amounts. Something to be said for cutting this into thirds as we did with a similar number of OFs (granting that they played different OF slots).
   483. Joey Numbaz (Scruff) Posted: February 06, 2009 at 03:48 PM (#3069264)
I'm really not a fan of Chris's group 2.

Anything less than 20 makes it easy for the ballot counters.

I just don't see any point to a 1968-69 cutoff at all.

Remember what we were shooting for?

1) Rules changes
2) pitcher usage trends
3) obvious cutoff points that make things neat and simple (WWII, leagues being added dropped, expansions, huge fluctuations in offensive levels, etc.)


As long as we have each group at 20 and under, it's easy for the ballot counters. I would like for the groups to be logical when we have to explain why we cut them off where we did.
   484. Joey Numbaz (Scruff) Posted: February 06, 2009 at 04:03 PM (#3069283)
Regarding Howie's post in 482, what about this as an option?

1871-1892: (6) Spalding Radbourn Galvin Keefe Clarkson Caruthers

1893-1923: (18) Young Nichols Griffith Rusie McGinnity Plank Waddell MBrown RFoster Mathewson Walsh JWilliams Alexander Mendez WJohnson Faber Coveleski Rixey

1924-1958: (21) Rogan Vance Rixey Grove Lyons Hubbell Ruffing BFoster Dihigo Paige Ferrell RBrown Feller Wynn Lemon Spahn Wilhelm Newhouser Roberts Pierce Ford

1959-1986 (plus post-1986 for now): (19) Bunning Koufax Drysdale Gibson Marichal Perry Niekro Jenkins Seaver Carlton Sutton (Fingers) Palmer Ryan Blyleven (Gossage) (Eckersley) Stieb Saberhagen

Whitey Ford is a tough call . . . I'd rather have him with the 1924-58 group, I think he fits better there, but it's super close, with credit for military service:

1950-58: 37.4 WAR, 1797 tIP, 120 DRA+
1959-67: 35.7 WAR, 1880 tIP, 116 DRA+

Individual years:

1950-58: 6.2, 6.0, 5.8, 4.1, 3.7, 3.3, 3.3, 2.8, 2.5
1959-67: 6.7, 5.4, 5.3, 4.9, 4.5, 4.0, 3.2, 1.4, 0.2

I would think he fits better in the 1924-58 period, but if 21 is a huge deal for the ballot counts, we can justify him in the later group.

I would volunteer to do the manual formatting for the 21 person group if necessary.
   485. Joey Numbaz (Scruff) Posted: February 06, 2009 at 04:20 PM (#3069307)
Wait, Ford can go in 1924-58, because Rixey was double counted (I thought there were only 63!) and should be in the 1893-1923 group, especially once he's given credit for 1918.

1912-23: 40.6 WAR, 2592 tIP, 110 DRA+
1924-33: 34.0 WAR, 1932 tIP, 113 DRA+

Hoyt Wilhelm on the other hand, would also get moved to the 1959+ group, I missed him earlier.

That would make the groups:

1871-1892: (6) Spalding Radbourn Galvin Keefe Clarkson Caruthers

1893-1923: (18) Young Nichols Griffith Rusie McGinnity Plank Waddell MBrown RFoster Mathewson Walsh JWilliams Alexander Mendez WJohnson Faber Coveleski Rixey

1924-1958: (19) Rogan Vance Grove Lyons Hubbell Ruffing BFoster Dihigo Paige Ferrell RBrown Feller Wynn Lemon Spahn Newhouser Roberts Pierce Ford

1959-1986 (plus post-1986 for now): (20) (Wilhelm) Bunning Koufax Drysdale Gibson Marichal Perry Niekro Jenkins Seaver Carlton Sutton (Fingers) Palmer Ryan Blyleven (Gossage) (Eckersley) Stieb Saberhagen

Does this work?
   486. DL from MN Posted: February 06, 2009 at 04:49 PM (#3069341)
That works. I'd propose allowing the most time to rank the 93-23 and 24-58 groups. The Negro Leaguers, Dihigo especially, make things really difficult. The 59-86 group is actually pretty straightforward. Maybe we rank that one first?
   487. DL from MN Posted: February 06, 2009 at 04:52 PM (#3069342)
A quick look at the 59-86 group for me is pretty surprising.

1) Seaver
2) Blyleven (!)
   488. Chris Cobb Posted: February 06, 2009 at 05:25 PM (#3069381)
I think going with four groups rather than five groups works well.

The break points all correspond to the most significant (post-1892) changes in offensive context and pitcher usage, and it gives us groups that (except for pre-1893) are similar in size to the position players groups. Feller is no longer a problem case, and the early 60s generation is together, which strikes me as proper.

Ford is a borderline case, but it looks to me that he ought to be in the pre-1959 group because a) with war credit, he would have more value 1950-58 than afterwards, and b) he was a more effective pitcher in this part of his career.

1950-58 1357.7 IP, 3.81 DERA
1959-67 1812.7 IP, 3.94 DERA

His IP in the first half of his career is lower partly because of the missed seasons and partly because of the way Casey Stengel used his pitching staff. That also means that his "biggest" seasons came later in his career under Ralph Houk and Yogi Berra in 1961-64, but they were bigger because he was used in a four-man rotation, not because he was a better pitcher than he was in the 1950s. The quality was about the same.
   489. Joey Numbaz (Scruff) Posted: February 06, 2009 at 05:49 PM (#3069409)
My quick ranking of 1959-86, just the numbers, no interpretation has:

RK  Pitcher       PA
x1
Seaver      1.722
x2
Carlton     1.509
x3
Niekro      1.462
x4
Blyleven    1.379
x5
Perry       1.346
x6
Ryan        1.331
x7
Gibson      1.314
x8
Sutton      1.183
x9
Jenkins     1.160
10. Eckersley   1.146
11. Drysdale    1.136
12. Reuschel    1.031
12. Bunning     1.026
13. Koufax      1.018
14. Marichal    1.006
15. Palmer      1.003
16. Saberhagen   .994
17. John         .979
18. Wilhelm      .979
19. Gossage      .965
20. Hershiser    .932
21. Appier       .930
22. Stieb        .891
23. Gooden       .879
24. Martinez     .873
25. Tiant        .863
26. Tanana       .858
27. Fingers      .851 


Using the more peak friendly Bill James type system (with my WAR as the inputs) I get:

RK  Pitcher      Score
x1
Koufax       32.08
x2
Gibson       29.47
x3
Seaver       29.22
x4
Carlton      28.40
x5
Marichal     27.23
x6
Niekro       26.51
x7
Drysdale     26.09
x8
Gossage      25.87
x9
Saberhagen   25.73
10. Bunning      25.37
11. Perry        25.35
12. Jenkins      24.96
13. Hershiser    24.93
14. Appier       24.80
15. Stieb        24.31
15. Gooden       24.12
16. Blyleven     24.07
17. Viola        23.50
18. Palmer       23.28
19. Eckersley    23.12
20. Fingers      23.08
21. Ryan         22.98
22. Reuschel     22.89
23. Guidry       22.53
24. Wilhelm      22.16
25. Messersmith  21.82
26. Rogers       21.81
27. Tanana       21.72
28. Valenzuela   21.62
29. W
.Wood       21.61
30. Sutton       21.50
31. Blue         21.48 


These do not account for the recent change in the BPro player cards regarding DERA.
   490. Joey Numbaz (Scruff) Posted: February 06, 2009 at 05:51 PM (#3069418)
DL, I'd rather go chronologically, but we could allow an extra week of discussion . . . would that be OK?
   491. Paul Wendt Posted: February 06, 2009 at 06:00 PM (#3069434)
What is the point at 1958?
Elroy Face? --no, that's facetious

If it is to put Wilhelm, Koufax, Drysdale, and Bunning in the later group it doesn't need to be so early as 1958. Call it 1960 if that will be easier to write up.
   492. Joey Numbaz (Scruff) Posted: February 06, 2009 at 06:09 PM (#3069451)
Actually - if you look at my post 467, there's a significant change in usage of top starters from 1958 -> 1959, and it makes our groups a little easier to deal with too.
   493. mulder & scully Posted: February 06, 2009 at 07:16 PM (#3069548)
I like Joe's suggestion from post 485.
   494. karlmagnus Posted: February 06, 2009 at 07:24 PM (#3069564)
I agree; I like Joe's 485 too.
   495. Paul Wendt Posted: February 06, 2009 at 09:20 PM (#3069694)
I looked at the whole distributions of pitcher starts (not innings), and at the distributions of pitcher starts restricted to at least 10 or restricted to at least 20.

The headline story is dramatic change from 1904 to 1909 and relative stability thereafter. Regarding the season-level distribution of starting assignments 1909 and 1960 are on the same street; 1909-13 and 1956-60 are in the same neighborhood.

The break between 1920-23 and what followed is clear. Because I have discarded war 1914-15 and 1918-19 for now, nothing is clear on the other side. Beside the wars that disrupted the labor supply, 1916-17 was the highwater mark for games played, maybe also for innings.

Workloads were light in 1946 and 1945. I suppose that is because baseball teams were required to rehire returning veterans (and maintained larger rosters in order to handle them in 1946); because ballplayers recovered their skills, or didn't, in a time-consuming process partly during the championship season; because there were many wartime replacements to assess under "normal" conditions.
   496. Paul Wendt Posted: February 06, 2009 at 10:03 PM (#3069736)
The second page of "1910 Ballot Discussion" includes a lot on playoff WPA (win probability added) for pitchers, by DL from MN plus discussion partners.
playoff WPA for pitchers (see #125-153)

152. DL from MN Posted: February 06, 2009 at 12:41 PM (#3069502)
. . .
For the 1880s guys I just followed the general rule of thumb that a shutout was worth about 0.6, giving up 1-2 runs was worth around 0.3, giving up 4 runs was worth 0 and more than that was progressively more negative. You have to eyeball the game too though, if you give up 5 in a 7-5 win then the pitcher typically gets no WPA while a 1-0 10 inning win is worth around 0.75. I certainly didn't calculate actual WPA for any playoff series pre-1903.

Without complete digitized play-by-play there will be no WPA calculations soon, so it won't be possible to handle the 1880s systematically. That's why we have segregated the 1880s pitchers! Maybe I'll be able to do something for Caruthers and others when the time comes.

DL, I'm sure that your back-of-the-envelope estimates are fooled by the run environment, probably because of the high rates of unearned runs.

Caruthers played in four Series. In 1885-87 --tie, win, lose-- the Browns faced teams with season OPS roughly .700, .800, .800 and held them to .630, .600, .600. On average Caruthers pitched better than his teammates. (The Browns returned in '88 but Caruthers returned in '89.) The 1889 Giants hit better during the Series than during the season but Caruthers alone among Brooklyn pitchers nearly matched his AA season ERA --up from the low threes to the high threes where his teammates all blew up over seven.

Clarkson pitched well too in 1885 and 1886. Despite the 2-3-1 WLT record (teammates 3-4-1) he led easily by ERA. McCormick 3-3 WL started 5 of 7 in 1885 and 1 of 6 in 1886.
During Caruthers's and Clarkson's time on the agenda here I'll try to do something that incorporates a check for unearned run illusions, perhaps take a step beyond WHIP by estimating OPS against each pitcher.

Radbourn was in one series, Keefe three (Caruthers five but he barely played outfield in 1890).
That's discouraging because it's a lot. We may soon have digitized play-by-play for the World Series, soon in statistorians time, but it won't be soon enough for this project.
   497. Paul Wendt Posted: February 06, 2009 at 10:12 PM (#3069751)
Quoting from "1910 Ballot Discussion"
This #122 is the local start of [a candidate for 2010 -Ed.]Pitcher Playoff WPA by DL from MN. See #122 to #153, at least.

122. DL from MN Posted: January 29, 2009 at 12:54 PM (#3063403)
I discussed my WPA method for adding postseason credit previously (normalize each postseason to 4 wins in a pennant, zero out negative postseasons, add up all postseasons). I've worked up several pitchers currently at the top of my ballot and some surrounding them. I plan on working up all the pitchers as we go through the voting. Feel free to doublecheck my numbers.

Pitcher PennantWPA
------- -----
Bridges: 1.36
Coveleski: 1.32
Tiant: 0.49
. . .
[others deleted. The point of entry was 2010 candidate Tommy Bridges -Ed.]
   498. DL from MN Posted: February 06, 2009 at 11:11 PM (#3069800)
Some of what we're calling the history of "pitchers" might be the history of fielding equipment. Modern looking gloves came into style in the 1890s including the catcher's mitt. In the 1920s (another breakpoint in the data) the Bill Doak model with webbing between the thumb and forefinger appeared. The Wilson A2000 with the full webbing came just after WWII and the first baseman's mitt arose around WWII.

http://www.popularmechanics.com/outdoors/adventures/1277471.html?page=2

Other big impacts:
1880s - chest protector, ash bats
1893 - 60'6"
1894 - pitcher's mounds, foul bunts are strikes
1895 - bat diameter increased to 2-3/4", evolution of gloves
1900s - shin guards and better catcher's mask
1911 - cork center ball
1920 - outlawing the spitter, umpires replace scuffed balls
1920's - open 'platform' catcher's mask, machine wound balls, Groh's bottle bat, Bill Doak glove
1930's to 1940's - the slider
1945 - first basemen's mitt
1950's - batting helmets, lights, 33 ounce bats, Wilson A2000
1960's - hinged catcher's mitt
1969 - lowering the mound, smaller strike zone
1970's - batting helmets with ear flaps

As you can see the biggest innovations were in the mid 1890s, in the 1920s and just after WWII into the 1950s. The only real change after that was lowering the mound. I think there is historical support for our current breakpoints.

http://www.baseball-almanac.com/rulechng.shtml
   499. Paul Wendt Posted: February 07, 2009 at 12:18 AM (#3069845)
As you can see the biggest innovations were in the mid 1890s, in the 1920s and just after WWII into the 1950s.

- perhaps only because that presentation doesn't really begin until the 1890s.

- "evolution of gloves" does not impress me


Rules, equipment, the rubber and the mound (a groundskeeping innovation) may be important in multiple ways. The pitching motion may become more or less tiring or destructive as some elements are required or forbidden or enabled or hindered. Thru any effect on the balance between pitcher and batter, one game or one inning may become shorter or longer. Also thru any effect on the balance, pitches may lose or gain opportunities to coast during the games.

Rules and innovations in schedules, roster sizes, and recruiting --beside indirect effect via the balance between pitcher and catcher-- may induce or force teams to use their staff pitchers differently.

Why did teams use more pitchers with lighter workloads 1909-1913 (vs ten years earlier)? Part of it may have been an economic boom for the game during the aughts. Griffith and McGraw began to use more pitchers not only because they saw that Chesbro, Matty, and McGinnity got tired but also because they could afford it. They could pay for larger traveling teams. (Did it become common during the aughts for the entire team to travel on road trips?)

Why did teams use fewer pitchers with heavier workloads 1920-1923 (compared with ten years earlier)? Part of it may have been the wartime shocks to player supply --Federal League and European wars. That is continuing to make along with what those shocks encouraged them to try.
   500. OCF Posted: February 07, 2009 at 12:55 AM (#3069865)
Griffith and McGraw began to use more pitchers not only because they saw that Chesbro, Matty, and McGinnity got tired but also because they could afford it.

Perhaps they also saw the results the Cubs and Pirates were getting by using each pitcher less, and moved to copy someone else's successful model.
Page 5 of 6 pages  < 1 2 3 4 5 6 > 

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Dingbat_Charlie
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.1052 seconds
49 querie(s) executed