As was requested . . .

On my top 500 lists, i have a yes/no column for if a guy is in the HOF.

When I update this year, I think I should give equal billing to the Hall of merit. What would help is a list of all HOM players, preferably with their retroID.

Thanks AROM! I can get you that tonight (unless someone else has it handy sooner?) . . . is RetroID part of the Lahman Database? If so it should be easy. If not, is still shouldn't be too hard . . .

Yes, it is, in the "Master" file. Just VLOOKUP!

That would be a nice addition.

One thing I've noticed is that ChoneWAR seems to really like Sal Bando, at least relative to our previously established consensus.

As a starting reference, here is AROM's article comparing Brian Downing and Jim Rice, in which he discusses some of his methodology.

Here and here are the original pieces describing the TotalZone system.

Re-reading the 2010 discussion last fortnight I re-discovered our first notice of that WAR and its discussion by OCF, DL, DanR, and EricJ. See 2010 Ballot Discussion #163-173 (March). Its published scope was then limited to 1955-2008 which made sense to me; I presumed that the rating method depends on play-by-play data.

OCF let drop that the author is also AROM here at BBTF.
AROM, are you also a'Rally o'Monkey and/or Alex Rodriguez 'o'm?

There is some more discussion whether or how this WAR is appropriate for the HOM project, in the current session of this thread. See 2010 Ballot Discussion #301-306 (November), which really begins on the previous page with epoc's preliminary ballot #286.

I guess I can get to my 1st set of questions:

How does Chone handle defense before 1955?
Is the WAR value schedule adjusted?
Can someone explain the DH replacement value more fully?
Is pitching value a FIP calculation?

I have a table with lahmanID(numeric), playerID, bbrefID, retroID, and numerous variables such as HOM membership and results of the last two HOM elections.

This is a quick and dirty join and select from much more. ... I'll include career Win Shares cWS (Bill James 1876-2001) and WAR (BaseballProjections Top 500 pitching and Top 500 no-pitching).

I haven't kept up with the assignment of IDs. Really I should call the first field [wendtID]. It's the lahmanID (positive numeric) extended by my assignment of zero and negative numbers to other people. In this selection they are nine, five HOM members and four who scored points in the last election.

lahmanID
-304 Carlos Moran
-205 Leroy Matlock
-168 Dick Redding
-157 Alejandro Oms
-156 Dobie Moore
-155 Bill Monroe
-140 Dick Lundy
-136 Grant Johnson
-105 John Beckwith
The 1583 records are sorted by [lahmanID] so these are the first nine records in the table.

Here is the link for your file, which will be available for 7 Days.
http://www.yousendit.com/download/TzY0eUNKTlFrWS9IRGc9PQ (BJCHONE.txt)

Format is csv but Windows extension is .txt, sorry about that.
Now I must run for eight hours. I'll check in again then or tomorrow morning.
Please confirm whether this works.

Thanks Joe, but no rush. I won't be able to do a WAR update for about 2 weeks. There's a ton of things to get together for it.

Next update will be for 2009, but also 1952, replacing the crude methods I used for earlier seasons with the better retrosheet data.

AROM, are you also a'Rally o'Monkey and/or Alex Rodriguez 'o'm?

Stands for Anaheim Rallymonkey of Maryland. A play on the team name change from back when.

How does Chone handle defense before 1955?

JAARF. Just another adjusted range factor. You've seen the type. I don't think it's any better or worse than the Davenport fielding ratings, or defensive win shares, or anything that more advanced than total baseball fielding runs (in other words, based on balls in play not innings). I won't spend a second defending it, it's just a crude estimation until I get better data.

Is the WAR value schedule adjusted?

No

Can someone explain the DH replacement value more fully?

Originally 10 runs worse than 1B (bad fielding 1b = DH) but adjusted to 5 on observation that there is a DH penalty, players hit worse there than when playing a fielding position.

Is pitching value a FIP calculation?

No, context adjusted runs allowed.

And the defensive spectrum going backwards in time?

No corrections for expansion or other standard deviation factors, I take it.

And the defensive spectrum going backwards in time?

Back through the Retrosheet years, I believe those are laid out here. Before that, I'll let Sean take the question.

You really just change the defensive spectrum that abruptly at each decade cutoff? So a 2B is +5 in 1989 but +2 in 1990? That can't be right...(my own work shows a rapid strengthening of 2B from 1975 to 1985, with the position roughly flat before and after that transition period).

I suppose there's no right or wrong answer to this, but I'd just add that I very strongly disagree with giving a "DH bonus." Tango agrees that replacement level for DH's is league average (as I use in my system).

Tango agrees that replacement level for DH's is league average

Citation, if it's not too much trouble? I was under the impression that Tango has always pushed the idea that it the 5 run DH penalty is a real effect.

EDIT: To be more precise, what I take Tango to believe is that replacement level DHs are league average hitters, but that when playing DH they will hit like -.5 win hitters, so someone who hits league average while playing DH is above replacement level.

Can anyone point to a good DH bonus thread? I've been fishing for opinions on this...
FWIW, Tango also gives a DH bonus (post 42) - right?

EDIT: Here's yer Coke.

You guys are way out ahead of me. I never heard of CHONE WAR until a few days ago. It's a new WAR put out by a guy named Chone or who loves Chone Figgins? Sort of like DanR's WAR, I mean in the sense that it's a system that a person other than BP put together. Is that right? And I also gather there's something idiosyncratic about it, well of course there is, but something objectionably idiosyncratic. Is that also right? Oh, and where is it.

Blackadder, I agree with that, and that is precisely how my system handles it.

I don't remember if I stuck with sharp position adjustment changes at decade ends or smoothed out the system. I'll check later.

Oh, and where is it.

Here.

Nope, no smoothing on the position adjustments. Might be worth doing for the update.

sunnyday, it's an all in one value stat put out by AROM that's quite similar to mine (and probably shows a .98 correlation or something). But there are some differences that I hope to explore on this thread.

Tango wrote me in a personal email just today that he uses league average as the rep level for DH's.

Sunnyday, I don't think there is anything objectionably idiosyncratic about it, in the way that, say, Win Shares' low replacement level was. I believe the substantive differences between Dan's and Sean's systems are a) Sean's system does not adjust for league standard deviations (although one could easily do so if one wished); and b) they determine their positional adjustments very differently, Sean using multi-position players and Dan using the observed overall performance of the worst regulars (of course with some complications in both cases); and c) Sean applies a DH bonus while Dan does not. These can add up to big effects; Sean has Edgar Martinez basically tied with Barry Larkin, while Dan sees a vast gap between them. As I see it, issue b) is possibly the largest outstanding unresolved conceptual issue in the sabermetric evaluation of value.

And the defensive spectrum pre-50's?

On the ballots that have been cast so far, there's near-unanimous agreement that Larkin belongs at or very near the top of the ballot; while many voters have mentioned the durability issues, there don't really seem to be many serious or systematic disagreements with respect to Larkin.

But Alomar is different. There is a sharp, systematic divide here, and as near as I can tell, Dan R and AROM (or at least those who have been quoting AROM) lie on opposite sides. The first position (and I think AROM is on this side) is that Alomar's value is very close to (possibly even ahead) of Larkin's and he belongs as one of the top two on the ballot. The second position (and I know Dan R is on this side) holds that, for reasons largely having to to with positional value and standard deviations, Alomar, while perhaps worthy of the lower part of the ballot, does not belong near the top.

Just above, Dan R said, But there are some differences that I hope to explore on this thread.

Alomar would seem to be a likely focus for that exploration.

Hmm, that is strange. I just sent Tom an e-mail asking him to stop by and sort this out.

I have Alomar 6 WARP1 below AROM. That might be enough to distinguish between the top and bottom of the ballot, but it's not enough to be an illustrative case of our methodological discrepancies.

Dan R also DOES adjust for season length. I believe this includes strike seasons.

Does Chone WAR catch the difference in replacement level AL v. NL due to the DH?

He appears to compare hitters to the pitchers-removed league average in both leagues, making that correction unnecessary, whereas I take average to mean quite literally the league average (as it's much easier to calculate :)). It's just an accounting issue.

I looked at his list - Bench 51st, Berra 98th. A catcher bonus is definitely necessary to use Chone's rankings.

Here are the pitchers

I must admit, I kind of like this ranking. Mullane and McCormick look really good. Cone drops down and Tiant moves up.

Top unelected pitchers from Chone

Rick Reuschel
Tony Mullane
<Kevin Brown>
Jim McCormick
Luis Tiant
David Cone
Vic Willis
Mickey Welch
Charlie Buffinton
Larry Jackson
Frank Tanana
Chuck Finley <Even with Jack Clark among hitters>
Silver King
Orel Hershiser
Tommy Bridges (no war credit)
Kevin Appier

Koufax is 61st. Why aren't the "Chone Voters" filling their ballots with pitchers? His system is saying they're more valuable than the hitters they're voting for.

Dan R, correct that pitcher hitting is removed, so the DH difference won't change anything. I adjust the replacement level when one league is stronger than another - AL now, NL in the 50's/60's.

I don't think using standard deviations are going to make a difference with contemporaries like Larkin and Alomar, and as Dan says, they probably are not going to show where a big difference in the systems lie. David Concepcion is probably the one to look at, along with his opponent in the 72 series, Dagoberto Campaneris.

Position adjustments before 1950: Use of the multiposition data (though suspect since the measurement is so crude), combined with some educated guesswork.

Catchers are at +10 for all of baseball history, except from 1920-1950, where I make it +5 as very few steals were attempted from the time Babe Ruth made them seem silly and Maury Wills brought them back.

I have the infielders relatively better than the outfielders as you go back further in time, as the combination of more groundballs, more errors, and bunting put infield at a premium (even first basemen).

I'm pretty conservative in the position adjustments as I don't go beyond +10 or -10 per season, at any point in time.

DL,

I agree with you completely. I'm putting together my consideration set based on the CHONE WAR numbers and do plan to have a relatively pitcher-heavy ballot. (Of course, I have typically leaned toward a HOM with reasonably proportional representation among positions.)

I've only put together numbers for 81 players in my initial consideration set of 126, and haven't yet reached the 19C guys, but Reuschel, Cone and Appier are very likely to make my final ballot. I imagine there will probably be three more pitchers on there.

(Also, 3/5-year peaksters should take a second look at Wilbur Wood.)

Even I wouldn't vote straight from the published totals. For one thing, the assumptions in the model as to replacement level for a pitcher vs that of a hitter could really change the ratings if you thought it should be different. Then there's peak value vs career to consider.

AROM,

Most certainly agree with you. I'm going with a combination of career and peak/prime (best 3/5/7/10 consecutive seasons). That gets the general scope of things, plus, as Joe Dimino so helpfully says, "you don't have to vote your system."

I'm going to take this opportunity to compare and contrast three players that my system and CHONE's strongly disagree on: Sal Bando, Willie Davis, and Dagoberto Campaneris. I hope a close analysis will help illuminate the differences between CHONE's methodology and mine (although it is worth emphasizing that a cursory comparison of our rankings shows that the systems are extremely similar). I have converted his numbers to a comparable format to mine--raw hitting, GIDP, and ROE are subsumed under batting; TZ, ifDP, OFarm, and Catch are subsumed under fielding, the subcategories are converted from runs to wins, and seasons are straight-line adjusted to 162 games. I have also adjusted my BWAA and Replc to use the pitchers-removed average as league average for non-DH years, to mirror CHONE's accounting.

Glossary

SFrac: Percentage of the league average plate appearances per lineup slot for that season, a measure of playing time (greater than 1 for leadoff men, etc.)
RBWAA: Rosenheck batting wins above average.
RBRWAA: Rosenheck baserunning wins above average.
RFWAA: Rosenheck fielding wins above average.
RRepl: Wins above average a replacement player would have accumulated in the same playing time, according to Rosenheck.
RWARP1: Rosenheck total wins above replacement, unadjusted for standard deviations.
LgAdj: Ratio of the 2005 standard deviation to the regression-projected standard deviation of the league-season in question.
RWARP2: Rosenheck total wins above replacement, adjusted for standard deviations.
SBWAA: Smith batting wins above average.
SBRWAA: Smith baserunning wins above average.
SFWAA: Smith fielding wins above average
SRepl: Wins above average a replacement player would have accumulated in the same playing time, according to Smith.
SWARP1: Smith total wins above replacement, unadjusted for standard deviations.
TXBR: Totals, excluding sub-replacement seasons.

Sal Bando



Sean and I clearly see Bando's offense the same way. We have a significant but not enormous disagreement on his fielding---Sean sees him as a slightly above average fielder, me as a slightly below one. (I'm really not sure what to make of this, since SFR thinks Bando was a brilliant 3B and DRA a meaningfully below average one; the truth probably lies somewhere in the middle). And we have a yawning gap on the intrinsic value of playing third base. I see 3B during Bando's prime as a dead mid-spectrum position, no different than it is today, while Sean has it as glove position, not far from where he puts modern SS. (Note that my view and Sean's converge starting in around 1977).


Willie Davis



Same issues as Bando. The offensive evaluation is idential. I see Davis's fielding as good, while Sean sees it as very good. Sean's decade-based cutoff for changing positional weights is very apparent here--he agrees precisely with me about the value of CF from 1970 onwards, but sees it as significantly more valuable in the 1960's (note the increase in SRepl/SFrac from -2.5 in 1969 to -1.75 in 1970--you'd never see that in my system). I also further ding Davis for the extreme ease of domination of the 1969-70 NL, a factor Sean does not take into account.


Dagoberto Campaneris



Once more, we have similar takes on overall value above average, with Sean slightly more kind to Campaneris. (It's worth noting here that SFR has Campaneris as an otherworldly +144 shortstop, and DRA also gives him a superb +109. With that kind of defense and baserunning value, Campaneris's case is far stronger than I make it here, where old BP FRAA and Fielding Win Shares thought he was merely good). It's that replacement level column that is night and day: Sean has SS a mere 2.5 runs per year more valuable in the 70's than it is today (which I think he himself recognizes is a fudge, due to the arbitrary +10 cap he places on positional adjustments...right, Sean? Kind of like Bill James's cap on Fielding WS). By contrast, as everyone in the HoM knows, I see 70's SS as the toughest position to fill at any point in MLB history. If you could even provide league-average offense at the position back then, you were an All-Star in my book. I have offered two potential explanations for this phenomenon: the mega-expansion of the 1960's, which hurt SS more than (say) LF/RF due to the intrinsic scarcity of the position, and the advent of turf ballparks.

As a reminder, I derive my positional weights by starting with Nate Silver's findings on the aggregate performance of Freely Available Talent (MLB players over age 27 earning less than twice the minimum salary), and adjusting them over time based on the performance of worst 3/8 of everyday players at each position. Sean gets his by studying the fielding of position-switchers over ten-year periods (e.g. 1970-1980).

The question I answered Dan is not exactly the question being asked here.

I do a valuation based on a multi-step process. The first step is to compare everyone to the same league average for hitting, and everyone to the positional average for fielding.

Then I apply a positional adjustment, which is this in wins:
+1.25 C
+0.75 SS
+0.25 2b, 3b, cf
-0.75 lf, rf
-1.25 1b
-2.25 dh

I reason that I set the DH at 1 win below the 1B, which is about what a bad fielding 1B would be.

Then I add replacement level (per 700 PA or per 162 G):
+2.25

(You can merge the last two steps into 1 if you like, so the C gets a pos+repl adjustment of +3.5 wins and the DH gets a 0 adjustment. I like to keep things separate... if you compare, you'll get numbers that are similar to Dan R.)

Finally, I give a +0.5 wins bonus per 700 PA for DH stats, because it's harder to DH, as per Andy's research in The Book.

***

Also note that in the THT 2009 annual, I found that catchers need at least a +0.5 win, if not +1.0 win bonus because it's harder to hit as a catcher. At the same time, perhaps my +1.25 positional adjustment is to aggressive, and maybe it should have been +1.00 to begin with.

So, you can make the original chart as +1.00 wins for the C, and then a +0.50 win bonus for catchers who hit.

I haven't decided yet.

Tango, what do you make of Nate's finding that Freely Available players in the DH-role do hit at the MLB average, not 5 runs below it?

If that's the same article where he did his other work, then I've already written that that process has selection bias. I think what he did there was a decent first step, but it's not the end product. If the presumption is that a player is a "SS" and therfore, that's his "position", then that's the wrong approach. A player, from the time he enters pro ball, is classified as a C, 2b/ss/3b, of, or 1B. That's what the pools are. So, when Nate looks for 27-yr old SS, well, what about the 27-yr old 2B/3B who WERE SS through his 20s? That guy falls out of his SS pool.

Anyway, his study has big problems like that, that run counter to the way I do things.

***

That said, it's possible that I should set my positional adjustment to -2.50 or -2.75 for DH, not -2.25.

At the moment, I'm generally comfortable with what I have, but, as you can see by my obvious gaps of .50 wins, I have a general error allowance, and I won't argue anything that is 0.25 wins or less. If it's at .50 wins or more, I would.

which I think he himself recognizes is a fudge, due to the arbitrary +10 cap he places on positional adjustments...right, Sean?

It wasn't arbitrary, at least for 1950's to the present. That is where the data led me. You can say it's arbitrary for pre 1950 as I was on shakier ground with the data, and didn't want to go beyond the limits of where the post 1950 position adjustments were.

For Campaneris, we're both rating him as an average player (within 5 wins) looking at batting, baserunning, and fielding above position average. His WAR/WARP1 rating is almost exactly what we give him for replacement + position adjustment.

You may be right about shortstop being dramatically tougher in the 1970's, but I'm not convinced. If so I would expect to see much greater differences - shortstops who played 2B were only 7.6 runs better, and those who played 3rd were actually worse as 3B. http://www.fangraphs.com/blogs/index.php?author=12

With such a great position adjustment, if you had an average hitting and average fielding 3rd baseman, you would be better off as long as he was less than a -25 defender at short. Now I'm curious, can I find players who played a full season at shortstop in the 1970's, after playing some minimum # of games at third base previously in their careers? And if so how did they do at both spots? I'd exclude players who played 3b after shortstop, because of aging. It might take some time to explore that though.

I looked at his list - Bench 51st, Berra 98th. A catcher bonus is definitely necessary to use Chone's rankings.

Here's an off-the wall question about catcher bonuses. If Joe Mauer never plays another game, does he have a shot at your ballot?

To me that doesn't have to do with catcher bonuses, but about peak value. I'd say Joe has a shot, but I won't commit to him if he never played a game. Does Nomar through 2003 get on your ballot? Does he now?

OK, here are some examples. I'm looking at anybody who played at least 200 3B games in their careers and had a year with 100 games at short some time in the 70's, if you want to replicate the query.

Bill Almon played short in 77, with a -5 TZ. He went to 3rd in 78 (only age 25, so don't blame aging) and had a -8 rating.

Dave Chalk played 3B in 75, at +9. He went to short in 76 and was +2.

Toby Harrah played short early in his career, and later went to third. For his career he was -7 per season 3B, and -5 per season SS. But as I said, I want to stay away from a guy who moved to 3rd as he got older. I find Harrah split time between the spots in 1973, and again in 1978.

1973: +3 ss, -7 3b. 1978: 0 3b, +5 SS.

Rico Petrocelli: Another who went from SS to 3B, so I must be careful with the data. I'm only looking at the year of the switch, 1970-1971, when he was at peak age (27-28). 1970: +2SS 1971: +2 3B

Dennis Menke: played some 3b and ss pretty much every season. Career: +1 3B, -4 SS.

Greg Pryor: Another utility guy. Career +6 3b, +5 SS

And finally, Maury Wills. Perhaps my best example. After his early years playing short for the Dodgers, Wills moved to third in 1967-1968. He was an average 3B, +3 67, -2 68.

In 1969, he was 36 years old, and they moved him the wrong way on the defensive spectrum, back to short. If the gap between the value of 3B/SS is so large, and he's getting old, shouldn't we expect him to be an unmitigated disaster at short?

Didn't happen. He played a full year at +2, then average in 1970, and +2 again at age 38 in 1971.

I realize the samples are already very small, but how many of those players were in turf home parks? The proliferation of turf is one of the major theories that's been advanced to explain the low SS replacement level Dan finds during that period.

AROM, what's your explanation, then, for the careers of Alfredo Griffin, Don Kessinger, Ed Brinkman, Tim Foli, Rafael Ramírez, and the like? Pure stupid groupthink by teams for 20 years? I agree that's part of it--I certainly think teams lacking SS would have done better to move a 2B/3B and take the defensive hit rather than call up a hitless wonder from the minors--but that doesn't explain why that was only the case in this one historical time period.

Ah, the people I took to calling "Rafaels" in honor of Rafael Ramirez, Rafael Santana, and Rafael Belliard. But I always thought of Alfredo Griffin as the type specimen.

Could, perhaps, some of what you're talking about be the inadequacy of offensive measurements? That the same thoughts you heard from writers and broadcasters also infected those who made decisions, and people actually thought that Alfredo Griffin was a good offensive player? Hey, he hits .280-.290, he's fast, he's exciting on the basepaths, he's an "igniter." We know he doesn't hit for power, but that's for outfielder McSlug to take care of. (Never mind that his power, or lack of it, does too matter, that his BA in is best years shouldn't be cited as "what he hits", that he may not actually be a good baserunner, and the fact that he never walks really, really matters.)

Sean, I think the problem with your method is it assumes that there is no intrinsic difference between playing shortstop and third; that shortstop is just a "harder" version of third but the same skills apply to both positions.

I think that might be a reasonable approximation in the current game (though not perfect; David Eckstein at his peak was an fine SS but he would have been a terrible 3B).

But back in the 70's? I think it breaks down.

When you look at position switchers, you have a biased sample; its the guys whose managers determined, based upon scouting/spring training observations, were capable of handling 70's-era SS, with all the attendant demands. Dustin Ackley is probably going to be a fine 2B but that's not because all CF could be 2B, it's because he's being moved because scouting indicates he can handle the position. Similarly for Biggio- not all catchers could handle 2B.

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

Extended discussion below:

One of the interesting things that I found when Dan was first developing his WARP was that the changes he was observing in offensive performance correlated with changes in the average size of players at the position over time. Thus, SS didn't just get more offense from 1975-1990- the players got larger too. There's a strong correlation between player height and player offense, so when you start to put bigger guys at the position, the offense increases. Its been oft observed that many of the great offensive SS of the current era are much taller than all but a couple of 1970's SS.

So if there was something intrinsic about playing SS in the turf-era that favored, even required a shorter SS, you would expect to see the propensity of shorter players and lower offense that you do, in fact, observe. You'd also expect that certain short guys could move between 3B and SS pretty much interchangably; you'd also expect that you'd have taller 3B who could not be moved to SS. (Nettles and Robinson would be great examples of that; it wouldn't surprise me if they'd be SS for their primes if they came up 30 years later.)

So if you see a player that can handle the switch, it doesn't tell you anything about what the guys who -couldnt- handle the switch would be like at the second position, even if they had the same defensive stats at the primary position. So you can't extrapolate from Maury Wills to Greg Nettles, or from Mark Belanger to Maury Wills, etc.

That's a *great* point, 'zop.

(Nettles and Robinson would be great examples of that; it wouldn't surprise me if they'd be SS for their primes if they came up 30 years later.)

In 1972, the Phillies had Larry Bowa playing SS and Don Money playing 3B. (And yes, Money was quite a bit bigger than Bowa.) They also had a cup-of-coffee prospect named Michael Jack Schmidt. Both Money and Bowa were still in their mid-20's. Money had just had two bad years in a row, but he'd been good before then and as it turned out would be good again. During the off-season, the Phillies kept Bowa and traded away Money, clearing 3B as a space for Schmidt.

The alternate-universe version: what if they'd kept Money, traded Bowa, and made Schmidt a SS?

Nice post, 'zop.

So if you see a player that can handle the switch, it doesn't tell you anything about what the guys who -couldnt- handle the switch would be like at the second position, even if they had the same defensive stats at the primary position. So you can't extrapolate from Maury Wills to Greg Nettles, or from Mark Belanger to Maury Wills, etc.

Maybe true, but I expect that if shortstop was a drastically more difficult position to fill in the 1970's, then we should see greater differences in players at 2b/ss and 3b/ss instead of the same pattern we see for modern players.

...and a good riposte. That's a giant unproven assumption of yours, Sean, but such postulates are probably unavoidable. My positional weights are based on an equally large and unsupported assumption: that the gap in wins between the worst 3/8 of MLB regulars at a given position and the true Freely Available Talent level remains fixed over time. There is, in fact, absolutely no reason why that should be the case. But I've managed to bring a bunch of HoM voters along for the ride. :)

Jay Bell and Buddy Bell (no relation) were about the same size.

I have heard several people comment on how Cal Ripken was the guy that allowed GMs to believe tall people could play SS. If the best SS talent in baseball was stuck playing 3B then it does make sense that Chone likes Buddy Bell and Ron Cey instead of Bert Campaneris and Dave Concepcion.

shortstop was a drastically more difficult position to fill in the 1970's, then we should see greater differences in players at 2b/ss and 3b/ss

Is this necessarily true? If there was a shortage of talent at 2B/3B/SS it makes complete sense that 2B and 3B get filled and the drought only applies to SS, the hardest position to defend. Yes, somebody should have traded for another team's 3B and stuck him at SS on his team, and one team was willing to shuffle Toby Harrah back and forth. This would mean the drought in "infield replacement talent" gets shared a little across all three positions.

Maybe true, but I expect that if shortstop was a drastically more difficult position to fill in the 1970's, then we should see greater differences in players at 2b/ss and 3b/ss instead of the same pattern we see for modern players.

Maybe, but I'm not persuaded. Here's what I mean.

Assume, arguendo, that SS in the 1970s requires some sort of unusual physical attribute that few players have. (Indeed, it does seem like there are SS of this short, acrobatic variety that would be physically out of place at any other position.) If that's the case, SS in the 70's moves "off" the defensive spectrum--what it really becomes akin to is catcher now.

Catcher is a good illustration of what I mean. Catcher is not "harder" per se than SS or 2B even in the modern game; if you look at average size and speed, you get some big lumbering guys who play catcher. Rather, catcher is detached from the defensive specturm- the skills required to play it, (1) super physicial durability, (2)very high arm strength, (3)"the catcher build", don't really translate to the other positions.

Nevertheless, you see really low offense from catcher. Why? Because the physical demands restrict the pool of possible players so very tightly.

But the important point is this: the offense at catcher doesn't vary up and down based upon its place on the defensive "spectrum". Its not on that spectrum, because it requires specialized skills. Offense at catcher is separated from the variation of the rest of the positions, and varies with the supply of catchers, the fertility of Molinas, and random chance.

Now, Sean comes in to analyze position strength. He sees Victor Martinez switching from C to 1B- he does pretty good. He sees Brandon Inge switching from C to 3B, and he's freakin' incredible. You see Eli Marrero getting starts in CF. And you conclude, "hey, this catching thing isn't so bad".

That's not how it works. In the 70's, you have a bunch of pixies playing SS in a game filled with huge, strong men. I think you could argue that SS had partially detached from the rest of the defensive spectrum (because no team was going to put a big guy at SS no matter how good he was with the glove). In those conditions, positions switching as proxy fails.

Not a good analogy. I don't use this method for catchers, because the skills to play the position are very different. A Molina is too slow to be a good defender at any other position except maybe 1st base. Elvis Andrus's or Endy Chavez's defensive skills are wasted at catcher, and he doesn't have the experience at the position. (Ichiro could be a good defensive catcher though, since Ichiro can do anything).

At third base, second base, and short, you have for the most part the same skills needed, with some variation: move laterally, field a ground ball, throw accurately to first. There have always been utility infielders who move between the three positions. Almost all of these guys were shortstops as amateurs and minor leaguers.

Assume, arguendo, that SS in the 1970s requires some sort of unusual physical attribute that few players have. (Indeed, it does seem like there are SS of this short, acrobatic variety that would be physically out of place at any other position.) If that's the case, SS in the 70's moves "off" the defensive spectrum--what it really becomes akin to is catcher now.

If that's the case, then when these big lumbering guys play shortstop (and Schmidt, Bell, Aurelio Rodriguez played a bit there.) then they should be terrible, if we postulate that shortstops of the time had some special skill and these guys, being career 3B, didn't have it.

I can't answer why teams would play guys like Oyler, Kessinger, Griffin, Foli, any more than I can answer why a team in that time period would play Enos Cabell at first base. As for people believing things that are flat out wrong, this is pretty minor compared to some of the ideas leading to our current economic crisis.

Toby Harrah is a pretty obvious example of teams doing things wrong. He had enough time at multiple positions to show he wasn't any better at third than he was at short. So if he's going to cost you 10 runs a year with the glove either way, why play him at a position with average hitters instead of the position filled with -25 hitters?

Toby Harrah is a pretty obvious example of teams doing things wrong. He had enough time at multiple positions to show he wasn't any better at third than he was at short. So if he's going to cost you 10 runs a year with the glove either way, why play him at a position with average hitters instead of the position filled with -25 hitters?

Harrah's case is interesting. He was the Rangers' regular SS in 1976 with Roy Howell (OPS+ of 92) at 3B. Howell became a Blue Jay, so they moved Harrah to 3B and brought in Bert Campaneris, who put up an OPS+ of 78 (or maybe they brought in Campy so they moved Howell, I don't know). Looks like a bad trade, right? Except the 1977 Texas Rangers won 18 more games than the 1976 version. The team ERA+ rose from 104 to 115. Now, they changed a bunch of other things: at 2B, they replaced Lenny Randle (OPS+ 63, most famous for fighting with his manager) with Bump Wills (OPS+ 110) and they changed managers mid-season in 1977, replacing Frank Lucchesi (31-31 in '77) with Billy Hunter (60-33 in '77).

But superficially, moving Harrah from SS to 3B and replacing him with a "real" shortstop looks to be an inspired move.

At third base, second base, and short, you have for the most part the same skills needed, with some variation: move laterally, field a ground ball, throw accurately to first.

You haven't proven this. In particular, MIF have to be shorter than 3B in order to turn the double play. The pool of players who are positions switcher and utility players are, by definition, the ones whose body types and skills are suited for multiple positions.

You have a biased sample. It's just that simple. Unless a random assortment of players switched positions then you can't make the extrapolation that you do.

The catcher analogy is an extreme one, for illustration's sake.


EDIT: One more point: my critique doesn't mean your method is wholly invalid. It just means that you're going to systematically underestimate the difference in difficulty of the positions, because the "best suited to move" guys are always going to be disproportionately represented in your sample. The relative strength through time should be OK, but the magnitude will be attenuated.

35. AROM Posted: November 24, 2009 at 06:57 PM (#3396147)
Even I wouldn't vote straight from the published totals. For one thing, the assumptions in the model as to replacement level for a pitcher vs that of a hitter could really change the ratings if you thought it should be different.

For another thing, the published ratings are distinct for Top 500 "pitchers" and for Top 500 "position players". --not merely distinct lists of 500 but distinct ratings, like Pete Palmer's TPI and TPR for "pitchers" and "players".

The scales may be the same, so that you have one rating of baseball players by simple addition, like Pete Palmer in the print Total Baseball, and recently Palmer/Gillette Baseball Encyclopedia. Even so, the published five hundreds are sufficient to do that only for a couple of players like John Montgomery Ward.


--
A couple of times this fortnight (probably re Edgar Martinez) I have alluded to the disagreement about DH batting which Blackadder has tried to clarify by inviting Tom Tango here. Here's the reference. It may be Tom Tango's introduction to Dan Rosenheck in absentia, by Blackadder.
The Book Blog: Tango: Edgar (mainly #35-46)
That thread is 15 months old. I visited this fortnight thanks to Dan Greenia.

In particular, MIF have to be shorter than 3B in order to turn the double play.

I don't know that they have to be "shorter", but turning the DP is definitely an additional skill set required up the middle that's not required at 3B. Also 3Bs need significant lateral movement only to their left.

The pool of players who are positions switcher and utility players are, by definition, the ones whose body types and skills are suited for multiple positions.

Here, I'm inclined to agree, and I think it does bias the results in the way the Dan R and others are suggesting.

-- MWE

The pool of players who are positions switcher and utility players are, by definition, the ones whose body types and skills are suited for multiple positions.

I have no doubt that this is true, but it's a big leap to go from there to a 2-win difference between shortstop and third base.

Rick Reuschel
Tony Mullane
<Kevin Brown>
Jim McCormick
Luis Tiant
David Cone
Vic Willis
Mickey Welch
Charlie Buffinton
Larry Jackson
Frank Tanana
Chuck Finley <Even with Jack Clark among hitters>
Silver King
Orel Hershiser
Tommy Bridges (no war credit)
Kevin Appier

Koufax is 61st. Why aren't the "Chone Voters" filling their ballots with pitchers? His system is saying they're more valuable than the hitters they're voting for.

As a peak voter of course I kind of prefer Koufax to Reuschel.

My main point here however is just to say that Tony Mullane also deserves a discount of some sort for throwing in the AA, which for much of it's history is probably analogous to the NgL in terms of relative quality, or the equivalent of the AL or NL during their relatively weakest periods respectively.

The big difference between the listed systems is really where replacement is set at. I don't have a good handle on where DanR's WARP sets it at, but for the other systems the expected win percentage of a team of replacement-level players is about:

Baseball Prospectus WARP: .230
Fangraphs WAR: .290
Baseball Projection WAR: .340
Win Shares: .110 (and then values are inflated to allocate absolute wins based on marginal wins above the .110 baseline)

All of that is from memory, so please don't hold my feet to the fire over any of those numbers in particular, just pay attention to the relative difference between the systems.

Now what's funny is (with the exception of Win Shares) the baseline for position players is basically the same between all three systems. Yes, there may be some modest disagreements here and there about things like positional adjustments, but the baseline for rep-level hitting is essentially the same.

The implications for pitcher replacement level, and comparing hitters to pitchers, are pretty drastic, I think.

I'm not understanding the problem with comparing the two lists side by side. Is it that pitcher hitting isn't included in the pitching list?

Differences in replacement value don't make sense. Sure a hitter's replacement value will be different than a pitcher's rep but a 3B already has a different rep than a RF. Both lists are denominated in WINS over REPLACEMENT. I assume the runs created / runs saved to equal a win are on the same scale. In that case a pitching WAR should be the same as a batting WAR.

In particular, MIF have to be shorter than 3B in order to turn the double play.

You haven't proven this.

You have a biased sample. It's just that simple. Unless a random assortment of players switched positions then you can't make the extrapolation that you do.

Yes, the sample is biased. But if we find a 5 run difference in this sample now, and believe that 5 runs is appropriate for the difference between 3rd and short, and then I use the exact same methods to find there was a 7 run difference in this biased sample 30 years ago, I cannot buy that the real difference is 25 runs. The evidence for this is that shortstops were 25 runs worse as hitters back then, and that doesn't convince me. I'd be interested in seeing evidence, on the defensive side, that justifies giving a player such a huge boost for playing 30 feet to the left.

But superficially, moving Harrah from SS to 3B and replacing him with a "real" shortstop looks to be an inspired move.

In that specific case, it just shows that Campaneris was a much better player than Howell. My system rates his 1977 as a 28 run defensive improvement over Howell's 1976 (assuming Harrah is a constant at either spot.) And even with the lower position adjustment I use, Campy comes in at 3.3 WAR and Howell -0.3. He was one of the first major free agents.

I assume the runs created / runs saved to equal a win are on the same scale. In that case a pitching WAR should be the same as a batting WAR.

They are on the same scale. But if you disagree with where I set replacement level, which is always going to be an arbitrary thing, you might think I have it set too low for pitchers and too high for hitters.

The pool of players who are positions switcher and utility players are, by definition, the ones whose body types and skills are suited for multiple positions.

Here's an idea if anyone is interested: Take the players who are primarily 3rd basemen. Select as your sample players who have 75% of their games played at third. Throw in a minimum height/weight. See how this sample does, as a group, in emergency games at short.

My global replacement level for hitters is 1.5 standard deviation-adjusted wins below overall average (batting vs. pitchers-included league average + baserunning vs. league average + fielding vs. positional average). So the starting position players in a non-DH league would be 12 wins below average per 162. With an average-hitting pitching staff and pinch-hitters, that's another 5.25 wins below average. And the pitching staff, I say 2.35 wins/200 IP * 1000 IP for the starters + say 3 wins for the relievers including leverage makes 32. 81-32 is 49 wins, or a .300 wpct.

AROM, Nate Silver found an eighteen-run difference between Freely Available 3B and SS from 1985-2005, not a five-run one. That is the base measure my system relies on (although my research on worst-regulars shows that gap as much higher in 1985 than in 2005--for 2005, the last year I have data, the 3B-SS gap is just eight runs, and it very may well have narrowed further since then).

Quoting myself #10
Here is the link for your file, which will be available for 7 Days.
http://www.yousendit.com/download/TzY0eUNKTlFrWS9IRGc9PQ (BJCHONE.txt)

Format is csv but Windows extension is .txt, sorry about that.
Now I must run for eight hours. I'll check in again then or tomorrow morning.
Please confirm whether this works.


EXTENT of that table: 1583 records, one for each player

This is only one way to describe its extensive scope:

1098 - Bill James rankings (2001) #1-100 for 100 pitchers and #1-125 for 998 players at eight other positions.
999 - career choneWAR ratings (1871-2008), Top 500 pitcher ratings except JMWard and Top 500 position player ratings

Jointly that is 1541 major league players.

34 other HOM members
: Spalding, McVey, Pike, Pearce
: Irvin, Paige, Willard Brown, Trouppe (4 "Negro Leagues" HOMers)
: 26 who never played in MLB (26 more "Negro Leagues" HOMers)

8 other "Negro Leagues" candidates named on HOM ballots 2008 or 2009
: Easter, Clarkson
: 6 who never played in MLB

Those 42 players without [BJrank] or [WAR] include 9 without [bbrefID] because they never played in the majors and they are not in the Hall of Fame. Thus they were not in the "Lahman database" and did not have appreciation pages at Baseball-Reference a couple years ago; I haven't checked recently. They do have [lahmanID] values in this table because that field is really the non-negative numeric lahmanID augmented by my assignments of numbers less than 1, or greater than 20000 iirc.

DanR - For the current era (1993-2009, give or take), what's the average WAR for a full-time position player (~650 PA)?

INTENSIVE SCOPE of that table - 23 fields

lahmanID
: numeric, never missing
playerID, bbrefID, retroID
: if you don't know, it doesn't matter

BJfieldpos, BJrank, BJ100
: classifications and rankings by Bill James, New BJHBA (2001). BJrank covers 1098 major league players. BJ100 is the "Best 100" which actually covers 110 including some from the Negro Leagues.

HOM
: ='m' for each of 237 Hall of Merit members
Xfieldpos, Xfieldvote, Xfieldrank, Xgroupvote, Xgrouprank, X09vote, X08vote
: seven fields with names 'X...' provide Hall of Merit data covering the 2008 and 2009 annual elections and the special elections of 2008-2009.

cWS
: career Win Shares 1876-2001 from Win Shares (2002) for 1546 major league players. (Somehow a few recent mlbers with choneWAR ratings have zero and a few are null like the NeLers who did not play MLB.)
WAR
: career choneWAR 1871-2008 for 999 mlb players by Sean Smith, BaseballProjections. This is pitcher rating for the top 500 pitchers except J.M. Ward; position player rating for the top 500 position players.

firstname, lastname
: never missing

IP
: mlb career innings pitched 1871-2008 (615)
ERA+
: mlb career 1871-2008 (515) - the difference (100) is Dolly Gray plus 99 "position players" who pitched fewer than 100 innings
PA
: mlb career plate appearances 1871-2008 (1551) - never missing; null only for 32 HOM members/candidates who did not play MLB
OPS+
: mlb career 1871-2008 (1464) - the difference (87) is 77 genuinely missing plus 10 pitchers with zero career PA

You have a biased sample. It's just that simple.

Yes, definitely.

But the question is if this bias is even more disproportionate in the turf years. If it is not, then we can say that the bias is the same throughout the Retro years. So, if you have a problem in the 1970s, you have the same problem elsewhere.

If on the other hand someone shows that this bias is more disproportionate in the turf years, then you need to apply a further adjustment. Basically, the argument of the specialized position, like C is, and what the three IF positions is (not lefthaded throwers), is being hypothesized to also apply to turf-SS.

We need evidence now to move the discussion forward.

***

AROM, Nate Silver found an eighteen-run difference between Freely Available 3B and SS from 1985-2005, not a five-run one.

And this is ridiculous for reasons I stated.

If SS were so valuable, then why would a SS move to 2B/3B while still in his late 20s? I mean, he's played SS all his life, he's got no competition there, and then, suddenly, he moves to 2B or 3B where there's more competition, and teams have a huge problem trying to find SS?

18 run difference? If that's true, if there was such a chasm of talent there, then a ton of 2B/3B would have moved INTO SS disproportionately in that era. Did this actually happen?

I mean, look at Eckstein, the prototypical 2B. He found a home at SS because, possibly, there was a huge dearth there. But, why was he the exception? Why would, say, Orlando Hudson move to 2B? Because he was in the same farm system as Izturis and other top SS. He should have had a ton of value as a SS to be traded for a high class 2B.

When you work out the implication of the 18-run gap, it simply doesn't make any sense. All the things you expect to find with such a huge gap simply doesn't happen.

Isn't it possible that Nate's pool of "freely available" "SS" is not really representative of what the actual bottom-level talent level that can be found there? Indeed, isn't it more likely that he's completely wrong to say that?

And one more thing: if you can compile a list of all the SS that were traded, and you show me their WAR in the next 3 years and compare that to the WAR for the players they were traded for, and if Dan is correct, then we should see a huge imbalance.

Basically, the idea is that if Rally is not giving enough value to the SS because he has a too-high replacement level, then we should see this in trades. Guys that Rally would say "why the heck was this crappy SS traded for that great 2B" should be very few, according to rally, and Dan would think "yup, that's cause that SS was not that crappy".

Show me the evidence from that marketplace.

You're going to see a huge imbalance in the Hardy for Gomez trade...

CW--for '93-'05 (the last year I have data), total WARP2 divided by total SFrac gives 1.957.

DL responded to my #60, I think.
65. DL from MN Posted: November 25, 2009 at 01:22 PM (#3396738)
I'm not understanding the problem with comparing the two lists side by side. Is it that pitcher hitting isn't included in the pitching list?

It goes both ways. The two Top 500 lists are what they say.
- Babe Ruth is listed in position players at 172 which doesn't count 18 for pitching
- Bob Caruthers is listed in pitchers at 52 which doesn't count 19 for batting/running/fielding
- J.M. Ward is listed in position players at 39 and in pitchers at 25 which don't count each other.

There is no "problem" with the dual choneWAR rating system here but it seems to me that many people use the Top 500 lists --the only compilation as far as I know-- rather than the player pages.

This publication decision is parallel to those made for Pete Palmer's ratings by Thorn & Palmer, Total Baseball (print, 1989 to 2001), and by Palmer & Gillette, Baseball Encyclopedia (2003 to ??). Hyperlinks permit one to flip quickly from the choneWAR top 500 listings to the individual player pages, and BaseballProjections puts the Pitching and Batting/Fielding on one player page. But the structure of the ultra-popular ranked lists of career leaders are "the same". (In BE2004, Babe Ruth is #2 batter-fielder at 112 wins, #130 pitcher at 17 wins, and #1 player at 129 wins.)
: not parallel, as far as I know Sean Smith has not published any overall player rankings or ratings


Somewhere I said that the publication in two Top 500 lists represents a "coup" by the author. Providing the two convenient lists is the reason why we suddenly (here in the HOM subsite, more gradual at baseball-fever) see WAR (one WAR with different attributions, I think) cited and quoted everywhere. There is a parallel to TPR, TPI and their sum TBR after the publication of Total Baseball 3 (1993) in paperback with price $35 (TB's only ppb or cheap edition).

Guys that Rally would say "why the heck was this crappy SS traded for that great 2B" should be very few, according to rally, and Dan would think "yup, that's cause that SS was not that crappy".

Ivan DeJesus for Ryne Sandberg!! I kid - that was DeJesus for Bowa with Sandberg as a throw-in, of course.

(edited because I forgot which SS the Cubs traded away)

I took a brief look at Tango's idea of traded shortstops and checked on some trades of the Herzog Cardinals. Small sample, and I don't think I learned anything.

In the mid 80s, the Cardinals system exported shortstops, for the obvious reason that none of them were ever going to take Ozzie Smith's job. Three of note:

Rafael Santana was simply released by the Cardinals, and signed as a free agent with the Mets. He went on to have several years as a major league regular.

Jose Uribe was a lesser part of a multi-player deal: Dave LaPoint, David Green, Rajsich, and Uribe to the Giants for Jack Clark. I like to short-hand this as LaPoint for Clark, which lets one pair it with the deal that sent George Hendrick to the Pirates for John Tudor. (And which always raises the point of why the Giants couldn't have just dealt Clark for Tudor and cut the Cardinals out of it.) Uribe doesn't seem to have been the focus of the deal. He also makes for a good line about being the ultimate "player to be named later," as he had been playing up until then under the name Jose Gonzalez. The Cardinals won that trade in terms of value, getting a genuine big hitter in Clark. Uribe was a major league regular for several years. I'm not sure what I can learn about the value of SS from this trade.

Argenis Salazar was traded to the Mets (with minor leaguers going in both directions) for another SS, Jose Oquendo. Salazar eventually became a starter for Kansas City, and turned in some of the worst offensive performances you'd ever want to see. Oquendo had been a starting SS briefly at a very young age, and had no bat then, but was in the minors reinventing himself at the time of the trade. Although he remained blocked by Ozzie, Oquendo became a useful player for the Cardinals, mostly playing other positions. Since the trade was SS for SS, it doesn't tell us much about the intrinsic value of the position. The only interesting tidbit is about the blocked SS contributing at other positions.

continuing the point (#60, 65, 77)

Top 500 Position Players
Bobby Wallace 60.4 ranks #106 on the choneWAR list of so-called Position Players, just ahead of Jimmy Wynn and Dave Winfield. Add his pitching WAR, +5.4 =65.8, and he would rank #75 just behind Gary Carter and Craig Biggio. Most of these players have pitching WAR=0 so Wallace probably ranks #75 among them if they all get credit for their pitching. No one will kick Wallace out of any halls either way but that's a big difference in another sense.

John Ward 39.1 ranks #290. Add his pitching WAR +25.4 = 64.5 and he would jump more than 200 ranks on that list, not quite enough to surpass fellow HOM shortstop Wallace.

Top 500 Pitchers
Consider this run of 15 from the "Top 500 Pitchers", ranks 25 to 39, plus Bob Caruthers #68 and John Ward #305.

'Rk' is rank and 'WAR' is pitching choneWAR.

Rk WAR
25 71.6 + Charley Radbourn
26 70.8 --- Don Sutton (down -5.2 to 65.6
27 70.5 ---- Pud Galvin (down -11.7 to 58.8 !
28 69.7 -- Curt Schilling
29 67.0 ++ Tom Glavine
30 66.3 - Rick Reuschel
31 66.0 -- Bob Feller
32 65.7 ++ Don Drysdale
33 65.1 ++ Tony Mullane
34 64.8 - Kevin Brown
35 64.7 - Jim McCormick
36 64.6 + John Smoltz
37 64.4 - Carl Hubbell
38 64.0 - Juan Marichal
39 63.5 - Jim Palmer
...
68 52.6 ++++ Bob Caruthers (up +18.8 to 61.4 !
...
305 25.4 * John Ward (up +39.1 to 64.5 !

There are 3-win gaps both ahead and behind this group of fifteen pitchers but it doesn't quite retain its integrity among the so-called pitchers if we add credit for batting, running, and fielding. Radbourn climbs over Pedro Martinez, at least, and Galvin probably drops under a few, out of the top 40, without considering Caruthers and Ward who also surpass him.

The symbols '+' and '-' roughly represent batting/running/fielding. According to the sum (total choneWAR?),
- Radbourn leads this group followed by the three big pluses (4); then, I think
- Schilling, Smoltz (the other plus), Reuschel and Sutton (tie), Ward (5)
- Feller and the five small minuses in some order with 63 to 64 WAR (6)
- Caruthers 61.4
- Galvin 58.8, out of the top forty "pitchers".

Coda
Again, this isn't about the WAR rating system, only about two coupious publications that many people misinterpret, I believe (some people, I know).

John Montgomery Ward is a HOVG hitter and pitcher, and if you add the two together a very deserving HOFer. I think he's unique in that regard among all players in history.

Well, I recall early 1970s NL thinking being that with the new stadiums and Astroturf, you needed a whippet to cover a lot of ground at SS. If he never hits a HR and rarely hits a 2B, oh well.

I was just a kid, but I'd laugh at the baseball cards of those guys.

To take a less obvious example than usual, Craig Robinson, 1974 Braves.
506 PA, 0 HR, 4 2B.
Yes, really.
.230/.280/.265
OPS+ 51
5-foot-10, 165 lbs

And this team won 88 games!


http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=360191288171&rvr;_id=&crlp=1_263602_263622&UA=WXF?&GUID=d4c433f111f0a04359621682ffff760a&itemid=360191288171&ff4=263602_263622

I think that any system that accepts the idea that GMs really did an efficient job of analyzing options in this era, and this really was the optimal play - I mean, c'mon.

He'd have to be Bugs Bunny on defense to make up for this, lol.

Oops!
and it's too late to edit #80. That work includes a 10-point clerical error against Bob Caruthers. At (52.6+18.8=) 71.4 he surpasses everyone on that list except Radbourn and seems to rank number 26 among the "Pitchers" between Pedro Martinez and Glavine.

The overall rating helps Babe Ruth, too, of course. It doesn't have the striking effect here that it has for the Palmer ratings in the Baseball Encyclopedia (2004, at least), because choneWAR rates the Babe slightly ahead of Barry Bonds by batting/running/fielding alone.

John Montgomery Ward is a HOVG hitter and pitcher, and if you add the two together a very deserving HOFer. I think he's unique in that regard among all players in history.

You should have said that he's unique in MLB history. Among all players in history, we shouldn't forget Martín Dihigo.

AROM # 33:

I adjust the replacement level when one league is stronger than another - AL now, NL in the 50's/60's.

I wasn't aware that you made this adjustment and will need to make some minor changes to my ballot (next year—I won't have a chance to update it over this holiday weekend). I had added my own adjustments for league quality and will need to undo some of them.

AROM, how did you calculate your adjustments for league quality—by comparing players' records before and after switching leagues? by using league records in interleague play? something else?

Should I infer that you make no adjustments for events that affect both leagues simultaneously (the 1969 expansion, WWII)?

Do the league quality adjustments only cover the recent (retrosheet) era? Or do you cover earlier history (Federal League, American Association, AL/NL quality gap during the 1910s, etc.)?

Is your method able to pick up temporary deviations (like the AL expanding in 1961 one year before the NL), or is it based on averages over longer periods of time?

Expansion: no

But WWII, fed, AA, and imbalances between AL and NL are accounted for by decade.

(see #60, 65, 77, 80, 83)

I looked up the batting/running/fielding for a few more of those pitchers. It seems to me that Galvin drops to rank #45 behind Red Ruffing, Lyons, Rusie, and Bunning as well as everyone whom I listed here, or #46 counting Ward. Caruthers, Ruffing, and Lyons break into the top forty, Lyons nudging Marichal out.


DL #32 posted "Top unelected pitchers from Chone". Here is my edition ranked by total choneWAR. The numerals represent their places in DL's list (by pitching choneWAR). My main point is another illustration that it makes a difference.

2 Tony Mullane
1 Rick Reuschel
3 Jim McCormick
8 Charlie Buffinton
4 Luis Tiant
# Tommy John
5 David Cone
7 Mickey Welch
10 Frank Tanana
11 Chuck Finley <Even with Jack Clark among hitters>
13 Orel Hershiser
9 Larry Jackson
# Jerry Koosman
+ Jack Stivetts
12 Silver King
6 Vic Willis (-6.7 down to 50.5 !
+ Urban Shocker (no war credit)
+ Dwight Gooden
15 Kevin Appier
14 Tommy Bridges (no war credit)
Jack Quinn

#, rank higher by pitching alone. DL's oversights or my clerical errors.
+, break up DL's selection thanks to strong batting/running/fielding records.

I tried to catch everyone who might break up DL's selection; that is, everyone whose total might surpass "Tommy Bridges (no war credit)" 50.7 - 1.0 = 49.7.

Jack Quinn is next at 49.5. Carl Mays, Wilbur Cooper, Dizzy Trout, George Uhle, Bucky Walters, Dolf Luque, and Burleigh Grimes have good batting/running/fielding WAR but fall short. Dave Foutz played a lot of outfield and firstbase, falls short too.

choneWAR picks up that Urban Shocker (no war credit) beats that crowd, so it's a good rating ;-)

Expansion: no

But WWII, fed, AA, and imbalances between AL and NL are accounted for by decade.

Comparing Pete Browning (AA) with Ned Williamson (NL), it looks like replacement for the AA is about 10 runs per 162 games versus 20 runs per 162 games for the NL. Research (starting with Cramer and including some work done here at the HoM) has generally indicated that AA quality was much lower than NL during the first part of the decade (1882-84), but was nearly equal to the senior circuit later in the decade. Your numbers appear to use the same adjustment (prorated by number of games) throughout the decade.

I feel obliged to add my two cents here, since I was pushing Rally's WAR in discussion of my ballot, but I don't really have anything to add. Rally's methods are pretty straightforward, and the only thing in any way controversial seems to be the replacement level and positional adjustment. As Dan's examples seem to show, the minor concerns (such as sharp changes in positional adjustments at the turns of decades) don't significantly affect the results. And of course as always defensive numbers (esp. before 1954) should be taken with a grain of salt.

I personally think the way Rally handles position + replacement is optimal, but if one disagrees, one will be suspicious of the results. Of course, because they're so straightforward and so clearly presented, it's very easy to adjust around any suspicions one might have. I personally fiddle with the positional adjustments for catchers and DHs, as I think he's a little lenient on the latter and a little harsh on the former. It's also very simple to adjust the defensive numbers or adjust according to Dan's standard deviation rates.

I'd be very interested in a discussion of Rally's WAR numbers for pitchers. I don't rely on those nearly as heavily, and his methodology isn't quite as clear there.

EDIT: Wow. Somehow I missed 42-88 when I posted this. I'll get caught up this weekend some time.

AROM, Nate Silver found an eighteen-run difference between Freely Available 3B and SS from 1985-2005, not a five-run one. That is the base measure my system relies on (although my research on worst-regulars shows that gap as much higher in 1985 than in 2005--for 2005, the last year I have data, the 3B-SS gap is just eight runs, and it very may well have narrowed further since then).

Yeah, about that. Using Silver's criteria for FAT, I searched the Lahman DB and found... 4 3B and 5 SS. That's a much larger sampling issue than the selective sampling issue from Rally's position-switching method.

And speaking of sample selection - the average SS in Silver's study has an age of 29. The average shortstop has an age of 27.6531, the lowest average age for any position. Silver is looking at older players because they're the ones that are easiest to identify as replacement players, but he's excluding younger replacement players in the process - and younger replacement players are disproportionately likely to be replacement SS. (This is why 40% of the players who meet Silver's criteria are first basemen.)

I'm sorry, but there simply aren't enough data points in the Silver study for you to place the weight upon it that you have.

Paul (post 83) corrects:

Rick Reuschel
Tony Mullane
<Kevin Brown>
Jim McCormick
Luis Tiant
David Cone
Vic Willis
Mickey Welch

To this:

2 Tony Mullane
1 Rick Reuschel
3 Jim McCormick
8 Charlie Buffinton
4 Luis Tiant
# Tommy John
5 David Cone
7 Mickey Welch

As some of you know, I've been obsessing over 1880s pitchers for the last week (with questionable results due to illness, but I did put in some hours). To these lists I would comment the following:

First, I think that Paul's list is MUCH better than the one DL quoted.

Second, Tony Mullane, out of this crowd, is the one most tainted by the American Association, by a large amount. I have him lower because of that. Compared to Mullane, McCormick pitched in the NL, had a much more serious workload at its peak (3 seasons over 550 IP), but pitched about 250 career innings fewer. That's why he's on my HoM list and Tony is not. The two of them, in career IP and ERA+, are in a pretty tight group with Hoss Radbourne, who pitched almost exactly Mullane's load (difference of 4 IP out of over 4500), but had a much higher peak than either Jim or Tony.

Third, where you place Pud Galvin depends almost entirely on how you weigh duration against rate. Galvin has the second-most IP ever (about 6000), but only a 107 ERA+. He's in the HoM with Radbourne, while Jim and Tony are not, and I think that's exactly correct ordering (although I think Jim should be in the HoM).

Fourth, a guy named Bobby Mathews is Pud Galvin light - or maybe not so lite. He "only" has 4956 IP, but he started in 1871, when schedules were scanty; Galvin started in 1875. He has the exact same 107 ERA+. He was a rookie at age 19 and retired at age 35; Galvin is 18 and 35. Mathews had a WL% of .545, Galvin is .540. Given the scantiness of those early schedules cutting down Mathews' IP, you can make a case that Mathews and Galvin are very very close.

Fifth, Mickey Welch is a weird outlier and hard to rank. His ERA+ is 114, which is very normal for an ace starter of the period. But his IP are 4802, and there's no one close (meaning within a THOUSAND) to his IP who has anything really close to the ERA+.

Sixth, Buffinton has a 115 ERA+, but only 3404 IP, which almost drops him out of the competition. Remember, Welch has 114 over 4802 IP. I don't understand why Paul has Charlie so high, although Paul always does have a good reason, and I'm looking forward to reading it. Best guess: the .605 WL%, but Paul can answer for himself.

As of right now, I would have the group of them (skipping the later guys) as:

Radbourne
(big gap)
Galvin
McCormick
Welch
Mathews
Mullane
Buffinton

And I tend to give the AA guys a break, unless they are at the beginning or end of that league.

- Brock Hanke

Um. That would be post 87, not 83. And the main reason that I think Paul's list is better is that it takes out Brown and Willis, who I don't think really belong in the group, and adds John, who probably does. The most interesting things about the lists are 1) Will White is not in either of them, and 2) both of them have Jim McCormick ahead of Luis Tiant. I have Tiant on my ballot ahead of McCormick, but it never occurred to me to compare the two of them. I'll get to that today. It should be interesting, because of the massive era adjustments. And it might give me a much better idea of where to place McCormick, which would make the whole task of placing 1800s guys much easier. I'm being conservative right now. The Tiant comparison should help my confidence, at least.

CW, you're saying there were a grand total of four third basemen and five shortstops in MLB over 20 years who were over 27 and made less than twice the league minimum???? I think something must be wrong with how your query was inputted...I'll double-check this as soon as I get a second.

Hey, epoc, glad to see you're back...did you ever get that prelim ballot moved over to the actual ballot thread? You need to do that for it to be counted.

#91,94
Brock, You are still overtired. This is "Chone's WARP" and I am on topic. Thus,

Paul Wendt #87
DL #32 posted "Top unelected pitchers from Chone". Here is my edition ranked by total choneWAR. The numerals represent their places in DL's list (by pitching choneWAR). My main point is another illustration that it makes a difference.

I don't mean to endorse that #87 ranking of eligible pitchers by "total choneWAR" as an adequate ranking for the Hall of Merit (neither does Sean Smith). DL doesn't mean that #32 ranking by "pitching choneWAR" that way, either; rather to deprecate it, I think. (DL: "Why aren't the "Chone Voters" filling their ballots with pitchers? His system is saying they're more valuable than the hitters they're voting for.")

I do agree with "total" ratings rather than "pitching" ratings for the Hall of Merit. Mullane and Buffinton rank second and eighth among eligible pitchers by pitching choneWAR alone (if DL didn't make any clerical errors in #32). They rank first and fourth by total choneWAR (if I didn't make any clerical errors in #87) because they were very good batters and the others weren't. That isn't a better ranking of Mullane and Buffinton, I agree with you, but it's appropriate here at the HOM to sum the ratings rather than to rank pitchers by pitching alone, in my opinion.

Here at the HOM "choneWAR" needs some adjustment for the numbers of team games played, as everyone is saying re "schedule adjustment". Evidently it also needs some adjustment for the practical demands on starting pitchers, who once worked all of a team's championship games, still commonly worked one-third of a team's games in the 1890s, and so on. "Overrating" mid-1880s and earlier pitchers, insofar as that is what we see in both #32 and #87, shows that for them it is more important to down-adjust early workloads at the pitching level than to up-adjust early schedules at the team level.


Paul's list is better is that it takes out Brown and Willis, who I don't think really belong in the group, and adds John, who probably does.

This mixes the issues of eligibility (it's Kevin Brown) and pure pitching. Sticking to the latter, Willis drops a lot in the move from ranking by a pure pitching rating (#32) to ranking by a comprehensive rating (#87). That's right on the mark, and that's the only move featured here and in #87:

pitch + other = sum
64 + 2 = 66, John Smoltz - not eligible, trails eligible pitcher Mullane alone by total WAR, trails three by pitching WAR
58 - 5 = 53, Jerry Koosman - #13 eligible pitcher by total WAR, #6 by pitching WAR
57 - 7 = 50, Vic Willis - #16 eligible pitcher by total WAR, #8 by pitching WAR
47 + 3 = 50, Urban Shocker - #17 eligible pitcher by total WAR, #24 by pitching WAR

In my opinion Willis and Koosman, say, need to drop further relative to Shocker and Smoltz, say. Willis and Koosman pitched when a lot of pitchers were pitching a lot, partly because of supporting conditions. Smoltz isn't eligible, of course. He's the example here because he is the only closest recent pitcher I see, in the vicinity of Koosman and Willis by pitching WAR, with at least a moderate plus by batting/running/fielding WAR.

The preface by Joe Dimino:
Please use this thread to discuss Chone’s WARP and how it relates to the Hall of Merit and other ‘uber-systems’, like DanR’s WAR, BPro WARP (1, 2 and 3), Win Shares, my (Joe Dimino’s) pitcher ratings, VORP, etc.. If I forgot one let me know and I’ll add it to the list as well.

Both Tom Tango's WAR and Fangraphs WAR have been mentioned recently. Evidently Fangraphs WAR is WAR by Tom Tango aka tangotiger (tangoWAR?), not necessarily the latest edition. See "Confused Says What?… Getting to Know FanGraphs Stats" by tangotiger.
>>
TangoTiger says:
December 10, 2008 at 12:13 pm

Scott: WAR exists by many people by many names in many forms. I’ve invented my own version, of which I call WAR. Whether someone else also uses that name, preceding me, or following me, I don’t know.
<<


I have mentioned Pete Palmer's ratings published in the Thorn/Palmer and Palmer/Gillette print encyclopedias of the last twenty years and explained someways that publication of choneWAR at Baseball Projection (no 's') is and isn't like the Palmer ratings in print.

Some recent edition(*) of the two Palmer ratings(#) has been published at Retrosheet.org in the season and career lines of all player pages. No components of Batter-Fielder Wins (BFW) or Pitcher Wins (PW) are given, however.
: John Ward at retrosheet
As I have explained, those two ratings must be summed to get a comprehensive rating of the ballplayer (career 20.7 for John Ward). The sums are unpublished, except the lists of all-time leaders in the print encyclopedias. In substance the two ratings cover pitchers and all other players where Sean Smith's two published WAR ratings cover pitching and all other functions. So Palmer allocates a player's batting to his pitcher games and all his other games --proportionally, I suppose, using innings and other data alongside games. It turns out that Babe Ruth afater 1920 gets about one batting run for each pitcher game, as the batting contribution to his pitcher wins (PW).

* BFW and PW for John Ward at retrosheet match those in the Palmer/Gillette 2005 ESPN Baseball Encyclopedia (print).

# BFW and PW are the current editions of TPR and TPI. Total Player Rating, Total Pitcher Index, and their abbreviations were trademarks of Total Baseball, later Total Sports, distributed who knows where after its bankruptcy.

I must say that this thread shows off the best of Baseball Think Factory. Kudos to all!

Paul, thanks for the post. I'm still not at my best, but I'm clearly getting over the last bout of the flu, and this wasn't as foggy as it might have been. The only real mistake I made was to not realize that you were not correcting the original system, but just adding batting and fielding wins to a pitching component. I failed to notice this mostly because I think adding in the hitting and fielding (if you can estimate a pitcher's fielding) is such a no-brainer that it didn't occur to me that "pitching ChoneWAR" did not mean "ChoneWAR for all pitchers, including their pitching, hitting, and fielding." So I ended up thinking you'd made a tweak to ChoneWAR. I know d*mn well that I'm not up to analyzing anyone's system right now, and so didn't try to look at ChoneWAR, and so didn't catch the mental error. Sorry about that. You are, of course, correct to add the extra wins in. Oh, and thanks for replying to my question about Buffinton. I've looked at his bat before, and it's good for a pitcher, but he wasn't Bob Caruthers or someone like that. I still find it hard to think that his bat can make up for the large IP weakness. And I wasn't so bad off as to miss which Brown it was; I just don't think he belongs in with the bulk of that group. And I would like to know where total ChoneWAR places Will White. I have him higher than most people, I think. It would be very useful to know where a system that actually does compare early pitchers to each other has him and why. - Brock (I ain't takin' no kudos until my brain works normally again) Hanke

Oh, and Kid Gleason asked me to complain that John Ward is not completely unique as a pitcher moved to position play after his arm blew out.

And Smokey Joe Wood.