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

Monday, June 27, 2005

1955 Ballot Discussion

1955 (July 3)—elect 2
WS W3 Rookie Name-Pos (Died)

278 73.8 1933 Dixie Walker-RF (1982)
263 77.0 1934 Augie Galan-LF (1993)
217 65.1 1938 Jeff Heath-LF (1975)
180 64.4 1933 Schoolboy Rowe-P (1961)
185 59.0 1935 Elbie Fletcher-1B (1994)
179 51.5 1936 Buddy Lewis-3B/RF (living)
169 44.5 1940 Stan Spence-CF (1983)
139 49.9 1938 Rip Sewell-P (1989)
145 45.2 1936 Harry Gumbert-P (1995)
138 42.0 1942 Whitey Kurowski-3B (1999)
110 41.1 1938 Nels Potter-P (1990)
113 38.5 1941 Tex Hughson-P (1993)
118 36.7 1940 Tiny Bonham-P (1949)
123 33.9 1938 Taffy Wright-RF (1981)
101 32.9 1940 Wally Judnich-CF (1971)
102 29.3 1940 Frank Gustine-2B/3B (1991)
080 31.0 1939 Mike Tresh-C (1966)
081 26.1 1939 Hugh Casey-RP (1951)

1955 (July 3)—elect 2
HF% Career Name-pos (born) BJ – MVP - All-Star

HF 33-53 Buck Leonard-1B (1907) #1 1b - 4 - 7*
64% 30-53 Ray Brown-P (1908) 5 - 2*
00% 37-49 Eugene Benson-CF (1913) #5 lf – 1 - 0*
00% 40-49 Booker McDaniel-P (1912) 2 - 2*
00% 40-49 Tommy Sampson-2B (1914)0 - 1*


Players Passing Away in 1954

HoMers
Age Elected

57 1943 Oscar Charleston-CF

Candidates
Age Eligible

87 1907 Hugh Duffy-CF/LF
86 1903 Sadie McMahon-P
76 1916 Bill Bradley-3b
70 1922 Chief Wilson-RF
70 1923 Chief Bender-P
64 1929 Jim Bagby-P
63 1934 Bill Doak-P
62 1939 Rabbit Maranville-SS
61 1931 Walter Holke-1B
55 1945 Earl Whitehill-P
53 1940 George Grantham-2B/1B
37 1954 Russ Christopher-RP

Thanks to Mr. Greenia and Mr. Cobb for the lists!

John (You Can Call Me Grandma) Murphy Posted: June 27, 2005 at 05:33 PM | 156 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 2 of 2 pages  < 1 2
   101. TomH Posted: June 30, 2005 at 07:28 PM (#1442263)
waaay too little arguin this week. Where's the Independence Day Passion?! When in the course of human events, it becomes necessary to break the bands of dissing underrated hurlers....

anyway, somebody 'splain this to me:

1954 ballots pts - votes
..R Ruffing 354 ..28
..W Ferrell 296 ..24
..E Rixey.. 246 ..21
..B Walters 125 ..11

Bucky got drubbed by the other 3 modern pitchers. Now, among these 4: by Win Shares, Walters is 3rd in career totals, 2nd in WS/yr. For you WARPians, Walters is 2nd in career (only to Ruffing), and again 2nd in rate (WARP/yr). He has the LOWEST 'translated ERA' among all of these. So what gives?? I accuse you redcoats of 'avoid the new toy syndrome'! I hereby dump all of your pure career total and peak season tea into the Hahbuh!
   102. TomH Posted: June 30, 2005 at 07:33 PM (#1442288)
and to Dolf Lucky, the Best FOBW..keep the faith, buddy!
   103. Howie Menckel Posted: June 30, 2005 at 08:15 PM (#1442482)
Walters is mostly 1939, 1940, and 1944.
And 1944 is a war year.
Yes, also solid in 1941 and 1942.

I defy anyone to actually LOOK at Walters' year by year career on baseball-reference.com and then tell me that's even a top 15 HOM candidate.

(How's that, TomH?)
   104. OCF Posted: June 30, 2005 at 08:16 PM (#1442485)
My own equivalent records for these four five pitchers, offense-adjusted for Ruffing, Ferrell, and Walters:

Ruffing 269-214; 27 big year points.
Rixie 275-224; 19 big year points.
Ferrell 177-115; 49 big year points.
Walters 197-148; 43 big year points.
Bridges 190-124; 17 big year pionts.

Ruffing and Rixie are the career candidates; Ferrell the peak candidate. Walters has big years, too - helped along by the fact that his best hitting years coincided with his best pitching years. Bridges has a peak/rate case but without the big years.

Chris C. has Ray Brown at something like 269-190, and with some big years. That does seem to go at the head of this particular line.
   105. Dr. Chaleeko Posted: June 30, 2005 at 08:53 PM (#1442641)
A LITTLE MINI-STUDY ON WARP AND WS AND HOW THEY VALUE THE DEFENSIVE SPECTRUM

So FRAR ends up really being deterministic of the final rating. Or maybe that is just an artifact of this particular case (these particular players)?

If I understand WARP correctly, and I don't know that I do...this isn't entirely accurate.

My impression has always been that FRAR itself is not the deterministic factor, but rather that something more basic is---position.

To rephrase something I said earlier but a lot more clearly...

WS does not adjust batters for position but does make adjustments for position in fielding. To WS every batter is the same in relation to one another, but not every fielder is.

WARP adjusts hitting based on position as well as fielding, creating distinctions between each position both at field and at bat. To WARP whether at the plate or in the field, each batter is only comparable to the other fellows playing his position.

[However, that being said, I'm not sure that's still current. Remember that BP took down BRARP from its player pages about a year ago. So unless they've returned BRARP or are simply choosing not to display it but still making it part of the calculation, then may no longer be basing hitter's performance on position.]

A couple years ago I did a little digging around and either found on the site or else reverse engineered some aspect of WARP to figure out what the replacement-level EQA for each position was, and it looked like this:
Pos       EQA    PCT
====     =====  =====
general  .230    .90
1b       .253    .99
2b       .2299   .90   
3b       .237    .93
ss       .2213   .87
lf       .244    .96
cf       .2365   .93
rf       .2475   .97
c        .2245   .88
dh       .255   1.00


The third column reflects how much offensive production each position contributes relative to the most difficult hitting position (difficult meaning having the highest replacement-level EQA).

Meanwhile, I also looked into what the "intrinsic weights" for each position looked like in WARP too. That intrisic weights vary season-by-season in WARP1, but...

In the all-time adjustments, an average catcher is set to 39 runs above replacement per 162 games, first base to 10, second to 29, third to 22, short to 33, center field to 24, left and right to 14.

Depending on how you approach this idea, the results are substantially similar to WS.

For instance, let's say that this set of numbers represents the true difference in the level of difficulty among the positions. OK, then catcher is the most difficult and therefore contributes the most runs at a time, and, therefore, has the largest difference between replacement and average. So now, we can just examine the ratio of FRARP for each of the positions relative to the most difficult postion to see to what degree WARP values one position over another. SS, for instance is about 97% as difficult as catcher.

Well, WS says a catcher builds 19% of his value from fielding and that SS builds 18% of his value from fielding or that SS is 94.7% as difficult as catching.

Now, we can then see how closely WS and WARP evaluate the relative difficulty of the positions

Pos   EQA   WS
===  ===== ====
1b    .31   .32
2b    .87   .84
3b    .67   .63
ss    .97   .95
lf    .51   .37
cf    .67   .73
rf    .51   .37
c     1.0   1.0
dh    0.0   0.0%


These are substantially similar conclusions about the value of one position versus another. Which is more "correct" is, of course, debatable.

Does this contribute to our discussion? I have no idea. I had hopes when I started that it would....
   106. KJOK Posted: June 30, 2005 at 10:07 PM (#1442821)
It helps me, and I thought I basically understood already! The CF values are particularly interesting in how they differ.
   107. sunnyday2 Posted: June 30, 2005 at 10:07 PM (#1442822)
Conceptually, to me, it all starts with the defensive spectrum.

Some positions are harder to field, no question. And the idea is that, the harder the position is to field, the less offense is expected out of that position.

What is confusing to me is the inference--incorrect inference--that there is something incompatible between fielding at catcher or shortstop on the one hand, and hitting on the other. Not the case. What is correct is simply that you've got a smaller pool of people who can field certain positions and therefore, with a more or less random distribution of hitting ability throughout the different levels of fielding ability, there are simply fewer who also can hit.

So far, so good.

But what are the implications of this insight for evaluating talent--whether you are doing it to build a winning team, for one thing, or build a HoM on the other? (Are they different for each of these purposes? I hope not, it's hard enough thinking about it once.)

I don't have an answer, just a question. But I would now add to my list of crucial issues as we approach the day of reckoning, when we start electing 3 and (maybe) go deep into our backlog of borderline candidates.

1. Sort out the Tier 2 NeL pitchers

2. Determine how to balance offense and defense on the defensive end of the spectrum (essentially among Cs and SSs)
   108. Jim Sp Posted: June 30, 2005 at 11:39 PM (#1442965)
Dixie Walker #62. Augie Galan, and Jeff Heath are HoVG candidates, well off the ballot.

Beckley, Rixey, Waddell, Cravath, Monroe, Bresnahan, Griffith, Joss, Jose Mendez, and Welch are in my PHoM but off my ballot.

1)Buck Leonard--
2)Ray Brown--
3)Beckwith-- A great hitter, he played a considerable amount at the difficult end of the defensive spectrum. I’m willing to give him the benefit of the doubt on his “unusual circumstances”. His selection as manager indicates to me that his intangibles weren’t all negative. He made my PHoM in 1940 over Coveleski and Faber.
4)Suttles--
5)Averill--Looks like a HoMer to me even without PCL credit, but I do give him some PCL credit as he was obviously major league quality before arriving in the majors. Compare him to Goslin: Averill has a higher OPS+ (133/128), and is an A+ CF vs. a C+ LF. Goslin has career length, mostly because Averill plays in the PCL for a while.
6)Mackey--#2 on my 1949 prelim, but more data on his hitting has dropped him to here.
7)Cool Papa Bell--If Max Carey is in, Cool Papa should be too.
8)Schang--His rate stats would put him in the HoM, but a look at each individual year isn’t impressive. Still, a hitting catcher with his career length isn’t common...Bill James rates him a C+ fielder in Win Shares, but says he was a good catcher in the NHBA.
9)Lombardi--Well, there goes my consensus score. A long career as a catcher with a big bat. I see 15 obvious catching electees: Gibson, Bench, Fisk, Carter, Hartnett, Dickey, Piazza, Berra, Simmons, Ewing, Cochrane, Campanella, Parrish, Rodriguez, Santop. I’m an advocate for what I see as the next tier: Freehan, Munson, and Porter will get strong consideration on my ballot too. You can’t have a baseball team without a catcher. Almost 2000 games caught including PCL.
10)Sewell--109 OPS+, reasonably long career, good shortstop (A- Win Shares). Yes, I am allowing for his switch to 3B at the end of his career.
11)Medwick--
12)Bob Johnson--A very underrated player. Usually I'm a WS guy but this time I think Warp has it right.
13)Billy Herman-- I’m still perplexed trying to figure out his career relative to the defensive spectrum shift at 2B. He looks good compared to modern 2B, not so great compared to early lively ball 2B. Gets two years war credit, that helps too.
14)Doyle— His hitting is legitimately outstanding, he played 2nd base, and a C+ defender by Win Shares. 126 career OPS+, compare to contemporary George Cutshaw, who was a regular 2B for 11 years with an OPS+ of 86. #19 all time in innings at 2B. Regularly in the 2B defensive Win Shares leaders, WS Gold Glove in 1917. Top 10 in Win Shares 1909-12, 1915.
15)Stan Hack--His time will come, I think. I like him better than Groh, who I voted for.

Ruffing#30, he’s HoVG but I don’t like him as much as the consensus.
Ferrell—one of the top 100 pitchers of all time, but not on my ballot currently.
Hughie Jennings—impressive peak, not enough career.
Rixey--#17
   109. Chris Cobb Posted: July 01, 2005 at 12:11 AM (#1443036)
Why Walters doesn’t do better.

His combined peak and career value isn’t high enough. He has 9 above average seasons according to win shares (four of which are 2 ws or less above average), in which he was a total of 95 season-and-WWII adjusted win shares above average. Ferrell, with a smaller career, also has 9 above average seasons and was 114.3 win shares above average. Rixey and Ruffing can’t match Walters’ 4 big seasons, but Rixey has 13 seasons above average, totaling 101 win shares above average, while Ruffing has 14, totaling 101.4 win shares above average. Given that Walters already trails these two pitchers substantially on career win shares, his pure peak advantage in his 4 great years (reflected in my own system in peak rate) isn’t nearly enough to catch him up. While he is ahead of Ferrell on cws about as much as Ferrell is ahead on ws above average, 266 adjusted win shares to Ferrell’s 245, Ferrell’s pure peak is better than Walters’ and puts Ferrell ahead.

I don’t actually use these win shares numbers; my home-grown pitcher win shares see Rixey as better, Ruffing as worse, Ferrell as better, and Walters as about the same as James’ win shares. But Walters falls short in both systems for the same reasons: his peak isn’t quite substantial enough to counterbalance his career shortfall. Ferrell’s superior peak is enough. I have them Rixey, Ferrell, Ruffing, Walters, and my full pitcher list, down as far as Walters, is Brown, Griffith, Rixey, Ferrell, Ruffing, Grimes, Mendez, Redding, Byrd, Harder, Mays, Shocker, Welch, Waddell, Walters, with the all-time in/out line probably running between Redding and Byrd.
   110. Mark Shirk (jsch) Posted: July 01, 2005 at 12:12 AM (#1443039)
I have Walters on my ballot based on his peak. He beats Dean and Waddell in peak and they are just off my ballot. I have his two best years as 1939 and 1940, and 1944 is still a damn good year even with a 10% discount.

I understand why people aren't voting for him but he was #14 last year and will be again this year unless I make some drastic changes.
   111. karlmagnus Posted: July 01, 2005 at 12:24 AM (#1443056)
Sunnyday2, I think that's right. There's no reason why a shortstop shouldn't be able to hit as well as Babe Ruth, he will simply have a greater value than the Babe if he does so. On the other end, however, Maranville can keep a job as a SS for 20 years, whereas he'd have washed out in 2 as a LF. This means that all SS hitting 125OPS are more valuable than LF hitting 125. Similarly, 1890s 1B hitting 125 are more valuable than outfielders hitting 125, so Beckley is more valuable than Duffy, Ryan and Van Haltren, and also as valuable as a 1930s RF and LF hitting 130 with a Beckley-length career (thus better than medwick/Averill, who had shorter careers.)

This is NOT however a rant about Beckley. It means that Rabbit Maranville and Dave Concepcion should not be elected for their fielding; it's just that managers were working off preconceptions and artificially reducing the value of great hitting good fielding players by making them 3B or outfielders. On the other hand, when a shortstop hits like a good outfielder, he is in fact more valuable than that outfielder, so Cronin should have been in the HOM by a mile, and if they have long enough careers, ALL of Jeter, A-Rod, Nomar Larkin and Tejada should go in.

The same applies to catchers except that a corrective needs to be made for greater wear and tear. Thus the Schnozz, even though a mediocre catcher, hit 125 OPS+ and hence was an extraordinarily valuable player (equivalent to an OF with a 140) and should be credited for that value, and not penalized for playing fewer games than an OF. Conversely, if you believe Mackey's OPS of 102 is correct or close, his career length doesn't qualify him when his OPS+ (equivalent to an OF with say 115) doesn't.

Carey was a mistake, and Rice and Hooper would be others. Sewell was MUCH more valuable than Carey, but probably still not good enough. Also, intermediate position players -- 2B,3B with 115 OPS+ -- shouldn't be inducted for position balance; almost all 3B can play the more valuable position of SS (albeit badly), and so it's not surprising if the HOM has more SS than 3B.
   112. KJOK Posted: July 01, 2005 at 01:12 AM (#1443114)
I apologize that this is long, but I'm copying this from High Boskage House regarding Fielding value vs. Batting Value. The figures sighted are for recent seasons of course, and past seasons will have a slightly different result weighted in favor of fielding, but I think the conclusion does still hold...


The average baseball defense will see maybe 4,500 opportunites--"balls in play" as we call them. Assume for discussion that the full-season difference in Fielding Efficiency from average is about plus-or-minus 2% (that is, ranges from 73% to 77% around a major-league average of 75%)--which experience suggests is actually an exaggeration of the team-to-team differences--we would see at the extremes of good and bad team fielding a difference of about 90 baserunners a year. If we take as a crude but passable rule of thumb (derived from experience) that 37% of base runners eventually score, we are suggesting a difference of plus-or-minus 33 runs a year from average to best- or worst-fielding team; that, in turn--from the games-won-from-runs formula--translates to about plus-or-minus three wins a season.

Let's put that in perspective. The difference in earned runs from the norm for the best or worst pitching staffs in baseball will be 155 to 160 runs a season--almost five times the effect from best to worst fielding. Recall that baseball analysis demonstrates conclusively what elementary baseball common sense suggests to, apparently, all but the professional baseball person: runs scored on offense and runs surrendered on defense have essentially equal value in determining seasonal games-won results. That means that in baseball defense as a whole--pitching and fielding--is exactly 50% of the game. Fielding is, at most, not even quite 20% of "defense," so pitching is about 80% of 50%, or 40% of the game (old sayings notwithstanding).

And keep in mind that if things haven't changed that much over the years, by season's end the variations from best-fielding to worst-fielding team may be more like plus-or-minus 1% than 2%, which would make fielding 5% of the game (not 10%) and pitching 45% (not the "80%" or "90%" or whatever that old-timers so glibly blather about).

Conclusion: we can now say for certain that team fielding ability just is not a significant component of the game of baseball relative to pitching (which is at least four to possibly nine times more important), or--more directly significant, since fielders bat and batters field--than batting, which is five to ten times more important.

In short, wood trumps leather, big-time.
   113. Chris Cobb Posted: July 01, 2005 at 02:24 AM (#1443175)
KJOK's post does a good job of explaining the line of argument that justifies seeing fielding as much less valuable than pitching or hitting, but it contains (it appears to me) one glaring error and a couple of small distortions of the case that leads the overall argument to significantly understate the real value of fielding.

Here's what I see. When they get a rough estimate of the runs saved by fielding efficiency, leading to a conclusion that the difference between best-fielding and average fielding is typically 33 runs a year, I think they are on pretty solid ground. However, if we were going to sweat the details we would want to observe that the percentage of baserunners who score will be lower for a good fielding team than an average one, so using a flat rate probably underestimates the difference by a few runs.

Next, the big problem comes in when they say that, ok, great fielding saves 33 runs above average, but the difference in earned runs between a great team and and average team is typically 155 to 160 runs a season, 5 times the effect of best to average fielding. Ergo, they say, pitching is 80% of defense and fielding no more than 20%. This neglects to notice (I think) that "earned runs" totals _include_ runs saved by the fielding efficiency. Therefore, one should _subtract_, say, 33 runs saved from 160 runs saved, to get the share saved by the great pitching, which then suggests that fielding value is 26% of defensive value overall. This is still not quite the 33% of defense that win shares assigns to the fielders, but it's closer, and if we sweated the details about the way improved fielding efficiency decreases the rate at which baserunners score and the way in which good fielding supports good pitching, and the way in which a run saved and a run scored are not entirely equal, I think there's a pretty good chance that we'd end up pretty close to the win shares 67/33 ratio. In which case, batting is about 3 times more important than fielding, not 5 to 10 times more important as the writer quoted by KJOK claims.

If that's an accurate analysis, then we're back to the question of how much, within a context in which fielding is less important but still a significan part of player value, does a great fielder contribute to a team's success at run prevention?
   114. Chris Cobb Posted: July 01, 2005 at 02:32 AM (#1443178)
More arguments on fielding:

Sunnyday2 wrote: What is confusing to me is the inference--incorrect inference--that there is something incompatible between fielding at catcher or shortstop on the one hand, and hitting on the other. Not the case. What is correct is simply that you've got a smaller pool of people who can field certain positions and therefore, with a more or less random distribution of hitting ability throughout the different levels of fielding ability, there are simply fewer who also can hit.

Why is this inference incorrect? Consider. (1) Bulk is, to some degree, advantageous for power hitting. (2) Bulk is, past a certain point, disadvantageous for middle infielders because it limits their mobility and exacerbates the pounding their bodies take during the ordinary course of defensive play. The fielding demands of certain positions, then, rule out or at least strongly discourage certain body types that select positively for hitting. It's not that it's impossible for anyone who can play shorstop to be able to hit, obviously. But the fielding criterion for shortstop does screen out disproportionately ball-players who would be good hitters.
   115. KJOK Posted: July 01, 2005 at 03:52 AM (#1443224)
What is correct is simply that you've got a smaller pool of people who can field certain positions

and then

The fielding demands of certain positions, then, rule out or at least strongly discourage certain body types that select positively for hitting

I think you two are actually AGREEING, just that Chris is bringing up one of the reasons WHY the pool is smaller.
   116. KJOK Posted: July 01, 2005 at 04:00 AM (#1443228)
Therefore, one should _subtract_, say, 33 runs saved from 160 runs saved, to get the share saved by the great pitching, which then suggests that fielding value is 26% of defensive value overall.

Not sure it works that way. That 'best' pitching team and the 'average' pitching team most likely differ in defense by ZERO, not by 33 runs, correct?

In other words, it's just as likely that the great pitching team that is 160 earned runs better could have a less than average defense while the average pitching team could have a better than average defense, making the spread GREATER than 160 for just pitching?! Or am I completely missing something here?
   117. Chris Cobb Posted: July 01, 2005 at 04:50 AM (#1443265)
Not sure it works that way. That 'best' pitching team and the 'average' pitching team most likely differ in defense by ZERO, not by 33 runs, correct?

In theory, perhaps, but not as defined by the analysis presented.
If you are identifying the best pitching team by picking the team with the best earned run average, as this study is doing, then you are wrongly attributing to the pitchers the runs saved by the fielders. ERA does not remove most of the effects of fielding efficiency, so the team with the lowest ERA will be the one that has the highest combination of pitching quality and fielding efficiency. For a team to be identified as a "great pitching team" by ERA, it will necessarily also be a great fielding team. It will be usual, rather than random, for any team that leads the league in runs saved to lead it in defensive efficiency.

Just as an empirical check, I looked at NL 1954-1958, and every year, if I have handled the park factors correctly, the team with the best ERA is also the team with the best defensive efficiency. Over a larger sample (and with verified correctness in the handling of park factors), I'm sure the correlation wouldn't be 100%, but it will be strong enough to prove that the runs saved by a great pitching team, identified by this method, will include a large chunk of runs saved that should be credited to the fielders.
   118. sunnyday2 Posted: July 01, 2005 at 12:17 PM (#1443347)
Chris et al,

Even if good team defense saves 33 runs a year, the question is how are those distributed? A really good SS might save all 33 and everybody else hover around zero, no?

Or pick whatever distribution you like. You still can't argue that 33/8 = 4 and 4 is the most that an individual player can possibly (or is likely to) save, I don't think.

Secondly, I'm not sure I agree re. body types. This rpesumes that fat guys are better power hitters when, in fact, it seems to me that a guy like A Rod is a pretty good body type for that. Maybe a better body type.

Has everybody forgotten, BTW, that WS caps defensive value? Certain teams and players max out defensive value and can get no more credit even though they were in fact better than what they're credited with. Maybe this is why WS is generally less friendly to the left end of the defensive spectrum. It a priori rules out certain values.
   119. Dr. Chaleeko Posted: July 01, 2005 at 01:13 PM (#1443389)
Sometimes I need to recap this stuff for my own comprehension, so I'm sorry if I'm just regurgitating....

I think there's a pretty good chance that we'd end up pretty close to the win shares 67/33 ratio. In which case, batting is about 3 times more important than fielding, not 5 to 10 times more important as the writer quoted by KJOK claims.


So we're back to square one. Dave Bancroft's glove is awesome, his hitting is mediocre, and so he's a borderline candidate because his biggest strength, his glove, does not offer enough value to counteract his bat. Or to more specifically address Sunnday's query, it's his glove that gets him into the position of being a HOM candidate in the first place, but his bat isn't potent enough to get him a bloc of votes, and he didn't play long enough to get the career-sympathy vote. Value is still value, it's just "easier" to stack up value if you can hit.

Even if good team defense saves 33 runs a year, the question is how are those distributed? A really good SS might save all 33 and everybody else hover around zero, no?

And this question takes us back to the chart I posted earlier. WS and WARP have thought it through, and they have roughly agreed on the average distribution of runs across the defensive spectrum. Of course, on a team-by-team basis things vary: A-Rod probably saves 90% of the runs "saved" by the 2005 Yankee defense, but that's becaus everyone else plays their position like a water buffalo.

On an excellent defensive team like, say, the 1985 Cards, the actual distribution of runs saved then ought to more closely mirror the WS/WARP model of the average team because so many of them were good defenders.

Which, to summarize the obvious, means there are two dimensions to fielding analysis in each of these systems:
1) runs saved in relation to the team
2) runs saved in relation to other players in the league at the same position.

So returning to the 1985 Cardinals, they're distribution of runs might well be similar to the distribution of runs saved across the defensive spectrum as suggested by the "intrinsic weights" in each system. But at most positions, their defenders' actual runs saved (as calculated by the system) will still be higher than would the runs saved by other players at their position throughout the league.

All of which is to say that defense-first candidates do appear to be getting a fair shake. Yes, it's tougher to accumulate value via defense, but their profile also tells us that they want for offensive value, which is an even more important aspect of any player's game.

Now returning to Bancroft versus Sewell, value is value. These guys are extremely close is value in every way via WS. IMNSHO, WARP is way overstating the relative strength of their leagues and so distorting the differences between them. It's a matter of Bancroft building his value on the glove side and Sewell with a little more of the stick. Value is value. Throw 'em in a hat.
   120. andrew siegel Posted: July 01, 2005 at 01:29 PM (#1443408)
I have a lot of comments but only time for one. So, here's the one I think is most important:

If you take the amount of runs saved by the best defensive team and start divvying up the runs saved to the various players and use that to cap defensive value, you will drastically underestimate the amount of runs saved by the very best defensive players. That is because the very best defensive team will likely be made up of a few great defensive players, a bunch of good defensive players, and one or two average or poor defensive players. The sum of runs saved by an all-star defensive team will always be higher than and may even be twice as much as the sum of runs saved by the best actual defensive team, depending on the distribution of defensive talent in any given league.
   121. Chris Cobb Posted: July 01, 2005 at 01:33 PM (#1443416)
Even if good team defense saves 33 runs a year, the question is how are those distributed? A really good SS might save all 33 and everybody else hover around zero, no?

Well, not exactly. If we are positing 33 runs saved as the de facto limit on fielding value, then we should infer that this is a team that has several exceptional fielders. I'm not saying that we should assume that the division of runs saved is anything like even, but I think the presence of a single "best-of-all-time" fielder at shortstop would not turn an average fielding team into the best fielding team.

Or pick whatever distribution you like. You still can't argue that 33/8 = 4 and 4 is the most that an individual player can possibly (or is likely to) save, I don't think.

This seems exactly right.

Secondly, I'm not sure I agree re. body types. This rpesumes that fat guys are better power hitters when, in fact, it seems to me that a guy like A Rod is a pretty good body type for that. Maybe a better body type.

Here's a bit of empirical evidence.

Average height and weight of the 36 players who have hit 50 or more home runs (players hitting 50 more than once counted more than once):

6' 0.5", 207 lbs.

Lightest is Willie Mays at 5'11", 180 lbs.
Shortest is Hack Wilson at 5'6", 190 lbs. (he is only player under 5'11").
A Rod is at 6'3", 190 lbs.

Average height and weight of the 20 post-1900 players who are the win shares all-time single season leaders in fielding win shares at shortstop, second base, and third base (for third post-1920 only):

5'10.5", 172 lbs.

Heaviest is Nap Lajoie, 6'1", 195 lbs.
Tallest is Bobby Grich, 6'2", 190 lbs.

So A-Rod is toward the small size for a 50-home-run hitte (which he is), and at about the maximum size for a historically great defensive middle infielder (which he is not).

Has everybody forgotten, BTW, that WS caps defensive value? Certain teams and players max out defensive value and can get no more credit even though they were in fact better than what they're credited with. Maybe this is why WS is generally less friendly to the left end of the defensive spectrum. It a priori rules out certain values.

WS caps fielding ws at the team level. This cap is reached quite frequently prior to the lively ball era, but it is seldom reached after that. It's already the HoM consensus that win shares underrates the fielding value of pre-1920 (and certainly pre-1900) fielders. Whether it underrates them from 1930 onward would depend more on the general weights it uses to divide value between pitching and fielding and the intrinsic weights assigned to each position than to the the presence of the cap, which is a smoking gun concerning WS underrating of early fielding.
   122. sunnyday2 Posted: July 01, 2005 at 02:04 PM (#1443497)
Doc, re. WARP, just go to WARP1 and you eliminate it's most egregious problem.
   123. Dr. Chaleeko Posted: July 01, 2005 at 02:16 PM (#1443519)
And when you go look at WARP1, it basically tells me the same thing: throw 'em in a hat.

Bancroft has nine to ten more WARP total, but Sewell has a slightly higher peak. Do it by WS, and it's a similar thing.

Which means just one thing....

Sewell voters, vote for Dobie Moore!
   124. sunnyday2 Posted: July 01, 2005 at 03:10 PM (#1443640)
You are corrrect sir!
   125. Mark Shirk (jsch) Posted: July 01, 2005 at 07:06 PM (#1444334)
One thing about WARP3 (and 2 I guess) concerning Sewell and Bancroft...

They aren't exact contemporaries. Many of Bancroft best seasons were in the 1910's NL (He was rookie of the year on the Phillies' pennant winning 1915 team), A league I think we can all agree wasn't the most competitive.

Sewell's career as a regular started with Ray Chapman's death on 1920 (I believe), so he doesnt' get anything taken off for playing in the mid 1910's.

I think they are close with Sewell being slightly better. This is a reason that I am not really behind either of them, Sewell is somehwere in the high thirties, Bancroft in the low 50's.
   126. EricC Posted: July 01, 2005 at 10:22 PM (#1444803)
just go to WARP1 and you eliminate it's most egregious problem.

When you deal a shuffled deck, is it any surprise that one hand comes out stronger than another? There are only about 200 regular players in a 16-team major leagues- "shuffle and deal" them, and, even by random chance, one would expect a big enough difference between leagues to matter when it comes to evaluating players. Now the question of quantifying this difference is unresolved, but, based on my own attempts to quantify league differences, I'd say that WARP3 gives a more accurate picture than WARP1 when it comes to comparisons between players in contemporary leagues.
   127. Brent Posted: July 02, 2005 at 01:08 AM (#1445084)
Augie Galan - Will he be the first player with three 30-WS seasons not to get a single HoM vote?

If the middle portion of his career had just been consistent with its beginning and end, he would have been a wonderful player.
   128. Brent Posted: July 02, 2005 at 01:27 AM (#1445109)
Howie Menckel wrote:

I defy anyone to actually LOOK at Walters' year by year career on baseball-reference.com and then tell me that's even a top 15 HOM candidate.

I don't know about that. Comparing Vance and Walters, they obviously were very different types of pitchers, but their results look pretty similar in peak, prime, and career, with a modest advantage to Vance. Vance acquired almost all of his value from pitching, but Walters was a much better hitter. It's hard for me to see how someone could have listed Vance number one and not have Walters somewhere in their top 15.
   129. Brent Posted: July 02, 2005 at 02:14 AM (#1445192)
Taking stock –

About ten “years” ago, Devin McCullen posted (on the Goose Goslin thread, # 21) a list of HoMers organized by the decade in which each player’s career was “centered.” Since we are now near the end of assessing the players of the 30s and just starting on the 40s, I thought it would be interesting to update his list. I’ve also added counts of each decade’s HoMers by primary position and parenthetical lists of the viable candidates (defined as players who received at least 100 points in the last election, with the top 10 candidates shown in italics).

Although it would be possible to divide several players’ careers either across decades or across positions, I’ve tried to avoid doing so--the only exception being that I’ve classified both Ward and Dihigo as half pitcher and half position player.

1860s - 1 (Pearce) (SS)

1870s - 9 (Anson, Barnes, McVey, Pike, Spalding, Start, Sutton, White, Wright) (P, C-2, 1B-2, 2B, 3B, SS, CF)

1880s - 17 (Bennett, Brouthers, Caruthers, Clarkson, Connor, Ewing, Galvin, Glasscock, Gore, Hines, Keefe, Kelly, O'Rourke, Radbourn, Richardson, Stovey, Ward) (P-5.5, C-2, 1B-2, 2B, SS-1.5, LF-2, CF-2, RF)
{Candidates – Browning, Welch, C Jones}

1890s - 13 (Burkett, Dahlen, Davis, Delahanty, Grant, Hamilton, Keeler, Kelley, McPhee, Nichols, Rusie, Thompson, Young) (P-3, 2B-2, SS-2, LF-3, CF, RF-2)
{Candidates – Jennings, Griffith, Beckley, Van Haltren, Duffy, Childs}

1900s - 16 (M Brown, Clarke, J Collins, Crawford, Flick, R Foster, Hill, G Johnson, Lajoie, Mathewson, McGinnity, Plank, Sheckard, Wagner, Wallace, Walsh) (P-6, 2B, 3B, SS-3, LF-2, CF, RF-2)
{Candidates – Leach, Waddell}

1910s - 15 (Alexander, Baker, Carey, Cobb, E Collins, Groh, Jackson, W Johnson, Lloyd, Magee, Santop, Speaker, Torriente, Wheat, Williams) (P-3, C, 2B, 3B-2, SS, LF-2, CF-4, RF)
{Candidates – Redding, Mendez, Cravath}

1920s - 12 (Charleston, Coveleski, Faber, W Foster, Frisch, Goslin, Heilmann, Hornsby, Rogan, Ruth, Vance, Wilson) (P-5, 2B-2, 3B, LF, CF, RF-2)
{Candidates – Beckwith, Rixey, Mackey, Sisler, Sewell, Schang, Grimes}

1930s - 20 (Cochrane, Cronin, Dickey, Dihigo, Foxx, Gehrig, Gehringer, Gibson, Greenberg, Grove, Hartnett, Hubbell, Lyons, Ott, Simmons, Stearnes, Terry, Vaughan, Waner, Wells) (P-3.5, C-4, 1B-4, 2B, SS-3, LF, CF, RF-2, UT-0.5)
{Candidates – R Brown, Suttles, Herman, Medwick, Ruffing, Ferrell, Averill, Bell}
Not yet eligible - Paige.

1940s – 0
{Candidates – Leonard, Hack, Walters}.


What do we learn? The 1930s have now moved ahead of all other decades in HoMers, and with most of the leading candidates in the backlog also coming from the 30s, it appears that this decade will ultimately be represented by 25 to 28 HoMers. Is this a problem? Not necessarily. Part of the reason the 30s appear strong is because many of the players whose careers are centered in the decade overlapped with the 20s or the 40s, both of which are likely to end up underrepresented by this measure. Also, the maturation of the Negro leagues around this time increased the depth of the candidate pool.

Comments?
   130. Chris Cobb Posted: July 02, 2005 at 02:39 AM (#1445249)
Although this "centered in a decade" measure is somewhat crude, it does suggest that the 1930s are significantly ahead of most other decades, even with overlap taken into account.

We shouldn't, and can't, try to balance things out immediately, but I would argue that we are by no means finished with the 1890-1930 period. We should end up with 17+ reps for each of the decades between 1890 and 1930 if we are to represent fairly the top players in those eras.

For the 1890s, I'd like to see three or four of the listed six get elected.

For the 1900s, I'd like to see at least Leach go in, and Waddell, Bresnahan and Moore should be on the radar.

For the 1910s, at least a pair of Redding, Mendez, Cravath, and Doyle should go in.

For the 1920s, Beckwith, Rixey, Mackey, and Sisler ought to go in, and at least one from the Sewell/Schang/Grimes group.

Looking farther back, I think we're finished with 1860-1890 until we catch up to the present, but Tommy Bond, Mickey Welch, Charlie Jones, Pete Browning, and maybe Tony Mullane, Ed Williamson, and Fred Dunlap should not be forgotten in the longer term.
   131. EricC Posted: July 02, 2005 at 03:43 PM (#1445594)
Comments?

Scaling the data by decade in #129 by the ratio of the 2000 U.S. population to the U.S. population at the beginning of the decade:

1870s: 64
1880s: 95
1890s: 58
1900s: 59
1910s: 46
1920s: 32
1930s: 46

My thoughts:

1. The biggest anomaly is the excessive number of 1880s players, which can be explained by the two-league system, statistical distortions in an era of high variance, and by the HoM having started in 1898 with the particular pattern of elect-by-year that it did.

2. The biggest deficit is in 1920s players. I suspect that this is due to the "Ruth shadow" effect plus the fact that we are not yet finished electing 1920s players.

3. Their is actually no shortage of 1890s players in the HoM. The belief that there is a shortage is an illusion that comes from the excess of 1880s players.
   132. Chris Cobb Posted: July 02, 2005 at 04:03 PM (#1445614)
Since the election schedules is based on the number of major-league teams, not on U.S. population, a population-based scale, while interesting, is not directly relevant, and results that don't agree with this scale can hardly be described as anomalous, as it is not the measure the HoM is designed to use. It's highly likely that we will elect fewer players from the more populous 1940s and 1950s than the 1930s, so these ratios are not going to show a consistent pattern, unless it turns out to be one of gradual decline, which we might generally expect, since the election schedule is scaled to the number of major-league teams, not to the size of the population.
   133. EricC Posted: July 02, 2005 at 04:20 PM (#1445637)
results that don't agree with this scale can hardly be described as anomalous

The anomalies listed in #131 are still hold true with respect to a gradual decline in HoMers per population per decade (which may represent a more "fair" allocation than a strict per-population trend, though I don't think that a strict per-team allocation is right in principle).
   134. Howie Menckel Posted: July 02, 2005 at 10:29 PM (#1446010)
Brent,
It's interesting that you note the Vance-Walters comparison. I think Vance may not even have made by ballot, and that he is the only HOMer ever elected who didn't make my ballot that year.
So yes, they may be a good comparison. Not a problem for me, as neither makes a ballot.
I don't know what the Vance-lovers do with it, however.
   135. Brent Posted: July 02, 2005 at 11:45 PM (#1446097)
And Vance was # 1 on my ballot the year he came up; I placed Walters # 9 on last year's ballot.

Howie, I guess you and I don't see eye to eye when it it comes to evaluating pitchers.
   136. Howie Menckel Posted: July 03, 2005 at 12:01 AM (#1446112)
Brent,
I guess you're right there.

Where does Jennings rate for you? Seems to me like he did a lot more in his best years than Vance did.
   137. Brent Posted: July 03, 2005 at 12:49 AM (#1446202)
Howie,

I've always support Hughie too - had him # 7 last year.

Did Jennings do more in his best years than Vance did? In terms of value to his team, yes I think he did. Their win shares for their top 5 years (adjusted to 154 games) were:
Jennings 42,34,34,32,28
Vance - 36,32,26,25,20

For pitchers I actually use a mix of Warp1 and WS, which helps Vance:
Vance - 41,35,26,26,24

But Jennings is still ahead. However, that will be true of peak comparisons of almost all top live-ball era pitchers with top position players... the peak comparisons will favor the position player. To give live-ball pitchers a fair chance at peak scores I boost their scores by an amount that is based on average innings pitched during the decade - informaiton I've posted on the pitchers thread. For Vance, I boosted his raw scores about 19 percent, a factor that is typical for post-1920 pitchers.

I understand that voters who are looking to use a pure measure of "value" might not do such an adjustment, which may explain why the HoM honorees currently are only 26 percent pitchers. Personally, I think that 30 to 33 percent pitchers should be what we aim for - that's a conservative estimate of the importance of pitching to the game.

I've talked about this more on the Dizzy Dean thread - Dizzy is a better match than Dazzy to Hughie. (Just try repeating that last sentence three times quickly!) Actually Dizzy's peak, relative to live-ball era pitchers, is almost an exact match to Hughie's peak, relative to pre-1940 position players. I see their HoM cases as essentially identical.
   138. Howie Menckel Posted: July 03, 2005 at 02:37 AM (#1446416)
Good stuff, Brent.
I have bounced Hughie on and off my ballot, usually on, currently off, but may return soon. I don't mind SP/hitter of era adjustments, but any understanding of 1890s baseball, as you know, shows Jennings to be a staggering force. And the game was so brutal in that era, more than any before or since, that almost no one from several fielding slots was able to manage a long career.

I had pitchers in the 3 and 5 slots last year, with Rixey and Suttles ahead of electee Wells.
   139. Mark Shirk (jsch) Posted: July 03, 2005 at 01:20 PM (#1446584)
Seems to be a pretty boring ballt this year as the two newbies replace the two electees from 1954. Augie Galan will be somewhere in my top 50 as I have a soft spot for high peak/prime OBP machines.

1. Ray Brown
2. Buck Leonard - This is really close and in some ways Leonard was probably the better player. However, i can't really distinguish him too much from Suttles and Beckwith as hitters, even though I think he was a little better than both. Right now there isn't another pitcher I deem to be as good as Brown. Wes Ferrell is next on my ballot but he is boosted a few spots for his hitting. Both are PHOM

3. Hughie Jennings (PHOM)
4. Mule Suttles(PHOM)
5. John Beckwith(PHOM
6. Ducky Medwick
6a. Willie Wells
7. Wes Ferrell
8. Stan Hack
9. Billy Herman
10. Cupid Childs (PHOM)
11. Hugh Duffy
11a. Ted Lyons
12. Dick Redding
12a. Bill Terry
13. Clark Griffith
14. Bucky Walters
15. Earl Averill
   140. Brent Posted: July 06, 2005 at 04:58 AM (#1452010)
I'll address this question to our experts on Negro League pitchers, Chris Cobb and Dr. Chaleeko.

Bill Byrd vs. Hilton Smith - you both have rated Byrd quite highly (Dr. C at # 15, Chris at # 23 I believe) and were dismissive of Smith.

I have trouble understanding. Smith's MLE support-neutral record is 174-123; Byrd's is 208-180. The difference is 34-57, which, for a .500 ballclub, should be below replacement level. Should an extra 4 seasons of sub-replacement level pitching really count for enough to put Byrd ahead?

Also, if you're giving that much weight to career length, shouldn't we be more certain that Smith really wasn't a major-league quality pitcher before 1936? Looking at the information that's available on the Bismark team for 1935, it sure looks to me like he was already a good pitcher then. And it seems possible he may have been a ML-quality pitcher even earlier, though the record isn't there to tell us.

Maybe Smith wasn't an HoM quality pitcher (I have to admit, it's hard to tell from the limited record). But if he wasn't, then I'm having real trouble seeing how Byrd could have been either.
   141. sunnyday2 Posted: July 06, 2005 at 12:54 PM (#1452208)
Interesting discussion in #129-132 about the chronological distribution of HoMers and candidates. I was out of town over the weekend and just saw it.

I've said many times that HoMers are outliers by definition, so distributions of this and other kinds are not a big deal. But interesting nevertheless.

Chris, the timing of our elections (that is, number of electees per year) is based on the number of teams and it's true that value is a function of opportunity.

It's an open question, however, whether merit is a function of opportunity more than of skill or ability, or more of skill and ability. I don't mean skill in an absolute sense as this would imply a timeline, but I mean skill and ability relative to the norm at the time.

Well, actually, that's not an open question to me. While HoMers are outliers, I would suggest that while value is a function of opportunity, merit is nevertheless a function of skill and ability. And skill and ability reside in the population, not in the team or league.

Trying to balance our selections on Eric's scale (#31) would be a mistake, of course, as it would lead to a large number of HoMers from recent years, especially after taking into account the populations of foreign countries now represented in the MLs. But for the period that we have been dealing with it is thought-provoking at least. And I agree with his conclusion that the 1890s are not underrepresented but rather the 1920s are. Not that I'm about to give up my support for Hughie Jennings by any means.

It might be useful (despite all of that) to "center" HoMer careers a little for scientifically, however, before worrying this much further.
   142. sunnyday2 Posted: July 06, 2005 at 12:57 PM (#1452211)
I would also mention to Brent (#40) that the "added seasons" analysis is ALWAYS negatory for the pitcher with the longer career, well, at least in cases where two pitchers are at least very very roughly comparable. Obviously that wouldn't be true if you were comparing Hilton Smith with Walter Johnson.
   143. Dr. Chaleeko Posted: July 06, 2005 at 01:15 PM (#1452227)
Brent,

I'll be completely honest with you. My feeling about Byrd is completely a gut-shot. I should revisit him in greater detail and work up my own MLEs to see if they agree with Chris's, but the thing is that I see him as like Lyons or Rixey or Grimes: a long-career guy with plenty of innings and effective throughout that career.

One of the things that's really hard for me about grappling with the Negro League pitchers is that with the exceptions of a Paige or a Williams, the pitchers with the longest tenures just don't have long careers (in the way MLB guys did). Byrd's career clocks in around 16 documented years. Cooper's career is really about 18 or 19 years, but we've only got good data on like 12 of them. So they look like 200-215 game winners with 3000 innings instead of 250-300 game winners with 4000 career innings. That doesn't mean that both Cooper and Byrd (and Foster and Brown) would actually have thrown that many innings or won that many games, but someone, perhaps all of them would have.

So then you lay that information on top of the war and on top of what we know (or what we think we know) about arm injuries, and the whole question becomes difficult to answer. It so happens that I've selected Byrd to represent the long, consistent career type of Negro League pitchers. When I work him back up, maybe I'll think it should be Cooper. Or maybe I'll find it should be neither of them???

I'm sorry I can't be more specific about it and offer you stats and backup, but that's just how I feel about it having read and worked through some of the issues on my own.
   144. Chris Cobb Posted: July 06, 2005 at 02:01 PM (#1452301)
Sunnyday2 is more or less correct, I think: the "subtract the careers" method of pitcher comparison just isn't all that accurate, unless you believe that the only value that counts is value above average, in which case it's obvious that Smith (20.5 wins above average) is better than Byrd (14 wins above average).

That's not how win shares sees it, obviously.

Two points to make here:

1) by the win shares estimate system, 34-57 for a .500 team is not below the zero point. This is a pitcher who would earn 11.4 win shares in about 792 innings. That may be below replacement level, but it is not without value in the win shares system. So even if Byrd had exactly Smith's career and then threw four more seasons as a bad pitcher with the above record, win shares would give him more win shares. If Bill Byrd = Hilton Smith + 11 win shares, wouldn't he have to rate higher?

2) Win shares are not linear, so unless Byrd actually was 34-57 for a particular set of seasons, extracting that record from his career is not going to represent his true value fairly. I think a better model is this:

Byrd was 14 wins above an average pitcher in a career of about 3376 innings.

Hilton Smith was 20.5 wins above an average pitcher in a career of about 2579 innings.

An average pitcher earns 195.8 ws in 3376 innings.
An average pitcher earns 149.6 ws in 2579 innings.

Byrd therefore earned 196 + 42 = 238 career win shares
Smith therefore earned 150 + 62 = 212 career win shares

This is not a small difference in career value, but it is not so large as to make it impossible for a voter to prefer Smith to Byrd. Whether you prefer the one to the other depends on a) whether you accept the win shares/WARP view of value (and the two systems are pretty close on replacement level/zero point these days) and b) how you deal with peak value in relation to career.

I'm not going to go into the full seasonal breakdown of value, but suffice it so say that my system, using the seasonal numbers, finds Smith having a small advantage on peak value, but not nearly enough to outweigh the difference in their career value. Smith had four great seasons, two of which were significantly better than any Byrd had. Byrd had three great seasons and four more in which he was significantly above average. I see these two peaks as close to equivalent.

Reviewing Hilton Smith here, perhaps he ought to appear in my top 70 candidates, but the difference, as I see it, between him and Bill Byrd is nevertheless substantial. But then I rate Mel Harder above Dizzy Dean and Lefty Gomez. I suggested earlier that I believe that Harder/Gomez are roughly comparable, as a pair, to Byrd/Smith. If you prefer Lefty Gomez (or, I suppose, Dizzy Dean) to Mel Harder, then you may still find my preference for Bill Byrd over Hilton Smith hard to understand.

As to Smith possibly being of major-league quality earlier in his career, I don't believe the record supports such an assumption. He broke into organized baseball in 1932, with the Monroe Monarchs in the Southern League. That league was counted as a top league that season because several teams from the defunct Negro National League joined it. Smith may have been with the team, but he was not being pitched against league opponents: Holway has no data for him, nor does Macmillan. So, he was not a major-league quality pitcher in 1932.

Of 1933-1935, Riley's bio says, "For the next three years he played with teams of lesser quality, including the New Orleans Crescent Stars, and pitched in the National Baseball Congress in Wichita in 1935-36, registering a perfect 5-0 record, indicating tha the was ready to step up to the black major leagues. In the fall of 1936 he began barnstorming with the KC Monarchs."

Maybe this bio is wrong, but it gives us an unusually clear account, for NeL playe bios, of Hilton Smith's development. I see it as justifying a view that Smith reached major-league replacement level sometime in 1936, and that he got his call-up at about the usual time -- late in the season. I have given him a full season of MLE credit for 1936 as an average pitcher; I see that as giving him some benefit of hte doubt, actually, about when he became a major-league quality pitcher.

If he were to get some MLE credit for 1935 (say 5-7 win shares for a part-season in the majors), that would not be enough, for me, to close the gap with Byrd.
   145. TomH Posted: July 06, 2005 at 02:37 PM (#1452394)
Bucky Walters post-season record is quite good, and since we seem to have little discussion here on World Series performance and how much that affects our rankings, I thot I'd detail it a bit.

1939 WS Cinci-NY. Won by NY 4-0.
Walters started and lost one game when his opponent tossed a shutout. He lost another game in relief; came on for a save but an error by shortstop Myers allowed 2 unearned runs.

1940 WS Cinci-Det. Won by Cinci 4-2.
Walters pitched a CG win (5-3) in game 2, doubling and scoring one run. In game 6, Walters threw a shutout and homered, allowing the Reds to get to game 7 and win the next day.
   146. Jeff M Posted: July 06, 2005 at 02:38 PM (#1452405)
I had Vance #3 in his first year of eligibility. If Walters had been eligible that same year, I would have had him #49, tied with Hack Wilson. Here's why (in summary, not mathematical, terms):

1. Vance lost a lot more wins due to poor run support (see Chris' RSI). In addition, Vance got poor defensive support and Walters got good defensive support (according to DERA on BP). I make adjustments to their win totals for these things, and Vance benefits more than Walters in the adjustments.

2. Vance's ERA is a little lower, but he did it in a context where league ERAs were about 1/4 of a run higher than in Walter's playing career. As a result, Vance is almost 50% better in linear weights (runs saved above average).

3. Vance has a much better winning percentage, and much better Wins Above Team.

4. Walters does not have an impressive ERA+ for his career, measured against HoFers. He actually had three years in his prime that were significantly below an average pitcher.

5. Vance's win shares per start is outstanding. Walters' is very good.

6. Vance is significantly better than Walters in WARP 1 on 3-year peak, 5-year consecutive peak, 7-year peak and WARP per start.

7. Adjusting for run support, defensive support and ERA context, Vance and Walters both get a boost in their Fibonnaci score, but Vance matches with the best HoFers, while Walters' lags significantly.

I wouldn't call myself a "Vance lover" -- in any context. :) I don't have pet players. I feed them into my syste (which is flawed like everyone else's) and the scores/rankings pop out.
   147. Chris Cobb Posted: July 06, 2005 at 03:40 PM (#1452553)
A shorter career NeL pitcher who I see as much more comparable to Bill Byrd than Hilton Smith is Leroy Matlock.

Here's his career line in my MLE stimates right now:

2816 IP, 3.96 DERA, 114 DERA + 182.2-141.5 sn W-L .563, 20.4 wins above average.

By my estimates, he has as many wins above average as Smith, plus another 240 league-average innings, and a stronger contiguous peak than either Smith or Byrd: he was, in actual games, 45-7 over a 4-year period. 14-3, 17-0, 9-3, 5-1. That counts 1937 Santa Domingo as Negro-League play, which may be overestimating the quality of competition there, but it's still a very impressive stretch of seasons, even pitching for a great team. He also has another five average or better seasons outside this peak.

Since Matlock is eligible now (and has been since 1948), I figure it's fair to put this on the ballot discussion thread, but I'd like to post the full set of data for Matlock on a thread of his own. John, would that be possible?
   148. TomH Posted: July 06, 2005 at 03:49 PM (#1452581)
The contrast between Vance and Walters you made, Jeff, which will also apply to Ruffing and Rixey, will be a test of our buy-in to the data which says the NL of 1940 was much stronger than the 1920s NL; a big jump particularly fromthe late 1920s thru the mid 1930s. BP's WARP2 gives Ruffing and Walters gains for their work, and discounts Rixey, Vance, et al. I believe the 1970s Palmer/Thorn/Adams study (Hidden Game of Baseball) reached somewhat the same conclusion. Our collective wisdom on 'league strength' issues will play a big role when we get to electing some of our backlog soon.
   149. Dr. Chaleeko Posted: July 06, 2005 at 04:53 PM (#1452736)
Chris, the question about 1937 and the quality of play in the Trujillo league is an interesting one.

I don't remember all the details precisely, but but I recollect reading in Riley and Holway that there were basically three teams in the league. Paige was given a suitcase full of money by Trujillo to hire the best NgL players for the dictator's team. Meanwhile, the other two teams also ramped up with NgLs too. The story goes that the pennant was hotly contested and the Trujillo had to win one game on the last day of the season to win the Island's championship.

If all this is true, then given a very small league with a tight race that includes a mother load of NgL regulars, I'm thinking that's a got to be a very competitive league environment, probably as competitive or nearly as competitive as Mexico in the early forties.

But again, I could have my facts wrong, so someone please correct me as needed.
   150. Brent Posted: July 07, 2005 at 04:21 AM (#1454457)
sunnyday2 wrote:

I would also mention to Brent (#40) that the "added seasons" analysis is ALWAYS negatory for the pitcher with the longer career, well, at least in cases where two pitchers are at least very very roughly comparable.

While I agree that the “added seasons” approach isn’t terribly reliable as a general approach to evaluating pitchers, in the case of Negro League MLEs it seems we don’t have much else to go on. Chris has observed that the single season MLEs are noisy and shouldn’t be given a lot of weight, so mostly it comes down to comparing career MLEs of pitchers with different career lengths. In the case of Byrd and Smith, I was surprised to see most voters favoring Byrd, even though it was clear that Smith had a better peak and it wasn't clear that Byrd's career value was very much better.

If we are comparing the career records of two pitchers with different career lengths, somehow we need to say that if pitcher A has x more wins and y more losses than pitcher B, his record is better/equal/worse. What are the standards? Fibonacci win points says that it’s a winning percentage of .414 – that is, in comparing two pitchers if the extra wins of the pitcher with the longer career divided by his extra decisions exceeds .414, then he will be credited with more Fibonacci win points. In Politics of Glory, James called this .414 threshold “replacement level.” From the numbers Chris posted, it appears that the zero level for win shares is quite a bit lower, about a .332 winning percentage.

Also, the added seasons analysis is relevant because sometimes it is actually approximately applicable to the seasons of the pitchers we are comparing. Byrd versus Smith is an interesting example – if we compare just their MLEs for the seasons the have in common, 1936-48, their records were Smith 174-123, Byrd 189-146. In his extra seasons 1933-35, Byrd was 19-34. If we think their records were comparable for 1936-48, then the relevant question is whether Byrd’s extra 19-34 from his three extra season were enough to push him ahead of Smith and into serious ballot consideration. (This importance of this question is amplified because Smith was also pitching during this period, but in leagues that are not as well documented.)

Chris Cobb wrote:

1) by the win shares estimate system, 34-57 for a .500 team is not below the zero point. This is a pitcher who would earn 11.4 win shares in about 792 innings. That may be below replacement level, but it is not without value in the win shares system. So even if Byrd had exactly Smith's career and then threw four more seasons as a bad pitcher with the above record, win shares would give him more win shares. If Bill Byrd = Hilton Smith + 11 win shares, wouldn't he have to rate higher?

Not if we want to measure value above replacement level. I think even Bill James concedes replacement level is higher than the zero point for win shares. I think it would be difficult to imagine a pitcher being able to maintain his job after earning only 11 win shares in 800 innings, so the difference between Byrd and Smith seems very likely to be below replacement level. On the “pennants added” thread, I believe that Joe Dimino decided to use a replacement level for pitchers of 5.7 win shares per 220 innings (or 20.5 win shares for 792 innings). It’s hard for me to see crediting a player for the HoM with career value for seasons that are below replacement level.

Byrd was 14 wins above an average pitcher in a career of about 3376 innings.

Hilton Smith was 20.5 wins above an average pitcher in a career of about 2579 innings.

An average pitcher earns 195.8 ws in 3376 innings.
An average pitcher earns 149.6 ws in 2579 innings.

Byrd therefore earned 196 + 42 = 238 career win shares
Smith therefore earned 150 + 62 = 212 career win shares


I don’t see where the 20.5 wins above average comes from. Smith’s support neutral MLE record was 173.7-122.8; shouldn’t that be 25.5 wins above average?

And wouldn’t that raise Smith’s career win shares to 227? (Again a difference of 11 from Byrd).

Of 1933-1935, Riley's bio says, "For the next three years he played with teams of lesser quality, including the New Orleans Crescent Stars, and pitched in the National Baseball Congress in Wichita in 1935-36, registering a perfect 5-0 record, indicating tha the was ready to step up to the black major leagues. In the fall of 1936 he began barnstorming with the KC Monarchs."

Chris, are you familiar with the material from http://www.pitchblackbaseball.com/northdakotabaseball.html on Smith’s 1935-36 seasons in North Dakota? Unfortunately, the information is anecdotal rather than statistical and in any case it would be difficult to evaluate the level of competition. But it’s clear that during those two years Smith was a respected member of a pitching staff that also included Satchel Paige, Chet Brewer, Double Duty Radcliffe, Barney Morris, and Ted Trent.

I'm not trying to argue for Smith over Byrd; in fact, the more I look at them, the less sold I am on either of them. I just am trying to understand how several influential members of the electorate decided Byrd was so much better than Smith, when to me he doesn't look better at all.
   151. Brent Posted: July 09, 2005 at 02:15 AM (#1460076)
Jeff M wrote:

Vance lost a lot more wins due to poor run support (see Chris' RSI). In addition, Vance got poor defensive support and Walters got good defensive support (according to DERA on BP). I make adjustments to their win totals for these things, and Vance benefits more than Walters in the adjustments.

My concern is that at least part of Vance's poor runs support (90.47) was self-inflicted (OPS+ of 10), while Walters, who appears to have had neutral run support (100.06) was a good hitter for a pitcher (OPS+ 69), so the run support he got from his teammates must have been well below average. I think the adjustment we would really like to make would be based on their run support excluding their own hitting, but I don't know how to make that adjustment. Has anyone figured out how to do it?

This is what's keeping me from using Chris J's run support indexes as a direct input into my own pitcher evaluations (though I do look at the numbers and sometimes will make an informal adjustment).

By the way, I agree that Vance was better than Walters -- I just think that if you take into account the differences in batting skill, the overall difference between the two is not large.
   152. Chris Cobb Posted: July 09, 2005 at 03:08 AM (#1460197)
On poor hitting affecting RSI:

I've done some modeling with runs created, and I've found that poor hitting does meaningfully affect RSI, but the effect is not huge. As a rule of thumb, I assume that for a pitching season in which the pitcher's OPS+ is 16-29, an average-hitting pitcher would have had an RSI 1 percentage point higher, for an OPS+ of 1-15, 2 points higher, and for an OPS+ 0 or less, 3 points higher. This adjustment would apply equally to career, but I adjust season-by-season. Obviously, these estimates could be more exact, but as it's pretty clear that the RSI difference between an average-hitting pitcher and a pitcher who adds nothing but outs is around 3 percentage points, I haven't found it worth it to fine-tune the measures any further.

I haven't worked much on estimates for seasons of OPS+ above 30, because I just add batting win shares for pitchers who earn them. I figure the 30-50 OPS+ range is more or less break-even for pitchers, with opportunities for accruing bws beginning above that.
   153. Esteban Rivera Posted: July 09, 2005 at 06:37 PM (#1460764)
To further add to Doc's post 49 about the 1937 Dominican league play, I have been reading Mark Ribowsky's negro leagues' history book. In the book there are more details about who were some of the players that were there.

Just before the defections started in the 1937 season, the famous Josh Gibson holdout occurred. Gus Greenlee did not want to give Gibson a raise and started badmouthing him in the press, hoping Gibson would back down and return. Greenlee's strategy back fired when Gibson did not back down, and he was forced to trade him to Cum Posey. The trade was $2,500, Pepper Bassett and Henry Spearman for Gibson and Judy Johnson. However, Johnson was so offended that Greenlee would trade him that he promptly retired, although Greenlee convinced him to temporarily unretire so the deal could go through. Soon after, representatives from the Cuban, Mexican, and Dominican leagues arrived to recruit players. (The book mentions Puerto Rico, but I believe they only had semi-pro leagues at the time). The leagues had decided to switch from winter to summer and knew they could entice blackballers into playing for them. Martin Dihigo, an agent for the Cuban league, signed Showboat Thomas, Spoony Palm, Bill Perkins and Thad Christopher. Soon after, Satchel Paige arrived for spring training, was signed to play in the Dominican for $6,000 and was also hired to serve as an agent.

The 1937 Dominican season was probably the professional baseball season with the heaviest political overtones ever played.
There were only three teams:
Los Dragones de Ciudad Trujillo, owned by the Dominican's dictator Rafael Trujillo; San Pedro de Macorís and Santiago de Los Caballeros were the other two teams.
Both of them were owned by some of Trujillo's still standing political opponents. Trujillo's aim with his team was to use baseball to improve his public image in the country by associating his name with the country's best team. His opponents wanted to give the same boost to their image so they could be in a favorable position to take over when the revolution came. As you would imagine, the owners could not afford to lose and spared no expense in importing players.

The players brought in were the following:

Ciudad Trujillo - Satchel Paige, Cool Papa Bell, Leroy Matlack, Sam Bankhead, Harry Williams, Herman Andrews, Josh Gibson, Pedro Cepeda and Lazaro Salazar. The team also included heralded locals Enrique Lantigua and Amable Sonlley Alvarado.

San Pedro - Pat Patterson, Showboat Thomas, Spoony Palm, Johnny Taylor, Clyde Spearman, Bertram Hunter, Thad Christopher, Luis Tiant and Martin Dihigo.

Santiago - Chet Brewster, Spoon Carter and George Scales.

The championship series was played between Trujillo and San Pedro. The series went seven games, with the seventh game being played with Trujillo's troops along the first base stands and the other owner's troops along the third base stands. As soon as the game was over, the players raced to their hotel, packed their bags and hurried to catch the first boat back to the states. Meanwhile, the powers that be in the negro leagues had given them lifetime bans for jumping. On their return, the outlaw players formed the Trujillo All-Stars and toured for a while, until the bans were lifted.
   154. Dag Nabbit is part of the zombie horde Posted: July 09, 2005 at 07:51 PM (#1460909)
I think the adjustment we would really like to make would be based on their run support excluding their own hitting, but I don't know how to make that adjustment. Has anyone figured out how to do it?

After I gave my presentation on RSI at SABR34, Sean Forman apparently mentioned to someone (I heard it third hand) that this could be accounted for without too much difficulty based on runs created (I think that was the metric).

Here's the problem though -- I'm not looking at how many runs each pitcher created, but at the difference between an indivudual pitchers's runs created and what a league average pitch would create. Otherwise, though I can adjust for a pithcer's bat, I'd de-lign it off its 100 resting point. And I'd rather know what the average is than make this adjustment. After all, for most pitchers the hitting difference is minor.

That's the problem, though. I'd need to know what an average pitcher did each year in each league. The frick I want to figure that out. Too much heavy lifting for too little payoff. Instead, I created my Pitcher OPS+. You can guesstimate off of that how much a pitcher gained/lost due to his run support. (The first bunch of lists are just the 192 guys on the site, but below that are the main lists of all SP with at least 200 GS going decade-by-decade.

When Vance played the average pitcher had an OPS+ of around 30, his was 10. Walters scored a 69 when the average was around 25.

Does that make up for the 10-point gap? I doubt it. I'd tentatively guesstimate Vance's bat-adjusted RSI around 92, and put Walters down to about 95-96, but those really are just WAGs.
   155. John (You Can Call Me Grandma) Murphy Posted: July 11, 2005 at 01:08 AM (#1462788)
Anybody want a thread for Max Manning?
   156. Dr. Chaleeko Posted: July 11, 2005 at 12:45 PM (#1463348)
Probably a good idea, John.
Page 2 of 2 pages  < 1 2

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
The Id of SugarBear Blanks
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 0.8985 seconds
49 querie(s) executed