Personally, Alomar and Larkin are givens, Martinez is most likely with MiL credit and McGriff and Ventura deserve careful analysis. The rest need not apply.

Martinez is most likely with MiL credit

I am of the opinion that Martinez doesn't deserve MiL credit, if you look at his career arc honestly and in context. In 1985 and 1986, he didn't hit especially well in the Southern League. In 1987, you need to consider that his numbers were posted (a) in the PCL and (b) in Calgary, one of the better parks for hitters in that league. I would argue that it was not until 1988 that he definitely established his credentials for a major league job - and he got one a year later, albeit as a backup with a .618 OPS.

-- MWE

I'd expected to be cynical about these guys, but on a first cut Alomar and Larkin are both in with a bit to spare, as is Edgar even without MiL credit; McGriff is probably just below borderline. Alomar's a very different kettle of fish from Ozzie Smith, who I didn't support; 29 points of OPS+ (116 vs.87) is worth WAY more than the SS/2B differential plus fielding bonus, IMHO.

183 56.5 1992 Eric Karros-1B

This makes me feel old, and also shows how long my Rotisserie league is. Eric Karros' rookie year sparked a heated fight at my annual roto auction- the owner who had Karros reserved as a minor leaguer insisted (this was 3 days before the season opener) that Karros had not made the Dodgers, wouldn't break camp with him, and that therefore he (the owner) didn't have to activate Karros (costing him a roster spot and $10)- he could simply keep him reserved as a minor leaguer.
Everyone else (literally everyone else, 11 out of 12 owners) said no, Karros was on the Dodgers, everyone knew that, he had "made" the team and was in fact going to start- accordingly Karros' owner either had to activate him (costing a roster spot and $10), or release him so other teams could bid on him at the auction.

The upshot was that the "Karros' Rule" was invented- if at auction there was a dispute as to whether a reserved minor leaguer was going to be in the MLB on opening day the owner had a choice
1: Activate him, which cost $10, a roster spot and a year of service time (you could get 3 years out of player)
2: Don't activate him - if on opening day he's on a MLB 25 man roster you lose him, he's FA and any team can claim him.

So the Karros owner reserved him- and promptly lost him on opening day as he took the field for the Dodgers- which was a completely dumbass move since Karros was and would be for quite some time well worth $10+ in an NL only Roto league...

and now Karros has been retired for 5 years and is eligible for the HOF...

I was always a huge Mark McLemore fan. Not a Hall of Famer, but a pretty underrated player in his day IMO.

Shouldn't Juan Gone be in? For a decade was one of the most FEARED hitters in the AL!

You might be right, Mike. I'd still like to see MiL translations for him when somebody gets the chance.

I'm not sure if I see the Edgar MiL credit argument, either. The only year he really has a cae for is 1988, and that's a partial season.

Looking closely at his record, which I haven't really done since he retired, I'm struck that he really looks more like a pure late bloomer. He still hadn't added any HR power into the late-80's, which was probably the reason he wasn't brought up quicker, but he had a ton of doubles; his slugging profile was more like a 21-23 year old than a 26 year old.

Strange development history. If he weren't from PR, I'd question whether or not he was a little younger than advertised- it would fit his career better.

Shouldn't Juan Gone be in? For a decade was one of the most FEARED hitters in the AL!

The supporters of Gonzalez's candidacy have their rallying cry now! :-)

The supporters of Gonzalez's candidacy have their rallying cry now! :-)

Hey, Rice got in with one MVP. Juan Gonzalez has TWO! That's more than every other new person on the ballot combined!

I am of the opinion that Martinez doesn't deserve MiL credit, if you look at his career arc honestly

I don't think some HOM voters are looking at career arcs in determining minor league credit- if a guy hit well enough in the minors to suggest he was already a good MLB hitter (as Edgar arguably did in 1988- then they'll give him credit for that year- even if it was entirely reasonable for that players team to have had such player in the minors that year.

In 1987, you need to consider that his numbers were posted (a) in the PCL and (b) in Calgary, one of the better parks for hitters in that league

yes but he still lead a AAA league in OBP his first go around in AAA...

Of course the team as a whole hit .287/.373/.435...

Seriously, in looking at Edgar's minor league numbers, and using 20/20 hindsight (ie we know what Edgar was capable of as a hitter)-
why was Edgar so underwhelming in AA? His K/BB was typically great and in line with the rest of his minor league career... his power was down, but only a little,- his BABIP just completely cratered- for 2 years in a row...
For what it's worth, when Edgar hit .258/.378/.353 in AA at age 22, the team hit .263/.341/.374- but that was an OLD team- Edgar was the youngest regular- he then went on to hit .353/.450/.485 in 20 AAA games that year- only to be sent back to AA the next year- one could argue that he should have spent 1986 in AAA, then 1987 would have been his 2nd full year in AAA...

as an aside- wow did the Mariners have an OLD farm system back in the mid 80s- just as a guess there average team age was seriously old for every league, AAA, AA, A etc... I think the concept of using the minors to develop "prospects" was not very familiar to them.

His minor league record was, in context, that of a player with excellent strike zone control and below average pop. His IsoP was around .150 which would place him 4th or 5th on his triple A team in Calgary. He was a poor defender at third base, and didn't run well. How many teams would give him a shot as a DH or third baseman or convert him to an outfielder with that skill set?

which was probably the reason he wasn't brought up quicker,

The main reason was Jim Presley who the Mariners variously thought
1: Was going to be a star (1984/85)
2: Was a star (1985/86)
3: Would rebound (1987/88)
4: Can't possibly be thus bad... the rest of his career

Basically Presley was a less athletic version of Jeff Francoeur, similar hitter, completely unable to adjust when pitchers adjusted to him...

Using 20/20 hindsight it's pretty obvious that Edgar was a better player than Presley in 1988- but in 1985 and 1986 Presley had pretty well hit in the majors when Edgar still APPEARED to be struggling in AA...

How many teams would give him a shot as a DH or third baseman or convert him to an outfielder with that skill set?

Not many, but this is a team that should have- they had nothing better to do with their roster/time (seriously)
Jim Presley, after a promising start, not only failed to develop quickly regressed both offensively and defensively to the point where he may have been the worst regular in MLB from 1987-1990.

The 1988 MAriners had 27 year old Brantly in LF: .263/.296/.399;
29 year old Glen Wilson in RF: .250/.286/.324

Of course 1B and DH WERE in fact reasonably productive for the 1988 Mariners- Al Davis hit .295/.412/.462, Phelps hit .284/.434/.547 and Balboni hit .251/.298/.480
They also had Darnell Coles, who still looked like he might have a career in 1988

I agree that the Mariners ought to have given Martinez a shot over Presley at third base. However, in my view, that might very well have affected his offensive output in his 30s. Placing a player at a position where he is stretched defensively has had impacts on offensive development.

Lurker here: is there no interest in Omar Linares, or is he in a category with the Japan-only players, or did he play somewhere after 2004?

Lurker here: is there no interest in Omar Linares, or is he in a category with the Japan-only players, or did he play somewhere after 2004?

We expect to have an international wing of the HoM sometime in the future, Boog. Linares and other quality candidates would be eligible for it then.

Instead of posting all of the player threads this week, I think I'll release them one at a time starting next week instead. Any problems with that?

Hey all. I spoke with John over email and am going to start participating in the HoM. Currently, I'm doing my own rankings of the HoM inductees by position, as you all did previously, and then I'll start getting cracking on the non-inductees to prepare by my first ballot.

Just a little info on how I rank players...I don't have a strict system, or any sort of system at all, really. I look at each of Dan R's WARP, Win Shares, BP's WARP, and OPS+, mostly (for position players. I haven't yet started looking at pitchers). I don't have any strong tendencies towards peak or career. For me, it depends entirely on the respective peaks and careers that I'm comparing. I don't care about consecutive peaks. I look at best 3 years, best 5 years, best 7 years, and overall career. As I said, I don't use any sort of system, but rank each individual player on my own.

I'm positive that Alomar and Larkin will both be on my ballot this upcoming election, and of the other new candidates I'm going to give close looks to McGriff, Martinez, and Ventura. I'm on the fence on minor league credit for Edgar, but lean towards no.

I'd be glad to answer any questions anybody may have about my thoughts on anything. I look forward to contributing to the HoM. :)

My favorite offensive tool is keyed to a book whose stats stop with 1999, so I'm increasingly uncomfortable rating the new candidates - but I can at least make estimates.

Alomar seems to fit in offensively somewhere clearly below Gehringer but clearly ahead of Sandberg and Whitaker. That should make his case simple.

McGriff I have with offensive value in the same general range as Terry, Cash, Cepeda, and Hernandez. We elected two of those four - and McGriff doesn't have Hernandez's defensive value. Not sure where that is going.

I haven't worked up Martinez or Ventura just yet. They were both on my "all-1990's" team, but as bench players. (I decided to switch a 2B over to 3B because of the general shortage of 3B for that decade, and the two 1B/DH spots went to Thomas and McGwire.)

Now, what about Larkin? My system, which has always been fairly forgiving of gaps in playing time, puts Larkin's offense among the top (non-Wagner) group of hitting shortstops. The knock on Larkin is playing time, so I wrote down the following table of total playing time for all of these hitting SS:



Larkin played at least 145 games or the strike-shortened equivalent in the following seasons: 1988, 1990, 1994, 1995, 1996, 1998, 1999, 2002. And that was closer to 145 than 162 most of the time, although he somehow played 161 games in 1999 at the age of 35.

Here are his seasons of less playing time than that:
1986: 41 games (August call-up as rookie)
1987: 125
1989: 97
1991: 123
1992: 140
1997: 73
2000: 102
2001: 45
2003: 70
2004: 111 (Final season, age 40.)

It is difficult or impossible to govern the pace and direction of discussion. The method of player threads with simultaneous release hasn't worked so well that there is a lot to lose.

The Hall of Merit will be around for a while so you may also justify it as an experiment. If you will experiment, I recommend a little more system. Post the schedule such as "every Sunday evening in the following order:"

Hey all. I spoke with John over email and am going to start participating in the HoM.

I signed up, but then actually participating in the ballot discussions and voting seemed too much work... If they made voting for the real HOF as hard as voting for the HOM...

Seriously, I started working on ranking players (all players) awhile back, one thing I decided to do was to do a runs above average/replacement calculation, but calculated my baseline by park and era adjusting every player's numbers- then comparing against an 80% baseline calculated by using all players with 3000 PAs. That gave me a very high baseline- it also moves some players up and down- Dave Concepcion for instance is not compared to 1970s SSs- he's compared to all SSs with 3000+ PAs (Actually all players who had 3000+ PAs and spent 40%+ of their time at SS)- that lowers him considerably (on the offensive end) from where Dan R would rank him.

I never got around to finishing the rankings or adding defensive #s (I was thinking about just borrowing someone else's defensive #s).

anyway, without considering defense (position yes, actual dee no), I have the current crop ranked thusly:
Edgar Martinez 608.1
Barry Larkin 533.2
Roberto Alomar 524.6
Fred McGriff 501.6
Ellis Burks 376.1
Robin Ventura 327.4

Edgar is in an area where virtually everyone who is eligible is in- but no one else was a career DH.
Larkin's offensive value is in acluster with these guys:
George Van Haltren
Eddie Murray 538.9
Jimmy Ryan 535.0
Barry Larkin 533.4
Joe Kelley 533.2
Willie Keeler 525.8

which overlaps the Alomar cluster:
Willie Keeler 525.8
Roberto Alomar 524.6
Sam Thompson 524.3
Jake Beckley 524.0
Johnny Bench 521.1

McGriff is bookended by HOFers (who are bookended by non-HOHers:
Mike Tiernan 504.3
Carlton Fisk 502.7
Billy Williams 502.0
Fred McGriff 501.6
Al Simmons 500.3
Hank Greenberg 500.2
Jack Glasscock 497.8

That gave me a very high baseline- it also moves some players up and down- Dave Concepcion for instance is not compared to 1970s SSs- he's compared to all SSs with 3000+ PAs (Actually all players who had 3000+ PAs and spent 40%+ of their time at SS)
. . .
Edgar is in an area where virtually everyone who is eligible is in- but no one else was a career DH.

What does the group of 40% career DH look like?
eg, How many are there with 3000 PA and 1000 PA?
eg, How does their batting compare with firstbasemen, etc?

What does the group of 40% career DH look like?
eg, How many are there with 3000 PA and 1000 PA?
eg, How does their batting compare with firstbasemen, etc?

I cheated, Edgar is compared to 1Bs, the DH group was too small.
I was planning on fixing that... So Edgar is compared to a 5.514 r/g baseline, but the short DH list actually averaged 5.946

If you give Edgar a 5.946 baseline... he drops from 608.1 to 544.8...
The he's surrounded by:
Luke Appling
Elmer Flick
King Kelly
Edgar Martinez
Joe Cronin
Robin Yount
Dave Winfield

He had 8672 PAs at a 147 OPS+
If you do something q & D like PA* (OPS+ -75) Edgar's hitting peers looks like this:
Ed Delahanty
Joe Morgan
Paul Waner
Edgar Martinez
Jesse Burkett
Joe DiMaggio
Johnny Mize
Billy Williams
Spots 22 through 31 of the guys with 7500+ PAs and a 147 or higher OPS+ looks like this:
22 Honus Wagner 150
23 Nap Lajoie 150
24 Jeff Bagwell 149
25 Jim Thome 148
26 Harry Heilmann 148
27 Mike Schmidt 147
28 Willie McCovey 147
29 Alex Rodriguez 147
30 Willie Stargell 147
31 Edgar Martinez 147

The problem with all my lists regarding Edgar and where he's placed- is that he's almost unique in his utter lack of defensive value- he roughly matched Stargell in both PT and quality (as a hitter) - but Stargell clearly provided his teams more value.

in fact here is the entire DH baseline list:
Cnt Player OPS+ PA
+----+-----------------+----+-----+----+----+
1 Frank Thomas 156 10074
2 Edgar Martinez 147 8672
3 Travis Hafner 142 3000
4 David Ortiz 138 5428
5 Jose Canseco 132 8129
6 Cliff Johnson 125 4603
7 Paul Molitor 122 12160
8 Hal McRae 122 8058
9 Andre Thornton 122 6293
10 Harold Baines 120 11092
11 Chili Davis 120 9996
12 Don Baylor 118 9401
13 Brad Fullmer 111 3065
14 Ron Kittle 110 3013

I will have Larkin at the top of the ballot, Alomar somewhere near 12, Edgar PHoM only because of minor league credit. Ventura is ahead of McGriff but neither will be close to PHoM.

This is one of the few years without a PHoM pitcher eligible.

Very, very prelim ballot:

1-Larkin
2-Alomar
3 to 15-The 2009 ballot holdovers.

I actually hope that the MLE-fiends among us can pursue the Edgar Martinez MiLB credit case.

Based upon a preliminary assessment of the MiLB numbers, the whole "OMG EGDAR HOF BLOCKED STOOPID MARINERS" meme looks really questionable- perhaps Edgar was the caliber of a MLB regular by 1988, but it wasn't obvious from the statistical data. It looks like the Calgary-effect created a bit of a stathead myth.

I'd be real interested, for example, to see what a ZIPS/PECOTA prediction system would spit out for the future if you fed in Edgar's career up to age 24/25/26/27.

We've got a good everyday team from this year's class, but we're short on pitching:

C: Santiago
1B: McGriff
2B: Alomar
3B: Ventura
SS: Larkin
LF: Gonzalez
CF: Lankford
RF: Burks
DH: Martinez

C/3B/1B/OF/PH: Zeile
1B/PH: Karros
UT IF: McLemore
Extra UT IF: Viña

SP: Hentgen
Closer: Beck
RP: Jackson

Still need the rest of the rotation and the bullpen - does the shortage of pitchers from an otherwise fine team tell us something?

Could use an actual 4th OF, but we do have backup C covered with Zeile. McLemore and Viña are both basically 2B, whereas the need is really for a backup SS.

I'd bat Juan Gone 6th or even 7th in the order - too many good OBP's that I'd want in front of him. Larkin and Alomar, in some order, bat 1-2.

I'd bat Juan Gone 6th or even 7th in the order - too many good OBP's that I'd want in front of him. Larkin and Alomar, in some order, bat 1-2.

Thinking of making a lineup out of that...

Larkin
Alomar (really, it is interchangeable with those two)
Edgar
McGriff
Juan Gone
Lankford
Burks
Ventura
Santiago

There are several good possible permutations. One could, for example, move Ventura or Lankford to 3rd and bump everyone else down a bit. Also, I didn't check handedness, so one could adjust for that as well

Suggestions made by Baseball Musings' Lineup Analizer

OBPs and SLGs are neutralized.

Preliminary analysis.

How does Ventura's fielding grade out with the modern stats?
He is +154 BP-FRAA, which is outstanding. He also won 6 Gold Gloves at 3B.
(Nettles was +135 total but that would have been higher if he hadn't been -30 FRAA for his last decade).

Given his fielding is for real, then I see him as #3 on my ballot.
1) Alomar
2) Larkin
3) Ventura

Edgar will make the upper half of my ballot though I'm not sure exactly where.
McGriff needs a careful look. He's got a nice peak 1988-92 and a good career, but I'm not sure he has enough prime. A lot of mediocre seasons there post-peak.

(December 7, 2009)— - So the HOM is having 50 weeks of discussion for 2010? Seems a bit much? Shouldn't we be discussing ranking the HOM pitchers first?

Ventura is arguably a top 20 3B all-time and that alone gets him a look.

As for Edgar, how much does it matter that you can't seriously evaluate him among a positional cohort? I suppose that it shouldn't (matter). But it does.

I haven't worked up Martinez or Ventura just yet. They were both on my "all-1990's" team, but as bench players. (I decided to switch a 2B over to 3B because of the general shortage of 3B for that decade, and the two 1B/DH spots went to Thomas and McGwire.)

Uh, not quite. I went back and found my all-decade team file from nine years ago and I specifically left Ventura off. Here's what I had (and yes, 1990-1999 is arbitrary endpoint, and yes, I would have done some of the analysis different now.) I was using a lot of raw stats, which raises big problems not just with ballpark effects but with the differences between the AL and the NL and the differences between 1990-92 and 1993-99. I also didn't hesitate to shift positions if I thought it was possible, hence Biggio as a 3B (because he was better than any of the 3B available) and Bonds as a RF. Here's what I had:

Starting lineup
C: Piazza
1B: McGwire
2B: Alomar
3B: Biggio
SS: Larkin
LF: Belle
CF: Griffey
RF: Bonds (MVP of decade)
DH: Thomas
Bench
C: I. Rodriguez
1B: Bagwell
IF: Ripken
OF/PH/PR: R. Henderson
OF/PH: Gwynn
PH: E. Martinez
Rotation
Maddux (CY of decade)
Clemens
R. Johnson
P. Martinez
Glavine
Bullpen
Wetteland (Closer)
Cone
K. Brown
Smoltz
Hoffman

In retrospect, that bench has too many hitters and not enough infielders: maybe I should have included Ventura (or Boggs) and not Gwynn.

Better to have Belle in RF; Bonds's one missing tool was his arm.

(December 7, 2009)— - So the HOM is having 50 weeks of discussion for 2010? Seems a bit much? Shouldn't we be discussing ranking the HOM pitchers first?

Hey, we can walk and chew bubble gum at the same time around here. :-)

I do agree we should start figuring out what the heck to do with the hurlers.

My very initial prelim etc. etc.

PHoM: Robbie, Barry, Fred

1. Alomar
2. Larkin
3. Leach
4. McGriff
5. Perez
6. Van Haltren
7. Welch
8. Duffy
9. Brock
10. Staub
11. Bonds
12. Martinez
13. Monroe
14. Willis
15. Singleton

16-20. Cash, Murphy, Johnson, Puckett, Grace
21-25. Redding, S.Rice, Grimes, Streeter, Gleason
26-30. Greene, Ryan, Bremmer, Coimbre, Robinson

Here's my (very) prelim ballot.

1) Larkin (PHOM)
2) Bridges
3) Tiant
4) Leach
5) Cone
6) Reuschel
7) Edgar Martinez (PHOM) - includes minor league credit of 1 season
8) Bob Johnson
9) Lee Smith
10) Urban Shocker
11) Alomar (PHOM)
12) Ben Taylor
13) Clarkson
14) Cash
15) Cravath

16-20) McCormick, Schang, Cey, Redding (END OF PHOM not HOM)
20-25) Quinn, Cross, Appier, Pesky, Dom DiMaggio, Rizzuto
26-30) Trucks, Trout, Hilton Smith, Bobby Bonds, Bob Elliott
31-35) Robin Ventura, Chuck Finley, Kiki Cuyler, Buddy Bell, Jose Cruz
77) McGriff - solid HoVG territory

I'm sure this will change by next fall. One change I'm intending to make is with postseason credit. Sean Forman has calculated WPA for all postseason games. I'm hoping this is available as a searchable spreadsheet so I can add in postseason credit for all players. This has the potential to slide people around since the biggest gap after Larkin-Bridges is 3 points.

Here's a relevant discussion:

http://www.baseballthinkfactory.org/files/newsstand/discussion/studeman_postseason_probability_added/

Lee Smith ahead of Alomar? That surprises me. How do you deal with relievers generally, DL?

Runs above average and runs above replacement from WARP2. Not pretty but there's no relief innings in Dan R's numbers.

I wouldn't advise using BP's numbers for evaluation of relievers, and in particular for comparison to starters. For one thing, the baseline has to be different (for reasons described in Steve Treder's articles in THT several years ago, the role of reliever leads to significantly better statistical performance in K/9, BB/9, HR/9, as well as ERA). Then, you have to consider leverage and chaining.

Rusty Priske has Larkin ahead of Alomar, and DL from MN has Larkin 10 places ahead of Alomar.

I haven't worked through that yet, and might arrive at the same order myself - but it's not clear to me yet. I think it's clear that when they were both on the field operating at full strength, Larkin was a better baseball player than Alomar. But Alomar was on the field considerably more often. Look at the career WS and WARP numbers at the top of this page - that shows the playing time difference.

I think the big question about Alomar is his fielding. By reputation he was outstanding, but Zone Rating and the Retrosheet defensive stats like SFR (which show exceedingly high correlations to an average of PBP metrics) find him merely average. As an average-fielding modern 2B (so the same positional value as 3B), his hitting and baserunning are just good enough to get him into the bottom 10% of the HoM. If you give weight to the sterling anecdotal reports of his defense, and/or if you compare him to 2B when the position was more valuable (roughly 1925-1985), then he becomes a no-brainer.

For the 50 zillionth time, WS and WARP use a replacement level that bears no resemblance whatsoever to MLB reality, leading to a gross overvaluation of mere "showing up" over quality while on the field. Clay Davenport is in the process of fixing WARP right now to reflect this, and WS hasn't been updated since 2001. There is a clear sabermetric consensus here, and continuing to cite raw WS and WARP totals in this context is nearly tantamount to advocating for a flat earth hypothesis in my view.

DanR,

I can understand why the positional adjustments for 2B relative to 3B might change significantly from 1910 to 1930 say, but I don't understand why the positional adjustments should have tilted from 2B toward 3B during the period from 1975 to 1995, say. Sacrifices were down, double plays were up. These are factors which tilt toward an increased positional adjustment for second basemen. Am I missing something?

OCF's point that Alomar played more than Larkin can be made without reference to WS and WARP. He had 1000 more PAs. Or to look at it another way, if you zero out Alomar's last 2 years, he has more PAs and he is a significantly better hitter than Larkin. The question is whether Larkin makes up the difference, when one accounts for defensive ability and position. I think that he does so easily, but it is not immediately obvious.

I haven't really thought about why the 2B/3B gap closed in the mid-1980's, but the evidence is incontrovertible that it did. My nine-year moving averages of the standard deviation-adjusted batting plus baserunning wins above average per full season played of the worst 3/8 of MLB starters at each position show that replacement 2B were between 0.8 and 1.2 wins a year below replacement 3B in every single season from 1946 to 1975. By 1986, that gap had narrowed to 0.1, and since then it has oscillated between 0.5 and -0.2 (e.g., replacement 2B being 0.2 wins a year *better* than replacement 3B). I guarantee if you look at positional average offense (just check Baseball Prospectus, or the study done by I believe Gassko at THT) you will see the same phenomenon. I thus regard it as an empirical fact that the intrinsic positional value of playing 2B as opposed to 3B shrunk dramatically between 1975 and 1985. Can anyone offer any possible explanations?

I don't want to interrupt this terrific discussion but I want to jot down my thoughts on postseason credit.

I just worked up Walter Johnson. Walter Johnson pitched in two different World Series. In the first World Series he had a cumulative WPA of 0.44 and in the second he had a cumulative WPA of 0.18 (thanks to a dreadful game 7). Add up those cumulative WPA and he gets a 0.62 bonus to his career numbers. Walter Johnson scores 267 points in my system without any postseason bonus, and he clearly doesn't six tenths extra to get inducted but there are borderline candidates that would clearly be helped by a postseason bonus (Dizzy Dean, Tommy Bridges) and in our ranking of pitchers it would help make the finer differences stand out more. The most WPA any pitcher has accumulated is Mariano Rivera and even he doesn't make a huge dent on his career numbers if you normalize to 4 games required for a pennant (divide the 12 games/pennant to normalize to the previous 4 games required) gathering a bonus of less than 4 points.

My method would be as follows.

Sum WPA for a given postseason
Zero out a negative postseasons (no deductions for making the playoffs)
Add all postseason numbers together for total postseason bonus.
Add to spreadsheet with a potential multiplier (<2).

This would make 1 postseason WPA equivalent to 1 career win above replacement.

The difficult part for me right now is I would have to add up all the WPA by hand from Baseball Reference (which has postseason WPA back to 1903 based on retrosheet play-by-play info). A searchable spreadsheet with the WPA data by player/series would make things a LOT faster.

I'm with Mike here. Realistically, it wasn't the Mariners that held Martinez back (and even if it was, that's not REALLY the kind of rationale that merits giving minor league credit); rather, it was Edgar holding Edgar back. He was pretty mediocre in two seasons in AA, and his numbers in AAA should be looked at skeptically, since Calgary had a tendency to make a lot of very marginal hitters look very good. Also, one of the downfalls of not having a defensive position is that you may get blocked for much longer than you would otherwise. There aren't that many DH spots in the league, and a lot of GMs still don't want to have someone who literally can't play a defensive position clogging up the 25-man roster unless they're sure they can make up the difference with their bat.

I'm also firmly of the belief that McGriff is more meritorious than Martinez. Maybe Martinez has an edge in peak offensive numbers, but not enough to make up the difference in defensive value between the two. I really wish that zone data existed for the entire retrosheet era (I suspect there will be someday, but not quite yet); if there was, I'm pretty confident that McGriff's WAR would blow Martinez's out of the water, and that their best seasons would look comparable once you factor defense into the equation.

I decided to personalize it a little bit. The replacement level second basemen from 1975 were indeed vile, and so I looked at one. Denny Doyle. He comes up at age 26 in 1970 and gets 450 PAs, with a 48 OPS+. He follows that with another 800 PAs in 1971 and 1972 with a 66 OPS+. There is no question that a second baseman now with this profile would not have a regular job after that. But, Doyle did, and actually hit a little better.

Maybe the early work of Bill James had some impact by 1985 on some front offices' perceptions of what might constitute acceptable offence from a second baseman (or in places like Texas, where Craig Wright was, they were doing the research on their own, and perhaps, changing their own perceptions).

A minor correction: though he hasn't officially published them, James is now computing "loss shares" for players, and is using them to compute wins above replacement; see, for instance, the article comparing Jeter and Bobby Abreu on Bill James online: http://www.billjamesonline.net/ArticleContent.aspx?AID=637&Code=James01074. I have no idea if he has dealt with the other serious issues with Win Shares, but it is a further instance of the emerging sabermetric consensus on player valuation.

Dan: The reference to WS/WARP was more of a throwaway line to reference the playing time difference. Using different measures with a higher baseline, I still like Alomar quite a bit - as in a significantly better offensive player than any 2B we haven't elected and some we have. The playing time issue is hard to avoid, one way or another, when talking about Larkin.

Mike Green: Go even a few years later and look up the career of Doug Flynn. I remember a Bill James rant in one of his Abstracts -probably after the 1983 season - in which he looked at the RC/(27 outs) for the LF + 2B of several teams. His point was that Flynn's offense was so bad that it essentially negated the goodness of Tim Raines. (Of course that means that all the other teams he looked in that piece had better-hitting 2B than Flynn.) Flynn managed a 1300-game, 4000-PA career with 2B as his majority position and a 57 career OPS+.

Can anyone offer any possible explanations?

Which position moved?
Did 3b get weaker offensively or 2b get better? (relative to other positions)
Or did they converge?

(By "converge", I meant "split the difference")

Rocco's Not-so Malfunctioning Mitochondria--We have Zone Rating data for all of McGriff's career. By combining it with Retrosheet-based metrics like SFR, we can get fantastically high correlations to an average of modern PBP metrics. My system places *a lot* of weight on playing 1B vs. DH--I give DH's zero credit for it supposedly being tougher to hit as a DH, which means that playing a league average 1B is worth nearly a full win a year (0.8-0.9 to be precise) over DH'ing. And even so, McGriff is notably behind Martinez. Here is how it breaks down in my WARP, using standard deviation-adjusted numbers, and translating batting wins above average and 1B replacement level in McGriff's NL seasons to the AL benchmark for consistency (this has no effect on overall totals). Figures may not add up exactly due to rounding.

McGriff



3-year peak: 15.3
7-year prime: 31.1
Career: 49.8


Martinez



3-year peak: 16.5
7-year prime: 35.4
Career: 56.7

As you can see, it's not really close. It's simply not true that Edgar didn't have a defensive position in his youth--he was, in fact, a perfectly fine 3B, who was moved to DH (like Molitor) to keep him healthy later in his career. That also explains why there isn't such a large gap in positional value between him and the Crime Dog--the early years at 3B help cancel out the later ones at DH.

Here's Dan's post #53 sorted by years from best to worst. (Please note: these are Dan's numbers, not mine.) I multiplied the WARP by 10 so I could leave out the decimal points. I also notice that Dan didn't put anything for Edgar's 1987 and 1988; those become the * in the chart below. (1987 must have been at good for something like +0.5, shouldn't it have?)

McGriff 60 59 57 55 47 40 35 35 34 19 16 12 11 7 0 -1 -6
Martinez 66 50 49 48 48 48 45 44 36 35 28 20 5 * * -1 -9[\pre]

You can see for yourself where the differences lie.

Ack! No edit function. Forward slash, not backslash

McGriff  60 59 57 55 47 40 35 35 34 19 16 12 11  7  0 -1 -6
Martinez 66 50 49 48 48 48 45 44 36 35 28 20  5  *  * -1 -9 

You can see for yourself where the differences lie.

I think the big question about Alomar is his fielding. By reputation he was outstanding, but Zone Rating and the Retrosheet defensive stats like SFR (which show exceedingly high correlations to an average of PBP metrics) find him merely average. As an average-fielding modern 2B (so the same positional value as 3B), his hitting and baserunning are just good enough to get him into the bottom 10% of the HoM. If you give weight to the sterling anecdotal reports of his defense, and/or if you compare him to 2B when the position was more valuable (roughly 1925-1985), then he becomes a no-brainer.

I have no idea if Alomar's fielding was, in actuality, closer to his reputation or his stats. That being said, I was [and still am] a Jays fan and avidly followed the Jays during Alomar's time there. Perhaps he had a slow couple steps that made his diving finishes better than they actually were, but his play often amazed me.

The one that I specifically remember is in the '93 playoffs ranging to the first base line [perhaps 30-40 feet behind the bag?] and diving to take away a flair hit; I believe it was in the World Series, but I can`t recall which game. Maybe I should be easier on old-timers who get their game recollections a little confused.

2010 Preliminary Ballot

Point totals are based on a sum of career WAR2, career WAR2 above average, and 5*peak rate (best career rate over five or more consecutive seasons of full time play)
All-Time in-out line is a little bit below 120.

1. Barry Larkin, 186.3. A top 100 player. I fear he will be Trammelled by the writers.
2. Rick Reuschel, 140.1. Way underrated, if BP WARP's assessment of his quality is anywhere near correct. Only eligible player besides Larkin who is not near the borderline.
3. Roberto Alomar. 129.44. Just barely earns the third elect-me spot. Overrated by BP's WARP and definitely within the bottom quintile of the HoM, but clearly worthy.
4. Bert Campaneris. 129.42. Best shortstop of his era. Could easily rank ahead of Alomar.
5. Phil Rizzuto. 127.9. Great defensive shortstop, and enough offense to go with it.
6. Dave Concepcion. 127.7. Ditto.
7. Luis Tiant. 126.5. Underrated because of overemphasis on single-season workloads in many voters' rankings of pitchers. Not an innings eater, but still of at least average durability for his era, and far above average in effectiveness.
8. Gavvy Cravath. 129.3. MLE credit beginning in 1906. Could rank higher, but I'm being a little conservative in ranking early players against modern players.
9. Johnny Pesky. 123.9. Needs a lot of war credit to get here, but it sure looks like he deserves it.
10. Fred Dunlap. 125.1. Scaling BP's pre-1893 WARP2 to Dan R's WAR shows Dunlap as worthy. Better than Richardson or McPhee among contemporary second basemen. Arguably the second best player in baseball 1880-85 after Big Dan Brouthers.
11. Urban Shocker. 123.3. Steadily excellent in a tragically shortened career. No extra credit for what might have been in these numbers, however.
12. Edgar Martinez. 122.7. Great hitter; best pure DH of all time. That's just enough to get him over the line without getting into MLE credit.
13. Tommy John. 122.2. He doesn't have a great peak, but his 12-year prime is as good as any eligible pitcher's, and he adds another 4-6 good years on top of that. A lot better than Kaat, better than Sutton, whom we elected.
14. Buddy Bell. 122.16. A half step behind the deservedly elected Nettles, to whom he is otherwise very similar.
15. David Cone. 121.7. Excellent peak, and just enough career added by several effective seasons after coming back from arm injuries.

Next Five. All will be PHOM when I finish constructing it sometime this year.

16. Dave Bancroft. 122.1
17. Tommy Leach. 121.7
18. Don Newcombe. 121.2
19. Burleigh Grimes. 121.2
20. Bob Johnson. 120.2

Other Notable newcomers.

35. Robin Ventura. 115.5. Just a little below the in-out line. I can see how a system calibrated even slightly differently from mine would show him as worthy.
82ish. Fred McGriff. 104.2. More than a little below the in-out line. Still an outstanding player, but as Dan R's WAR shows, he was not much better than an average player outside of his peak, which wasn't all that high. He's in Harry Hooper territory. Among first basemen, a little better than Jack Clark, a little more better than Tony Perez. Probably has as good a chance as Barry Larkin and Tim Raines of being elected to Cooperstown.

Out of the Top 100.

Juan Gonzalez. 86.0
Ellis Burks. 84.6

Jim Rice scores 88.5 in my system; Mark Grace scores 86.6. Gonzalez and Burks are in that value neighborhood. Very good players, but no one with sense would mistake them for Hall of Famers.

Did I say 15.3 for McGriff's 3-year peak? I meant 17.6. Not sure how I screwed that up.

Dan, on the same scale, what was Will Clark's 3-year peak? (Same position, nearly the same time.)

20.6 (8.6 in 1989, 6.7 in 1988, 5.3 in 1991).

That's all? I would have thought he'd be further out in front.

Conversely: if Mcgriff has 17.6 to Clark's 20.6, then maybe we can say that McGriff had a pretty good peak after all.

(My own stuff does show a bigger gap than that.)

McGriff's top 3 years:




Clark's top 3 years:



Just using my standard equations, these BWAA numbers suggest that McGriff's OPS+ in his three best overall years "should" have been 163 in 1989, 151 in 1994, and 150 in 1988, while Clark's "should" have been 176 in 1989, 160 in 1988, and 152 in 1991. In fact, McGriff's figures were 166, 157, and 157, so in fact my numbers are less friendly to his peak than his "eyeball" offensive stats, while Clark's were 175, 160, and 154, so right on.

One thing that might help explain it is that the numbers show McGriff as having paired his two best defensive seasons ('88 and '94) with two of his top hitting seasons. Clark did the same in '89, but he only has one of those years to McGriff's two. Then again, McGriff is hurt by baserunning in two of his top years, while Clark is helped (leaving that ghastly '87 out of it)...

I'm not sure why your numbers would differ substantially from mine here, OCF. Clark's '89 was obviously outstanding and his '88 was very good, but he didn't have a third Great Year to really differentiate himself.

In response to jimd, I have Ventura at +15.7 fielding wins, which is nothing short of brilliant. However, I still don't think he has enough offense or career length to get in. He's the same profile as Brooks, Nettles, and Bell, but below all three of the them, and we've set the in/out line after Graig. I would, however, have given him the 1999 NL MVP.

DanR,

If the positional adjustments for 3B relative to 2B have increased since 1985, does that not change the Ventura to Brooks/Nettles/Bell comparison? It does seem a stretch to say that Ventura had much more defensive value than Alomar, but that seems to be the case. Where Ventura suffers is in career length. His prime from 91-99 is pretty impressive though, and perhaps better than that of the other three if one accepts the positional shift.

Incidentally, the worst 3/8 of major league starters is a high standard for replacement level. You may be looking for the WAR of the worst 3/8 of innings by second basemen, but that is something different, and more difficult to calculate. To give an example, in 2008, Robinson Cano might have been within the bottom 3/8 of major league starters at second base. He really is not relevant to the discussion. You want to look at part-time players like Sean Rodriguez instead, or alternatively look at the bottom 20% (say) of starters.

Incidentally, the worst 3/8 of major league starters is a high standard for replacement level.

Using a high standard for replacement level is necessary, IMO, for this context - otherwise you don't get enough of a degree of separation between the best players and the HoVG. You're trying to identify the best players in an arena where the difference between best and next-best may be very small.

-- MWE

Mike Green--No, 3B didn't really move much on the spectrum. The main changes from 1978 to now are that both middle infield positions have gotten much easier and catcher has gotten much harder.

Ventura indubitably had much more defensive value than Alomar--by a TON if you trust the numbers.

I don't define the bottom 3/8 as replacement level. I use the bottom 3/8 as an anchor to track the evolution of the defensive spectrum over time (rather than the average, which is skewed by star gluts and droughts), and keep the gap between the bottom 3/8 and the replacement level constant. For example, the worst 3/8 of corner outfielders averaged 0.5 standard deviation-adjusted hitting + baserunning wins below average per 162 games from 1985 to 2005, but according to Nate Silver, Freely Available corner outfielders were 1.0 standard deviation-adjusted hitting + baserunning wins below average per 162 games from 1985 to 2005. Thus, my corner OF replacement level in any given year is equal to the average performance of the worst 3/8 of starting MLB corner OF in the surrounding nine-year period, measured in standard deviation-adjusted hitting + baserunning wins above average per 162 games, minus 0.5.

I don't agree with Mike Emeigh. Your replacement level merely establishes your preference for rate vs. playing time. If you take that too far, you'll wind up inducting someone like Mike Donlin. One question is how big your Hall is, another is how much you want to reward quality vs. quantity. They are entirely separate.

I always liked quality. I always thought the HoF and HoM are for "great" players, and you can't become great just by being good for a long time. Any methodology that operationalizes that idea, I like it.

He was pretty mediocre in two seasons in AA

That would be the factor that I think y'all should consider wrt Edgar and the minors. If you just extrapolate his MLB record backwards, à la Monte Irvin, Minnie Minoso, or someone like that, he will look great. But he hit .258 at age 22 in a full AA season, and then .264 at age 23. That can't add up to much if translated into major-league performance.

sunnyday, then where is Frank Chance on your ballot? If I recall correctly, you tend to value in-season durability quite highly, which would mean you would advocate a LOW replacement level combined with a strong peak preference.

Bob "Jugement" Dernier--no one's suggesting Edgar should be credited for those seasons. The issue is whether he deserves it for 1987 and 1988, where his numbers would MLE to major league caliber.

Dan - Thanks for running those numbers. Guess my theory doesn't hold up that well, although I wouldn't mind seeing a different set of translations on the defense. I have a tough time believing that McGriff's defense was below replacement level his entire career, and ZR (without additional information) isn't necessarily the best means of measuring a player's defensive worth, particularly for 1Bs.

I wasn't necessarily advocating McGriff on the ballot either, I just don't see that strong of a case for Martinez.

Also, when I initially said that Edgar wasn't great in AA, it wasn't to say that he shouldn't get MLE for those years (of course he shouldn't), but that they would have given Mariners management good reason to be cautious about bringing him up even after mashing in AAA, considering that pretty much anyone who was a AAAA player could mash in Calgary. It was like Vegas or Albequerque today - numbers from a park that hitter-friendly should be taken not with a grain, but with a pound of salt.

Responding to Mike Green re comparing Ventura to Robinson, Nettles, and Bell:

If the positional adjustments for 3B relative to 2B have increased since 1985, does that not change the Ventura to Brooks/Nettles/Bell comparison? It does seem a stretch to say that Ventura had much more defensive value than Alomar, but that seems to be the case. Where Ventura suffers is in career length.

Dan R has already addressed the changing positional adjustments issue. I will add that career length matters a great deal here, because Robinson and Nettles were definitely elected on the basis of career, not peak. If Ventura, during his prime was very similar to Robinson and Nettles, but his career was shorter, then that might well put him below the in-out line, because if the productive careers of Robinson and Nettles had been shorter, they might well have fallen below the in-out line also. (It's important to note that both Robinson and Nettles had unproductive seasons at the ends of their careers that don't add to their case, but even without those seasons they had longer and more valuable careers than Ventura).

Responding to Bob "Jugement" Dernier -- when the HoM looked at Irvin and Minoso, we (at least, most voters) did not extrapolate their MLB records backwards. Instead, we used translations of their Negro League and minor league play into a major league context. If Edgar Martinez is given credit for minor-league play, it would be on that basis as well (at least, for most voters).

Catcher has gotten much harder, relative to the middle infield, than it was in 1968 or 1988? Really?

Base-thievery is way down. There are no more knuckleballers than there were before. It is true that Ks are up and balls in play are down, at least relative to 1988, but if you tell me that catchers have more positional value relative to shortstops or second basemen now as compared to 1968, that is something that I ought to check.

I have a tough time believing that McGriff's defense was below replacement level his entire career

He wasn't. You are misinterpreting the numbers in Dan's chart. "Replacement level" indicates the number of wins below average that a replacement player at that position would have been, if he had played the same fraction of a season that McGriff did that year. Every player's "replacement level" column in Dan's WAR is negative, and that negative amount is subtracted from the player's wins above average to determine his wins above replacement. Dan's numbers show McGriff as having been a slightly above average defensive first baseman for his career.

It looks to me like Ventura's prime is considerably ahead of that of Robinson and Nettles. His offence was definitely noticeably better.

Rocco's Not-so Malfunctioning Mitochondria--Who said "McGriff's defense was below replacement level his entire career?" I don't even know what that means. As I (and Tangotiger) repeat ad infinitum, there are no replacement statistics, only replacement players.

McGriff's fielding was three wins worse than a league average defensive first baseman for his whole career. Since replacement DH's hit at the league average, you can determine total defensive value (the analog in my system to Fielding Win Shares) as the sum of FWAA and Replc (with the Replc sign reversed). That would give McGriff 9 wins of defensive value (a replacement player at his mix of positions and playing time would have been 11.8 wins below average for his career, and McGriff's fielding was 2.8 wins below average), and Martinez 7.6 wins of defensive value (a replacement player at his mix of positions and playing time would have been 6.5 wins below average for his career, and Martinez's fielding was 1.1 wins above average). The basic conclusion to draw is that a guy who spent 30% of his career as a slightly above-average 3B and 70% as a DH has nearly as much defensive value as a guy who spent 90% of his career as a slightly below average 1B and 10% as a DH.

The fielding numbers aren't just Zone Rating--they are the mix of Zone Rating, Simple Fielding Runs, and TotalZone that produces the highest correlation to a weighted average of UZR, Plus/Minus, and PMR for the 2004-07 period. I think it is safe to say that they measure 1B *range* pretty damn well. I don't know where to find Rallymonkey's 1B scooping numbers on the Internet anymore, but I don't remember McGriff being notably good or anything.

Mike Green--yes, *substantially* so. From the dawn of the live ball era until about 1985, catcher was a mid-spectrum position, similar to modern-day 2B/3B/CF. Since then, the hitting + baserunning of the worst 3/8 of MLB starting C has been declining rapidly and steadily, to the point where it is now (in my system) equal to SS and in Tango's system significantly harder still. Again, I'm sure you would see the same phenomenon in positional average offense. I'd be eager to hear potential explanations.

Chris Cobb--not so! I have replacement first basemen as precisely 0.026 wins above average in 1972, the year before the DH created 12 new jobs for them and brought it back down into negative territory. And my numbers show McGriff as slightly BELOW average defensively at 1B (2.8 wins to be exact, ignoring sub-replacement seasons).

One possible explanation for declining performance by replacement-level catchers is the increase in stolen base efficiency over time. Although stolen bases and attempts are lower now than in 1968, efficiency is way up. That process was well underway by 1985. I do remember that in the late 1960s, there was not much discussion about the efficiency side of the running game, but that this did change beginning in the 1970s. The influence of early sabermetrics might also have been felt there.

The basic conclusion to draw is that a guy who spent 30% of his career as a slightly above-average 3B and 70% as a DH has nearly as much defensive value as a guy who spent 90% of his career as a slightly below average 1B and 10% as a DH.

The problem as I see it is you are saying McGriff was -3.6 wins worse than league average, but Martinez is 0.9 wins _above_ league average. Martinez gets zeroes for most of his career, presumably because you are comparing him to other DHs. That doesn't make sense to me -- in essence crediting Martinez compared to McGriff because he didn't take the field.

If the average player saves 40 runs defensively in a year (or whatever), Martinez should be penalized at -40 for every season he was a DH (prorated for the number of games played).

The problem as I see it is you are saying McGriff was -3.6 wins worse than league average, but Martinez is 0.9 wins _above_ league average. Martinez gets zeroes for most of his career, presumably because you are comparing him to other DHs. That doesn't make sense to me -- in essence crediting Martinez compared to McGriff because he didn't take the field.

Check out the replacement levels they're being compared to. Martinez goes up against a replacement level of league average while he's a DH; McGriff seems to gain 7-10 runs a year on him there. That's where Dan factors in positional adjustments.

Although stolen bases and attempts are lower now than in 1968, efficiency is way up.

The "although" is awkward

it should be

Efficiency is way up because stolen bases and attempts are lower now than in 1968.

or

Although stolen bases and attempts are lower now than in 1968, the running game is more effective now since efficiency is way up.

DanR #53
Total Career BWAA=52.0 for Edgar (2.0 is a typo error).
The difference from TX version BWAA=52.5 (correctly given) is 0.5, his net value as a batter during 1993 and 2004, the two seasons that TX discards.

It's interesting that 1989 is no discard despite OPS+=74, thanks to good fielding in 61 games at thirdbase.

In 1987-88 he played 27 games including 25 fielding 3B. Why pass over those two years?
(By the way, should we count September call-ups full weight?
Has anyone estimated the league quality beginning -09-01 or -09-11 or -09-21, for any epoch in mlb history?)


DanR #62
Just using my standard equations, these BWAA numbers suggest that McGriff's OPS+ in his three best overall years "should" have been 163 in 1989, 151 in 1994, and 150 in 1988

What do you mean, Dan?
Does the "standard equation" use non-pitching SFrac to express BWAA as an index for all non-pitchers, with BWAA=0 expressed as 100?


77. David Concepcion de la Desviacion Estandar (Dan R) Posted: January 21, 2009 at 12:02 PM (#3057012)
Mike Green--yes, *substantially* so. From the dawn of the live ball era until about 1985, catcher was a mid-spectrum position, similar to modern-day 2B/3B/CF. Since then, the hitting + baserunning of the worst 3/8 of MLB starting C has been declining rapidly and steadily, to the point where it is now (in my system) equal to SS and in Tango's system significantly harder still. Again, I'm sure you would see the same phenomenon in positional average offense. I'd be eager to hear potential explanations.

The "obvious" hypothesis is that mlbclubs finally responded to the return of base-thievery by jacking up the standard at pitcher (no more Nolan Ryans and Mark Clears) and catcher. I suppose that it happened partly by scouting and partly by training.

George Case and Jackie Robinson. Luis Aparicio, Maury Wills, and Lou Brock. Rickey Henderson, Tim Raines, and Vince Coleman. The number of good thieves keeping company with the marquee names swelled. For some players and some teams, base-stealing frequencies and success rates grabbed attention. At the same time some clubs figured out how they could put a blanket on it, by scouting and coaching partly for defense against the running game.


At second and third, the standard would have increased if positioning and training helped make adequate fielders out of more bigger and stronger players; if training, tactical, and technological adjustments helped more of the traditional body types to catch up as batters; or if sabrmetric understanding led teams to put greater weight on batting.

Today is trite day.

The phrasing was awkward, but the substance was right (door #2, then). Stolen base attempts and successes could be at lower levels than in 1968, but the running game could be at comparable levels of efficiency. This was the case in the 1950s, for instance.

when the HoM looked at Irvin and Minoso, we (at least, most voters) did not extrapolate their MLB records backwards. Instead, we used translations of their Negro League and minor league play into a major league context

Thanks, Chris. I seem to remember this issue raised in some discussion of Jackie Robinson (though it may not even have been a HOM thread), where the vexed question was what to do with Robinson, who didn't play very much baseball before his first year in the NeL (and didn't play well at UCLA, at that). Almost every such player could have been an ML star if given a chance. What's odd about Edgar is that he was so mediocre in AA, at a typical age for AA non-star prospects.

Check out the replacement levels they're being compared to. Martinez goes up against a replacement level of league average while he's a DH; McGriff seems to gain 7-10 runs a year on him there. That's where Dan factors in positional adjustments.

Okay, thanks for explaining that to me, but I am still a bit confused. Martinez is 9 wins better than McGriff with the bat. Okay, I can accept that (more or less; I don't think you should penalize a player for poor seasons, but that's another matter). But when you factor in baserunning and defense, the gap only narrows to 7 wins? That just seems wrong to me. By definition, Martinez was a below-average fielder; he didn't field at all!

53. David Concepcion de la Desviacion Estandar (Dan R) Posted: January 20, 2009 at 05:48 PM (#3056450)
My system places *a lot* of weight on playing 1B vs. DH--I give DH's zero credit for it supposedly being tougher to hit as a DH, which means that playing a league average 1B is worth nearly a full win a year (0.8-0.9 to be precise) over DH'ing.

Is that zero credit because you don't yet believe the finding, the jury is out?
Or do you believe the finding should not be applied in that way and in this context?
--in "hall of fame conversation", I suppose

oops, I should have asked that next door in Dan Rosenheck's thread

I don't think you should penalize a player for poor seasons, but that's another matter

I agree. A Bill James comment in regard to this was a major factor why I abandoned my initial love of Pete Palmer's Linear Weights years ago. Hypothetically, an inner-circle HoMer could negate every great season he had this way if a GM was stupid enough to keep him in the lineup after his effectiveness disappeared. That doesn't make any sense to me.

Zero value I can handle (though my system gives inconsequential value for subpar seasons), but not negative value.

Since then, the hitting + baserunning of the worst 3/8 of MLB starting C has been declining rapidly and steadily, to the point where it is now (in my system) equal to SS and in Tango's system significantly harder still.

My guess is that enough teams actually looked at and/or were aware of studies showing that good hitting young catchers do not develop as hitters or last as long as good hitters as comparable hitting 1Bs, OFs, 3Bs, SSs, etc... the result was that if you had a good young hitter who was reasonably athletic- organizations might think of moving him out from behind the plate in order to maximize his development as a hitter.

Hypothetically, an inner-circle HoMer could negate every great season he had this way if a GM was stupid enough to keep him in the lineup after his effectiveness disappeared.

Pete Rose says "hi"
seriously, some statheads were of the opinion that Rose was not a HOFer (you know what I mean) because many of his later seasons actually had negative value- dragging down his overall wins above replacement/average/ whatever.

I don't doubt that Pete was a HOFer (if he wasn't an ineligible gambling fool), but I don't think his HOF (or HOM) worthiness was impacted by his crappy last few seasons- but if Maris' 61 homer mark deserved an asterisk then so did Pete's all time hit record.

A Surfeit of Peaches Graham--Eric J's comment is correct. A 1B would have to be about 8-9 runs below average per year with the glove before he would have less value than a DH with the same number of PA in my system.

This is the key misconception--Martinez did field at all! He spent a third of his career playing third base at a slightly above average level. The -6.5 in his Replc column means that a replacement third baseman would have been 6.5 wins below average over the games that Martinez played at third base (plus the 0.0 wins below average that a replacement DH would have been over the games Martinez played at DH). If Martinez were truly a full-career DH, then his career value would have been roughly equal to McGriff's: Martinez was nine wins better with the bat, and McGriff has nine wins of total defensive value (12 for playing 1B, minus 3 for playing it at worse than league average). But he wasn't. He was a 30% above-average 3B/70% DH, and that is worth nearly as much as a 90% below-average 1B/10% DH.

How do I penalize players for poor seasons? I explicitly discard below-replacement years in the TXBR row.

Paul Wendt--yes, that's a typo, nice catch.

I only count seasons with at least 50 PA to keep my spreadsheet manageable. Doing them manually, here are the missing lines for Edgar:



That brings his career total to 55.9 and his TXBR to 57.0.

My "standard equation" (which I think was calculated by Jim Sp, actually) is OPS+ = 11.43*BWAA/SFrac + 90.8 for non-DH years, and 11.43*((BWAA/SFrac)+.6) + 90.8 for DH years. It's just a way to re-state BWAA on the familiar OPS+ scale, and to spot-check where my BWAA diverge substantially from a player's OPS+ (thus providing significant/useful additional information).

It is zero credit because Nate Silver finds that replacement DH's hit at roughly the league average. Regardless of whether a full-time DH is a "true" +3 win hitter who is reduced to +2.5 by the "difficulty" of DH'ing, or whether he is a "true" +2.5 win hitter who is unaffected by this phenomenon, he is still 2.5 wins above what a freely available minor league veteran would produce in the same playing time.

Eric J's comment is correct. A 1B would have to be about 8-9 runs below average per year with the glove before he would have less value than a DH with the same number of PA in my system.

Right. It's just that that seems low. How do you calculate the number? I can think of a few ways, all defensible. In order from most to least penalizing:

1) Calculate how many additional runs a team would give up if it played with only three infielders. Penalize a DH some fraction of those runs (50% maybe?)

2) Figure out who the worst fielder in the majors is, and set DH defense to that level.

3) Set DH defense at the tenth percentile of all major leaguers.

Arguments can be made for all three, but I would be loath to be more generous than method 3).

According to Nate Silver, freely available replacement 1B are 8.7 runs below league average per year (-10.8 hitting, -0.3 baserunning, +2.4 fielding), while freely available replacement DH are 0.5 runs below league average per year (-0.4 hitting, -0.2 baserunning, and some rounding), for a gap of 8.2 runs. This empirical replacement level is what determines actual value to a team--if you want a guy who can play first base, you give to give up 8.2 runs of hitting + fielding compared to a guy who can only DH.

http://www.baseballprospectus.com/article.php?articleid=4891

Is LF replacement value above or below 1B?

Replacement corner outfielders are significantly worse hitters + fielders than replacement 1B, as you can see in the above-linked article or in my StDevs and Rep Levels.xls spreadsheet.

Pete Rose says "hi"
seriously, some statheads were of the opinion that Rose was not a HOFer (you know what I mean) because many of his later seasons actually had negative value- dragging down his overall wins above replacement/average/ whatever.

I don't doubt that Pete was a HOFer (if he wasn't an ineligible gambling fool), but I don't think his HOF (or HOM) worthiness was impacted by his crappy last few seasons- but if Maris' 61 homer mark deserved an asterisk then so did Pete's all time hit record.

I agree with all of this, JPWF13.

Thanks for the tutorial guys. I've lurked here long enough that one would think I'd know how to read the charts by now.

I guess Martinez was better than I ever remember him being. I have trouble seeing him as an HOM or Hall type, but I guess when you have 70 or so guys in the HOM who are deserving rather than some of the scuzz in the Hall, then it gets a lot of very good players closer to the fringe, if you're prone to ignore the really bad Hall choices when determining your average.

Rusty Priske has Larkin ahead of Alomar,

No I don't. I have Alomar first and Larkin second on my prelim ballot.

According to Nate Silver, freely available replacement 1B are 8.7 runs below league average per year (-10.8 hitting, -0.3 baserunning, +2.4 fielding), while freely available replacement DH are 0.5 runs below league average per year (-0.4 hitting, -0.2 baserunning, and some rounding), for a gap of 8.2 runs. This empirical replacement level is what determines actual value to a team--if you want a guy who can play first base, you give to give up 8.2 runs of hitting + fielding compared to a guy who can only DH.

I think I've figured it out. You're treating DH as a position. I don't really see DH as a separate position. Under your system if you had two identical players with poor fielding skills, and one played the field and the other was a DH, player 1 would rank higher. Whereas I would look at it differently. I would rate a DH as the equivalent of Helen Keller in the field -- i.e., he made no plays, and therefore saved no runs, while even a player who dropped every routine popup would get some fielding credit, because he at least would prevent some triples and inside-the-park home runs.

More to the point, if you treat DH as a position, then of course Martinez is in. But that obviates the comparison between McGriff and Martinez. You've got what amounts to a quota.

Edit: Above, I mean player 2 would rank higher.