Mike posts, about Paul Waner's playing fields, "In Pittsburgh, this was more true of left field than it was of right field. The grandstand in right cut the coverage area down considerably. Right field wasn't "small", mind you, as the grandstand angled out pretty quickly, but it wasn't as tough as was left field."

Thanks, Mike. That might argue that Paul's contemporary rep was a bit overblown. While I know the reputation, I'm not completely trusting, because of the Cuyler thing. It looks like Paul's defense rep may be based mostly - perhaps close to entirely - on the 1927 season. This season, which may be the most famous in all of baseball history for reasons having to do with the Bucs' World Series opponents and their 60-homer man, has Paul, as well as Lloyd, at their best. It is possible that contemporary observers fixed their idea of Paul's defense in that season and never bothered to update it, because the memory was so strong. This happens to memories of sports figures. Stan Musial's contemporary rep in St. Louis, for example, is largely driven by his sudden and unexpected breakout in 1948. Joe D's and Teddy Ballgames' reps would be largely from 1941, except that Ted hit a homer in his retirement plate appearance. Contemporary reps are much more trustworthy if the player never did anything superspecial in a very high-profile season or postseason. That's my worry. It's easy to overblow a defensive rep if what you're basing it on is an early season when the player was fully healthy and his team was hot.

Mind you I'm not willing to discard the rep. It really does appear to be monolithic, and you have to respect that. It's also possible that Paul's defense skills were really great, but having less room to go chase fly balls in his particular home outfield spot kept his range factors down. This is, of course, what did NOT happen to Clarke and Sheckard. They had loads of room to roam in their left fields. But Bill James, whose defensive Win Shares are a very good system (though not currently exact state of the art), essentially see Paul as Stan Musial defensively. A guy who could certainly play center if you needed that, but who would not be your first choice. Your first choice would be Lloyd or Carey or Terry Moore or someone like that. Overall grade of B. That's pretty much Musial, and the B grade applies to Waner, too. But Dan asked specifically about contemporary rep, and few are as unwavering as Paul Waner's on defense.

I'd really be interested in what particular features of defense cause Dan's one rating to be so high. That might help sort things out a bit.

Thanks again, Brock

Here's a King Kelly defense comment, as opposed to a Paul Waner one. If you read Bill James' comments in the New Historical, you get a picture of a defender with enormous range and assists quality, but really lousy error prevention. Overall, Bill gives Kelly a B+. I have a problem with that, but the problem stems from a quirk of my analytical methods. I ignore defensive errors. I treat them, as you all know by now (see Harry Stovey), as an offensive stat, when I can get data. I do that because I decided, about 20 years ago, that the only important point about a defender was how many outs he recorded amortized to his field of realistic opportunity. I don't care whether the reason that the player did not record an out was because he lacked range or had a bad glove. For example, consider a fly ball down the left field line. Greg Luzinski doesn't get to it and it bounces to the wall for a double. Lonnie Smith gets to it but it bounces off his glove and skids to the wall for a two-base error. What's the difference? Nothing, except that the official scorer gives Smith an error while not penalizing Luzinski's stats at all.

While I'm certainly aware that not all errors have the same value as all hits, I decided that the overall value of such was so close that you could actually safely ignore errors if you knew the player's adjusted range factors. This would, of course, vault King Kelly into the A category of defenders, and that is exactly what I intend to give him when making out my ballot. The reasons for my posting this are two:

1. I'm trying to convince people of this, particularly in ranking Kelly, because he's the poster boy.

2. I'm looking for counterarguments, to test my theory out.

So please feel free to go ahead and rip. I wrote this post for exactly that reason. And yes, I'd love to see some of the advanced number crunchers here look at Kelly's stats and tell me what their methods would do to him if you amended the method to simply ignore errors.

Thanks again, Brock

I think we should ignore all errors in fielding and only count throwing errors as a stat. Your judgement seems to jibe with that.

My current FWAA are just an average of BP FRAA and Fielding Win Shares for the pre-1987 period. You'd have to dig into those systems to know what drives it.

Chris says that Crawford had some ordinary years by OPS+ mixed in, and I see that he has Waner and Kaline ahead of Wahoo Sam. Well, Crawford's 5th best season by OPS+ was a 157, Kaline's was 149 and Waner's 144. 10th best season Crawford was 147, Kaline's was 142 and Waner's was a whopping 131 (lower than Winfield's). 15th best: Crawford 124, Kaline 127, Waner, well he didn't have 15 years of ?100 games and ? 100 OPS+.

OPS+ doesn't really sell Waner too short. I get his "fair" OPS+ at 136, just two points above his real-life figure, and just about what one would guess given his OBP-heaviness. And there's no hidden baserunning value or anything. The keys are the defense and the durability. Waner accumulated a tremendous amount of PA's each season--from 1926-38, he averaged more PA per season than a league-average player playing all 154 games--and both numbers and (according to recent posts) reputation suggest he was a knockout right fielder. So when you combine the 150ish OPS+ seasons with high Gold Glove defense and rock-solid durability, you get a really, really high peak. It's basically the Billy Williams model, but taken to the max.

Chris says that Crawford had some ordinary years by OPS+ mixed in.

Yes, in the context of five-year consecutive peak, which is what I was discussing, explaining the reasons why Crawford does less well than many other right fielders at this particular measure. Both Kaline and Waner (not to mention Clemente) had their best seasons by rate more consecutively than Crawford did.

For his career, Crawford was obviously a better hitter than either Kaline or Waner. For all I know, he may have been a better hitter for his five-year consecutive peak, but he was not near the fielder either of them were at their best

I rate Waner and Kaline ahead of Crawford because their total packages, which includes baserunning and defense, were better than Crawford's.

Waner, well he didn't have 15 years of ?100 games and ? 100 OPS+.

Well, actually, he did. His 15th best season by OPS+ in seasons in which he played more than 100 games is only 102, so it's not like it helps his case, but if you are going to say that he didn't have 15 seasons with 100 games played and an OPS+ of 100, I am going to have to disagree.

If we're going to take a career perspective on Waner vs. Crawford vs. Kaline, it is worth noting that Waner does less well by career partly because he got a later start than they did.

It has been noted earlier on the thread that he starred in the PCL for three years before being bought by the Pirates. Baseballlibrary.com notes that he hit .369, .356, and .401. in those three seasons, so it's highly likely that he would have reached the majors no later than age 21 had he lived in the east. (A debut season at age 23 with a 147 OPS+ suggests he would _probably_ have been ready earlier, in addition to the likelihood that his PCL MLEs would be outstanding).

My ranking of Waner does not include MLE credit, but if you're going to go looking at 15th best seasons, Waner's PCL years ought to be kept in mind. His 15th best season, with his real quality of play 1923-25 taken into account, was probably about a 120-25 OPS+, since his three lowest over 100 in 100+ game seasons were 102, 108, and 120.

Um, *does* Waner deserve minor league credit? It would be really helpful, I think, if there were some sort of spreadsheet/database of MLE's for ALL electees and backlog candidates, so that each voter could determine whom to credit and how much. I have the nagging feeling it's only the borderliners (Keller, Averill, Bob Johnson etc.) whose minor league records have been scrutinized. Certainly Joe DiMaggio merits a year, as I discussed on his thread.

I don't think MLEs for Waner were ever made, at least I didn't find them in his thread (which does include his raw PCL stats). It might be interesting to see them, even though they wouldn't affect my ranking in this ballot, I'm a completionist like that.

Now... checking out my spreadsheet, depending on the amount of MLE credit he gets, he might catch Delahanty in a combined corner outfield ranking.

Using the data posted on Waner's thread and doing a very rough-and-ready translation from AAA to ML level (using .93/.86 as the conversion factors) and not adjusting for league offense level or park effects or anything, Waner's OPS+ totals for 1923-25 look to be in the vicinity of

1923 - 127
1924 - 121
1925 - 140

Those could easily be off by 10 points or so, depending upon PCL league offense levels.

But the only scenario whereby he doesn't deserve MLE credit is if these are 10 points too high, in which case 1925 is the breakout year, though 110 OPS+ would probably have been good enough for a starting job.

I just finished reading the Enos Slaughter thread, and it seemed to hang up on Enos' defensive rep. Of course, I know that one, at least in St. Louis. It's not good. Enos played right field, starting when he was a young fast kid, in a park where that was the easy field. He had a rep for an arm, but nothing else. Bill James gives him a B-, and I would think that's, if anything, too high. The main positive I have to offer is that he was playing in a tough peer group. During Enos' career, the Cards' farm system was delivering kid outfielders wholesale, and they could all run. Being the worst of a lot consisting of Terry Moore, Stan Musial, Johnny Hopp and people like that isn't a disgrace. But Enos was indeed the worst of the peer group. If Slaughter hadn't had the big arm, I doubt that even Eddie Dyer would have had Stan Musial playing first base (Dyer was the manager during Musial's first tour at first). - Brock

Thanks for that, bjhanke. I hereby disavow my published ratings on Slaughter and encourage voters to consider him a roughly league average defensive outfielder (albeit very good in 1939, and pretty good in 1942, his peak offensive year). I will have him around 170M, near Winfield--another long career, low peak guy.

Dan R says, "Thanks for that, bjhanke. I hereby disavow my published ratings on Slaughter and encourage voters to consider him a roughly league average defensive outfielder (albeit very good in 1939, and pretty good in 1942, his peak offensive year). I will have him around 170M, near Winfield--another long career, low peak guy."

Dan, are you sure? I mean, I've just delivered a reputation from my teenage years, while you've actually crunched hard data. Your hard data crunches have proven to be pretty convincing. I don't want to discount what I just posted, because it did address an issue in the Slaughter thread, but your work is pretty darn good, if you ask me. And Slaughter did have a tough peer group in St. Louis. In fact, I trust your work enough that I wasn't sure, when I read your post, that you weren't being sarcastic, since you had done hard data work and I was repeating other people's opinions from an era where hard defensive analysis was not exactly an art form. I do hope you weren't, but in any case, I'd be more inclined to, at least, average my old rep and your hard data, rather than just discarding your work. Now, if you had some questions about what your results were from the start, that's different. But your work is good. I tend to trust it.

- Brock

It would be really helpful, I think, if there were some sort of spreadsheet/database of MLE's for ALL electees and backlog candidates, so that each voter could determine whom to credit and how much. I have the nagging feeling it's only the borderliners (Keller, Averill, Bob Johnson etc.) whose minor league records have been scrutinized. Certainly Joe DiMaggio merits a year, as I discussed on his thread.

That is a great idea. I was pretty critical of the MLE credit for Averill, e.g., not so much on principle, but on the basis that we weren't comparing apples to apples--IOW we had MLEs for Averill but not for other players who might have deserved them and for whom it might have made a difference. Obviously we can't MLE everybody, but I just thought we were a bit selective.

Now, for this exercise, it woulda been good to revisit what the consensus was about MLEs for all of these guys. Oh well. Not a perfect world.

I do think MLEs (specifically MiL MLEs) are one of the several great innovations that this project has utilized and which deserve wider dissemination. (Other forms of MLEs, mostly notably MLEs for NeLers, are even more important but not on-topic at the moment.)

BTW, Crawford was an all-timer for durability, too. Led the lead in games played how many times? Six, maybe?

BP FRAA loves Slaughter's D; Win Shares likes it a good bit. Both are quite poor defensive statistics that show weak correlations to modern play-by-play metrics. Defensive Regression Analysis, which is a far superior fielding stat, thinks Slaughter was just about average for his career, as well as average over the 41-42-46-47 span that would determine his war credit. Now, there IS positive fielding value there that's hidden by his decline phase; DRA has him at +57 through 1949, and then giving it all back in the 1950's. But my FWAA (the FRAA/Fielding WS hybrid) doesn't see him start to tail off until 1954, and then only mildly so. If I use DRA instead of the FRAA/Fielding WS hybrid, I get Slaughter at $175M including war credit, a rather different story from the $200M he was at before, although still waay over the HoM in/out line. That's similar to Willie Keeler, Dwight Evans, Joe Kelley, Minnie Miñoso, and Dave Winfield, as well as ineligibles Sammy Sosa and Larry Walker. I was just looking for anecdotal reports to see if there was anything to back up FRAA and Fielding WS here, and it doesn't seem like there is--not just judging by your comment, but also by the absence of rave reviews. This is where I will vote on Slaughter, and I would encourage voters who make use of my WARP to do the same. Here are Slaughter's seasonal WARP2, including war credit, if we use DRA instead of FRAA/Fielding WS:

1938 0.7
1939 7.0
1940 4.1
1941 2.6
1942 7.4
1943 4.8
1944 4.8
1945 4.8
1946 4.0
1947 3.3
1948 5.8
1949 6.0
1950 0.7
1951 1.6
1952 4.4
1953 1.4
1954 -0.2
1955 0.5
1956 0.5
1957 0.7
1958 1.0
1959 -0.6

3-year peak: 20.4
7-year prime: 40.5
Career: 66.0
Salary: $175,870,543

I have Waner in striking distance of 5th place with 3 extra years, up from 7th. I'd love to see proper MLEs done in his case. If Slaughter is only an above average glove (not a terrific glove) he slides from 9th to 13th behind Reggie.

DL from MN, according to DRA and apparently to reputation, he was not even above average for his career--well above average pre-1950, and then a good bit below post. Seems like both FRAA and Fielding WS screwed the pooch on him in particular for some reason.

Slaughter's rating by Fielding WS could be influenced by the Cardinals excellence during his years there. During 1938-53, minus his war years, I see the Cards with 1 sub-.500 record, his rookie year of 1938. Most years are decently above the .500 level. If an individual's fielding numbers in Win Shares is influenced by overall team quality (i.e., good teams have more Fielding WS to divvy up, larger pie raises everone in context) then it doesn't hurt if your team is always good.

Unless the team consistently overperforms its component statistics, that shouldn't make a difference, should it? Yes, there are more WS to go around, but those wins were the product of a higher level of production, so the WS-stats relationship (whether those are, say, doubles for hitters or putouts for fielders) shouldn't change.

Dan, thanks for the post. It all makes sense now. As you can tell from the Waner comment and this one, I don't want people to be TOO influenced just by the fact that I "was there" in St. Louis in the 1950s and had access to a few people who were there in the 1920s. I wasn't even 21 when we're discussing most of the guys I've posted reps for. My judgment or my memory could play tricks.

When you posted up in detail here, I checked a couple of things out. Win Shares does exactly what you said it does. It vastly downgrades Slaughter's defense as soon as he's traded to the Yankees, but not one season before, even if you adjust for Enos' lower playing time. That is, Slaughter's defense looks like a ballpark effect to WS, as Al Peterson suggests. And yes, that shouldn't happen, as you say. Team quality is not supposed to affect WS, and I think it usually doesn't. Besides, the Yankees are a real good team when they have Slaughter, so team quality doesn't change except for his time on the Kansas City Shuttle. WS is just doing something weird.

But if you go over to Total Baseball, Pete Palmer's Linear Weights, usually a really flawed defensive system, has the drop right where your DRA has it, in 1948-49, one year before the system sees a big offense drop. Your DRA and Linear Weights both have what looks like a normal defensive progression for a player. WS sees Enos retaining a lot of his defensive value long after most outfielders have declined. Normally, I would trust WS over LW, but in this case, with your DRA and the rep, I think you've got it right. In terms of career value, LW, which uses the average as a baseline, has Enos with exactly zero defensive value for his career. That is, Pete Palmer thinks his overall career defense was exactly average for a right fielder.

Thanks again for the post, Brock

IIUC, Dan's reply to Alan is correct.

Alan's idea is on target if
(a) Bill James's allocation of outfielding to LF, CF, and RF (is it arbitrary?) gives a share to RF that is "too much" at this time and place in baseball history. Then all 1940s mlb RFs are overrated but those with excellent teammates at LF and CF are overrated relative to their fellow RFs, or
(b) the same for allocation of fielding to outfield and infield, or
(c) the same for allocation of defense to fielding and pitching.

Effect (b) overrates all outfielders. Effect (c) overrates all fielders. So their impact on the RF ratings would be smaller than (a) and smaller than that.
(a) > (b) > (c)

Effect (a) depends on excellence of team fielding at LF and CF only, perhaps only at CF. Effect (b) depends on excellence of team fielding. Effect (c) depends on excellence of team defense.

So effect (a) is a more plausible cause of overrating Enos Slaughter's fielding in 1940s mlb (St Louis). Indeed, in effect, in magnitude, I suppose that
(a) >> (b) >> (c)

If Win Shares is overrating Slaughter, to whom should it be allocating those outs instead?

Paul/Dan,

My thoughts on defense were more of the (a) type listed in #124. Not that I can prove anything but just a rationale as to the apportioning shares by position and how Slaughter might be a case that doesn't conform well.

Yes, there are more WS to go around, but those wins were the product of a higher level of production, so the WS-stats relationship (whether those are, say, doubles for hitters or putouts for fielders) shouldn't change.

Of course, this assumes that WS is allocating the defensive responsibilities correctly. If the positional weights are not correct, then value moves from the underrated positions to the overrated ones, but that's team-independent. If the positional rating systems are not calibrated properly, say they are too conservative, then value moves from the above (team) average performers to the below (team) average performers, which is a team effect that will cause good teams to subsidize poor performers, and bad teams to penalize good performers.

We had some discussion of this on the Roger Bresnahan thread. The discussion started out being about Win Shares, (BP) WARP, and Bresnahan in CF, and then about Roy Thomas, a contemporary of Roger with an excellent fielding reputation for the Phillies. He had the distinct pleasure of playing a quality CF for pennant contenders before Lajoie, Delahanty, Flick, etc. left for the AL, becoming a "terrible" fielder when the Phillies went down the toilet in all departements, and becoming excellent again after the team rebuilt (fielding assessments from Win Shares). I calculated a correlation coefficient of .84 between his rating and the overall team fielding rating, which leads me to suspect that Win Shares may not isolate the individual's impact as well as it might.

Dan R, I know you're a busy person, but you have access to another fielding rating system, DRA. Have you had occasion to calculate ratings for Roy Thomas? It might be interesting to compare them with WS and FRAA.

Here are the seasons I have DRA data for for Thomas. Note that this is an old version of DRA. The first column is putout runs, the second is arm runs.

1899 -7 0
1900 -38 -1 (ouch!)
1901 -12 -6
1902 -22 3
1903 10 1
1904 5 3
1905 15 6
1906 12 -5
1907 10 -2

Needless to say, a much higher standard deviation than in my published WARP. But both systems see 1900 as his worst year and 1905 at his best, and have a .64 correlation over these 9 data points.

I calculated a correlation coefficient of .84 between his rating and the overall team fielding rating, which leads me to suspect that Win Shares may not isolate the individual's impact as well as it might.

It doesn't.

There are three fundamental issues with Fielding WS:

1. James allocates them at the team level based on a hard percentage allocation between pitchers and fielders; fielders *always* get around 18% of the total team win shares. This has the effect of overvaluing fielders in low-BIP eras (like today's game) and undervaluing them in high-BIP eras (pre-1920).

2. James allocates DWS by position based on another hard percentage allocation, with little regard for variations in the distribution of BIP, and the distributional adjustments he DOES make are (mostly) too small. This has the effect of overvaluing infielders on GB teams and outfielders on FB teams (among other things).

3. James won't allocate negative win shares, no matter how bad the fielders are at a position. This has the effect of limiting the effect of good fielders on bad defensive teams - a fielder who might be worth, say 4 WS might not have four WS available to him when the positional allocations are made on a bad team because a player who is calculated to be worth -2 WS is bumped up to zero.

The first and last effect have their biggest effect on "really" bad teams. Really bad teams usually have lots of balls in play in part because they don't have pitchers who get strikeouts (and in part because they have bad defenses as well). Really bad teams also usually have at least one set of fielders who SHOULD get negative win shares.

-- MWE

Win shares is more likely to misvalue outfielders based on the quality of the team than other defensive positions, because 30% of the weighting given to outfield defense at the team level is team defensive efficiency. So outfielders on an above average defensive team will always be above average by this measure, while outfielders on a below-average defensive team will always be below average by this measure.

the WS-stats relationship (whether those are, say, doubles for hitters or putouts for fielders) shouldn't change.

For fielders, it does, because balls in play generally vary inversely as a function of team quality. A typical fielder will generally make fewer plays on a good team than he will on a bad team, because he'll have fewer opportunities. James recognizes this and tries to account for it by an up-front assumption that a good team has a good defense and a bad team has a bad defense - but he counters it to some extent by his caps and allocations - which are intended to transfer value from position players to pitchers, as he himself admits (he wanted a system where the best pitchers were close in value to the best position players, and the only way he could do that was to put arbitrary limits on the amount of value that could be given to fielders).

-- MWE

1899 -7 0
1900 -38 -1 (ouch!)
1901 -12 -6
1902 -22 3
1903 10 1
1904 5 3
1905 15 6
1906 12 -5
1907 10 -2


1900 is the one season where WS sees Thomas as below-average relative to his team. (It sees him as considerably below-average relative to league in 1903 and 1904, when the Phils fielding was putrid, but him as one of the better fielders on the team, relatively speaking. WS sees Thomas as worse in 1904 than in 1900, relative to league.) Win Shares sees 1901 as his 3rd best fielding season behind 1905 and 1907, but that happens to also be the best team defense that he was a part of, according to WS.

Thanks for the prompt response Dan. And thanks for the WS comments Mike and Chris.

Note: see post 60 on the Bresnahan thread for the last of the Roy Thomas posts.

MIke Emeigh (post 129) says, among other things, "2. James allocates DWS by position based on another hard percentage allocation, with little regard for variations in the distribution of BIP, and the distributional adjustments he DOES make are (mostly) too small. This has the effect of overvaluing infielders on GB teams and outfielders on FB teams (among other things)."

Mike is dead right about the WS mechanics here, and this issue may be the one governing the Enos Slaughter thing. Essentially, the differences among the various methods seem to be almost entirely focused on Enos' defensive value from 1949 through 1953. This happens to be the "Robin Roberts" period in baseball history, where pitchers responded to the huge numbers of walks from 1946-49 or so by throwing high fastballs and just living with the occasional homer. Well, doesn't that imply that they are throwing a LOT of fly balls? I haven't broken down fb/gb ratios for these years, because I don't have a historical baseline to make a context from, but this would make a lot of sense. WS has Enos' value remaining artificially high during this period simply because he's an outfielder. LW and DRA don't have that particular bias, so they don't fall into that exact trap.

The point about the center fielder's quality being a factor may very well apply, too. Terry Moore retired after 1948, and the next run of Cardinal center fielders is, well, not the best. From 49-53, there's Chuck Diering, Peanuts Lowrey, one year of Musial, and then Rip Repulski. What happened was that Branch Rickey's farm system was unravelling, the man himself having left the organization years before, and so the system was no longer doing what I said it did in the 1930s and 1940s (turning out fast kid outfielders wholesale). Lowrey, in particular, was old and none too good. Musial actually ran him off the position for the one year. - Brock

Preliminary ballot, LF/RF combined:

Preliminary ballot:

1. Babe Ruth
Ted Williams
Stan Musial
2. Hank Aaron
3. Mel Ott
4. Frank Robinson
Rickey Henderson
Ed Delahanty
5. Reggie Jackson
6. Sam Crawford
Carl Yastrzemski
7. Pete Rose
8. Joe Jackson
9. Harry Heilmann
10. Al Kaline
11. Paul Waner
12. Tony Gwynn
Tim Raines
Jesse Burkett
Fred Clarke
13. Roberto Clemente
Willie Stargell
Al Simmons
Billy Williams
14. Enos Slaughter
15. King Kelly
Sherry Magee
16. Elmer Flick
17. Dave Winfield
Goose Goslin
Zack Wheat
Monte Irvin
Joe Medwick
Joe Kelley
18. Willie Keeler
Minnie Minoso
19. Dwight Evans
Charlie Keller
Ralph Kiner
Jimmy Sheckard
Harry Stovey
Reggie Smith
20. Sam Thompson
Charley Jones

As you all know, I spend time trying to figure out career lengths relative to their peers for early players. I ran through the five early RFs here (Crawford, Keeler, Flick, Kelly, and Thompson), FSEd them, applied some of my plausibility deductions, and came up with the following:

In actual games played, among 50% RFs, career through 1920, BB-Ref generates Crawford first in the group at 2517 G. Then Keeler, 2nd at 2123. Flick, 6th at 1483. Kelly, 8th at 1455. Thompson, 9th at 1407. That is, these guys are all among the top ten of their peers - even Thompson - and Crawford and Keeler sort of lap the field.

But if you look at all outfielders in the same time period, including the CFs and LFs, the rankings change a lot. Crawford is still first - the most games played by any 50% outfielder up through 1920. Keeler still holds down 4th place. But Flick drops to 32nd, Kelly to 36th, and Thompson to 39th. So it is clear that the really big careers are Crawford and Keeler and no one else in this peer group.

Then I just ran BB-Refs FSEs. Crawford and Keeler don't move much, to 2685 and 2395 games, respectively, so I just used their actual games played as plausible. Nor does Flick, to only 1614, so I used that. Kelly, though, goes all the way up to 2210, and Thompson climbs to 1723. I estimate that there's about 200 games of FSE fluff in Kelly's numbers (which is low for his time period, but he doesn't have any of the big red plausibility flags that plague the 1870s guys like O'Rourke), so I give him credit for 2000 games FSEd and plausibilitied. Thompson, because of his late age start, has almost no fluff. I credit him with 1700 games, losing only 23 to plausibility. This is important for you Thompson career length haters. If you don't buy my approach, Sam ranks lower. Flick gets no plausibility deduction. His career is late within the group, and the deductions are greater the earlier you go. MUCH greater in the 1870s. Remember, for this purpose, I'm making plausibility deductions for these guys relative to each other. If I were to use these numbers to compare them to actual games played by 154-game schedule guys, that would be wrong, and I would drop them all down because BB-Ref FSEs to 162 games.

What happens, then, is this list, which I do trust as a reasonable comparison of these particular five guys:

Crawford 2517
Keeler 2123
Kelly 2000
Thompson 1700
Flick 1614

As you can see, Kelly benefits a lot by this analysis, while Flick just gets hammered. Thompson ends up holding his own, for whatever you think that's worth. He ends up ahead of Flick, and I am completely comfortable with that, even though I am an Elmer Flick fan. His career is just short and there's no FSE credit to pull it up.

I hope this helps someone other than just me.... - Brock

bjhanke, did you really just say "plausibilitied?"

I plausibility
You plausibility
He/she/it plausibilities

I have plausibilitied
They plausibilitied
You will plausibility

...

:)

Kid, have you plausibilitated yourself? - Brock (the unreconstructed hippie)

Brock #134 replied to me #124, perhaps among others
The point about the center fielder's quality being a factor may very well apply, too. Terry Moore retired after 1948, and the next run of Cardinal center fielders is, well, not the best. From 49-53, there's Chuck Diering, Peanuts Lowrey, one year of Musial, and then Rip Repulski. What happened was that Branch Rickey's farm system was unravelling, the man himself having left the organization years before, and so the system was no longer doing what I said it did in the 1930s and 1940s (turning out fast kid outfielders wholesale). Lowrey, in particular, was old and none too good. Musial actually ran him off the position for the one year. - Brock

In #124(a) I meant the opposite effect. Some of the credit due an excellent CF such as Terry Moore may be assigned to the RF because there is a kind of ceiling on Moore. Evidently that doesn't fit the particular time points: win shares apparently overrates Slaughter in baseball middle age, immediately after he teamed with Moore.
I suspect that my point doesn't fit the win shares system, either, but I don't have time to study it now.

On the other hand, I don't see that Mike Emeigh #129, presumably (2) and (3), supports any intra-outfield effects unless one of the team (Slaughter in this application) deserves negative win shares.

Full Seasons equivalent games, ordered and ranked by f.s.e. fielding RF

Paul - You're right. Your point "a" leads to the opposite of what I was saying. But I wasn't looking at what you wrote when I thought that paragraph up. I was thinking about Mike Emeigh's post #131, which talked about "opportunity." What happened was that I put up a paragraph of a post that I hadn't thought all the way through, but that I knew might support Slaughter's defense. Here's what I was starting to think out, now thought through much more.

It's possible for Slaughter defense defenders (heh) to use the Terry Moore timing to argue that Enos' defense is underrated. Here's the possible scenario. Up front, I want to say that there is no "insider" info here. I've never talked with anyone who "was there" about this. In fact, I'd never thought about it at all until this thread.

Terry Moore was not only a superglove in center, he was captain of the Cardinals. This put him in a position sort of like Nap Lajoie's in Cleveland, although at a much lower level. Say Moore went to Slaughter and said, "Enos, as you know, all teams have a system in place to deal with who takes fly balls hit in between outfielders. Ours works like this: if a fly ball is hit between you and me, you take off for it, but listen to hear if I'm yelling. If you do NOT hear anything, the ball is MINE. Your job is to veer off deep and be the backup in case it somehow gets past me. If you DO hear me yelling, don't wonder what I'm saying. Just the fact that I'm yelling says that I can't get to this one. It's YOURS. Go for it. Don't pull up and hold it to a single. Go for it all out. I'll be there deep, backing you up."

If you think about it, this is a simple and elegant approach to the Alphonse and Gaston collision issue that occasionally arises between outfielders, where both of them can get to the ball, and they're trying to figure out what to do, yelling to each other above the crowd noise. Enos doesn't have to worry about what Terry is yelling. Just the fact that he is or is not yelling tells Enos what to do. And the result? Terry Moore gets every ball he can reach, regardless of whether Enos could have gotten to it or not. Terry just doesn't yell, and the ball is his. A Slaughter defender could say that Moore was vulturing some of Enos' putouts away from him. The reason that Enos' range doesn't decline in 1949 is not that WS has a bad system. It's that Terry Moore stopped vulturing Slaughter putouts in 1949, and the next wave of Cfs weren't good enough to vulture. Enos' defense should be upgraded, because it's artificially suppressed when Moore is there. You should come up with an early defensive evaluation that makes the WS between 1949 and 1953 a realistic decline. That upgrades Enos' defense a lot.

I, personally, don't buy this. I don't think there are enough fly balls that really can be caught by either of two fielders to make a serious difference. I think the reason that we can all remember several outfield collisions and Alphonse and Gaston routines is that they are flashy way out of proportion to their actual rate of occurrence. But a Slaughter defender could mount this argument, and it's not a silly one. I thought I ought to include it just out of intellectual honesty.

That's what I was trying to get out in my paragraph. I hope it helps, even though I don't buy the argument myself.
- Brock

Preliminary RF ballot.

1. Babe Ruth
2. Hank Aaron
3. Frank Robinson
4. Mel Ott
5. Pete Rose
6. Reggie Jackson
7. Al Kaline
8. King Kelly
9. Sam Crawford
10. Tony Gwynn
11. Enos Slaughter
12. Dave Winfield
13. Joe Jackson
14. Paul Waner
15. Roberto Clemente
16. Harry Heilman
17. Dwight Evans
18. Elmer Flick
19. Willie Keeler
20. Sam Thompson

MWE: "1. James allocates them at the team level based on a hard percentage allocation between pitchers and fielders; fielders *always* get around 18% of the total team win shares."

I haven't read the details in a long time, but this is not quite my understanding. I too have been critical of James' lack of flexibility in allocating WS to team defense, but there IS some movement, based on things like DER and pitcher KOs and hits allowd in play, is there not? Or am I missing what you are saying?