Anybody want to bet me that we won't have two more Negro Leaguers inducted into the HoM two weeks from now? :-)

Probably no takers on that one John. The main question seems to be whether Leonard and Brown jump to the top or if Suttles and Beckwith go in now. Looks like all four will be in by 1958.

On my ballot I think Leonard and Brown will be at the top this year.

I'm certainly not betting against that. It's just not clear which two. It's not impossible for Negro Leaguers to finish 1-2-3-4.

Most HOMers, minimum 10 G

1926 - 38
1927 - 37
1925 - 35
1924 - 33
1928 - 34
1929 - 32
1930 - 31
1931 - 31
1932 - 31
1916 - 31
1891 - 30

I don't know the answer to this, but does there come a point where we start to wonder if we're overdoing it on a particular generation?

I've got to say that I'm worried about Howie's numbers. I'm not going to take it out on Suttles or Beckwith, but Earl Averill might want to look over his shoulder.

I'm not going to take it out on Suttles or Beckwith, but Earl Averill might want to look over his shoulder.

Why is Averill behind Medwick anyway? Comparable hitters, about the same length careers (if you give Averill minor league credit), plus Averill was the better defender at a tougher position.

Is Muscles' 1937 season weighted that much that it pushes him ahead of Rock?

Well, part of the reason for the "spike" in the mid-20s is just generational crossover - there's a bunch of guys who started in 22-26, and another bunch who hung on until 28-29, perhaps aided by the heightened offensive levels of the '20s. And there's a higher proportion of Negro Leaguers as we move into the 1920s, and they tend to have longer careers.

It's not something I would ignore, but I think these kinds of issues are more important when you're dealing with the marginal candidates. I don't really think this is an issue with this particular election, but other people might differ.

I had Medwick slightly ahead, but I agree with John. I'm starting to think of Medwick, Averill, and Bob Johnson as Glut II: Clogged Up Again.

I'm starting to think of Medwick, Averill, and Bob Johnson as Glut II: Clogged Up Again.

I remember seeing the first movie during the aughts. Same damn plot. :-)

Looks like we're making the same mistake as the HOF -- too many from the 20s/30s and not enough from the 90s (only more of ours are NEL). Beckley, Griffith, Childs, van Haltren, where are you when we need you?

Devin is exactly eight on the 'hang-arounds.'

In 1890, with the 3rd league/expansion and about the 20th anniversary of 'organized ball,' baseball's earliest studs were able to hang on a little longer.
In the 1920s, you had the Cobb-Speaker-Johnson-ECollins legends finishing up around the same time.

But I suspect the 1890s and 1940s may suffer similar fates: not enough adjustment for the one-league/viciously played era on one hand and on the career-jumbling WW II on the other.

Why is Averill behind Medwick anyway? Comparable hitters, about the same length careers (if you give Averill minor league credit), plus Averill was the better defender at a tougher position.

From the perspective of someone with Medwick at #4 and Averill at #47:

1. Medwick is better at gray ink, which I adjust to add OBP and Power/Speed Number (deleting AB and Games) and reducing hits from 3 points to 1 (and moving OBP up to 3).

2. Medwick kills him on the Hall of Fame monitor, adjusted to normalize their stats to a specific run environment (same for all candidates). Medwick is an "A" measured against HoFers. Averill is a "D". Black ink factors in there, and Medwick has a huge advantage there too.

3. Medwick beats him on Hall of Fame standards, adjusted to normalize their stats to a specific run environment. That's primarily because Medwick played 300+ more games. Minor league credit might make up for this category, but I'm not as quick to give minor league credit as some. Too susceptible to manipulation.

4. All of Medwick's WS measures (3 year peak, 5-year consec peak, 7-year peak, WS/162 and career) are higher, and no amount of outfield defense is going to make up for that. Averill's defensive advantage is about 1.5 WS per year. That defensive advantage is already built into the WS measure, so it doesn't matter which position he played.

5. Pretty even in WARP1 adjusted for season length, but Medwick's got him on the career number. Again, Averill's defensive advantage is already built into the WARP measure and it doesn't propel him ahead.

I don't weight all of those things equally (more weight to WS and WARP than the Politics of Glory measures) -- and I consider a few other things --but I weigh everything against where all HoFers stack up in the same categories.

In the long run Averill meets 59% of my HoM standards and Medwick meets 78%. That's just my take. I'm sure it would be different if I tried to convert his minor league stuff to the majors.

Dixie Walker is the #30 RF in TNBJHBA: Galan and Heath are #42 and #44 in LF.

Walker 278 CWS 33-28-27/133/23.64
Galan 263 32-32-30/130/24.45
Heath 217 28-24-24/103/25.41

Walker 121 OPS+ in ~7500 PAs played thru WWII
Galan 122 in ~ 6950 PAs played thru WWII
Heath 140 (!) in ~ 5600 PAs played thru WWII

Walker at #30 is bracketed by Gavy Cravath (#29) and Ken Griffey Sr. (#40) with King Kelly at #31. He hit .401 in the South Atlantic League in 1930 and .350 with 105 RBI at Newark in 1932 and started with the Yankees. The Yankees gave up on him because, James says, Joe McCarthy didn't like southerners. So Walker, ironically, was a victim of discrimination himself. He spent 1935 in the minors. He didn't quite catch on with the White Sox or Tigers, finally catching on in Brooklyn at age 30. "He put himself on the wrong side of the race question, and this probably cost him a chance to manage in the major leagues."

Galan, James says, is "very underrated," in part because he had leadoff type hitting skills which are underappreciated.

Roger Peckinpaugh, who managed Jeff Heath when he broke in, said he "could have been one of the greatest players. But he just had no hustle, no nothing." Then in 1947 Willard Brown got a few PAs with the Browns, but the Browns, James says, didn't really want him to succeed. They told him not to bring any bats, that they would supply him with bats. Well, guess what. No bats, and no teammate would even loan him one. Brown eventually picked up the heaviest bat he could find, which was Heath's, except the knob had broken off, and Heath had discarded it. "He (Brown) hit a home run--the first home run by a black player in the American League, and the only home run of Brown's 'amjor league career.' When he got back to the dugout, Jeff Heath reclaimed the bat, and shattered it against the dugout wall."

Jeff Heath, wherever you are, you need to know that we here at HoM know that you had no hustle, no nothing and you were a racist bastard. 140 OPS+? I personally don't care if it was 240.

This may just be an aritfact of the last couple classes, but the arrival of 4 guys who jumped right to the top of the backlog makes me wonder if we're actually ever going to elect anyone other than Suttles/Beckwith from the pre-1950 backlog.

Well, like I said, it explains the "spike" - the really high numbers from the mid-20s. But there is definitely a greater number of HoMers for the era starting in about 1923/24. However, if you look at these players, there are a couple of questionable candidates (Terry, Faber), but not that many. When we're done, just about every era will have a couple of iffy inductees, I would think. And there are always going to be some vagaries in the course of history - there's a very definite dip in the period immediately before this (1918-22).

Finally, while we might be overrepresenting this period somewhat, we're a long way from the HoF. Checking my copy of Politics of Glory, they had 55 active HoFers in 1928-30 without counting the Negro Leaguers, although that count includes everybody who came to bat even once that year.

Devin, my count of HOMers so far counting even token appearances for 1928-30 is 34/33/34

As for Medwick, Bob Johnson and Earl Averill:

Career OPS+

Averill 132 in ~7160 PA
Johnson 139 in ~7995
Medwick 133 in ~8072

Best Seasonal OPS+ over 140

Averill 159-49-47-43
Johnson 157-56-58-55-42
Medwick 179-57-49-40

Career WS

Averill 280
Johnson 287
Medwick 312

Top 3 Career WS

Averill 33-30-30
Johnson 31-29-26
Medwick 40-36-33

I agree that Averill looks good in comparison to Medwick, but I'm not sold on Johnson even by the numbers but especially knowing that one of his career years was against weak WWII competition (I've adjusted the OPS+ numbers for his peak accordingly but no adjustment to the WS, just didn't get around to it.

But this OF glut is murder.

As for Medwick, Bob Johnson and Earl Averill:

Speaking of Glut II, remember the good old days of Glut I? Those guys, as far as a know, are still hanging around, ballot eligible, and playing in an UNDERrepresented era. Before I vote for ANY of Medwick, Johnson, or Averill (and in 1954, I voted for none of them), perhaps we can see how they stack up against Jimmy Ryan, George van Haltren, Tommy Leach, Hugh Duffy, and Mike Tiernan?

That makes eight REALLY REALLY GOOD outfielders that I am not voting for. I'm open to convincing that one is clearly better than the others, but right now I just don't see how to vote for one and leave the other seven off.

And if anyone wants to take me up on my offer to be convinced, please throw Jimmy Sheckard into the mix, too. He was not on my ballot when he was elected, essentially for the same reasons that the other eight aren't. I'm not saying that he's a "mistake," just that it's not at all clear to me that he's the cream of this particular polygluttic crop.

As for the outfield glut I had Medwick at #5, Duffy at #11 (I believe), Averill at #15, and GVH just off ballot. With as strong a pool of candidates we ahve there isn't much seperating these four. I count Leach as a 3B and he is in the mid 30's, Ryan and Tiernan are both out of my top fifty. I keep going over Ryan's case and I can never justify putting him over someone like Browning or Cravath to bring him up to the glut.

I do believe that Medwick is the best of the new glut and that Duffy was the best of the old one baseed on their peaks, but then again I am a peak voter. I am not a big fan of Bob Johnson as his peak isnt' as good as Medwick or even Averill (and was effected by WWII) and he doesn't have the strong prime of Averill. His career totals aren't awe inspiring, or at least aren't neough to overcome peak conerns the way that say, CP Bell's are. Overall I just don't see him as a special player. Yeah he coudl hit and yeah he was consistently very good but so are alot of other guys in MLB history. I also give no bonus points for consistency.

As for an overabundance of 20's and 30's players in the HOM, what this reallly looks like is an underrepresentation of players from the period 1892-1919. The period before 1892 I believe is very well represented, especialy for the number of teams and players that were playing during this era. I do believe that the 1890's are udnerrepresented and I sincerely hope that we get to elected a few of these guys in the future (Preferrably Jennings, Childs, and Duffy, the guys on my ballot :-)). Some of this dead lock may be an extreme of career guys (GVH, though I like his priem as well, and Beckley) to peak guys (Jennings, Childs, McGraw) that is splitting the ballot.

However, I am coming ot the conclusion that the period between 1903-1919 simply had fewer great players. A good portion of this is that the NeL was in its nascent stages so a lot fo the great black players from this time are either unknown or have VERY scant statistical records. I am comfortable not electing these players if we only have very sketchy images fo them.

But, as for MLB players from this period, I only have three in my top 30 in Vic Willis, Roger Bresnahan, and Gavvy Cravath and those three are all between 20 and 30. I also see no reason to really significanltly raise these players. Could it be possible that this period was just had a few less great players in it?

I also have a question about Hughie Jennings. I see a lot of voters mentioning Hughie time at 1B in their explanations as to why they aren't voting for him. I broached this with jimd (I believe) a while back and he said that it was mostly just put there as something to say, which is fine. If this is the case with the rest of you, fine. But...

Even for a guy like me, one of the three who put Jennings at #2 in the last election, what Hughie did after 1898 bears very little on his HOM case, or at least on the positive side. His case is based 95% on 1894-1898, with the rest of his career serving only to get him above 200 WS. Some people think that is not enough and dont' vote for him, fine. I disagree with this and think that Hughie is a special case of player with a peak so high as to drown out career length worries, but I understand it somewhat.

My question is then, why does what position he played from 1898 to the end of his career matter in anyway?

First of all, if you use WS or WARP, position shouldnt' really matter. A 41 WS seasons at SS is just as valubale as a 41 WS seasons at 1B, same with seasons of 30, 20, or 8 WS. Second of all, Hughie played SS for the period where he was a great player and didnt' switch to 1B until after the shoulder injury. Isn't taking points away from him because he played 1B sort of like double penalizing him for his bad seasons?

When it comes to19th century MIers, we seem to have elected the guys with really long careers in McPhee, Dahlen, and Glasscock while ignoring the guys with the nice peaks like Jennings, Childs, and even Dunlap. With the 1890's such an underrepresented era, shouldn't we be looking to balance this out a little. If we elect Jennings and one of the OF glut that would go a long way toward balancing out this era.

Also,

Many people liken Jake Beckley to an 1890's version of Rafael Palmeiro since both had long careers without ever being dominant. While agree that they both had long careesr and that neither was particularly dominant I dont' believe this is a very apt comparison. Below are the top 7 WS seasons of both players in descending order and their career totals.

Palmeiro 31,31,30,26,25,24,24 372
Beckley 24,23,23,22,21,20,20 364

Now before you all jump on me, a few clarifications. First of all, Palmeiro's career totals are only through 2003, he is missing 2004 and whatever he has this year. You could probably
add 20 WS to his career total.

Second, yes Beckley's numbers aren't adjusted to the 162 game schedule. BUT they are adjsuted to a 154 game schedule, I am just a little lazy right now. How much could this add to Beckley's totals? Maybe 20 to his career totals and one, MAYBE 2 to some of his season totals. This still gives Palmeiro a clear advantage in peak/prime with equivalent career value.

Beckley major problem in my eyes is that he never was a top 10 player and rarely ever a top 15 player in a league with only 250-300 players in it. Maybe Palmeiro was never a top ten player, though I doubt that since he had three 30 WS seasons.

But remember that Palmeiro played in a time with 700-750 players in it, making it about three times harder to be on of the top 10-15 players in baseball through WS. Even if you dont' believe that (and since WS is based on actual wins and there were about three times as many wins when the seasons are schedule adjusted, you should believe that) you must admit that there are more good players (in total numbers, not exactly in percentages) today than in the 1890's and 1900's due to the expanded pool of players.

And for those of you who say that a pennant is a pennant, during Palmeiro's career pennant weren't won during the regurlar season. We are looking at regular season stats and the goal of the regular season in Palmeiro's time was to make the playoffs. This allows for mroe players, htough roughly the same percentage wise, to have HOM type value. So instead of 2 teams playing for teh title out of 10 in Beckley's day, you have 4/26, 4/28, 8/28, and 8/30. With Wild Cads currently dominaing the playoffs, one could argue that wins in the regular season today are MORE valuable than wins in the 1890's.

So Palmeiro and Beckley have a equal career value (though viagra man is still adding on) and Palmeiro is clearly ahead on peak/prime. I guess you could make the argument that Beckley had more decent seasons (in the 10-16 WS rang or so) but anyone who says that is inherently, though maybe not conciously, punishing peak players.

one more modern comparison...

Mark Grace 27,25,25,23,22,21,20 294

A little less career with a very similar peak/prime. Beckley seems to have been Mark Grace with about four average seaons tacked onto his career. He was better but not wildly so.

Re: The Glut-tony of outfielders...

For me, as a WS voter with a peak-leaning stance, Medwick's real competition is not Averill, nor even Johnson, but rather Cravath (with MiL cred) and Burns.

Anyone who loves Medwick should be very seriously considering both of them. Here's their annual WS, ranked best to worst, prorated to 162 games, and, in Medwick's case, with a 10% deduction to his 1943-1945 seasons. Cravath is listed with MiL cred for 1907, 1909-1911 per his thread.

        JM  GB  GC
       --- --- ---
        42  37  37
        38  36  36
        35  33  32
        25  30  30
        25  26  29
        25  25  27
        25  25  27
        23  23  22
        22  23  21
        20  2119
        18  2016
         9   6  14
         5   6  12
         5   2   2
         4   0
         3 
         0 

       === === ===
career 324 314 325
  3 yr 114 105 105
  5 yr 165 161 165
 10 yr 280 279 281
 15 yr 321 314 325

The only point of differentiation here is that Medwick's two strongest seasons are worth a combined ten WS more than George and Gavy's. AFter that they each make up that 10 WS difference in a big hurry. Those 10 WS are why Medwick's ahead of Cactus and not-Tioga on my ballot, but he's not very far ahead.

Averill, by contrast, is off the end of my ballot. Not by much, but off, while Johnson is well off the end.

Matt, I would have to guess that most voters already have decided where Glut II stands against Glut I. As for me, following are the OF and 1B on my Hot 100 list of eligibles that goes through 1959.

1. Joe DiMaggio
4. Johnny Mize
6. Buck Leonard
8. Joe Medwick
10. George Sisler

13. Mule Suttles
20. Edd Roush

22. John Beckwith--I make a special case listing Beckwith among the "hitters"
24. Chuck Klein
28. Gavy Cravath

31. Pete Browning
32. Willard Brown
38. Hugh Duffy
39. Cool Papa Bell

43. Hack Wilson
44. Earl Averill--so I'm with Jeff on Medwick vs. Averill though I agree there is a case to be made the other way
49. Charley Jones

Second 50: (Keeler), (Stovey), Wild Bill Wright, Cuyler, Veach, (Sheckard), Keller, (Hill), Leach, (Kelley), Berger, Beckley, Johnson, Poles, Van Haltren, Burns, Rice, Chance, Ryan, Estalella.

HM: Tiernan, Bottomley, Cy Seymour, Dave Orr.

Yes, this is tricky, when there is a case for #44 over #4. I have been working on the cornermen recently with more emphasis on plain old OPS+ than before, and it scrambles things. Based on OPS+ (mainly) I am thinking about the following changes.

Top 10

Medwick could move down 133 career/179-57-49-40-38

Second 10

No problems

Third 10 (20s)

Beckwith really should be higher. I don't have his annual/peak/prime numbers but his 137 career looks better and better. Cravath looks better than Klein and they probably should swap positions, though they're very close.

Cravath 149/170-69-57-51-46
Klein 135/168-58-55-49-49

Thirties

Browning should move up based on OPS, obviously
Bell--shouldn't move based on OPS but 400+ WS is a lot

but if they move up, who moves down?

Forties

Wilson, Jones and Averill all look like they should rate higher, but again, who moves down if they do? Duffy, I guess.

Klein 135/168-58-55-49-49
Duffy 121/172-49
Wilson 145/177-60-59-55-50
Jones 150/182-75-66-63
Averill 132/159-49-47-43

Second 50

Stovey is the guy who really jumps up for me as he is not PHoM.

Stovey 141/182-65-62-60-56-54

Keller and Chance are the guys among HoM eligibles or pending eligibles who jump up on OPS+ peak.

In the end, I don't think my cornermen/"hitter" rankings are going to change much. But clearly we can all come to vastly different conclusions by our choice of key data points.

As for Burns, he is extremely underwhelming on OPS+.

Burns 115 in ~8,113 PA (AB + BB + HP)
Extended Peak 149-48-42-29-26-21-21

Certainly a valuable guy, but in a consideration set of 25 FO/1B his career best of 149 ranks 23rd, ahead of Cuyler 148 and S. Rice an eye-popping 125. His (Burns') fourth-best of 129 is also way on the low end of this list. Cuyler's was 133, Veach 135, Rice 121 among the other lower. Even Jimmy ryan is 170-49-44-44. Edd Roush is 162-53-49-47-42.

Heck, Billy Herman is 131-28-28-28-21 and Stan Hack 143-43-32-28-25-20.

I'd be curious to know if anyone thinks that Gene Benson and Eugene McDaniel might merit deeper inspection than their 00%s above suggest. We've got some data on McDaniel in Mexico that could be helpful, and Benson might be they type whose career was a bit truncated on the end by integration, lessening his lengend.

32. Willard Brown

Sunnyday,

How confident are you about this placement of Brown?

As I looked through Holway's yearly rankings, I'm seeing a guy who was among the NAL's leaders in everything but walks year in and year out, and who played shortstop and centerfield in addition to the corners.

So basically: tons of power, high average, speed, tons of EXBH, probably a CF or very good corner fielder, very few walks. So his ceiling is a Vlad Guerrero, but he feels like a Medwickian sort who generates lots of value from his AVG and SLG despite his lack of plate discipline.

Plus he'll qualify for War credit and possibly for Minor League credit for his time in the 1950s in the Texas League.

I'm already anticipating him landing in my top ten, but that's without any translations to work from.

On the Palmeiro/Beckley comparison, I agree Palmeiro is noticeable better than Beckley. But I would have Palmeiro easily at the top of this upcoming ballot (and I like Brown and Leonard enough to have them 1 and 2). So having Beckley 4th or 5th seems reasonable to me.

I think there are a good number of voters (I'd ballpark 25 to 35) who would have Palmeiro in the top 5 on this coming ballot. Being a level worse than Palmeiro is not much of a slight.

Jschmeagol, this is the same rubbish argument you have put forward several times already. First, WS gets 1B wrong in the 1890s. Second, Beckley has 40% more career than Grace, a HUGE difference.

Third, you need to adjust to the same season lengths (not difficult, you add 8/154, you could have done it if you'd wanted to do so and not just be intellectually dishonest) and if you do that Beckley has 384WS (and is then undervalued) versus Palmerio's 372.


Your argument is boring, trvial and wrong, and I am extremely irritated that simple souls among the other voters appear to be being polluted by it. We are VERY short of 1890s players and have far too many from the 20s and 30s; hopefully the electorate will come to its senses during the "elect 3 years" or at least after the present glut.

Doc,

I am not confident about Willard Brown at all. All I know is his reputation. I am waiting to see his actual record.

Keep in mind, he is #32 on my Hot 100 eligibles as of 1953 and continuing through 1959. That probably means that he is right around #13-18 when he becomes eligible. (Medwick is #8 on the Hot 100 and was #2 on my ballot. Buck Leonard and Ray Brown are #6 and 12 on the Hot 100 and will probably be #1-2 on my 1955 ballot.)

So if he Brown the player I think he is, he could be on my ballot. If he is better than what I know now, he could move up. And of course he could move down.

karl,

I'm not sure that we are "very" short of 1890s players. Couldn't find the list just now. But as I recall it, we might be a "little" short or we might not be short at all, in the eye of the beholder.

And the underlying point holds. Beckley's peak is surely not an all-timer. Pro-rate to 162 and add in some defense and it just as surely becomes VG.

But of course that's must me, and I'm no big Raffy fan either. I'd rather have 10 years of Donny Baseball than 20 of Raffy. Of course, I'd rather have 20 years of Eddie Murray than either one.

Beckley is still #78 in my current consideration set and he could come back someday. But I don't find rehashing his story now trivial or boring.

karlmagnus:

I sympathize with your position regarding Beckley. He was clearly a bigger force in his era than Grace ever was at first base. No argument will persuade me otherwise.

But if you wish to persuade other people of that view, a little less argumentation might help. Besides, I doubt Mark has some type of vendetta against Beckley that a reasonable argument couldn't persuade him. :-)

There was a very funny British series in the 1930s called "Letters to the Secretary of a Golf Club" in which the chief trouble-stirrer was General Sir Armstrong Forecursue, who was difficult to deal with because he exploded with rage every time he found something he disagreed with. Always seemed a highly sympathetic character to me :-))

Always seemed a highly sympathetic character to me :-))

LOL

Sounds a bit like Basil Fawlty to me, who I always thought was a pretty sympathetic character--in fact, probably the most sympathetic character on the show, other than all the other characters ;-)

'and if you do that Beckley has 384WS'

Prorating to 162 games season and upping his def WS by 15% get Jake up to 371 according to my calcs. Do you think that WS is so bad at 1890s 1B Def that upping DefWs by 15% doesn't correct it? I've said before and I'll say again that I think you seriously overcorrect for this 1B def issue with WS. Even if it is as serious a problem as you say(and you've never given me any data to back it up), you're still seriously overestimating how much it affects his overall WS total. This is coming from someone who had Beckley 8th on his 1954 ballot and believes he eventually should get in. I don't think you're helping get him elected by vastly overrating him. Things like the above statements and putting him ahead of Arky Vaughan on your ballot just turn people off. It turned me off and I'm pro Beckley.

"Very" is open to interpretation, but I've got the trough at 1897, with 19 semi-regulars.

Anson; Burkett; Clarke; Collins; Dahlen; Davis; Delahanty; Grant; Hamilton; HR Johnson; Kelley; Keeler; Lajoie; McPhee;Nichols; Rusie;Wagner; Wallace; and Young. Token appearances by Roger Connor

That's essentially half of the peak.

Part of the problem is that there are only three pitchers on the list (Rusie, Nichols, Young) -- an artifact of the 1893 rules change that led to the pre-1893 stars petering out by 1897 and the newcomers just revving up. Another part is the only two Negro Leaguers due to the fact that there wasn't a Negro League.

If you assume that structural factors cut out three pitcher and three or four black players, that's still a ten-plus player trough.

My Top Ten "eligible who played in 1897" include (roughly in order of quality) Jake Beckley, Clark Griffith, Cupid Childs, Hugh Duffy, Bill Monroe, George van Haltren, Jimmy Ryan, Mike Tiernan, and Herman Long.

We could elect the top half, and still not really be scraping the bottom of the barrel.

Karlmagnus,
Did you just call me a simple soul? That would be a first. I've been called lots of bad things in my life, but never simple.

karl,

In what way does Beckley have 40% more career than Palmeiro? It surely isnt' hits as Palmeiro is closing in on the 3000 hits that Beckley just missed out on.

Even if Beckley has 384 WS when schedule adjsuted, entirely possible, the actaul totals aren't 384 to 372 but 384 to 372 plus whatever Palmeiro had in 2004 plus whatever he has accumulated so far this year plus the fact that he is still adding to his total. I would bet a lot of money that Palmeiro will clear 400 WS with room to spare.

I also dont' think my argument was tired or anything of the like as I have yet to read anything like that here. And of course Beckley was better than Grace, I just wanted to throw that in there.

As for the defense, I would love to see some numbers on how much you boost 1890's 1B defense and at the expense of what position you would take that credit from. Pitching? 3B? 2B? OF?

Daryn,

I dont' see Palmeiro and Beckley as just a level apart, but that is a difference of opinion. I see a few levels between then based on Palmeiro's peak and prime, small levels that probably add up to about 40 to 50 spots on the next ballot.

I really coudlnt' justify putting Palmeiro behind Jennings, even with Jennings obvious peak advantage, but I am not sure that I like him more than Leonard or Brown. Likely #3 were he on this ballot. Then again, he may be as low as #6, behind Leonard, Brown, Jennings, Suttles, and Beckwith. Not that any of this matters.

Palmeiro v. Mattingly will be interesting (assuming Donnie Baseball is still around by then). Mattingly was my favorite player as a kid and the closest thing I have to s real sports hero as he stopped playing when I was just 14. In fact he may be part of the reason why I am biased toward peak players.

Back to the Outfield Glut II.

Joe Medwick's peak is overrated by win shares. Bob Johnson's peak and career is underrated by win shares. While the fact that Medwick was usually playing on good teams and Johnson on bad teams probably makes some contribution to win shares missing the mark on these two, Pythagorean overperformance/underperformance is the key issue.

Consider these two pairs of seasons, important to the peak assessment of both players.

1937 Medwick 156 g, 180 OPS+, 36.9 bws. Team -1 vs. Pyth.
1944 Johnson 142 g, 174 OPS+, 28.2 bws. Team -6 vs. Pyth

1936 Medwick 155 g, 156 OPS+, 32.4 bws. Team +10 vs. Pyth.
1939 Johnson 150 g, 155 OPS+, 27.0 bws. Team -3 vs. Pyth.

Medwick is ahead of Johnson slightly by games and by OPS+ in both seasons, but he is _way_ ahead in batting win shares, much farther than the differences in playing time and batting rate would justify. Johnson's 1944 does need a wartime discount, too, so I'm not trying to argue that Johnson's peak was better than Medwick's.

However, I do make two claims:

1) Medwick's peak superiority to his fellow outfielders should be reduced by voters relying on ws because his 1936 season is significantly overvalued in win shares. (His other peak seasons nearly zero out, but he gains 10-15% more ws than his playing time and rate stats suggest that he should).

2) Johnson's peak inferiority to his fellow outfielders should be reduced by voters relying on ws because his 1939 and 1944 seasons are being undervalued in win shares. His career value also should be increased by win-share voters to compensate for his unluckiness in his teams. Overall (as has been pointed out before) his teams underperformed Pythagorean expectations by 25 games over the course of his career, which costs Johnson, by my calculations, about 10 career win shares.

I'm working on a fuller study of the win-share evaluations of Johnson, Medwick, and Averill along these lines that I hope to post in a few days.

(His other peak seasons nearly zero out, but he gains 10-15% more ws than his playing time and rate stats suggest that he should for 1936).

Fewer than expected HoMers from the early 20th century? Possibly. I may chalk it up to a talent distribution quirk. We also had a Very high % of all-time greats during the deadball era; most people would put Collins-Lajoie-Wagner-Cobb-Speaker-Johnson-Young, all active in 1910, in the highest tier of the HoM. Just as I wouldn't denigrate these greats too far for simply showing up at the same time, maybe there were somewat fewer 'near-greats' then.

OF glut - analysis of VanHaltren-Duffy-Beckley-
(I know, Jake played 1B) Roush-Averill-Medwick-Johnson
By Win Shares above high replacement level
Career WS per year minus 13, multipled by length
of career: (WS-per-yr - 13)*(careerWS) / WS-per-yr

van haltren166
Roush156
Duffy156
Averill153
Medwick153
Johnson138
Beckley128
By WARP3, using career value of each season of (WARP3 above 2.5), where seasons of less than 2.5 WARP are discarded
Johnson61
Medwick53
Averill50
Duffy49
van haltren48
Roush44
Beckley42
Averaging these two methods and dividing WS by 3 gives a ranking of
van haltren55
Johnson54
Medwick52
Averill51
Duffy50
Roush48
Beckley42
--
1. Beckley should be higher based on stronger value of 1B then; I would put him just above Duffy.
2. WS has no timeline, WARP3 does. Given the lack of 1890s players, I'm comfy with averaging the two measures.
3. Small MLB credit for Averill would put him fighting for the top of the list.
4. None of the others I could have listed here (Ryan, H Wilson, etc) would show up as high, and of course this ignores NeLers and Cravath.

I see a lot of voters mentioning Hughie time at 1B in their explanations as to why they aren't voting for him. I broached this with jimd (I believe) a while back and he said that it was mostly just put there as something to say

Couldn't have been me. I've been a longtime FOHJ.
(Background crowd chants: Hughie! Hughie! Eeyah! ;-)

PS: is it EE-yah or ee-YAH?

Pursuant to Chris Cobb's point about Pythags viz WS, this brings Medwick closer to Burns and Cravath, close enough to very nearly be a wash.

It wont' bring Johnson high up enough to compete with them (IMO), but it now makes those three rather difficult to distinguish from one another at a glance.

This suggests that they are all HOVG because they are not particularly unique, or that they are all equally meritorious.

TomH,

Burns? Also, I can't imagine enough defensive credit would boost Beckley up over Duffy. But mainly: Go GVH!

43:

It was Jim Spencer, not jimd.

PS: is it EE-yah or ee-YAH?

Never mind that. Can someone tell me how to pronounce Augie Galan?

Tom H's numbers do a good job, I think, of pointing out differences in career value in the glut. I've been thinking about this is in somewhat more simple terms like this:

George Van Haltren: 15 years as an average or above-average starting player (10 years OPS+ above 120, 15 years above 100)

Jimmy Ryan: 15 years as an average or above-average starting player (8 years OPS+>120, 15 years>100)

Bob Johnson: 13 years at or above average starter (13 years OPS+>120)

Averill: 13 years at or above-average starter (2 MiL seasons, 8 ML seasons at >OPS+ 120, 3 seasons > 100)

Medwick, Roush, 11 years at or above average starter (Roush 9 above 120 OPS+, Medwick 8)

Joe Medwick's peak is overrated by win shares. Bob Johnson's peak and career is underrated by win shares. While the fact that Medwick was usually playing on good teams and Johnson on bad teams probably makes some contribution to win shares missing the mark on these two, Pythagorean overperformance/underperformance is the key issue.

Ah, love the Bob Johnson talk. Chris Cobb points out the Johnson was on generally bad teams and Joe Medwick on good teams and this slightly affects Win Shares. Let me throw up some info on Johnson, Averill, and Medwick

                     Games Played by  Team Record
Player  Time Period  Player in Period  in Period
Johnson 1933-45         1863           826-1156
Averill 1929-40         1660           992-846
Medwick 1933-47         1938           1281-1021 

Averill and Medwick have some split team seasons so theirs are approximate team records those years. Totals then might be slightly off. Anyways, Indian Bob played on bad, BAD teams whereas Averill and Medwick benefitted from good teams.

In general for Win Shares its stated good players get their fair shake on win shares regardless of the team record. But does it work at the extremes?

Bob Johnson had it extreme. Average record for 13 seasons computes to approx. 64-89. Then we compare him to a couple of guys who might get a slight uptick due to excellent teams.

The gap within Outfield Glut II isn't so much that Medwick gets on 30 ballots, Averill 22, and Johnson is on a respirator...

Good and bad teams don't generally affect WS. Performance relative to pythag affects WS. Chris has shown us that in a couple of cases Johnson was hurt by the Sox' performance below pythag, and in one case (one season) Medwick was helped by his team outperforming pythag. Not at all clear how conclusive this is, though it is obvious that WS and OPS+ for the two of them don't track.

So?

Well, the basic insight of WS is that if a team overperforms its pythag that there may or may not be a cause-and-effect other than luck, and yet the wins are there, they happened and value resides in wins. It doesn't make a lot of sense to me to use WS and then grouse about departures from pythag.

WS chooses to track wins and NOT PYTHAG. If you've got a big problem with that, don't use WS! Just use LWTWS (and re-set replacement value)!

More numbers...

From Retrosheet comes the following. The latest update of material on the website included loading Batter-Fielding Wins (BFW) for players. That stat is from the 2005 ESPN Baseball Encyclopedia and was added to each players stats.

Here are the 25 highest eligible players in 1955 based on career BFW. Of course you lack the NeL players, it doesn't account for league quality/WWII deductions, etc. Still fun to look at -

 Dave Bancroft    36.2
 Bob Johnson      35.7
 Joe Sewell       35.2
 Billy Herman     31.6
 Cupid Childs     30.4
 Stan Hack        29.4
 Dick Bartell     28.2
 Fred Dunlap      28.1
 Pete Browning    28.0
 Hughie Jennings  26.8
 George Sisler    26.5
 Harland Clift    26.3
 Joe Medwick      25.6
 Jack Fournier    24.4
 Jake Beckley     23.0
 Roy Thomas       22.7
 Frank Chance     22.6
 Roger Bresnahan  22.5
 Del Pratt        22.5
 Miller Huggins   22.3
 Ed Konetchy      21.2
 Chuck Klein      21.1
 Wally Berger     21.0
 Ernie Lombardi   20.6
 Wally Schang     20.4 

The metric might be slanted toward middle infielders or have fielding issues since Bancroft and Bartell are way off the electorate radar. And looks who is #2 :)

The metric might be slanted toward middle infielders or have fielding issues

In this I concur. ; )

If this list is representative, then it looks like their BFW is suffering from the dreaded TB-Throwing-Infielder-Subluxation. Only a registered Sabrmetric Statistical Chiropractor can properly manipulate the backbone of this system and release the pinched values within to restore it to health.

Good and bad teams don't generally affect WS.

For the most part, this is true for individual season totals. It is not true for career totals.

Good teams play easier schedules than bad teams. The bad team plays a tough opponent (the good team) while the good team plays an easy opponent (the bad team). This is fair for determining the pennant because they are playing each other but it does have an affect on the player statistics due to the diffence in competition quality. (This is compensated for in the Total Baseball/baseball-reference.com "park" factors.)

It is not compensated for in Win Shares. Good teams usually win a couple of bonus games per year relative to an average team, who win a couple more than the bad team. The career effect is that a player who accumulates 300 Win Shares on an average team can expect to accumulate 306 Win Shares for a good team (.600 career WPct) and only 294 Win Shares for a bad team (.400 career WPct), everything else being equal.

Note: when teams are playing balanced schedules, calculating the size of the effect for an individual team is easy. Complete the schedule. That is, for the 1927 Yankees, add 22 games against themselves, in which they go 11-11. Their "true" winning percentage is 121-55 .688 instead of 110-44 .714, so they really only deserve 318 team win shares, not 330.

Never mind that. Can someone tell me how to pronounce Augie Galan?

I've heard it pronounced guh-LON (I assume you know how to pronounce Augie :-)

Rather than throwing the FBFW numbers out with the bath water--and BTW, thanks Al for the numbers--I would want to think a bit about whether it is in fact WE who are underestimating the importance of defense.

In my little world, the best available "hitters" would be Leonard, Suttles, Sisler, Beckwith, Medwick, C. Jones, Browning, Klein and Cravath. Again, hitting only. (With BFW, we seem to have the opposite problem as some other measures, and that is a rather HIGH eval. of dead-ball 1B defense and a LOW eval. of CF defense, but that aside... And besides, if you use WS, this will balance it out.) Otherwise we have elected all the reasonable "hitter" candidates.

But we have not elected all the reasonable "glove" candidates, and this list serves to remind us of who some of them are. If defense really matters, then Bancroft in particular is being very much underestimated, and by me too. IOW I think his being at the top of the list of 2B and SS is not an unreasonable place for him--if defense matters.

PRELIMINARY 1955 BALLOT

1. Ray Brown: Best pitcher available. Picking up 30+ WAT when you play for a majorly dyanstic team is pretty impressive.

2. Buck Leonard: Best hitter available. He edges Suttles by virtue of a bigger prime, extended prime, and career.

3. Mule Suttles: Has Leonard on short and extended peak, but that's it, and it's reasonably close. Leonard's glove is also a difference-maker here.

4. John Beckwith

5. Hugh Duffy

5a. Martin Dihigo (on track for a 1956 pHOM induction)

6. Joe Medwick: He does deserve some of the credit for his teams overachievement, doesn't he?

7. Jose Mendez

8. Gavy Cravath: Right behind Medwick.

9. Geo. Van Haltren: I've always felt like there should be a NYC bridge called the Van Haltren.

10. Bucky Walters: In this week's episode of Everyone Loves Wes, guest star Bucky Walters says: "Holy Crap, no one loves me!"

11. Wes Ferrell

12. Geo. Burns: Speaking of bridges in NYC, I like putting Geo. because it reminds me to always take the T-Z or the Ham Fish when I'm driving from Philly to Portsmouth, or vise verse. Reminds me in that the GWB is often abbreviated as Geo Wash Br or Geo Washington Br, and just below the Geo is usually a whole mess of people foolishly thinking one of the decks will be traffic free. And anyway, shouldn't the GWB span the Delaware instead of the Hudson?

13. Spot Poles: In need of more reconsideration from this electorate by gum!

14. Stan Hack: Hacktastic!

14a. Ted Lyons: No discernable ETA for his pHOMing at this point. Which means Bill Terry and Red Faber might want to go out and get a few hundred blocks of briarwood and a sharp knife so they can start whittling away the time until they make it....

15. Bill Byrd: If he somehow switched with Hilton Smith, do you think he'd be called "Satchel Paige's Legs?"

HEY, NEW GUY!!!!
Gene Benson: Anyone thinking meritorious thoughts about him? He's got that Cool Papa kind of feel to me.

Booker McDaniels: He was pretty acey for a while there. Does he have any kind of case?

depends on how much defense matters, i guess....

Question: What exactly is "Batter Fielding Wins" measuring? Is it against a single baseline, or positional baselines? How is it weighted re batter batting wins?

Question: On what analysis is this measure based?

Question: Is it me, or, if this is supposed to be a measure of fielding value, are there some dubious rankings here? Pete Browning above Hughie Jennings? Ernie Lombardi on the list and Ray Schalk not? Jack Fournier ahead of anybody?

All the guys on this list are pretty good hitters, and none of the really weak-hitting, great glove players like Maranville, Tinker, Long, and Schalk are showing up? Or have only players currently getting votes been listed??

From retrosheet:

"On each player's page, we now included BFW and PW, which are
Batter-Fielder Wins and Pitcher Wins, the categories presented
in the 2005 ESPN Baseball Encyclopedia, edited by Pete Palmer
and Gary Gillette. We thank these gentlemen for generously
making this information available to us."

I can't find more info -- maybe you have to buy the ESPN encyclopedia to get a glossary for that stat?

Looks like a proprietary third party derived statistic has snuck its way into retrosheet's otherwise amazing collection of raw data. Personally, I don't think stuff like that belongs there.

IOW BFW is a combination of batting wins and fielding wins. It is the next generation of TPR.

Does BFW use FR? And are these the same TPR or whatever from the latest version of Total Baseball? If so, I am not a big fan of this measure because of the use of fielding runs. I had thought that they were fixed but Don MAttingly is still rated as a 0 in FR. He is an A- (I believe) WS grades and one of the top 10 all-time in FRAA. In other words I still think taht FR is a very flawed measure.

Yikes, what hath I wrought, lol?

Figures I have an Olympic-sized workload today, too. Well, maybe I'll get to this later....

Well, the basic insight of WS is that if a team overperforms its pythag that there may or may not be a cause-and-effect other than luck, and yet the wins are there, they happened and value resides in wins. It doesn't make a lot of sense to me to use WS and then grouse about departures from pythag.

_Team value_ resides in wins. Player-value resides in what the player does to help his team. That correlates strongly with team wins, but not entirely. Luck is a factor; teammates are a factor. Most of the time, we who use win shares assume that if the system finds that one player gets two more win shares a year more than another player over the course of their careers, the player with the higher win share total is significantly better. What I'm arguing is that this assumption isn't necessarily true in all circumstances, and when we face the challenge of making choices among very similar players, we ought to test our assumptions.

James himself says about win shares: "If one player in this system is credited with 20 Win Shars and another with 18, we can state with a fair degree of confidence that the one player contributed more to his team than the other . . . not that we are always right; there will always be anomalies and there will always be limitations to the data, but I would be confident that we had it right a high percentage of the time."

For the HoM, I think it's not "grousing" to say that a metric is generally sound, but that we want more than "a fair degree of confidence that one player contributed more to his team than the other." We want the highest degree of confidence we can achieve, which means doing more than simply accepting the conclusions about value presented to us by comprehensive metrics.

Chris, I don't disagree with your conclusion. Of course we need to look at other data.

I guess my point was that departures from pythag just don't strike me as the biggest issue with WS or in the HoM discussions generally. Rather it is exactly what the BFW and other data suggests--that WS systematically overvalues CF defense, and certainly places less importance on middle IF defense that (say) LWTS or whatever its latest iteration is. Does that make WS wrong and LWTS right, or vice versa?

Definitely maybe.

IOW, given the choice between an indisputably great hitter and an indisputably great fielder, the great fielder will quickly go invisible. Bring Joe Medwick back to earth due to overperformance vs. pythag, fine, and then replace him atop the ballot with Mule Suttles or Buck Leonard or Bob Johnson or Earl Averill? Are we sure this is the answer? What about Dave Bancroft?

Good and bad teams don't generally affect WS. Performance relative to pythag affects WS.

AND IN RESPONSE

For the most part, this is true for individual season totals. It is not true for career totals.

Also, should keep in mind that "good" teams generally have a positive performance relative to pythag, so in reality playing on "good" or "bad" teams DOES have some impact on Win Shares, and COULD make a difference where Win Share values are close.

Does BFW use FR? And are these the same TPR or whatever from the latest version of Total Baseball? If so, I am not a big fan of this measure because of the use of fielding runs. I had thought that they were fixed but Don MAttingly is still rated as a 0 in FR. He is an A- (I believe) WS grades and one of the top 10 all-time in FRAA. In other words I still think taht FR is a very flawed measure.

Just because one player seems to be 'mis-evaluated' doesn't mean a fielding measurement system is "very flawed". ALL fielding systems will have 'outliers' that differ from other systems.

Fielding Runs have been greatly improved, and should probably be looked at ALONG WITH Win Shares/1000 innings and the BP fielding metric where 100 = League Average.

I still see fielding evaluation as in the fairly early stages. Yes, improving greatly, but it never fails to shock me how many people count on a stat that incorporates fielding measures as if they are as reliable as hitting ones.

To me, it starts with the hitting. Position is a bonus, as logic dictates that a guy with 15 yrs at a position was competent. Reputation and stats (ideally, combined as in Jimmy Collins' case) then measure in with fielding.

It's not that I think fielding is irrelevant, just that some of those metrics are still in flux, which would force some voters to change their votes in a few years. It's not so hard to dock Doyle and Browning, credit McPhee and Collins, and so on.
Vaughan deserved his pts because he was a freat-hitting SS who did not appear to hurt his team at all defensively. I'm not convinced he helped them there, though.

"great-hitting"

The shortest measurement of time known in the universe is the 'nano-HOM', which is the time between when you hit the "submit your comment" button and you realize that you made a typo.

One thing to remember about the batting-fielding wins list is that the statistic treats all leagues as equal. Bancroft and Bartell (and many of those other guys) played in a very weak NL and deserve a substantial discount. Though I don't swear that WARP has it right, the translations from WARP1 to WARP3 are instructive. Bancroft, for example, leads Sewell substantially on WARP1 (something like 112 to 102; I'm too lazy to look it up), but trails him on WARP3 (something like 83 to 78).

Tentative prelim:

(1) Buck Leonard (new)--Pretty similar to Beckwith with the bat, and would normally trail him slightly based on position, but edges him when Beckwith's troubles are taken into account.

(2) John Beckwith (2nd)
(3) Mule Suttles (4th)

(4) Ray Brown (new)-- Certainly an HoMer. Only debate is whether he is a Hubbell or a Lyons. I lean slightly to the latter but am willing to be convinced I'm wrong.

(5) Hughie Jennings (5th)
(6) George Van Haltren (6th)

(7) Dobie Moore (10th)-- For different reasons, Averill, Duffy, and Medwick lose a little ground. Moore (and the next guy) benefit.
(8) Wes Farrell (11th)

(9) Earl Averill (7th)-- In a deep OF field, he is close enough to the line that you begin where to wonder where he'd rank if the excluded players had been let in.

(10) Red Ruffing (15th)-- I overreacted last week. He's got a lot of quality bulk.

(11) Hugh Duffy (8th)-- Like Medwick benefits a bit too much in WS from Pythagorean overachievement; unlike Medwick, those overachieving teams were historically great.

(12) Cupid Childs (12th)-- A smidge ahead of the Herman-Sewell-Hack gang, largely b/c/ of the difficulty of the conditions he played under.

(13) Eppa Rixey (14th)
(14) Billy Herman (15th)

(15) Joe Medwick (10th)-- Minor adjustment for SNT syndrome and b/c/ his big year seems s slight WS fluke. Still think he should be inducted eventually.

I wish I could find room for Sewell (up a few slots to 16) and Mackey. There are about 40 guys on this ballot who I think would make worthy HoMers, but I only feel strongly about my top 17.

PHOM: Leonard and Brown (though Lyons has an outside shot at nosing out Brown).

andrew, the question with Bancroft is not whether he is better than Sewell. I have never had him ahead of Sewell. The question is whether we are underrating both of them. Just as an e.g., I see that the highest rated 20C IF on your ballot is Herman at 14th. Of course you've got Jennings, Moore and Childsin there, and my ballot is very similar. Jennings and Moore up high, Herman in the double digits. But I am still not confident that I/we are not undervaluing defense.

Sunnyday--

Fair enough as far as it goes.

There are two tasks we need to perform in the coming years. First, we need to figure out how to rate 2B/SS/3B vs. the hitting positions. Second, we need to rate those guys against each other. Frankly, I think ordering Boudreau, Sewell, Stephens, Reese, Herman, Doerr, Gordon, Hack, Lundy, Dandridge, Bancroft, et al. is going to make sorting out the OF glut look like child's play.

As the person who first threw out the Batting-Fielding Wins material I'll chime in. I did so more as a break from the standard rehashing of backlog candidates. It was different and fun to see the Bancroft, Bartell, Clift and Huggins type of players with perceived value, maybe above the hitters we constantly mention.

BFW gives another opinion of how someone did in their own time and place. I never anticipated it would cure world hunger. :)

Prelim with a re-eval of "hitters."

1. Ray Brown (nerw, PHoM 1955)--best pitcher by a wide margin

2. Buck Leonard (new, PHoM 1955)--best hitter by a slight margin

3. Hughie Jennings (was 3)--my personal BFW favorite

4. Dobie Moore (4)--the black Jennings, except that Hughie met the pope twice and Dobie met the madam one time too many

5. Mule Suttles (7)--movse ahead of Medwick and Sisler

6. Joe Medwick (2)--still, a mighty peak

7. George Sisler (5)--still, a mighty prime

8. Tommy Bond (6)--a mighty peak/prime

9. John Beckwith (13)--jumps ahead of Herman and Roush

10. Charley Jones (last on ballot 1945)--returns due to hitter re-eval

11. Rube Waddell (9)
12. Jose Mendez (10)--fall victim to hitter re-eval

13. Gavy Cravath (last on ballt in 1952)--jumps Klein and Roush

14. Billy Herman (11)

15. Chuck Klein (14)

Drops out: Edd Roush (12)--on ballot every year since 1945, actually falling victim to Charley Jones who dropped off my ballot in 1945

And Ed Williamson (15)--in and out over the past few years

16-20. Bell, Doyle, Roush, Joss, Browning
21-25. Dean, Williamson, Ruffing, Sewell, Averill

The other really big beneficiary of my hitter re-eval is Harry Stovey who is HoM/not PHoM. He goes in the first rank of hitters somewhere with Medwick, Sisler, Beckwith and Jones, ahead of Cravath and Klein. He goes with Teddy Lyons and Willie Wells as the next backlog PHoMer, not sure what order, however. One or two by 1958, I would thin, though I still have Suttles and Beckwith queued up as well.

Point-counterpoint

On the other hand, it's not like we're hurting for SS or 2B. We've just inducted two from the 20s and/or 30s, Beckwith will make it in by 1958. Lloyd's in, so's the Rajah, so's Frisch, so's Gehringer, so's Cocky Collins. Appling (coming to a ballot near you) and Cronin played in the 1920s.

The knot of Doerr, Gordon, Stephens, Boudreau, Sewell, Bancroft, Long, Doyle, Rizzuto, (and Pesky??) et. al. is going to be trouble not just now, but in the future when it's joined by the likes of Ozzie Smith, Alan Trammell, Jim Fregosi, maybe Al Dark, Jim Gilliam, maybe Bonnie Serrell, maybe Bobby Avila, and certainly Nellie Fox and Lou Whitaker. Heck you could even toss Chuck Knoblauch on the heap just for giggles.

In my world (which is small and not usually very tidy), the line looks like it's going to be drawn in the vicinity of Rizzuto and Fox, and it won't be pretty because these guys are all close, and there won't be room for more than four or five of them. Much will depend on how much war credit certain guys get and how much wartime discount others get.

Maybe the reality is that their lack of distinction is a point against them. Middle infielders with OK peaks and OK career length are epidemic, making the Morgans, Griches, Sandbergs, and Larkins look all the better.

Now 3B, that's a whole different kettle of tea, where we've really only got Jud Wilson and Heiniemann Groh to speak of, with a smidge of HOMrun Baker and Frisch, and some of Beckwith's career when he's elected. Is there a Hack-ing Mass out there? Will Ray Dandridge look as good as his reputation says he should? Is Bob Elliot going to get creamed by wartime discounts? Will McFly ever get back to 1985?

Point-counterpoint-point.

I am not clear in my own mind, so how would my posts be clear? But my response to your post would be that, yes, we're talking about middle IF here, but we're also talking about defense, and some of our HoM middle IF are not there for their defense.

My real point, I think, is, What about the Bancrofts and Maranvilles whose value is clearly more defense than offense?

Way back in 1986, when LWTS was new, David Neft did an article in the BRJ rating SSs. He came up with a batting factor, a running factor and a fielding factor. The Batting Factors ranged from .484 (guess who) to .289 for Bobby Wine. The Running Factors only varied by .34 and so didn't affect the ratings much. The Fielding Factors varied from +.174 (Ozzie) to -.90 (Jack Barry) for a total of .246 or 20 percent more than the range in the Batting Factors. So in theory the fielding may have influenced the final outcome of his rankings more than the batting.

Now, for SSs, I wonder if this is wrong? Whereas I would guess that traditionally we tend to rank players mostly on hitting and then make modest adjustments for fielding, kind of like they do for the Gold Gloves.

So anyway, his rankings come out like this:

1. Honus Wagner
2. Dave Bancroft--#2 fielder behind Ozzie
3. Bobby Wallace
4. Ozzie Smith
5. Ray Chapman--these are obviously all rate stats (problem)
6. Joe Tinker
7. Rabbit Maranville
8. Donie Bush
9. Dick Bartell
10. Joe Cronin

11. Cal Ripken--still active at the time but these are rates
12. Luke Appling
13. Lou Boudreau
14. Robin Yount--ditto Ripken (so was Ozzie still active BTW)
15. Joe Sewell
16. Luis Aparicio
17. Kid Elberfeld
18. Art Fletcher
19. Garry Templeton
20 Arky Vaughan--highest rated SS with a negative fielding rate

21. Travis Jackson
22. Ernie Banks--negative fielding rate
23. Pee Wee Reese
24. Rick Burleson
25. Dave Concepcion


Having seen, as a Twins fans, both Zoilo Versalles and Greg Gagne, I am quite sure that Gagne was the more valuable SS. By analogy, I'm not sure that Dave Bancroft is not pretty much in the ballpark with Arky Vaughan. Vaughan is the #2 hitter and #84T fielder, while Bancroft is the #2 fielder and #22 hitter. (The Batting Factor is adjusted for era.)

I guess it comes down to the questions of:

1. How many opportunities at bat and in the field do SSs get (at different times in history)? If more FC than PAs, then that is important.

2. What is the differential between excellent and average or replacement, etc? My sense is the differential is greater at bat--e.g. BA range from .200-.225 to .350-.400, e.g., while FA only range from about .900 to .990 or so (over time, though the real question is what is the range at a given time). At a given time, I am quite sure the differential among hitting is greater than the differential among fielding.

In theory, the various uber-stats make these sorts of calculations...but do they? What is the right balance of bat and glove in evaluating SSs? 50-50? 60-40? Whereas I would guess that most evaluations take defense at 33 percent or less, but I could be wrong.

Here's how Win Shares's intrinsic fielding weights look:

41.5 fws for avg team with 243 WS (81*3)

pos  weight  fws on avg tm
c    19%        7.9
1b    6%        2.5
2b   16%        6.6
3b   12%        5.0
ss   18%        7.5
lf    7%        2.9
cf   14%        5.8
rf    7%        2.9
               41.1

There's some rounding issues, but that's the picture that James paints of his own system. This nets out to , approx, 1/3 of the 52% of WS he awards to the defense, and about 17% of an average team's total Win Shares.

To look at it by position, however, we could say that according to WS, SS is
3.0 times as valuable as 1B
2.6 times as valuable as LF and RF
1.5 times as valuable as 3B
1.3 times as valuable as CF
1.1 times as valuable as 2B
0.9 times as valuable as C.

TOTALL TANGENTIAL, BUT INTERESTING QUESTION (depeding on your point of view)

Let's say that you had to impersonate Tim Keefe on Saturday afternoon for about four hours. What kind of accent would he have?

He was born in Cambridge, Mass, and he seems to have had some education and was known as Sir Timothy, presumably as a nod to his air and stature.

Anyway, what's anyone think? Working class Boston accent? Erudite John Kerry from the 1970s accent? New York accent???

Doc,

Here's another answer. Bancroft's career WS total of 269 consists of 39 percent fielding and 61 percent offense--or 105 fWS and 164 bWS.

Ozzie gets 43 percent fielding (140 fWS) versus 57 percent hitting (185 bWS) for a total of 325.

As you all know, Bancroft and Ozzie both hit <100 OPS and were perhaps the two best fielding SS in history. Yet their ratio is 60 percent offense anf 40 percent defense.

As Howard Hughes said in the movie, Does that look clean to you?

Of course, there's always Mark Belanger who was a worse hitter and not as good of a fielder, and he does get 64 percent of his value with his glove.There are a bunch of guys in the '70s with ?50 percent of their value on defense. that rarely happened any other time, not even deadball. Maranville is 47 percent defense.

As (not) an aside, Sewell has 277 WS, Bancroft 269. Sewell has sunk but Bancroft never floated at all, not even as a SNT.

So comparing Bancroft to WS's version of the theoretically average SS, his WS totals would represent double the percentage of his value than we would expect.

So then there's two questions to ask.

1) IF WS captures the essence of the value of fielding correctly, then what difference does it make how a player generates value? Value is value.

2) Is WS caputring the essence of the value of fielding correctly? That is, in the examples that Sunnyday2 has cited, does it seem plausible that the players he cites do derive the amount of value from fielding that WS suggests they do?

Here's one other thing to look at since someone brought it up before. How many creditable plays does a SS make? In Bancroft's case...

4623 putouts
6561 assists
660 errors
1017 DPs
====
12861 plays

Bancroft accumulated 8244 plate appearances.

Although I know it's for post-1955, Jim Furtado's xruns might be useful here. It weights outs as worth -.09 runs, DPs as -.37. Assuming that errors are roughly worth something like half-way between a single and a double, xR would suggest .61.

So multiplying it all out...

-the outs Bancroft accounts for are worth -1007 runs to opposing offenses
-the DPs he participated in are worth -376 runs to opposing offenses
-the errors he made are worth 403 runs to opposing offenses

His defense might then be said to be worth -980 runs to opposing offenses.

Meanwhile on offense, he created 930 x-runs.

By this alone we could suggest that Bancroft should be 50/50 or 55/45 defense to hitting as a SS. But what I would really want to know was how routine the defensive plays were.

Is every ground ball where Bancroft fields cleanly and throws the runner out as important as every time he HITS that ground ball?

Here's a thought experiment:

Which has more potential for run scoring? Any given at-bat by a hitter or any play where an infielder has to handle a chance?

I would bet my stack of 1988 Topps Gary Carters that any given at-bat would, of course, be more important because it has more different ways to work out positively (more outs as they say in poker), whereas the ball fielded can only realistically result in three things: error, infield hit, out, none of which will typically score a run by itself, nor advance a baserunner more than a single base (obviously errors can, but I don't know enough about them to know how many bases runners and batter-runners usually get on them).

Anyway, I guess I'm circling around the point that I think Win Shares has it closer than B-FW because there's a certain level of specialized expectation associated with fielding events that isn't associated with batting events.

Or I could be wrong, it's been known to happen.

In Neft's LWTS ranking (post #75) the weighting is really based on the range from high to low of each item.

The batting factor was sort of like a BA or EQA (.289 to .464) and the fielding factor was a much lower number (from negative .090 to +.174). So at a glance you might infer that the batting factor was, oh, twice as important as the fielding factor (2.3).

But no, the ranges are about .180 and .260 respectively. So two players who are at the top of one range and the bottom of the other will score about .374 and .463 with the great fielder (and bad hitter) scoring more highly.

So I guess what would be interesting is the range of scores in offensive RC and defensive R saved. I mean it's one thing to add up Bancroft's runs and say he's worth 1910 runs, but that's not what we're doing. We're saying he's worth 269 WS.

His 104 dWS in 12,000+ plays equals about 1 WS for everyt 124 plays.

His bWS (165) in 8244 PAs equals about 1 WS for every 50 PAs.

In other words an average PA (of Bancroft's) is 2.5 times more valuable than Bancroft's average play in the field. It's only the fact that he made 50 percent more plays in the field than PAs that enable his defense to be 39 percent of his value.

But even then what defines a great player is not how many bases or outs he accounts for, it is how many bases or outs he accounts for that other players wouldn't earn.

I guess I should just go back and re-read WS The Book and remember how he came up with the ratios he did. But it would seem to be one thing to say that half of the game is two-thirds mental... no, wait, that the game is half offense, half defense. It is another thing to evaluate indi vidual players at different positions based on offense and defense., and it seems to me that nobody really knows how to properly value great defense at important defensive positions.

remember how he came up with the ratios he did.

IIRC, James never says how those ratios were derived. They are just presented as a given (though there is a comment about boosting the defensive value of catchers).

I believe they are derived from the concept that the absense of offense implies the defensive value. That is, managers will only grudgingly give up offense to get an equivalent value on defense. IOW, FWS + BWS should be about the same at each position when summed over all time.

IOW, FWS + BWS should be about the same at each position when summed over all time.

Jim D,

Because you've got facility with WARP....

Wasn't this (or isn't it) somewhat true of WARP as well, but in reverse? IIRC, BP uses positional batting weights to figure the replacement level for hitters at each position. So aren't they ultiamtely arriving at a somewhat similar point? (That is that there's total value at each position but that it's divvied up differently based on the defensive demands of the position as reflected by the hitting abilities of its members. If you know what i mean...)

IOW, FWS + BWS should be about the same at each position when summed over all time.

Wasn't this (or isn't it) somewhat true of WARP as well,

I'm sure it is. However, there's room for play in the following equation (C+F)+B where C is a constant. The larger the constant C, the higher the emphasis on fielding relative to batting.

WARP places a higher emphasis on fielding, at least and in particular during the 19th century and deadball eras when errors were a much larger part of the game and the variance amongst players at a position was much larger.

Well, I think most people here realize that "closers" are not put to optimal usage these days, to use one example.
You don't need a mathematics degree from Harvard to figure out, for instance, that it makes more sense to use closers with 'leads' of 0, 1, and 2 rather than 1, 2, and 3. Of that group, the 3-run leads are ones that anyone on your staff can handle. And the tie games are the ones that should most demand the best pitcher available (along with the one-run leads).

However, I do agree that hitters who play a generally weak-hitting position deserve a 'bonus' for outpointing their rivals by a great margin.

I just wouldn't want to overdo the cerebral aspect of 60-year-old men who wear uniforms. Sometimes they do dumb things.

Howie, a good philosophical and methodological discussion could be based on the idea that "hitters" who play a generally weak-hitting position deserve a "bonus" for outpointing their rivals by a great margin.

Or rather, what is "value"?

1. Outpointing rivals at your same position? (Presumably including both offensive and defensive values, but defined as the differential vs. those rivals at your same position.)

2. Or, simply earning bases on offense and contributing to outs on defense regardless of position? (The assumption here being that the generally weak-hitting are probably playing at positions where they have a fair opportunity on defense to make more plays and, thus, to make up the difference in offensive value [of a C or SS vs. a 1B or LF, e.g.].

So IOW is a C or a SS or a relief pitcher "great" because he is the best C or SS or RP in the league at a particular time? Or does that player have to be able to make a case for being better than the LFers and 1Bmen and starting pitchers in order to be "great"?

Probably the biggest impact of your choice of eval. model is on C who historically couldn't get enough innings to make enough defensive plays to overcome their weaker offense. (Unless you can find a way to quantify his contribution to successful pitching, which has of course been tried.) But if you use model 1, then you wouldn't have to give a C bonus, it seems to me. A C bonus implies model 2.

Players at other positions appear to have an opportunity (though maybe not an equal opportunity) to overcome offensive weaknesses. The Bancroft discussion was based on the perspective that, well, it is indeed not an equal opportunity. And if the defensive spectrum was a perfectly ideal model, then everybody (every position) would have an equal chance at 'greatness." And C clearly do not. Hard to know what to make of 3B.

Of course the 2005 NY Yankees are an exception to both of the models in post 86, in that they're playing everyone out of position, which would apparently be how they could spend 200 mills and not be a winning team....

In all seriousness, I think the truth lies somewhere between models 1 and 2 because different styles of player acquisition and deployment come in and out of vogue. In the early 20th C., first basemen didn't hit as much as today but were better fielders. Today SSs hit nearly as well as traditional power positions do, but in the 1970s they more closely resembled the Concepcion model than the Ripken model. In this same way, we know that 3B and 2B underwent a major change.

(Is catcher the next position that will undergo a major change? It's happened once when the equipment finally allowed guys like Schalk, Schang, and Ferrell to start catching 5 days a week. Maybe it's the next frontier in finding bats????)

Anyway, so maybe the flux between the two points of view is part of what makes the question a moving target?

I'm so nerdy that...

I look up Holway numbers for kicks at 9:00 PM on a weeknight.

Anyway, I noted his batting averages for each season, then I noted how many games his teams got decisions in. For the years he was listed as one of the main three outfielders, I multiplied the decisions by 3.5 to get his at-bats and figure his hits. After throwing in some VZ, Cuba, and P.R. nubmers, summing it all up, and figuring the combined average, I got

2405 bats, 670 H, .278 average.

As Dana Carvey once said while in the guise of GHW Bush: "Not gon' dewit."

Doc, are you talking about Oms here?

I'm sure it is. However, there's room for play in the following equation (C+F)+B where C is a constant. The larger the constant C, the higher the emphasis on fielding relative to batting.

What I meant by this is as follows: Once we figure out that if we add 1 to 1B and 5 to SS as fielding value this evens out their total value (fielding + hitting), then we could also add 4 and 8 and it would still even out. The difference lies in how this effects the percentage of value in hitting and in fielding.

When comparing Win Shares to WARP, WARP uses a larger constant so that the hitting/fielding balance has more emphasis on fielding; Win Shares emphasizes the hitting. Which is closer to the truth has yet to be determined.

For defensive value, it is the plays that Bancroft makes that other SS's dont' that give him value. No matter how bad a defense a team has it will still get 27 outs and the SS is bound to get some of those outs. You coudl do this using fancy computers like UZR or by using James' method of estimations. And if you believe that a replacement level SS is one that is replacement level offensively and average defensively, as I do and WARP doesn't, more defensive value is taken away.

As for Bancroft V. Sewell, they are nearly identical in WS in my system, one reason I never voted fro Sewell. In WARP, however, Sewell comes clearly ahead.

Interesting that WARP emphasizes defense and WS emphasizes hitting, and WARP likes Sewell and WS likes Bancroft. Can anybody explain that?

Meanwhile, the Minnesota Twins have run guys named Bartlett, Punto, Rodriguez, Castro and Rivas out at the SS position this year, and I agree that they are more than willing to settle for a guy who is BOTH average defensively AND replacement offensively.

OTOH it seems to me that a SSs actual value would be equal to how far above replacement he is on offense and on defense. IOW the threshold (minimum value) to get playing time is one thing, but measuring the value that is contributed during that playing time is another. Two different standards.

Sorry, forgot to mention that I was talking about Gene Benson.

FWIW, batting-fielding win for peak voters.

Here are the players who were in the top 5 in their league in batting-fielding wins the most times (eligible through 1960 in CAPS, * means eligible but unelected). Take this with usual disclaimers about not adjusting for league size or strength:

15--Aaron
14--COBB, RUTH (not including pitching), Musial, Mays, Bonds
13--WAGNER, SPEAKER, HORNSBY, Williams
11--GEHRIG, OTT
10--G. DAVIS, LAJOIE, E. COLLINS, Mantle, Scmidt
9--Rickey
8--DELAHNTY, Grich
7--GLASSCOCK, MCPHEE, FOXX, Mathews, Morgan
6--BARNES, G.WRIGHT, BROUTHERS, DUNLAP*, J.JACKSON, BANCROFT*, VAUGHN, P. WANER, BOUDREAU, MIZE, Berra, Kaline, Santo, Boggs, Yount, Ripken, Thomas, ARod
5--HAMILTON, DAHLEN, FLICK, BAKER, FLETCHER*, SISLER*, SEWELL*, HEILMANN, GREENBERG, DIMAGGIO, DOERR, Banks, F. Robinson, Yaz, Allen, Brett, Raines, Larkin, Edgar, Griffey

The eligible but not yet elected are Dunlap, Bancroft, Sisler, Sewell, and the unheralded Art Fletcher. (Also note both Boudreau and Doerr.)

Make of it what you will

Interesting that WARP emphasizes defense and WS emphasizes hitting, and WARP likes Sewell and WS likes Bancroft. Can anybody explain that?
-----
WS 'likes' Bancroft only in that he has the higher career total, which is a function Rabbit's lonnng career & WS' low replacement level.

WARP gives Sewell a HUGE advantage in league strength - almost 30 'wins' in going from WARP1 to WARP3. Even I, a FOJS, would not deem the AL-NL difference to be that large. But only using WARP1, Sewell is ahead by 209 batting runs above replacement, and Rabbit only has a 64 fielding runs above average advantage of Joe.

>The eligible but not yet elected are Dunlap, Bancroft, Sisler, Sewell, and the unheralded Art Fletcher. (Also note both Boudreau and Doerr.)

I like everybody in that group, which is only to say they (except Fletcher) are in my top 50.

TomH, can you provide the WARP numbers for Bancroft, Sewell and Maranville?

..BP stats ...WARP1 (raw) .......WARP3 (adj)
shortstop BRAR/BRAA FRAR/FRAA BRAR/BRAA FRAR/FRAA
Bancroft ..319/..69 .701/.106 .226/.-24 .449/..72
Maranvil ..213/-147 .991/.166 .115/-245 .646/.123
Sewell.....422/.183 .514/.102 .315/.112 .420/..89

actual G ..WARP1/WARP3
Bancroft ..1913 ...111.1/77.0
Maranvil ..2670 ...131.0/85.8
Sewell... ..1903 ...102.9/87.4

trying to line them up betterer...


..BP stats ...WARP1 (raw) .......WARP3 (adj)
shortstop BRAR/BRAA FRAR/FRAA BRAR/BRAA FRAR/FRAA
Bancroft ..319/...69 ..701/.106 ..226/..-24 ..449/..72
Maranvil ..213/.-147 ..991/.166 ..115/.-245 ..646/.123
Sewell.....422/..183 ..514/.102 ..315/..112 ..420/..89

........actual G ..WARP1/WARP3
Bancroft ..1913 ...111.1/77.0
Maranvil ..2670 ...131.0/85.8
Sewell... ..1903 ...102.9/87.4

Thanks TomH. My ignorance is showing certainly. I quit looking at WARP a year or more ago when the numbers kept changing all the time, so I forget exactly how all of these numbers are calculated. As you know, my re-interest in WARP is rekindled by its apparent valuing of defense more highly than WS.

Anyway, the numbers above are very interesting. I never cared about WARP3, BTW, so I'll stick to WARP1.

• A question, which actually would apply to WARP1, 2 or 3: Can one just add together batting and fielding runs?

If so, here is what I get.

Player...B+FRAR...B+FRAA...WARP1

Bancroft 1020 175 111
Rabbit 1204 19 131
Sewell 936 285 103

• Second question: WARP is apparently based on value above replacement. Is that correct?

• Then, why bother then to publish value above average?

• And BTW, RAR (whether B or F) must be based on positional replacement, right? And I'm guessing that RAA is also based on positional average?

• Most importantly, to paraphrase Howard Hughes (at least in the movie): Does this look clean to you?

E.g. Maranville is 400 FRAR ahead of Sewell and only 200 BRAR behind? Bancroft is right in between on both FRAR and BRAR, but that means he is +/-200 FRAR and +/-100 BRAR.

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

For BRAA + FRAA of course Sewell ends up way ahead with a total of 285 RAA versus Bancroft 175 and Maranville 19. This seems to have no bearing on the uber-rating. But frankly this is the one number that agrees with my subjective sense of where these players should rate. That doesn't make it right, just interesting. How is it that RAA seems intuitively to get to the right result?

So many questions, so few answers.

probably best answered by the WS versus WARP thread archive, but in the main, WARP1 is value above positional replacement.