The Grove and Hubbell Era

February 17, 2019
 

When we left off in this series last Friday, we were reviewing the best pitchers in each league each year, and we had made it through the American League in 1925.  

 

 

1925 NL—Dolf Luque (D-WAR) or Eppa Rixey (R-WAR)

This is an interesting one.   The Cincinnati Reds in 1925 were just 80-73, but had three of the four best pitchers in the National League.  Luque was just 16-18 but led the National League in ERA (2.63); Eppa Rixey was 21-11 with a 2.88 ERA, and Pete Donahue was 21-14 with a 3.08 ERA, all for Cincinnati.  Dazzy Vance for Brooklyn was 22-9 with a 3.53 ERA in a perhaps tougher park.   Both systems agree that those are the top four pitchers in the league, but D-WAR ranks them Luque-Vance-Rixey-Donahue; R-WAR ranks them Rixey-Vance-Luque-Donahue. 

Let’s just talk about Rixey and Luque, who were teammates and are the leaders in the two systems.    R-WAR likes Rixey.  The R-WAR differences, however, are too small to be meaningful.  Baseball Reference has the four pitchers at 6.2, 6.1, 6.0 and 5.9.  The difference between first and fourth is only 3 runs, in effect, which is not a meaningful difference because our systems of measurement are not precise enough for a 3-run difference over the course of a season to be reliable. 

D-WAR, on the other hand, sees Luque as being quite significantly better than Rixey, leading him 9.0 to 6.7, in effect a 23-run difference.  Focusing on Luque and Rixey, both men were 34 years old in 1925, pitching for the same team.   Luque pitched more innings (291-287) with a better ERA (2.63-2.88), and Luque had twice as many strikeouts (140-69).   We could spend another half-hour here looking at small statistical discrepancies between the two, but I would have to say that my system (D-WAR) gives what would seem to be the better answer in this situation. 

Luque, as mentioned earlier, was a small Cuban right-hander, and would seem to be more comparable to Pedro Martinez than anyone else that you might have seen.   Rixey, on the other hand, is EXTREMELY comparable to Jim Kaat.  Like Kaat, he was a big left-hander with a decent/not outstanding fastball.  And like Kaat, he was unusually quick and graceful for a big man, and was regarded as the best fielding pitcher of his time.   Rixey had a career record of 266-251, but made it into the Hall of Fame; Kaat was 283-237 but has not yet made the Hall of Fame, primarily because Kaat retired at a time when there was a backlog of high quality starting pitchers waiting to get in. 

 

1926 AL—Lefty Grove (D-WAR) vs. George Uhle (R-WAR)

Both systems have Grove and Uhle as the 1-2 pitchers in the league.  R-WAR has it Uhle 8.7, Grove 7.6, while D-WAR has it Grove 8.5, Uhle 8.3.   I am missing a little data for both pitchers, and I don’t doubt that Uhle was better, meaning that Baseball Reference has it right.   Grove was 13-13 with a 2.51, while Uhle was 27-11 with a 2.83 ERA in 60 more innings.  

 

1926 NL—Charlie Root (D-WAR) vs. Hal Carlson (R-WAR)

We’re missing 9 starts for Hal Carlson (9 out of 34), so he isn’t near the leaders in my data.   Both Root and Carlson seem like kind of odd choices.   Root, a rookie in 1926, tied for the league lead in losses, 17 (he was 18-17), although he justified his selection by winning 26 games the next year.  Root was second in the league in strikeouts (behind Dazzy Vance) and second in ERA (behind Remy Kremer), 4th in innings pitched.

Carlson, who for all I know may well have been the best pitcher in the league in 1926, was 17-12 with a Philadelphia Philly team that was the worst in the league (58-93).   Get it. . .Carlson, Carlton, Phillies, worst team in the league, best pitcher in the league anyway.  He still seems like a bit of an odd choice; he was among the league leaders in appearance numbers (starts, complete games, innings, batters faced) but struck out only 55 batters all year, gave up 293 hits in 267 innings, and was not among the league leaders in anything like ERA or WHIP or strikeout/walk ratio.   Carlson died in mid-season a few years later.   My memory is that he had breathed mustard gas in World War I, and it eventually got him.  Nobody in the National League in 1926 is clearly better than the other guys.

 

1927 AL—Ted Lyons (D-WAR) vs. Tommy Thomas (R-WAR)

Ted Lyons and Tommy Thomas were the two top pitchers on a bad Chicago White Sox team (72-81).  Both D-WAR and R-WAR regard Lyons and Thomas as the top two pitchers in the league, although in different order.  

This phenomenon of two of the league’s best pitchers on a bad team, rating one-two in the league, happens in part because, when a team is short of pitchers, they may push those pitchers that they DO have to work harder, to pitch more innings, than they would normally pitch.   That may make pitchers like Lyons and Thomas, who are AMONG the best in the league, to show up as the BEST in the league, because they are pitching more innings than the comparable pitchers. 

As to Lyons vs. Thomas, that’s interesting.  They pitched 307.2 innings each.  Lyons allowed fewer earned runs, 97 to 102, so he has a little better ERA, but more total runs, 125 to 110.  That’s like the Jack Kralick/Camilo Pascual issue, so that partially explains why R-WAR prefers Thomas.   My data is missing 4 starts for Lyons and 6 starts for Thomas, so that partially explains it.

Lyons has a better won-lost record (22-14 vs. 19-17) and a better ERA, a better WHIP (1.164 to 1.186), and allowed less than half as many home runs (7 to 16).   D-WAR scores them at 8.9 for Lyons, 7.7 for Thomas; R-WAR makes it 8.5 for Thomas, 7.4 for Lyons.  Just based on what I know about them, Lyons would seem to me to have an edge.   

1927 NL—Jesse Haines (D-WAR) vs. Dazzy Vance (R-WAR)

My data is missing only 3 starts for Haines, 9 starts for Vance, so that probably explains the difference.  I would guess that Vance might come out ahead if we had full data for both players.

1928 AL—Lefty Grove (both systems)

1928 NL—Dazzy Vance (both systems)

1929 AL—George Earnshaw (D-WAR) vs. Willis Hudlin (R-WAR)

D-WAR sees Earnshaw as much better than Hudlin (6.8 WAR to 4.8), while R-WAR sees Hudlin as much better than Earnshaw (7.5 to 5.4).  It’s not a case where it’s really too close to call, like Lyons and Thomas, and it isn’t a matter of missing data; it’s something else.

Earnshaw was 24-8 with a 3.29 ERA, led the league fewest hits per nine innings but also in walks (125); Hudlin was 17-15 with a 3.34 ERA, didn’t strikeout or walk nearly as many.   Both Earnshaw and Hudlin pitched quite a bit of relief, which my system does not deal with. 

I think Hudlin was the more valuable pitcher, probably.   He pitched more innings in a tougher park for a pitcher, with basically the same ERA and same un-earned runs allowed.

 

1929 NL—Red Lucas (D-WAR) vs. Watty Clark (R-WAR)

Lucas was 19-12 with a 3.60 ERA; Clark was 16-19 with a 3.74 ERA.  Lucas and Clark are 1-3 in D-WAR, 1-2 in R-WAR.   Neither man was really at what you would normally think of as a Cy Young level.  Pat Malone was the only 20-game winner in the league (22-10), and led the league in strikeouts (166); he’s about the same level of value as the other two guys, actually also had a better ERA (3.57) than Lucas (3.60) or Clark (3.74).  I kind of think that Malone was probably the best pitcher in the league.

Red Lucas, a right-handed pitcher but a left-handed batter, was a very good hitter, and was used hundreds of times in his career as a pinch hitter.   He was a singles hitter, but a career .281 hitter.   He hit "just" .293 in 1929 after hitting over .300 in 1926, 1927 and 1928, and he would hit .336 in 1930. 

 

1930 AL—Lefty Grove (both systems)

1930 NL—Dazzy Vance (both systems)

1930 is like 1928; Grove and Vance are the best pitchers in their leagues by both methods.

 

1931 AL—Lefty Grove (both systems)

Grove was 31-4 with a 2.06 ERA, and won the American League MVP Award.   His war is 11.1 by one method, 10.5 by the other.   He was really good.

 

1931 NL—Bill Walker (D-WAR) vs. Ray Benge (R-WAR)

Walker was 16-9 with a 2.26 ERA, while Benge was 14-18 with a 3.17, so the presumption would be that Walker was better, and if you think Benge was better, you’d have to explain why.  Walker’s ERA+ was much better (163-134); Walker led the league.  Their innings pitched are comparable, and their strikeout/walk data is almost the same.   Their un-earned runs are about the same.  I don’t really understand why R-WAR thinks that Benge was better, since the difference between them in ERA was much larger than the park adjustment.   It must have something to do with a fielding adjustment, I guess.   I’d certainly take Walker.

 

1932 AL—Lefty Grove (both systems)

1932 NL—Carl Hubbell (D-WAR) vs. Lon Warneke (R-WAR)

I think it is pretty clear that Warneke was in fact the best pitcher in the league in 1932.  I have too much missing data to rely on my D-WAR study. 

 

1933 AL—Orel Hildebrand (D-WAR) vs. Lefty Grove (R-WAR)

In 1933 Connie Mack’s starting rotation, other than Grove, fell apart on him.  Rube Walberg, 20-12 in 1931 and 17-10 in 1932, dropped to 9-13 with a 4.88 ERA.  George Earnshaw, a 20-game winner in 1929, 1930 and 1931, and 19-13 in 1932, finished 5-10 with an ERA just short of six.  Popeye Mahaffey, 15-4 in 1931, was 13-10 but had an ERA over 5.00, and Tony Freitas, who appeared ready to step into a starring role in the rotation after going 12-5 in 1932, was hurt all year, won only two games and had an ERA over 7.00.

With the other starters all struggling, Grove was frequently asked to come out of the bullpen to clean up the mess.  He made 28 starts but also 17 relief appearances, and had a 0.81 ERA in 44 innings of relief.   He had 6 Wins and 6 Saves in relief; Grove finished 24-8 with a 3.20 ERA. Hildebrand pitched only five times in relief, but with an 8.44 ERA.  My system deals only with starts.  I’m pretty sure that if you included the relief work, Grove would be more valuable. 

To complete the digression, Tony Freitas went back to the minor leagues, and won over 300 games in the minor leagues.   Also Popeye Mahaffey pronounced his name Ma-HAY-fee, rather than Ma-HAFF-ee, like the 1960s pitcher with the same last name.   I’m always interested in that, when there are players who have the same last name but pronounce it differently.   I think there are fewer than ten instances of that in major league history, like Frank Saucier (SAW-see-err) and Kevin Saucier (SO-see-ay) and Jim Lefebvre (Le-FEE-ver) and Joe Lefebvre (Le-Fay).  There was also a Bill LeFebvre, back in the 1930s, but I don’t know how he pronounced it, and he capitalized the F. 

1933 NL—Hal Schumacher (D-WAR) vs. Carl Hubbell (R-WAR)

Schumacher and Hubbell were teammates on the Giants, who won the National League pennant.   Hubbell, much like Grove, also pitched 12 times in relief, 35 innings in relief, and had a 0.78 ERA as a reliever.  I am also missing 7 starts for Hubbell.   I’m sure Hubbell was the #1 pitcher if you include the missing starts and the relief work.

 

The Unluckiest Pitcher of All Time

One of the virtues of this approach is that it makes "lucky pitcher" or "unlucky pitcher" a hard concept, rather than a soft concept.   A soft concept is something you know in general but don’t know in specific terms.  Roger Craig in 1962-63 was a very unlucky pitcher, finishing 10-24 and 5-22 despite pitching not too bad.  Soft concept. 

A hard concept is something you know in very specific terms.  You can measure it; you can compare one to another.   This approach converts "lucky pitcher" or "unlucky pitcher" into a hard concept.  Of course, you can accept my method or argue with it, but. . .

By my method, the unluckiest starting pitcher of the last 100 years was Paul Derringer in 1933.   Derringer did not pitch great; his Deserved Won-Lost Record, by my math, was 12-13, a Deserved Winning Percentage of .491.   But he finished 7-27.   The chart below shows him as 7-26, but he also took a loss in relief (despite not allowing an earned run in the game.) 

In his second start of the season, April 26 against Cincinnati (for the Cardinals), the game was tied 1-1 going into the ninth.  Cincinnati had the bases loaded, two out, got a infield hit on a soft ground ball to second base and won the game, 2-1.  

About ten days later Derringer was traded to Cincinnati.  Cincinnati at the time was the worst team in the league.  They had finished last in 1931 and 1932, would finish last in ’33 and ’34.  Derringer lost games that year 3-1 (May 21), 2-0 (June 9), 1-0 (July 3), 3-2 in 10 innings (July 9), 1-0 (July 16), 3-2 (July 26), 2-1 (August 3), 3-2 (August 12), 2-0 (August 24), and 2-1 (September 14). 

Derringer lost 11 quality starts that year, assuming a Quality Start is 6 innings or more/3 runs or less.  He lost three other games in which he gave up more than 3 runs, but no earned runs at all.   I have Derringer, based on how he pitched in each game and who the opposition was and where the game was played, as deserving of 12.26 wins, 12.72 losses.  He actually was credited with 7 wins, so that is minus 5.26, and was charged with 26 losses (as a starting pitcher), so that’s minus 13.28.   Adding them together, that’s -18.54.   That’s the worst "luck" figure in my data.  These are the bottom ten, the ten toughest-luck seasons in my data:

Year

First

Last

Actual W

Actual L

D-Win

D-Loss

D-WPct

D-Luck

1933

Paul

Derringer

7

26

12.26

12.72

.491

-18.538

1987

Nolan

Ryan

8

16

18.17

8.14

.691

-18.038

1977

Jerry

Koosman

8

20

14.50

9.43

.606

-17.070

1985

Jose

DeLeon

2

18

8.38

9.57

.467

-14.805

2007

Matt

Cain

7

16

15.12

9.36

.618

-14.759

2015

Corey

Kluber

9

16

16.24

8.51

.656

-14.732

1922

Dolf

Luque

9

19

13.16

8.52

.607

-14.639

1962

Turk

Farrell

8

17

13.93

8.32

.626

-14.618

1963

Roger

Craig

5

21

10.94

12.54

.466

-14.407

1996

Roger

Clemens

10

13

18.93

7.88

.706

-14.052

 

DeLeon was 2-19 in 1985, not 2-18, because he also had one Loss in relief.   Dolf Luque was 13-23 in 1922, not 9-19; missing significant data there.   Turk Farrell in 1962 was 8-17 as a starting pitcher, also was 2-3 as a reliever.   Roger Craig in 1963 was 5-21 as a starting pitcher, also 0-1 as a reliever. 

 

 

 

1934 AL—Lefty Gomez (both systems)

1934 NL—Dizzy Dean (both systems)

1935 AL—Lefty Grove (both systems)

1935 NL—Hal Schumacher (D-WAR) vs. Cy Blanton (R-WAR)

Blanton was clearly better.  I only have data for 17 of Blanton’s 30 starts, but he almost finishes first in the league just based on those 17 starts.

1936 AL—Lefty Grove (both systems)

Grove in 1936 was "only" 17-12, 2.81 ERA, but had the highest R-WAR of his career (11.1).   I would find that conclusion to be somewhat suspect—that 1936 was the best season of his career.

  

Nice Outing There, Hod

On September 11, 1936, Rip Radcliffe had four hits and scored four runs.  Mike Kreevich on the same day went 5-for-6 with four runs scored and three RBI, while Zeke Bonura went 5-for-5 with three runs scored and four RBI.  Luke Appling on the same day went 5-for-6, scoring three runs, and Jackie Hayes on the same day had three hits including a three-run homer. 

Not too unusual, I suppose; sometimes several hitters will have big days on the same day, but Radcliffe, Kreevich, Bonura, Appling and Hayes were all teammates on the Chicago White Sox.   And even more unusual, the Athletics never changed pitchers.   Hod Lisenbee pitched a complete game, giving up 26 hits and 17 runs in eight innings. 

It’s the worst start in my data, by any standard.  His Game Score was -35. 

That was the 104th career start for Hod Lisenbee, and it would be his last for nine years.  In 1945, with World War II still raging, Lisenbee returned to the majors and made three more starts, giving up a total of 18 runs in 17.2 innings. 

 

 
 

COMMENTS (5 Comments, most recent shown first)

KaiserD2
A funny footnote to this data.

I have my own data base of all leading pitchers in each league since 1901 based on my own version of WAA (not WAR) as explained in my book. Obviously I was curious to check on how my leaderboard compared to the two Bill is using. In the vast majority of cases, we agreed.

I was very surprised, though, checking the NL for 1925, to find that I showed the best pitcher as Hal Carlson (barely--4.1 WAA to 4 for Rixey and 3.9 for Luque), and Bill didn't mention him at all. Looking a little more carefully at my data, I saw that the reason was that the Phillies show up (using DRA) as having dreadful fielding in that year, which means their pitchers, led by Carlson, were better than they looked. DRA shows the Phillies' fielding cost them -134 runs in 1925, while the baseball-reference measurement showed them at -65 runs.

I wasn't going to bother to comment on that, until I saw that, mirabile dictu, for the next year, 1926, Bill did show Carlson as a top NL pitcher. DRA shows Phillies fielding for that year to be even worse-- -160 runs, compared to -83 in baseball-reference.
8:02 AM Feb 23rd
 
CharlesSaeger
@bjames That’s the answer I was expecting, leading to my next one: how much more? I assume there’s some way of studying this, like matched pairs of pitchers with the same average Game Score, but one with a high standard deviation, the other with a low one.
6:12 PM Feb 22nd
 
bjames
CharlesSaeger
Is there a tendency for pitchers who are more consistent to have better won-lost records?


I believe that there is, yes, if studied on a season-by-season basis. If studied on a career basis the differences in consistency would disappear or become so small as to be nearly invisible. Within a season, I'm fairly sure that the pitcher who is more consistent is likely to have a better won-lost record, other things being equal.

4:04 AM Feb 22nd
 
CharlesSaeger
Is there a tendency for pitchers who are more consistent to have better won-lost records?
10:57 PM Feb 21st
 
dburba
Hod Lisenbee was the last major leaguer born in the 1800s. No real point, just thought that was interesting.
1:10 AM Feb 21st
 
 
©2019 Be Jolly, Inc. All Rights Reserved.|Web site design and development by Americaneagle.com|Terms & Conditions|Privacy Policy