(IOW Yaz gets one hit every 3 months over his career more than Larry Doyle. There, the case against Yaz!)

Let's compare cranberries and canteloupes while we're at it.

What is crappy is the conversion to the all-time context.

Which is the only context that Win Shares offers for its pitching/fielding split.

In this view, Doyle is horrible defensively.

WARP works out what the fielding stats appear to say about the player in terms of runs saved or lost and goes with that. Win Shares hands out grades on a number of exams, each on a different facet of defensive play at the position, and then weights those exams to produce a final rating. Unlike batting and pitching, there is no conversion to runs. Players who are extremely good or horrible at one facet get capped, the excess either way is ignored.

People would not accept that on offense (Barry Bonds has maxed his HR's at 40, and his BB's at 100, no further credit is allowed. Any OPS+ below 50 is set to 50; Bill Bergen is overjoyed.)

Why do we accept that on fielding?

On the FRAA I posted above:

These are fielding runs above average in WARP1. No all-time context adjustment or competition adjustment included.

When there's a general discussion of "this guy was great, good, average, indifferent, horrible with the glove," this is the relevant measure from WARP to use because it shows how much above or below average that player was at his position, compared to his contemporaries.

It doesn't get into the adjustment to an all-time context that goes with WARP3.

Likewise, it doesn't get into the adjustments for historical context that are the flip side of the WARP3 adjustment. Because WARP treats the distribution of fielding and pitching value as highly variable over time and accounts for that difference in the distance between average and replacement level, to compare players from different eras by WARP1 FRAR will give you a reading of value, but it isn't a useful, cross-period eyeball measure of quality.

When I use WARP actually to rank Larry Doyle against his contemporaries, I use FRAR as part of the comprehensive WARP1 stat. But when I wan to see if Larry Doyle was as good a defensive second baseman as Joe Gordon or Joe Morgan, I look at FRAA, WARP1. It is the best WARP equivalent of a win-shares fielding letter grade or career fielding win-share rate.

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

I am in substantial agreement with jimd's comments above about the problems with the win shares fielding assessment.

I would posit, however, that one reason to accept the win-share approach to fielding while also recognizing that other alternatives may be as good or better, is that, given the uncertainty about fielding evalution, there is good reason to exercise caution in assigning fielding value by avoiding extremes.

I am increasingly convinced that voters would do well to look carefully at both WARP and win shares when ranking candidates, using careful examination of the context to decide which system offers us a more accurate view in cases where the two systems disagree, and making adjustments for identifiable systemic flaws.

Don't they have FRAR, FRAA, and BRAA in a player's era AND in an all-time context? I thought that david had posted those as well.

However, I don't believe that Davenport's translations into an al-time context are worthless, they need to be taken with a grain of salt, but they can be informative. There are statistical reasons for the different era adjustments, even if the are balanced with subjective beliefs as well.

Didn't realize there was another page...

Jim,

Are you suggesting that WS hands out its grades and then figures out WS? It is the opposite of that I believe. Could you please clarify what you meant above?

jimd wrote:

Win Shares hands out grades on a number of exams, each on a different facet of defensive play at the position, and then weights those exams to produce a final rating.

jschmeagol wrote:

Are you suggesting that WS hands out its grades and then figures out WS? It is the opposite of that I believe. Could you please clarify what you meant above?

jimd can speak for himself, of course, but let me point out quickly that you are confusing the "Letter Grades" that James provides on pp. 137-154 of _Win Shares_ with jimd's characterization of the method James uses to assign fielding win shares as "hand[ing] out grades," a phrase jimd uses to contrast the win shares system of fielding analysis to the WARP method for assigning its "fielding runs." Jimd is pointing out that the win share fielding system never attempts to figure out the actual run value of the actual fielding events that are recorded in fielding statistics. Instead, it compares a player's performance to league average on a variety of measures (which jimd calls "exams") to generate a number that describes, percentage-wise, how good the player's performance was. This percentage (which jimd calls a "grade"), allows the player (or more accurately, the player-position on a team) to claim a proportional amount (weighted by position) of the team's fielding win shares.

James' system is essentially saying, "It is either too hard or not possible to figure out the run value of fielding events recorded in our available statistics. However, we can measure fielding quality in certain ways, and use our assessment of that quality to come up with a decent estimate of that player's share of the team's defensive value," based on our estimate of the relative importance of each defensive position.

The WARP system, contrariwise, tries to figure out the run value of fielding events, tying its conclusions about fielding more directly to actual runs allowed.

Just my personal opinion -- though I'm aware of the faults of both systems, I've come to trust the WS fielding ratings ahead of those based on WARP. Too many times with the WARP fielding ratings I wind up scratching my head and wondering how their rating could be so different from the player's defensive reputation. I realize that the WS system is pretty ad hoc, but James also seems to have really good intuition about these things.

Chris, I stand corrected on Doyle's FRAA. Sorry I jumped on you, not as sorry I jumped on WARP. I am so frustrated with WARP I can hardly stand it, 1) because they're always changing and 2) because the all-time adjustments in WARP3 are inscrutable, at least to me, which may be my loss, but if so I can live with it.

I agree with Chris Cobb re Gordon and Childs being better than Doyle with fielding taken into account.

But I've voted for Doyle several times in the past, too. He's probably in my top 25, one of the guys who will rate another look as soon as next 'year.'
I don't necessarily accept that Doyle's fielding was THAT bad, in part because most guys who could hit as well as he did but are lousy fielders at their current spot tend to try other places. Yet Doyle never played a single game anywhere else.

Some may suspect that Doyle would do better in the voting if he had a lower OPS+ and a higher fielding grade, even if the 'trade' wouldn't actually have made him a more meritorious player. That's what I suspect, anyway, though he'd still rate behind Gordon and Childs in my book.

jimd can speak for himself, of course,

Chris did a very good job explaining what I was trying to get at.

I used a metaphor that I assumed most of us were familiar with, multiple exams weighted to produce the final grade for a course. I kinda forgot that James also produces a final grade for the career, which means that my metaphor for the annual marking system and his metaphor for the career rate could get mixed up, producing confusion all around. Sorry about that.

I got it now.

Since WARP fielding has come up here, I'm going to pose a set of questions I've been meaning to ask jimd for a while that I think could use some general discussion. I'll crosspost it to the WARP v Win Shares thread, too. It concerns WARP1's calculation of fielding runs above replacement level in relation to fielding runs above average by position. The question is: how should one attempt to reconcile WARP's representation of period-specific fielding spectra with the evidence of the shifting offensive production at defensive positions and/or with other views of period-specific fielding spectra (win shares) that receive fairly wide acceptance?

Background:

WARP generates FRAR by taking FRAA and adding an amount of FRAR per game that is fixed for each position. This amount changes over time to reflect 1) shifts in defensive responsibility between pitchers and fielders and 2) shifts in defensive responsibility between positions.

For example, in 1895, an average defensive player at each position (FRAA = 0) would receive something very near to the following FRAR for a full season (132 games):

P - 9 (obviously no pitcher would play 132 games, but this shows the fielding importance of the position relative to other positions)
C - 44
1B - 16
2B - 43
3B - 32
SS - 47
LF - 30
CF - 31
RF - 14

In 1965 the FRAR for an average defensive player at each position for a 162 game season would be very near to these amounts

P - 8
C - 29
1B - 13
2B - 33
3B -23
SS - 34
LF - 17
CF - 25
RF - 15

This readiness to shift fielding value around is one of WARP's potential points of superiority to win shares, which sticks to a constant set of "inherent weights" to distribute fielding value among the positions. The one change James acknowledges in the defensive spectrum involves the 2B and 3B, which he sees as switching places on the defensive spectrum. The "inherent weights" in the fielding win share system are

C - 19%
1B - 6%
2B - 16%
3B - 12%
SS - 18%
OF - 29% (James treats OF as one position and then uses the distribution of win shares among individual players to sort out the relative value of each outfield position, but we can estimate that CF will typically land between 2B and 3B and that LF and RF will fall between 3B and 1B).

Before 1920, the weights for 2B and 3B are reversed.

WARP shows us a shifting defensive spectrum over the history of the game, where win shares does not.

However, jimd's study of average OPS+ by position also suggests a shifting defensive spectrum over the history of the game, one in which the shifts are rather different from the ones WARP presents. I'll reproduce his famous table once again:

Decade 1B LF RF CF 3B 2B Ca SS Pit
1870's +1 +4 -1 +4 +2 +2 +0 +1 -13
1880's 13 +6 +1 +5 +1 -1 -7 -2 -17
1890's +6 +9 +7 +7 +0 -2 -6 -2 -22
1900's +6 10 +9 +8 +0 +2 -9 -1 -29
1910's +6 +7 +9 10 +1 +1 -7 -4 -31
1920's +9 10 10 +8 -3 +1 -4 -7 -32
1930's 13 +8 10 +5 -1 -3 -3 -4 -36
1940's +8 11 +9 +7 +2 -3 -4 -4 -37
1950's +9 10 +7 +7 +4 -3 -1 -5 -40
1960's 11 +9 11 +7 +4 -5 -3 -6 -46
1970's 10 +8 +8 +5 +3 -5 -2 -11-45
1980's +8 +6 +6 +2 +3 -4 -4 -8 -48
1990's +9 +4 +6 +1 +1 -3 -4 -7 -50
Mean.. +9 +8 +7 +6 +1 -2 -4 -5 -36

The premise here is that the defensive importance of a position is suggested by the amount of offense the management is willing to give up at a position in order to play a competent defender there.

Avoiding, for the moment, any question of the overall weight given to fielding value in any system, let me line up the defensive spectrum for the 1890s and the 1960s as represented by WARP, WS, and the OPS+ study. (Pitchers will be left out.)

1890s
W1 -- SS C 2B 3B CF LF 1B RF (Top three spots are is 3+ times more valuable than bottom spot, 3B is twice as valuable as 1B)
WS -- C SS 3B CF 2B LF/RF 1B (Top 3 spots are 2 2/3+ times more valuable than bottom spot, 2B is twice as valuable as 1B)
OPS+ -- C SS/2B 3B 1B CF/RF LF (Top 3 spots have OPS+ below avg., 3B is 0, 1B-LF +6 to +10)

1960s
W1 -- SS 2B C CF 3B LF RF 1B (Top three spots are 2.2+ times more valuable than bottom spot, 3B & CF are almost twice as valuable as 1B)
WS -- C SS 2B CF 3B LF/RF 1B (Top 3 spots are 2 2/3+ times more valuable than bottom spot, 3B is twice as valuable as 1B)
OPS+ -- SS 2B C 3B CF LF RF/1B (Top 3 spots have OPS+ below avg., 3B is +4, RF/!B +11)

Parallel representations of the proposed defensive spectrums for other decades would show different discrepancies. One I am especially concerned about right now is 2B/3B pre 1930. OPS+ has these two positions always close in average offense, and they shift back and forth as to which is slightly higher or lower decade to decade, but WARP _always_ gives second base more defensive value. The treatment of pre-1930 first base is also a fraught issue, as is the relative importance of infield vs. outfield positions.

The big questions:

Where there are disagreements in these lists, which assessment should one accept, and why?

If one wants to use WARP or win shares, but trusts the OPS+ assessment more, how might one adjust the results of these systems?

If WARP's calculation of FRAA is run-based, is their estimate of FRAR also run-based, or is it offense-based (like the OPS+ study), or theoretical (like win shares)? If it is offense-based, what measure does it use and how does it get from offensive value to defensive value? If it is theoretical, what is the theory?

And have I ever mentioned about WARP thinking that Barry Bonds's total value is around 17-20% in his fielding above replacement?...As a career left fielder?

I know, that's just one of those pesky, silly Chaleekoisms.... ; )

Just because it's not intuitive doesn't mean it's wrong, though it doesn't mean it's right, either.
What's the evidence that this is an incorrect assessment?

Also, notice that Barry's playing profile has changed a lot.

1986-1989 40% of value is defense (includes a year and a half in center field)
2001-2004 6% of value is defense

For career, it's 17% at present.

Ralph Kiner's % of value as defense is 15%, so WARP doesn't see a huge difference between a great-hit, great-field left-fielder and a great-hit, poor-field left fielder.

What is the problem with its setting of replacement level where it does?

6% of value at defense seems much closer to realistic than does 40%, let alone 17%.

Look at it this way. Let's say that Bonds creates 2500 runs, of which 1650 or so are above replacement. Assuming all things to be equal/relative, and considering the run values in Jim Furtado's extrapolated runs, about 680 of his BRAR come from homers (hr = 1.44 runs/hr; 1650 BRAR/2500 EQR*713 HR*1.44 R/HR, rounding to the nearest ten). What's the implication for his fielding? Let's pretend that we could express Bonds's defense the same way we just did with his HR. He has 365 FRAR, so...



So if you'll forgive the conceit between homers and defense, WARP is suggesting that Bonds' created 150 homers' worth of FRAR PLUS another 210 FRAR. I don't see it, personally. It seems to steep to me. Left fielders aren't that important defensively. If he was a SS I might feel differently, but he's not.

How about looking by FRAA?



That's somewhat more like where I'd think we'd end up: 1 to 1.5 season's worth of homers, instead of 3-5 years' worth of them.

Like I said, this is kind of a little toy, but it can be useful to contextualize a guy's defense by comparing how it compares to the offensive components of his game.

Why would 850 of Bonds RC be belwo replacement? That seems a little high to me.

Why would 850 of Bonds RC be belwo replacement? That seems a little high to me.

As I understand it, that's the result of how WARP calculates BRAR.

On the WARP card, EQR is "equivalent runs," calculated as follows: EQR = 5 * OUT * EQA^2.5.

Batting runs above replacement is the number of EQR above the EQR for an EQA of .230. Since EQA is a normalized stat for which average is set at .260, there is not so very much distance between average and replacement level in the WARP batting evaluation. This means that a lot of runs created fall below replacement level.

I don't know the rationale for setting batting replacement level at a .230 EQA.

It should be possible by analysis of the win share formulas to determine whether the WS 0 point is higher or lower than this, but I don't have time to mess with that much math at the moment. If anybody does, it'd be nice to see how the WS 0 point might be expressed in EQA terms!

Dr. Chaleeko, let me play with your formulas a bit.

Since the ratio of EQR/BRAR for Bonds career is 1679/2527, and you calculate that a home run is worth 1.44 EQR, then each home run = .9567 BRAR

Since 1 BRAR = 1 FRAR in the production of wins, Bonds 365 FRAR = 382 home runs.

What happens if you turn Jim Furtado's extrapolated runs around to studying Bonds' defensive stats in order to find a rough FRAR rate per putout?

Bonds does have 5235 put outs in his career, plus 163 assists. A back-of-the-envelope calculation starting from the other ending, using his career rates in lf and cf suggests that an average outfielder in the same conditions would have 5094 putouts and 152 assists.

So 141 putouts + 11 assists above average = 122 FRAA

Bonds is 365 FRAR, so that would mean that his career totals are 422 putouts and 33 assists above a replacement level outfielder.

Obviously, with 2 variables here we can't rigorously set the value of an average put-out and average assist here, but lets assume that 1 assist = 2 put-outs. Bonds then is 488 putouts above replacement.

Do 488 extra outs created on defense equal 382 home runs added on offense?

What would an application of extrapolated runs here suggest?

These are not rhetorical questions. I have no idea what the answer might be, and I am now quite curious.

Win Shares has Bonds' defense as 8 percent of his total career value.

Here are some others:

Ken Griffey 16%
Raffy Palmeiro 10
Mark McGwire 6

Rickey Henderson 11
Eddie Murray 9
Tim Raines 14

Reggie Jackson 9
Bobby Bonds 14
Bobby Murcer 15

Hank Aaron 11
Willie Mays 16
Frank Robinson 10

Mickey Mantle 10
Stan Musial 11
Duke Snider 16

Ted Williams 8
Dixie Walkeer 16
Joe DiMaggio 16

Mel Ott 10
Lou Gehrig 7
Jimmie Foxx 11

Babe Ruth 6 (plus, pitching 14)
Harry Heilmann 9
Tris Speaker 19

Ty Cobb 11
Joe Jackson 10
Clyde Milan 21

Sam Crawford 11
Fred Clarke 16
Elmer Flick 11

Billy Hamilton 15
Ed Delahanty 12
Hugh Duffy 20

Roger Connor 11
Dan Brouthers 7
Cap Anson 11

Jim O'Rourke 15
Paul Hines 17
Charley Jones 13

Chris,

Dr. Chaleeko, let me play with your formulas a bit...These are not rhetorical questions. I have no idea what the answer might be, and I am now quite curious.

Ditto, let's figure this out.

Since the ratio of EQR/BRAR for Bonds career is 1679/2527, and you calculate that a home run is worth 1.44 EQR,...
Just in the interest of citation, that's Furtado's run estimation, not mine. (See http://www.baseballthinkfactory.org/btf/scholars/furtado/articles/IntroducingXR.htm.)

then each home run = .9567 BRAR
I think I've opened a wormhole here. I referenced Furtado since he's regressed the value of various events over time. EQR surely weights each value differently than XR does. I admit to playing the apples/oranges game here by mixing formulae. So let me back up a second.

Using Furtado's formula on Bonds's current career totals:

This total is NOT adjusted for league, season, or park, which is why it's lower than BP's estimates. They are about 100 runs higher (for all-time and season), or about 5% higher, most of which I think is explained by his playing in PacBell which is a tremendous pitcher's park.

OK, now let's turn those WARP apples into oranges...


The calculated ratio of BRAR/EQR or XRAR/XR is about .66, which means that...

...all else being equal, .66 of Bonds' XR on homers are above replacement. So

.66 XRAR per XR * 1026.7 XR on HR = 677.6 XRAR on HR only. Dividing back out by original multiplier (1.44):

677.6 XRAR on HR
---------------- = 470.6 HR
1.44 XR per HR

So 471 of Bonds's 713 homers are, if you will, HR above replacement.

Since 1 BRAR = 1 FRAR in the production of wins, Bonds 365 FRAR = 382 home runs.
(Chris, we're differing at this juncture, but I think the above calculations explain why. Previously, when I just did it back-of-the-envelope and mixed and matched my systems, I screwed up and led you down a garden path. Let's see if I can get us out it...gulp...)

1 BRAR = 1 FRAR. Bonds has 365 FRAR. I don't think we can simply divide his FRAR by the HR multiplier because that's an apples/oranges comparison. Bonds does not only accumulate runs by his HR. So instead....

First we should make sure that by using XR I'm comparing a similar pct of to value for fielding. So...

Which does simplify believe it or not, and ends up at 350 XFRAR

Which leads us to figure out how much of his fielding could be communicated as HR...


Those 147 runs would then be broken down by the XR HR mulitplier:


So we're still talking about two to three years worth of homers, depending on which phase of Bonds's career we're dealing with.

What happens if you turn Jim Furtado's extrapolated runs around to studying Bonds' defensive stats in order to find a rough FRAR rate per putout?....

So 141 putouts + 11 assists above average = 122 FRAA

Bonds is 365 FRAR, so that would mean that his career totals are 422 putouts and 33 assists above a replacement level outfielder.

What would an application of extrapolated runs here suggest?

There's at least two ways I can see of using XR to look at this question (and I'm not an expert, just a dilletante).

Here's the first way:
An out is an out is an out. All outs are worth .09 in XR. By that method, Bonds' traditional defensive stats would be worth 486 runs total.

Here's another way:
1) An out is an out when it's a put out.
2) The vast, vast majority of Bonds's assists are likely to occur at second base (because he's a left fielder, he's unlikely to throw many runners out who are advancing to third on balls hit by them or in front of them) with a few kills at home. XR says a double's worth .72 runs, so let's call all his baserunner kills .85 since those few home-base kills are very high leverage despite their infrequency.

By this method, Bonds is at 471 runs on the putouts (5235 * .09) and 139 runs on the assists (163 * .85). 610 total XFR.

Using Chris's guess at an average outfielder, we've got 5094 putouts and 152 assists. By method one that's 472 runs for the average LF/cf. By method two, it's 587 runs for an average LF/cf.

Using his guess at a replacement outfielder, we've got 4813 potouts and 130 assists. By method one that's 445 runs, and by method two, 544 runs.

So Bonds, unadjusted for park, league, etc, is ahead of the average fielder at his position by about 15-25 runs per the XR methods I've just toyed with. He's ahead of the replacement fielder by 40-65 runs.

That's not nothing, but it's also not more than a couple months' worth of roundtrippers as compared to the three seasons's worth that WARP suggests. Which isn't to say one is right or the other based on this rinky-dink analysis, but rather that there's a big difference out there in how to look at it. I prefer the more conservative view that Bonds's fielding is a lower portion of his total value (probably around 10%) because even using a tool designed to measure something completely different (offense versus defense), I start to suggestions that even the wildest result would be closer to the conservative position than not.

By the way, I'm a flaming liberal, so you must understand that it pains, pains, pains me to use
the c-word in polite conversation.

Actually, let's take a quick look at the question of putouts and assists. The wildest way you could look at them would be to take method two above and count every putout as if it were going to be a single .

5235 * .50 + 163 * .85: 2618 + 140 = 2758 Runs. Funny, but that's almost exactly how many EQR Bonds has. But the thing is that we know that's totally unreasonable. Most of the balls a fellow catches are trackable, lazy flies. IIRC, 90% of all flyballs are caught. So an exceptional OF is really hauling in LDs. LDs represent around a quarter of all hits (IIRC), and they are the main engine of extra-base hitting. OK, so let's take every putout over replacement and treat it like a double in XR. So for Bonds...

tPO = 5235
rPO = 4813 * .09 = 433
POar = 422 * .72 = 304
========================
PO TOTAL 737
AST = 163 * .85 = 139
========================
BB TOTAL 876 FR

rOF TOTAL 544 FR
=========================
BB FRAR 332 FRAR

Ah hah! Now we've either stumbled onto something interesting (i.e. a confluence between WARP's FRAR and this FRAR), or into a one-time mathematical coincidence. Hoping for the former, you can see that if we bonus each outfielder for his POs above the observed replacement level, we might be on the way to explaining where WARP and linear-weighted run estimators meet for evaluating defense. Or maybe not.

well, it wil tkae me way to long to figure THOSE posts out....

I will say that comparing WS to WARP in terms of "% of value" will be a apples/oranges (or maybe apples/chipmunks or apples/helium) difficulty, because while WS alows no performance below zero, WARP does. Hence, more than 100% of Frank Thomas' value is in his hitting according to WARP. Obviously systems like TPR will reach this conclusion much of the time, since it compares to average. And check out Bill Bergen; over +300 FRAR, but BRAR worse than 200 runs below replacement. WARP describes Thomas abd Bergen's value more closely on a runs basis, but WS is a better tool for determining "% of value".

Like Tom, it's going to take me a little while to digest the good doctor's latest post in response to my questions, but a BIG tip of cap for the inspired investigative number-crunching!!

I am inclined to think that this analysis is finding a way to a confluence between WARP and a linear weights-type analysis, which is only to be expected, since (I think) some version of linear weights underlies WARP, as DOS underlies Windows . . .

DOS underlies Windows?

This explains a lot.

Okay, now for the real tough question: What does all of this have to do with Dobie Moore?

Okay, now for the real tough question: What does all of this have to do with Dobie Moore?

That's easy! It's all about comparing Larry Doyle to Dobie Moore. In order to make that comparison, all we had to do was totally reassess the distribution of fielding value throughout the history of baseball and reanalyze the WARP fielding system. Typical HoM linear thinking :-) .

Comparing 1910s NL fielding to NeL fielding is sort of like an episode of Lost.

Comparing 1910s NL fielding to NeL fielding is sort of like an episode of Lost.

Over on the Heavy Johnson thread, sunnyday2 asked me to read this thread and add some comments, which I am happy to do.

Leaving aside most of the tangential discussions that appear among the posts, this thread is a very thinly-veiled argument between the peak voters and the career voters. If I could sort that one out for you, I'd be a lot smarter than I actually am. So we will inevitably end up comparing him to other players who had short careers, like pre-headache George Sisler, or Chino Smith and Heavy Johnson on the blackball side. Like most of you, I feel one has to give him some reasonable amount of credit for the ball he was playing in the 25th Infantry. But it's a short career nonetheless.

I would start by pointing you to posts 23 and 28 by gadfly, who gives a very good summation of Moore from a historian's point of view. By only quibble, and it's pretty minor, is that gadfly makes the mistake of saying that Chino was 5'7" tall. That's one of those factoids that's been repeated so many times that nobody questions it anymore. Actually, Chino was a 5'11" and very muscular.

But, given Doc Chalek's post 46, we have to have a little talk about Moore's defense. I have stated, in virtually every comment I've made on this site, that what I can contribute is an overview of how Negro Leagues historians view a career; as a sabermetrician, I'm a beginner. So, in that context, let me just make the point that "A" defense for Moore is simply not justified. I can see that the numbers *seem* to justify it, and I certainly respect the erudition of the people who computed those numbers. But I approach the problem from a different perspective.

Basically, it's this: the NeL players who were notable for their defense are almost always consensus picks. At the SS position, there are guys like J.H. Lloyd of course, and Dick Lundy, and also some no-hit glovemen like Jake Stephens and Horacio Martinez (who was elected to 4 E/W games, despite the fact that he couldn't hit a lick). Willie Wells was probably in this group, even with his notoriously weak arm, because he was both quick and a positioning genius.

At the next level are players like Moore, who could absolutely play the position, and who wouldn't have been moved to any other defensive position just to keep their bats in the lineup. So I would think that would be B-grade defense: good but not in the top tier.

Below this level are guys who played the position, but would not have been shortstops at the MLB level. There are no better examples than Monte Irvin and Jackie Robinson, neither of whom got any consideration at all as ML shortstops, despite being shortstops in the NeLs. In fairness, not all of that had to do with their abilities. I think either could have been made into a decent SS in MLB. But MLB managers had an image in their minds of what a SS should look like, and neither Irvin nor Robinson fit that image. And yes, racism may well have been a component in these decisions.

But also, realize that SS at the MLB baseball level is considered very much a elite position. The Yankees, for instance, made sure the Crosetti tutored Rizzuto extensively before they let Phil go out there and make embarrassing mistakes.

The same cannot be said of blackball. Generally speaking, if you could hit, a NeL team would find a place for you to play. That's how the Homestead Grays tolerated an infield that had George Scales at 2B and Jud Wilson at 3B. No white manager would, for an entire constellation of reasons, have done so.

The one tangent I will take up is ("koufax") Fred McNeil's somewhat stormy presence in this thread. The way in which it's relevant is in terms of weighting Moore's performance in the California Winter League. Somewhere, in some thread, someone (oh, good research skills, burnis) talked about how the CWL seemed to be about white teams playing against black teams. Well, it didn't just "seem" to be--that was the whole purpose of the league. Virtually every year, what was essentially a black all-star team competed in games against a variety of hastily thrown-together white outfits, which varied in makeup from a few big-leaguers down to some warm bodies. There is, for instance, a box in McNeil's book in which Walter Johnson (yes, that one) played the outfield and batted cleanup. I don't remember anything quite like that taking place in Washington.

The giveaway is that the white teams rarely played each other; that's why the stats for the black players are huge compared to most of the white players.

Actually, this appeal of black v. white baseball is hardly unique to the CWL. That's how Dean, Feller, Paige and that crew raked in the money in post-season barnstorming. And it also explains why booking agents like Nat Strong in New York didn't see much of a future in black league baseball. After all, Strong thought, where's the drama in seeing blacks play other blacks?

My point is, I may be grateful to Fred McNeil for uncovering this interesting material, but I would put a very low reliability value on the CWL stats.

OK, back to the difficult business of trying to evaluate Dobie Moore.

Grant Johnson, J.H. Lloyd, Dick Lundy and Willie Wells are the great NeL shortstops of their respective eras. How Moore fits in depends on how you want to weigh offensive value versus defensive value at that position. Johnson and Lloyd were outstanding offensive players. Wells and Lundy were good, but not great (Wells' HR stats were completely distorted by the car barn that was 230 feet away in St. Louis; you have to look at his Newark career to judge him correctly as a hitter).

All four of them were superior to Moore defensively--even Wells with his weak arm. If I had to pick out a comparable situation, rather than a comparable player, it would be the question of how you rate George Scales relative to Newt Allen and the other secondbasemen with big reputations among blackball historians. Scales was much more dangerous than Allen at the plate, even though Allen was a pretty good hitter. But that doesn't seem to have moved him past Allen (or even Hughes or DeMoss) in many people's minds. And Moore was clearly superior to two of the four "greats" above, and maybe the equal of Johnson and Lloyd as well. He just wasn't their equal on defense. In a word, that's why he's not in the HOF, although I think that evaluation may need revisiting.

One of the ways in which I feel this site is really superior to the HOF voting is that you guys are capable of taking an issue like Moore's army ball seriously, and making an earnest effort to factor it into the mix. I don't see that happening at the HOF votes at all; Moore's evaluation starts in 1920 for them, and that really cripples his candidacy.

In the Moore/Scales analogy, I would give a substantial edge to Moore, since I believe Moore could have been a SS in MLB, while Scales would have been moved to 1B.

As to white comparables, there simply were none in the 1920s. Overall, I think Cronin is as close as you're going to get. Banks is a bit of a stretch, because Moore hit for a higher average, but didn't uppercut the ball the way Ernie did.

Moore is not in the HOF, although he was considered in the special election of February, 2006. If I had been a HOM voter at the time, he would have been high on my ballot.

One more thought about our friend Dobie (and you guys hardly need to learn this from me): I think the evaluation of artificially-shortened careers is about as hard as anything in this biz ever gets.

The first reason is the reality that there's a huge comfort factor in evaluating long careers. It's so damn easy to look at a stats line for somebody like Musial--who put out essentially the same numbers from 1941 through 1962 and only had one bad year at the end--and know with absolute certainly what his value was. The same for Cobb, Mays, Aaron, and all the others in this category.

And the second is that the manner in which careers get shortened varies so wildly from case to case. Inevitably we're called upon to compare apples with oranges, pears, and kiwis.

And it's not just the obvious cases like Moore, Keller and Hal Trosky. Hell, Bill James really likes Dick Wakefield, pointing out that he was never the same after he returned from his WW2 service. The same as what? The 1.5 years when he was really good?

I'm not saying that James is wrong; in fact just the opposite. He probably makes an excellent point. It's just that it's SO far down the coulda-woulda-shoulda trail.

When you see the percentage of baseball fans (darned close to 100%) who never go beyond the counting stats stage (Bonds' 756th is worth $750K, but his 757th is worth ten bucks), I have to marvel at the intellectual courage of people willing to work hard to figure out the really tough stuff.

As a batter, I think burniswright means,
Moore was clearly superior to two of the four "greats" above [Wells & Lundy], and maybe the equal of Johnson and Lloyd as well. He just wasn't their equal on defense. In a word, that's why he's not in the HOF, although I think that evaluation may need revisiting.

One of the ways in which I feel this site is really superior to the HOF voting is that you guys are capable of taking an issue like Moore's army ball seriously, and making an earnest effort to factor it into the mix. I don't see that happening at the HOF votes at all; Moore's evaluation starts in 1920 for them, and that really cripples his candidacy.

I agree so much, I can't resist pointing out that it undermines the preceding paragraph (partly quoted). Perhaps both points are alone sufficient to sink him in the special election of 2006. I feel sure only about the second one: Moore's evaluation starts in 1920. Note that they didn't elect Johnson. That's all I have time for, although the truth isn't so simple. They did elect Johnson's contemporary Pete Hill and the early icons Frank Grant and Sol White.

I saw this book in the SABR Negro Leagues committee newsletter, was wondering if anyone has read it?

Forgotten Heroes of the Negro Leagues and Latin Baseball by Layton Revel, Ed. D, and Luis Munoz.

Forgotten Heroes is a series of biographies about unheralded stars who played both stateside and in Latin America such as; Perucho Cepeda, Pancho Coimbre, Alejandro Oms, Tetelo Vargas, Alonzo Perry, Al Pinkston, Lazaro Salazar, Wild Bill Wright, Dobie
Moore, and Adolfo Luque.

The 200 plus page book is loaded with vintage photographs and statistics.

Dobie Moore's Real Stats

CORRECTED LINK:

Dobie Moore's Real Stats

Here's my latest MLEs for Dobie Moore. They will be updated as new information becomes available.

That new MLE is 9 more WAR than Ernie Banks ages 20-30.

If you look at Moore's known record, it's frickin' amazing. It's not just the 147 OPS+ from a shortstop, it's also the tremendous DRA. The info that's missing for Moore include his entire pre-NNL career, his 1926 partial season, and his California Winter League performances (where he was quite good, but we don't have complete stat lines nor leaguewide numbers to use). He was also quite durable. About the only thing Moore didn't do well was run. I also haven't reckoned any Gdp info for him, which as a righty is probably slightly negative. Here's what I wrote about him:

Per the Negro Leagues Database, from 1920 through 1925, he was third among all Negro Leaguers in Wins Above Replacement. His batting average was the eleventh highest in the leagues, his slugging percentage fourteenth. Finishing third in games and fourth in plate appearances, he placed second in hits, doubles, and triples to the incomparable Oscar Charleston, not to mention tenth in homers. His 91 fielding runs led all Negro Leaguers at any position during that same stretch and puts him second all-time to keystone man Bingo DeMoss who racked up 101 in 200 more games. If you asked someone in the Negro Leagues who the best player in the game was in the early 1920s, Charleston would have been the answer. But Dobie Moore was second with a bullet.

I hope you'll forgive the pun at the end. ;)

Moore was simply a tremendous player, and he hit the ground running in the NNL as a 24 year old. That said, I use his career average performance for his 1916-1919 and 1926 seasons. The playing time in those years is based on a group of about a dozen comps' early careers, and the PA totals are adjusted to reflect the shorter MLB schedules of 1918 and 1919. This is the same routine I used for Rogan (and will use for Heavy Johnson). Obviously, if/when more info becomes available I'll adjust the projections to accommodate it, but we probably know at least 75% of his overtly professional career, none of which appears to include a decline phase. He's not a substantially different player in the Cal Winter League than we see here, and Riley mentions that in 1926 he got off to a very fast start.

I'm guessing that young Lou Boudreau might be a better comp to look at than Ernie Banks. Shufflefoot stuck at age 22, and from then to his age-30 season, he racked up
5594 PA
203 Rbat
-5 Rbaser
99 Rfield
86 Rpos
384 RAA
40.2 WAA
166 Rrep
550 RAR
56.7 WAR

For his entire career, I'm showing Moore with
5380 PA
239 Rbat
-1 Rbaser
80 Rfield
75 Rpos
412 RAA
43.8 WAA
183 Rrep
595 RAR
64.0 WAR

My MLEs do appear to be tamping down the R/W conversion figure a little. I'll take a look at the R/W conversions this weekend to be sure they are correctly calculating. If they aren't, then Moore's WAA probably wouldn't drop by more than one, but his WAR could drop a little more than that. But in general, Moore looks like a slightly better version of Boudreau. Which is one hell of a player.

Got a chance to look at this before leaving the house tonight.

I took Boudreau's exact WAR components as published on his BBREF page from age 20-30, and I ended up with 40.7 WAA and 59.8 WAR using my formulae. The actual BBREF totals are 40.3 WAA and 57.6 WAR. I suspect that those are both influenced by some rounding issues since BBREF shows whole numbers and no decimals. I do notice that I'm getting different Rpos each year than BBREF, but that's probably because I'm showing Moore at just one position, and Boudreau at least appeared at other positions. My formulae would actually give Boudreau a couple fewer than he actually earned.

More important, I'm getting slightly elevated totals for Rrep. My formula would show 200 Rrep for Boudreau rather than the 175 he actually got. I checked back on the WAR how-to page, and, to be honest, if you go to the replacement runs area, it's kinda cloudy. I was previously using 20.5 runs / 600 PA. BBREF appears to use an empirical to align to reality, but it's not entirely clear upon reading. When I included that empirical, the Rrep total went down to 182, just seven more than actual, and knocked two WAR off of my predicted Boudreau total. Looks like I should add this to all batters, so I'll go back and make that revision to all the currently published MLEs by the end of this weekend if not sooner. You'll know if the edits have been made because I'll note it in the article.

Outside of that, I think that rounding and the lack of multiple positions will account for some of the remaining discrepancies, though I won't rule out discovering something else. Thanks to DL for helping me improve this!

...And I also updated all the MLEs on my site, so you can check 'em out. Not huge changes, but a couple-few WAR here or there.

Dobie Moore bio project by Adam Darowski