We'll have the 2020 MMP election first but I had this done so I thought I might as well post the discussion.

There is a big difference in run scoring standard deviations between the AL and NL this season. LgAdj is .982 for the NL and .922 for the AL.

Dr C NGL MLEs

Player Name Age Lg Pos G PA Rbat Rbaser Rdp Rfield Rpos RAA WAA Rrep RAR WAR
Turkey Stearnes 32 NL CF 150 630 36 0 0 2 -2 37 4.1 20 57 6.4
Martin Dihigo 28 NL CF 136 570 25 1 0 6 -2 30 3.4 18 48 5.5
Oscr Charleston 36 NL 1B 143 600 23 2 0 1 -6 20 2.2 19 38 4.4
Sam Bankhead 22 NL SS 143 600 3 2 0 4 8 18 2.1 19 37 4.2
Dick Lundy 34 NL SS 131 550 4 0 0 5 8 17 1.9 17 34 3.9
Josh Gibson 21 NL 1B/C 83 350 21 0 0 2 -1 22 2.5 11 33 3.8
Buck Leonard 25 NL 1B 134 560 18 0 0 2 -6 15 1.7 17 33 3.8
Willie Wells 27 NL SS 148 620 -4 1 0 6 9 11 1.3 19 30 3.5
Jud Wilson 37 NL 3B 103 430 12 0 0 3 2 17 1.9 13 30 3.5
Newt Allen 32 NL 2B 141 590 -8 0 0 14 4 10 1.2 18 29 3.3
John Beckwith 33 NL 1B 134 560 22 0 0 -7 -6 10 1.1 17 27 3.1
Mule Suttles 32 NL 1B 138 580 12 0 0 2 -6 8 1.0 18 27 3.0
Alejandro Oms 37 NL CF 112 470 11 0 0 -1 -1 9 1.0 16 25 2.8
George Scales 32 NL 3B 129 540 5 0 0 -3 2 4 0.5 17 21 2.4
Biz Mackey 35 NL C 107 450 1 0 0 1 3 5 0.6 14 19 2.2
Cool Papa Bell 30 NL CF 150 630 0 2 0 -1 -2 -1 -0.1 20 19 2.2
Lazaro Salazar 20 NL CF 67 280 5 0 0 1 -1 5 0.6 10 15 1.7
Bullet Rogan 39 NL RF 88 370 6 0 0 -3 -4 0 0.0 12 11 1.3

Pitcher Name Age Lg G IP R RA9 lgRA9 RAA WAA pWAR Rrep RAR PA Rbat Rpos RAA bWAR WAR
Satchel Paige 26 AL 44 280 126 4.06 5.03 30 3.0 5.9 29 59 93 -8 9 1 0.1 6.0
Willie Foster 29 NL 36 240 96 3.61 3.97 10 1.1 3.5 20 29 80 -7 8 1 0.1 3.6
Martin Dihigo 28 NL 34 210 86 3.70 3.97 6 0.7 2.8 17 24 70 1 7 8 0.6 3.4
Lazaro Salazar 20 NL 24 150 55 3.32 3.97 11 1.3 2.7 12 23 50 1 5 5 0.4 3.1
Ramon Bragana 24 NL 36 240 106 3.99 3.97 0 0.0 2.4 20 19 80 -3 8 5 0.3 2.7
Bill Holland 32 NL 31 190 79 3.75 3.97 5 0.5 2.4 16 20 63 -7 6 -1 -0.2 2.2
Hilton Smith 26 NL 34 210 95 4.09 3.97 -3 -0.3 1.8 17 15 70 -1 7 6 0.4 2.2
Webstr McDonald 33 NL 44 280 129 4.14 3.97 -5 -0.6 2.2 23 18 93 -10 9 -1 -0.1 2.1
William Bell 35 NL 36 240 123 4.59 3.97 -17 -1.8 0.6 20 3 80 -2 8 5 0.4 1.0 
Ray Brown 24 NL 35 210 108 4.63 4.60 -1 -0.1 2.1 20 19 70 -1 7 6 0.4 2.5
Bill Byrd 25 NL 30 160 78 4.36 3.97 -7 -0.8 0.8 13 6 53 -2 5 3 0.2 1.1

A couple things about the MLEs:

Dihigo was playing in Venezuela this season and I can't find any stats so that is probably a smoothed projection.

Josh Gibson actually played 69 recorded games so projecting him for just 83 seems really low. I am going to add 50% to that to get him to 120 games.

1933 Prelim Ballot

1) Chuck Klein - NL Triple Crown winner gets the edge over his AL counterparts due to the league run scoring standard deviations
2) Carl Hubbell - Top NL pitcher is very close top the top NL position player. Hubbell is going to end up in the top spot once I figure in 20 scoreless postseason innings
3) Jimmie Foxx - best bat, less glove than Klein
4) Arky Vaughan
5) Joe Cronin - his superior fielding doesn't top Vaughan's better bat
6) Lon Warneke
7) Ed Brandt
8) Wally Berger
9) Mickey Cochrane - still the top C in baseball but competition is improving
10) Charlie Gehringer

11-15) Lou Gehrig, Josh Gibson, Turkey Stearnes, Hal Schumacher, Bump Hadley
16-20) Bill Rogell, Martin Dihigo, Lefty Grove, Pepper Martin, Mel Ott

Best pitching seasons in the history of the MMP project by my point score

1913 W Johnson 22.6
1920 Alexander 22.2
1985 Gooden 21.5
1995 Maddux 21.4
1972 Carlton 21.2
1994 Maddux 20.6
1918 W Johnson 20.4
1923 D Luque 20.2
1928 Vance 19.9
1997 Clemens 19.9
1924 Vance 19.6
1968 Gibson 19.5
2000 P Martinez 19.5
2009 Greinke 19.1
1902 Waddell 19.1
1912 W Johnson 19.1
1921 Faber 19.0
1944 Trout 19.0
1945 Newhouser 18.8
1963 Koufax 18.7
1933 Hubbell 18.7 (after postseason credit)
1946 Feller 18.6
1919 W Johnson 18.5
1972 Seaver 18.3
2004 Santana 18.2
1971 Seaver 18.0
1999 P Martinez 18.0
1971 Wood 17.9
2015 Greinke 17.9
1901 Young 17.8
1916 Alexander 17.8
1978 Guidry 17.7
1972 Perry 17.7
1995 R Johnson 17.7
1980 Carlton 17.5
1953 Roberts 17.5

Maybe I will try my hand at an MMP vote.

1) Jimmie Foxx - best overall bat in MLB
2) Carl Hubbell - best pitcher in MLB
3) Martín Dihigo - strong performance as two-way player is too impressive
4) Chuck Klein - second best bat, NL bump
5) Joe Cronin - slightest edge over Vaughan
6) Arky Vaughan - just behind Cronin
7) Lon Warneke - great run prevention and lots of innings, just behind Hubbell all around
8) Lou Gehrig - fantastic season, 2nd best bat in AL
9) Charlie Gehringer - fantastic defensive season just behind the offensive power of Gehrig
10) Mickey Cochrane - 150 wRC+ from your catcher in 130 games

11-20) Wally Berger, Babe Ruth, Turkey Stearnes, Dizzy Dean, Lefty Grove, Ed Brandt, Mel Harder, Josh Gibson, Mel Ott, Larry French

You are welcome to cast a ballot

I had occasion to take a look at BB-Ref's Defensive WAR a couple of days ago. I read the explanation of it in the popup window. It seems to say that, unlike every other Replacement Rate stat in existence, BB-Ref is using the league average as their "replacement rate" for defenders. Did I misread something? That can't be right. You don't call something Wins Above Replacement, but mean Wins Above Average. And you sure don't put them both into the computation of the player's total WAR. So, what IS the deal with dWAR? Does anyone know?

dWAR is always misinterpreted because it includes the positional adjustment. They don't double-count the positional average for the final WAR.

Use RField - that's defensive runs above or below positional average.

Hi. This is Brock Hanke’s MMP ballot for 1933. Again, I have no plans to change this, so you can just shunt it over to the Ballot Thread when that happens.

A short essay about Chuck Klein: Most of you know that I have little respect for Gavy Cravath’s homer numbers in the Baker Bowl, because he has a home/road homer split, for his career, of 3.85 to 1. The NEXT-highest such ratio I have found is the 1B on Cravath’s teams, Fred Luderus, who is an almost-exact contemporary of Cravath’s. His is 3-1 exactly. Well, Chuck Klein was very famous for hitting all his homers in the Baker Bowl. He was so famous for this that my dad, who was a Browns (AL) fan, told me about it when I was a kid just learning about baseball. Dad said that the Phillies traded Klein to the Cubs, but the Cubs traded him right back after the season, because even Wrigley wasn’t the Baker Bowl. This is true. It is also true that Chuck Klein’s career home/road homer splits are 1.7-1, which is less than half of Cravath’s. So, what happened?

What happened is Babe Ruth and the Clean Ball (not made any easier to hit, but just kept clean, which has the same effect). In this new environment, a good power hitter could hit homers in almost every park (except Griffith, which is in the AL). So, the Baker Bowl didn’t become any harder to hit homers in, but the rest of the parks became possible to hit homers in. Klein’s splits are extreme for his era, but not impossible. What was impossible was for anyone to duplicate Cravath and Luderus’ experience. So, I slipped Klein behind Wally Berger because of the ballpark, but didn’t push him any further down.

Jimmie Foxx is ranked first by both Win Shares and WAR, or I might have put Carl Hubbell above him, based on the postseason. Carl’s dead now; he won’t mind finishing second. This year, the WAR/Win Shares disconnect regarding pitchers shows up strongly. Lefty Grove ranks 5th in WAR, but 22nd in Win Shares, and Win Shares loves Lefty Grove. Bump Hadley is 9th and 30th. Meanwhile, Wally Berger is 2nd in Win Shares, but 10th in WAR, Lou Gehrig is 3rd and 11th, and Mel Ott is 8th and 16th. This sort of thing is why I use BOTH Win Shares and WAR when figuring these rankings.

For the 21-year-old Josh Gibson to be the best player in the Negro Leagues may be hard to imagine, but Gibson is in a class with Ted Williams and Stan Musial, who were top-end superstars by the time they were 21, too. My placement of Gibson at #8 is, really, just a guess.

So, enough of this; here are the rankings:

1. Jimmie Foxx
2. Carl Hubbell
3. Joe Cronin
4. Wally Berger
5. Chuck Klein
6. Arky Vaughn
7. Lou Gehrig
8. Josh Gibson
9. Lon Warneke
10. Charlie Gehringer

Brock, yes, the baseline on defense is league average. There is no such thing as "replacement level" defense. Replacement level is a concept that applies to overall value.

dWAR includes positional adjustment as DL says. dWAR a bad name because there's no coherent meaning of replacement level (the way we usually think of when referring to replacement) when applied only to defense. Replacement runs (in the WAR formula) are based on offensive playing time, and offensive replacement level under the assumption of average defense. I can try to explain this but it may come across clumsy.

You want to identify what is "replacement level." The way WAR does this is to set a level of offense that could be freely found from league minimum FAs or minor leaguers who can play average defense. You can see (I hope) that you have to pin either offense or defense to an absolute baseline (the most obvious candidate being league average) in order to define overall replacement level. A guy who was "replacement level" on both offense and defense would be useless - he'd be a defensive first baseman who hits like a backup shortstop. But that is not a replacement level player - you can do better than this hypothetical guy easily, just get the actual backup shortstop who hits like one, or the iron-glove first basemen who hits better than that but still below average. The iron-glove 1B who hits like the SS is not replacement level, he's a guy who washes out of baseball after high-A. Theoretically, you could set the defensive baseline at something below league average defense, but that just means you've increased the threshold for what counts as replacement level offense. If you set the defensive threshold at bad-but-not-awful corner outfield defense, the level of offense you can expect from the pool of players who can provide at least that level of defense is much higher than the level of offense you could expect from guys whose overall defensive contribution is MLB average.

I hope that makes sense. Setting defense to average was a practical consideration for determining the replacement level of offense. They are not independent concepts, they are directly (inversely) related to one another. If you want to change the replacement level for defense, you have to adjust replacement level for offense, so that the total value calculation comes out the same. More accurately, so that the level of offense that should be expected for the given defensive requirement lines up with how stringent that defensive requirement is.

By the way, the idea of "replacement level defense" is widely misunderstood. People talk about replacement level offense frequently, it makes intuitive sense to assume there's a replacement level defense as well. Which technically there is: it's been defined as average. So, not what people mean at all when they refer to replacement on defense. "Replacement level defense" is a misunderstanding of how WAR is defined mathematically.

kcgard2 - What I do not understand is 1) why dWAR is called that when it clearly is not, and 2) why is it impossible to fix a replacement rate on defense, but possible on offense. The arguments you use can be flipped the other way, and be just as reasonable. There is also the "Zero-Point" problem. Many players are hurt badly in WAR by having NEGATIVE dWAR numbers which offset the positive hitting numbers. It's like dealing with Linear Weights. Of course, players are hurt by having negative numbers at the plate offsetting positive glove numbers, but when the offensive numbers are being compared to the replacement rate, while the defensive numbers are being compared to the average, well, it's not the same order of magnitude of offsetting the strength with the weakness. To give an example that I, who am from St. Louis, know well: Lou Brock. Take a look at his BB-Ref page. Do you really think that his glove was so bad that it offsets a full third of his offense? Compared to the Replacement Rate? Really? That's what BB-Ref seems to think. But it's entirely false, although it has trashed Brock's reputation among analysts. But, really, it's just the Zero-Point problem.

2) why is it impossible to fix a replacement rate on defense, but possible on offense.

Probably because the defensive players not in MLB are probably as good or better than the guys in MLB. If you had no hitting requirement, i.e. two platoon baseball, you'd probably find a lot of fantastic gloves that can't hit at all. Entry into MLB is determined by bat, not glove. So, replacement level D is meaningless, unless you condition it on some minimum offensive requirement.

This may not directly apply to Brock's point, but I hope it clears up exactly how much of Lou Brock's WAR is because of his defence.

WAR - offensive WAR (O) + defensive WAR (D) + WAR from positional adjustments (P)

From Baseball reference's Lou Brock page:
O + D + P = 45 WAR (total WAR)

O+ P = 49 WAR (offensive WAR). on baseball references site the positional adjustment is counted on both the offensive and defensive WAR (where I believe Most of the confusion arises.
D + P = -16.8 WAR (defensive WAR)
Add: O + D + 2P = 49 -16.8 = 32 .2 WAR
Subtract: 0+ D + P = 45 WAR
Therefore P = -12.8 WAR. (the adjustment made for positional adjustment is -12.8

Substitute : O = 49 - (-12.8) = 61.8 WAR. (WAR from offence is 61.8 WAR)
and D = -16.8 -(-12.8) = -4 WAR. (WAR from defence is -4 WAR)

The defence does not reduce his WAR by 1/3, but by about 5- 6%. The rest of the reduction is because of the position that he played.

1) why dWAR is called that when it clearly is not, and 2) why is it impossible to fix a replacement rate on defense, but possible on offense.

1) Technically it is still dWAR. Average defense is replacement level on defense.
2) It's not impossible to fix a replacement level on defense. That level in the way WAR is formulated now, is average.

You could think of it this way - if you set offense at "replacement level", you can expect defense to be average among the pool of players freely available at that level of offense. If you set defense to "replacement level", you can expect offense to be average among the pool of players that are freely available. If you did this, there would be replacement level defense, and you would asking how on earth there can't be a similar replacement level on offense.

There is replacement level overall. If replacement level on offense is low, the pool of freely available players who can match that offense can be expected to play average defense. If replacement level offense is a .280 wOBA, the pool of players out there freely available who can match that offense can be expected overall to play defense at an MLB average level. Not worse. If you switch frames of reference, and make replacement level defense (a bad corner outfielder, say), the pool of freely available players who can match that defensive performance can be expected overall to provide offense at an MLB average level. Therefore, replacement level on offense would be average, like .315 wOBA instead of .280. A player who is really bad at both offense and defense is not replacement level, he is worse than replacement level.

I hope this is clearer.

1933 prelim ballot:

1. Josh Gibson - Easily the best hitter in the Negro Leagues, while playing a premium defensive position. While still unsure about his defensive value at catcher, that's enough for me this year.
2. Jimmie Foxx - Best ML hitter and player
3. Carl Hubbell - Best pitcher, NL MMP. Gets a World Series boost on top of a great season.
4. Chuck Klein - Baker Bowl deflates some of the raw stats, but still the best NL position player.
5. Mickey Cochrane - Pretty good season for a non-Gibson catcher
6. Lon Warneke - Makes up some ground on Hubbell with the bat. Not enough, of course...
7. Lou Gehrig - Passed by Foxx as the best 1B, but still an outstanding bat.
8. Sam Streeter - Best NgL pitcher
9. Turkey Stearnes - Great season in CF.
10. Arky Vaughan - Very strong season at a good defensive position.

Next tier: Babe Ruth, Oscar Charleston, Charlie Gehringer, Ed Brandt, Willie Foster, Wally Berger

Also, Mel Harder would be my AL MMPitcher.

Prelim:

1) Jimmie Foxx 1B
2) Wally Berger CF
3) Lou Gehrig 1B
4) Carl Hubbell SP
5) Josh Gibson C
6) Turkey Stearnes CF
7) Arky Vaughan SS
8) Joe Cronin SS
9) Lon Warneke SP
10) Pepper Martin 3B

test

1933 World Series

Player Name G AB R H 2B 3B HR RBI BB SO BA OBP SLG OPS SB CS E WPA
Carl Hubbell 2 7 0 2 0 0 0 0 0 0 .286 .286 .286 .571 0 0 1 -0.10
Mel Ott* 5 18 3 7 0 0 2 4 4 4 .389 .500 .722 1.222 0 1 0 0.56

Joe Cronin 5 22 1 7 0 0 0 2 0 2 .318 .318 .318 .636 0 0 1 -0.16
Joe Kuhel* 5 20 1 3 0 0 0 1 1 4 .150 .190 .150 .340 0 1 0 -0.19
Buddy Myer* 5 20 2 6 1 0 0 2 2 3 .300 .364 .350 .714 0 0 3 0.00

Pitcher Name G GS ERA W L SV CG IP H R ER BB SO WHIP WPA
Carl Hubbell* 2 2 0.00 2 0 0 2 20.0 13 3 0 6 15 0.950 1.20
Hal Schumacher 2 2 2.45 1 0 0 1 14.2 13 4 4 5 3 1.227 0.08

Earl Whitehill* 1 1 0.00 1 0 0 1 9.0 5 0 0 2 2 0.778 0.39