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

Thursday, February 04, 2021

Most Meritorious Player: 1935 Discussion

The Detroit Tigers defeated the Chicago Cubs 4 games to 2 in the World Series.

The Pittsburgh Crawfords defeated the New York Cubans 4 games to 3 in the East-West championship series.

Vote for 10.

Player				bWAR
Arky Vaughan			9.7
Jimmie Foxx			8.1
Lou Gehrig			8.7
Mel Ott				7.3
Hank Greenberg			7.7
Charlie Gehringer		8.5
Buddy Myer			6.4
Augie Galan			5.9
Billy Herman			7.0
Luke Appling			7.0
Joe Vosmik			5.6
Joe Medwick			6.0
Gabby Hartnett			4.9
Mickey Cochrane			5.1
Ripper Collins			5.1
Bill Rogell			5.4
Stan Hack			4.9
Bill Terry			5.0
Paul Waner			3.8
Jo-Jo Moore			3.2
Red Rolfe			4.1
Hank Lieber			4.6
Wally Berger			5.8
Pete Fox			4.2
Cecil Travis			3.6
George Selkirk			4.1
Ernie Lombardi			3.7
Bill Dickey			3.0
		
Josh Gibson			2.6
Ray Brown			4.0
Martin Dihigo			3.0
Cool Papa Bell			2.2
Willie Wells			2.7
Buck Leonard			2.1
Turkey Stearnes			1.6
Mule Suttles			1.4
Leroy Morney			1.3
Rap Dixon			1.3
Lazaro Salazar			1.0
Ray Dandridge			0.8

Pitcher				bWAR
Wes Ferrell			8.2
Lefty Grove			9.4
Dizzy Dean			7.1
Cy Blanton			7.2
Curt Davis			7.1
Mel Harder			7.3
Schoolboy Rowe			3.8
Bill Swift			4.9
Syl Johnson			5.2
Pat Dean			5.7
Ivy Andrews			6.4
Red Ruffing			3.8
Hal Schumacher			4.6
Willis Hudlin			5.1
Johnny Marcum			4.1
Lon Warneke			4.2
Carl Hubbell			4.7
Ted Lyons			4.9
Larry French			3.9
John Whitehead			4.9
Lefty Gomez			4.1
Tommy Bridges			3.1

Satchel Paige			2.3
Leroy Matlock			3.7
Willie Foster			2.6
Leon Day			2.1
Ted Trent			1.9
Roosevelt Davis			2.9
Johnny Taylor 			2.0
DL from MN Posted: February 04, 2021 at 01:36 PM | 37 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. DL from MN Posted: February 04, 2021 at 01:46 PM (#6003764)
Unfortunately Baseball Gauge has shut down so I don't have an easy source for Win Shares. In the interest of time I created the thread without that data.
   2. DL from MN Posted: February 04, 2021 at 02:42 PM (#6003782)
Dr C NGL MLEs
Player Name Age Lg Pos G PA Rbat Rbaser Rdp Rfield Rpos RAA WAA Rrep RAR WAR
Willie Wells 29 NL SS 151 650 27 1 0 6 9 43 4.2 20 63 6.3
Turkey Stearnes 34 NL CF 151 650 37 0 0 2 -2 38 3.8 20 59 5.9
Josh Gibson 23 NL 1B/C 147 630 31 0 0 4 -4 31 3.0 20 50 5.1
Sam Bankhead 24 NL SS 144 620 11 2 0 4 9 27 2.7 19 46 4.7
Martin Dihigo 30 NL CF 140 600 13 1 0 7 -2 18 1.8 19 37 3.8
Buck Leonard 27 NL 1B 133 570 19 0 0 2 -6 16 1.6 18 34 3.5
Cool Papa Bell 32 NL CF 147 630 15 2 0 -1 -2 14 1.4 20 34 3.4
Mule Suttles 34 NL 1B 135 580 19 0 0 2 -6 15 1.5 18 34 3.4
Lazaro Salazar 22 NL CF 142 610 8 1 0 1 -2 9 0.9 21 30 3.0
Newt Allen 34 NL 2B 140 600 -7 0 0 14 4 11 1.1 19 30 3.0
George Scales 34 NL 3B 121 520 7 0 0 -3 2 6 0.6 16 22 2.2
Jud Wilson 39 NL 1B 82 350 12 0 0 0 -3 9 0.9 11 20 2.0
Oscr Charleston 38 NL 1B 102 440 6 1 0 1 -4 4 0.4 14 18 1.8
Dick Lundy 36 NL 2B 119 510 -2 0 0 0 4 1 0.1 16 17 1.8
Ray Dandridge 21 NL 3B 82 350 2 0 0 2 1 6 0.6 11 17 1.7
Burnis Wright 21 NL CF 114 490 3 1 0 -1 -1 1 0.1 15 16 1.6
Alejandro Oms 39 NL RF 47 200 7 0 0 1 -2 6 0.6 7 13 1.3
Biz Mackey 37 NL C 65 280 -3 0 0 1 2 -1 -0.1 9 8 0.8
Quincy Trouppe 22 AL C 50 220 -6 0 0 0 1 -5 -0.5 7 2 0.2

Pitcher Name Age Lg G IP R RA9 lgRA9 RAA WAA pWAR Rrep RAR PA Rbat Rpos RAA bWAR WAR
Lazaro Salazar 22 NL 38 190 30 1.43 4.71 69 7.4 9.1 19 88 63 1 6 7 0.5 9.6
Satchel Paige 28 AL 41 270 109 3.62 5.14 46 4.5 7.3 29 74 90 -8 9 2 0.1 7.4
Willie Foster 34 NL 32 200 79 3.54 4.71 26 2.7 4.7 20 46 67 -8 7 -1 -0.2 4.6
Martin Dihigo 30 NL 38 190 86 4.10 4.71 13 1.3 3.2 19 31 63 1 6 7 0.5 3.7
Ramon Bragana 26 NL 38 250 124 4.47 4.71 7 0.7 3.3 24 31 83 -3 8 5 0.3 3.6
Ray Brown 26 NL 39 230 115 4.50 4.68 5 0.5 2.9 22 27 77 -2 9 7 0.5 3.3
Hilton Smith 28 NL 35 220 111 4.54 4.71 4 0.4 2.7 22 26 73 -1 7 7 0.4 3.1
Wbstr McDonald 35 NL 41 270 148 4.93 4.71 -7 -0.7 2.2 26 20 90 -9 9 0 0.0 2.1
William Bell 37 NL 34 150 116 6.95 4.71 -37 -3.5 -1.9 15 -23 50 -2 5 4 0.2 -1.7
Bill Byrd 27 NL 33 170 129 6.84 4.71 -40 -3.8 -1.9 17 -24 57 -2 6 4 0.2 -1.7

   3. DL from MN Posted: February 04, 2021 at 02:53 PM (#6003783)
In the spring of 1935, Greenlee refused Paige's request to raise his $250 per month salary, so Paige decided to return to Bismarck for the same $400 per month and late model used car that he got before. Churchill added other Negro league players to the team—pitchers Barney Morris, and Hilton Smith, catcher Quincy Trouppe, and pitcher/catcher Double Duty Radcliffe. Paige dominated the competition, with a 29–2 record, 321 strikeouts, and only 16 walks. In Wichita, Ray "Hap" Dumont was establishing a new national baseball tournament, the National Baseball Congress. Dumont invited 32 semi-pro teams, paying $1,000 for Paige and his Bismarck teammates to attend. The tournament was held at Lawrence–Dumont Stadium in Wichita, Kansas and offered a $7,000 purse. Churchill added yet another Negro league star to his team—Chet Brewer, the Kansas City Monarchs' ace pitcher. Bismarck swept the tournament in seven straight games. Paige won the four games he started, pitched in relief in a fifth game, and struck out 60 batters—a record that still held 74 years later.[52][53][54]

In September, Paige could not return to the NNL because he was banned from the league for the 1935 season for jumping to the Bismarck team. J. L. Wilkinson, owner of the independent Kansas City Monarchs, signed Paige on a game-by-game basis through the end of the season.[55]
- Wikipedia

Paige is at the top of his abilities in 1935 but isn't playing much NGL baseball because the pay is better elsewhere.

   4. DL from MN Posted: February 04, 2021 at 02:56 PM (#6003785)
Salazar still doesn't have many documented NGL innings so that MLE has to be entirely based on Cuban Winter League stats.
   5. DL from MN Posted: February 04, 2021 at 02:59 PM (#6003786)
Ray Brown's MLE is looking at just his pitching and not giving him credit for a great year in CF when he wasn't on the mound. He was a 2 way threat in 1935.
   6. DL from MN Posted: February 04, 2021 at 04:21 PM (#6003808)
1935 Prelim

1) Arky Vaughan - big gap between him and the rest of the field. Average glove but 190 OPS+ from a SS.
2) Wes Ferrell - led the league in batters faced (with great results), hit very well. 35 pinch hit appearances.
3) Jimmie Foxx - 2nd best bat (tie with Gehrig) and decent fielding statistics
4) Mel Ott - great defensive season for Ott
5) Lou Gehrig - a little less fielding than Foxx. 3-5 are very close
6) Josh Gibson - best NGL player, best C
7) Lefty Grove - Best pitcher in baseball, didn't hit like Wes Ferrell and 50 fewer innings
8) Satchel Paige - Best NGL pitcher
9) Hank Greenberg - MLB talent has shifted to 1B
10) Charlie Gehringer

11-15) Martin Dihigo, Buddy Myer, Dizzy Dean, Babe Herman, Gabby Hartnett
16-20) Augie Galan, Luke Appling, Willie Wells, Mickey Cochrane, Johnny Vosmil
21-25) Turkey Stearnes, Curt Davis, Joe Medwick, Mel Harder, Ray Brown
   7. DL from MN Posted: February 04, 2021 at 04:28 PM (#6003812)
1935 Pittsburgh Crawfords vs New York Cubans (Negro National League)
The Pittsburgh Crawfords had the best record in the league with a 42-15 (.737) record and easily
out distanced the competition during the 1935 baseball season. The Columbus Elite Giants had
the second best won-loss record with a 28-21 (.571) record and the New York Cubans finished in
third place with a 29-24 (.547) record. Pittsburgh won the first half of the season and the New
York Cubans rebounded from a dreadful start to win the second half of the season.

Game one of the Play-Off Series was held in Paterson (New Jersey) and played on September 13th. The New York Cubans started Frank Blake who had only seen limited action during the regular season. Blake responded to Dihigo’s confidence in him and pitched a four hitter for 9-3 victory.
Game two of the series was moved to New York City and pitted John “Neck” Stanley (New York) against Sam Streeter (Pittsburgh). Stanley silenced the powerful Crawfords bats and coasted to a four hit 4-0 shutout victory.
Crawfords ace Leroy Matlock who was undefeated during the season took the mound for game three. Leroy pitched a five hit 3-0 shutout with Johnny Taylor taking the loss for New York.
Cuban manager Martin Dihigo named himself the starter for game four which was played in Pittsburgh on September 18th. Leroy Matlock started game four for the Crawfords on only two days rest. The Cubans scored a 6-1 victory to win the game and end Leroy Matlock’s pitching streak at 22 wins (4 in 1934 , 17 in 1935 and 1 in the play-offs) without a loss.
The Pittsburgh Crawfords entered game five was a must win situation. Pittsburgh scored the winning run in the bottom of the ninth inning to take a 3-2 victory over New York. Roosevelt Davis was the winning pitcher and Frank Blake took the loss.
Down 6-3 going into the bottom of the ninth of game seven, the Pittsburgh Crawfords scored four runs off Martin Dihigo for a 7-6 victory. The win by Pittsburgh evened up the series at three games apiece. Game seven was a slugfest and went back and forth. It was highlighted by four homeruns. The Cubans
got two (one each from Rap Dixon and Clyde Spearman) and the Crawfords got two (one each from Oscar Charleston and Josh Gibson). Pittsburgh scored three runs in the top of the ninth inning to take an 8-5 lead. New York could only score two runs in the bottom of the ninth to come up one run short and an 8-7 loss. The victory by Pittsburgh gave them the Negro National league championship title.

Top Hitters for the Series:
Pittsburgh – Josh Gibson (.407), Oscar Charleston (.370) and Sam Bankhead (.269).
New York – Clyde Spearman (.345), Rap Dixon (.333) and Showboat Thomas (.323).
Leading Pitchers for the Series:
Pittsburgh – Roosevelt Davis (2-0), Jimmie Crutchfield (1-0) and Leroy Matlock (1-1).
New York – Neck Stanley (1-0), Frank Blake (1-1) and Martin Dihigo (1-1).


Source
   8. Michael J. Binkley's anxiety closet Posted: February 05, 2021 at 08:44 PM (#6004046)
1935 Win Shares

Arky Vaughan - 39.5
Mel Ott - 34.8
Wes Ferrell - 34.3
Lou Gehrig - 34.3
Hank Greenberg - 33.2
Augie Galan - 32.9
Joe Medwick - 32.3
Buddy Myers - 32.1
Billy Herman - 31.9
Charlie Gehringer - 31.1
Dizzy Dean - 30.0
Jimmie Foxx - 29.9
Lefty Grove - 29.4
Joe Vosmik - 27.8
Hank Lieber - 27.7
Ripper Collins - 27.4
Mel Harder - 27.1
Gabby Hartnett - 25.7
Carl Hubbell - 24.7
Cy Blanton - 24.0
Jo-Jo Moore - 23.9
Luke Appling - 23.5
Mickey Cochrane - 23.5
Hal Schumaker - 23.5
Bill Terry - 23.3
Schoolboy Rowe -23.0
Bill Rogell - 22.5
Pat Dean - 22.4
Paul Waner - 22.1
Stan Hack - 22.1
Wally Berger - 22.1
Red Rolfe - 21.9
Earl Averill - 21.9
Lon Warneke - 21.5
Willis Hudlin - 21.0
Pete Fox - 20.9
Paul Derringer - 20.8
Ivy Andrews - 20.8
Ben Chapman - 20.8
Red Ruffing - 20.7
Curt Davis - 20.5

If anybody has requests for any other 1935 WS numbers for certain players, let me know.
   9. kcgard2 Posted: February 06, 2021 at 11:47 AM (#6004100)
prelim

1. Arky Vaughan: not even that much of his almost 10 WAR this year was from defense. 194 wRC+ from a plus defensive shortstop
2. Wes Ferrell: lots of innings at a good not great rate, the bat puts him over Grove
3. Lefty Grove
4. Lou Gehrig
5. Jimmie Foxx
6. Hank Greenberg (is the whole league just 1B??)
7. Josh Gibson
8. Satchel Paige
9. Mel Harder
10. Mel Ott

11-21) Dizzy Dean, Charlie Gehringer, Billy Herman, Cy Blanton, Willie Wells, Martin Dihigo, Luke Appling, Joe Medwick, Turkey Stearnes, Augie Galan, Curt Davis
   10. DL from MN Posted: February 07, 2021 at 01:18 PM (#6004200)
Leroy Matlock has a good argument as a rival to Satchel Paige this season. This is his career year. Alex King's MLE from 2011 gives him 9.3 WAR. Chris Cobb's first MLE was 41 Win Shares. https://www.baseballthinkfactory.org/hall_of_merit/discussion/leroy_matlock

Dr. C's MLE for Matlock is a bit more "smoothed" out (https://homemlb.wordpress.com/tag/leroy-matlock/)

Leroy Matlock
Negro Leagues Stats | Mexican League Stats | Bio
Career: 1929–1942
Destination: NL 1929–1942
Missing data: 1929–1931, 1937, 1939
               PITCHING          |   BATTING   |  TOTAL
YEAR  AGE   IP  RAA   WAA   WAR  |   PA   WAR  |   WAR
========================================================
1929   22  150    6   0.6   2.2  |   50   0.1  |   2.3
1930   23  180    6   0.5   2.5  |   60   0.1  |   2.6
1931   24  220    7   0.8   3.0  |   73   0.1  |   3.1
1932   25  210    5   0.5   2.7  |   70   0.1  |   2.8
1933   26  240   15   1.7   4.1  |   80   0.1  |   4.1
1934   27  230   19   1.9   4.3  |   77   0.1  |   4.4
1935   28  220   39   4.2   6.4  |   73   0.1  |   6.5
1936   29  270   31   3.3   6.0  |   90   0.2  |   6.2
1937   30  220    7   0.7   2.9  |   73   0.1  |   3.1
1938   31  220  - 9  -0.9   1.4  |   73   0.1  |   1.5
1939   32  210    5   0.5   2.7  |   70   0.1  |   2.8
1940   33  230   15   1.6   3.9  |   77   0.1  |   4.0
1941   34  190    8   0.9   2.8  |   63   0.1  |   2.9
1942   34  180  - 3  -0.4   1.4  |   60   0.1  |   1.5
-------------------------------------------------------
TOTAL     2970  151  16.0  46.2  |  989   1.5  |  47.7


   11. DL from MN Posted: February 07, 2021 at 01:30 PM (#6004203)
Matlock will be near Gehrig on my ballot.
   12. bjhanke Posted: February 09, 2021 at 06:53 PM (#6004500)
MJB anxiety - Where did you get those Win Shares from? Everybody from 1935 has WS printed in Win Shares, but those don't have a decimal point number. I just found out that you can get WS, including both Offensive and Defensive WS, at Bill James Online. Did you get them there? In any case, thanks for the list, Mr. Brinkley!

I should add that, this year and until the problem is fixed somehow, I am going to be weighing Win Shares MUCH more strongly than WAR, because WAR is badly broken and Defensive WAR is broken even worse than Offensive. Right now, WAR is trying to claim that 1) They can get away with using DIFFERENT Replacement Values for offense and defense. That's badly broken right there. 2) But also they have decided, somehow, that the Replacement Rate (which is, of course, also the zero point in their math) on defense is actually right at .500!!! This is just the Linear Weights mistake all over again. A .500 Zero Point is catastrophic to any system, just as it is for Linear Weights. I'm not going to get into their Positional Adjustments, because it would involve a separate and long analysis.

But I am going to count Win Shares twice to WAR's one this year, and every year until WAR gets its act together.
   13. Eric J can SABER all he wants to Posted: February 09, 2021 at 11:29 PM (#6004531)
Right now, WAR is trying to claim that 1) They can get away with using DIFFERENT Replacement Values for offense and defense. That's badly broken right there. 2) But also they have decided, somehow, that the Replacement Rate (which is, of course, also the zero point in their math) on defense is actually right at .500!!

That's not how replacement level works. Replacement level isn't a level of offense or defense; it's a level of overall ability combining both. You can have a replacement level player who's average as a hitter and a lousy fielder, or vice versa, or someone who's below average at both.

I can understand not wanting to use oWAR and dWAR as presented on B-R, because they are kind of weird (especially in that they don't add up to WAR, because both include the position adjustment). But WAR isn't intended to measure purely offense or purely defense (which is why it behaves strangely when you try to force it in that direction); it's meant to measure the player as a whole, in comparison to the level of other entire players who you can pick up on the cheap.
   14. DL from MN Posted: February 10, 2021 at 09:28 AM (#6004556)
until WAR gets its act together


I wouldn't hold my breath. I think you are misunderstanding how WAR works. I wish they wouldn't even list oWAR and dWAR since I think it confuses people.

For position players the components of WAR (pretty much any system) are RBat, RField and RPos. RBat is versus an average hitter. RField is versus the average fielder at that position. RPos is where the replacement value adjustment ends up. There's also RBaseR (baserunning vs average) and RDP (GIDP vs average) but those are smaller components.

Waiting "until WAR gets its act together" doesn't even make sense.
   15. Michael J. Binkley's anxiety closet Posted: February 10, 2021 at 02:03 PM (#6004589)
Brock-

I got the WS numbers with decimals from the now-defunct thebaseballgauge.com. Before its demise, it had excel downloads available of bWAR, gWAR, WS, and any customizable metric you might have created from their numbers for 1871-2019. DL just asked me to post some of the top WS performers for 1935.

I can send you the full WS file if you are interested. I believe I still have your email address.
   16. bjhanke Posted: February 12, 2021 at 09:29 AM (#6004837)
Michael - Thanks! I have found a source for Win Shares, but it isn't robust in terms of how many kinds of sorts you can do. But it is actually on Bill James Online, so it's probably Bill's Win Shares as opposed to BaseballGauge's, which I understand are not exactly the same. You CAN quickly summon up the page for any player, and it will give you his whole career. I would, however, be REALLY happy if you emailed that file to me. It would save me some serious work.

Eric and DL - I know what WAR is trying to do. But there are consequences to naming a Replacement Rate. One of those is that your Replacement Rate, whatever it is, becomes the zero point in your mathematical system. You can't get out of that. And setting the Replacement Rate too high leads to problems. For one, you will end up with players whose numbers say that, if they had retired 3 or so years before they did, they would have had better CAREERS. I've been working up a comparison of Bobby Abreu to Lou Brock (Abreu isn't doing well). But BOTH of them suffer from this problem, at least on BB-Ref. Both players would have higher WAR numbers if they had retired three years before they actually did. Brock's last three years include two negative numbers and one positive one, but the positive one isn't large enough to outweigh the negatives, so he shows as, very literally, having HARMED his teams by playing during those last three years. The same thing is true of Abreu, although his numbers have one quixotic item - he didn't actually play, if I'm reading the numbers right, at all, at least in the Majors, the year before his last year in the Majors. That year is surrounded by two negative ones. So, again, he would have a better career if he had retired three years before he did. But what's worse, that year that he did not play, because it counts as zero, is the BEST year of the three, according to WAR. It's zero; the other two are both negative. This happens (or used to happen, back when I was buying Total Baseball every year) a lot in Linear Weights, because the zero point is .500. WAR doesn't have the zero point set THAT high, but it is set high enough so that the career ending problem exists.

I agree with the oddity of oWAR and dWAR and the positional adjustment (let's call it pWAR). The oddest thing is that it isn't necessary. What you do to find the Final WAR (fWAR) is to add oWAR and dWAR together and then subtract out pWAR, because it's included in both oWAR and dWAR. What I want to know is what it's doing in oWAR at all. Positional adjustments apply to defense; no one hits as a second baseman, everyone hits as a hitter. If you just didn't include pWAR at all in oWAR, you could save the subtract-out part of the process, and your oWARs would make more sense, since they would not be complicated by this irrelevant defensive adjustment.

What I mean by WAR getting its act together is that the .500 zero point in defensive analysis is obviously wrong (it's the Linear Weights problem), and having a different Replacement Rate for offense and defense is indefensible, and eventually the people who do WAR will recognize this, and then they will do something to fix it. That's what I mean by getting its act together. Right now, it's a mess.

I know nothing except the basic concepts about how WAR deals with runs, except that they apparently reconcile everything to the team's Pythagorean W/L, instead of the team's actual W/L. I do know that RBat, RField, etc, are measured in runs, not wins, and so some calculation must be going on that renders the runs into wins. It's the Wins that make no sense.
   17. DL from MN Posted: February 12, 2021 at 02:48 PM (#6004895)
you will end up with players whose numbers say that, if they had retired 3 or so years before they did, they would have had better CAREERS


You can always zero out those years when making the comparison. Every year there SHOULD be players who get playing time where the numbers say they hurt their team with their performance. The Orioles would have been better off cutting Chris Davis in 2017, for example. Hall of Fame caliber players often get to play beyond when they are able to contribute because they have name recognition that will drive ticket sales (Ichiro) or they're still playing under a big contract and they're good enough to be a bench player (Pujols). There are always teams that aren't actively trying to win the pennant that year but they still have to field a team. Why not bring in a famous player who might be able to pass a milestone and sell some tickets?
   18. DL from MN Posted: February 12, 2021 at 02:54 PM (#6004900)
What you do to find the Final WAR (fWAR) is to add oWAR and dWAR together and then subtract out pWAR, because it's included in both oWAR and dWAR.


No, what you do to find WAR is add RBat + RField + RBaseR + RDP + RPos. That gives Runs above Replacement which is converted by a formula from Runs to Wins. oWAR and dWAR are customized reports of the components, not actual inputs to the calculations. Quit looking at oWAR and dWAR; look at RBat and RField, since that's what you're really interested in.
   19. Michael J. Binkley's anxiety closet Posted: February 12, 2021 at 05:46 PM (#6004924)
Brock - I sent the file to you. Let me know if you didn't get it.
   20. Eric J can SABER all he wants to Posted: February 12, 2021 at 06:38 PM (#6004934)
What I mean by WAR getting its act together is that the .500 zero point in defensive analysis is obviously wrong (it's the Linear Weights problem), and having a different Replacement Rate for offense and defense is indefensible, and eventually the people who do WAR will recognize this, and then they will do something to fix it. That's what I mean by getting its act together. Right now, it's a mess.

But none of this is meaningful to WAR as a whole; it only matters to oWAR and dWAR, which, as mentioned, are unnecessary. WAR as a whole doesn't have different zero points for offense and defense because it doesn't apply the replacement-level baseline until offense and defense have already been combined.

Also, a side question: How exactly do you set offense and defense (or to be more specific, batting and fielding) to the same baseline, if that baseline is not average? They are completely different tasks that operate in different directions and under different sets of circumstances. If you were to set them to a common, non-average baseline (which again, WAR is not doing because it only applies the baseline to the player's entire contribution), what would it be and how would it be applied?

For one, you will end up with players whose numbers say that, if they had retired 3 or so years before they did, they would have had better CAREERS.

As DL said in 17, you can zero out the negative seasons if you want; that's what I do. (I have a whole spiel on this if anyone is interested.) The negative values exist for a reason and I agree that they should, but I also don't consider them when evaluating an individual player's career by WAR, because they reflect more on the team than on the player.

On top of that, though, the negative WAR scores you occasionally see from players who had good careers aren't usually especially large; Brock has one of the larger negative totals I can remember, and they come out to all of -3 spread over three years ('61, '77, '78). 3 WAR over the course of more than 11,000 plate appearances is basically rounding error.
   21. bjhanke Posted: February 13, 2021 at 02:24 AM (#6004957)
DL - No, I'm not really looking for anything stated in Runs, because I do not know the conversion formula to get from Runs to Wins. I have to look at the Wins. And it's the right place to look, anyway, because the WAR are the end product. And, if you look at the Wins, what you end up with is that Final WAR is computed as oWAR + dWAR - Positional WAR = Final WAR. Translating that to a formula that gives you the Positional WAR (which are not displayed on the player's page, at least anywhere that I can find) is just algebra. The Final WAR computation I listed above is oWAR + dWAR - pWAR = fWAR. It's just algebra that gets you to oWAR + dWAR - fWAR = pWAR.

The results can be astonishing. I was looking at Lou Brock, and I realized that he has three contemporary left fielders who are in the Hall and whose careers are very close contemporaries of Brock's: Yaz, Stargell, and Billy Williams. And everyone wants to compare Brock to Rickey Henderson and Tim Raines, who played later - in the 1980s and 90s, but who were even better at stealing bases than Lou was. Here are the results of the pWAR computations, listed in order of pWAR (they are all negative because Left Field is not the most difficult position to play, but it's obvious that the higher the number, the harder LF was to play):

Rickey: -8.3
Raines: -8.7
Stargell: -12.9
Brock: -13.2
Williams: -14.2
Yaz: -19.2

What do these numbers say? Well, they are supposed to measure how difficult Left Field is to play on defense. My friend Don Malcolm thinks they may be ballpark adjusted, and I tend to believe Don when he says things like that. But, still. These are supposed to be measures of how difficult Left Field was to play in the player's time period. So, I guess Left Field must have been MUCH harder to play in the 1980s and 90s than the same position was in the 1960s and 70s. I have never seen this massive change in difficulty noted anywhere else. Except for in pWAR, it's completely undocumented. And I don't believe it for a second. Also, the numbers say that the EASIEST Left Field was Fenway Park in the 1960s and 70s. Left Field in Fenway Park does not hold a reputation for being easy to play. There's a concept called "playing the monster" that says the exact opposite.

At this point, I'm probably better off just telling you where to find the Bill James essay that goes through the problems of using the average as your mathematical zero point, using Linear Weights as the model of a .500 zero point. It's in the book Win Shares, on page 102. The essay itself is untitled, but it's the first essay in the Whys and Wherefores section of the book. It makes the case against using .500 in some detail. I can't really copy the whole 4 pages of this essay (for one thing, BTF won't allow a comment that long), so I just have to give you the place where you can find it. I am fully aware that there are people here who are nowhere near as old as this 73-year-old geezer, and that Win Shares has not been reprinted since its release in 2000. Hopefully, anyone who is interested either has a copy or can find someone who has one and will lend it to them. Best I can do, right now.
   22. kcgard2 Posted: February 13, 2021 at 08:43 AM (#6004962)
Brock,

These guys did not have equal playing time in the corner OF, although you are glossing over that when trying to make a point about pWAR. Rickey had 400 fewer innings in the corner outfield, and 2600 more innings in CF than Brock did. That is a pWAR swing of roughly 3 wins. Raines had 3800 fewer corner outfield innings than Brock did, again a pWAR value of roughly 3 wins. Just accounting for their playing time gets them basically to a rounding error with Brock on pWAR.

As Eric tried to explain here, and I tried to explain in a previous MMP thread, there is no such thing as replacement level on defense (or offense), there is only replacement level as a player overall. You can be replacement by being a godawful corner outfielder who hits well, or a Mendoza level hitter who is a genius defensive catcher, or anything in between. If you set replacement level defense as "bad corner outfielder" and replacement level offense as "bad middle infielder," as you wish to do, what you will get as a result is not a replacement level player. You get a guy who washes out in A ball. This is patently obvious. Would you like to address this observation?

Replacement level defense does not make sense as a concept (in the way that you wish it to). If you insist on defining it in isolation, you would be forced to recognize that among the pool of freely available players, in aggregate they would be able to play defense at a roughly league average level. Therefore, the only reasonable "replacement level" on defense would be average (again if you insisted on deciding replacement level of only one facet of the game which doesn't make sense). Teams don't replace a player's defense, they replace the whole player. If you replace his defense, you are also replacing his offense. Thus you can come out ahead by replacing his defense with worse defense, as long as the offense more than offsets that loss. Replacement level therefore refers to a total contribution in runs from both offense and defense.

In football, there would be a replacement level for offense, and a replacement level for defense.
   23. DL from MN Posted: February 13, 2021 at 09:18 AM (#6004964)
I do not know the conversion formula to get from Runs to Wins


It's roughly 10. RAR and WAR are in columns right next to each other on the player's page.

https://www.baseball-reference.com/about/war_explained_runs_to_wins.shtml

oWAR + dWAR - pWAR = fWAR


Not at all.

oWAR = RBat + RBaseR + RDP + RPos + RRep --> run through the runs to wins calculator
dWAR = RField + RPos (Does NOT include RRep) --> run through the runs to wins calculator

In your calculation what you have listed as "pWAR" would be RPos ONLY, not RRep which gets put on the oWAR side but NOT on the dWAR side.

What do these numbers say? Well, they are supposed to measure how difficult Left Field is to play on defense.


No, your numbers above are strictly the positional adjustment for those particular fielders based on games played at various positions. It has nothing to do with how difficult it is to play defense but instead how difficult it is to find a left fielder who will give you a total offensive and defensive contribution above replacement level. It's a measure of scarcity of talent and it is calculated mostly by how much OFFENSE teams are willing to give up to fill the particular position.

From the definition of WAR:
WAR for position players has six components:

Batting Runs
Baserunning Runs
Runs added or lost due to Grounding into Double Plays in DP situations
Fielding Runs
Positional Adjustment Runs
Replacement level Runs (based on playing time)


There is a lot more information here:

https://www.baseball-reference.com/about/war_explained.shtml

This feels like an argument made in bad faith. Lots of claiming "I don't know" when the links are readily available and the information is out in the open, appeals to authority (Bill James), misrepresenting what the numbers are trying to say and then arguing against a straw man instead of what the numbers actually represent. I don't think there is some agenda here, other than maybe liking Lou Brock too much and trying to find a system that makes him look good.
   24. Mefisto Posted: February 13, 2021 at 09:42 AM (#6004967)
I think it's unfair to accuse Brock of bad faith. I think instead that he's let James' (bad) arguments influence him when he shouldn't. Is Bill arguing in bad faith when it comes to WAR? Well let's just say his bias in favor of his own system heavily influences his writings.
   25. Eric J can SABER all he wants to Posted: February 13, 2021 at 05:42 PM (#6005006)
It has nothing to do with how difficult it is to play defense but instead how difficult it is to find a left fielder who will give you a total offensive and defensive contribution above replacement level. It's a measure of scarcity of talent and it is calculated mostly by how much OFFENSE teams are willing to give up to fill the particular position.

I was about to say that this was wrong, but then I looked back at the WAR explainer; they must have changed this at some point, because the positional adjustment definitely used to be based solely in changes in defensive performance among players who played multiple positions. That must at least still be a strong factor, because center fielders outhit right fielders for several seasons in the '50s and the difference in adjustment between the two positions remained +6 for the whole decade. (It's probably good to bring the offense in for some edge cases, such as catchers, since emergency catchers are basically a non-factor in modern MLB. But I think the defensive method is usually better, especially when comparing positions with significant overlap.)

On the other hand, as you and others have noted, comparing Rickey's Rpos to Yaz and calling the position adjustment nonsense is silly. Games by defensive position for those two in particular:

Yaz:
LF 1912
1B 765
DH 412
CF 165
3B 33
RF 8

Rickey:
LF 2421
CF 446
DH 149
RF 27

That's almost 1200 extra games at 1B/DH for Yaz, and not quite 300 extra games in CF for Rickey. Which would explain a lot of difference in position adjustment.
   26. Eric J can SABER all he wants to Posted: February 13, 2021 at 06:02 PM (#6005011)
To expand on the position adjustment a bit more, this is actually one of the things WAR is REALLY good at. Take, say, Joe Mauer and Brian McCann. If you're looking at them casually, it would be easy to just classify both as "catchers who played from about 2005-2020" and move on to other aspects of the comparison. But while they were both primarily catchers, Mauer played almost 900 games at other positions (1B and DH), while McCann played less than 100. (In this specific case, moving out from behind the plate allowed Mauer to extend his career; combine that with being able to DH on his days off while he was still catching, and you get his extra 1000 plate appearances.)

Mauer is still the better player between the two of them, but by less than you might expect from a difference of 1000 PA, 14 points of OPS+ and four Gold Gloves. Because while both of them were "catchers," Rpos gives McCann a +96 and Mauer a flat 0 thanks to the differences in the amount of time they spent elsewhere. That's not a flaw in the system; it's reasonable quantification of a complicated question.
   27. Mefisto Posted: February 13, 2021 at 09:13 PM (#6005023)
I think the defensive method is usually better, especially when comparing positions with significant overlap.


Absolutely.

BTW, is anyone else having difficulty getting onto HOM pages? It took me hours to be able to post this.
   28. DL from MN Posted: February 15, 2021 at 10:23 AM (#6005165)
I think it's unfair to accuse Brock of bad faith. I think instead that he's let James' (bad) arguments influence him when he shouldn't.


I agree.
   29. Brent Posted: February 15, 2021 at 11:42 PM (#6005284)
I have a few comments on the NeLg MLEs.

1. Dr. C has Josh Gibson listed as 1B/C and the positional adjustment suggests he's been evaluated primarily as a 1B. According to Seamheads fielding data for 1935, Gibson caught 60% of the Crawfords' recorded defensive innings, so he should be regarded as a full-time catcher who also played other positions to keep his bat in the lineup. In addition to catching, he played 1B for 15%, RF for 5%, 3B for 1%, and even pitched a couple of innings. Altogether, he played at some position for 83% of the Crawfords' defensive innings. Clearly, though, catching was his main position.

2. Lazaro Salazar is evaluated as both a pitcher and as a position player, and as a pitcher he has an off-the charts pWAR of 7.4 and total WAR of 9.6. I assume this is entirely based on his Cuban Winter League performance, as he barely did any pitching at all for the New York Cubans in the NNL. In the CWL, his 1.55 ERA was second in the league, and as a batter he also led the league in average (.407) and triples (6 in 86 AB). It looks pretty impressive, but there are also some big problems. Just a little more than a year after the Cuban Revolution of 1933, the Cuban League did not invite any foreign players to participate; usually the league included a number of top NeLg stars. Also, some of the top Cuban players, like Dihigo and Oms, skipped the season. So the competition was quite watered down. We can see this from looking at the *best* pitcher in the CWL that year, who was 44-year old Dolf Luque, who had a 1.27 ERA and pitched a few more innings than Salazar (his teammate). While Luque was still pitching in the majors in 1934 (and a couple of games in '35), he was at that point just a league average relief pitcher. That should suggest the low quality of CWL play that season. Also, if Salazar had actually been considered a top pitcher, I assume the NY Cubans would have used him more than the 2 innings they had him pitch that summer. BTW, On the other MLE as a position player, Dr. C has him listed as a CF. In fact he played 50% of his team's innings at 1B, 31% in LF, 2% in RF, and only 3% in CF.

3. I'll also mention that Ray Brown leads the league in Seamheads WAR and may be underrated by Dr. C's MLEs if you like two-way players. He pitched 27% of the Grays' innings with an ERA+ of 162, and also played CF for 35% of their defensive innings and LF/RF for 8%. His OPS+ was 136. As a major leaguer, he probably wouldn't have been a two-way player, but in the context he was actually playing that was a very valuable skill.

4. I was disappointed that Paige didn't do better in the 1934 election, which was his best recorded season. I suspect that Paige was just as good in '35, but because he jumped the Crawfords to play semi-pro in North Dakota, we can't really know how good he was.

These comments aren't intended to disparage Dr. C's MLEs, which I suspect may have been done before the most recent iteration of the Seamheads data. I'm just suggesting that it's worthwhile to spend some time looking at the actual Seamheads data to make sure it's telling the same story as the MLEs
   30. DL from MN Posted: February 16, 2021 at 10:20 AM (#6005315)
Agree with all of this Brent
   31. MrC. Posted: February 19, 2021 at 10:25 AM (#6005796)
1935 prelim

1. Arky Vaughn 10.24 WARR
2. Wes Ferrell 9.43 WARR
3. Lou Gehrig 9.22 WARR
4. Lefty Grove 9.05 WARR
5. Cy Blanton 8.13 WARR
7. Jimmie Foxx 7.80 WARR
8. Martin Dihigo 7.48 WARR
9. Augie Galan 7.48 WARR
10. Luke Appling 7.43 WARR

Rest of top 20
11. Mel Ott
12. Charlie Gehringer
13. Billy Herman
14. Josh Gibson
15. Joe Vosmik
16. Leroy Matlock
17. Mel Harder
18. Pat Patterson
19. Syl Johnson
20. Willie Wells
   32. DL from MN Posted: February 21, 2021 at 09:00 PM (#6006147)
1935 World Series
Player Name G AB R H 2B 3B HR RBI BB SO BA OBP SLG OPS SB CS E WPA cWPA
Mickey Cochrane 6 24 3 7 1 0 0 1 4 1 .292 .393 .333 .726 0 0 1 -0.18 -7.21%
Pete Fox 6 26 1 10 3 1 0 4 0 1 .385 .385 .577 .962 0 0 1 -0.01 -0.50%
Ch Gehringer 6 24 4 9 3 0 0 4 2 1 .375 .423 .500 .923 1 0 0 0.20 7.10% 
Hank Greenberg 2 6 1 1 0 0 1 2 1 0 .167 .375 .667 1.042 0 0 3 0.01 -0.12%
Billy Rogell# 6 24 1 7 2 0 0 1 2 5 .292 .346 .375 .721 0 1 0 -0.25 -1.90%

Augie Galan# 6 25 2 4 1 0 0 2 2 2 .160 .222 .200 .422 0 0 1 -0.17 -5.09%
Stan Hack* 6 22 2 5 1 1 0 0 2 2 .227 .292 .364 .655 1 0 0 -0.25 -6.79%
Gabby Hartnett 6 24 1 7 0 0 1 2 0 3 .292 .292 .417 .708 0 0 0 -0.19 -6.28%
Billy Herman 6 24 3 8 2 1 1 6 0 2 .333 .333 .625 .958 0 0 1 0.25 9.61%

Pitcher Name G GS ERA W L SV CG IP H R ER BB SO WHIP WPA cWPA
Tommy Bridges 2 2 2.50 2 0 0 2 18.0 18 6 5 4 9 1.222 0.50 17.72%
Schoolboy Rowe 3 2 2.57 1 2 0 2 21.0 19 8 6 1 14 0.952 0.16 4.75%

Larry French 2 1 3.38 0 2 0 1 10.2 15 5 4 2 8 1.594 -0.33 -14.67%
Lon Warneke 3 2 0.54 2 0 0 1 16.2 9 1 1 4 5 0.780 0.65 14.64%
   33. Yardape Posted: February 22, 2021 at 12:31 AM (#6006169)
1935 prelim

1. Arky Vaughan - Pretty clear cut above. Great hitter, good defense.
2. Jimmie Foxx
3. Lou Gehrig
4. Hank Greenberg - Trio of slugging first basemen.
5. Satchel Paige - Difficult to really evaluate. It's clear that he was very good, and that makes sense in context with the rest of his career. But the competition makes it tough to judge just how good. I do think he's the best pitcher on the planet in 1935.
6. Mel Ott - A tier below the first basemen.
7. Gabby Hartnett - Best offense/defense combo aside from Vaughan.
8. Wes Ferrell - Best ML pitcher, although a lot of that rests on his bat. Still counts!
9. Joe DiMaggio - A terrific MVP season in the PCL. Obviously there are translation issues, and I could be overrating him somewhat, but I think his performance in 1936 suggests it's not totally out of the question.
10. Josh Gibson - Best NgL position player (again).

11-20: Leroy Matlock, Charlie Gehringer, Dizzy Dean (best NL pitcher), Lefty Grove, Lon Warneke, Willie Wells, Wally Berger, Buddy Myer, Joe Medwick, Mel Harder.
   34. DL from MN Posted: March 01, 2021 at 11:23 AM (#6007093)
5 elections remain, including this one. Then we will be complete 1901-2020
   35. Qufini Posted: March 02, 2021 at 09:54 PM (#6007335)
test
   36. DL from MN Posted: March 03, 2021 at 12:42 PM (#6007414)
1) Leroy Matlack: Best ML/NeL player and pitcher - I held back my ballot for a while due to this, but I stand by it.


I think #1 makes more sense than leaving him off the ballot this year
   37. John (You Can Call Me Grandma) Murphy Posted: March 03, 2021 at 01:42 PM (#6007437)
I think #1 makes more sense than leaving him off the ballot this year


Yeah, I totally understand anybody not agreeing with my placement of him, but his numbers are too good not to be ballot worthy, IMO.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Dynasty League Baseball

Support BBTF

donate

Thanks to
Rough Carrigan
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 0.5777 seconds
61 querie(s) executed