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

Thursday, January 06, 2005

Credits and Deductions for WW II Players

David Foss felt that a thread should be created so we can figure out what we need do about the wartime players.

I concur.

John (You Can Call Me Grandma) Murphy Posted: January 06, 2005 at 12:30 AM | 201 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 3 pages  < 1 2 3 > 
   101. Joey Numbaz (Scruff) Posted: May 24, 2005 at 10:07 AM (#1357659)
Just to anticipate a question on the 'excluded' phrasing, with regard to players that were banned from the game . . . we cover that specifically in a later paragraph.
   102. KJOK Posted: May 24, 2005 at 10:32 PM (#1358790)
"This has nothing to do with ability over value."

What??? It has just about everything to do with ability over value. Williams and Musial had close ability, but Musial was many times more valuable to the Cardinals during WWII eran than Williams was to the Red Sox.

"It's credit for what they would have done, had WWII not happened."

It's a little bit different from the Negro League players in that we factor in those player's playing time, missing time for injuries, etc. when we do MLE's for Negro League players.

For WWII players, we don't know if they would have been injured or missed playing time for some other reason if they had been in MLB during those years, but it's statistically likely that some of them would have, so you can't just assume 'what they would have done' just based on what they might have done before or after WWII.
   103. The definitely immoral Eric Enders Posted: May 24, 2005 at 10:47 PM (#1358818)
"It's credit for what they would have done, had WWII not happened."

Exactly. And the Negro Leaguers get credit for what they did do, not what they would've done. Therein lies the difference.
   104. David C. Jones Posted: May 24, 2005 at 10:49 PM (#1358822)
Yeah, sorry Joe, but I completely disagree with everything you have to say on this issue, including your interpretation of the constitution, even though you wrote it.

With Negro Leaguers, we evaluate their performance. Period. With WWII players, some people are guessing as to what their performance would have been had they been allowed to play?

The error in your thinking is that you are assuming that the Major Leagues are THE STANDARD, and that every player excluded from the majors for whatever reason is equivalent. You are basically asking me to pretend that the Negro Leaguers didn't actually play, but that we should evaluate them as if they did play. Well, that's just wrong. We evaluate the Negro Leaguers based on what they did; not based on what they might have done. At least, that's how I evaluate Negro Leaguers. Others have chosen to evaluate them based on what they would have done if their style of play had been forced into a major league context. This is an approach I have little to no sympathy for, as it completely distorts history, i.e., what actually happened.

So, to summarize, you are wrong about everything on this issue. Sorry.
   105. David C. Jones Posted: May 24, 2005 at 10:50 PM (#1358824)
Sorry, there should have been no question mark after "play" in the second paragraph of the last post.
   106. David C. Jones Posted: May 24, 2005 at 10:55 PM (#1358836)
This has nothing to do with ability over value. When you are comparing players across eras, you have to adjust for things that are different in each era. Adjusting for players missing time during WWII is no different than adjusting for the fact that more HR are being hit in 2005 than in 1914.

That is an absurd comparison. In the one case you are comparing the value of a home run in one era with the value of a home run in the other, in the other instance you are making an educated guess as to what a player WHO DIDN'T ACTUALLY PLAY BASEBALL "would have done." If rejecting that reasoning is borderline unconstitional, then the constitution needs to be amended.
   107. Jim Sp Posted: May 25, 2005 at 12:46 AM (#1359369)
This was also meant to include others that are 'excluded', such as players who miss time due to player strikes or lockouts.

Wait a minute...does anyone else have a problem with that?
   108. karlmagnus Posted: May 25, 2005 at 12:59 AM (#1359434)
Mr Jones seems to be looking for a $5,000 fine for arguing with the Commissioner!

Seriously, I agree with Joe here, and indeed that's how we've always treated it, including giving WW1 credit, so it's an issue that's already been around for 20 "years." If we can convert stats from the NEL to the majors, we can interpolate on a conservative basis between 1941-42 and 1946-47 -- mathematically, it's a much solider procedure.

The moral beauty of being the wrong color for pre-47 MLB may be exquisite, but it's no more exquisite than that of going off to fight to save civilisation, even if you were drafted. NEL and WWI/II are in my view qualitatively different from being stuck in the minors, kept out of ML baseball by its temporarily lousy economics, etc.
   109. Kelly in SD Posted: May 25, 2005 at 02:33 AM (#1359730)
The Constitution does NOT need to be amended. People spent a hell of a lot of time ironing out the Constitution. I wasn't here when it was being set up and I have some issues with certain things - especially Black Sox eligibility. But so what. I read the Constitution before I started this activity, I knew what the rules were, and I still decided to play the game. By participating in this activity, we tacitly agree to play by the Constitution's rules, even if we disagree with them.
That is my 2 cents. I am assuming I am a hardliner about this, but that is my opinion.

So, are voters allowed to interpret the Constitution as they see fit and vote accordingly, or do they have to follow the Constitution? And what does the Constitution require?
Specifically asking the founders.
   110. Chris Cobb Posted: May 25, 2005 at 03:34 AM (#1359842)
I've said my say about the principles of awarding compensatory credit for missed time due to WWII, so I won't repeat what I've said before.

On the subject of constiutionality, I would add three points, however.

1) A voter who does not make some effort within his or her rankings to ensure that players from the 1940s are not underrepresented in comparison to other eras is clearly violating the spirit of the Constitution.

2) The Constitution defines major-league baseball as the primary concern of the HoM in its discussion of eligibility.

3) The Constitution does not mandate any form of compensatory credit, and its language even discourages assigning such credit. This language runs against the Constitution's statement of purpose on electing the best players from each era.

Thus, I think the Constitution as written supports both the awarding of compensatory credit and the withholding of compensatory credit. I think that the principle of fairness to all eras is the crucial point, and that as long as voters try to adhere to this principle, the specific methods that they use to do so are at their discretion, as long as they are reasoned.

Some selections from the Constitution, with comments:

We will start with the 19th century players on the first HoM ballot, and then step through baseball history one year at a time. Our goal is to identify the best players of each era and elect them to the Hall of Merit.

If a denial of compensatory credit for missed playing time were to lead to signficantly fewer of the best players of the 1940s era being elected, that would certainly violate the spirit, if not the letter of the Constitution's statement of the HoM's goal.

Eligibility:

All major league players are eligible for the Hall of Merit. Also eligible are all “excluded” players, most notably Negro Leaguers, and pre-MLB players. Following the timing of Hall of Fame ballots, players are generally eligible for the Hall of Merit five years after their last MLB (or equivalent) season.


The eligibility statement makes it clear that the focus of the HoM imitates the focus of the HoF, which is major-league baseball. Players who were _excluded_ are also eligible, but because the focus of the HoM is major-league baseball, it is clear that major-league baseball is, generally, the baseball that matters to the project and the level of competition that sets the standard for evaluation.

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

This sentence argues against compensatory credit for excluded players who were not actually playing baseball. Since the sentence that follows it specifically excludes _managerial credit_, I think that the framers were thinking about managers and coaches here, not about excluded players, so the force of this sentence with respect to compensatory credit is weaker than it appears in isolation.

Since voters are "strongly encouraged" and not required, this statment does not forbid compensatory credit any more than the statement of purpose categorically requires it.
   111. Chris Cobb Posted: May 25, 2005 at 03:39 AM (#1359848)
KJOK wrote:

For WWII players, we don't know if they would have been injured or missed playing time for some other reason if they had been in MLB during those years, but it's statistically likely that some of them would have, so you can't just assume 'what they would have done' just based on what they might have done before or after WWII.

KJOK, are you awarding WWII credit? If so, how are you taking into account the probability of injury and suchlike?

This is what I am currently trying to decide how to do. Right now, my plan is to handle it similarly to the regression I've been using for the NeL MLEs: take the average of the player's playing time for 2 years before the missed seasons and 2 years after (if available) to set playing time, setting the player's WS/game based on their rate over those same 4 years, and then deducting 15% from the total as a nod to uncertainty.

I'd be very interested to hear how you, and others, are planning to award compensatory credit.
   112. Kelly in SD Posted: May 25, 2005 at 04:09 AM (#1359900)
Chris,

Thank you for your response.

Regarding WWII credit, I used a similar system to yours, except I used, if available, the three previous seasons and three seasons after. I weighted the years as follows: 3rd year away as 1, 2nd year away times 2, year closest times 3. Add the totals and divide by 12. Then fill in the blanks.

As for players who played while others were away, I looked to see how much their performance was out of character for their career - using a similar system as above.

I am still wondering about a deduction for uncertainty. So far I have not incorporated one, but it is a factor I want to incorporate.

So far Charlie Keller has proven to be the most difficutl to deal with because he had a fantastic year in 1946, then his congenital back problems surfaced and he was never the same. Any thoughts?

Augie Galan is another problem player. He had a 30+ WS season in the 30s, battled injuries forever, then was healthy in WWII and had 2 more 30+ seasons.

<sarcasm alert> Lastly, since many or most baseball players in the military played in military leagues or exhibitions, should we try to find their records during their games against other major leaguers and do MLE's? After all, they were playing other major leaguers (or minor leaguers). <sarcasm alert over>
   113. Paul Wendt Posted: May 25, 2005 at 09:57 PM (#1361331)
Phillybooster #5
there certainly seem to be a lot of players in 1946 and 1947 who, after missing a few years, came back and pitched 50-100 innings very well (ERA+ over 100), but then were done.

Did the demilitarization rules require their old clubs required to employ them for a time at their old salaries? If able to pitch only 100 innings, they may have been overpaid, relative to baseball value or relative to young players. Did they find good work elsewhere, one or two years after military service? Good explanations may be very complicated.

Dr. Chaleeko #15
one can argue about the phallacy of the predetermined outcome

what is that?

Phillybooster #48
the "slope" for baseball players has always been relatively steeper in the majors than in the minors. The difference between a median and replacement level Major Leaguer is a lot bigger than between a median and replacement level AAA-level player.

I suppose this is true of expected player values in a reasonable axiomatic model of organized baseball. There is more uncertainty in the minors, hence maybe greater differences between expected and actual values, hence maybe a greater range of actual performances in the minors.

CarlG #73
I don't think most people giving war credit are crediting players with the best years of their careers while they were gone.

I think you will see this with Country Slaughter.

Chris Cobb #
Edd Roush is the only serious candidate at present [scap] that I am aware of who is affected by hold-out considerations. Are there any others I should know about?

I don't know of any. Several voters think Amos Rusie should be another, but he is in the HOM.

. . . In cases of black-listing because of contract disputes, I do give credit. Charley Jones is the [only scap] affected by black-listing considerations, as far as I know.

Joe Jackson is in the HOM, Pete Rose in the future. Eddie Cicotte is the next scap after Jones, I think.

A few years ago, Mike Crain with some help from SABR 19c Cmte members compiled a list of 60 men banned from employment in organized baseball.
shady.csv (one of the baseball-databank Yahoo Group Files)
The most prominent other names on that list are Hal Chase, Candy Cummings (no action), Jim Devlin, George Hall, Benny Kauff, Mickey Mantle, Willie Mays, Lip Pike (HOM), Heinie Zimmerman, George Steinbrenner.

Brent #96 (and Joe Dimino #100 concurs)
"extra credit" - I dislike this term. First, there's the school grading connotation, which does not match how I think about assessing merit.

Several voters maintain formal systems derived from Win Shares and its cousins. Career "pennants added" is a sophisticated example. Negro League players &c get credit that is "extra" because it is outside the formal system.
   114. KJOK Posted: May 25, 2005 at 11:52 PM (#1361696)
KJOK, are you awarding WWII credit? If so, how are you taking into account the probability of injury and suchlike?

Yes and no is probably the most accurate answer. Yes, IF I'm comparing two contemporary players at the same position, such as Gordon and Pesky, I do discount Gordon's 1943 season by about 10-15% and "give" Pesky about 85-90% of what he would have been expected to contribute.

But in general, for WWII players, I evaluate them similarly to how I evaluate Frank Chance or John McGraw, which is "did they exhibit that they were one of the best players on the field when the DID play." If the answer is yes, then they'll make my ballot WITHOUT any explicit "WWII credit".
   115. Howie Menckel Posted: May 31, 2005 at 03:05 AM (#1372671)
(crossing over from the Willie Wells thread)

Wow.

WAY too emotion about the WW II credit, if you ask me.

I've mentioned how much I respect the balance we've found re Negro Leaguers. It doesn't take much to clench one's fists and grit one's teeth thinking about the unfairness of the whole thing.
But we're not doing this to be bleeding hearts, and we're also not doing this to be automotons who can't handle voting for players who don't have precise stats for us to work with.
We do the best we can with what we have. I can't say for sure that we've picked exactly the right players, but we're surely offering the fair chance they never got.

I hope we can do the same with war-affected careers. The issue isn't about the (often im)morality of war, it's about figuring out the best players in history. Mostly we look at actual results, but there are other ways to go in some circumstances.

I don't box myself in in with exact numbers; maybe that helps me here.
When I get to Hank Greenberg, for instance, I'll see if he's in the ballpark of HOMiness. He'll get neither 'full credit' from me for missing time nor non-credit for the war years. I'll factor in the raw truth of the era, as I will somewhat discount those who played with so many great ones not in the picture.

Not sure that politics is going to enter into anyone's voting, but I sure as hell hope not. There are plenty of other places to vent those spleens, one way or another.


C'mon, guys. Let's keep on getting this project right, for the right reasons.
   116. Howie Menckel Posted: May 31, 2005 at 03:11 AM (#1372683)
David Jones,

I like a lot of what you bring to this project, so I'll assume that the tone produced by your messages is not at all that what you have intended. Please tell me that's true.
   117. John (You Can Call Me Grandma) Murphy Posted: May 31, 2005 at 03:48 AM (#1372734)
He'll get neither 'full credit' from me for missing time nor non-credit for the war years.

That's what I'm doing also, Howie. I'm not going to project those missing seasons as if they would have been his finest seasons prior to military service, but it's going to be a hell of a lot more than what Snuffy Stirnweis actually did because the latter happened to be lucky enough to be 4-F.
   118. Paul Wendt Posted: June 01, 2005 at 05:58 PM (#1374378)
all quotation is from the Willie Wells thread

Joe Dimino, quoting David Jones, Willie Wells #86
"Finally, this reflexive instinct to give players "war credit" carries the faint odor of #### patriotism to me."

This has nothing to do with it. For me at least (and I'm sure most others).


Joe, the next paragraph undermines this protest. Indeed, it reminds me that you have argued the general case for war credit in terms of 'defending his country' time and again for three years. That is the case for giving extra credit (a term favored by one side) rather than docking or punishing (terms favored by the other side).

In real life I'm pretty much against all war, except for things like Hitler where you have to stop a madman. I'm against the war in Iraq. But if Khalil Greene is drafted or enlists and loses three prime years defending his country, I'm not going to dock him relative to Dave Concepcion for this.

jschmeagol
I also don't think this has anything to do with whether or not you agree with war, WWII, or what your political leanings are.

I'm not sure whether that is a personal statement of values or an empirical claim about the HOM. As an empirical claim, the proper reply is
This is an election. It matters to some voters and not to others.

or perhaps
This is a discussion. It matters to some discussants.

Kelly San Diego
The voters for the HoM have generally adopted one of Bill James' comments: "I make adjustments for any player who is clearly a major league player, but who is prevented from playing in the major leagues by forces beyond his control."

That is an empirical claim, I'm sure. I think it's true, if generally means a clear majority.
   119. Mark Shirk (jsch) Posted: June 01, 2005 at 06:03 PM (#1374383)
Dragging the Willie Wells discussion back over here again...

The one difference between a religious mission, enlisting in the army, etc. and WWII is that these players had no choice. There choices were a) get drafted and be assigned to a post, or b) sign up early and get to choose your branch, post, etc. There wasn't anyone who said to the military, "No, I would like to continue playing baseball." and then did continue to play baseball. Those playing were the ones who were either a) too young, b) too old, or c) were classified as 4F.

A religious mission is just as important as a military one but it is a choice. There won't be any cops or army personnel or whatever banging down your door if you decide not to go. Enlisting in the army is different as well because that is a choice. WWII wasn't really a choice.

One interesting thing to think about is if we would have given credit to say, Hank Greenberg had he resisted being drafted and was put in jail for his stance for the years 1941-1945 (or maybe longer). Right now I may be inclined to give him credit, assuming he was allowed to have a career after the war.
   120. Paul Wendt Posted: June 01, 2005 at 06:28 PM (#1374421)
(continued without reading anything posted in the meantime)

Bill James' particular list of 5 adjustments must be understood in terms of his particular system --the answer to the question, adjustments to what?-- where official major league baseball is fundamental. Is omlb fundamental by his conceptual framework or his dataset? Maybe both. Anyway, he doesn't have Win Shares for NAPBBP and NNL seasons, so two of his adjustments concern pre-1876 baseball and segregation-era black baseball.

If official mlb is conceptually fundamental for someone, then the nature of some adjustments is to give credit for hypothetical official mlb play. Where to draw that line is a crucial question. Why give credit for hypothetical omlb play to this group and not to that one? Eg, to Negro National League players but not to established omlb players occupied in military or other war work?

If the NAPBBP or some of the major Negro Leagues are conceptually par with the official majors then no argument for considering play in those leagues is necessary. Along these lines, one is not giving major credit to one group outside the majors, but simply giving credit for major play.

Anything so brief as those two paragraphs is extremely limited, of course. That said, I think the former fits several HOMeboys reasonably well and the latter fits a few reasonably well.
   121. John (You Can Call Me Grandma) Murphy Posted: June 01, 2005 at 06:49 PM (#1374449)
One interesting thing to think about is if we would have given credit to say, Hank Greenberg had he resisted being drafted and was put in jail for his stance for the years 1941-1945 (or maybe longer). Right now I may be inclined to give him credit, assuming he was allowed to have a career after the war.

Using that scenario, I would still give him credit. Anyway you slice it, he wasn't going to be able to play ball whether he had served in the military or was serving a jail sentence for those years.

Now, if Greenberg had decided to join up with the Abe Lincoln Brigade to fight Franco during the mid-thirties, I wouldn't give him any credit. He would have made a conscious choice without being coerced into it.
   122. sunnyday2 Posted: June 01, 2005 at 07:14 PM (#1374504)
Without tooting my horn too much here, I think my rule about individual experiences versus experiences that affect a whole cohort works here. WWII affected everybody. The Abe Lincoln Brigade affected a few who made an individual choice and had an individual experience.

The idea that a player might have resisted the draft or whatever is a tricky one, I gotta admit. But I would say that that would be an individual decision and an unusual, individual experience. Part of the bigger social phenomenon, true. But not an experience that the masses were subject to. But purely a hypothetical anyway. (Oh oh, that word.)
   123. John (You Can Call Me Grandma) Murphy Posted: June 01, 2005 at 07:26 PM (#1374535)
Without tooting my horn too much here, I think my rule about individual experiences versus experiences that affect a whole cohort works here. WWII affected everybody. The Abe Lincoln Brigade affected a few who made an individual choice and had an individual experience.

That encapsulates my thoughts pretty succinctly, Marc.
   124. Michael Bass Posted: June 01, 2005 at 07:45 PM (#1374578)
While I certainly don't wish to cut off the discussion, especially important and interesting discussion about the level of war credit to be given to specific players (and possible deductions from "holdback" players)...

I think the discussion of whether or not to give war credit on the whole is played out, and has now reached the point of emotional sniping back and forth. I don't feel further discussion on the issue is going to change anyone's mind, and simply will lead to further unpleasent exchanges.
   125. Thane of Bagarth Posted: June 14, 2005 at 06:09 PM (#1403829)
In case anyone is interested, a while back I calculated the ratios of WARP2/WARP1 and WARP3/WARP1 for 1940-1950. I have not updated the data for the latest WARP changes. I was hoping it would provide some insight into changes in league quality during wartime.

Given that I recall someone mentioning a Clay Davenport article looking into the effect of the war and estimating a 10% or so drop in EQA, the decline in AL "league quality" seems severe, particularly in '45.

The first set of numbers is broken out by league. The second set is both leagues combined. In each case I just added all of the team totals for WARP2 & WARP3 and divided by the sum of the team WARP1 totals.
Year NL W2/W1  AL W2/W1  NL W3/W1  AL W3/W1
1940  0.971     1.029     0.994     1.053
1941  0.948     0.993     0.967     1.013
1942  1.058     0.826     1.087     0.850
1943  0.935     0.845     0.955     0.866
1944  0.791     0.708     0.807     0.725
1945  0.794     0.617     0.817     0.634
1946  0.945     0.979     0.965     1.000
1947  1.030     0.980     1.052     0.999
1948  0.986     1.035     1.009     1.059
1949  0.996     1.006     1.016     1.030
1950  0.988     1.002     1.012     1.021 

      W3
/W1  W2/W1
1940  1.023  1.000
1941  0.990  0.970
1942  0.969  0.942
1943  0.911  0.890
1944  0.766  0.750
1945  0.727  0.706
1946  0.983  0.962
1947  1.026  1.005
1948  1.034  1.010
1949  1.023  1.001
1950  1.016  0.995 
   126. EricC Posted: August 27, 2005 at 10:11 PM (#1578301)
I tried to answer my own question (#49) on the percentage of players during WWII that were major-league quality. The methodology was to look at how many regular players in year X were still regular players in year X+N. If, for example, 50% of the regular players in 1937 were regular players in 1940, but 40% of regulars in 1940 were regular players in 1943, that would be evidence that about 80% of the players in 1943 were major league quality.

I compared sets of years for N <=5 and for X from 1937-N to 1945-N to come up with some numbers. I wanted to use post-WWII years in the comparisons too, but the trends were so different that they suggest the normal player development cycle was affected by the war.

What I came up with was:

1942: 89%
1943: 78%
1944: 68%
1945: 62%

Much higher than Bill James' 40% estimate, even for 1945.

Note that these numbers are NOT "adjustment factors" for performance.
   127. Howie Menckel Posted: August 27, 2005 at 11:23 PM (#1578420)
Eric, that about squares with my more informal sense of things. I ignore any 1942 discount, do just a little for 1943, but then much more for 1944 and then 1945.
Particularly problematic for me are those whose best year (or two years) are either in 1944 or 1945; Boudreau would have been off my ballot entirely instead of 10th if not for his big 1948.

Still, I mainly like to work the discounts (and creidts) to compare war-era players. It's a little bit dicier to do it vs. a guy from the 1920s-1930s, since only one of the two gets an 'automatic' discount.
   128. Dr. Chaleeko Posted: February 24, 2006 at 06:59 PM (#1874017)
Hey, everyone,

I've come across two references lately to players who served in the military during the Vietnam era (usually stateside): Bobby Murcer and Ken Holtzman. In addition, I believe Al Bumbry served during this time.

Are there any other players who served during this era?

War credit could make a difference for Murcer who has enough peak to attract a following. He put in two years I think. Holtzman gave Uncle Sam about 1/2 of 1967 (his 9-0 year), so his creadit should be simpler to figure out. Don't know much about Bumbry except that he debuted at 25.
   129. KJOK Posted: February 24, 2006 at 07:50 PM (#1874109)
There are a whole list of players who had at least SOME service time in the late 60's and early 70's. For example, Ted Simmons missed some weekends due to National Guard duty, although in his case it's probably not enough to matter.
   130. KJOK Posted: February 24, 2006 at 07:59 PM (#1874120)
Here are the guys who retired after 1964 that has "significant" military duty:

BillCampbell
DustyBaker
BobJones
GarryMaddox
AlBumbry
JerryKoosman
JimBibby
BobbyMurcer
EdFigueroa
CarlosMay
DaveSchneck
ChuckGoggin
WillieMays
PhilHennigan
JimMagnuson
MelQueen
HoytWilhelm
ErnieBanks
GeneBrabender
EddieMathews
GeneMartin
SmokyBurgess
CurtSimmons
WhiteyFord
WaltBond
DelCrandall
RobinRoberts
WarrenSpahn
CarlWilley
NellieFox
YogiBerra
DukeSnider
BillyPierce
HalBrown
CalMcLish
   131. Brent Posted: February 25, 2006 at 04:37 AM (#1874592)
I had missed the fact that Billy Pierce spent time in the military. Was it his age 19 and 20 seasons?
   132. Howie Menckel Posted: February 25, 2006 at 06:17 AM (#1874661)
I was a Mets fan as a kid in the late 1960s.
I seem to remember Tug McGraw and/or Nolan Ryan doing some time in the reserves, probably both.
The papers would always show the guys in 'boot camp,' doing chin-ups and such in t-shirts and fatigues.
I have a scrapbook somewhere with some of these photos.
   133. KJOK Posted: February 25, 2006 at 06:47 AM (#1874679)
I had missed the fact that Billy Pierce spent time in the military. Was it his age 19 and 20 seasons?

No exact info listed for Pierce, other than it was WWII, which would certainly be around 1945 I would think.
   134. DanG Posted: February 25, 2006 at 01:29 PM (#1874734)
Pierce was born April 1927. Per Who's Who in Baseball, his 1945, split between Buffalo and Detroit, totaled 20 G, 93 IP. In 1946 with Buffalo he had 10 G, 56 IP. If he spent time in the military, it wasn't much. He was up in MLB to stay in 1948 at age 21, so it's hard to see him getting any extra credit.
   135. Dr. Chaleeko Posted: February 25, 2006 at 06:18 PM (#1874894)
Davey Lopes apparently did a couple years too per the Biographical Encyclopedia, I think between A and AA, pushing his debut age back to 27.

So far Murcer, Koosman, Bumbry, and maybe Lopes look like the most important players we'll need to consider for possible war credit during Vietnam. Maybe Holtzman if anyone thinks the half-year would make a difference.
   136. sunnyday2 Posted: March 14, 2006 at 12:32 PM (#1898140)
Per Joe's comment on the 1972 ballot thread, I'd love to see somebody list the 5 players (or so) who gain the most by getting full WWII credit. Show their record before (WS, PAs, etc.) and after and where they rank among the backlog at their position with and w/o the credit. Something like that, if anyone wants to advocate for some of these players. I'm willing to reconsider the issue.
   137. andrew siegel Posted: April 24, 2006 at 03:00 PM (#1985170)
This past week, I just finished running all the conceivable candidates for WW II credit through the objective part of my system with full credit for seasons lost to war. My system is WS-based and is divided 50-50 between career value and 7-year non-consecutive peak, so take it FWIW. Anyone five players moved from the not qualified to the qualified category based on war credit: Charlie Keller, Joe Gordon, Phil Rizutto, Johnny Pesky, and Mickey Vernon.

That having been said, I will only be voting for 2 of the 5 guys.

(1) Pesky's results are something of a fluke--based on his age and the timing of his best two seasons, my methods project him to have had his 2nd, 3rd, and 5th best seasons during the war. I'm not willing to project nearly an entire peak. He falls into the might have been category.

(2) Vernon goes over the line because his career value becomes astronomical. He is a better candidate than I thought, but his career value is overrated by lots of mediocre seasons. I subjectively lower the rankings of guys who accumulate high career WS totals in that way. Doing that drops Vernon well out of my PHoM.

(3) The Scooter is a tougher case. With full war credit, he end up with 1.02 points on a scale where 1.00 is meant to signal HoM-worthiness. His case is similar to Pesky's in that I am projecting prime seasons (though in his case, not peak). Between the fact that I am reluctant to do that, the fact that he barely crosses the line even with that help, the WS advantage of playing for the Yankees, and the fact that much of the value of his prime comes from one slightly fluky season, I can't quite support the Scooter. Still, he is now somewhere in the late 20's for me and might make my ballot one day.

(4) Joe Gordon--I've already been giving him full war-credit and slotting him mid-ballot. Running the numbers again confirms my general ranking of him, though it does drop him a few slots on my ballot.

(5) Charlies Keller--I missed the balll on him. He scores at .65 on my system without war credit, but moves to 1.23 with full war credit (and 1.45 if you give him credit for his last minor league season). I think he needs war quality and Yankee inflation discounts, but even applying those, he scores out at 1.10 (or 1.33 with minor league credit). He'll be in the top 10 on this year's ballot.
   138. Joey Numbaz (Scruff) Posted: April 26, 2006 at 06:37 AM (#1989870)
"(2) Vernon goes over the line because his career value becomes astronomical. He is a better candidate than I thought, but his career value is overrated by lots of mediocre seasons. I subjectively lower the rankings of guys who accumulate high career WS totals in that way. Doing that drops Vernon well out of my PHoM."


Huh? That makes no sense. You are saying that if Vernon was injured for those years, instead of fighting the war, he'd score higher?????

If that's the case, I believe your system is flawed.

Also what is this 'WS advantage of playing for the Yankees'?

If anything, playing for great teams would hurt you, in that you can only win so many games, and you have all of the great players taking WS from each other.

WS is much less affected than other systems for not facing your own pitchers or hitters. None of the other players on your team got to face those hitters and pitchers, and those are the guys you are competing with for a finite number of Win Shares.

I think the effect of facing your own pitching/hitting is a red herring with regards to Win Shares.
   139. DL from MN Posted: April 26, 2006 at 05:38 PM (#1990337)
Does Dick Bartell deserve any war credit? He's not far from my ballot (thanks to whomever pointed him out in comparison to Sewell) and a little war credit would put him on next year.
   140. Chris Cobb Posted: April 26, 2006 at 05:55 PM (#1990373)
Huh? That makes no sense. You are saying that if Vernon was injured for those years, instead of fighting the war, he'd score higher?????

Andrew can certainly speak for himself, but I believe what he was saying is that he docks players who accumulate a great deal of their career value in mediocre seasons, which is true of Vernon. Without war credit, Vernon doesn't have enough value to get into contention. With war credit, he has enough value to get into contention, but that then calls for closer examination of his career, which reveals that he was mediocre _a lot_ of the time, which drops him back out of contention. That's how I interpreted Andrew's statements. Thus, if Vernon had been injured, rather than at war, he would never have come up for serious consideration in the first place.


Also what is this 'WS advantage of playing for the Yankees'?

If anything, playing for great teams would hurt you, in that you can only win so many games, and you have all of the great players taking WS from each other.

WS is much less affected than other systems for not facing your own pitchers or hitters. None of the other players on your team got to face those hitters and pitchers, and those are the guys you are competing with for a finite number of Win Shares.



Joe, you should check out the discussion of these points on the main ballot and WARP v. WS thread, but you are wrong here on both points.

First, jimd has pointed out that the "diminishing returns" problem only kicks in for teams with winning percentages of .750 or better: a player doesn't lose out by "competing" with his teammates at any lower level of performance, because the ratio of runs created to wins is closer to linear. You are not "competing" with your teammates -- all of the teammates are adding value, and getting wins in return in proportion to the value that they add.

Second, they actually get wins more easily than players on other teams precisely because they don't have to face their teammates. By facing easier competition, they create runs more easily, win more, and get more win shares.

_Every_ study I've ever done that takes players with similar levels of performance by any moderately sophisticated study of batting value that adjusts for team context -- OPS+, EQA -- has found that players on better teams, especially players on outstanding teams, consistently earn more win shares for a statistically identical level of performance than players on average teams, and that players on bad teams consistently earn even fewer win shares than the players on average teams.

One might decide that this isn't a problem, that a win is a win, but it is simply wrong to say that playing for great teams hurts a player in a win-shared-based analysis. The opposite is consistently the case.
   141. DL from MN Posted: April 26, 2006 at 06:50 PM (#1990489)
I'll partially answer my own question: Dick Bartell served 2 years in WWII. I'm more interested in opinions of how good he would have been in his early 30's during that hypothetical stretch of his career.
   142. jimd Posted: April 26, 2006 at 07:04 PM (#1990531)
First, jimd has pointed out that the "diminishing returns" problem only kicks in for teams with winning percentages of .750 or better

I do want to qualify this by pointing out that this refers to Batting WS. The effects are worse for DWS, which affects pitchers, mostly.

A balanced .700 team loses 3 BWS to "diminishing returns". A balanced .750 team loses 6 BWS to "diminishing returns". A balanced .800 (1031RF-515RA) team loses 13 BWS to "diminishing returns". (Assumes 162 G schedule.)

At .700, a couple of players get dinged. At .750 most of the regulars get hit for 1 BWS. At .800 players are losing two or more. This affects the Boston NA teams, the best of Cap Ansons's teams, and that's about it, until we start calculating WS from scratch for NeL teams.
   143. sunnyday2 Posted: April 27, 2006 at 12:13 AM (#1991638)
> (#40): he docks players who accumulate a great deal of their career value in mediocre seasons, which is true of Vernon.

This pattern is very much pronounced for almost all long career pitchers. I never "docked" anybody but I certainly have not given a lot of credit for these types of seasons and these types of careers. For whatever reason, I have now come around to selecting Red Ruffing to my PHoM while Early Wynn continues to un-impress me.

I have Vernon ranked pretty low--around #75 maybe. So I still don't value the long career by itself. OTOH if the 7th or 10th best season of a Red Ruffing or a Mickey Vernon is better than the 7th or 10th best season of his peers (however defined), that is worth something. My sense is that Vernon's are not, nor Early Wynn's.

Sisler actually fares OK from this standpoint, though I need to withdraw my recent comment that his post-sinus years are as good as Jake Beckley's shoulder or declining years. They're not.
   144. jimd Posted: April 27, 2006 at 02:13 AM (#1992034)
This pattern is very much pronounced for almost all long career pitchers.

And it's even more pronounced for long career hitters. It has been estimated (by TangoTiger) that WS replacement level is at around .200 WPct for hitters, and around .300 WPct for pitchers. Hitters get even more WS than pitchers do for just showing up.
   145. andrew siegel Posted: April 27, 2006 at 06:33 PM (#1992754)
Thanks to Chris for translating my post.

Let me explain my view of long career players a little bit more. My system takes a player's top 7 seasons (non-consecutive) and scales them so that 140 win shares (scaled to 154 games) is worth nothing to an HoM argument and 225 win shares makes you fully qualified. I then take career WS and scale them so that 200 WS (again adjusted to 154 games seasons) is worth nothing and 400 makes you fully qualified. Adding together your score on both 7-year prime and career value gives me my total objective number. I then adjust for lots of things subjectively. One of the dozens of adjustments I make is to take some air out of the career number if the player used an especially large number of games or outs or innings to get there. This is made necessary by the fact that WS uses an artificially low replacement level. I could have used WS over replacement value as my objective measure of career value and not made this subjective adjustment, but it would have made eyeballing my calculation impossible and wouldn't have made a difference for the vast majority of players. For guys like Vernon and Rice, however, such an adjustment is imperative or their career score will be overblown.

(As should be obvious, this is my system for position players. Pitchers are on a different scale.)
   146. Joey Numbaz (Scruff) Posted: May 04, 2006 at 11:05 AM (#2003537)
Taking another look at Mr. Rizzuto, by way of WARP1 - comparing him to similar HoMers and candidates, only removing their age 25-27 seasons from the discussion.

I'll use WS later, but right now I only have access to WARP. For this, I'll include all ML HoM and HoF shortstops except John Ward, Dickey Pearce and George Wright. I'll also throw in Joe Sewell, Vern Stephens, Rabbit Maranville, Herman Long, Dave Bancroft, Johnny Pesky, Cecil Travis and Dick Bartell. Ernie Banks and Luis Aparicio also will be added, since they are retired in 1976. I'll adjust everyone to a 154 game season and show best 3, 5 and 7 years.

For simplicity the 154 game adjustment will only be for players who played outside the typical 154 game season. I'm not going to adjust 1941 for the Yankees because they played 156 games instead of 154. I will adjust the non-154 game seasons to 155.5 figuring that accounts for the typical team having one or two ties a year in the 154 game schedule.

I gave others reasonable war credit, since we are trying to compare Rizzuto. So Appling gets credit for 1944-45, Boudreau gets docked a little for 1943-45, etc.. I also gave a little bit of a dock for AA seasons, UA seasons. Maranville gets credit for WWI, etc.. These lists drop everyone's age 25-27 seasons.

Let's see how this looks. I can't bold within preformatted text now, it just shows the tags. If anyone knows how to get around this, please let me know. I'll use * instead.

Career WARP1
*Wagner       202.2
*Dahlen       152.1
*Davis        151.0 (credit for 1903)
*
Glasscock    135.6 (docked some in 1884/90)
*
Appling      121.7 (docked 1943credit 1944-45)
*
Wallace      121.6 (doesn't include prior to age 23)
*Vaughan      115.5 (docked 1943, credit 1944-46)
 Maranville    111.4 (credit 1918)
*Reese         93.0 (credit 1943-45)
 Banks          92.7
*Cronin        89.2
 Bartell        88.8 (docked 1943, credit 1944-45)
 Tinker         88.3 (docked 1914-15)
 Bancroft       87.8
 Long           85.7 (docked 1889)
 Sewell         72.5
*Boudreau      70.5 (docked 1943-45)
 Aparicio       70.0
 RIZZUTO        69.1 (no credit 1943-45)
 Travis         60.2 (credit 1942-45)
 Stephens       56.4 (docked 1943-45)
 Pesky          56.0 (credit 1943-45)
*Jennings      53.3 (docked 1891)
 Jackson        49.1 


Best 3 WARP1
*Wagner        48.1
*Glasscock     45.6
*Vaughan       42.8
 Banks          38.4
*Dahlen        37.2
*Davis         35.3
*Jennings      34.9
*Cronin        34.9
*Wallace       33.9
 Bancroft       33.6
*Appling       32.3
*Boudreau      32.3
 Bartell        31.9
 Sewell         31.2
 Maranville     30.0
 Tinker         29.7
 Pesky          28.2
*Reese         27.3
 Travis         27.2
 Long           26.5
 RIZZUTO        26.2
 Jackson        25.1
 Stephens       24.1
 Aparicio       23.3 


Best 5 WARP1
*Wagner        76.9
*Glasscock     71.0
*Vaughan       65.9
*Dahlen        59.4
*Davis         57.1
*Wallace       55.0
*Cronin        53.1
 Banks          52.7
*Appling       52.3
 Bancroft       51.4
*Boudreau      50.9
 Maranville     48.4
 Tinker         48.2
 Sewell         47.1
 Bartell        46.7
*Jennings      43.6
*Reese         43.5
 Travis         43.0
 Long           42.4
 Pesky          41.0
 RIZZUTO        40.7
 Stephens       37.7
 Aparicio       36.8
 Jackson        36.1 


Best 7 WARP1
*Wagner       104.9
*Glasscock     92.1
*Vaughan       82.7
*Dahlen        80.4
*Davis         78.6
*Wallace       75.6
*Appling       69.4
*Cronin        68.3
 Bancroft       66.3
 Tinker         65.8
 Banks          64.8
 Maranville     64.7
*Boudreau      64.5
 Sewell         59.3
 Bartell        57.9
*Reese         57.6
 Long           56.6
 RIZZUTO        54.2
 Travis         53.5
 Pesky          52.0
*Jennings      51.0
 Stephens       48.4
 Aparicio       47.9
 Jackson        44.3 
   147. Joey Numbaz (Scruff) Posted: May 04, 2006 at 11:16 AM (#2003540)
Remember, those lists only apply to Rizzuto. It doesn't work to say that Jennings' 5 year peak was equivalent to Reese, becuase 3 of Jennings' best years aren't included. I think that is applicable when comparing him to Rizzuto.
   148. DL from MN Posted: May 04, 2006 at 02:21 PM (#2003635)
Dick Bartell is looking pretty good on those lists, or at least noticeably better than Rizzuto.
   149. DanG Posted: May 04, 2006 at 02:46 PM (#2003665)
Bancroft looks particularly good here. Especially vs Boudreau and Reese.
   150. sunnyday2 Posted: May 04, 2006 at 03:01 PM (#2003680)
Lemme get this straight. The lists above delete year 25-27 play for everybody? (Rizzuto's already being missing, right?)

Why not just add in WWII credit for Rizzuto and leave everybody else as is? Does Rizzuto still come out as badly if you do it that way?
   151. jimd Posted: May 04, 2006 at 06:08 PM (#2003874)
Why not just add in WWII credit for Rizzuto and leave everybody else as is?

Because there is always a debate about how much credit is appropriate, especially when it comes to the seasons around a player's peak. Sending all the competition off to war at the same age provides an alternative perspective that might be revealing.

The comparisons in the above are only useful as "playerA vs Rizzuto". They have NO value as "playerA vs playerB" (eg Sewell vs Bartell), because we know what A and B both did 25-27, and that is part of their comparison. To move ahead of just Jennings after throwing out most of Jennings' peak does little for Phil, because that was expected. If Rizzuto had moved ahead of Reese and Boudreau, it would have made a much stronger case for Rizzuto, at least in my mind.
   152. jimd Posted: May 04, 2006 at 06:19 PM (#2003886)
Bancroft looks particularly good here. Especially vs Boudreau and Reese.

Because Bancroft was not at his best during ages 25-27. He was an average offensive player with a great glove for the Phillies in Baker Bowl 1916-18. He became a much better hitter after being traded to the Giants and the Polo Grounds, and had his best seasons there, even though his defense was declining. It may have been the park or it may have been coaching, because he brought the improved bat with him to Braves Field when the Giants were done with him.
   153. John (You Can Call Me Grandma) Murphy Posted: May 04, 2006 at 06:50 PM (#2003924)
It may have been the park or it may have been coaching,

...or maybe the changes in the game?
   154. jimd Posted: May 04, 2006 at 09:01 PM (#2004134)
...or maybe the changes in the game?

The big improvement is in 1920, the year he moves to NY. The improvement is largely BA driven. Temporary rise in slugging (triples and HRs), but that only lasts two years; the BA/OBP improvement lasts until his batting falls apart in 1927. A good walker his first two seasons, that part of his game had deteriorated during 1919-20, and came back in 1921.

From NL league statistics, this is a transitional year, closer to 1919 than 1921.
1919 .257/.311/.337 / .648
1920 .270/.322/.357 / .679
1921 .289/.338/.397 / .735

1921 is the clean ball rules. 1920 is the lively ball in the AL (Babe hits 54!); not sure when the NL livens there ball (each league has their own official ball). Also not sure when the spit-ball ban takes effect; that is a more gradual affair anyway as the practitioners retire.

So, yes, it could be that also. But I'm leaning towards coaching.
   155. Dr. Chaleeko Posted: September 17, 2006 at 05:17 AM (#2179274)
So, as mentioned in the, I think, ballot thread for 1985....

File under: I shoulda done this years ago, but I didn't.

So to get a handle on how likely it truly might have been for players to experience an injury that either jeopardized their career or could make them miss substantial portions of a season, I did a little SBE research.
1. I found every player from the years 1932-1941 who had 300 or more PAs. That is, any player who could have been considered a regular.
2. Then I looked to see whether those players had 300 or more PAs in the next season.
3. If they did not I used biographical sources (Biographical Encyc., Baseball Library, 1969 Mac with injury info) to take some stab at what the cause of the drop in playing time was.
4. I noted injuries, ineffectiveness, military, retirement, or other. I did my best here, often injury and ineffectiveness are alike, and there were several players whose biogarphical sources didn't say anything about injury, so I had to list it as ineffective. When in doubt, I listed ineffective. Retirement and ineffective are in essence the same thing, I think, but what the hey. Ineffective is essentially a catch-all for actual ineffectiveness or any non-injury reason why a guy didn't get the PAs the next year, coulda been blockage at position, roster shuffling, whatever.
5. I further noted all seasons where a player played his final game to see how many of those final seasons resulted from catastrophic injury.

So, here's the dope.
-390 players had seasons with 300 or more PA.
-Which includes 1383 individual seasons
-There were 364 instances of players with stoppages in their career.

What was the nature of those 364 stoppages?
-314 of them were stoppages in the midst of a career.
-50 stoppages ended a player's career.

Looking first at the termination points of careers.
-80% of career endings appear to be from ineffectiveness or retirement.
-12% of career endings appear to be from injury.
-8% of career endings appear to be from military service.

So then,
-the likelihood of a player's career ending after a 300 PA season prior to 1942 was about 3.6% (50 endings on 1383 seasons)
-the likelihood that a given player's career would end after a 300 PA season was about 13% (50 endings on 390 players)
-the likelihood of a career-terminating injury appears to be around 0.4% (6 injury endings on 1383 300 PA seasons)
-the likelihood of a career-termintating injury affecting one specific player is around 1.5% (6 injuries among 390 players).

Ineffectiveness (or early retirement) is by and far the biggest reason for a player departing the league after a 300 PA season.

Breaking down the interruptions....
Of the 312 major career interupptions which prevented a player from having consecutive 300 PA seasons:
-78% were from ineffectiveness
-15% from injury
-5% from military service
-1% from converting to pitcher (Walters and Ch. Dean)
-0.3% from skin color (Estallela)
-0.3% from returning to school for a couple years (Billy Sullivan)

Rephrasing this information in terms of the number of 300 PA seasons recorded (excluding terminal seasons):
-18% of all 300 PA seasons were followed by an interruption caused by ineffectiveness
-4% of all 300 PA seasons were followed by an interruption caused by injury

So, executive summary level stuff....
14% of all players had an injury that caused them to fail to repeat a 300 PA season.
10% of those injuries were career-ending.
1.5% of all players had their careers ended by injury.

The limitations of this little study are
-my sources were not thoroughgoing about injuries or didn't even include the player in question
-nagging, nondisabling injuries are often a cause of ineffectiveness and are difficult to seperate.
-it's late.

That said, I think that to err on the side of caution, one could estimate that as many as 20% of players would have an injury that causes them to miss substantial action in a given year and that a maximum of 3%-4% of all players have a catastrophic career-ending injury.

Seem reasonable? Too much? Too little? Too late?
   156. sunnyday2 Posted: September 17, 2006 at 10:21 AM (#2179304)
Now if you can tell us which players would have had those injuries had the NOT been in the military in WWII, we can fine tune our WWII credits accordingly :-)

Seriously the import of this is that by giving no WWII credit, one would over-estimate the likelihood of injury by 5X. We'd be wrong 4 times out of 5.
   157. rawagman Posted: September 17, 2006 at 11:37 AM (#2179314)
I give WWI/II credit when applicable, but as I don't use WS or WARP(s), I simply add the necessary amount of time to my rate/career index.
For those who use WS/WARP(s), maybe it would be best to discount 5-10% of totals added on to career WS?WARP(s) numbers.
   158. Dr. Chaleeko Posted: September 17, 2006 at 03:47 PM (#2179404)
Seriously the import of this is that by giving no WWII credit, one would over-estimate the likelihood of injury by 5X. We'd be wrong 4 times out of 5.

I think that's a fair assessment.

Let's take player X for example.

X's career begins in 1939 and his WS through 1941 are
15
25
30

Then he's in the war for four seasons.

After the war he goes

35
35
25
25
15
done for a total of 205.

If you averaged the four seasons surrounding his military service, you'd get 31.25 WS per year. I think the incidence of career-ending injury is probably an overall wash with the incidence of career-ending shrapnel wound in the services, so take that 20% figure and apply it. That's 25 WS a year. So now

15
25
30
25
25
25
25
35
35
25
25
15
done for a total of 305.

That takes X from borderline Kiner type player to an excellent peak/prime candidate.

OK, let's look at some real players using the simple average of nearest four surrounding seasons (or three if needed). All seasons that are war credits are asterisked.
Slaughter  Williams  Rizzuto  Pesky  Travis  Greenberg  Judnich  S Gordon  J Gordon
    
---------  --------- -------  -----  ------  ---------  -------  --------  --------
1930    --        --       --       --     --         0        --       --       --            
1931    --        --       --       --     --        --        --       --       --     
1932    --        --       --       --     --        --        --       --       --     
1933    --        --       --       --      1        14        --       --       --     
1934    --        --       --       --     10        31        --       --       --     
1935    --        --       --       --     20        34        --       --       --     
1936    --        --       --       --     15         3        --       --       --     
1937    --        --       --       --     22        33        --       --       --     
1938     9        --       --       --     20        34        --       --       19      
1939    23        32       
--       --     13        24        --       --       25   
1940    22        30       
--       --     22        31        20       --       26
1941    20        42       21       
--     34        20*       17        1       24
1942    37        46       25       28     13
*       20*       25        1       31
1943    21
*       36*      17*      23*    13*       20*       15*      10       28
1944    21
*       36*      17*      23*    13*       20*       15*      10*      19*
1945    21*       36*      17*      23*    13*       20*       15*      10*      19*
1946    29        49       12       34     10        31        18       13        9
1947    20        44       26       25      1        14        13       13       25
1948    26        39       15       20                          8       25       24
1949    29        40       22       23                          0       19       19
1950    16        19       35       19                                  29       12
1951    14        34       23       21                                  22
1952    23        21
*      21        5                                  25
1953    17        21
*      18        9                                  19
1954     5        29        6        3                                  16
1955    12        23        6                                            6
1956     7        25        1
1957     7        38 
1958     6        25 
1959     1         9
1960              21
=======================================================================================
old    323       555      231      187    169       267       101      199      242
new    386       705      282      256    221       367       146      219      280 


Obviously I'm doing this very indelicately, there's a couple places where I could have used the information a little more subtly (for instance with Greenberg's and William's partial war seasons), and there's lots of different ways of figuring out the per annum WS rate other than averaging the four surrounding seasons. But that gives an idea at least.

Which brings up another question for me. How much nuance could you apply? and with what other information?
-Following up on the first point, you could see whether there's "momentum" to having a certain number of consecutive 300 PA seasons and determine the likelihood that a player gets through an injury nexus with a certain number of healthy seasons in a row.
-Recognizing that some positions are more dangerous to your health than others, you could look at injury rates for positions and apply them to people at the appropriate position.
-Remembering that health is in some measure a skill, you could look at players in terms of the remainder of their career to see how likely they are to fall into the zero/one/two/three interruptions groups and then recalibrate from there (as in, did they have any interruptions after? If not/so , how likely is it they'll have a first or subsequent interruption in one of those seasons? and then how likely would it be to follow an injury versus merely ineffectiveness?
-Recognizing that some players may be more or less vulnerable to injury by dint of their age or years of experience, you could look at injuries by age, years of experience and apply findings that way.

What's anyone else think?
   159. Dr. Chaleeko Posted: September 17, 2006 at 03:52 PM (#2179405)
Oops!
If you averaged the four seasons surrounding his military service, you'd get 31.25 WS per year. I think the incidence of career-ending injury is probably an overall wash with the incidence of career-ending shrapnel wound in the services, so take that 20% figure and apply it. That's 25 WS a year.

Shoulda noted that the 20% figure didn't include career-enders and was the conservative figure I threw out in my previous post.

Double Whoops!

My dashed list should have been in this order!

Which brings up another question for me. How much nuance could you apply? and with what other information?
-Remembering that health is in some measure a skill, you could look at players in terms of the remainder of their career to see how likely they are to fall into the zero/one/two/three interruptions groups and then recalibrate from there (as in, did they have any interruptions after? If not/so , how likely is it they'll have a first or subsequent interruption in one of those seasons? and then how likely would it be to follow an injury versus merely ineffectiveness?
-Following up on the first point, you could see whether there's "momentum" to having a certain number of consecutive 300 PA seasons and determine the likelihood that a player gets through an injury nexus with a certain number of healthy seasons in a row.
-Recognizing that some positions are more dangerous to your health than others, you could look at injury rates for positions and apply them to people at the appropriate position.
-Recognizing that some players may be more or less vulnerable to injury by dint of their age or years of experience, you could look at injuries by age, years of experience and apply findings that way.
   160. Paul Wendt Posted: January 22, 2007 at 03:31 PM (#2284407)
.
Joe Dimino, 1993 Ballot #125, a comment appearing in other ballots too
If a guy was a 25 WS a year player before and after the war, a zero is a much bigger mistake than giving him three 25s.
[I agree.]
As far as injury you just credit a guy based on his playing time before and after the war. There's no reason to assume he would have been any more (or less) injury prone during those years.


But injury-prone doesn't imply an injury every year. In other words, for example, how can we say that Charlie Keller was such-and-so injury-prone? or disability-prone? One might ask whether minor injuries caused him to miss about 40 games 1940-43, but I mean to raise the issue as if he played every day during those four seasons. Does the back trouble, decisive in 1947, show that he was prone in 1944-45 or not?
(I would give Keller credit for 1938 and he would be on my ballot. He might be #5 on my ballot with full credit for 1938-47, but I wouldn't grant that.)

I wonder whether Phil Rizzuto would be in, had he enjoyed his 1947 season in 1946 and vice versa. Would it then be obvious that he missed half of his six-year peak as a league-average batter and dynamite overall player --his first six seasons, 1941-46?
   161. John (You Can Call Me Grandma) Murphy Posted: January 22, 2007 at 03:38 PM (#2284418)
I wonder whether Phil Rizzuto would be in, had he enjoyed his 1947 season in 1946 and vice versa. Would it then be obvious that he missed half of his six-year peak as a league-average batter and dynamite overall player --his first six seasons, 1941-46?

That, in a nutshell, is the quandry we face as we dole out credit, Paul. That's why I tend to lean toward a more conservative approach.
   162. Paul Wendt Posted: January 22, 2007 at 03:47 PM (#2284428)
.
Chris Cobb #111 and KJOK #114 describe different "15% rules" of thumb --15% below established achievement level and 15% below actual achievement against wartime competition.
   163. Paul Wendt Posted: July 06, 2007 at 04:12 PM (#2431024)
I composed this elsewhere. Apparently this is a version of the same point I have made before and once again in reply to a Joe Dimino ballot. SO I may be overdoing it.

--
Joe Dimino in "2001 Ballot"
I give full war credit, and I think it's a major mistake not to when comparing players across eras. My biggest regret on this project is that we didn't require all voters to give war credit like we did with Negro League credit. I see no difference, both were a circumstance of the player's birthday that was beyond his control. I also follow similar philosophy on strikes. I think it's a cop out to say we don't know so it's a zero. If a guy was a 25 WS a year player before and after the war, a zero is a much bigger mistake than giving him three 25s.
[I agree.]
As far as injury you just credit a guy based on his playing time before and after the war. There's no reason to assume he would have been any more (or less) injury prone during those years.


Because most players miss time sporadically, rather than at a consistent annual rate, this approach magnifies the swings in playing time. Someone who missed four seasons to the war and played every game in four seasons (or two or six depending on how many pre-war and post-war season you incorporate) becomes Billy Williams. Someone who missed 60 games to injury just before the war loses 10% off each of four wartime estimates.

Some regression to the mean should be imposed. For number of games, unlike quality, the player's own career mean --or moreso his "4-season before and after" mean-- is too variable. That method creates too many Billy Williamses and Frank Chances.

Of course this problem is severe if you use only one season before and one season after, on the argument that proximity in time makes the best estimate; including years -2 and +2 will only bias the estimate of quality. By playing time and quality combined, probably Enos Slaughter is thereby overrated more than anyone else. Taking two seasons before and two after incorporates one season that is 40 games short, which "fortunately" brings his estimated playing time almost in line with his 8-year postwar average. That is, straight line average of two seasons before and two after is a fortunate choice for the analyst in the case of Slaughter's playing time.

Of course, two seasons before and two after is a much better basis for straight line estimate than is one and one. The net effect is certain to be small where 1/3 of one season is an estimate (1981 or 1994) in contrast to 3 seasons (1943-1945). How to estimate wartime seasons --not whether to give credit at all-- is certain to have greatest effect on the rankings by peak/prime voters of players who missed peak/prime time.

Slaughter, Henrich, Rizzuto, and Pesky are some players in the relevant range whose war credit is probably both high and sensitive in the way discussed here. Regarding playing time, whether Joe Dimino's method is too generous to Henrich and Rizzuto ("makes Billy Williams" of them, to exaggerate the point) depends on whether Henrich departed the 1942 season early and Rizzuto returned to the 1946 late, or whether they "fortunately" missed games during those seasons.
   164. Paul Wendt Posted: July 10, 2007 at 05:09 AM (#2435417)
DanR in "2001 Ballot"
6. Phil Rizzuto. I love my shortstops. With appropriate war credit, Rizzuto has a strong prime to go with the huge MVP year and fistful of rings. All-world glove and a premium bat for the position.

premium bat? I infer that you have a forgiving interpretation of Rizzuto's 1946 return season, and he turns out to be a "premium bat" based on a 1941-1947 prime. Many others stepped back up to the plate and didn't miss a beat. Is there evidence for a forgiving interpretation in this case? Eg, some injury? Or he had a combat job --or a mere desk job!-- where others played G.I. baseball?

If no evidence, that's OK. I have a forgiving interpretation of Eppa Rixey's return.
But you might convince someone else to consider Rizzuto a 100 OPS+ batter in his prime --as "no one" else considers him, I guess.
   165. David Concepcion de la Desviacion Estandar (Dan R) Posted: July 10, 2007 at 01:22 PM (#2435496)
He wasn't much of a hitter in '48 either. I was actually referring more to his premium bat for the position during the Yanx' string of 5 straight titles from 1949-1953, not his performance in the mid-'40s. That said, if you average 41-42-46-47 you get some preeettttty nice SS offense for those war years, even if he was dragged down by the '46.
   166. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 13, 2008 at 11:32 PM (#2901942)
I know it's awfully late in the project to revisit this issue, but I've done some research on war credit that I think some in the group might find helpful. The discussion of Williams vs. Musial made me think long and hard about the war credit I was giving Teddy Ballgame--the five years he missing were, in my system, worth comfortably more than a Hall of Merit career in and of themselves. Yes, Williams's '41-'42-'46-'47 were jawdroppingly great seasons, and there's no reason to expect he would have done any worse in '43-'44-'45. But could I really say with confidence that he would have been ringing up 1.150 OPS's during those years? What if he had gotten hurt? This is the same uncertainty that some voters have expressed by saying that "giving peak war credit"--as in Johnny Pesky's case--makes them uncomfortable.

Well, I just realized that there's an empirical way to go about this. I just did a study of all players who had seven consecutive seasons with at least half the season played from 1940-1960, excluding the 1943-45 war years. There were 757 such rolling 7-year periods, involving 228 different players. I then performed a multiple regression of the seasons x-2, x-1, x+3, and x+4 against the average of the seasons x, x+1, and x+2 in the following categories: playing time (measured by SFrac), batting rate (measured by BWAA2/SFrac), baserunning rate (measured by BRWAA2/SFrac), and fielding rate (measured by FWAA2/SFrac). I got highly significant findings in all four categories. The equations are (as applied to a player who missed 1943, 44, and 45):

Playing time: The player's average SFrac for the years 1943, 44, and 45 is equal to .155 times his 1941, plus .198 times his 1942, plus .247 times his 1946, plus .112 times his 1947, minus .002 times his age in 1943, plus .33.

Batting rate: The player's average BWAA2/SFrac for the years 1943, 44, and 45 is equal to .21 times his 1941, plus .264 times his 1942, plus .243 times his 1946, plus .225 times his 1947, plus .42.

Baserunning rate: The player's average BRWAA2/SFrac for the years 1943, 44, and 45 is equal to .152 times his 1941, plus .204 times his 1942, plus .2 times his 1946, plus .156 times his 1947.

Fielding rate: The player's average FWAA2/SFrac for the years 1943, 44, and 45 is equal to .122 times his 1941, plus .181 times his 1942, plus .21 times his 1946, plus .124 times his 1947, plus .16.

This is all completely intuitive, of course. In each case, we weight the '42 and '46 the highest, the '41 and '47 less, and we add regression to the mean. The group's collective instincts on how to approach war credit are borne out by the data. But this allows us to answer the "What if?" question much more rigorously and precisely than we have been doing. Here's how it applies to The Kid in WWII (leaving the doubts about his 1942 and 46 fielding statistics aside for now):

Year SFrac BWAA BRWA FWAA Replc WARP
1941  0.91 10.8 
-0.1 -0.3  -0.7 11.1
1942  1.04 10.5  0.0  1.3  
-0.9 12.8
1943  1.00 10.0  0.0  0.6  
-0.8 11.4
1944  1.00 10.0  0.0  0.6  
-0.8 11.4
1945  1.00 10.0  0.0  0.6  
-0.8 11.4
1946  1.03 10.4  0.0  1.2  
-0.8 12.5
1947  1.06  9.4  0.0  0.0  
-0.9 10.4 



And Joltin' Joe:

Year SFrac BWAA BRWA FWAA Replc WARP
1941  0.93  7.4  0.2  0.8  
-0.9  9.3
1942  1.05  4.8  0.2  0.1  
-1.0  6.1
1943  0.94  5.1  0.2  0.3  
-0.9  6.4
1944  0.94  5.1  0.2  0.3  
-0.9  6.4
1945  0.94  5.1  0.2  0.3  
-1.0  6.5
1946  0.87  3.6  0.2  0.4  
-1.0  5.1
1947  0.92  4.5  0.3 
-0.8  -1.0  4.9 


And the outfield-mate and brother of the aforementioned:

Year SFrac BWAA BRWA FWAA Replc WARP
1941  1.02  1.9  0.2  0.3  
-1.0  3.4
1942  1.08  2.8  0.2  2.1  
-1.0  6.1
1943  0.98  2.4  0.1  1.1  
-1.0  4.6
1944  0.98  2.4  0.1  1.1  
-1.0  4.6
1945  0.98  2.4  0.1  1.1  
-1.1  4.7
1946  0.94  2.5  0.2  1.6  
-1.0  5.3
1947  0.91  1.2  0.1  1.4  
-0.9  3.7 


Clearly, these equations are not going to cause anyone to radically revise their ballots or PHoM. Compared to a straight '41-'42-'46-'47 average, Williams loses 0.3 wins a season, while neither DiMaggio moves a whit. Nonetheless, at the very least, this should provide reassurance that our standard approach to granting war credit has a strong statistical basis. As I apply these equations to more cases, if I find some where they produce results very different from a simple four-year average I will post them. (Just eyeballing it, it seems that you'd see the biggest discrepancies on guys whose value is heavily in D and baserunning, and whose '41 and '47 were quite different from their '42 and '46. These factors cancel each other out in the case of Dom D).
   167. Paul Wendt Posted: August 14, 2008 at 03:06 AM (#2902433)
Among HOM members, not pitchers, Rizzuto and Gordon put up notably poor seasons when they returned. Of course that tends to undermine one of your premises, iiuc, which is to take '41-42 and '46-47 for granted. It seems natural to see direct effects of the military service in their '46 achievements (poor quality) as well as their '43-44-45 (nil).

For a moderate case they may fit all the suggested criteria, consider Enos Slaughter.
   168. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 14, 2008 at 05:26 AM (#2902573)
Rizzuto played through malaria which he had contracted in the Pacific in '46. He's a special case.
   169. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 16, 2008 at 04:11 PM (#2905407)
My recent work on war credit has also made me wonder whether I might be able to develop a more empirical way to deflate the stats compiled during the war by players who didn't fight than the one I'm using now. I've taken a very straightforward approach: for each of the year-to-year transitions (1942-43, 43-44, 44-45, and 45-46), I looked at every player who had at least 50 PA in both seasons. I took his BWAA2/SFrac and added 8.7 as a quick-n-dirty estimate of batting wins created (since a guy in a 4.5 R/G league playing every game with 0 runs created costs his team 8.7 wins) and called it Batting Wins Rate (BWRt). I then weighted each player by the harmonic mean of his SFrac in the two seasons in question, added up the products, and took the ratio. The results:

1942-43: The group's weighted BWRt for 1943 was 1.027 times its weighted BWRt for 1942.
1943-44: 1.056
1944-45: 1.023
1945-46: .888

If we go forwards in time from 1942, that means that a BWRt of 1 in 1942 is equivalent to a BWRt of 1.027 in 1943, 1.085 in 1944, and 1.109 in 1945. If we go backwards from 1946, it means that a BWRt of 1 in 1946 is equal to 1.126 in 1945, 1.101 in 1944, 1.042 in 1943, and 1.015 in 1942. Averaging these values, I get adjustments of 1.035 for 1943, 1.093 for 1944, and 1.118 for 1945. These then need to be inverted to get multipliers of .966 for 1943, .915 for 1944, and .895 for 1945.

Repeating the same process for BRWAA and FWAA, here are the adjustments for all three categories:

Year BWRt BRWRt  FWRt
1943 .966  .996  .984
1944 .915  .998  .973
1945 .895  .998  .957 


Clearly, baserunning was not affected by the war enough to matter. Fielding, on the other hand, most definitely was, showing a very similar pattern to hitting.

I have added a sheet to the file posted in the Yahoo group with war-adjusted stats for every player-season in question.
   170. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 16, 2008 at 10:37 PM (#2905582)
Just a further thought on war credit--I've been dishing it out evenly across all the missing seasons. But, just like Negro League MLE's, doesn't this unfairly flatten guys' peaks? If we think a guy would have had 18 WARP from 1943-45, wouldn't it be fairer to give him 8-6-4 rather than 6-6-6? I suppose you could even try to extrapolate the proper degree of year-to-year variance by looking at how volatile the player's performance was for the rest of his career. What do all of you think?
   171. Eric J can SABER all he wants to Posted: August 16, 2008 at 11:37 PM (#2905679)
I don't know if it would make much difference, but you could try running a separate regression for each year between '43 and '45 (if you haven't already). Apart from that, it seems like flattening things out is just going to be a feature of war credit; since it's inherently an estimate of what would have happened, it'll be an average of several possible outcomes.

If you want to look at volatility, I suppose you could go further and look at variability in performance in general, or across different age time frames (24-26, 31-33, whatever years the player misses).
   172. Brent Posted: August 17, 2008 at 02:07 AM (#2906036)
An old comment by EricC on the Rizzuto thread provides an example of how variability could be associated with war credit.
   173. Joey Numbaz (Scruff) Posted: August 19, 2008 at 09:18 PM (#2908768)
Great stuff guys.

Dan would it be a lot of trouble to ask you to work up all of the cases of anyone who is potentially in the HoM discussion (say anyone who received a vote) if I could get you a list?

Is this an automated type thing, which it's just a question of formatting the answers (which I could do, given a spreadsheet) or does each player have to be 'worked' up?

Would be nice to get this 'on the record' somewhere . . .
   174. DL from MN Posted: August 19, 2008 at 09:41 PM (#2908801)
Do players get war credited to the playing conditions during the war or before and immediately proceeding the war?
   175. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 19, 2008 at 09:45 PM (#2908806)
Well, the adjusted stats for guys who did play in '43, '44 and '45 are now available in the file posted to the Yahoo group. Yes, I could post them for anyone relevant to the HoM--in fact, I have been doing so for a number of guys as you can see on their threads--but you can do it just as well as I can. The equations are:

For guys who missed all three years:

SFrac: .155*41 + .198*42 + .247*46 + .112*47 - .002*Age in 43 + .33
BWAA/Yr: .21*41 + .264*42 + .243*46 + .225*47 + .4
BRWA/Yr: .152*41 + .204*42 + .2*46 + .156*47
FWAA/Yr: .122*41 + .181*42 + .21*46 + .124*47 + .16

For guys missing two of the three years ('44 and '45 in virtually all cases):

SFrac: .115*41 + .067*42 + .255*43 + .24*46 + .127*47 - .004*Age in 43 + .286
BWAA/Yr: .128*41 + .158*42 + .257*43 + .21*46 + .199*47 + .31
BRWAA/Yr: .109*41 + .173*42 + .154*43 + .159*46 + .148*47
FWAA/Yr: .057*41 + .124*42 + .197*43 + .239*46 + .127*47 + .09

For guys missing just one year (1945 in this example):

SFrac: .124*41 + .29*44 + .358*46 + .075*47 - .006*Age in 43 + .31
BWAA/Yr: .083*41 + .09*42 + .159*43 + .223*44 + .172*46 + .216*47 + .2
BRWAA/Yr: .186*42 + .124*43 + .159*44 + .193*46 + .107*47
FWAA/Yr: .079*42 + .134*43 + .238*44 + .205*46 + .153*47 + .026

Remember that the equations for BWAA, BRWAA, and FWAA are to be applied to *rates*, not to totals. You take the guy's BWAA/BRWAA/FWAA *divided by his SFrac* in each season, plug those into the formula, and then multiply the result by the new SFrac calculated with the above equation. Make sure to use the war-adjusted numbers if you are using stats from '43, '44, or '45. And yes, I know it's not intuitive that '42 and '43 fell out of the one-year SFrac equation, but that's just what the regression told me.
   176. Joey Numbaz (Scruff) Posted: August 20, 2008 at 07:01 PM (#2910373)
Excellent, thanks for the info Dan! Hopefully I can get something going in the next couple of weeks, as I'm going to be out of town the next two weekends, but will try to find the time as I'd really like to see this and add the numbers to the DB I made of your WARP.
   177. David Concepcion de la Desviacion Estandar (Dan R) Posted: September 02, 2008 at 04:22 PM (#2925921)
I figured I'd take my own crack at one of my favorite candidates for '09, Johnny Pesky. I can't seem to find a thread for him, so I'll put this on the war credit page. It's tough to calculate war credit for him because he spent 1941 in the minors. Following Dr. Chaleeko's lead, I've calculated an MLE for his 1941 season and used that to determine war credit.

Using a version of my BaseRuns estimator customized for the 1941 American Association, I find Pesky with 86 Runs Created in 405 AB-H, or .212 runs per out. Knowing nothing about Louisville's stadium, I'll assume it's neutral (Dr. Chaleeko figures it was a slight pitchers' park). Borrowing Dr. Chaleeko's .90 conversion factor for this league (is .90 right for double-A?), that would be reduced to .191 runs per out. The 1941 AA averaged .185 runs per out, while the 1941 AL averaged .187, so .191*.187/.185 = .193 runs per out in the 1941 AL. Pesky's OBP that year was .372, and OBP varies at the .45 power of run scoring, so .372*(.193/.212)^.45 = a .356 MLB OBP. Pesky played 146 of his team's 154 games; an average player in the 1941 AL had 667 PA, so 667*146/154 = 632 PA for Pesky, which at a .356 OBP is 407 outs. At .193 runs per out, that's 78 major league RC for him. An average team in the 1941 AL had 3954 AB-H, so with Pesky consuming 407, that leaves 3547 for his teammates, who at .187 runs per out will generate an additional 663 runs, maing a total of 742 for the team. An average team in the 1941 AL scored 738 runs, giving a Pythagorean winning percentage of .502, which straight-line adjusting to 162 games is 81.4 wins, or 0.4 batting wins above average. We need to multiply this by the 1941 AL standard deviation adjustment of .95, but then also to divide it by the fraction of the season Pesky played (also .95) to get an annual rate, and those two factors cancel out, making Pesky an 0.4 BWAA2-per-year player in 1941.

To get a handle on his fielding for that year, I looked at every shortstop who played at least half a season in the majors in the liveball era in each of his age-21, 22, and 26 seasons. I then regressed their FWAA2 per year rate in their age-21 season against the average of their rates in their age-22 and age-26 seasons, and found that the rate in the age-21 season was equal to .74 times the average of the rates in the age-22 and age-26 seasons minus 0.3. Pesky averaged 1.7 FWAA2 per year in 1942 and 46, which according to this equation makes him .96 FWAA per year in 1941. Since he played 95% of the season in 1941, that's 0.9 FWAA2.

He had 0.2 BRWAA2 in 1942 and 0.1 BRWAA2 in 1946, so I'll give him 0.1 in 1941.

Plugging all this data into my war credit regression equation, I get Pesky with 2.4 BWAA, 0.1 BRWAA, and 0.9 FWAA for each of his missing seasons. Adding that in to the rest of his MLB performance, he looks like this:

Year SFrac BWAA BRWAA FWAA Replc WARP
1942  1.03  2.5   0.2  1.8  
-3.2  7.7
1943  1.02  2.4   0.1  0.9  
-3.1  6.4
1944  1.02  2.4   0.1  0.9  
-3.1  6.4
1945  1.02  2.4   0.1  0.9  
-3.1  6.4
1946  1.06  3.3   0.1  1.8  
-3.1  8.2
1947  1.09  2.2   0.1 
-0.5  -3.2  5.0
1948  1.01  0.9  
-0.1  0.7  -1.4  2.9
1949  1.07  1.5   0.1  1.7  
-1.5  4.8
1950  0.90  1.3   0.1  0.7  
-1.1  3.2
1951  0.86  1.6   0.1 
-0.1  -2.6  4.1
1952  0.46  0.0  
-0.2 -0.9  -1.4  0.3
1953  0.52  0.4  
-0.2 -0.6  -1.2  0.8
1954  0.29 
-0.6   0.0 -0.2  -0.6 -0.1
TOTL 11.35 20.2   0.5  7.1 
-28.5 56.2
TXBR 11.06 20.8   0.5  7.3 
-27.9 56.3
AVRG  1.00  1.8   0.0  0.6  
-2.5  5.0 


3-year peak: 22.3
7-year prime: 45.0
Career: 56,3
Salary: $166,659,796, comfortably over the in/out line. Among shortstops, this is similar to Campaneris; overall, it's near Magee and Stargell.
   178. Joey Numbaz (Scruff) Posted: December 13, 2008 at 08:49 PM (#3028457)
Dan, I just tried running the numbers on Bobby Adams, who I assume would have started his big league career in 1943, had he not been in the military. He played until he was 37, was a regular for a few years and made AAA by the time he was 20 in 1942.

I used the 3-year equation, with zeros for 1941 and 1942. Do you think this will work?

He ends up with a .46 SFrac per season, after .51 and .38 in 1946 and 1947. His WARP1/YR and WARP2/YR both end up being higher for 1943-45 (4.16) than his 1946-47 (3.3, 3.2), I assume because of the regression to the mean that is built into your equations?

Just want to make sure I'm doing things correctly for these type of players. Thanks for any help.
   179. Joey Numbaz (Scruff) Posted: December 13, 2008 at 09:33 PM (#3028472)
Also all of your formulas above, even the ones for 1944-45 and just 1945 say to use age in 1943 for SFrac, just want to make sure that's not an oversight. Thanks!
   180. Joey Numbaz (Scruff) Posted: December 13, 2008 at 09:46 PM (#3028475)
OK, figured one part out, I wasn't multiplying the rates by the new SFrac. Now I get Adams as having rates of 3.07 WARP2/YR from age 21-23 (estimated) and actual rates of 3.3 and 3.2 WARP2/YR at age 24-45 (majors 1946-47).

I think I need to add in some lower weights for 1941-42 estimated, but not really sure how to go about guessing what to use.
   181. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 13, 2008 at 09:50 PM (#3028477)
You DEFINITELY don't want to just use zeros for '41 and '42! What you would need to do is translate his minor league performance in '41 and '42 to its major league equivalent, and use those numbers in the equations.

And yes, it is always the age in 1943, even for the formulas for 1944-45. That's just a product of how I did the calculations.
   182. Joey Numbaz (Scruff) Posted: December 13, 2008 at 09:54 PM (#3028478)
I used his worst number from 1946-47 for each specific skill at the weights suggest above for 1941-42, and this drops his WARP2/YR rates for 1943-45 to 2.49, 2.59, 2.49, his projected age 21-23 seasons; as compared with the 3.3 and 3.2 for his age 24-25 actual seasons.

Seems reasonable, but obviously not perfect.
   183. Joey Numbaz (Scruff) Posted: December 13, 2008 at 09:55 PM (#3028479)
I don't have time to come up with MLE's for everyone from 1941-42 . . . don't even have access to the stats if I wanted to try . . . is something like what I suggested in #182 reasonable? Any other ideas?
   184. Joey Numbaz (Scruff) Posted: December 13, 2008 at 10:09 PM (#3028485)
Also what if a guy is missing just 1944 - is it the same formula as your 1945 above, with all of the years moved back by one?

So for SFrac, would I use 1940, 43, 45 and 46 instead of 1941, 44, 46 and 47?
   185. Joey Numbaz (Scruff) Posted: December 13, 2008 at 10:15 PM (#3028487)
And for someone like Luke Appling, who came back late in 1945, I'm going to figure the projection for 1944-1945 SFrac (.94), subtract out his SFrac that he actually played (.11). Then I'll base the projection on the .83, and back in his actual 1945 and make this his final 1945. I assume this is fine, but figured I'd check just to make sure.
   186. Paul Wendt Posted: December 14, 2008 at 04:04 PM (#3028638)
I suppose you mean that SFrac-projected for Appling 1944 and 1945 is 0.94 and 0.94. You will his projected record in playing time 0.94 and and 0.83, plus his actual record in playing time 0.00 and 0.11. That seems reasonable to me. Alternatives would make use of the date that he returned in 1945 (late, you say, perhaps based on the small share 0.11).
   187. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 15, 2008 at 12:43 AM (#3028884)
Yes, if a guy is missing just 1944, you would use the same formula as the above 1945, with all of the years moved back by one.

The way I do partial war seasons is:

1. I add the war-adjusted statistics for 1945 to his 1946, and calculate new yearly rates for 1945-46 combined, using them in the role of the 1946 rates for the equations.
2. I calculate war credit for 1945 using the equations.
3. I prorate the 1945 war credit by the fraction of the season he DIDN'T play: e.g., if my projected war credit SFrac is .7, and the guy actually had an SFrac in 1945 of .3, then I would reduce his war credit by 30% (bringing his war credit SFrac to 0.49). Then I add on the real war-adjusted numbers to that (bringing the total SFrac to 0.79).
   188. Joey Numbaz (Scruff) Posted: December 15, 2008 at 03:41 AM (#3028945)
Dan, I'm running the numbers for Phil Rizzuto, and I'm getting much different numbers than what you posted on his thread.

You posted 1.7 FWAA per year for him, and I'm getting 1.1. I've triple checked and I believe I have the correct formula in the spreadsheet. Is it possible there's some kind of error in post #175?

I checked, and I have the same numbers you used for 1941-42, 1946-47 going into the calc, so I'm sure that's not it.

I also came up with .89 SFrac, not .87, so I'm wondering if something is going on there as well.

I get .4 BWAA, not .3. We both get .2 BRWAA.

You are getting -2.6 for replacement, I get -2.53 for 1943-44 and -2.47 for 1945. I assume you still apply the league adjustment, right? Either way, changing it to 1 isn't enough to explain the differences.
   189. Joey Numbaz (Scruff) Posted: December 15, 2008 at 04:03 AM (#3028950)
Except for replacement level, I've got Pesky matched to your numbers above, using the exact formulas I did for Rizzuto.

For replacement level, I'm taking the number from your spreadsheet: -2.987230322 for 1943 SS, -2.987141066 for 1944, -2.951629646 for 1945. Giving the full decimal as you have it to make sure we're on the same page.

I take that, and multiply it by SFrac to get Rep1.

Then I take Rep1 and multiply it by the league adjustment to get Rep2.

For Pesky, I'm getting his Rep2 as -2.9, -2.9, -2.8 for 1943-45, but above you have -3.1 for all three years.
   190. Joey Numbaz (Scruff) Posted: December 15, 2008 at 06:21 AM (#3028986)
Found this on your 2009 Ballot Dan:

"7. Phil Rizzuto
With proper war credit--and I've increased it upon finding out that his poor 1946 was due to a malaria infection--"


I assume that increased credit is why we differ so much? If so, great. I actually gave him more credit than you then, but that's OK.

Still wondering why I'm off of the replacement part of your equation though.
   191. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 15, 2008 at 07:22 PM (#3029476)
OK, Rizzuto walkthrough, using his actual 1946 and pretending he never had malaria:

SFrac = .155*41 + .198*42 + .247*46 + .112*47 - .002*Age in 43 + .33. Rizzuto's SFrac were .81 in '41, .93 in '42, .79 in '46, and .94 in '47, and he was 25 in 1943. Thus, each war credit year has SFrac of .89. No idea why I apparently had .87...

BWAA/Yr = .21*41 + .264*42 + .243*46 + .225*47 + .4. Rizzuto's BWAA2 per 1.00 SFrac were .12 in 1941, .86 in 1942, -1.65 in 1946, and .85 in 1947, so his war credit BWAA2 per 1.00 SFrac is .44. Multiplying that by a SFrac of .89, we get 0.4 BWAA2 per war credit year.
BRWAA/Yr = .152*41 + .204*42 + .2*46 + .156*47. Rizzuto's BRWAA2 per 1.00 SFrac were .25 in 1941, .43 in 1942, .13 in 1946, and .11 in 1947, making him .167 BRWAA2 per 1.00 SFrac in war credit. Multiplying that by a SFrac of .89, we get .149 BRWAA2 per war credit year.
FWAA/Yr: .122*41 + .181*42 + .21*46 + .124*47 + .16. Rizzuto's FWAA2 per 1.00 SFrac were 2.35 in 1941, 2.47 in 1942, 0.38 in 1946, and 2.34 in 1947, making him 1.26 FWAA2 per war credit year. Multiplying that by a SFrac of .89, we get 1.1 FWAA2 per war credit year.
Rep2: The StDevs and Rep Levels sheet has SS rep level at 3.0 wins below average per 1.00 SFrac for each of '43, '44, and '45, so multiplying that by .89 gives 2.67 Rep2 per war credit year.

Adding these up, Rizzuto gets three years at 0.4 BWAA2 + .149 BRWAA2 + 1.26 FWAA2 + 2.67 Rep2 = 4.5 WARP2, or 13.4 WARP2 total. If my prior post said otherwise, then I made a mistake in the calculations. This total should, in my view, be subjectively increased given that Rizzuto was playing sick due to a war-related illness in '46, but how much you want to bump it is up to you.
   192. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 15, 2008 at 07:24 PM (#3029483)
I just checked--the chart I posted in Rizzuto's thread was done via guesswork/averaging rather than this more rigorous regression-based approach, which I just developed this year. That said, I had Rizzuto at 4.7 WARP2 a year there, and I put him at 4.5 here using the "scientific" method before giving him malaria credit, so clearly my instincts weren't that far off!
   193. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 15, 2008 at 07:26 PM (#3029485)
Pesky's Rep2 are just the 3.0 per 1.00 SFrac for SS listed on the StDevs and Rep Levels.xls sheet, multiplied by a 1.02 SFrac.
   194. Joey Numbaz (Scruff) Posted: December 15, 2008 at 08:16 PM (#3029574)
Thanks Dan,

But don't you also have to multiply that by the league adjust factor? If not, why is Rep1 and Rep2 slightly different for many players?

I will go back and check on my Rizzuto numbers once I get home tonight . . . and yes, I was referring to what was posted on his thread.
   195. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 16, 2008 at 04:19 PM (#3030481)
Rep1 is equal to Rep2 divided by the league adjustment factor. The replacement level figures in the StDevs and Rep Levels.xls spreadsheet are Rep2/already-standard-deviation-adjusted numbers. Just take the replacement level number right off StDevs and Rep Levels.xls, multiply it by your war credit SFrac, and that's your Rep2.
   196. Joey Numbaz (Scruff) Posted: December 16, 2008 at 08:25 PM (#3030903)
Yeah, I was doing that, and we're still getting different numbers.

Have the replacement levels on StDevs and Rep Levels.xls changed, could I have an old version? Do the numbers I quoted from the table in post #189 match yours?
   197. David Concepcion de la Desviacion Estandar (Dan R) Posted: December 16, 2008 at 10:22 PM (#3031060)
Yes, they match mine. But as I said, you do NOT multiply the replacement levels in the StDevs and Rep Levels.xls file by the league adjustment, because they are ALREADY adjusted. (To calculate Rep1 for a given year, you would *divide* the listed replacement level by the LgAdj). You simply take the listed figure as the standard deviation-adjusted replacement level per 1.00 SFrac, and multiply that by the player's SFrac to get his Rep2. So in Pesky's case, his MLE SFrac in 1941 is .95, his actual SFrac in 1942/46/47 were 1.03, 1.06, and 1.09, and he was 23 in 1943, making his war credit SFrac .155*.95 + .198*1.03 + .247*1.06 + .112*1.09 -.002*23 + .33, or 1.02. The StDevs and Rep Levels.xls file lists the standard deviation-adjusted SS replacement level as -3.0 wins above average per 1.00 SFrac for each of 1943, 1944, and 1945. Thus, Pesky's Rep2 are -3.0*1.02 = -3.06 per war credit year, which rounds to the -3.1 listed above. Is that clear?
   198. Joey Numbaz (Scruff) Posted: December 16, 2008 at 10:46 PM (#3031096)
Ah, OK, I missed that step . . . thanks!
   199. Paul Wendt Posted: May 16, 2009 at 04:19 AM (#3181267)
Pitchers

Here is a selection from the controversy about last week's election
(Election Results: Top 5 HoM Moundsmen For Group 3 - Grove ...)
where the big deal is the strong showing by Hal Newhouser.

Do most of us overlook or underestimate the proper quality discount for 1944-45?

[begin edited quotation]

#29. Paul Wendt
...
Further, I believe there is a majority [here at the Hall of Merit], still far short of consensus, on the revisionist side regarding WWII. I guess that 15% is a typical discount here, implicit for many participants.

> JWPF13
> He WAS that good in 1946, but if he was as good in 1944/45 as he was in 1946, he would not have gone 54-18 with ERA+s of 161 and 195, he would have cleared 200 with daylight behind him.
>
> IF you look at Hal as someone who should get credit for 140-150 ERA+ season in 1944/45,

[Newhouser 1944/45] At 15% ... ERA+ 137 and 165, two year average 151.
That is very close to what JWPF13 suggests, in criticizing the group moderately.


#60. Paul Wendt Posted: May 12, 2009 at 04:01 PM (#3176255)
["Jolly Old St Nick" asks]
> how you would compute the strength of the American League today if an average of 18 players each (none of whom were 4-F's, teenagers or geezers) were removed from the rosters?
<

After the season was in the books, meaning posted at baseball-reference player pages, we would probably compare good old OPS+ and ERA+, 2008 and 2009 (say), for the players who would remain. In practice most of us would cherry-pick a handful of players :-) but some would take everyone with 502 PA or 162 IP in both seasons (I might), or everyone with 300 PA or 100 IP in both seasons (Brent?). Dick Cramer might take every player weighted by his minimum playing time across the two seasons.


#66. JPWF13
...
49 pitchers had 90+ ip in both 1945 and 1946
those 49 had an average* ERA+ of 115 in 1945 and 108 in 1946 (quite frankly I expected a grater divergence)
I need to look at 1941/42 versus 1943/44/45, etc.
There is also a selection bias- the guys with 90+ ip in 1945 who were bad, or teams knew were bad irrespective of 1945 performance likely did not get 90 ip in 1946...

* "average" being all 49 ERA+s added up and divided by 49.

#67. JPWF13 Posted: May 12, 2009 at 06:07 PM (#3176420)
63 pitchers pitched 120+ ip in 1944/45 and also in 1946/47

Those pitchers lost an average of 14 points

1944/45: Average ERA+ =115
1946/47: Average ERA+ =101

Next I'll look at 1941/42 compared to 1944/45....
   200. Paul Wendt Posted: May 16, 2009 at 04:25 AM (#3181282)
Good show, JPWF13.

You posted these calculations so promptly after my suggestion. If that was a response rather than a coincidence, I wonder whether you have a season-level database that includes ERA+. If you have both IP and ERA+ at the season level then I (among others here) may have some good suggestions for weighting the average. --that is, for weighting the pitchers who qualify both during wartime and during normal time.

Just for instance, Dan Rosenheck #169, using Batting Wins Rate,
"weighted each player by the harmonic mean of his SFrac in the two seasons in question."
Page 2 of 3 pages  < 1 2 3 > 

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
rr
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

Syndicate

Page rendered in 1.0998 seconds
43 querie(s) executed