Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Hall of Merit > Discussion
Hall of Merit
— A Look at Baseball's All-Time Best

Sunday, July 29, 2007

Eddie Murray

Eligible in 2003.

John (You Can Call Me Grandma) Murphy Posted: July 29, 2007 at 02:22 PM | 98 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

   1. John (You Can Call Me Grandma) Murphy Posted: July 29, 2007 at 02:25 PM (#2460150)
Which season was his best? Damned if I know.

BTW, he's an easy HoM selection.
   2. John DiFool2 Posted: July 29, 2007 at 02:47 PM (#2460162)
Odd relationship with the press. At the beginning of his career as he started to gain the "clutch" moniker the sportswriters worshipped him. Then he stopped talking to the press and they sort of turned on him, implying several unsavory (but completely unprovable) things behind his back, as sportswriters are wont to do when a high- profile player refuses to play their little games. Later on tho near the end of his career he was back to being revered again.
   3. Yeaarrgghhhh Posted: July 29, 2007 at 02:55 PM (#2460170)
When I was a kid in the early and mid 80s, my best friend was a Yankees fan, and we used to get into endless arguments about Murray vs. Mattingly. Too bad I didn't have OPS+ or any other advanced stats to help my case...it was hard to overcome Mattingly's BA and RBIs.

Still my favorite player of all time.
   4. OCF Posted: July 29, 2007 at 03:07 PM (#2460177)
Had the highest batting average in all of baseball in 1990 but wasn't recognized as a batting champion. Most of you understand that; if you're sitting there scratching your head wondering how that's possible, I won't spill the beans - I'll let you have the pleasure of figuring it out for yourself.

My impression when he came to the Dodgers was that he'd been run out of town in Baltimore and that Baltimore fans - or maybe it was just the local press? - had a "good riddance" attitude about that. Frankly, I found that very hard to understand. Of course, I am talking about the same things as in John DiFool's post #2.

As for his quality: he's got a very strong career case, but he doesn't have any really standout years. Will Clark had a much higher peak than Murray. Of course, Murray kept going and going for a thousand more games than Clark.
   5. Pasta-diving Jeter (jmac66) Posted: July 29, 2007 at 03:11 PM (#2460179)
Which season was his best? Damned if I know.

OPS+ from 78-85

140
131
138
156
156
156
156
149

I wonder if ANY player (let alone any great player) ever had 4 years in a row with OPS+ within 3 points

according to Bill James in an early abstract, this is why Murray would never win the MVP--he was so mind-numbingly consistent, that he was just doing what was expected of him
   6. OCF Posted: July 29, 2007 at 03:22 PM (#2460193)
In my scaled RCAA system, the same 8 years as in jmac66's post come out as

38
35
40
31
50
58
61
53

Note that the 31 is for the 1981 strike year. In this system, his top year was actually his 1990 with the Dodgers, at 62.

I mentioned Will Clark. This system has his 1988 and 1989 at 81 and 99 (up there in Norm Cash 1961 territory.)
   7. Shooty Survived the Shutdown of '14! Posted: July 29, 2007 at 04:08 PM (#2460216)
I loved Murray's baseball cards. Even as a wee tike, I knew there was something special about the consistency of those numbers on the back. Steady freakin Eddie! Has there ever been a more appropriate nickname?
   8. Dan The Mediocre Posted: July 29, 2007 at 04:15 PM (#2460218)
Had the highest batting average in all of baseball in 1990 but wasn't recognized as a batting champion. Most of you understand that; if you're sitting there scratching your head wondering how that's possible, I won't spill the beans - I'll let you have the pleasure of figuring it out for yourself.


That is an interesting situation, but we usually dislike the AL/NL stat separation for players that get traded because they hurt the player that was traded, not enable him to win a batting title that would have otherwise gone to someone we like.
   9. The Clarence Thomas of BBTF (scott) Posted: July 29, 2007 at 05:06 PM (#2460239)
the reverse happened to Placido Polanco a couple years ago if i remember correctly. lead the league with a .331 combined batting average but didn't have enough AB in either league to qualify.
   10. The District Attorney Posted: July 29, 2007 at 05:49 PM (#2460332)
Will "Norbit" hurt his chances?
   11. Shooty Survived the Shutdown of '14! Posted: July 29, 2007 at 05:55 PM (#2460362)
Will "Norbit" hurt his chances?

If Best Defense and The Golden Child didn't slow him up, nothing will.
   12. John (You Can Call Me Grandma) Murphy Posted: July 29, 2007 at 06:01 PM (#2460395)
Murray is the Scottish variant of Murphy, BTW. :-)
   13. vortex of dissipation Posted: July 29, 2007 at 06:38 PM (#2460534)
Michael Young was AL batting champion at .3308; Polanco's totals with Philadelphia (173 PA) and Detroit (378 PA) gave him an average of .3313. He didn't have the highest ave in baseball, though, because Derrek Lee hit .335 for the Cubs.
   14. The Clarence Thomas of BBTF (scott) Posted: July 29, 2007 at 07:00 PM (#2460613)
ah, right. my bad. i forgot Lee, i knew that he was a touch better than Young.
   15. 47YOUNEVERKNOW47 Posted: July 29, 2007 at 07:41 PM (#2460710)
I saw Eddie at least 100 times on his way into Memorial Stadium when I was a kid hanging out in search of autographs. I'd usually just wish him good luck unless I had something special to get signed, and the rudest thing the man ever said to me was "Not today, sir"

I still hate what the Baltimore media did in running him out of town. Many fans aren't innocent either. I'm talking about the dozens and dozens who called talk shows suggesting that the O's were better off with Jim Traber at 1B, or that Larry Sheets would certainly make up for Eddie's lost production.

The O's owner mouthed off about Eddie during a rain delay in the 1988 (?), implying a lack of hustle and the media just ran with it. I never belied it to be true, and it's worth noting that the paragon of baseball's spartan hard-working virtue -Cal Ripken, Jr., himself- always cites Murray as the man who taught him how to be ready on a daily basis to give his best effort to the team.

The coolest thing about Eddie to me was his "JR" necklace. No, he's not Eddie Clarence Murray, Jr. or anything like that. He said it stood for "Just Regular". Great, great player with that coiled cobra stance from either side. EDD-IE!!! EDD-IE!!!
   16. Mark Shirk (jsch) Posted: July 29, 2007 at 10:37 PM (#2460907)
I guess that means that Eddie the baseball player was Scottish and Eddie the comedian was Irish...oh

Still, I think that he will top my ballot (over sandberg), many 28-33 WS years which adds up to a pretty nice peak.
   17. Devin has a deep burning passion for fuzzy socks Posted: July 29, 2007 at 10:39 PM (#2460911)
Played more games at 1B than any other player in baseball history (Yesterday's Mets trivia question, of course I got it wrong.) I didn't write down the numbers, but the top 5:

1. Eddie Murray
2. Jake Beckley
3. Fred McGriff
4. Mickey Vernon (my guess)
5. Mark Grace

I don't know if it was deliberate*, but it certainly helped that when he left the Orioles, he went to the NL and couldn't be a DH.

(*That he wanted to play in the field, not that he wanted to set the record.)
   18. BDC Posted: July 29, 2007 at 10:59 PM (#2460930)
Played more games at 1B than any other player in baseball history

If Sean would put up fielding leaderboards, we could check in an instant, please Sean :)
   19. Dr. Chaleeko Posted: July 29, 2007 at 11:07 PM (#2460936)
Why are we talking about the old Detroit placekicker?
   20. Dr. Chaleeko Posted: July 30, 2007 at 01:24 AM (#2461036)
Let's say that you wanted to form up a team of Eddie Murray kind of players.
-Long careers
-Very consistent
-But not huge peaks
-But not low peaks either

C: Bill Dickey
1B: Murray
2B: Charlie Gehringer
3B: Chipper Jones? Or Darrell Evans?
SS: George Davis (SSs just don't seem to have the consistency as a rule)
RF: Tony Gwynn
CF: Alejandro Oms?
LF: Billy Williams
DH: Hal Baines
SP: Eddie Plank
RP: Lee Smith

You'd win some games with that squad.
   21. OCF Posted: July 30, 2007 at 04:08 AM (#2461151)
The second of a cluster of players who grew up in the African-American neighborhoods of the south part of Los Angeles. Those neighborhoods don't produce baseball players any more, and those high schools may well be majority Latino now. Here's the cluster:

Ozzie Smith: Born Dec. 1954, career 1978-1996.
Eddie Murray: Born Feb. 1956, career 1977-1997
Darryl Strawberry: Born March 1962, career 1983-1999
Eric Davis: Born May 1962, career 1984-2001

If you look up birthplaces, you'll see that Smith was born in Alabama, but he grew up in Los Angeles. Smith went to college and was drafted out of Cal Poly SLO; the other three were drafted out of high school. None of the four spent particularly long in the minor leagues.
   22. Alex meets the threshold for granular review Posted: July 30, 2007 at 04:19 AM (#2461154)
#20 - Chipper Jones looks very out of place on that list. He went 146-175-142-162-155 OPS+ at third base, and is now in a 147-157-169 stretch. Of course, this does depend on some extent how you view Larry's fielding, but Dial, for example, has done work that show that conventional metrics are way off on his glove due to the makeup of the Braves staffs he played behind.

Darrell would be a good choice.
   23. Alex meets the threshold for granular review Posted: July 30, 2007 at 04:20 AM (#2461156)
If I could edit, I would add the caveat that the latter stretch has playing time issues. And that the first stretch had a 141 after that which should certainly be included. Oh well.
   24. vortex of dissipation Posted: July 30, 2007 at 04:53 AM (#2461179)
3B: Tony Perez
   25. thok Posted: July 30, 2007 at 07:16 AM (#2461233)
More fun with junk stats

Eddie Murray is 4th in career putouts (behind Beckley, Anson, and Kotcheny) and 3rd in career total chances (behind Beckley and Anson).

Does it feel strange to anybody else that Murray has more career games at first than Gehrig?
   26. DL from MN Posted: July 30, 2007 at 02:02 PM (#2461312)
I have him between Greenberg and McCovey among 1B. My 1B list is Gehrig, Foxx, Anson, Connor, Brouthers, Greenberg so that's pretty high up the list.
   27. DavidFoss Posted: July 30, 2007 at 02:30 PM (#2461336)
Had the highest batting average in all of baseball in 1990 but wasn't recognized as a batting champion. Most of you understand that; if you're sitting there scratching your head wondering how that's possible, I won't spill the beans - I'll let you have the pleasure of figuring it out for yourself.

Interleague trade to a contender on Aug. 29th. Does this mean the eventual batting champion cleared waivers? Or were the rules different back then.
   28. Chris Cobb Posted: July 31, 2007 at 12:00 AM (#2462202)
Great, great player.

I have him behind Gehrig, Anson, Foxx, and Connor, just ahead of Mize and Brouthers among first baseman through 2003. Jeff Bagwell and Frank Thomas probably goes ahead of him on peak among the great 90s first basemen, but he's still top 10 all-time.
   29. Dr. Chaleeko Posted: July 31, 2007 at 12:10 AM (#2462215)
my keltner-based method shows murray as a top-ten guy

Gehrig
Brouthers
Connor
Thomas
t-Leonard
t-Mize
t-Anson
Bagwell
Foxx
Murray

Foxx gets dinged a bit for being in Gehrig's and Ruth's league. Mize includes war credit.
   30. David Concepcion de la Desviacion Estandar (Dan R) Posted: July 31, 2007 at 12:23 AM (#2462238)
Yeah, that ranking of Foxx just looks like a typo, Eric. If that's what your system has to say about Double X, I'm surprised your system would allow you to vote for this year's shortstops at all, given that they happened to play at the same time as Ripken and Yount. Foxx is far and away the 2nd best 1B of the century, and I don't see anyone else close. Pujols could give him a run for his money, though.
   31. David Concepcion de la Desviacion Estandar (Dan R) Posted: July 31, 2007 at 12:43 AM (#2462268)
I thought I'd like Murray more, given my appreciation for the low-stdev 1980's. But the only season post-'85 that my system is particularly impressed by is 1990. 115-130 OPS+'s from guys at the wrong end of the defensive spectrum--with notably poor non-SB baserunning, no less--are just not tough to come by. So you're left with the eight straight All-Star caliber years from '78 to '85, plus 1990, none of which could conceivably be described as MVP-caliber. That's a nifty prime, to be sure, but it's nothing to make you go wow. Murray is definitely a deserving HoM'er, but I have him virtually even with the leading backlog shortstops. I'll put him above them (which probably means an elect-me spot) merely as a nod to consensus.
   32. AJMcCringleberry Posted: July 31, 2007 at 01:28 AM (#2462322)
My top 11

Gehrig
Anson
Foxx
Connor
Bagwell
Brouthers
(Thomas)
Murray
Palmeiro
McCovey
Clark
   33. Dr. Chaleeko Posted: July 31, 2007 at 01:51 AM (#2462337)
We've been over Foxx before, he's better than my system shows because he's got Gehrig and Greenberg to deal with at his own position and Ruth and Gehrig and Greenberg to deal with in his own league, and my system looks for top-of-league performance at both the positional and leaguewide levels. That's why I acknowledged his low ranking. In reality when I someday learn how to do databasing stuff and then put a new engine in my system, I'll be in a position to deal with issues like this one.

Trammell and Smith show up as borderline SS HOMers but I don't believe they are nearly so affected by the positional question as Foxx would be. Nor again the other borderliner SS. Among NL SS of the 1965-1985 period, they are generally helped by the weak intrapositional competition and they come out like dirt in leaguewide comparisons because they can't hit a lick. Trammell's case is a little different in that he's got the Ripken/Yount blockade issue, but Yount disappears in the midst of Tram's prime, so no problemo for Alan there, and thanks to the 12-team league, he gets a couple extra chances to make up ground, which he does.
   34. Kiko Sakata Posted: July 31, 2007 at 06:54 PM (#2463173)
Eddie Murray's my alltime favorite baseball player. My two favorite statistics of Murray's:

(1) Most MVP shares of any player who never one won - 3.33 shares (21st alltime), including 2 2nds, a 4th, and 3 5ths

(2) Murray's career batting line with the bases loaded: 302 PA, .399/.387/.739, 19 Grand Slams, 299 RBI
   35. Dr. Chaleeko Posted: July 31, 2007 at 07:14 PM (#2463227)
Kiko, he's positively Tableresque! Although methinks you have either his AVG or his OBP wrong.
   36. Kiko Sakata Posted: July 31, 2007 at 07:19 PM (#2463238)
methinks you have either his AVG or his OBP wrong.

Actually, that's one of my favorite aspects of it. According to Retrosheet, he has 22 walks and 42 sacrifice flies with the bases loaded. SFs count against OBP but not BA. Weird but true.
   37. Pasta-diving Jeter (jmac66) Posted: July 31, 2007 at 07:22 PM (#2463249)
methinks you have either his AVG or his OBP wrong.

no, they're correct--he had 42 SFs in bases-loaded situations
   38. Dr. Chaleeko Posted: July 31, 2007 at 08:10 PM (#2463373)
wow, that's pretty neat.
   39. David Nieporent (now, with children) Posted: July 31, 2007 at 08:34 PM (#2463413)
The O's owner mouthed off about Eddie during a rain delay in the 1988 (?), implying a lack of hustle and the media just ran with it. I never belied it to be true, and it's worth noting that the paragon of baseball's spartan hard-working virtue -Cal Ripken, Jr., himself- always cites Murray as the man who taught him how to be ready on a daily basis to give his best effort to the team.
It was 1986, the (only) year Eddie got hurt, and the year local hero Jim Traber came up to temporarily replace him.
   40. sunnyday2 Posted: August 06, 2007 at 02:38 AM (#2473163)
All-time (retired)

1. Musial--played more games at 1B than any other position (unless you lump all OF together)
2. Gehrig
3. Foxx
4. Brouthers
5. Anson
6. Mize
7. Murray
8. McGwire
9. Greenberg
10. McCovey

11. Buck Leonard
12. Killebrew
13. Pete Rose
14. Connor
15. Terry
16. Sisler
17. Allen
18. Perez
19. W. Clark
20. Cepeda

21. Hernandez
22. Mattingly
23. McVey
24. Hodges
25. Dave Orr
   41. Howie Menckel Posted: August 06, 2007 at 11:07 AM (#2473552)
Murray was Winfield-esque - and vice versa - yes?
   42. sunnyday2 Posted: August 06, 2007 at 12:11 PM (#2473559)
Murray was quite a bit better, I think. Probably coulda/shoulda won at least one MVP award. Winnie was never that kind of great.
   43. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 06, 2007 at 03:29 PM (#2473767)
Oh sunnyday, I beg to differ. I don't see any of Murray's seasons as close to MVP caliber--although many of them were high All-Star--while Winfield was clearly a top contender in 1979 (although I would have given the award to Schmidt thanks to his fielding).
   44. andrew siegel Posted: August 06, 2007 at 03:46 PM (#2473785)
I have Gehrig and Foxx one-two like everybody else and then Brouthers and Connor. After that, it gets very messy; numbers five through sixteen at 1B are harder to sort out than at any other position. FWIW, I have them: (5) Mize; (6) Leonard; (7) Bagwell; (8) Anson; (9) McCovey; (10) Thomas; (11) Murray; (12) Greenberg; (13) McGwire; (14)Allen; (15)Killebrew; (16) Suttles. (Musial, Rose, and McVey are rated elsewhere.) Big break and then a knot of Hernandez, Terry, Sisler, Will Clark, Cash, Ben Taylor, and Beckley. Palmiero probably belongs at the end of the 5-16 group but I need more perspective on him. Pujols ranks 17th or 18th on raw numbers but will obviously go much higer. Chance, Cepeda, and Tony Perez aren't that far behind either. Joe Start is the only HoMer I haven't mentioned yet; he is somewhere between 29th and 33rd. Wow, there are a lot of 1B I like.
   45. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 06, 2007 at 05:15 PM (#2473883)
Post-1893, MLB only, I have:

1. Gehrig
2. Foxx
3. Mize
4. Greenberg
5. Bagwell
6. F. Thomas
7. McGwire
8. Pujols
9. Thome
10. Murray
11. Allen
12. McCovey
13. Giambi
14. W. Clark
15. Palmeiro
16. Sisler
17. K. Hernandez
18. Killebrew, if you put him at 1B
19. Beckley (with estimates for pre-1893 play)
20. Cash
21. Chance
22. Camilli

Makes you realize that the "steroid era" was really a golden age for 1B (juiced or otherwise), no? #'s 5 through 9, plus 13 and 15. Helton and Delgado are around too, although neither will make my PHoM without a late-career surge, as are Olerud and McGriff. It makes sense, with the game moving more towards the Three True Outcomes and away from speed and contact hitting, that the era would be more favorable to burly, immobile 1B. My system doesn't like McCovey because WS/WARP consider him a poor fielder and because the replacement level for 1B was at an all-time high in the 1960's--there often wasn't one starting 1B in baseball who hit below the league average in those days.
   46. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 06, 2007 at 05:24 PM (#2473895)
Andrew Siegel, no love for Thome? 149 career OPS+ is 30th all-time among players with 10 or more MLB seasons, and he's still going strong...
   47. sunnyday2 Posted: August 06, 2007 at 06:14 PM (#2473955)
Yes the 1990-ff era is a golden age for 1B. I doubt that any position has seen the concentration of stars like the ABC boys, then Gehrig-Foxx-Greenberg-Mize, and then now (if you know what I mean by "then now").

I don't rate active players and I haven't spent any time yet figuring out where Palmeiro goes. But here (DanR) we've got 8 '90s-ff 1B in the top 15 all-time. I can't quite get there from here. The 5th best 1B of the past decade as the 9th best ever?

That suggests that it was a time when it was "easy to dominate" if you were a hulk with little or no knack for fielding but you could do 3 true outcomes. But if it was so easy to go out there and play 1B and so easy to dominate with the bat, then they should all slide down a bit. Or if it wasn't that easy to dominate, then how come they did?

Obviously it matters how good they are--i.e. the 3rd best 1B of the 1880s, 1930s and 1990s is better than the best of many other times such as the 1900s and 1950s, no question. But again, I can't quite get comfortable with the 8th best of the '90s being better than the best from other eras.

And if, as is rumored, pitching was especially good/dominant in the ' 60s, the I don't see how you can knock McCovey (and probably Killebrew) down that far. A 180 OPS+ in the '60s had to be better than one in the 1990s, didn't it?

I understand that it correlates, but I still don't get the flow of causation. You say players dominated because the structure of the game allowed them to. I still wonder if the structure of the game appears to be what it was because some guys dominated. I mean, it's easier for great players to dominate than for very good ones, and great players come along in random fashion. In other words, you look at SDs and see the underlying structure of the game. I look at SDs and I see clusters (and voids) of better players.
   48. sunnyday2 Posted: August 06, 2007 at 06:16 PM (#2473956)
And I'm not saying I'm not an idiot. I'm just saying I still don't get the basic concept that the SDs seem to be revealing to others. Not unlike astrophysics. Just 'cause I don't understand string theory doesn't make it wrong. But if there's a simpler theory, well, I'm just gonna go there.
   49. Willie Mays Hayes Posted: August 06, 2007 at 06:40 PM (#2473973)
Am I the only one who sees Murray about even with Beckley?
   50. DL from MN Posted: August 06, 2007 at 06:46 PM (#2473977)
> Am I the only one who sees Murray about even with Beckley?

Maybe you and Karlmagnus but I'm pretty sure that's not what you're saying.
   51. Willie Mays Hayes Posted: August 06, 2007 at 06:55 PM (#2473986)
Maybe you and Karlmagnus but I'm pretty sure that's not what you're saying.


LOL, no. Maybe I just have an artificial underrating of first basemen. I mean, Murray was a heck of a ballplayer, but a 5-year peak of 156, 156, 156, 156, 149 OPS+'s doesn't strike me as awesome, especially when the first 156 came in 2/3 of a season. Granted that trumps Beckley's offensive peak, but the glovework and league adjustments bring the two close enough, don't they?
   52. Yeaarrgghhhh Posted: August 06, 2007 at 06:59 PM (#2473993)
Granted that trumps Beckley's offensive peak, but the glovework and league adjustments bring the two close enough, don't they?

How so? Murray was a good defensive player, and Murray should get a big boost from a timeline/level of competition adjustment. I guess I'm just not following your argument.
   53. Willie Mays Hayes Posted: August 06, 2007 at 07:29 PM (#2474025)
How so? Murray was a good defensive player, and Murray should get a big boost from a timeline/level of competition adjustment. I guess I'm just not following your argument.


Right - integration. I thought I was missing something. Thanks.
   54. DavidFoss Posted: August 06, 2007 at 07:51 PM (#2474057)
And if, as is rumored, pitching was especially good/dominant in the ' 60s, the I don't see how you can knock McCovey (and probably Killebrew) down that far. A 180 OPS+ in the '60s had to be better than one in the 1990s, didn't it?

Isn't it already normalized? I mean, McCovey 1968 is .293/.378/.545 while Thomas 1996 is .349/.459/.626 but roughly the same OPS+ (175-178) so comparable value. And doesn't the extra pop up and down the lineup make it harder for the 1B to put up great OPS+'s?
   55. DavidFoss Posted: August 06, 2007 at 08:04 PM (#2474083)
And I'm not saying I'm not an idiot. I'm just saying I still don't get the basic concept that the SDs seem to be revealing to others.

To tell you the truth, I don't fully understand some of the SD arguments either. Especially the "lower SD means comparable OPS+ seasons are worth more pennants". 30 WS is worth 10 wins no matter what the standard deviation is. If the standard deviation is small, then you win fewer pennants -- not more -- because its harder for one star to carry a team. More dependence on roster depth rather than clusters of superstars. It also might explain why there are fewer dynasties now.

Now, if you want to say that great seasons in lower SD eras are more impressive and imply a higher level of talent, then sure... but from a pure value perspective, I don't think they result in "more pennants". Not that we should go pure value, of course. There's era balance issues to consider.

Am I missing something?
   56. Yeaarrgghhhh Posted: August 06, 2007 at 08:18 PM (#2474110)
Even putting aside any timeline or integration issues, Murray was easily superior to Beckley for both career and peak. Murray had ~2000 more PAs and had a higher career OPS+.

Murray's top 10 full seasons by OPS+: 159, 156, 156, 156, 156, 149, 140, 138, 136, 135.

Beckley's top 10 full seasons by OPS+: 152, 144, 138, 133, 131, 128, 127, 127, 126, 124.

So Murray beats him every year (in many years, by a significant amount).
   57. Dr. Chaleeko Posted: August 06, 2007 at 08:26 PM (#2474127)
Dan,

I can't quite remember. Do you do SDs for the season or for position as well? Just season, right?
   58. Willie Mays Hayes Posted: August 06, 2007 at 08:29 PM (#2474131)
The evolution of the position, and Beckley's defensive edge make it much closer than that, before you account for integration and the like, though. I never particularly cared for Beckley, which is what surprised me when my spreadsheet had the two comparable. After proper adjustments, Murray is a few notches above Eagle Eye.
   59. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 06, 2007 at 08:30 PM (#2474135)
sunnyday2, thanks for your response.

Well, sure, and I have the third-best 1B of the 1930's as the fourth-best overall since 1893. And the 1930's were a *high* standard deviation era. I agree with you that there is no reason to believe that the distribution of talent should be equal across eras and across positions. What I said about the three true outcomes and the 1B bulge of the 1990's/2000's is mere speculation--I haven't done a study on how changes in the way the game is played have affected the relative values of the various positions.

I am not sure that you understand my stdev methodology, sunnyday2. I do not use the actual standard deviation of any given year or group of years to determine its ease of domination. I did a regression analysis of standard deviation, and found the various league factors (run scoring, expansion etc.) that correlate to it, and got an r-squared of about 30%. I use the regression-predicted standard deviation--how difficult to dominate we would expect the league to be, based on its characteristics--rather than the actual standard deviation to adjust for ease of domination.

My system doesn't "know" anything about the concentration of stars in the game or in a league at a particular period in time. If you look at the stdev chart for the 1910's, you'll see the AL's actual stdev as extremely high (Collins, Cobb, Speaker, Lajoie, Jackson), the NL's actual stdev as extremely low (just Cravath and Burns, really), and the regression line--which is the number I use to adjust for ease of domination--running right down the middle between them. I do not penalize Collins/Cobb/Speaker for the high stdev of their league generated by the fact that they all happened to be in it, nor do I reward Burns and Cravath for the low stdev of their league generated by the absence of Collins/Cobb/Speaker (if I did, I'd probably vote for Cravath). Instead, I look at the characteristics of each league-season in terms of run scoring, expansion, and a few less important factors, and calculate an adjustment based on that. So there is no way that your suggestion that "the structure of the game appears to be what it was because some guys dominated" can be correctly applied to describe my system. The stdev adjustment I use has no "knowledge" of whether some guys dominated or not. All it "knows" are the independent league factors that have correlated to standard deviation over the course of baseball history.

There have been seasons that were, by my measure, rather easy to dominate, but that no one took advantage of: the 1950 AL is a good example (only two qualifiers had an OPS+ over 150). In fact, the AL stdev was lower than its regression-predicted value every year from 1948 to 1952, which suggests the league had a "star drought" during this period. There were only two qualifying seasons by AL players with a 170 OPS+ in those years, both by Ted Williams. Yet players from those years like Vern Stephens and Phil Rizzuto do *not* receive any extra credit from me, because the low stdev was caused by an absence of superstars, not because it was actually difficult to dominate.

Conversely, there have been seasons that were, by my measure, rather difficult to dominate, but which produced numerous megastar seasons nonetheless. The biggest example is the 2001 NL, where Bonds, Sosa, and Luis González all went nuts, but we can presumably attribute some of that gap to a variable that is not included in my regression: steroid use. (This is, in my opinion, a great way to measure the impact of steroids on the game, by the way). A better example would be the 1992 NL, a low-scoring league 23 years removed from expansion that Barry Bonds, Gary Sheffield, Darren Daulton, Ryne Sandberg, and Andy Van Slyke still all made their b-!-t-c-h. Those remarkable seasons all get an additional boost in my system due to the difficult conditions in which they were achieved.

McCovey and Killebrew are hit in my system for three reasons. The first is that they played during modern baseball's first hyperexpansionary period (the second was the 1990's NL), which means that they got to beat up on expansion pitching and had the league averages they are compared to dragged down by expansion hitting. This has a massive effect on standard deviation--just look at all the guys who had career years in 1961 (Maris, Cash, Gentile) and 1969 (Petrocelli, Reggie Jackson, Wynn, Staub, McCovey, Frank Howard, Bando, Cleon F'-in Jones). (Interestingly, the only guy who really took advantage of the 1962 NL expansion was Tommy Davis. Mays, Frank Robinson, and Aaron were great, but they were always great). Clearly, McCovey and Killebrew (and all players whose careers spanned those years) need to be regressed for this effect. A given OPS+ in 1969 is *definitely* not better than the same OPS+ in the 1998 AL; quite the opposite, it's substantially less valuable. Thanks to the expansion, a 160 OPS+ in the 1969 NL would have been 7th in the league; in the 1998 AL, which had two more teams than the 1969 NL, it would have been 2nd. This is *not* due to a "star drought" in the 1998 AL; the league featured peak seasons from Albert Belle, Alex Rodríguez, Ken Griffey Jr., Derek Jeter, Bernie Williams, Nomar Garciaparra, Manny Ramírez, Iván Rodríguez, Mo Vaughn, Rafael Palmeiro, Juan González, Edgar Martinez, Carlos Delgado....it was due to the fact that 1969 was an expansion year, while the 1998 AL was only softened by the minor dropoff from the hapless Brewers to the slightly worse Devil Rays.

The second reason is that replacement level for 1B was at an all-time high in the immediate pre-DH era. The absolute worst starting 1B in those years were close to league average hitters (Denis Menke in 1971, Gail Hopkins in 1970, Ed Kranepool in 1969, etc.), far better than they were before the war or after the arrival of the DH created 12 new jobs for good-hit no-field players. You could basically pick a league-average hitter off the scrap heap to play 1B in those days, so the relative value of having a monster bat at the position was lessened.

The third reason is that WS and WARP both consider them to be poor defenders. I am sure this is true in Killebrew's case, but less so in McCovey's, if nothing else just because of the nickname "Stretch." If McCovey were an average rather than poor fielder as the uberstats suggest, he would quickly move up in my rankings.

Basically, the reasons why I am unimpressed with McCovey and Killebrew (high standard deviation, high replacement level, bad fielding) are the mirror image of the reasons why I support Concepción (low standard deviation, low replacement level, good fielding). At least I'm consistent.
   60. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 06, 2007 at 08:37 PM (#2474146)
Whoa, lots of posts while I was typing that. Responses:

DavidFoss: I think it's pretty straightforward...in a low-stdev league, a 160 OPS+ leads the league, in a high-stdev league, it's 10th. Yes, wins above average are wins above average, but everything is relative--if you're contributing 5 wins when the leaders are contributing 10, that doesn't produce as many pennants as a 5-win contribution when the leaders are contributing 6. This seems intuitive enough to me, but to express it mathematically I would have to review the Pennants Added methodology. Let me know if my doing that might influence your voting; if so, it's definitely worth my time to do so.

Dr. Chaleeko: I most definitely only do stdevs by league-season, not by position. If I did them by position, then guys like Traynor and Mickey Vernon would be on my ballot. It *might* be worth doing a study on positional stdevs to see if they show any statistically significant correlations to league factors above and beyond those that affect the whole league--is it tougher for infielders to dominate relative to outfielders when there are more/fewer ground balls? etc.--but that's not research I've done, and it's not included in my analysis.
   61. DavidFoss Posted: August 06, 2007 at 08:58 PM (#2474169)
Yes, wins above average are wins above average, but everything is relative--if you're contributing 5 wins when the leaders are contributing 10, that doesn't produce as many pennants as a 5-win contribution when the leaders are contributing 6. This seems intuitive enough to me, but to express it mathematically I would have to review the Pennants Added methodology.

I think I was more concerned about the equivalence of the leaders in the example above. 6 wins is less than 10 wins no matter what the stdev's are and teams in the low STDEV era must rely on a deeper collection of talent (with fewer holes).

But, I guess what you are doing is saying that the "Pennant Bar" is higher in a high STDEV era. For example, you'd need a .650 WPCT to win a pennant in a high-stdev league and a .600 WPCT to win in a low STDEV league. Anecdotally, that seems true.

You do a very thorough analysis of stdev in relationship with replacement level, though. Plus, statements like "the low stdev was caused by an absence of superstars, not because it was actually difficult to dominate" tell me that you are thinking of all the details. I'll take another look at your data to see if you have metrics demonstrating these concepts.
   62. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 06, 2007 at 09:12 PM (#2474183)
DavidFoss, yes, that is exactly correct. Ceteris paribus--assuming that the distribution of talent among teams doesn't change--in a higher-stdev league, the best team in the league will have a higher winning percentage than in a lower-stdev league.

The file that should interest you, DavidFoss, is the "StDevs and Rep Levels" spreadsheet in the .zip archive in the Yahoo! group. It includes all the regression results (coefficients, r-squared, etc.), and a graph of the actual and regression-projected standard deviations for both leagues from 1893 to 2005. It is fascinating to look at the residuals, the places where the regression projection is not close to the actual league stdev. These usually correspond to star gluts and droughts (like the 1910's AL/NL disparity, where the AL stdev is super-high, the NL is super-low, and the regression line runs right through the middle), but not always. The huge gap between the actual and projected stdev around the years 2001-02 seem to me the best evidence of the impact of steroids on the game that I have seen--talk about unexplained variance.
   63. Dr. Chaleeko Posted: August 06, 2007 at 11:42 PM (#2474284)
Dan, thanks for clarifying.

One quick question:

which means that they got to beat up on expansion pitching and had the league averages they are compared to dragged down by expansion hitting.

I'm a little confused by this. The way you phrased this it sounds like the expansion hitters are worse than the pitchers because they drag down the league averages. But the league averages don't show this.

In 1960 the AL scored 4.39 R/G. .255/.328/.388
In 1961 the AL scored 4.53 R/G. .256/.329/.395

In 1961 the NL scored 4.52 R/G. .262/.327/.405
In 1962 the NL scored 4.48 R/G. .261/.327/.393

In 1968 the AL scored 3.41 R/G. .230/.297/.339
In 1969 the AL scored 4.09 R/G. .246/.321/.369

In 1968 the NL scored 3.43 R/G. .243/.300/.341
In 1969 the NL scored 4.05 R/G. .250/.319/.369

Let's skip up to 1997-1998

In 1997 the AL scored 4.93 R/G. .271/.340/.428
In 1998 the AL scored 5.01 R/G. .271/.340/.432

In 1997 the NL scored 4.60 R/G. .263/.333/.410
In 1998 the NL scored 4.60 R/G. .262/.331/.410

So in our six examples here, the league R/G
-increased four times.
-stayed put once.
-decreased once.

The league AVG stayed nearly identical in four instances, but in 1969 it leapt up in both leagues.

The league OBP stayed nearly identical in four instances, but in 1969 it leapt up in both leagues.

The league SLG
-increased four times
-went down once
-stayed about the same once.

Now we know that the K zone was redefined in 1969 to be a little more accomodating to batters, so that's explaining all or most of the uptick in production there, but in general the league figures go up slightly or stay the same, they don't go down.

So I'm still unsure what you mean about the expansion hitters/pitchers. Can you clarify? Thanks!
   64. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 07, 2007 at 12:55 AM (#2474359)
Actually, if I'm not mistaken, the big offensive spike from '68 to '69 was due to the lowering of the mound, no???

When leagues expand, people always talk about established hitters beating up on expansion pitchers. But that in and of itself is NOT enough to change their uberstat value--since every established hitter gets to face the expansion pitching, every established hitter should increase his production by the same amount, run scoring should go up, and neither the standard deviation nor OPS+/WS/WARP should change. What accounts for the increased standard deviation is the introduction of weak expansion hitters, who pull down the league average run scoring to roughly where it was before the expansion (as your data from 60-61, 61-62, and 97-98 demonstrate). That means that the gaudy numbers produced by the established hitters against expansion pitching are not only compared to each other, but to the poor expansion hitters as well, resulting in much higher OPS+/WS/WARP scores for the leaders.

Example:

We have a league with 11 hitters, producing 3, 3.5, 4, 4.5, 5, 5.5, 6, 6.5, 7, 7.5, and 8 runs per game. The league average is 5.5 runs per game, the stdev is 1.66 runs per game, and the leader is 2.5 runs above average per game.

The league expands. All 11 hitters get to clobber pitchers who were formerly in the minor leagues, and all add 0.5 RC per game. Thus, they now produce 3.5, 4, 4.5, 5, 5.5, 6, 6.5, 7, 7.5, 8, and 8.5 runs per game. The league average moves up to 6.0 runs per game, but the standard deviation does not change, and the leader remains 2.5 runs above average per game. The introduction of expansion pitching has done nothing to the relative value of the established hitters (OK, it actually reduced it a bit, because 2.5 runs above average is worth fewer wins in a 6 R/G environment than in a 5.5 R/G environment, but you get the idea).

However, we also add two expansion hitters, who produce 2.5 and 3 RC per game in the expanded league. Now what happens? The league average falls back to 5.5 R/G, right where it was. But the standard deviation moves up from 1.66 to 1.95, and the leader is now 3.0 runs above average per game, instead of 2.5. Thus, while the addition of expansion pitching is what causes established hitters' offensive production to increase, it is the addition of expansion hitting that causes their uberstat value to rise. So while it might be fair to "blame," say, Ron Moeller for Jim Gentile's .302/.423/.646 line in 1961, it's really Coot Veal's "fault" that Gentile rang up a 184 OPS+. (Yes, I had fun looking those guys up).
   65. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 07, 2007 at 01:06 AM (#2474366)
By the way, sunnyday, despite the expansion and replacement level factors, if I make McCovey an average fielder, he'd move up to #8 on that list. If I call him an average fielder *and* put him in a non-expansionary league *and* give him a 1990's 1B rep level, he'd essentially tie McGwire for 7th and be extremely close to Frank Thomas. Bagwell, however, is well out of reach--his defense and baserunning separate him from the pack.
   66. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: August 07, 2007 at 02:23 AM (#2474480)
Just for the record, Murray soured on the media in 1979 after a paper (IIRC it was the LA Times) ran a big story about a brother of his who'd been in trouble with the law. Murray saw it as an unforgiveable violation of his family's privacy and barely said a word to the press after that, at least for many years.
   67. Cblau Posted: August 08, 2007 at 02:10 AM (#2476221)
Actually, if I'm not mistaken, the big offensive spike from '68 to '69 was due to the lowering of the mound, no???

No, it was due to the shrinking of the strike zone to its 1962 size. From what I can find, everytime the mound was raised, scoring increased, and the one time it was lowered, scoring was lower than it was the previous year with the same strike zone.
   68. Paul Wendt Posted: August 09, 2007 at 05:10 AM (#2478246)
OCF #4
My impression when he came to the Dodgers was that he'd been run out of town in Baltimore and that Baltimore fans - or maybe it was just the local press? - had a "good riddance" attitude about that. Frankly, I found that very hard to understand. Of course, I am talking about the same things as in John DiFool's post #2.

Charles Steinberg is a longtime member of the Larry Lucchino team, presently Boston Red Sox VP for PR and all that. He praises Eddie Murray as Cal Ripken does, and more; Murray may be his favorite player. According to Steinberg (as brilliant guest speaker for the Boston Chapter SABR meeting last fall), Murray criticized the Orioles sharply in the late 1980s, for abandoning "the Oriole way", and he clearly wanted out of that no longer serious organization. Steinberg is a deeply interested party yet this may have some value as counterpoint to the view that Murray was "run out of town".


DanR #60
Dr. Chaleeko: I most definitely only do stdevs by league-season, not by position. If I did them by position, then guys like Traynor and Mickey Vernon would be on my ballot. It *might* be worth doing a study on positional stdevs to see if they show any statistically significant correlations to league factors above and beyond those that affect the whole league--is it tougher for infielders to dominate relative to outfielders when there are more/fewer ground balls? etc.--but that's not research I've done, and it's not included in my analysis.

In another thread here today, I asserted casually and invited correction, that 1950s integration improved the batting (and fielding) relative to the pitching.
   69. Paul Wendt Posted: August 09, 2007 at 05:18 AM (#2478261)
Top Tens, OPS+
10 8 3 2 2 2 5 7 2 - Murray
7 1 9 9 3 4 9 - Winfield

Top Tens, plate appearances
6 10 8 7 8 10 3 - Murray
7 - Winfield

Perhaps Winfield in 1979 was a better MVP candidate than Murray in any season. But I'm with Alpen Marc: Murray was clearly better.
   70. Chris Cobb Posted: August 14, 2007 at 11:48 PM (#2484910)
This is a continuation of a discussion of Murray’s peak and Melky?’s ballot from the 2003 ballot thread.

John Murphy wrote:

Now, within the parameters of peak/prime/career, if you want to show inconsistencies regarding a voter's ballot (i.e. if Marc, as a peak voter, had a Beckley-type player at #1 on his ballot), then you will make an inroad.

Let’s see about making some inroads.

On the top of his ballot, Melky? wrote the following:

1. Pete Browning - Yes, I firmly believe he is the best player on this "year"'s ballot. His peak, to me, was more valuable than Murray's career. I see Murray as a rich man's Beckley, and I find that to be worth a bit less than the monstrous peak of Browning.
2. Eddie Murray - Still a fine player. Very consistent, missing that one or two huge seasons.


Farther down his ballot, Melky? also wrote:

13. Don Mattingly - Clark, Murphy, and Donnie Baseball are essentially tied in my system. Excellent defender, great 3 year peak.

So Eddie Murray is missing those one or two huge seasons, but Don Mattingly has a great three-year peak.

During Don Mattingly’s 3-year peak from 1984-86, he put up the following EQA2s:
.331, .330, .338

During Eddie Murray’s peak 3-year peak from 1982-84, he put up the following EQA2s:
.330, .329, .338

Over his five best seasons, 1981-85, Eddie Murray’s performance totals, by WARP2, are as follows:

Games 724
Outs 1909
BRAR2 349
BRAR2/400 outs 73.12
FRAR2 86

Over his five best seasons, 1984-88, Don Mattingly’s performance totals, by WARP2, are as follows:

Games 759
Outs 2131
BRAR2 370
BRAR2/400 outs 69.45
FRAR2 91

Murray was a better hitter over his best five consecutive years than Mattingly was over his, at least as WARP sees it. Mattingly leads in counting stats because he played more, but let’s look at Murray’s numbers again, with 1981 adjusted from 108 to 162 games:

Games 773
Outs 2045
BRAR2 371
BRAR2/400 outs 72.65
FRAR2 94

I must raise the question inconsistency on a ranking that docks Murray for not having monster years while praising Mattingly for a great peak, when their peak performances were nearly the same. Mattingly’s top 3 were indeed better than Murray’s because his fielding peak was a little higher and lined up exactly with his batting peak, but if you go out even to five years, he passes Mattingly on both hitting and fielding.

It’s also worth mentioning that Eddie Murray, over his five-year peak, was, by WARP2, the best hitter in the American League.

Here are the top rates of offensive production for those five years:

Eddie Murray 73.12 BRAR2/400 outs
Rickey Henderson 72.56 BRAR2/400 outs
George Brett 68.44 BRAR2/400 outs
Dwight Evans 57.00 BRAR2/400 outs
Robin Yount 56.79 BRAR2/400 outs

Now, as the four below Murray were probably all good to outstanding baserunners and Murray was not, I could certainly believe that a full accounting of base runs could move the other players ahead of Murray.

Moreover, all of these players had more defensive value than Murray, for sure.

But as a hitter, surely Murray was the best, and he was competing directly with some outstanding offensive players.

I dislike pounding on Melky’s? ballot like this, but his interpretation of Murray’s record seems significant.

I hope that we will, as sunnyday2 has put it elsewhere, use imagination and intuition when they are called for, but I hope that we will not use them at the expense of meaningful data, where it exists.
   71. OCF Posted: August 15, 2007 at 12:16 AM (#2484966)
In my posts #4 and #6 above and in my ballot commentary, the peak-value candidate I was using to compare to Murray wasn't Mattingly but the not-yet-eligible Will Clark. I'll add that in my system, Clark just clobbers Mattingly all across the board. If I had to choose between Clark and Murray, I've got enough career in my system (hey, I once voted for Mickey Vernon) that I'd prefer Murray. But I could understand the hard-core peaksters in the electorate (you can diagnose them by their votes for Al Rosen) might go for Clark. But Murray vs. Mattingly? I agree fully with Chris's sentiments.
   72. Mark Donelson Posted: August 15, 2007 at 12:31 AM (#2485003)
But Murray vs. Mattingly? I agree fully with Chris's sentiments.

So do I. And I'm one of those guys who vote for Al Rosen! (I don't think there are any others anymore, actually...)
   73. Willie Mays Hayes Posted: August 15, 2007 at 12:56 AM (#2485059)
Chris - You're either cherry-picking my words, not reading my whole posts, or disregarding my words, to back your case up. I get it, you don't find Browning meritorious. Plenty of people disagree with me, and I'm ok with that. You bring up Mattingly vs. Murray. Did you not see the point where I mentioned the chasm between Singleton and Jack Clark on my ballot? Or that I mention Mattingly's 3-year peak? What does 5-year peak have to do with that? You waste time and space comparing the five-year peak, I get it. Mattingly had a 3-year peak befitting the 14th slot on my ballot. Murray is 12 spots, and untold measures of value higher! Did you miss that? I never ever said Murray wasn't a fine player. You are piecing together an argument without looking at the relevant points, and you're putting words into my mouth.

Did I ever say Mattingly had that one or two "wow" seasons?

Did I ever say Murray was a worse player than Mattingly?

Did I have Mattingly higher than Murray on my ballot?

Did I mention anything about Mattingly's five-year peak?

The answer to all of these questions, of course is no, and furthermore, none of these questions are pertinent to your major beef, namely my ranking of Browning over Murray. Rather than dwell on this anymore than necessary, I'll list some reasons I have Browning higher.

RC/27 (Best 5 seasons)
Murray: 8.13, 8.01, 7.78, 7.44, 6.68
Browning: 11.31, 10.14, 9.50, 8.79, 6.84

Players with similar RC/27, same season and league, top 3 seasons:

Murray:
1990: Ryne Sandberg (7.36), Barry Bonds (8.24), Lenny Dykstra (7.06), Kevin Mitchell (7.09)
1982: Robin Yount (8.26), Dwight Evans (7.77), Hal McRae (7.44)
1984: Dwight Evans (7.44), Don Mattingly (7.87), Cal Ripken (7.08), Mike Easler (7.37)

Browning:
1887: (Browning 11.31) Tip O'Neill (16.00), Bob Caruthers (10.50), Denny Lyons (9.27), Oyster Burns (8.70)
1890: (Browning 10.14) Roger Connor (10.20), Dave Orr (9.34), George Gore (8.44), Dan Brouthers (8.24), Buck Ewing (9.04)
1882: (Browning 9.50) Ed Swartwood (7.31), Hick Carpenter (6.19), and thats about it.

It's close, but an edge here to Browning, both in quality of contemporaries, and distance between he and the nearer players.

Browning finished Top 6 in OPS+ 9 out of 10 years, Murray finished Top 10 in 9 out of 13 years. Edge goes to Pete there, league adjustment or not.

Black Ink: Browning 21, Murray 11

Beckley finished Top 10 in OBP 9 years out of 10, Murray wasn't even close.

Browning finished top 3 in Slugging from 1882 - 1888, way better than Murray's five year run of top sixes.

EQA Top 3 seasons:
Browning: .403, .359, .349
Murray: .333, .326, .322

Even adjusting for league, that's advantage my boy.

I can't quantify it, and no one can, but Browning's fielding was not as bad as the uber-stats make it appear, which negates Murray's better defense at a less important defensive position.

Long story short, yes Browning played in easier leagues. But he dominated them. Murray played in tougher leagues, and was among the better players in those leagues. He did not have the level of domination at his peak that Browning did, and that counts a lot in my system. I'm sorry you don't see that, and you are choosing to pound my ballot. In my opinion many others have chosen to let far more egregious ballot positions go with nary a mention. I wish I were here Josh Gibson's year to figure out what the hell he needed to do to be unanimous. How does Tony Oliva warrant a first place vote, ever? I understand your concern about Browning. I never cared for the election of Beckley. Things like this happen under the set up we have.

Let's call a spade a spade, though. This would be irrelevant if I'd have had Singleton above Murray, because Singleton doesn't have a chance at election. You're making an example out of me an my ballot. I'm fine with that, I can handle it. But stop pretending your motives are pure, when they clearly are not.
   74. Chris Cobb Posted: August 15, 2007 at 04:27 AM (#2485533)
Got Melky? wrote:

Chris - You're either cherry-picking my words, not reading my whole posts, or disregarding my words, to back your case up.

I'm sorry you feel that I am misrepresenting your words. I can only say in response that I was attempting to interpret your choices, based on what you say on your ballot. I picked statements that seemed salient, and quoted them. As your ballot comments are, like most of our ballot comments, anecdotal rather than systematic, it is hard for me to discern from your comments how you have actually made your decisions.

Did you not see the point where I mentioned the chasm between Singleton and Jack Clark on my ballot? Or that I mention Mattingly's 3-year peak? What does 5-year peak have to do with that?

Your description of "chasms" separating players on your ballot did not catch my attention because I can't see how much meaning it has. How big is a chasm, and what gets a player across it? I brought up both 3- and 5-year peak in my commentary because your description of Murray, particularly with the comparison to Beckley, makes it sound as if he does not have a peak worth speaking of, when, as far as I can tell, he has as much of a peak as Mattingly, a player whom you have ranked lower but praised for his peak. That appeared to me, as I looked at your ballot, to be inconsistent. By showing something of the actual number you are working with, you make it much easier for me to understand your reasoning. From words, I cannot know how you differentiate between a "wow" season, a monster season, and a great three year peak. And when you are (1) going against what is quite near the unanimous position of the electorate in ranking Browning over Murray and (2) doing so in what is likely to be a closely contested election for a player about whom there is wide disagreement, I think it is only right that you make your reasoning clear, and I appreciate your taking the time to do so.

But stop pretending your motives are pure, when they clearly are not.

I am also sorry if I gave the impression of "pure" motives. When the Browning dustup was proceeding on his thread, I quite explicitly disavowed being above interest in my choice of arguments. I repeat here that I am not.

To avoid the appearance of disinterest, let me make my position on Browning in all this clear, which I should have done in the preceding post on Murray. By way of excuse, let me say that my intention had been to say something about the Browning issue. I gathered the data about Murray and Mattingly in a Word file, copied it, then added a few sentences on Browning, hit submit, and went immediately to dinner. When I returned, I discovered that I had lost my internet connection before submitting and the post had not gone through. I immediately repasted from the word file and submitted, forgetting that the Browning bit was not in the original text. When I realized the omission, I considered writing another post on the Browning bit, but I decided that the Browning connection was obvious enough that I probably didn't need to bring it up again. Clearly that decision was incorrect, and I apologize if I appeared to create an insufferable image of self-righteousness.

Here is my view on Browning's possible election. My reconsideration of his case has raised him in my estimation, not so much that I support his election, but to the point where I can see clearly the evidence that is meaningful to some voters. If he is to be elected, however, I would like to see that election take place together with a strong justification of his merits, and without his election depending on highly questionable ballot placements. Your placement of him #1 ahead of Murray seemed an example of such questionable placement, particularly given your description of Murray's qualifications.

I did not say anything more about Browning because I thought I have said what I can say about him, and because the issue in your rankings was not, on its face, how you were placing Browning, but how you were placing Murray. So I looked at your ballot to try to figure out how you really are evaluating peak, and that brought me to Mattingly, who as a peak candidate and an immediate contemporay of Murray's seemed a particulary relevant case for seeing if you had properly recognized Murray's peak.

Let's call a spade a spade, though. This would be irrelevant if I'd have had Singleton above Murray, because Singleton doesn't have a chance at election. You're making an example out of me an my ballot.

Well, it wouldn't be irrelevant, because I would still see it as a large mistake in ranking, but it wouldn't be worth the effort for me to take the time and suffer through the pain that is almost invariably created from seriously challenging the reasoning underlying someone's rankings, both of which are typically considerable, as our current discussion demonstrates. For example, I view rawagman's ranking of Hugh Duffy above Murray as pretty comparable to your ranking of Browning over Murray in terms of undervaluing Murray, but that error is not going to affect the outcome of the election, so I let it pass.

That does not mean, however, that I am doing this to make an example out of you. As I hope I made clear above, my purpose is not to keep Pete Browning from being elected by any means necessary! If he is elected, I won't say it was a huge mistake. My intention in questioning your rankings is to try to make sure that if he is elected, it is for the best reasons we can produce. The problem, I say again, is more with your ranking of Murray than with your ranking of Browning. Let me explain more fully why I say that by a brief glance at the voting so far. Ballot-counting along the way is frowned on, but I think I can say this much without crossing the line: rawagman's ballot, which I have already mentioned, is the only one besides your own where I have had a problem with the ranking of Murray. Mostly he's been #1, though a few peak or high-fielding voters have had Sandberg #1, and that's a reasonable position to take. Nobody besides yourself and rawagman has placed Murray below anyone else. Some people have had Browning high on their ballots. Sean Gilman's ballot went 1. Murray, 2. Sandberg, 3. Browning. I don't agree with that, but from what I know of the case, I think such a ranking could be defended with some persuasiveness, so I didn't question it. My questions about your ballot are not to make you an example, because, frankly, I doubt anyone else was planning to rank Browning over Murray, anyway. Your position on Browning/Murray is probably unique, and its uniqueness indicates that no one else has found this position persuasive. I think that might be reason enough for you to reconsider your thinking. I would like to see you change your mind, so I questioned your ballot because I thought it ought to be questioned on its own merits and because of its undoubted importance for this election. Unfortunately, direct argument almost never leads participants to change their minds -- it is more likely to just entrench the positions already held. But the only other option was not to bother, and I decided I should bother, anyway, and see what happened.

If you've read this far, you''ve read more than enough from me, but it would be unfair of me not to respond to the data you have thoughtfully presented. The way you are looking at RC/27 outs makes sense, though I would disagree with your conclusion that Browning has an edge in quality of contemporaries. That may be partly just my view of the players you list, but it is also due to the fact that I think single-season leaders list can give a misleading view of who a player's contemporaries really are and what their relative merits are, unless quite a few are considered, preferably over a continuous span of seasons.

When I looked at for Murray's continuous peak, it was clear that it was in the early 1980s, when players like Henderson and Brett were in their primes, as well as Yount and Evans, who show up in your lists. Henderson and Brett didn't happen to have great years in 1982 and 1984, which are years that RC/27 sees as Murray's best two. They were great in 1983 and 1985. But they were among the players Murray was being measured against, and over even a five-year continuous period, the ups and downs of little injuries or luck are evened out, and all the great players rise to the top, Murray among them. Murray, in his own league, was competing against a very deep pool of strong hitters. Browning's pool was a lot shallower at the top. In 1981 to 1985, Murray placed 3-2-2-2-5 in OPS+. Every player who ranked ahead of him is an elected or certain HoMer, except Don Mattingly, whose peak is very similar to Murray's. They are Grich and Evans in 1981, Yount in 1982, Brett in 1983, Mattingly in 1984, and Brett, Henderson, Mattingly, and Boggs in 1985.

Browning has 2 #1 finishes in OPS+ that Murray doesn't, 1882 and 1890. I'll give Browning 1890: he was the best hitter in his league and in major-league baseball that year and was measured in his league against the top hitters, though Connor does slightly better in RC/27, I note. I'll ask to set 1882 out of consideration for the moment: Browning was undoubtedly great, but there were no other even _good_ hitters in the league at that time, by the standards of the NL.

But let's look at a group of seasons where Browning's OPS+ finishes match Murray's and see who he was up against that topped him in those years. If we look at 1888, 1883, 1885, 1887, and 1884, we get finishes of 3, 2, 2, 2, 5 in OPS+, corresponding to Murray's, and in a mostly continguous set of seasons. Who was ahead of Browning?

(A question I will answer in the next post -- this one was slightly over quota, so I'll split it in two here)
   75. Chris Cobb Posted: August 15, 2007 at 04:27 AM (#2485535)
Players ahead of Browning by OPS+
1888 John Reilly, Harry Stovey
1883 Fred Swartwood
1885 Dave Orr
1887 Tip O'Neill
1884 Dave Orr, John Reilly, Harry Stovey, Fred Fennelly

Here there's one HoMer, Harry Stovey. Dave Orr was consistently a masher, could have been a HoMer with more career. The other guys, no. They had a couple of big years, but that's it. If we thinned Murray's pool of competition down to a couple of HoMers instead of the set of HoMers who topped him 1981-85, (plus late Reggie Jackson, Dave Winfield near his prime, Cal Ripken, and Alan Trammell, all of whom placed in the top 10 in OPS+ during Murray's peak), it seems very likely to me that Murray would have stood out a lot more than he did. At the very high end, quality of competition isn't just about how one's OPS+ number ought to be adjusted, it's about who is in the league who has a serious chance of bettering you. Murray's league was _very_ deep.

On the other hand, you might want to flip it around and go back to a value argument, and then we are plunged back into the undecidables about how to adjust OPS+ or RC/27 outs numbers.

(I'll also note that short seasons make it easier for a good player to have an outlying great year, so that may be a factor in who topped Browning: it would be good to look at five consecutive years together to get another view of who the top hiters really were. But I'm guessing that it would be mostly the same group of players who showed up among the league leaders over a five-year span -- I don't think the league leader lists are leaving out strong hitters in Browning's pool of competitors.)

But the point I would return to is that if you removed Grich, Evans, Mattingly, Henderson, Brett, and Boggs and replaced them with the equivalents of Stovey, Orr, O'Neill, Reilly, Swartwood, and Fennelly, Murray's dominance would look a lot like Browning's. He probably would have topped his leagues a couple of times. Given all the career that Murray is carrying around, I can't see how Browning's peak advantage, with context duly considered, can possibly outweigh it. That's all I'm saying. I'm not saying Pete Browning was a scrub. I'm not arguing that you ought to drum him from your ballot. I'm not even arguing that you should drop him below Ryne Sandberg (though I think that would be a sound idea, I don't think I can make an undeniable case about that). I am saying that if you look at the preponderance of the evidence about competition quality, about height of peak, about who the contemporaries were, about career, and everything, there's just no way Pete Browning has more merit than Eddie Murray.

That's all I have to say about that. Thanks for your patience in reading it; I hope at least that you feel, after reading it, that I have not misrepresented your position in order to make an example of you. I've tried in this post to clarify my position, to understand your position, and to offer a reasonable response to at least most of the points in your analysis.
   76. rawagman Posted: August 15, 2007 at 01:29 PM (#2485666)
Chris - I suppopse now would be the time for me to chime in on why I have placed Murray 4th, behind Sandberg, Duffy and Ben Taylor.
I don't have the numbers with me, so please bear that in mind.
The direct comparison for Murray was with Taylor and when I lined them up side by side, it was scary how close they really were. I decided to give Taylor the edge as I beleive that we are missing information on how good he probably was in the teens. The information that we have of him, the information that got him into the Hall of Fame and has sustained him near the top of my ballot for this long is mainly from the second half of his career.
That decision could have gone either way, but I am comfortable with Taylor ahead of Murray.
That being said, I now should say that I look at Murray as a long prime type. My system gives only minimal weight to counting numbers, and no real weight to uberstats (as I point out every week). I prefer the peak/prime types to the prime/career types. The peak/prime/career types are the inner circle. Not only do I prefer peak/prime to prime/career, but I also prefer glove types (such as Sandberg and Duffy) to pure bat types (although I do credit good 1B defense more than others). The types who will always rise in my system are those who combined a stick that was demonstrably better than league average with a glove that was heralded for excellence. Which is why I am Hugh Duffy's best friend.
Please also bear in mind that coming in 4th place, Murray is easily in what I consider to be an elect me spot. If Duffy were already elected, Murray would have come in 3rd. If Taylor were also already elected, Murray would be in 2nd.
   77. DL from MN Posted: August 15, 2007 at 01:56 PM (#2485701)
I'm not sure an 1890s outfielder could be considered a "glove type" especially when you are comparing to a terrific fielding 1B - Ben Taylor.

http://www.hardballtimes.com/main/article/historical-hitting-by-position/

That article says Hugh Duffy played in a time where CF was about the same as LF and RF in the defensive spectrum. The shift to CF as a glove position happens around 1930, the same time as the swap of 2B and 3B.

I'm probably not giving enough credit to modern CF, though I think it is minor.
   78. sunnyday2 Posted: August 15, 2007 at 01:59 PM (#2485704)
>I'm probably not giving enough credit to modern CF,

Kirby Puckett A+
   79. Chris Cobb Posted: August 15, 2007 at 02:11 PM (#2485718)
rawagman,

I wouldn't try to talk you out of Taylor over Murray. In my view, Taylor was on Murray's level as a hitter, but until I get off my butt and finish full MLEs for Taylor to lay out the case for that, I'm not in a position to make an argument, and my incomplete analysis could be wrong at this point, anyway.

For Duffy, there are two questions: how good a hitter was he, really, and how valuable was his stellar outfield defense (I don't think there's much question that his outfield defense _was_ stellar)?

It looks to me like your argument for having him #2 after Sandberg would be that they were highly similar as hitters, both stellar defenders at their positions, but Sandberg gets the nod because second base defense is more valuable than outfield defense.

For a ranking between those two players, I think that is sound analysis.

However, I would guess that you are overrating both as hitters, since if I recall correctly your (to my mind mistaken) rejection of uberstats even extends to OPS+. Sandberg and Duffy are both overrated by non-normalized stats because they played in great hitters' parks. They were very good hitters compared to all of their contemporaries, but there's no way I would say of Hugh Duffy, as you did on your ballot, "Amazing bat!" Sandberg was an amazing hitter _for a second baseman_, certainly the best at his position during his prime. Duffy was nowhere near the best hitter at his positions during his prime. Thus, Duffy isn't truly close to Sandberg in value, and, as he isn't close to Sandberg, he isn't close to Murray, either.

Since this is the Murray thread and we are talking the 1980s, a 1980s hitter who I see as a good comp for Hugh Duffy is Chet Lemon. Would you describe Chet Lemon as an amazing hitter? Would you see _any_ argument for ranking him ahead of Eddie Murray?

Now I think Chet Lemon was a terrific player, who really has not gotten the credit he deserves from the electorate, and though I don't think he's quite a HoMer, he could be getting votes. I would say the same things about Hugh Duffy. And about Cy Seymour, for that matter. What _real value_ does Hugh Duffy have that Chet Lemon and Cy Seymour don't? (That's not a rhetorical question, btw, and I think there are some positive answers to it, but I have to do some work this morning, so I'll leave that for later discussion, if rawagman and/or other folks are interested in taking the subject up.)
   80. Chris Cobb Posted: August 15, 2007 at 02:14 PM (#2485721)
Taylor was on Murray's level as a hitter

Correction to this misstatement: Taylor was _not_ on Murray's level as a hitter. The negative makes a difference in the argument . . . I don't think Taylor was better than Murray because he didn't hit as well for most of his career. In a few peak seasons, he may have hit as well as Murray, but not for the bulk of his career.
   81. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 15, 2007 at 02:14 PM (#2485722)
Chris Cobb--again, what you are saying in practice is that Browning played in extremely high standard deviation, easy-to-dominate leagues, while Murray played in extremely low standard deviation, hard-to-dominate leagues, enabling Browning to put up much higher OPS+ scores than Murray. My quantitative research into this supports your contention. Also, I will have John McGraw an easy #1 on my ballot, well ahead of both Murray and Sandberg. If you'd like to challenge me on that, the time is now, since I post my ballot on the final day.
   82. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 15, 2007 at 02:19 PM (#2485727)
Also DL from MN, thanks very much for posting Gassko's article; I hadn't seen it. It's nice to see he has found the exact same patterns as I have--including the narrowing of the 2B/3B and SS/C gaps in the modern era (Jim Sp, you and I have discussed this) as well as the SS drought of the 1970's. I may use his data to tweak my OF numbers pre-1930.
   83. TomH Posted: August 15, 2007 at 02:23 PM (#2485731)
I'm a staunch FOMugsy, but as has been pointed out previously, John McGraw ain't goin in to the HoM this week, and Murray will no longer be competing with Mugsy next week, so it isn't all that relevant to the project as a whole. Even if you lowered McGraw to 3rd [ where he belongs :) ], he would still get your elect-me bonus, so it's not as if you're giving him a pile of extra points by listing him #1.
   84. Chris Cobb Posted: August 15, 2007 at 02:25 PM (#2485733)
Also, I will have John McGraw an easy #1 on my ballot, well ahead of both Murray and Sandberg. If you'd like to challenge me on that, the time is now, since I post my ballot on the final day.

Thanks for the heads up. I'll think about it. Since I know how your ranking system works and that it is internally consistent, I doubt there is an argument that I could make that would be valid on its terms, and I am probably not up to arguing that you should rethink your system on the basis of this result.

if I have a question about the system that I think is worth looking into, it lies in how the rate set by salary, prorated to playing time works. If two players are equal in a season by total WAR2 earned, but one had a higher rate and less playing time than the other, how much does your setting salary by rate rather than production favor the high rate player? As a GM, I can see favoring rate because rate is an indicator of a player's upside, but in retrospective analysis I doubt that is a valid consideration. If value is understood to be _value above replacement_, then the two players' performances were functionally equivalent and ought to be recognized as such in a retrospective evaluation. As I say, I simply don't know how much effect the issue I am describing actually has in your system or how much effect it has on the case of John McGraw in particular. It may not matter at all, and if it does matter, you may disagree with my objection.
   85. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 15, 2007 at 03:09 PM (#2485796)
Chris Cobb, you are right that my salary estimator prefers half a season at a 6-win rate to a full season at a 3-win rate. My position, basically, is that just as I prefer 5 great seasons to 10 average ones, I prefer 5 great half-seasons to 5 average full seasons. Rate is always treated exponentially, and playing time is always treated linear-ly, whether it's games played within a season or number of seasons. I like that consistency.

That said, my estimator's preference for peak rate and ignorance of durability vs. longevity (for me, it's all just playing time) is not a deal breaker in McGraw's case. My arguments for McGraw are the following: his game (unprecedented OBP, speed, and defense) was perfectly suited to 1890's baseball in a way that BP's and WS's run estimators do not capture; his lack of playing time is excessively penalized by those systems due to their low replacement levels, and 3B in the 1890s was much tougher than the uberstats give it credit for. As a result, I have his total career standard deviation-adjusted wins above replacement equal to Frank Baker's. Using my rate-based salary estimator, I have McGraw at $124 million, Sandberg at $104M, and Murray at $102M. But if I apply the salary estimator to seasonal totals, rather than to rate times the fraction of the season played, I have McGraw at $107M, Sandberg at $104M, and Murray at $100M. Close, but Muggsy still comes out on top.
   86. DL from MN Posted: August 15, 2007 at 03:25 PM (#2485818)
> a 1980s hitter who I see as a good comp for Hugh Duffy is Chet Lemon

Actually, I think Kirby Puckett could be a fair comp - both hit for high average and were great defenders. Still, Kirby is no Eddie Murray.
   87. DL from MN Posted: August 15, 2007 at 03:30 PM (#2485826)
Are all the run estimators using linear assumptions or are there 2nd order effects computed?

I'm just about ready to swap using BP WARP to using Dan R WARP but there are 2 problems for me right now.

1) No pitchers - this is a huge sticking point for me
2) No WAA information - This goes to my definition of merit and is essential to my spreadsheet. The initial results I did using the Dan R data didn't make as much sense as I would like and I think an overall WAA is what would help align this for me. I can work around this one but it strikes me as a "nice to have".
   88. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 15, 2007 at 03:45 PM (#2485840)
DL from MN--

What do you mean there is no WAA information? The spreadsheet in the Yahoo group has batting wins above average, baserunning wins above average, and fielding wins above average, in both standard deviation-adjusted and raw versions. WAA is just the sum of those three values. Just bear in mind that all baserunning WAA are positive in the pre-CS era (with a reduced SB weight so the league runs still add up), and that my DH adjustment is applied to the replacement level rather than to the vs. average numbers, so you have to make that correction (0.6 Rosenheck standard deviation-adjusted wins per 162 games) manually to the vs. average numbers.

BP's run estimator is calculated via the ratio of a team's raw EqA (a number similar to OPS) to the league's raw EqA. WS is just based on RC, no? I use the non-linear BaseRuns for pre-integration seasons.

You might want to check with Joe Dimino, whose pitcher WAR numbers seem to be on a similar scale to mine. I can give you standard deviation adjustments for pitchers for every league-season since 1893.

I haven't come out with my own pitcher numbers because I am still struggling with some thorny conceptual issues (correcting for park and defense and translating innings and career lengths, mainly).
   89. DL from MN Posted: August 15, 2007 at 03:50 PM (#2485844)
Are the batting/fielding/baserunning WAA v. all batters/fielders/baserunners or vs the position?

I like Joe D's numbers, I'd like them more if they were integrated into the same spreadsheet.
   90. Chris Cobb Posted: August 15, 2007 at 04:00 PM (#2485855)
DL:

My situation was almost exactly the same as yours vis a vis Dan's WAR, and I have recently come up with work-arounds that are satisfactory for me, so they may work for you as well.

On wins above average, which I use in my system. Dan R provides information on how many wins above replacement an average player is at each position for each season. I figured that an average player overall would be the number of wins above replacement that is the mean of these averages. I calculate the wins above replacement of an average player for each season, then subract it from the player's WAR total to find wins above average. For example, in 1916, replacement level for catcher was -2.5 wins below average, 1B was -0.9, 2B -1.4, 3B -1.7, SS -3.0, LF/RF-0.7, and CF -1.1. Averaging these, I get 1.51 WAR as average, which I subtract from player's seasonal total to get WAA. I'm not sure this process is entirely sound methodologically, but the results seem sensible so far and have for me the sort of "aligning value" you say you are looking for. "Sensible" for me means that the results are not inconsistent with my understanding of merit and reasonably consistent with the best measurements of value that we have available.

On pitchers: I have found that if I scale my pitchers' numbers (which is based on BP WARP and home-grown win shares) to WAR, using them in tandem with WAR-based rankings for position players changes the relative ranking of the pitchers very little. Here's how I did the scaling. I have 3 measures in my system career wins above replacement, total wins above average (taken by summing wins above average in all seasons with above average totals) and peak rate (wins above replacement/162 games over 5 best seasons) times 5. I rank players against their contemporaries by summing these three numbers and putting the players in order from highest to lowest. Using the 1970s as a test case, I created totals for a bunch of players using WAR and added them up, then added up the WARP1 (and WS, since I also use WS) totals for the same players, found the difference between the totals, and divided by the number of players included to find the average difference between the results. It was 60 for both WARP1 and WS/3. So I can scale my pitchers' numbers for WARP1 and WS to my batters' numbers from Dan's WAR simply by subtracting 60 from their totals and interfiling the rankings.

(It's actually a bit more complicated than that, since I have not _replaced_ WARP and WS in my evaluation of position players with WAR, but I have added a heavily weighted WAR factor to my larger system. My totals for position players is now the sum of .25 the scaled WS total + .25 the scaled WARP1 total, + .5 the WAR total. For pitchers, the total is .5 the scaled WS total + .5 the scaled WARP1 total. It seems to work.)

This approach to scaling might be adapted to your system, though you'd need to find the right constant to fit your own system.
   91. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 15, 2007 at 04:06 PM (#2485859)
Batting WAA: versus all batters, including pitchers hitting in non-DH leagues.
Baserunning WAA: The SB/CS component is measured vs. a guy with 0 SB and 0 CS (so if the average player generates nonzero SB/CS runs, then the league average for this component will not be 0). The non-SB/CS component, which starts in 1972, is measured against an average baserunner.
Fielding WAA: versus all fielders at the same position.

The Rep column is the gap between the overall league average (a guy who hits, runs, and fields his unspecified position at the league average) and a replacement player at the given player's position in the given player's playing time. E.g., since Rep for SS is usually around 3.4 standard deviation-adjusted wins per 162 games, a SS who hits, runs, and fields at the league average and plays 81 games would have BWAA of 0, BRWAA of 0, FWAA of 0, and Rep of 1.7, for a total of 1.7 WARP).

Joe Dimino, are your numbers available in spreadsheet form? With Lahman identifiers? If so, I could merge them....
   92. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 15, 2007 at 04:09 PM (#2485860)
Chris Cobb, you can get the actual wins-above-average values in my updated spreadsheet in the Yahoo group.
   93. TomH Posted: August 15, 2007 at 04:23 PM (#2485885)
My arguments for McGraw are the following: his game (unprecedented OBP, speed, and defense) was perfectly suited to 1890's baseball in a way that BP's and WS's run estimators do not capture

While I agree that BP (EqA) misses this, Win Shares I believe does capture McGraw's "game" in his era properly; James' RC formulae was matched to the run environment. Do you have a reason why you assert this, Dan?
   94. Chris Cobb Posted: August 15, 2007 at 04:24 PM (#2485887)
Thanks. I should get the updated spreadsheet.

I'm not sure, though, how I would use the actual wins-above-average in my system in addition the the Replacement level number, which is the one that matters to the way I think about what overall average is.
   95. DL from MN Posted: August 15, 2007 at 04:30 PM (#2485899)
Great discussion. I believe marginal wins above average are worth more than marginal wins above replacement so I'm essentially giving extra credit for wins above average. The inconsistency in my result seems to be from the fielding WAA being position adjusted. An average fielding SS who is a league average hitter and baserunner is getting no extra credit when I add things up currently even though that player is _clearly_ an above average player. If I add in Rep to the BWAA, FWAA, BRWAA numbers, does that get what I want or should I be looking at WAR-Rep and not considering the component numbers?
   96. DL from MN Posted: August 15, 2007 at 04:43 PM (#2485921)
About my thought process - this goes to a John McGraw type player. If I'm a GM do I want the average player (assume 4 WAR) or do I want a guy who gives me 4 WAR in half a season and nothing the rest of the year? A rational GM would choose the average player and use the other roster spot upgrading another part of the roster. Due to the nature of the bell curve, each win above average gets harder and harder to find (though I think it's geometric, not exponential). Wins between replacement and average can probably be approximated with a straight line. Also, I don't want to reward players for position scarcity if they're really only an average ballplayer (see Campaneris) just because player development people converted all the good athletes to centerfielders. Droughts in one position seem to line up nicely with surpluses in other positions.
   97. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 15, 2007 at 04:53 PM (#2485939)
TomH, well, it still generates very different results. An average player gets .48*3*81/9 = 13 batting win shares per 162 games, right? So in 1899 McGraw played in 117 games and had 26.1 batting WS, meaning he was (26.1-(13*117/162))/3 = 5.6 offensive wins above average, which adjusts for season length to 5.9 offensive wins above average. I have McGraw with 7.7 offensive wins above average that year. So there's still a big discrepancy...I can easily walk anyone who's interested through my methodology for calculating McGraw's 7.7 season length-adjusted BWAA+BRWAA for 1899. Could a WS expert perhaps do the same for his 26.1 batting WS?

DL from MN, WARP minus Rep is equal to the sum of BWAA, FWAA, and BRWAA, there's no difference. It sounds like what you are looking for is wins above positional average. I don't track that number, since I don't like it--it makes guys like Trammell, Tejada, or Greenberg look much worse than they really were--but you should be able to calculate it pretty straightforwardly if you want. Just add the total BWAA, FWAA, and BRWAA for each position in a given league-season and divide that by the total SFrac of the players at that position to get positional WAA per 162 games for that league-season. Then just multiply that by a given player's SFrac and subtract the resulting value from his BWAA + FWAA + BRWAA to get wins above positional average.
   98. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 15, 2007 at 05:11 PM (#2485977)
By the way, I think I figured out why BP gets guys like McGraw wrong. It first translates UEQR, the result of its run estimator, to EQR, which is scaled to a 4.5 R/G, 2.0 Pythagorean exponent league. Then it calculates how many "standard league" runs an average player would have generated in the same number of outs and subtracts to get BRAA. The problem is that the win value of out avoidance changes with the run environment--OBP is much more important than SLG in a high-scoring league, much less so in a low-scoring league. So this will cause it to overrate Buck Freeman and underrate McGraw in 1899, and should theoretically cause it to overrate Ron Hunt and underrate Frank Howard in 1968.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
BDC
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.1554 seconds
49 querie(s) executed