hot topics

John, I think you might want to re-title this thread "1958 Ballot Discussion," yes?

Somehow, it got knocked off. Thanks for spotting it, Chris.

A note for the Commish: If it's at all possible, it would be great to get the updated Pennants Added for this election. I can use all the help I can get. (And if there's anything I can do to help out, just let me know.)

Is Boudreau an SNT beneficiary this year? He's a fine player, but he comes in at #4, Bobby Doerr at #21, and Joe Gordon at #33??

No way there is that much difference between these three players, especially between Boudreau and Gordon.

A few quick notes: Gordon played in only 100 fewer games than Boudreau, despite missing two seasons in WWII. Both had career OPS+ of 120, with Boudreau getting in a couple of peak seasons in 1944 and 1945, while Gordon lost two peak seasons. (And OPS+ accounts for the Yankees thing.) Boudreau was an A+ shorstop, while Gordon was an A second baseman. Slight edge to Boudreau there, but Gordon has a slight edge on hitting and durability.

I'm not saying that Gordon is necessarily better than Boudreau (though I have him ranked higher), but I just can't see how Boudreau makes 32 ballots and Gordon makes 10.

Our first three-HoMer election!

1958 (August 22)—elect 2
1958 (August 14)—elect 2

A couple 'elect 2's are still up there.

Clay Davenport has an article at BP titled: "An Objective Hall of Fame: Part One". I'm not a subscriber, but am curious as to what he has to say.

Boudreau was an A+ shorstop, while Gordon was an A second baseman.

Depends on whose word you take on this. BP has him certainly as an A 2B pre-war, maybe even better. THey also have him as a giant turd defensively post-war. Boudreau was great at SS by all accounts for his entire career.

Boudreau's OWP is .603, Gordon's is .583, plus Boudreau was his era's Ozzie Smith.

I would question how BABE HERMAN is so much higher than Gordon, Childs, Monroe, even Doyle....

"Boudreau was an A+ shorstop, while Gordon was an A second baseman. Slight edge to Boudreau there,"

I consider this a large advantage.

Good point Devin. I will see what I can do. I'm so far behind at this point, it will be considerable work to get updated, but I should have a decent amount of time available this week, after tomorrow.

I would certainly agree with you on Babe Herman KJOK.

"Clay Davenport has an article at BP titled: "An Objective Hall of Fame: Part One". I'm not a subscriber, but am curious as to what he has to say."

I'm not a subscriber either, but someone sent me the article, so I read it.

There are some striking similarities. Same number of HoMers and HoFers (he's not doing Negro Leagues). Annual elections, though he's starting in 1936. 5-year after retirement you are eligible, but we both stole that from the Cooperstown. He's even using team seasons to set the number of electees, though not according to our algorithm.

Differences. He has only one voter, and it's WARP3. He's doing it strictly mathematically, using WARP3, but not straight. He's counting peak years as 14x, 2nd best year at 9x, 3rd best at 8x, etc.. All years after 10 get multiplied at 1. He's electing as many as 5 per year until the Hall is 'full' (1954), then he elects one or two per year. He'll have the 15-years of eligibility rule, after he's met the quota, so starting in 1959.

I'll reserve comment for now. Both on his system on it's own merit (pun intendend) and on the similarities to our project.

I like his method for adjusting for military service. He basically removes the missing years for every player in history and then checks to see where the player would rank. He then adjusts the player based on the gain that group (player at that rank + or - 2) had over the group around the player that missed time. That's pretty darn cool, but impossible to calculate without an easily manipulatable database of WARP scores . . .

"I would question how BABE HERMAN is so much higher than Gordon, Childs, Monroe, even Doyle.... "

Were you actually referring to Billy Herman?

If so, then I disagree. Billy Herman is clearly the best of this group, especially if you give war credit. His peak isn't quite as high, but he's way, way ahead on career value. He'd have about 9500 PA at that point with rate stats not terribly far behind.

Hard to believe our first elect 3 comes in a year with no obvious newbies. The task before us is easily defined, however.

Are Herman, Boudreau and Hack the best 3 players available? I hope everyone will seriously consider the question rather than just copy and paste last year's entry.

Or are Medwick and Ruffing, the next two holdovers, perhaps more deserving? (Jennings is next but trails by enough of a margin that, even though he is likely to be my #1, I don't presume that he is really in the mix for induction in 1958. But of course it was very recently that Earl Averill looked like a '58 selection. Things happen.)

I love middle IF and I think we have undervalued their skills a bit overall. I will have to take that back if we indeed elect Herman, Boudreau and Hack. I personally agree that Herman is the best 2B among a strong group, but I am skeptical that Boudreau and Hack are the best at their positions. And the comparison of Boudreau and Gordon (elsewhere) is very provocative.

Boudreau's career record is a bit weak even without a WWII discount but his peak is pretty darn good and I'm usually a peak/prime voter. But if it's peak you want, I can't see Lou over Hughie or Dobie Moore.

nice 'state of the ballot' discourse, sunny!

I have Herman, Hack, Boudreau and Ruffing all in my top 8, so I would be OK with any of them.

The guy I have off-ballot is Medwick. A few comments on Ducky Wucky:

If his 1937 (1055 OPS) and 1943 (686 OPS) years had been averaged out instead of one Triple Crown year and one poor one, wouldn't he exactly be Bob Johnson? And Indian Bob only got on 4 ballots.

He won the TC in 1937, but only by tying Mel Ott in HR, with an unusually low HR total of 31. DiMaggio hit 46 in the AL that year. Greenberg would hit 58 next year. Ott typicaly hit 35 HR a year; he was just a few lower in 1937. Medwick's 31 would have finished 7th in the AL. It's a veyr good year, but not a historic one.

Medwick led in OPS+ only once. Never again in the top 3. Medwick's 37 was his only year of OPS+ over 160.

For a guy who basically is only a hitter, it appears he has a very good prime and a very good career. We have LOTS of others who might fit most of these really-good-stick criteria; besides Johnson, there is Averill, Sisler, Browning, Oms, Cravath, Arlett, Beckley, Charley Jones, Klein, Keller, Roush, Hack Wilson.

It seems to me we are ready to induct someone based on one great year. Yes, that should count a good deal, but I'm not ready to say changing two good seasons into one poor one and one super one makes someone rise from the HoVG to the HoM.

Prelim. My PHoM includes someone who was inducted into the HoM 9 years earlier (breaking Elmer Flick's record).

Willard Brown, Stan Hack, Carl Hubbell

HoM Prelim

1. Willard Brown
2. Red Ruffing
3. Stan Hack
4. George Van Haltren
5. Joe Medwick
6. Mickey Welch
7. Jake Beckley
8. Billy Herman
9. Biz Mackey
10. Cool Papa Bell
11. Eppa Rixey
12. George Sisler
13. Tommy Leach
14. Hugh Duffy
15. Dobie Moore

16-20. Trouppe, Boudreau, Rice, Roush, Averill
21-25. Powell, Ryan, Childs, Doerr, Mullane
26-30. White, Griffith, Jennings, H.Smith, Streeter

Prelim

1. Jennings
2. Moore
3. Medwick
4. Sisler
5. Bond
6. Waddell
7. Mendez--all are PHoM

7a, Wells
8. Herman
8a. Lyons--these are my 3 prelim PHoM 1958

9. Willard Brown
9a. Beckwith
9b. Harry Stovey
10. Boudreau--these 4 are also in PHoM 1958 consideration set, Yes, Boudreau and Gordon are pretty comp, as stated elsewhere, but that is more of a minus for Lou than a plus for Gordon right now. I mean, Boudreau is surely closer to Gordon than he is to Herman, even for me, a peak/prime voter (but ~1/3 career. Some have argued that Medwick's case is based on one year alone. The same can be said (should be said) for Boudreau. Take away 1948 and what do you have, really? I'm also going to include Hack in my PHoM eval just in case I am missing something there. I also have to say re. Stovey that I never had him above Browning or C. Jones. He is the top HoM/not PHoM among position players but I may have him too high here, so then I guess I also have to include Browning in my PHoM consideration (Jones is already in).

11. Joss
12. Williamson--already PHoM
13. Gordon
14. Browning
15. Charley Jones--already PHoM--but it could just as well be:

16-19. Doyle, Doerr, Klein, Cravath
20. Duffy

Hack is about #27, Ruffing is about #32. Among the 5 runners-up in '57 I would prefer Medwick, Herman and Boudreau despite what I said about Lou above.

My consensus scores are due to set a new negative record. I note that NONE of the consensus top 8 returnees are in my top 15. The highest ranked returnees that are on my ballot are Rixey, Sisler, Griffith and Beckley, all of whom I would strongly urge the electorate to consider carefully before enshrining the mediocrities (except Jennings, to be fair) that are currently 3-10

WHAT a poor year for our first elect-3, and no more to come until 1972.

Re. Medwick's "tainted" Triple Crown, I would just note that Carl Yastrzemski "only" won the Triple Crown by tying Harmon Killebrew for the HR title, but for some reason they still called it a Triple Crown. Two-and-a-Half if you ask me ;-)

Various points on Boudrea vs. Gordon:

"Boudreau was an A+ shorstop, while Gordon was an A second baseman. Slight edge to Boudreau there,"

I consider this a large advantage.

To quantify:

Gordon 5.19 ws/1000 at second
Boudreau 6.49 ws/1000 at ss

This gives Boudreau about a 1.6 ws/season edge over Gordon, given their typical games played.

The fine print: Boudreau played 1/4 of his time at 3B and 1B from 1949 on. Gordon played 30 games at 1B in 1941.

The WARPed view:

Depends on whose word you take on this. BP has him certainly as an A 2B pre-war, maybe even better. THey also have him as a giant turd defensively post-war. Boudreau was great at SS by all accounts for his entire career.

To be precise, they show him as a below-average second baseman starting in 1947, his age 32 season. It should also be noted that they show Boudreau's defensive value dropping from excellent to average in 1949, his age 31 season. Average is better than below-average, but Boudreau was also not a full-time player after 1949. WARP gives him 9 years of great defense, 1940-1948. It gives Gordon 7 years of excellent defense, 1938-43, 46, with the 2 war years interrupting his run.

What this suggests is that Gordon is _very_ close in value to Boudreau if you give him fair credit for 1944 and 1945 and if you give Boudreau a discount for the war years. Even WARP, which hates Gordon's late-career defense, shows this.

On offensive winning percentage:

Boudreau's OWP is .603, Gordon's is .583, plus Boudreau was his era's Ozzie Smith.

How, if at all, do these career rates adjust for WWII? Is anything shaved off for Boudreau in 1944-45? How would Gordon's rate be affected if two good years, equivalent to his avg. rate 1941-43 were added in for the war years? If the effect of the war years is not being addressed, a career rate stat like this will exaggerate the difference between the two players.

I'm not trying to run Boudreau down here (much). But a lot of the commentary I've seen on Boudreau has been on the order of "A+ defense! 120 OPS+! Wonderful!" without sufficient effort at contextualizing and quantifying the value of those undoubtedly wonderful features of Boudreau's career.

If Boudreau is so wonderful, then Gordon and Doerr are almost equally wonderful.

I prefer all three to Medwick, fwiw.

If Boudreau is so wonderful, then Gordon and Doerr are almost equally wonderful.

I prefer all three to Medwick, fwiw.

Without a doubt, Chris. I honestly don't understand the love for Ducky Wucky.

58 Scatterbox Prelim

---- “clearly in” “borderline” “HoVGood”
-C --------------Mackey Bresn/Lomard/Schang
SS ----Boudreau ------Sewell ----Jennings
2B –Herman---------the 5 secondbasemen
3B ----Hack ----McGraw Traynor/Leach
1B ---------------------Chance/Sisler/Beckley
OF ---------VanHaltren ----B Johnson Oms
OF -----------CP Bell, Averill Medwick Cravath
-P Griffith Walters------Rixey –--Mendez
-P ---------- Ruffing --------Dean/Welch
-P -------- Ferrell --------------Waddell


as for Boudreau/Gordon:
Lou's career WS per year - 27.3
Joe's career WS per year - 21.1
WWII diffs won't come anywhere near making up that difference

To be precise, they show him as a below-average second baseman starting in 1947, his age 32 season. It should also be noted that they show Boudreau's defensive value dropping from excellent to average in 1949, his age 31 season. Average is better than below-average, but Boudreau was also not a full-time player after 1949. WARP gives him 9 years of great defense, 1940-1948. It gives Gordon 7 years of excellent defense, 1938-43, 46, with the 2 war years interrupting his run.

The general jist of this is fair, but some minor clarifications.

First, it's worth pointing out re: age 31 vs. age 32 that Boudreau's excellent defense started a year earlier than Gordon's.

Second, a countdown comparison of FRAA:

G: 21, 18, 14, 10, 9, 8, 1, -10, -13, -14, -18
B: 22, 22, 19, 15*, 13*, 12, 12, 12, 10, 3, 3, 1, -1

The two years Boudreau played that Gordon did not are starred. Even if you completely strike them from the record, by BP at least, Boudreau is a much higher quaility defender *before ever taking SS vs. 2B into account*. Once you do take that into accout, I would contend that the two are no longer all that close at all in overall value.

-------------------------

FWIW, I think we all agree on Medwick. :) Not that I hate him, but I see no way that he is better than Sisler, who had a simlar career shape, but was obviously better in his prime.

Here's my prelim:
1-Billy Herman-Fantastic from 32-40. Good during the war, but I took a little credit away for 43-45.
2-Red Ruffing-Nice long(but not spectacularily high) peak with the Yanks, plus 1 strong year with the Sox. More Career value than Rixey, even giving Rixey wartime credit. I'm willing to say he's the best pitcher on the ballot right now.
3-Earl Averill-With PCL credit, you can add career value to an already nice peak.
4-Jake Beckley-He's not inner-circle, but definitely 'in' when the back-log-clearing years come around.
5-Eppa Rixey-Great Long Career; long enough that his near total lack of peak doesn't kill him.
6-Bobby Doerr-Strong peak and career player. His war years seem to even out well.
7-Joe Sewell-Slick fielder, above average hitter. I've been underating him. I can't keep Hughie over him because his career value is just too strong.
8-Dick Redding-One of the great Negro League pitchers
9-Lou Boudreau-Strong Peak. I think most are overrating him. I'm placing him here for now.
10-Joe Gordon-Nice Peak. Gets a little war credit for 44-45 and a slight downgrade on '43; although '43 is in line with the years before so my downgrade is very slight.
11-Rube Waddell-Nice peak
12-Addie Joss-Great Peak
13-Hughie Jennings-5 phenomenal years. Its enough, I think, but he'll need to wait.
14-Stan Hack-Pretty slick fielder and a good hitter. I took a little away for his 43-45 numbers or he would be higher.
15-George Sisler-The peak is hard to ignore.
16-Indian Bob Johnson-Solid for the years he was in the MLB. Deserves a couple years of prior credit as well, I think.
17-Cupid Childs
18-Sam Rice
19-Wes Ferrell-He's not Grove, Hubbell, Ruffing, or Lyons, but he's 5th this period. I upped him in 29-31 for pitching against much tougher offenses than Grove did.
20-Mel Harder-Nice Prime from 32-39, but not a lot else.

Lou's career WS per year - 27.3
Joe's career WS per year - 21.1

This rate for Gordon's career is wrong. Once again, the NBJHBA proves utterly UNRELIABLE on career rate stats.

The career rate for Boudreau is correct 277/1646*162 = 27.26

The career rate for Gordon is INCORRECT 242/1566*162 = 25.03

25.03 is much closer to to 27.26 than 21.1.

Moreover, if you give Gordon war credit and give Boudreau a competition discount for the war years on the model I use, their career rates shift to

Boudreau 26.57
Gordon 25.45

Once again, Gordon is _very close_ to Boudreau.

A couple things to say after a four-day break.

1) As mentioned elsewhere, 451 points for Boudreau and 123 for Moore is really bogus. I'm coming out go-guns on this one because I think the discussion of Moore, the statistical information on him available through our usual sources, and the oral history on him support him receiving comparable votes to Lou Boudreau, yet our electorate as a whole is convinced that Boudreau is somehow a few orders of magnitude better.

No way.

I can see how Boudreau's better than Gordon, Doerr, Sewell, and the rest, but Dobie Moore was among the best five players in the NgLs during his NNL career: a big bat, a strong glove at SS, contributing to a winning team. If the uncertainty about his time with the Wreckers is suppressing his vote totals relative to Boudreau, then every one of us needs to return to the Moore thread and re-read it and reconsider Moore relative to Boudreau.

2) You all may be surprised to hear this, but Pete Browning is about to enter my ballot. I've done ample reconsideration on him, and I believe that I have unfairly excluded him on my ballot due to my own uncertainties regarding the AA's early years. After review, he's going on the ballot ahead of Averill/Poles/Roush.

3) The news on Browning is bad news for GVH who is once again sliding down my ballot. He may, in fact, slide off of it. It'll be very close.

Ah, like a bad penny, Joe Gordon's incorrect WS/162 returns. I wonder how many have Gavvy Cravath too low based on the same Bill James error.

Doh! Yes, I took the 21.1 right from the NBJHA. If the real number is 25.0, Gordon moves from off my ballot to 12th-ish.

Any other glaring numerical mistakes / misprints of which I should be aware?

can someone point me to the threads/links where Ruffing's RSI and his MOWPs are given and discussed?

If Bill James were voting this years, this might be his ballot, based on NBJHBA rankings:

1 - Cool Papa Bell
2 - Joe Medwick
3 - Stan Hack
4 - Billy Herman
5 - Dizzy Dean
6 - Lou Boudreau
7 - Joe Gordon
8 - Wally Berger
9 - Earl Averill
10- Charlie Keller
11- Bobby Doerr
12- Edd Roush
13- Carl Mays
14- Wes Ferrell
15- Pie Traynor
or- Tony Lazzeri

I would suspect he would have Biz Mackey (#2 NL catcher in the NBJHBA) in the top 15, based on that ranking and his comment about the high quality of Negro League catchers. Obviously, not above the guys listed in his top 100, but somewhere after that. Otherwise looks about right. His timelining makes me look like sunnyday. :)

Absolutely right on, Mike. A couple other NeLers could be here, as well.

hey, if you think BJames timelining is strong, wait until you see my take on the 1950s-60s NL! :)

seriously, I don't 'timeline' as much as the NBJHA does from 1870-1950, but I DO see a good jump upon the integration of the NeLers, and since by far most of the quality players went to the National League, I would put the 1960 NL on par with the level of play in 1995. Instead of the BJames straight-line method,
..../
.../
../
./
/
mine would look more like
....._
..../
...|
._/
/

2B/3B/SS:

Our consensus, from 1957:

3. Herman (583)
4. Boudreau (451)
5. Hack (417)
8. Jennings (325)
20. Childs (162)
21. Doerr (155)
22. Sewell (154)
28. Moore (111)
30. Leach (100)
33. Gordon (96)
38. Doyle (64)
40. McGraw (47)
44. Monroe (38)
53. Traynor (17)
63. Lundy (8)

My own vote:

3. Hack
5. Doyle
7. Boudreau
11. Sewell
18. Herman
20. Childs
21. Gordon
26. Doerr
(Traynor recently dropped from my top 30.)

The person I should take another look at is Moore, but there's a lot you have to accept to put him up there. I'm not getting anywhere arguing for Doyle - but the man could hit. I do understand why Boudreau ranks high.

But the one I'm having trouble with is Herman. I do not see what distinguishes him from Sewell, Gordon, and Doerr - or for that matter, from Lazzeri.

On Herman:

I wrote up a detailed comparison of Herman vs. Gordon and Doerr on the second base positional threads. (I've never mastered the art of linking threads here, but you can find it by clinking on the discussion of position players on the HoM homepage and then on the second base link).

The most obvious thing Herman has going for himself is career length. With war credit, he has 15 years to Doerr's 14 to Gordon's and Sewell's 13 to Lazzeri's 12. Assuming their peaks are the same, career value is a tie-breaker, yes? And their peaks aren't the same: Herman's is the best of the lot: not by much, but it's still the best.

OCF, how does it appear to you that Herman is not above S, G, D, L in both career and peak value?

Tom, the correct anweer re. BJ WS/162 data is that you should compute them yourself every single time you want to quote one. I couldn't possibly list all the errors. They are endemic.

As to Dobie Moore, he is #2 on my prelim and unlikely to move. My tough choices are all about #8-10, which actually includes 7 players including HoM/not PHoMers.

We know that Moore had 6.5 years in the NeLs, during which time he was clearly one of the top players in the league, more Oscar Charleston than anything. i.e. .365 with power and an outstanding SS with a great throwing arm, specifically. He was forced to retire while still in his prime if not at his peak.

I guess a comp might be Hughie Jennings or better yet George Sisler if he had retired due to his eye trouble. Or maybe Kirby Puckett. I mean all olf these in terms of career shape.

Even I would agree that 6.5 years does not and cannot make a HoMer. But we know for a fact that Dobie played at least 3 and possibly as much as 6-7 years with the Wreckers. Whether he was recruited to play with them or not, I don't think we know. But they appear to have been a AAA caliber team which, in fact, beat up on PCL teams as often as not.

He apparently was an unpolished player when he joined the Wreckers, and we see him played 3B, not SS, and batting 7th during at least one period with them. So he did not have a 10 year peak/prime but a 10 year career with about 6.5-7 years of peak/prime.

Ultra-careerists will not like this, fair enough. But if you look at the WS analysis on his thread and also at the performance data, you will see that for 6-7 years he appears to have been Joe Cronin or Ernie Banks. Add 3-4 years as a young, developing player and that is good enough for me.

In fact, the final comp for Dobie is perhaps Banks had he, too, retired instead of moving to 1B.

"The fine print: Boudreau played 1/4 of his time at 3B and 1B from 1949 on. Gordon played 30 games at 1B in 1941."

You can't be serious right? Boudreau played 76 games in his entire career that weren't at SS or C (where he played 3 games). 1/4 of his time from 1949 makes it sound like he was turned into Eric Hinske or something. He played 57 G and 3B, 16 at 1B and 3 at 2B in his entire career.

Didn't mean that sound as scalding as it may have, sorry about that Chris :-) I stand by the point though . . .

I also agree with you on the peak/career thing with Herman. Except for maybe Boudreau on peak, Herman is a clear choice to me.

Dobie Moore is a tough one. I have him in my top 40, but I'm not high on Jennings, and I can't see placing him above him. Marc, can you convince me that is a mistake? I really want to get this ballot right this year, and I'm open to wholesale changes, and elect-3 backlogger elections don't come along often . . .

Joe, oddly enough, our info base on Dobie Moore hasn't changed from the very first few posts of his first year of eligibility. To wit:

Chris Cobb's WS Estimate

1920 24
1921 34
1922 36
1923 26
1924 31
1925 28
1926 15 (part season)

194 total

Chris goes on to say that this is based on Moore's (let's be honest) fabulous hitting record (documented) and fielding reputation.

John responds that his total of 194 makes it tough to support him. "His army years will be the deciding factor for me."

A long discussion of his army years ensues. It turns out he played for an estimated 7 years (though there are suggestions that it was as little as 3 or as much as 10).) I wrote the following:

"We know Moore played baseball, and we know he played it well. He was hand-picked, though of course we do not know by whom, to play on an elite squad. I can't just dismiss it.

He played 7 years in the army, age 19-26. Taking Chris' suggestion of giving him half credit (three years)--at least for the moment, until we learn more about that army squad--seems conservative enough."

Unfortuntely we have never learned more about that army squad. But I digress:

"Then, what quality do you assign to those three years? Well, saying he matched his three worst years in the NL (24-26-28 WS by Chris' MLEs) may NOT be conservative. It may be a bit generous. Let's even reduce that by 25%. Dobie may have been a bit rough, unpolished; may have been a late bloomer. (This is a large assumption, of course, but I'm trying not to assume the guy INTO the HoM.) You get three more years at 18-20-21 WS.

"Career total now = 253
Top 3 still = 96
Top 5 still = 155

"Career comps Joe Tinker, Maury Wills, Dick Bartell (all in the 250s). This is a conservative total (IMHO) and you also end up with completely different kinds of ballplayers. More comp if you consider the style and strengths might be Boudreau (277), Jennings (214--not normalized) or Stephens (265).This is good company but not NB territory.

"Peak comps Jennings (97-150), Joe Cronin (102-152) or Ernie Banks (96-143). Now we are cooking with gas."

There you have it, the argument FOR Dobie Moore. At his peak he was Joe Cronin or Ernie Banks.

Why not support such a player? Well, 1) maybe you don't believe that he was really Joe Cronin or Ernie Banks. I think he was. 2) Even if he was Cronin or Banks at his peak, it was appended to too short of a career. Unlike #1, I will allow this as a valid objection.

So it comes down to a philosophical issue. For me, greatness is a quality a player possesses at a particular time and place, you can't accumulate greatness over time. Once you distinguish yourself as a great player, of course, then any and all career value that you can accumulate is a plus.

But Dobie Moore WAS a great player for a time, greater for 6.5 years than any other player available, I think, almost as great for his best 5 years as Hughie Jennings. The only other eligible player who was that great for 5 to 6 to 6.5 years was George Sisler, and that's why they're at the top of my ballot.

Further down my ballot, as I consider players who were not "that" great--not ever a Jennings or a Moore or a Sisler or a Cronin or a Banks--well, then, career value becomes an important tie-breaker. But for players who accumulated 350 to 400 WS without ever being great, well, sorry, after 1970 or 1980 or so, with modern training methods, etc. etc., there are just too many such guys for me to be all that impressed. I'll take Mark McGwire over Raffy Palmeiro every day, and ditto Moore over Billy Herman or Stan Hack.

I wrote: Boudreau's OWP is .603, Gordon's is .583, plus Boudreau was his era's Ozzie Smith.

How, if at all, do these career rates adjust for WWII? Is anything shaved off for Boudreau in 1944-45?

I don't think War Credit would help Herman vs. Boudreau, as Herman had peaked back in 1937 while Boudreau was coming into his peak:

Here are Herman's OWP before and after WWII:

1938 - .429
1939 - .568
1940 - .487
1941 - .527
1942 - .507
1943 - .644
1946 - .648
1947 - .137

1946 was a bit out of context for the 2nd half of his career, and 1943 was a war competition year.

Boudreau:
1939 - .406
1940 - .575 (Higher than Herman)
1941 - .509 (Higher than Herman)
1942 - .551 (Higher than Herman)
1943 - .675 (Higher than Herman)
1944 - .695
1945 - .611
1946 - .591
1947 - .674
1948 - .779

Since Boudreau was better the 4 prior seasons, whatever number I would estimate for Herman for '44 and '45 would be something less than Boudreau's number...

1941 should not say "Higher than Herman", which makes it only 3 out of 4 seasons, but I still stand by adding in '44 and '45 numbers will not bridge the gap between Herman's .563 OWP and Boudreu's .603..

I should also have said that merely having a great day or two is not really greatness (i.e. Mark Whiten was not a great player). And having a great year or two is not enough because there are plenty of players who are great for 3 to 5 years and more. Three years is generally a threshold that makes sense to me. Dobie Moore passes that threshold.

Comments on this 'n' that:

JoeD wrot, responding to me:

"The fine print: Boudreau played 1/4 of his time at 3B and 1B from 1949 on. Gordon played 30 games at 1B in 1941."

You can't be serious right?</i>

"The fine print" was a poor choice of phrase. I meant something like "the little things" that are necessary for a complete picture, but aren't all that significant. I think it is relevant to understanding Boudreau's career to notice that he was no longer a great defensive shortstop after 1948: people have a tendency to take win share letter grades as a fixed evaluation of career defense, when of course players' defensive value changes significantly over their careers, influencing their shape and the nature of a player's peak.

KJOK: I agree that war credit will not close the gap between Herman's OWP and Boudreau's, but what difference would it make between Boudreau and Gordon?

Sunnday2 wrote:

Joe, oddly enough, our info base on Dobie Moore hasn't changed from the very first few posts of his first year of eligibility. To wit:

Chris Cobb's WS Estimate:

It's my plan to revisit Moore's MLEs as soon as I get win shares finished for Willard Brown. I'll redo them according to current best practices, addressing playing time more systematically, using the regression formulas rather than eyeball estimates, and working out OBP, OPS+, and more firmly grounded win shares. Whether that will help or hurt his case I have no idea, but we ought to look at his record anew in light of the great deal we've learned about the translation business in the year (real time) since Moore became eligible.

KJOK: I agree that war credit will not close the gap between Herman's OWP and Boudreau's, but what difference would it make between Boudreau and Gordon?

That entirely depends on how you estimate Gordon's 'missing' time. Gordon played 112 games in 1946, and has his worst OWP of his career at .349 OWP:

1942 - .716 (career high)
1943 - .627
1944 -
1945 -
1946 - .349 (career low)
1947 - .645
1948 - .648
1949 - .505
1950 - .489

1946 was an outlier, but not sure you can just assume approx. .630 seasons for Gordon in '44 and '45 either, BUT:

if you assume two .630 OWP, 630 PA seasons for '44 and '45, Gordon's career OWP would go from .583 to .587.

>I suggest we call them Genetically Normal Equivalents (GNOMES) or the Biologically Adjusted Discount (BAD), but I'm open to other suggestions.

Re. Jeff's comment from 1957, I don't assume he is kidding unless he says he is. I don't think lots of people are kidding when you get poll results like this. (Unfortunately I cannot find the poll so the figures may not be quite accurate.)

The poll asked if the following players should be elected to the HoF or not. The percentage is the number who answered yes.

Mark McGwire 60%
Sammy Sosa 59%
Barry Bonds 57%
Rafael Palmeiro--don't remember but he was last among the 4

Note that none received 75%. Of course the poll does not distinguish no votes based on a belief that the player did not perform at a HoF level versus those who voted no because of alleged steroid use.

But Barry Bonds at 57% is clearly a steroid-based response. Not as clear on the others.

After reading the Moore thread and Sunnyday's summary in post#40(thanks btw, that was much easier to follow than the actual discussion), I am convinced that I was underrating Moore. I still look at him as an all-peak guy though and I can't justify putting him ahead of Hughie Jennings. As my ballot for '58 is currently structured, I have Jennings #13. I haven't decided exactly where I'll slot Moore with the overall field(probably 14-20), but he will be on or near my 15-man ballot(especially after we elect 3 this year). I have also decided I was underrating Sewell. His combo of career/peak and hitting/fielding deserves recognition. My new SS rankings go Sewell-Boudreau-Jennings-Moore; honorable mention to Lundy and Long.

Note to all ballot counters using the, um, ballot counter:

If you change the number in cell A1 from "Elect 2" to "Elect 3" (or whatever), the ballot counter will take care of the rest.

Thanks for the heads up, Evan!

Carl, I agree he can't be rated ahead of Jennings. I have Jennings #1 and Moore #2!

My methodology is basically 1/3 peak, 1/3 prime, 1/3 career. If one's method is more like 1/2 peak/prime and 1/2 career, I can see Boudreau (I have him approx. #9-11). But Sewell?

My problem with Sewell is that he only played SS for 8 years. If he was a regular SS for his entire career then I could see him in Jennings-Moore-Boudreau territory. But no. And we was regarded as only a fair fielder at 3B, not to mention by the 1930s a mediocre hitter for thge position.

Ironically his career pattern is not unlike Moore's, only backwards.

Moore
• 3 to 7 years Wreckers, modest value
• 6.5 years superstar NeLs

Sewell
• 8 years star ML SS
• 5 years mediocre 3B

I see Moore as having the higher peak and Sewell the edge on shoulder seasons. Total WS Sewell 277, Moore 253. But Moore has so much the better of it on peak that using James' method Moore would clearly end up higher than Sewell at #23.

Jennings is at #18 and maybe that's about where Moore goes? Or at his peak, he was Banks (#5) or Cronin (#8), though for his career he was only Wills (#19) or Tinker (#33) or Bartell (#37). Still, average them and you get 6.5 on peak and 30 on career, or about 18 again. So #18 versus Sewell's #23. Boudreau is #12.

All of this assumes you buy into WS, of course.

I'd be interested to know if anybody else among the voters weights career value at the 90% level or higher, as compared to peak and prime. I consider the Hall of Merit (and Fame) to be like a lifetime achievement award. They already got their awards for peak -- those are called Gold Gloves, Silver Sluggers, MVPs, All Star selections and Cy Youngs (actual or retroactive for the pre-awards players).

Am I alone among the electorate in this view?

Daryn, I have never regarded one year as a proper peak. I think James has it right conceptually that a peak is somewhere in the 3 to 5 year range. This is not arbitrary in the sense that if you look at 3 to 5 years, you will find plenty of players who were at an all-star level (roughly or by analogy or whatever, I don't mean that as a literal statement) from which to build a reasonable consideration set. If you add in everybody who won a seasonal award, you've got too many players, too much complexity, too many who are never going to make a ballot anyway. Probably one-third to one-half of the single season list make it to 3 to 5 years at that level, as a wild guess, and it is that one-third to one-half that is really worthy of our attention. At least that's how I see it.

Daryn,

I tend to be very peak/prime oriented. But I've incorporated more career and extended prime into the mix as I've gone along. Probably 65% peak/prime, then 25% extended prime, then total career on whatever's left.

Career tends to be icing for me, rather than cake. This is because I'm double and triple counting peak, prime, extended prime by using best any 3, 5, 10, 15, career to gague players.

1958 Prelim

This could certainly change in the very near future.

1-15
----
Duffy
Oms
Medwick
Mendez
Cravath
Matlock
Walters
Ferrell
Burns
Browning
Hack
Herman
(Lyons)
Willard brown
Ruffing
Jennings

16-25
-------
Averill
Rixey
Van haltren
Moore
Roush
(Terry)
Poles
Childs
Monroe
Boudreau
Sisler

Daryn, I am 90% career, but my 'replacement level' that determines career value is higher than many other 'career voters', so I suspect in essence I am often closer to 'prime' voters.

Marty Marion is eligible this year, so he may be worth a few comments.

He had 1572 career games, 1549 of them in the field, 1547 at SS. If you want me to pay attention to a light-hitting glove wizard, career length becomes a vitally important factor, and ~1550 games really isn't enough, even if all of it is at SS. So he's not a serious HoM candidate.

But there is 1944 to talk about. I've heard it said that Marion was the only player ever to win an MVP award "almost entirely with his glove." So what was it about the 1944 MVP vote?

1. The Cardinals ran away from the league, winning the pennant by 14.5 games. There's no doubt that they were the best team in baseball.

2. In a near-neutral park (Sportsman's park, near-neutral!), in a league that scored 4.25 R/G, they scored 4.92 R/G, and allowed 3.12 R/G, leading the league in both by comfortable margins. Both sides were outstanding, but the pitching/defense side seems more outstanding.

3. Stan Musial was the team's best hitter, and arguably the best hitter in the league. However, Musial didn't lead the team in either HR (Whitey Kurowski) or RBI (Ray Sanders.) Of course, Sanders was obviously benefitting from batting behind Musial.

4. Musial was the 1943 MVP. His 1944 performance approximately repeated his 1943 performance (OPS+ for the two years: 180, 175) but wasn't clearly better. Although it was legal for a previous MVP to be voted the award again, there was clearly a strong sportswriter prejudice against repeat winners - especially when they are "only" having another year consistent with their careers.

5. The Cardinals had two bad pitchers who totalled 8 IP between them. If you remove these two "temps", every pitcher on the staff had an ERA+ above 100. Clearly, the pitching was outstanding - but having every pitcher look good starts making you ask how good the defense was. Mort Cooper won 22 games and was the best Cardinal pitcher, but Cooper didn't really stand out among all pitchers in the league.

6. Bill Nicholson of the Cubs led the league in RBI, by a considerable margin, and had led the league in RBI the previous year. There was, and is, a considerable sportswriter prejudice involving RBI. Among three big-hitting flank outfielders - Musial, Nicholson, and Dixie Walker - Nicholson played every game, which was 10 more than either of the other two. (Mel Ott was also pretty good, but in only 120 games.)

7. This being the heart of WWII, I would think that character issues loomed large with the writers. I have no idea what those issues were, or why they would have fixed upon certain players as being more worthy than others, but there must have been some influences.

8. With the voters not being willing to push for Musial, but with many valuable Cardinals, whom to focus on? Make defense the story, and you come up with Marion (OPS+ 91). Walker Cooper had a 137 OPS+ in 112 games as the primary catcher; he got attention, too.

For all that, the vote totals:

1. Marion, 190 (7 first place votes)
2. Nicholson 189 (4)
3. Dixie Walker 145 (3)
4. Musial 136 (3)
5T. Bucky Walters 107 (3)
5T. Bill Voiselle 107 (0) (Giants pitcher, 21-16, ERA+ 122)
7. Ray Mueller 85 (2) (caught 155 games)
8. Walker Cooper 72 (1)
9. Mort Cooper 63 (0)
10. Bob Elliot 57 (0)

Other Cardinals were Ray Sanders with 25 (the RBI thing), Johnny Hopp 10, Ted Wilks 4, Max Lanier 2, Whitey Kurowski 1.

I didn't check either WS or WARP for the year. I'm sure there are things to say from those perspectives.

O,

Indeed.

Over 30

Musial 38 WS
D. Walker 33
Walters 32
Galan 32
Nicholson 31
Jim Russell 31

Among MVP Top 10

Marion 20
Voiselle 24
Mueller 26
W. Cooper 18
M. Coopeer 24
Elliott 27

Also of Note

Holmes 27
Cavaretta 25
McCormick 29
Ott 25
Barrett 25
Hopp 28

So if you're not going to select Musial, you might as well just get out the dart board. Which is what they appear to have done. Even just among the Cardinals, Marion is only #6 in WS after Musial, Hopp, M. Cooper, Kurowski and Sanders.

Actually this has to rival Ted Williams in 1942 as among the biggest snubs ever. Granted he was only #1 by 5 WS, but by 18 over the guy who got the award.

O, BTW, I count 21 1st place votes. Were there others? Who?

according to bbref, one vote for Babe Dahlgren, which makes 24.

The one other first place vote went to Babe Dahlgren.

Babe Dahlgren???

Granted the Pirates contended for the first time since 1938. But Dahlgren, their 1B? Compare him to Elliott, their 3B.

Dahlgren 12-101-.289/.347/.419/OPS+ 110/19 WS (5th on team)
Elliott 10-108-.297/.383/.465/132/27 WS

But Elliott was just a 27 year old kid, while Dahlgren was a 32 year old veteran who had been unwanted (17G in ChiN, 2 in StLA, 17 in Bro in 1942, and 1943 in PhiN) who made a "comeback" in '44, leading the league in G. Actually he had the same year in '44 that he had had in '43.

'43 6-56-.287/.354/.362/111
'44 12-101-.289/.347/.419/110

He did increase his XBH from 26 to 47 in 88 more PAs.

Dahlgren is probably one of the worst players ever to get a regular turn on a truly great team, with OPS+ of 76 and 86 with the Yankees of '39 and '40. Well, they were great in '39. His comp may be Doug Mientkiewicz, as he may have been a fair glove. Well, OK, B-.

But a 1st place vote in '44. Good god.

Joe Gordon and war credit...

I will admit that Joe Gordon is a player that I envisioned I would be supporting, but currently he ranks #18 on my ballot behind Jennings, Boudreau, Herman, Childs and Moore, just ahead of Doerr.

why?

I feel as if I need to be conservative with war credit. In the two years prior to his entering the service, Gordon posted 28 and 31 WS, respectively. His 31 WS came in a war year, with my 10% discount this equals 28 WS. SO he has 28,28,x,x through the war.

Then in 1946 he is, in HOM terms, awful. 112g, 11 HR, 47RBI, .210/.308/.338, 79 OPS+, 3.25 RC/27, 9 WS. Yech! He rebounds somewhat in '47 and '48 with 25 and 24 WS, a drop of about a win from his '42,'43 peak. This gives us 28,28,x,x,9,25,24. Make this to 162 games, he gets 29,29,x,x,10,26,25

*as an aside, why did he only play 112 games in 1946? Was he still in the service? Could he have another 2-3 WS added onto his total?

Granted that 1946 is a bit of an aberration, but it does give me pause before saying that Gordon missed two 'peak years' to the war. I am slightly conservative with war credit in that I dont' usually imagine seasons among player's best three, but I even feel a little uncomfortable giving Gordon as much as I do. I give him, again in WS, 25 and 22 for the two war years. This adds virtually nothing to his peak (WS >25, and 3 and 5 year totals) but does add to his prime (WS>15, seasons w/ 15+ WS, 7 year totals). As a peak voter, he falls a bit and his prime still isnt' as good as those of contemporaries Boudreau and Herman.

Those of you who have Gordon in the top ten, how much credit are you giving him for the war? Had 1946 not happened or had he been as good in 1946 as he was in '47 and '48, I probably would have given him 28,25 for the war. This would help him significantly on my ballot (i.e. put him on the ballot). But I can't bring myself to do it.

Oh, this is in WS because it is something I am comfortable and well versed in. If you prefer RCAA, OPS+, WARP, TPR, the same applies.

I think it's fair to say that along with Marty Marion, three other candidates who clearly are not HoMers are nevertheless "household names" among baseball history fans. Those would be:

• Tommy Holmes, who had a big year in 1945 and then a not-so-big-year yet helped lead the Boston Braves to a completely unexpected pennant in 1948

• Pete Reiser, who was touted as a potential all-time great until forced into early decline and retirement by injuries. Now we sometimes hear that he was overrated by his contemporaries but, hey, his OPS+ was 163 and 142 in his first two years, and he led the league in both 2B and 3B in 1941. And he spent '43-'44-'45 in the military, then came back at 122 and 117 in 100+ games in '46 and '47.

• Sam Jethroe, who has his own thread.

The name that jumped out at me, as in "who's that?" was Johnny Hopp. Hopp played through WWII, 14 years overall, but only 1393 games. BL/TL.

He played 100+ games in 1941 and 1944 through 1950. He was already 25 in 1941 when he played 134 games for a good Cardinals team that was a year away from really jelling, however. (He had played 86 games prior to this.)

1941 OF91G, 1B39 .303/.378/.436/122
1942 1B88, total 95, platoon with Ray Sanders .258/.334/.382/102
1943 OF52, 1B 27 .224/.297/.307/71

His playing time declined and Sanders' increased as his performance declined pretty dramatically. He went 3-for-21 in the '42 and '43 WS.

Harry Walker went into the service in '44 and '45 (and Musial in '45, not even to mention Country Slaughter) and Hopp took Walker's place and he performed well again in a full-time role.

1944 CF .336/.404/.499/150
1945 OF104, 1B 15 .289/.363/.395/108

In the '45 WS he hit .185.

In 1946 the Cardinals had too many hitters and Hopp went to Boston.

1946 1B68, OF58 .333/.386/.440/133
1947 CF .288/.376/.358/98

Unfortunately he was sent to the Pirates and missed the Miracle Braves WS appearance in '48.

1948 OF80, 1B25 .278/.345/.385/95
1949 1B71, OF20, 113 total .306/.367/.408/105
1950N 1B76, 106 total .340/.420/.522/141!

He went to the Yankees late in the year:

1950A 35 PAs, .333/.486/.593/180! (0-for-2 as PH in WS)
1951 1B25, 46 total, 72 PAs .206/.306/.317/53 (0-for-1 in WS)

In 1952 he was a PH with the Yankees, then Tigers, but with an OPS+ of 43 at age 36.

Totals .296/.368/.414/113
WS for the 1940s at least: 3-19-10-4-28-16-21-14-12-12
Total 155

A valulable man, I'd say, a lefty who seemed to be competent at 1B and all around the OF. WS rates him an A- OF, which is exceptional, I think, for a guy who played the corners more than he did in the middle. I don't have a clue about his splits, Oddly enough, he seemed to platoon with Ray Sanders, also BL (though TR), early on and in '46 the Cards kept Sanders rather than Hopp.

Hopp (again) .296/.368/.414/113/155 WS in 1393 G
Sanders .274/.370/.401/115/77 WS in 630 G

This doesn't prove the Cards didn't keep the right guy but it doesn't prove they did. If Hopp had established himself as a regular in 1941--if Ray Sanders had never lived--he might have played 2000 games and we might be looking at Hopp as a candidate right now. But I doubt it.

Stan Hack....again in WS

Hack's five year peak is 34,33,30,26,25 and he earned 316 career WS.

He gets my war deductiosn for 1943,44,and 45. So instead of 21,13,34, he gets, 19,12,30 (10% deduction, and I only round up at .7). His five year peak is now 33,30,30,26,25 and he now has only 309 career WS.

Now to move him to a 162 game schedule (which I have now done for everyone in my consideration set) and he goes to 34,31,31,27,26 and 324.

Bill James says that 325 Win Shares is a rousg estimate for a HOFer. Hack has 324.

I measure peak in three year totals (96 for hack), five your totals (149 for Hack) and WS>25 (24 for Hack). None of these are eye popping unless compared only to eligible 3B. I have him slightly above McGraew on peak, but he has much mor career. I have him roughly equal to Leach, but Leach had a lower peak, league strength issues, and played half his career at a position that WS seems to overrate (CF). No one has a prime (WS >15) as high as Hack at the position.


I think that Hack is the best 3B by far. However, I have to admit taht i fyou take hack resume and put it right up against those of a few other players he is a very borderline case. But he played a position that has been underrated hsitorically and that has very frew HOMers (not that we have missed anyone per se). For this I give Hack a position bonus that will raise him up to #4 on my preliminary ballot. I think that he is far better than a mere borderline case and better than guys like Terry, Wells, Lyons and Carey (my four HOM, not PHOM, along with Boudreau). I also think he was better than Heinie Groh, whom I supported for HOM.

Gordon received a severe spike injury in a spring training game on March 19th - "a completely severed extensor tendon of the middle finger of his left hand and the bone of the finger was slightly chipped." Looks like he missed the first 3 weeks of the season, and might explain why his batting was so poor that year?

On May 8th, Gordon had a "charlie horse" and sat out for a whole week. Beginning to look like he had a lot of injury problems in 1946...

preliminary ballot...

As of now I am pretty ocnfident in Stan Hack and Billy Herman as PHOMers, the last PHOM spot will come down to Medwick, Well, and Duffy. I will also take a hard look at Lyons and Redding.

1. Jennings (PHOM 1938)
2. Boudreau (PHOM 1957)
3. Ferrell (PHOM 1956)
3. Hack
5. Herman
6. Medwick
6a. Wells
7. Childs (PHOM 1939)
8. Duffy
8a. Lyons
9. Redding
9a. Terry
10.Walters
11. Keller
12. Averill
13. Moore
14. Griffith
15. Rixey

Gordon, Doerr, and Dean lurk right outside my top 15.

KJOK, are you suggesting that I shouldn't read too much into his 1946 season when I am figuring out how much war credit to give out? I am open to giving him more.

On June 9th, Gordon pulled a muscle in his right leg.

Also, (all from NYTIMES) July 12th article:

"Gordon appears to have made no attempt to conceal his attitude towards a manager (Bill Dickey) with whom he once was the closest of friends. His rancor is believe to be shared by many of the players who have been with the club since the late 30's"

Being the ringleader of rebellion against the manager may have cost him playing time, especially considering his performance.

KJOK, are you suggesting that I shouldn't read too much into his 1946 season when I am figuring out how much war credit to give out? I am open to giving him more.

I'm just suggesting that injuries probably played a part in his poor performance, as is usually the case when a player has a year that is way out of line from his surrounding seasons, AND that player rebounds in subsequent seasons...

Estimated win shares for Willard Brown are now posted on his thread.

Thought I'd mention it here Just to make sure nobody misses 'em, in case that thread drops out of view quickly.

Brown certainly deserves a close look.

Not that I hate him, but I see no way that he [Medwick] is better than Sisler, who had a simlar career shape, but was obviously better in his prime.

When I first read that, I thought maybe you got your players backwards. "Obviously" and "no way" are pretty strong opinions. But...

Sisler is slightly better on peak under WARP1 (using 3 and 5 year measures), but pretty far behind Medwick on peak under WS. Sisler is also far behind in WS per 162 games, but about the same in WARP1 per 162 games. Medwick beats Sisler on career measures under both systems (although the advantage under WARP1 is negligible).

I basically see Medwick as quite a bit better under WS, and Sisler as slightly better under WARP1. I think it's awfully hard to say that Sisler wins the battle of the uber-stat systems.

You've got to take into account the run environments also, which were pretty "normal" during Medwick's years in the NL but pretty high during Sisler's years in the AL. Basically, if you normalize their stats Medwick has just about the same BA and OBA, but much higher SLG. That's why, in part, Sisler's RC/27 was about 35% better than the league average, and Medwick's was up around 50% better. Their OPS+ are lower than the RC/27 relative measure, but there's a similar gap between them (+25% vs. +35%).

Medwick's got more black and grey ink. Medwick was a better defensive left fielder than Sisler was a defensive first baseman, for whatever that's worth -- which isn't much.

They are essentially the same on HoF Standards and HoF Monitor if you take those into account.

All in all, Medwick seems like the better player to me, but even if I'm evaluating that incorrectly, I'd have a hard time saying Sisler is obviously better.

Jeff, a few questions, all of which are about small details, but in evaluating Sisler vs. Medwick, we have to get down into the small details:

Medwick was a better defensive left fielder than Sisler was a defensive first baseman, for whatever that's worth -- which isn't much.

By what measure? And are you taking peak vs. career fielding value into account?

Are you adjusting Sisler's 1918 and 1919 seasons for the games lost to WW1?

And now the big question: Can you justify the difference in the WS evaluation of Sisler's and Medwick's top seasons?

Sisler 154 games, 181 OPS+, .348 eqa, 30.0 bws
Medwick 156 games, 180 OPS+, .348 eqa, 36.9 bws

Sisler's run environment was higher, 4.76 to 4.51. Is that enough to account for 7 batting win shares difference here?

The other major difference I see is that Sisler was second in the league in OPS+ whereas Medwick was first. Sisler, of course, was miles behind Ruth, who put up a 255 OPS+ that year . . .

I'll agree that Medwick's peak was better if you can show me why I should believe win shares in this instance.

Hello folks,

Just sticking my head in here. I've read through all the discussions up to 1952 (skipped some of the more recent ones) and all the ballot results. I've learned a lot about baseball. Thanks for your efforts.

Now, for a blast from the past, from a long-ago thread:

Top tier all made 36 ballots or more: Burkett, Start, McPhee, McVey, Stovey, Bennett, Duffy. These guys have the broad support needed to get elected some day.

Second tier, 29-32 ballots: Grant, Ryan, Thompson, Van Haltren. Not impossible to imagine a player getting elected with 30-32-35 ballots someday, as the pool gets bigger and votes more spread out.

Third tier, 20-22 ballots: Pike, Jennings, Childs. A very tough row to hoe. I don't see us electing guys with 20-22-25 ballots.

Fourth tier, 10-16 ballots: Tiernan, Browning, Pearce, McCormick, Caruthers. Fuhgedaboudit! Unlikely to ever make more than about 20 ballots.

It's interesting to see who did, didn't (and who still might) make it from that list.

Oh, one more thing:

Free Hughie Jennings!

John,

Would you mind creating a thread for Quincy Troupe? There's enough info available at this point (or soon to be available) and enough questions about his record that it's probably time to open up a more specific discussion of him.

In addition, with him being a catcher, I suspect that there will be much interest in his case among those who offer catcher bonuses.

For the record, James rates the NeL catchers:

Gibson
Santop
Mackey
Radcliffe
Petway
Campanella
Trouppe

James says his was a 22 year career, he was a switch-hitter and a "perpetual .300 hitter who walked a lot." He made the all-star team in 23 leagues over the years and played in the majors in 1952. So why is this man only #7?

As for his ML experience, I only have numbers. He played for the Indians, 6 games at age 39.5, going 1-fo-10 with 3 K and 1 BB. He is listed at 6-2 1/2, 225 lbs.

There's no thread on the main page yet, but folks here will probably want to know that Double Duty Radcliffe died yesterday.

Also in case you missed it, Brent found a nice web page on the integrated North Dakota teams that Radcliffe, Paige and others played on in the '30s.

http://www.pitchblackbaseball.com/northdakotabaseball.html

Death - we could use another Jennings supporter. Submit a ballot.

and if Death is a Clark Griffith proponent, vote early and often!

I am also a strong career-value voter with a low-to-middling replacement level. So I tend to like medium-peak long-career guys more than other voters.

Ugh, falling hopelessly behind in the discussions.

Almost want to gripe about chasing a tiger and his balls around in the hot sun all day, but I suppose there are tougher tasks. It is hot, though. :)
(Wargo will like that one, at least)

Seriously, I'm off next week and look forward to re-calibrating my whole process. It's good timing for me....

New stat: Average Opponent Winning Percentage. I never bothered to figure it before because I thought it would be too much work, but I belatedly realized that 90% of the work is done just figuring MOWP.

24 guys figured for. 6 have links to their page because for those six not only do I have their career AOWP figured, but I also got single-season info up on the site -- it's in the chart on top, alongside RSI, adjusted W/L record, and defensive adjustment.

If you get a chance, check out the AOWP for Joe McGinnity in 1908. To put it in perspective the next lowest AOWP for a pitcher with at least 20 starts in .443. Just a little bit behind that.

QUINCY TROUPE'S MEXICAN LEAGUE PERFORMANCES IN CONTEXT

MEXICAN COUNTING STATS
YEAR TEAM AGE  G  AB    H  TB  2B 3B HR  BB SB 
----------------------------------------------
1939 MON   26 38  137  42  71  11  3  4  24  7
1940 MON   27 76  276  93 150  25  7  6  48  9
1941 MON   28 98  363 111 171  25  4  9  66  9
1942 MEX   29 70  269  98 171  27  5 12  62 20 
1943 MEX   30 76  276  83 143  18  3 12  60  5
1944 MEX   31 57  197  47  82  12  1  7  40  4
1950 JAL   37 67  206  58  92  10  0  8  55  3
1951 JAL   38 63  177  45  70   6  5  3  52  2
==============================================
     TOTAL   545 1901 577 950 134 28 61 407 59 

POSITIONS: Cisneros offers no defensive splits, and he lists Troupe as being a 3B-C. The Negro Leagues book that KJOK posted a page from the other day shows his positions this way:

1939 C 2B SS OF
1940 3B
1941 C 3B
1942 C
1943 C
1944 C 3B 1B OF
1950 C 3B
1951 C

The Mexican season typically lasted between 80 and 100 games, though I think that the 1939 season was probably more like 50-60 games long.

MEXICAN CAREER RELATIVE AVERAGES
                      LG   LG   LG
YEAR  AVG  OBP  SLG  AVG  OBP  SLG OPS+
--------------------------------------
1939 .307 .410 .518 .261 .319 .358 173
1940 .337 .435 .543 .290 .379 .420 144
1941 .306 .418 .471 .288 .390 .396 125
1942 .364 .483 .636 .289 .380 .394 189
1943 .301 .426 .518 .273 .366 .367 157
1944 .239 .367 .416 .284 .380 .387 104
1950 .282 .433 .447 .272 .361 .335 153
1951 .254 .424 .395 .279 .384 .367 118
=======================================
     .304 .426 .500 .281 .374 .382 145

MEXICAN CAREER RELATIVE AVERAGES
AVG+ 108
OBP+ 114
SLG+ 131
OPS+ 145

ROUGH TRANSLATION FROM MEXICAN LEAGUE TO MLB CONTEXT
     MLE AVG  MLE OBP  MLE SLG
-------------------------------
1939 .296    .406      .476
1940 .284    .367      .415
1941 .252    .344      .366
1942 .295    .390      .471
1943 .263    .372      .417
1944 .205    .299      .338
1950 .255    .375      .424
1951 .218    .349      .332
===============================
MEAN .259    .363      .405

NL
MEAN .268    .338      .385

NL
C ONLY
MEAN .256    .322      .364

This rough translation considers the Mexican League to have a conversion rate of .90/.82 for each season through 1951. That might not be reality, but I think it probably averages out to be around there over the full scope of the 8 and 10 years of Troupe's and Wright's careers. I arrived at it by
-first indexing Troupe's averages relative to his league (to MxL that is)
-second applying that index to the NL context
-finally applying the 10%/18% conversion rate.

I'm surprised by how well Troupe comes out in this. I've been double-checking, and I don't think it's due to any computational error, though I'm not immune to them, so I'm not guaranteeing anything....

Perhaps one thing that might explain some of this surprising result is that he had an outstanding season in 1939, which was a weaker league season than the next few; and the same goes for 1950 and 1951 versus their preceeding seasons.

On the other hand, Wild Bill Wright wasn't down there in 1939, but was there for the first and second waves of imported hitting talent (1940-1941, then again in 1946-1947). But that is counterbalanced by fewer imported hitters in the surrounding seasons, where the number averaged about ten imported players.

Troupe missed that second wave altogether, but was very much on hand for the first wave, and was, in fact, leading it with C.P. Bell and Chet Brewer.

As mentioned in the Wright comments, I'm going away for a few days, which means I won't be able to do much with Troupe at this point, but with the information supplied by KJOK and others, there's lots more to go on now than before.

I don't know at this point that he's a HOMer, but based on his Mexican performance, his (roughly) translated performance relative to his MLB positional peers, and based on the high averages that Holway reports for him, I think he deserves a full workup by Chris C.

I want to second Doc's recommendation that Quincy Trouppe get his own thread. I've been too busy lately to pay close attention, but Eric C brought my attention to the fact that Trouppe was eligible. Basically Trouppe is a bigger better stronger version of Wally Schang.

1958 absentee ballot.

I will be unable to submit a ballot next week, so could somebody please post this on the '58 ballot thread?

A definite backlog year. Even with war credit, Dom DiMaggio falls short. Willard Brown has a strong case as one of the best hitters in the NeL "West" during his prime. He doesn't make my ballot in my first pass through his case, but I'm definitely not writing him off.

1. Wally Schang Pops back to the top by default. Tremendous career value in his long career, many seasons with > 120 OPS+ as a C. An OBP-heavy OPS, ranking 2nd all time among C in career OBP. Bresnahan and Lombardi are the other 20th century ML catchers who belong in the discussion.

2. Joe Sewell 87.4 WARP3 in 1903 games is a HoM-worthy career. The number of seasons among the top N players in league demonstrate a HoM-worthy peak. Like with Schang, one's point of view on AL vs. NL league strength in the 1910s-1920s has a big effect on his standing in the all-time rank-o-meter.

3. Red Ruffing We haven't seen anybody quite like him: IMO, a tad below Eddie Plank and a tad above Rixey s the best description. A lower peak than any other pitcher I've voted for so strongly, but his great career more than makes up for it. More career WS as a Yankee than any other pitcher.

4. Joe Gordon With all the talk about giving WWII credit, will it actually make any difference between anybody getting elected or not? Comparisons with the performance of similar players by age suggests that he may have missed out on some great seasons.

5. Bobby Doerr Another great 2B.

6. Stan Hack The 2nd greatest ML 3B to date, behind Baker, which is both a credit to Hack and an indication of the historic weakness of the 3B position.

7. Billy Herman Worthy.

8. Earl Averill Top ML CF of the 1930s.

9. Charlie Keller Peak, peak, peak, but will end up one of the few 30 WS/162 players in history not to make the HoM. Very few full-time seasons due to war and injuries, but was actually one of the most durable players when available 1940-1946. For those who thought that Greenberg, with a 158 OPS+ in 1394 games and 31.0 WS/162 was a shoo-in even without war credit, why not at least consider Keller, with a 152 OPS+ in 1170 games and a 30.0 WS/162 (although true, his war year stats are inflated)?

10. Tommy Bridges 126 ERA+ in the 1930-1943 AL.

11. Cool Papa Bell Long career, low peak, like Sam Rice, but with outstanding speed.

12. Lefty Gomez With "dominant season" pitcher bonuses, his 2 Cy-Young type seasons give him a big boost.

13. Lou Boudreau Why so much lower than the consensus, despite my regard for IF, and all of the sabermetric evidence in his favor? Probably the "comparison with peers at the same position" factor in my ratings. Not only was he a contemporary of Appling and Vaughan, but also strong upcoming SS candidates Pesky, Reese, Rizzuto, and Stephens.

14. Jose Mendez Evidence of a HoM worthy peak.

15. Joe Medwick Back on my ballot. Played like an inner-circle HoMer for 3 years, but like a non-HoMer the rest of his career.

Mackey is 16th. I wouldn't complain if Jennings , Ferrell , or Rixey eventually made it, but would prefer that my top 8 got in first.

Got to admit that this made me laugh. Eric C e-mailed me, asking about Quincy Trouppe. I posted two comments for him about Trouppe, one in Ted Radcliffe's thread, and post #84 above. I comped Trouppe as a bigger better version of Wally Schang. Then, in post #85 above, Eric C posted his 1958 ballot with, of all people, Wally Schang as number 1. I knew Wally Schang had his supporters but was unaware that Eric C was his main man.

Note to John Murphy: Eric C and I need a Trouppe thread to get the Quincy band wagon rolling.

Yeah Baby, Yeah (Sorry, Austin Powers on TV in background).

A few posts back, I asked Jeff M. about the real difference between Sisler's and Medwick's top seaons:

And now the big question: Can you justify the difference in the WS evaluation of Sisler's and Medwick's top seasons?

Sisler 154 games, 181 OPS+, .348 eqa, 30.0 bws
Medwick 156 games, 180 OPS+, .348 eqa, 36.9 bws

Sisler's run environment was higher, 4.76 to 4.51. Is that enough to account for 7 batting win shares difference here?

I think that we would all agree that, any way we examine the component stats of Sisler's and Medwick's top seasons, their production was **** near equal, if their OPS+ scores differ by 1 and their eqa is identical to the third decimal place.

But I think it is the case that Medwick is getting credit from win-share voters for a truly historic, 40 win-share season, while Sisler is getting credit only for a basic MVP-caliber 33 ws season, and Medwick is getting that credit mostly due to noise, random variation in the WS system that pushes Medwick up a few win shares and Sisler down a few win shares.

To see if the noise theory was correct, I looked at all the seasons between 1907 and 1967 in which a hitter achieved an OPS+ of 175 to 185. I then recorded the EQA, the games played, and the batting win shares for all these seasons as well, so that I could compare three rate stats for offensive production: OPS+, EQA, and bws/g.

There were 51 such seasons. The EQAs ranged from a low of .336 (Eddie Collins, 1914), to a high of .361 (Lou Gehrig, 1937). The batting win shares ranged from a high of 41.6 (Ty Cobb, 1915) to a low of 22.6 (Goose Goslin, 1928). Differences in playing time were a factor in total batting win shares of course, but these two seasons also happened to produce the highest and lowest rates of bws/g: .267 and .167 . Goslin's rate per game may be depressed by the fact that he appeared in 7 games only as a pinch-hitter, but his rate is still very low.

The study fixed the range between the highest and lowest OPS+ scores considered. The lowest were 94.5% of the highest. The range between the highest and lowest EQAs was similar. The lowest was 93.1% of the highest. The range in win shares was much larger: the lowest rate was 62.5% of the highest rate.

I am troubled by variations in win shares rates that are four times larger than variations in OPS+ or EQA. Are there good reasons why the win shares assessment that these seasons are really of very substantially different value, or can this range of values only be justified by putting the shovel into the bullshit dump?

Sisler's 1920 season, incidentally, has a bws/g rate of .195, the seventh-lowest out of the 51 seasons. Medwick's 1938 season has a bws/g rate of .235, the tenth-highest of the 51 seasons.

I conclude from this study that

1) Although Medwick's 1938 season was undeniably great, he nevertheless gets lucky with the win shares, while Sisler was unlucky iwht his undeniably great 1920 season;

2) Given the luck involved in Medwick's 40-win-share season in 1938 and the big boost from Pythagoras to his 1937 win-share totals, I'm really concerned that it would be a mistake to elect Medwick with his great win-share peak as the Exhibit A justification;

3) I'm really dubious about the value of any win-share assessment of players that includes fewer than five seasons;

4) I'm beginning to think that for the kinds of fine disinctions we're having to make between players now, win shares is not a sufficiently accurate tool.

Chris, I don't consider this luck at all.

I'd guess (don't have the time right now to check for myself) that the differences are caused by those guys playing for 'over or under achieving' teams if you'd want to call it that. OPS+, EQA, etc. - they are just guesses as to how many runs the stats would produce in theory. WS looks at the actual runs the team scored, and then gives out the credit proportionally. Neither is right or wrong, they are just different and both are guesses. But I wouldn't call it luck - if anything, WS is a better number to use if you care about what actually happened, IMO.

"Ugh, falling hopelessly behind in the discussions"

Tell me about it Howie . . . but doing the best I can to catch up now. Visiting in-laws, said I 'couldn't sleep' so that I can get some late-night computer time . . .

******

Very cool post Death to Immobile Things . . . I love looks back at the past like that.

******

I'm definitely a 90% career guy Daryn . . . The Pennants Added are a lot of data entry, which is why I'm behind, but the one thing they've shown me is that the impact of a 'peak' season is less than most people think. Generally a '10' and a '0' is only about 10-15% more valuable than two '5's, for example. I'm a huge fan of long-career pitchers, and this is one of the main reasons why.

******

Back to the Win Shares thing - especially on offense, I think Win Shares are very accurate. Everything is accounted for. If you care about how many games were actually won and lost, whether a season actually had 'Pennant Impact' instead of theoretically could have had that impact, they should be your choice. If a team comes up 10 games short in the standings, for whatever reason, a 'look-back' accounting system needs to deduct it from somewhere. I seriously have zero issue with the things Chris mentions regarding Medwick and Sisler.

Eric C e-mailed me, ..... Then, in post #85 above, Eric C posted his 1958 ballot

Gadfly- I think that there are 2 different "Eric C."s participating in this project. Somebody correct me if I'm wrong, but I believe that the one who e-mailed you is the person who posts as "Dr. Chaleeko".

Neither Win Shares nor WARP are remotely accurate; they are themselves attempts to estimate things that are close to unquantifyable, they have "favorite" fielding positions at different times in history, and they overstate the achievements of players on winning teams. Finally, as Chris pointed out: they leave in the "dumb luck" element in a team's season, which with only 154/162 entries, instead of say 600 AB, or batters faced, is far more important than in individual stats.

Neither WS nor WARP are anything more than a clever but inaccurate gimmick, and should not be used for serious analysis.

All-time 'vote points totals' leaders, through 1957

JENNINGS 15763
VAN HALTREN 15105.5
DUFFY 14540.5
BECKLEY 14516
Pike 13399
GRIFFITH 13380
BROWNING 12558.5
Thompson 12349
Bennett 11503
WADDELL 11263

Caruthers 10704
WELCH 10520
(RYAN 10507.5)
CHILDS 10354
Beckwith 9920
H Stovey 9576
Start 8378.5
McGinnity 8232
Pearce 8073
McVey 7985.5

Grant 7969.5
Suttles 7696
BRESNAHAN 7301
T LEACH 7105
RIXEY 7054
(Sisler 6567, C Jones 6160, Sewell 5338, Monroe 5149, Ferrell 4549, Williamson 4262, Doyle 4250, Roush 4206, Redding 3946, Mendez 3924)

Notes: Beckley likely will pass Duffy for the No. 3 spot next year.... Griffith will cruise into 5th.. Welch grabs 12th over dormant Ryan, Childs may pass Ryan next year... Beckwith checks out of the 'race' at 15th, only 80 pts shy of 10,000.... There are now more than 3,000 pts between 10th and 11th place among unelected players, from Childs to Bresnahan... No new top 25 members expected for several more years.

Besides Trouppe, we may also need a separate thread for Sam Jethro. He probably needs to be evaluated on the basis of is Negro League career.

Besides Trouppe, we may also need a separate thread for Sam Jethroe. He probably needs to be evaluated on the basis of is Negro League career.

Neither Win Shares nor WARP are remotely accurate; they are themselves attempts to estimate things that are close to unquantifyable, they have "favorite" fielding positions at different times in history, and they overstate the achievements of players on winning teams. Finally, as Chris pointed out: they leave in the "dumb luck" element in a team's season, which with only 154/162 entries, instead of say 600 AB, or batters faced, is far more important than in individual stats.

Depending on what you think is 'luck', this is supposedly the "strength" of Win Shares!

For example, the 1987 Cardinals, based on their and their opponents batting stats, "should" have had a record of 81-81, but actually had a record of 95-67. That's 14 wins, or 42 wins shares, that have to be allocated to the 9 positions on the field, above and beyond what the "stats" would dicate.

The 1984 Phillies are the reverse, as their stats indicate a 95-67 team, but they finished 81-81. That's 42 win shares that have to be subtracted from among the 9 positions.

Again, depending on your point of view, that's either a strength or weakness of win shares, but regardless, that's how they work.

Personally, I think the ultra low replacement level of Win Shares is much more problematic....

Neither Win Shares nor WARP are remotely accurate; they are themselves attempts to estimate things that are close to unquantifyable

That is why I only use TRIPLES! Anything else is just codswallop!

And don't get me started on the way WARP undervalues TRIPLES! Everyone knows they are the most exciting play in baseball, so they should be given even more extra-special double-choclatey goodness value. I am chomping at the bit for the day I can place Lance Johnson and Christian Guzman at the top of my ballot.

EricC is correct, I am also an Eric C, but not THE EricC.

"Again, depending on your point of view, that's either a strength or weakness of win shares, but regardless, that's how they work."

Good explanation KJOK - I consider it a strength, I can see why others don't.

"Personally, I think the ultra low replacement level of Win Shares is much more problematic.... "

I'd agree, though it isn't all that problematic to correct for, just subtract 6-10 WS (depending on where you think replacement level should be placed - I think TomH uses 12 - too high for me) for each full season. WARP has the same issue though, it just isn't quite as low. Generally I'll deduct 8 WS or 2 WARP (about 6 WS) from each player to get a realistic replacement level.

Dizzy Trout has a better resume than I thought. Seeing that 80+ WARP made me take a second look.

His 1944 is freaking amazing, even in a war year. 14.4 WARP3 is a pretty damn good season in modern times for a pitcher, and that takes into account the quality of competition issues. His 1946 is nothing to sneeze at either. Throw in 7 other years of at least 5 WARP3 and you've got someone that should be somewhat intriguing to peak voters.

ChrisJ - think you could RSI him??? A 124 ERA+ does not compute to a 170-161 record. Prospectus has his translated record at 196-123.

Attempting to clarify the win-shares issue here:

KJOK addressed the "actual wins vs. pythagorean wins" issue in win shares:

Depending on what you think is 'luck', this is supposedly the "strength" of Win Shares!

The data I've been analyzing actually point towards quite a different problem, one which jimd has raised as a problem with win-shares career evaluations, though not as a problem for seasonal evaluations. This is that win shares systematically overrates players on good teams and underrates players on bad teams, because it does not adjust for the fact that good teams don't have to play themselves.

I’ve done some more analysis of the data on 175-185 OPS+ seasons between 1907 and 1967 (57 player-seasons put up by 31 different players), comparing win-share values to team performance. Having adjusted the win-share/game rates for pinch-hitting appearances, I grouped the players’ rates against team winning percentage, categorizing three groups of player rates--the 10 lowest rates, the middle rates, and the ten highest rates—by team record, again divided into three groups--below .500 teams, .500-.625 teams, teams above .625.

Win %   Low 10  Mid 37  Top 10
<.500      5      3       0
Middle     5     27       4
>.625      0      7       6

There seems to be a strong correlation between team performance and the win-share rates for players with similar high levels of offensive production. Someone with more sophisticated statistical skills than I could tell us the probability that this pattern shows a significant correlation, but I’m pretty sure that it does.

Here’s the data regrouped to look at the Low 10 and Top 10 bws/g rates vs. team wp.

Group   <.500   .500-.625   >.625
Low 10    5         5          0
Top 10    0         4          6   

The low rates range from .174 ws/g to .202 ws/g; the high rates range from .237 ws/g to .261 ws/g.

Now, we should expect to see some correlation between team performance and the performance of its top players, so perhaps the correlation we see here is appropriate. To test this hypothesis, I sorted OPS+ scores and EQAs for these 57 seasons in the same way: separating out the top and bottom groups and comparing them to team winning percentage. Here are the tables for that study:

OPS+   <.500   .500-.625   >.625
Low 16   2         14         5
Top 9*   2          3         4

(Because OPS+ numbers are heavily rounded, I was unable to get neat 10-player groups. Low 16 are OPS+ 175-6; Top 9 are 185-4)

EQA    <.500   .500-.625   >.625
Low 10    1        7          2
Top 12    2        6          4

(Low 10 range from .336 to .342; Top 10 range from .361 to .354)

These tables support the idea that there is a correlation between what OPS+ and EQA+ measure as fairly small differences between great offensive performances and team winning percentage. The top years appear more frequently on teams with top winning percentages, than do the low years, though not to the degree shown in win shares. It’s important to note, however, that the low years show up with the bad teams at about the same rate as the top years: a rate in keeping with their percentage of the overall seasons.

I think it is important that, while there is a correlation between great team performance and greater seasons here, there is no observed correlation between poor team performances and slightly less great seasons. That makes intuitive sense to me: the better seasons will have an impact, but occasionally a player will have a great season for a bad team. If the team is bad, the difference between putting up a 175 OPS+ and a 185 OPS+ isn’t enough to change the team, but if the team is good, the difference between these types of seasons may be enough to push the team from good to great.

I conclude that, because great seasons earn such a large percentage of their teams’ value, the competition issue in win shares (plus the distortion that starts to appear at high and low winning percentages anyway due to the non-linear relation between runs and wins) shows up very strongly in these seasons, so that the effects are noticeable even at the level of single-season performance.

But, you may ask, so this pattern is discernable. But does it really matter? How big are these differences, anyway?

I’ve done a number of comparisons of player-seasons on this, which I haven’t had time to write up neatly yet, but I’ll give one example that I think shows that in particular cases the differences can be large enough to affect player ratings. I could give you another 4-5 very similar examples just from this one data set.

Here are two seasons put up by Jimmie Foxx:

1935 – 147 g (7 ph), 182 OPS+, .358 EQA, 26.3 bWS, .185 ws/g, team wp .389 (-3 pyth)
1938 – 149 g (0 ph), 182 OPS+, .357 EQA, 32.7 bWS, .219 ws/g, team wp .591 (+1 pyth)

Now, in assessing these two seasons by Foxx, we have two options. One, we can conclude that there were all kinds of factors not captured in either OPS+ or EQA that made Foxx’s performance 15% worse in 1935 than it was in 1938, despite the fact that he is the same player with the same skill set in both seasons. Two, we can conclude that the win shares analysis greatly exaggerates the difference between Foxx’s performances in these two seasons. The win shares system may be correct that Foxx’s performance had less much value, as measured in wins per unit of performance, in 1935 than in 1938, but I submit that it is a mistake to attribute that difference to Foxx’s merit as a player. Rather, it derives chiefly from contextual differences in competition level, differences in team ability to turn Foxx’s production into wins (runs/wins ratio non-linear), and luck.

Now, the difference between a 182 OPS+ season being worth 27.5 bws (Foxx’s 1935 value projected into 149 full games) and 32.7 bws (Foxx’s 1938 value) doesn’t really matter much for Foxx. He has 6 or 7 other seasons as good as these or better, most of them on good teams. But for players who spent a long time on bad teams (Bob Johnson, Ralph Kiner), losing up to 5 win shares off your top seasons in comparison to a player with the same statistical line on a good team could matter a lot. For players who have one or two great seasons (say, Joe Medwick and George Sisler), it matters a lot whether you are get those 5 win shares or not. A difference of 10 win shares in your three-year and five year peak is going to affect how you are rated! A difference of 5 win shares between two player-seasons is one that James would say, I think, is meaningful – not within the margin of error in the ws system. But I submit that we in fact take a considerable risk of being misled by systemic error if we assume that win-share value equals merit for great seasons, unless we cross-check against measures that are less strongly influenced by team context.