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

Friday, February 18, 2005

Turkey Stearnes

Turkey Stearnes

John (You Can Call Me Grandma) Murphy Posted: February 18, 2005 at 03:01 AM | 78 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: February 18, 2005 at 03:15 AM (#1151552)
hot topics
   2. Chris Cobb Posted: February 19, 2005 at 03:04 AM (#1153562)
Turkey Stearnes data

Mostly from Holway, with career data from MacMillan 8th & 10th editions.

Teams: 23-31 Det Stars, 32-35 Chi American Giants, 36 Phil Stars, 37 Det Stars, 38-41 KC Monarchs

Park Factors, as far as I know: Detroit played in a hitter’s park, though not an extreme one. The Chi Am Giants park was an extreme pitcher’s park. KC Monarchs’ park, if it is the same as their late 1920s park, probably played as a pitcher’s park.

Seasonal Data from Holway

1923 .360 for Det Stars; ba 5th, 17 hr (2nd), 33 hr/550 (5th), 16 2b (5th), 15 3b (1st); cf; all-star
6-13 vs. major-league competition
1924 .359 for Det Stars; ba 5th, 10 hr (1st) 23 hr/550 (2nd), 13 3b (1st); lf, all-star
1925 .364 for Det Stars; 18 hr (1st), 27 hr/550 (5th), 22 2b (3rd) 11 3b (3rd); cf; all-star
1926 .375 for Det Stars; ba 3rd, 20 hr (3rd), 36 hr/550 (2nd), 24 2b (2nd), 10 3b (3rd), 13 sb (4th); cf; all-star
1927 .339 for Det Stars; 20 hr (2nd), 35 hr/550 (2nd), 24 2b (1st), 12 3b (1st), 11 sb (3rd); cf, all-star
1928 .324 for Det Stars; 24 hr (1st), 42 hr/550 (1st), 18 2b (3rd), 7 3b (2nd); cf, all-star, MVP
1929 .378 for Det Stars; ba 3rd, 19 hr (3rd), 40 hr/550 (3rd); cf, all-star
1930 .339 for Det Stars; 10 3b (2nd); lf
.377 for NY Lincoln Giants; 5 3b (1st), 4 sb (1st); ut
1931 .370 for Det Stars, .095 for KC; 11 hr (1st), 28 hr/550 (2nd), 14 2b (1st), 6 sb (1st); cf, all-star
4-27 in World Series vs. Grays
1932 .297 for Chi Am Giants; 7 hr (1st), 20 hr/550 (1st), 12 2b (1st), 7 3b (1st), 14 sb (1st); cf, all-star, MVP
7-10 in playoff vs. Nashville
1933 .358 for Chi Am Giants; 11 hr (4th), 33 hr/550 (5th); cf
1934 .398 for Chi Am Giants; ba 3rd; 13 hr (2nd), 37 hr/550 (2nd), 13 2b (2nd), 5 3b (2nd); cf, MVP
6-16 in playoff vs. Phi Stars (also won best player award for Denver Post Tournament)
1935 .409 for Phi/Chi; ba 2nd, 8 hr (5th); cf, all-star
7-15 vs. major-league competition
1936 .308 for Phi Stars; 17 hr (1st), 47 hr/550 (5th); cf
1937 .382 for Det Stars; 3 hr (4th), 24 hr/550 (1st); cf
6-23 in playoff vs. KC
7-17 in World Series vs. Homestead
1938 .292 for KC, .280 for Chi Am Giants; 3 hr (2nd), 22 hr/550 (3rd), 6 2b (2nd), 2 3b (3rd); rf
1939 .453 for KC, 2 hr (2nd), 9 hr/550 (4th), 8 2b (2nd), 3 sb (1st); rf, all-star, MVP
1940 .287 for KC, 5 hr (2nd), 25 hr/550 (2nd); cf, all-star
1941 no data

Career Batting

1308-3937, .332
197 hr, 27/550 ab
37-98, .378 vs. major-league competition
.355 mean avg. for 18 documented seasons

There’s something odd in Turkey Stearnes’ career record. With only 5 seasons listed as below his career avg., it’s hard to believe that unevenness of games recorded could produce a career avg. that low. Either some of his seasonal averages are his “published” averages rather than statistically documented ones, or Holway’s career data is in error, or something else is strange.

Career data from MacMillan 8th edition
905 g, 3372 ab, 1186 hits, 199 2b, 106 3b, 185 hr, .352 ba, .638 slg
(note: ba printed as .359 . . . The NeL players really need some devoted historians who are deeply deeply deeply detail-oriented . . . )

Career data from MacMillan 10th edition
903 g 3358 ab, 1183 hits, 201 2b, 107 3b, 181 hr, .352 ba, .638 slg

Black/Gray Ink
51/178

Black Ink total is #4 among Negro-League players, trailing Gibson, Suttles, and Charleston.
Gray Ink total is #1 among Negro-League players.

Career Fielding

lf 24, 30
cf 23, 25-29, 31-37, 40
rf 38-39

Brief comment: I haven’t done MLEs yet, and I look forward to Gary A’s data, especially for walk totals that will give us some sense of Stearnes’ plate discipline. However, this one’s easy, folks. Stearnes was a great player. Unquestionably better than Goose Goslin, whom we’re poised to elect with a majority of the #1 votes in 1945. Among 1945 eligibles, Suttles was a better pure power hitter, but Stearnes was the more complete player.
   3. Dag Nabbit is part of the zombie horde Posted: February 19, 2005 at 03:32 AM (#1153587)
One quickie comment I can make about Stearnes - I did some Negro League reading about a year ago, and from the stuff that came out before the last round of Negro Leaguer HoF inductions (that put in Stearnes and others), Turkey was the most commonly mentioned "How could they not put this guy in?" player by experts and former Negro Leaguers. His reputation as a player was in line with the info Cobb gave up above.
   4. David C. Jones Posted: February 19, 2005 at 05:56 AM (#1153696)
Stearnes is a no-brainer. He'll move to the top of my ballot, probably.
   5. Dag Nabbit is part of the zombie horde Posted: February 19, 2005 at 06:53 AM (#1153784)
Stearnes is a no-brainer. He'll move to the top of my ballot, probably.

Careful -- people have gotten in fistfights over that phrase around here before.
   6. Gadfly Posted: February 19, 2005 at 04:30 PM (#1154230)
Chris Cobb-

Mack Park, the Detroit Stars home park from 1923 until it burned down in 1929, had an extremely short right field porch and probably extremely inflates Stearnes' home run total (Stearnes was an extreme pull hitter) as well as Edgar Wesley.

In 1930, the Detroit Stars moved in to Hamtramck Stadium which had a 407 foot right field line and Stearnes power became all triples.

In a normal stadium, Mule Suttles would have out-homered Stearnes by a considerable margin. But, of course, Stearnes was better than Suttles in just about every other aspect except drawing walks.

I already stated this in the Beckwith thread, but basically Stearnes is a dead ringer for Stan Musial, right down to the funky lefty stance. I think Stearnes was actually a little better than Stan, less average, more power, better defense.
Suttles is a dead ringer for Hank Greenberg.
   7. Howie Menckel Posted: February 19, 2005 at 05:24 PM (#1154271)
from baseballlibrary.com

Cool Papa Bell once said, "If they don't put Turkey Stearnes in the Hall of Fame, they shouldn't put anybody in."

Satchel Paige called Stearnes "one of the greatest hitters we ever had. He was as good as anybody ever played ball."

Stearnes was the complete player, a lefthanded batter who, said Paige, "hit with his right foot in the bucket and twisted his right heel and pointed his big toe up."


Not endorsing the statements as much as I am simply putting them into the record...
   8. Chris Cobb Posted: February 19, 2005 at 07:49 PM (#1154445)
In a normal stadium, Mule Suttles would have out-homered Stearnes by a considerable margin. But, of course, Stearnes was better than Suttles in just about every other aspect except drawing walks.

Their time together for the Chi Am Giants, 33-35 provides pretty clear evidence of this (not that Schorling was a "normal" park, but it was affected right-handed and left-handed hitters equally, as far as I know).

According to MacMillan 10th (easier to use for a quick check than the larger-sample data in Holway), Suttles had 21 hr in 325 ab during these three seasons, or 35.5 hr/550 ab.

Stearnes had 19 hr in 372 ab, or 28.1 hr/550 at bats. The sample size is fairly small, but the difference is about what we would expect.

An eyeball inspection of Holway suggests that the difference would be similar, but with a somewhat higher rate for both players with more recorded at bats for each.

Both Stearnes and Suttles were 32-34 during these seasons, so it's a pretty good snapshot comparison. I'd also guess, without having studied the numbers yet, that with the park effect of Schorling suppressing their batting (and assuming the park the Giants were in at that time was still Schorling park), esp. the power numbers, these hr rates are probably pretty close to what they would have done in a neutral park in the majors, maybe a little higher, maybe a little lower.
   9. sunnyday2 Posted: February 19, 2005 at 08:12 PM (#1154489)
Right now, depending on whether Al Simmons is elig. or not, I have Stearnes 3rd and Suttles 14th (without Simmons who IMO is an obvious #1) but I think both are too low. Of course if they move up Stearnes doesn't really have very far to go while Suttles does. It seems that 11 ballot spots is too big of a difference, but we've all been there before, with comparable players farther apart than seems intuitive.

Averill is another toughie. Better than Goslin at his peak but short career. The CF glut just gets worse and worse all the time.

Right now if I had to:

1. Stearnes
2. Roush
3. Averill
4. Browning
5. H. Wilson
6. Duffy
7. Van Haltren
8. Berger
9. Poles
10. Berger

But I am not sure, especially re. Averill whom I have not fully considered. Roush keeps looking better and better BTW. Peak, career, offense, defense...he wasn't the master on any one dimension but he was not really weak anywhere either.
   10. Gary A Posted: February 19, 2005 at 08:21 PM (#1154503)
1928 Turkey Stearnes
NNL Detroit Stars

Batting
G-80 (team 80)
AB-314
H-101
D-17
T-7
HR-23 (1st in league)
R-82
W-31 (tied for 7th; leader had 38)
HP-5
SH-5
SB-5
TB-201 (3rd)
AVE-.322 (NNL .278)
OBA-.391 (NNL .333)
SLG-.640 (NNL .384; 3rd in league)
   11. TomH Posted: February 19, 2005 at 08:28 PM (#1154518)
basically Stearnes is a dead ringer for Stan Musial
---
Well, I gues that's POSSIBLE, but I would argue that
1. The Man is one of the 12 best MLB players ever.
2. Stearnes is generally not considered as one of the best four NeL players ever.

Equating the at-the-very-best-#5 NeL player pre-1947 with the #10 or #12 MLB player of all time is way too much of a stretch for me.
   12. Gary A Posted: February 19, 2005 at 11:53 PM (#1154767)
1928 Turkey Stearnes

Fielding (cf)
G-78 (team 80)
DI-672.3
PO-191
A-7
E-8
DP-0
RF-2.65 (NNL cf 2.48)
FPCT-.961 (NNL cf .963)

Detroit Stars' center fielders (almost entirely Stearnes) accounted for 46.8% of the team's outfield putouts; the NNL average was 42.7%.

Stearnes also played two games in left field.
   13. Gary A Posted: February 20, 2005 at 12:07 AM (#1154781)
To recap from the Beckwith thread, here are PFs for Detroit's Mack Park (raw park factor, by which I mean simply home runs/game divided by away runs/game, with no further adjustments):

1920: 121
1921: 107
1922: 122
1923: 79
1928: 93

For 1921, I have more extensive park information (haven't compiled it yet for '28, though I can):

Both teams at Mack Park in 1921 hit: .262/.323/.397, with 21 triples and 42 home runs in 2169 PA.

Both teams in Stars' road games hit: .269/.323/.358, with 37 triples and 12 home runs in 2322 PA.

AVE: .975164
OBA: 1.001216
SLG: 1.107255
HR: 3.746888
   14. Mark Shirk (jsch) Posted: February 20, 2005 at 12:09 AM (#1154783)
I always thought of Searnes as a LFer. So which is it?

This will make a difference because a CFer that can hit like him is all-world, a left fielder that hits like him is a definite HOMer. It could mean the difference between #1's 1,2,or 3 next year.
   15. Chris Cobb Posted: February 20, 2005 at 01:00 AM (#1154855)
I always thought of Searnes as a LFer. So which is it?

According to Holway's seasonal information, Stearnes played centerfield in 14 of 18 seasons.

Riley's _Biographical Encyclopedia_ lists Stearnes' primary position as centerfield.

James ranks him as a left-fielder, but I suspect that's a ploy to spread out the great players among more positions. He lists Mule Suttles as a left-fielder, too.

Unless the resident experts have information to the contrary, the evidence that Stearnes was primarily a centerfielder is clear. Riley doesn't rhapsodize over his defense the way he does over Charleston and Torriente, but he says Stearnes was an excellent defensive outfielder.
   16. Gary A Posted: February 20, 2005 at 01:36 AM (#1154918)
Stearnes was primarily a center fielder.

Btw, he batted cleanup for Detroit in the 20s, but he was the leadoff hitter for the 1934 American Giants (I think also for the Monarchs when he played with them in the Denver Post tournament that year).
   17. Mark Shirk (jsch) Posted: February 20, 2005 at 02:43 AM (#1155041)
Okay, thanks guys. This changes things a lot now. Unless WS estimates and other evidence is overwhelming, Stearnes will be over Suttles on my ballot. Had Stearnes been a LFer, I may have chosen Suttles bat.
   18. Chris Cobb Posted: March 01, 2005 at 11:33 PM (#1173717)
In preparation for doing MLEs for Turkey Stearnes, I need a park factor for Detroit's Mack Park.

Here are the raw park factors we have from Gary A.:

Mack Park (Detroit):
1920: 121
1921: 107
1922: 122
1923: 79
1928: 93

On the basis of these factors, my plan is to use a park factor of 102 for Stearnes' seasons in Detroit, which fits (if I am correctly informed) with its reputation as a hitters' park, but not an extreme hitters' park.

I won't be running Stearnes' numbers until tomorrow night, so any comments on this park factor between now and then woudl be a great help.
   19. jimd Posted: March 02, 2005 at 01:01 AM (#1173811)
Gary A's Data     After Wendt  After Bal-Sched (8 teams)
1920: 121             118         108
1921: 107             106         103
1922: 122             119         108
1923:  79              81          92
1928:  93              94          97
5 Yr average 104      104         102

No consistency. However, Shibe Park (Connie Mack Stadium) was another park whose factor jumped all over the place.
   20. Chris Cobb Posted: March 03, 2005 at 02:39 AM (#1176254)
Turkey Stearnes MLEs


Year MLE BA MLE SA REG BA REG SA
1923 0.325  0.583  0.325  0.569
1924 0.323  0.513  0.326  0.554
1925 0.331  0.528  0.329  0.536
1926 0.341  0.610  0.335  0.593
1927 0.312  0.574  0.314  0.571
1928 0.290  0.545  0.297  0.550
1929 0.334  0.564  0.325  0.555
1930 0.297  0.503  0.305  0.514
1931 0.306  0.485  0.318  0.507
1932 0.313  0.532  0.330  0.545
1933 0.371  0.611  0.359  0.597
1934 0.408  0.688  0.384  0.639
1935 0.415  0.605  0.388  0.595
1936 0.277  0.468  0.308  0.511
1937 0.351  0.505  0.330  0.496
1938 0.275  0.478  0.304  0.457
1939 0.315  0.383  0.307  0.409
1940 0.266  0.414  0.278  0.416
car. 0.325  0.545  0.326  0.547



Stearnes’ career rates change a good deal, like Suttles, when projected into a major-league career, because the records of his early seasons are so much fuller.

Here are his projected career totals and career rates. Given the interest in on-base percentage on the Suttles’ thread, I’ve done bb and obp projections for Stearnes, but these are based on a single firm data point, so it’s quite speculative.

2590 g, 9785 ab, 3190 hits, 5253 total bases, .326 ba, .537 sa, 1087 bb, .393 obp

By these projections, Stearnes does not quite have Simmons’ peak, but for his career, he was a better hitter. His career shape as a hitter was more like Paul Waner’s, I think.

Stearnes’ 1933-1935 seasons look a bit out of keeping with the rest of his career, and they pull his career totals up a bit. Were the American Giants playing in Schorling Park in those seasons? Suttles and the rest of the team were hitting like they probably were, but Stearnes is putting up some of the best totals of his career. If it’s still Schorling and it’s playing like its old self, perhaps these are a fluke of short seasons that the regressions can’t quite smooth out?

He’s a HoMer regardless, but this is one point at which the projections look questionable, so I feel I should note the oddness and see if there’s a reason for it.
   21. Mark Shirk (jsch) Posted: March 03, 2005 at 02:50 AM (#1176273)
i ewas wondering how Stearnes BA and SLG cold be higher in the MLE's for those years. Does something like this happen with Beckwith? His career rates look very close to his MLE'.
   22. karlmagnus Posted: March 03, 2005 at 03:02 AM (#1176307)
Stearnes is really very good, isn't he -- streets ahead of Suttles, who I don't think is quite as good as Beckwith. Not as good as Charleston, but not that far off Pop Lloyd, right towards the top of the NL cadre. Will now post my ballot, which relies heavily on Chris' excellent work, but of course includes a lot of my own interpretation.
   23. KJOK Posted: March 03, 2005 at 04:10 AM (#1176456)
Stearnes’ 1933-1935 seasons look a bit out of keeping with the rest of his career, and they pull his career totals up a bit. Were the American Giants playing in Schorling Park in those seasons?

Chicago played in Schorling Park until the 1941 season, when they moved into Comiskey, another pitcher's park...
   24. KJOK Posted: March 03, 2005 at 04:15 AM (#1176459)
Cleveland and Detroit both changed parks around 1932-33. Not sure what type of parks they were, but if they were more hitter friendly than the previous parks, they could have an impact on Stearnes' stats being better.
   25. Chris Cobb Posted: March 03, 2005 at 04:21 AM (#1176465)
i ewas wondering how Stearnes BA and SLG cold be higher in the MLE's for those years. Does something like this happen with Beckwith? His career rates look very close to his MLEs

Yes, it also happens with Beckwith's 1922 and 1923 seasons, and for Suttles' 1933-35 seasons. The Chi Am Giants' park factor is larger than the conversion factor. The effect is actually more pronounced for Beckwith, because he was playing in Schorling at a time when the NeL offense levels were 5% below the majors, whereas in the early 1930s, the offense levels in the NeL equal or exceed slightly the offensive levels in the majors (at least for batting average -- the case on slugging is unclear due to lack of comprehensive evidence.

The years are less likely to be _noticed_ for Beckwith, however, because they fit smoothly into his overall career shape. Stearnes' own offense (perhaps like Torriente before him) doesn't decline markedly in Schorling the way Beckwith's and Suttle's (and most players') do. His combination of power and speed perhaps enabled him to take advantage of the park in ways that few players could?
   26. Gary A Posted: March 03, 2005 at 07:59 AM (#1176749)
I compiled 1934 stats a couple of years ago, but lost most of it when my hard drive crashed. I can, however, report these NNL totals (but must caution that these haven't been audited or balanced):

G: 201
AB: 11638
R: 1645
H: 3039
D: 321
T: 109
HR:104
W: 591
HP: 57
SB: 209
AVE: .261
OBA: .300
SLG: .334
R/G: 4.09

Walks, hit batsmen, and extra base hits are missing for many games played in Philadelphia and Newark (hit batsmen are missing for many other games, too). Unfortunately, I don't know exactly how large an effect the missing data has--clearly, the OBP and SLG are somewhat lower than they should be. There appear to be quite a few unaccounted-for plate appearances. I don't have sacrifice totals, so that's a big part of it.

The most reliable information here is that run-scoring was actually significantly lower than in the majors at the time (4.68 r/g in the NL, 5.13 in the AL).

I do have the schedule and scores, so I can work up basic park factors for 1934 in a few days.
   27. Gary A Posted: March 03, 2005 at 08:10 AM (#1176757)
By the way, the effect of the missing data wouldn't be enormous--I'd say that the correct final averages might be, say, 261/315/350, or something like that.

The 1919 AL averaged 4.09 runs/game and .268/.329/.359. I don't have fielding totals for the 1934 NNL, but there were certainly many more errors than in the majors.
   28. Chris Cobb Posted: March 03, 2005 at 02:49 PM (#1176953)
Walks, hit batsmen, and extra base hits are missing for many games played in Philadelphia and Newark (hit batsmen are missing for many other games, too).

Do you have any idea if failure to report extra-base hits (possibly excluding home runs) is a long-term problem with records for Philadelphia and Newark? I ask because I notice that Stearnes, who kept good speed throughout his career, has an anomalously low doubles total for his season in Philadelphia in 1936, and Suttles has very low doubles totals for a number of his early Newark seasons. In Suttles' case, it seemed possible that he simply wasn't hitting many extra-base hits aside from home runs (which are clearly being reported in _some_ fashion), but I wonder if it's just that the data is missing for his home games?

If that's the case, should his totals be adjusted? It wouldn't make a huge difference, but if he lost, say, half his doubles over three or four seasons, it would have a noticeable impact on his slugging percentages.

If this is indeed a major gap in the statistical record, it would affect Willie Wells' later career also, so it would be well to identify it.
   29. Chris Cobb Posted: March 03, 2005 at 03:10 PM (#1176982)
A two more questions about this data:

1) When you say you don't have sacrifice totals, does that mean that plate appearances in which a sacrifice may have taken place have not been recorded, or that sacrifices may have been counted as outs, or both?

2) What does the "games" total mean? If Newark played Philadelphia, is that counted as one game, or two? (My guess is that you are counting it as one game, because otherwise the AB totals seem impossibly high, but if it does mean "two teams played a game against each other" then a lot of plate appearances are missing from the data, aren't they?)
   30. Gary A Posted: March 03, 2005 at 03:18 PM (#1176991)
I don't have much material for the later 30s or 40s, so I can't say for sure. But the Newark box scores I have for 1934 didn't have bottom sections, so no information on extra base hits, walks, strikeouts, etc., unless the accompanying story mentioned something (of course, you can track pitchers' k's through catchers' putouts). This goes, I believe, for nearly all Newark home games, which weren't well-reported elsewhere.

Philadelphia's a different case, because a good number of those games were reported in the Pittsburgh Courier, Chicago Defender, or Baltimore Afro-American. The Phi Tribune, though, published horrible box scores in 1934, with just "r h e" for batters, and usually no bottom section (the paper did have a couple of play-by-play accounts, iirc). I don't know offhand how many games we're talking about, though I can at least give a decent idea when I do the park factors (I also have recorded the number of box scores for each game and which papers they come from). Eventually I will reconstruct the season (I still have all the box scores photocopied), but that's way down the road.
   31. andrew siegel Posted: March 03, 2005 at 05:07 PM (#1177215)
Those MLE's establish him as essentially Mel Ott's equal: careers of equal length, same era, similar power, 20-30 less OBP point, but above-average CF versus good RF. That puts Turkey's All-Time major league ranking somewhere btween 20 and 50. Right where Bill James had him (#27).
   32. Gary A Posted: March 04, 2005 at 04:47 PM (#1180064)
1) Sac hits -- I don't have sac hit totals because the data I had printed out before the hard drive crashed didn't include them, for whatever reason. Most Newark and many Phi games lack at bats AND sac hits in the box scores, so there's no way to know for sure whether you've estimated at bats correctly (though you can be fairly close), nor do you have any way of knowing how many sac hits occurred (though you can estimate them, given totals for the team in other games). Other than those games, I believe I've got correct at bats, but I'm just not sure how many sac hits there were. IOW, no, I don't think there are a lot of at bats that were really sac hits.

2) Games are the total number of games, not team games. You need to double them (402). As I said, what's missing from plate appearances are: sac hits for the whole league plus about one-fourth of walks plus an indeterminate number of hit batsmen.
   33. Gadfly Posted: March 04, 2005 at 11:29 PM (#1181025)
Chris Cobb-

1) Turkey Stearnes' BA jump in your MLEs from 1932-1935 at Schorling Park can easily be explained. In 1932, Stearnes joined the Chicago American Giants and his new manager was Dave Malarcher. Malarcher talks at some length about Stearnes in Holway's "Voices" book.

Malarcher, an extremely intelligent man, was aware that Stearnes' style of play, i.e. left-handed, extreme-pull, home-run hitter was not exactly suitable for the park. So Malarcher set about changing Stearnes style of play.

Aware that Stearnes was fast as hell, Malarcher told Stearnes to bunt to utilize his speed and, evidently, to also practice place hitting. Basically, Malarcher taught Stearnes the Rube Foster system of doing things.

In other words, Malarcher changed Stearnes from a home run hitter to a high average hitter. And, of course, the evidence is right there in the statistics, where Stearnes' BA rises every year he played for Malarcher:

1932: 297
1933: 358
1934: 398
1935: 409

The difference between your MLEs for Stearnes for these years and the rest of his career is simply one of adaptability. For most of his career, Stearnes was a power hitter first; but, for that time peroid, he was a high average hitter.

If Stearnes had actually played in the Majors, his talent would have probably landed somewhere in the middle of these two extremes; i.e. he would have hit for a higher average than his MLE indicates with a comparable reduction in isolated SA.

Although he gets no credit for it, Stearnes came fully loaded; he had the talent to thrive under any baseball condition.

Of course, as always, it goes without saying that I think the conversion factors are too low and Stearnes is greater than the MLE actually indicates; but, even at these reduced numbers, the quality and diversity of Stearnes' talent comes through.

2) The difference in Stearnes and Suttles home run rates in Schorling is interesting, but I think it still understates the true difference in their power.

Suttles had power to the alleys and Stearnes was, as someone said above in another context, exactly like Mel Ott (i.e. an extreme pull-hitter). On top of this, Stearnes was almost surely hit some inside-the-park homers (especially in Schorling)and Suttles almost surely did not.

In other words, the park crippled Suttles talents while only forcing Stearnes to simply rely on different talents.

Interestingly, statistics from a park that rewarded pure power to a numbing extreme and was graced by both Suttles and Stearnes is available.

Both men played in the Los Angeles Wrigley Field bandbox in the California Winter League. The true difference in their power really comes out there (despite the fact that the park favored lefties - Stearnes - over righties - Suttles):

G-AB-H-HR-BA

Stearnes
207 754 281 56 .373
154 561 209 42 .373 (pro-rated)

Suttles
126 450 170 64 .378
154 550 208 78 .378 (pro-rated)

According to these stats, Suttles had almost twice as much power as Stearnes, though I think that is an overstatement. In all reality, Suttles true power advantage is probably somewhere in the middle.

Interestingly, Suttles is not recorded as hitting a single 3B in LA while Stearnes hit 16. The difference in the players' speed was possibly even greater than the difference in their power.

Tom H-

Of players who began their career before 1950, Stan Musial is obviously not the best white outfielder in either peak or length of career. Babe Ruth, Ty Cobb, Ted Williams, and arguably Tris Speaker all have higher peaks and longer (i.e. more valuable) careers.

Two other guys, Willie Mays and Hank Aaron were also active during Musial's career and were better than him in both ways.

Of all the black players that started their careers before 1950, only one undoubtedly had a better peak and longer, more valuable, career than Stearnes: Oscar Charleston.

In fact, Stearnes has a career profile just like Musial. He was not the greatest outfielder of his time in his leagues; but he was damn close for an incredibly long time. His career has great weight even though there were other players who shined more brightly, but more briefly (i.e. Joe DiMaggio, Ralph Kiner, Monte Irvin, Rap Dixon, Torriente, etc.).

However, Musial was not the defensive player that Stearnes was or, I suspect, as fast. Both men were helped to a greater than ordinary degree by their home parks. I think equating Stearnes to Musial is probably a compliment to Musial.

Of course, I could be wrong but, if I am, it's not by much.
   34. Gary A Posted: March 05, 2005 at 04:59 AM (#1181424)
I may have said this earlier, but Stearnes batted cleanup throughout the 20s for Detroit, but in 1934 he led off for Chicago (with Wells usually hitting third and Suttles fourth).
   35. Chris Cobb Posted: August 27, 2008 at 03:09 AM (#2917946)
Turkey Stearnes Offense MLEs
These were prepared using exactly the same system as the recently posted MLEs for Oscar Charleston, except that for Stearnes all the NeL data used came from _Shades of Glory_.
Playing time projections were based on an examination of Stearnes games vs. team decisions, but I did not use a formula for making the projections.

Age Year  Team  G    PA    Hits  TB    BB  SB   BA     OBP    SA    OPS+
22  1923  Det  142   568   173   319   39   3  0.327  0.372  0.602  153
23  1924  Det  119   476   143   237   38   2  0.325  0.379  0.540  143
24  1925  Det  151   634   190   329   68  13  0.335  0.406  0.581  149
25  1926  Det  154   647   203   358   64  21  0.348  0.413  0.614  173
26  1927  Det  146   613   178   330   72  15  0.328  0.407  0.610  169
27  1928  Det  135   567   155   294   49   7  0.299  0.360  0.569  139
28  1929  Det  148   622   186   324   76  20  0.341  0.421  0.593  148
29  1930  Det
151   634   183   340   65  29  0.322  0.392  0.597  135
30  1931  Det
138   580   158   280   57  16  0.302  0.371  0.537  141
31  1932  Chi  140   588   166   274   62  25  0.316  0.388  0.522  142
32  1933  Chi  152   638   186   318   51   0  0.317  0.372  0.542  158
33  1934  Chi  154   647   200   343   52  24  0.337  0.391  0.578  156
34  1935  Chi  147   617   202   324   72  20  0.370  0.444  0.593  176
35  1936  Phi  134   563   156   250   47   3  0.302  0.360  0.485  126
36  1937  Det
111   466   121   187   45   4  0.287  0.356  0.445  116
37  1938  KC
*   82   328    80   134   24   9  0.264  0.318  0.440  106
38  1939  KC   115   437   107   165   39  13  0.268  0.334  0.413  100
39  1940  KC   110   418    92   152   47   6  0.248  0.332  0.410  104
career
....... 2429 10043  2878  4958  968 230  0.317  0.383  0.546  144 


*1930 Stearnes also played for the NY Lincoln Giants
*1931 Stearnes also played for the Kansas City Monarchs
*1937 Stearnes also played for the Chicago American Giants
*1938 Stearnes also played for the Chicago American Giants

Career hit distribution
1795 1b
477 2b
215 3b
391 HR
2878 hits

If anybody has commentary to share on Stearnes' fielding prowess, now would be an excellent time to do so! There's a little info in the thread above, and Holway shows Stearnes shifting to the corners in the late 1930s in favor of a young Willard Brown in center field, but that's about all I have to work with so far.
   36. Scott Simkus Posted: August 27, 2008 at 10:48 AM (#2918308)
Chris, I can share a small nugget from the Strat-O-Matic research regarding the Turkey's outfield defense: We've looked at over 500 games from Stearne's prime, and he has a range factor and fielding percentage ranking up among the very best defenders. His arm, though, scores at about average, to slightly below (depending on one's interpretation of base-runner kills, etc.)...
   37. TomH Posted: August 27, 2008 at 12:09 PM (#2918332)
so, Turkey was basically as good a hitter as Oscar, and as good a fielder? With two fewer years of career play. That's how I read the MLEs.
   38. Gary A Posted: August 27, 2008 at 12:26 PM (#2918339)
so, Turkey was basically as good a hitter as Oscar, and as good a fielder? With two fewer years of career play. That's how I read the MLEs.

Without Charleston's peak, though.
   39. Scott Simkus Posted: August 27, 2008 at 12:56 PM (#2918357)
One thing to keep in mind with Turkey Stearnes is the Mack Park factor. Everybody here knows it was a good place to hit, but perhaps you don't know how good it actually was. One more nugget of info, and I'll jump back off-line before violating my research contract: I conducted a rather extensive ballpark study of Negro League parks this summer. A nice cross-section of over 3000 games from 1920 to 46. My goal was to determine the "historical norms" for home runs hit in particular ballparks; i.e. how many Hr did righties hit there, how many did lefties hit there, and how does that compare to splits in other ballparks. I was shocked to discover Mack Park was neck-and-neck with the Catholic Protectory Oval in NYC, in terms of home runs hit by left-handed batters per game. In other words, Stearnes hit for power in the PERFECT ballpark for lefties, whereas Oscar jumped around a bit and played in different places. Hope this helps.
   40. Chris Cobb Posted: August 27, 2008 at 02:09 PM (#2918423)
Scott raises a good point.

In response, I should mention that Stearnes' MLEs are park-adjusted, with Mack Park recognized as a hitters' park. The park adjustments don't, of course, address lefty-righty splits, and they may not be quite large enough.

It's worth noting, however, that Stearnes was just as successful a hitter when he played in Schorling Park for the Chicago American Giants in the 1930s, which was a completely different hitting environment, and he adapted his style to take advantage of both conditions.

Gadfly has a good post on this point at #33 above.

Voters may legitimately account for these matters in different ways, but my view is that Stearnes had the broad and flexible skill set that enabled him to take advantage of his parks: his greatness was not a product of his playing in an environment that was uniquely and fortunately favorable to his (limited and inflexible) set of skills.
   41. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 27, 2008 at 05:31 PM (#2918694)
Chris Cobb, I imagine the odd-numbered years are translated to the AL here as usual? Why do you do that? It just about doubles the time it takes me to get WARP out of these.

Any progress on Fielding WS, or are we awaiting more information?
   42. Chris Cobb Posted: August 27, 2008 at 07:12 PM (#2918870)
Yes, odd numbered years through 1929 are translated into the AL. I do that now mainly because that's how I got started, and I want to keep it consistent. I did in the first place because I wasn't certain competition levels in the AL and NL were equal during the period in question, so I thought I could diminish the effect of that possible discrepancy by switching leagues. Also, when I started, I did not have the data necessary to do simple OPS+ calculations, so I didn't want the "counting stats" part of the projection to be too affected by league context, and the two leagues were quite different. By the time I started doing MLEs for players with significant playing time in the 1930s, I was doing OPS+ and I knew the NL was generally viewed as the stronger league after about 1932, so I figured just projecting the NeLers into the NL would be fine. But the alternations up to 1920 are a legacy of my original approach. If I were ever to sit down and re-do all the MLEs, so as to make them all perfectly consistent, I could at that time just project into one league or the other.

I'm starting to work on FWS, using the information that we have so far to start identifying likely models. I won't have much time to work on them until tomorrow night, however, so I'm leaving that as a window for more commentary from those with knowledge of the NeL's history and statistics.
   43. Gary A Posted: August 28, 2008 at 01:07 AM (#2919437)
Quoting my post above (#13) on Mack Park:

Both teams at Mack Park in 1921 hit: .262/.323/.397, with 21 triples and 42 home runs in 2169 PA.

Both teams in Stars' road games hit: .269/.323/.358, with 37 triples and 12 home runs in 2322 PA.

AVE: .975164
OBA: 1.001216
SLG: 1.107255
HR: 3.746888
   44. Gary A Posted: August 28, 2008 at 01:08 AM (#2919440)
And quoting from post 12, on Stearnes's fielding in 1928 (the only fielding stats I have for him):

Fielding (cf)
G-78 (team 80)
DI-672.3
PO-191
A-7
E-8
DP-0
RF-2.65 (NNL cf 2.48)
FPCT-.961 (NNL cf .963)

Detroit Stars' center fielders (almost entirely Stearnes) accounted for 46.8% of the team's outfield putouts; the NNL average was 42.7%.

Stearnes also played two games in left field.
   45. Paul Wendt Posted: August 28, 2008 at 01:59 AM (#2919595)
Both teams at Mack Park in 1921 hit: .262/.323/.397, with 21 triples and 42 home runs in 2169 PA.

Both teams in Stars' road games hit: .269/.323/.358, with 37 triples and 12 home runs in 2322 PA.


Measured by PA that is 7% greater playing time on the road. What is the league average = league total ratio of road to home play? If there were almost no neutral site games (road for both teams) then that league ratio will be 1 and the league-average home park factor will be very close to 1.00 or 100.
   46. Chris Cobb Posted: August 28, 2008 at 11:13 PM (#2920937)
Turkey Stearnes Fielding Win Shares

Age G    Rate  FWS
22  142  3.72  4.8
23  119  3.81  4.1
24  151  3.96  5.4
25  154  5.02  7.0
26  146  5.02  6.6
27  135  3.81  4.6
28  148  3.81  5.1
29  151  2.04  2.8
30  138  3.16  3.9
31  140  2.23  2.8
32  152  3.91  5.3
33  154  4.65  6.4
34  147  4.00  5.3
35  134  3.63  4.4
36  111  3.10  3.1
37   82  2.52  1.9
38  115  2.27  2.3
39  110  2.88  2.9
   2429  3.6  78.6 


Rate is WS/1000 innings

Notes. Given what little we knew of Stearnes’ defense, I rated him at about 90% of the Richie Ashburn/Willie Davis level of defensive excellence: about even with Larry Doby, as win shares sees it: an above average but not great defensive center fielder. I used Willie Davis’s career shape, more or less, moving the high years and low years around a little bit to match what appear to be great years and slight injury years for Stearnes. Davis seemed like a good match because he had a long career at the position and was particularly noted for speed, as was Stearnes.

Now on to Cristobal Torriente for more MLEs!
   47. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 29, 2008 at 04:43 PM (#2921528)
Chris Cobb, do you really mean to make Stearnes an absolute superstar in CF in '26 and '27? Those are league-leading Fielding WS totals in both years; they translate to +18 and +15 runs above average, which is REALLY pickin' it.

Assuming those are correct, here is Stearnes:

Year SFrac BWAA    BRWAA FWAA Replc WARP
1923  0.89  4.2      0.1  0.8  
-0.9  5.9
1924  0.75  3.2      0.1  0.5  
-0.7  4.5
1925  0.97  4.3      0.2  0.4  
-1.0  5.8
1926  1.02  6.7      0.2  1.8  
-1.0  9.8
1927  0.96  5.1      0.2  1.4  
-1.0  7.7
1928  0.88  3.2      0.1  0.2  
-1.0  4.5
1929  0.97  5.3      0.3  0.3  
-1.0  6.9
1930  0.96  3.2      0.2 
-0.7  -1.0  3.7
1931  0.90  3.6      0.2 
-0.2  -0.9  4.6
1932  0.90  4.0      0.3 
-0.7  -1.0  4.5
1933  1.01  5.6      0.0  0.5  
-1.1  7.1
1934  1.01  5.2      0.3  1.3  
-1.1  7.8
1935  0.95  6.7      0.2  0.5  
-0.9  8.4
1936  0.85  2.3      0.1  0.5  
-0.7  3.6
1937  0.72  1.5      0.0  0.1  
-0.6  2.3
1938  0.51  0.5      0.1 
-0.1  -0.5  1.0
1939  0.68  0.4      0.2 
-0.1  -0.7  1.1
1940  0.64  0.7      0.1  0.2  
-0.6  1.6
TOTL 15.59 65.5      3.0  6.7 
-15.6 90.8
AVRG  1.00  4.2      0.2  0.4  
-1.0  5.8 


3-year peak: 26.0
7-year prime: 53.6
Career: 90.8
Salary: $273,999,551, the cusp of the inner circle. This is below DiMaggio but well above Griffey and then Hamilton among CF, and on a par with Mize, Mathews, and Bench overall. Does this placement jive with Stearnes's reputation? It would definitely make him the #4 NgL position player after Gibson, Lloyd, and Charleston. A more even distribution of the Fielding Win Shares would knock a bit off his peak, but not enough to substantially affect his placement.
   48. DL from MN Posted: August 29, 2008 at 06:21 PM (#2921666)
No kidding, it bumps him ahead of Paige and Dihigo in my spreadsheet and from 50th to 40th all-time. Still behind DiMaggio.
   49. Chris Cobb Posted: August 30, 2008 at 12:58 AM (#2922343)
Chris Cobb, do you really mean to make Stearnes an absolute superstar in CF in '26 and '27? Those are league-leading Fielding WS totals in both years; they translate to +18 and +15 runs above average, which is REALLY pickin' it.

My examination of the career trajectories of centerfielders who are, as far as I can tell from what we know (which is not much) about Stearnes as an outfielder have tended to have a peak of a couple of seasons where they were league-leading fielders. Looking at those models, I try to create a career shape for the NeL player that fits the facts of the career as we know them and that fits usual patterns of fielding value progressions, so since the models for Stearnes have peaks about like that, he has a peak about like that, but 10% lower. How that ends up lining up season by season with the leaders in the National and the American leagues is not something that I look at. I do try to make sure that his totals do not violate the range of possible values for a given era, but I don't think that's a problem here.

All I can say is that given what I see of Stearnes as a player, and given what I see of players that seem similar, a fielding peak like that does not seem improbable.
   50. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 30, 2008 at 01:24 AM (#2922382)
Chris, I was putting Stearnes in CF for his whole career. Did he in fact play some corner outfield? If so, which years? It's only around an 0.2 win per year difference at that point, so it won't move his rank much, but just in the name of accuracy...

I should have mentioned that I was adding in a league-average HBP rate to those MLE's in extra PA's, as you said I should do with Charleston.
   51. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 30, 2008 at 02:01 AM (#2922436)
Chris, .421/.593 in the 1929 AL would be a 166 OPS+, not a 148.
   52. Chris Cobb Posted: August 30, 2008 at 02:30 AM (#2922490)
Chris, I was putting Stearnes in CF for his whole career. Did he in fact play some corner outfield? If so, which years? It's only around an 0.2 win per year difference at that point, so it won't move his rank much, but just in the name of accuracy...

He appears to have been a corner outfielder for 1/2 of 1938 and all of 1939. The positional designations in Holway indicate that Stearnes played centerfield for the Chicago American Giants in 1938, but when he joined the KC Monarchs mid-season, he played right field while Willard Brown played center. This arrangement continued in 1939. In 1940, Brown left for Mexico, and Stearnes moved back to center field during his final season.

Thanks for the catch on 1929. If it's an issue for Stearnes, it will be an issue for Charleston as well. I'll take a look and see if I can find the error.
   53. Chris Cobb Posted: August 30, 2008 at 03:52 AM (#2922561)
I'll take a look and see if I can find the error.

Well. "Error" in the singular is a bit mild of a word. I reviewed the league average numbers that I have been using for (1) the OPS+ calculation at the end of the process and (2) the league context adjustment earlier in the process, and I found that I have failed to maintain consistency in the league data I have been using for these adjustments in a couple of significant ways.

(1) I have stuck by the alteration of NL and AL in the league context adjustments, but I have been using straight NL numbers for the OPS+ calculation at the end.
(2) Although I have switched from applying the .81 slugging adjustment to slugging percentage as a whole to ISO only, I found that I have still been adjusting the league context for ISO with the same numbers I was using when I was adjusting slugging as a whole.

This is what comes of trying to work quickly with a cobbled-together system that has evolved over several years, after not using it at all for a couple of years :-( .

I am sorry for introducing such errors into the process. I expect that inconsistency (1) above has not had a huge effect on the results of my calculations. I fear that inconsistency (2) has had a significant effect.

I will make it my project over the weekend to go through both the league context adjustment and the OPS+ calculation and make sure that I have proper numbers for each season from 1916 to 1948.

Since it looks to me like the best way to do this will be to undertake a true redesign of my spreadsheet (going from system 1.5 or whatever to 2.0), I will also take the opportunity to implement some changes in the way that I handle the regression, so that the multi-season average established is less influenced by big changes in offensive level from one year to the next.

Since I am going to be redoing all this data and the spreadsheet design, I could, at this point, start projecting the players exclusively into one major league or the other. That would be more convenient for Dan R. Are there reasons to stick with the old, alternating system, or should I make this change? If I make it, which league should I use? Does it matter?
   54. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 30, 2008 at 01:28 PM (#2922688)
Use the NL. Aesthetically, it was the first league to integrate.

Why would the multi-season regression be influenced *at all* by offensive level? Shouldn't you be regressing some form of OPS+ (really Runs Created-plus)?

I can't quite follow the math--will this make NgL'ers look better or worse in translation, on the whole?
   55. Paul Wendt Posted: August 30, 2008 at 01:44 PM (#2922691)
I think there are some good reasons to go straight NL and none to go straight AL so it does matter some if you do make the change.

The National League is available as a point of reference for all such work from 1876 to date. The American League is not available before 1901 and it is a DH league after 1972. Directly that supports choosing the NL in any work with a timespan that overlaps 1901 or 1972. Indirectly it supports the NL in general because there it is valuable to choose the same point of reference that others do. Here there is no conceivable overlap with 1972 and we aren't close to a useful database that overlaps 1901. There may yet be a direct argument for the NL in its earlier and faster integration, so that more careers of Negro Leagues players include moves to the NL than moves to the AL after 1947.
   56. Chris Cobb Posted: August 30, 2008 at 02:22 PM (#2922715)
Why would the multi-season regression be influenced *at all* by offensive level? Shouldn't you be regressing some form of OPS+ (really Runs Created-plus)?

Because I built the regression equations before I had access to reliable league offensive levels, before I was even attempting to calculate OPS+. Dan, you were away from the project, developing your own WAR, during the years in which the MLE translations were built: you have no idea how many iterations and patches they went through as ways were found to refine the system and as more complete data became available. I'm still not sure it makes sense to regress by OPS+ itself because we still don't have league walk rate information for most NeL seasons. I am pretty confident at this point that we are getting BA and SA close to right, so I want to have regressed seasonal figures for them independent of the more speculative walk data. So I regress the three elements independently and then combine them, rather than combining them and then regressing them. Are there statistical arguments against that approach?

You also have no idea how weak my statistical and technical foundations are. I had never _used_ an Excel spreadsheet prior to starting to do MLEs, and I learned how to calculate regression to the mean in order to incorporate regression into the projections. I don't have the tenth of your knowledge of statistics in general or of baseball statistics in particular. I developed the MLEs because I thought they needed to be developed, because I was willing to do the work, and because I thought I could think through the basic problems of developing translations clearly enough to achieve useful results. I do not have the knowledge, the data, or the time (at least not in the "let's do it this week" sense) to make the MLEs as rigorous as they could be on my own, though I am resuming work on them.

All that means that _many_ aspects of the procedure for calculating MLEs have been done in ways that, in retrospect, were not ideal, and that it has not always been possible for me to make improvements in straightforward ways because I haven't had time, as I now hope to find, to rebuild the translation spreadsheet from scratch.

Please keep these factors in mind when asking questions about the MLEs.

I can't quite follow the math--will this make NgL'ers look better or worse in translation, on the whole?

I think that the league context issue will have little cumulative effect. The offensive levels of the NL and AL were not that far apart in the 1920s overall, though season-to-season they swung quite a bit. I expect NeL players may go up in some seasons and down in others, and that this will have little effect on their career values.

I don't know what effect changing the ISO adjustments will have. I think it will either not make much difference or it will make the Negro-Leaguers look better in translation. When Brent suggested switching from applying the SLG competition-level adjustment from SA to ISO, I did some tests that led me to believe that the change was needed to deal appropriately with very low slugging environments, but that it would not have much effect on higher-slugging ones. I still think the change is appropriate, I don't know if my sense of it having little effect was accurate. I will have to run the numbers on Charleston, and Stearnes, and Torriente and see what we find.

I think there are some good reasons to go straight NL and none to go straight AL so it does matter some if you do make the change.

This all makes sense. NL it will be then.
   57. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 30, 2008 at 04:50 PM (#2922846)
Chris, I hope I didn't come across as criticizing your work!! I just didn't know what approach you were taking. Obviously, we are all indebted to you for making it possible for us to even get a handle on these Negro Leaguers; we should all probably express our gratitude more regularly.

Hah! I am hardly a quantitative wizard--my formal background in this consists of one semester of college statistics. In fact, rather often--both in my WARP work and for my New York Times columns--I crash head-on against the limits of my knowledge. There are plenty of mathematical things I'd love to learn how to do--working with multinomial probabilities and doing logistic regressions leap to mind--that I simply don't have the tools for, and I can't understand the tutorials available on the Internets. I'm pretty good at toying with Excel, just basically through extensive parsing of the Help file, and I know how to do simple multiple regressions. That's really about it, sadly, but it's enough for most of my purposes.

As for whether you should be doing a "tripartite" versus a "unified" regression, I can see merits in both approaches. It's basically similar to the question about minor league MLE's: do you translate the components or the value? For example, minor league hitters with high BB and little power very rarely succeed at the big league level, but they definitely win games for their AAA employers. By contrast, minor league groundball pitchers tend to translate quite well to the majors (due to better fields and defense lowering BABIP on GB, and the higher HR per flyball rate in The Show increasing the contextual value of FB prevention.) Do we want to know how these players would do if they were in the bigs, or do we just want to apply an overall competition discount to how valuable they actually were in their own context? To recycle a distinction I've made in the past, this seems to me to be a normative rather than positive question: there is no One Right Answer, there are two equally valid answers, depending on each voter's definition of Merit. (My own instinct leans more towards the what-would-they-have-done-in-the-majors side, which is why I don't feel comfortable giving my full catcher bonus to Josh Gibson, for example, since I don't think he would have been a full-time catcher for his whole career had he been permitted to compete).
   58. Chris Cobb Posted: August 31, 2008 at 05:29 PM (#2923526)
Chris, I hope I didn't come across as criticizing your work!! I just didn't know what approach you were taking. Obviously, we are all indebted to you for making it possible for us to even get a handle on these Negro Leaguers; we should all probably express our gratitude more regularly.

Not at all. I just wanted to make sure that you (and others) aren't expecting too much from my creaky MLE translation contraption.

Your questions, which are perfectly legitimate and sensible questions, tend to address limitations of the translation system that I have been aware of but haven't gotten around to trying to fix, so I feel a combination of grumpiness and frustration that I haven't done all that I could to get the system right, and that my explanation of why things are the way they are will have to include the caveat that I could have fixed it but I didn't, or if I had been smarter, I would have designed it the right way in the first place.

I think the corrections and upgrades your questions prompted me to make have substantially improved both the equations and the functionality of the translation spreadsheet, to the point that I am thinking about uploading them to the HoM Yahoogroups area, so that others can use them, if they want, and begin to work out improvements.

The system could still be refined in many ways, but its modeling is more precise than it has been before.
   59. Chris Cobb Posted: August 31, 2008 at 07:25 PM (#2923796)
Turkey Stearnes MLEs, version 2.0

Age Year Team G     PA   Hits TB    BB  SB   BA     OBP    SA    OPS+
22  1923  Det 142   568  176  313   34   3  0.329  0.370  0.586  149
23  1924  Det 119   476  143  238   37   2  0.325  0.377  0.542  144
24  1925  Det 151   634  191  333   67  13  0.337  0.407  0.587  151
25  1926  Det 154   647  200  346   69  21  0.346  0.415  0.599  170
26  1927  Det 146   613  175  314   75  15  0.326  0.408  0.583  163
27  1928  Det 135   567  154  289   57   7  0.301  0.371  0.566  142
28  1929  Det 148   622  189  338   86  20  0.353  0.442  0.631  162
29  1930
Det 151   634  187  371   71  29  0.332  0.406  0.659  152
30  1931
Det 138   580  155  282   60  16  0.299  0.372  0.544  143
31  1932  Chi 140   588  165  301   61  25  0.314  0.385  0.572  153
32  1933  Chi 152   638  183  315   49   0  0.311  0.364  0.535  154
33  1934  Chi 154   647  201  386   53  24  0.339  0.393  0.650  174
34  1935  Chi 147   617  202  334   72  20  0.371  0.444  0.612  180
35  1936  Phi 134   563  156  255   50  13  0.303  0.365  0.496  130
36  1937
Det 111   466  120  186   48  14  0.286  0.359  0.444  117
37  1938
Chi  82   328   79  132   25   9  0.263  0.320  0.436  106
38  1939  KC  115   437  107  164   42  13  0.271  0.340  0.415  102
39  1940  KC  110   418   91  150   49   6  0.246  0.335  0.406  104
career       2429 10043 2873 5046 1005 250  0.318  0.386  0.558  147

1930
Also played for NY Lincoln Giants
1931
Also played for KC Monarchs
1937
Also played for Chi Am Giants
1938
Also played for KC Monarchs 


Unlike Charleston, who hardly changes at all with the new MLEs, Stearnes shows better. Not night and day better, but better.
   60. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 31, 2008 at 08:19 PM (#2923911)
Again, assuming there are no changes to the Fielding Win Shares or the ratios of 2B, 3B, and HR among XBH, and adding in hit-by-pitches in extra plate appearances at the league average rate, here's the new Stearnes:

Year SFrac BWAA BRWAA FWAA Replc WARP
1923  0.88  3.8   0.0  0.4  
-0.8  5.0
1924  0.75  3.1   0.0  0.5  
-0.7  4.3
1925  0.99  4.8   0.1  0.7  
-1.0  6.6
1926  1.02  6.5   0.2  1.8  
-1.0  9.6
1927  0.96  5.6   0.2  1.2  
-1.0  8.0
1928  0.88  3.5   0.1  0.2  
-1.0  4.8
1929  0.95  5.7   0.1  0.5  
-1.0  7.3
1930  0.96  4.5   0.2 
-0.7  -1.0  5.0
1931  0.90  3.7   0.2 
-0.2  -0.9  4.7
1932  0.90  4.7   0.3 
-0.7  -1.0  5.2
1933  1.01  5.2   0.0  0.5  
-1.1  6.7
1934  1.01  6.6   0.3  1.3  
-1.1  9.2
1935  0.95  7.0   0.2  0.5  
-0.9  8.6
1936  0.85  2.7   0.1  0.5  
-0.7  3.9
1937  0.72  1.6   0.0  0.1  
-0.6  2.4
1938  0.51  0.4   0.1 
-0.1  -0.5  1.0
1939  0.68  0.6   0.2 
-0.1  -0.7  1.3
1940  0.64  0.7   0.1  0.2  
-0.6  1.6
TOTL 15.59 70.6   2.4  6.6 
-15.6 95.2
AVRG  1.00  4.5   0.2  0.4  
-1.0  6.1 


3-year peak: 27.4
7-year prime: 56.0
Career: 95.2
Salary: $292,976,849, an inner-circle Hall of Famer. DiMaggio needs credit for 1935 to beat Stearnes. Overall, this is similar to Arky Vaughan, worse than Foxx, a cut above Mize and Mathews. *Extremely* impressive. Does Stearnes's reputation support a placement alongside this magnitude of immortals?
   61. Chris Cobb Posted: August 31, 2008 at 08:51 PM (#2923939)
Does Stearnes's reputation support a placement alongside this magnitude of immortals?

My sense is that it does.

Everybody saw him as a complete player, though he never had quite the defensive rep of Charleston.

Gadfly above suggested that Stearnes was comparable to Stan Musial. Gadfly's comparisons were always a bit over the top, but he was also immensely knowledgeable about black baseball, and he placed Turkey Stearnes among the upper echelon of Negro Leaguers.
   62. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 31, 2008 at 09:30 PM (#2923976)
Your MLE's show Stearnes as a markedly *superior* fielder to Charleston...

Well, this suggests we've got six inner circle Negro Leaguers: Gibson, Lloyd, Charleston, Stearnes, Williams, and Paige. That compares to 18 white inner circle players from 1900-47: Ruth, half of Williams, Wagner, Cobb, Speaker, half of Musial, Hornsby, Collins, Lajoie, Gehrig, Ott, Foxx, Vaughan, DiMaggio, Johnson, Young, Alexander, Mathewson, and Grove. By contrast, post-integration, we have seven inner circle minorities (Bonds, Mays, Aaron, Morgan, Henderson, A-Rod, F. Robinson, and no pitchers unless you count Gibson or Martínez) compared to nine inner circle whites (half of Williams, half of Musial, Mantle, Schmidt, Ripken, Seaver, Spahn, Maddux, Clemens, and Randy Johnson). This tells me a few things:

1. Six inner-circle Negro Leaguers is certainly within the realm of plausibility.
2. My standard deviation adjustment *really* is no substitute for a timeline adjustment--it does not pass my smell test that there were 24 inner-circle players from 1900-47 and only 16 from 1948 to the present.
3. Why have no white position players over $400M debuted since 1951, and no black pitchers of that magnitude debuted since integration? Is it because of racist "stacking" that channels minorities away from "mental" positions like pitcher?
   63. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 31, 2008 at 09:50 PM (#2923989)
Man, Stearnes really looks like Joe DiMaggio without crediting 1935 in these MLE's:

Stearnes: 95.2 WARP2 in 15.6 seasons (70.6 BWAA, 2.4 BRWAA, 6.6 FWAA)
DiMaggio (with credit for '43-'45 but not '35): 94.2 WARP2 in 14.4 seasons (72.4 BWAA, 2.6 BRWAA, 4.9 FWAA)

Stearnes does come out with a higher peak, but that's likely because Chris has distributed his best fielding seasons so that they line up with his best hitting seasons, an advantage DiMaggio lacks. This is a similarity score of like 990 or something.
   64. OCF Posted: August 31, 2008 at 09:57 PM (#2923993)
...and no pitchers unless you count Gibson or Martínez...

One of the odd little things about Gibson's biography is that he has a sliver of his life in the institutions of segregated sports. To be exact, he played a season for the Harlem Globetrotters. Now, that's basketball and not baseball, and the Globetrotters lived on by emphasizing show over sport, but if you look at the culture around them - who promoted them, how they made money, and so on - there's a kinship with the Negro Leagues.
   65. Blackadder Posted: August 31, 2008 at 11:23 PM (#2924039)
I know his career value is lacking, but somehow keeping Pedro out of the Inner Circle seems wrong.

Pujols, assuming he does not get hit by a bus, looks like a lock to eventually be Inner Circle. I can't think of anyone else today, not currently there, who has much of a shot.
   66. Mike Webber Posted: September 01, 2008 at 01:49 AM (#2924153)
Have I told you my Bob Gibson story? I was at a Doctors convention with my wife, and some drug company had brought in Bob Gibson to sign balls with the drug company logo on them. I'm in line with one of the DR.s and we are talking about Gibson, and I told him about Gibson's Globetrotter days. A couple of minutes later a lady comes up to lady in line right behind me and asks who she is waiting to get an autograph from.
"Well, I'm not sure who he is exactly, but he used to be a Harlem Globetrotter," she explained.


I always thought Turkey Stearnes was kind of a Duke Snider type, clearly a HOMer but not really the best CFer of his time. Of course that is a perception based on reading not on stats or anything.
   67. David Concepcion de la Desviacion Estandar (Dan R) Posted: September 01, 2008 at 02:28 AM (#2924193)
It's not just the lack of career value--it's that except for '97, he was only throwing 215 innings a season at his absolute peak, and that in an era of very high standard deviations for ERA+. Koufax didn't quite have Pedro's ERA+, but he was leading the league in innings, and a 180 ERA+ in the '60s was more impressive than a 180 ERA+ in 2000. Pedro's highest rank in innings was fourth.

There's no doubt that if you want one guy to pitch a Game 7 in all of baseball history, it's '99-'00 Pedro. But first you have to make the playoffs, and for that, volume counts just as much as quality.

Yes, I agree that Pujols is on track for the inner circle, but its lower reaches--is he really going to be greater than Foxx, for example? He might last longer, although I think he will need major surgery at some point for this condiditon that he miraculously plays through. I suppose if you take Probabilistic Model of Range at face value, which suggests that he's something like +25 with the glove, then yes, he could get pretty high in a hurry.
   68. David Concepcion de la Desviacion Estandar (Dan R) Posted: September 01, 2008 at 02:29 AM (#2924197)
Well, Stearnes was a contemporary of Charleston--more Foxx to Charleston's Gehrig than Snider to Charleston's Mantle, according to these MLE's.
   69. TomH Posted: September 01, 2008 at 07:55 PM (#2924806)
thanks, guys, for your incredible efforts in attempting your best to give us data for Stearnes, Oscar, etc. I wish I had the time (of course, I DO, but like everone else I choose to work on other things...) to analyze your framework. I hope others make efforts to sanity-check your studies.
   70. Paul Wendt Posted: September 01, 2008 at 08:14 PM (#2924873)
Regarding the 1900-1947 analysis

debut years
1906 Lloyd
1910 Williams
1916 Charleston
1923 Stearnes
1926 Paige
1931 Gibson

That is rather concentrated within the timespan. No 1932-1947(48?) arrival in the Negro Leagues is "inner circle"(?). Meanwhile approximately Young and half each of Lajoie, Wagner, Mathewson precede Lloyd.

When did the post-1947/48 black inner-circlists debut? Three in the 1950s, one each in the 60s, 70s, 80s, 90s. Joe Morgan 1963 is the median. Meanwhile we have one "surprise" (for those who have not paid close attention) on the white side. Five of the nine are pitchers and three of those five debuted in the 1980s, after the passing of occasional 300-inn and commonplace 250-inn seasons. Otherwise we have two outfielders (one-and-two-halves) at the beginning of the period, Mike Schmidt, and Cal Ripken.

--
I'm not sure why posterity lost Turkey Stearnes. Long-time SABR Negro Leagues Committee chairman Dick Clarke made him a username and a license plate and sometime in the last decade that cmtee has marked his unmarked gravesite. Why were those opportunities still available to Clarke and the NLC?

Regarding the Hall of Fame the explanation for Stearnes as for Williams must begin in the 1970s. Why did Cooperstown's original Negro Leagues Cmte close up shop after electing only nine, and why not Stearnes among the first nine?

When Cooperstown agreed to a five-year second chance it was a special ballot under the jurisdiction of the famous old veterans committee. They didn't elect Stearnes quickly, only under a two-year extension.

Whom did they honor?

HOFyr birthyr
1995 ; 1916 Leon Day
1996 ; 1904 Bill Foster
1997 ; 1908 Willie Wells
1998 ; 1889 Bullet Rogan
1999 ; 1885 Joe Williams
and finally on reprieve,
2000 ; 1901 Turkey Stearnes
2001 ; 1907 Hilton Smith

Day died that winter, March 16. Wells was about ten years gone, the others longer.
   71. Brent Posted: September 01, 2008 at 09:28 PM (#2925128)
Why did Cooperstown's original Negro Leagues Cmte close up shop after electing only nine, and why not Stearnes among the first nine?

I've assumed that their decision to elect only nine was influenced by the epilogue of Peterson's Only the Ball Was White, which put forward an argument for electing eight. (Peterson's argument was based on representation proportional to population--blacks represented 10% of the population and, at the time he finished writing the book, the Hall of Fame included 68 white players from the period 1900-1947.)

It's been noted that the original NeLg Cmte consisted almost entirely of players, executives and writers from the East and that they elected some marginal Eastern players while overlooking several more prominent Western players.
   72. Gary A Posted: September 01, 2008 at 10:31 PM (#2925210)
This is the original committee appointed by Bowie Kuhn, with the team (or newspaper) they are most identified with:

--Negro League Players--
Monte Irvin (Newark Eagles)
Roy Campanella (Baltimore Elite Giants)
Judy Johnson (Hilldale)
Bill Yancey (NY Black Yankees)

--Promoters / Officials--
Eddie Gottlieb (Philadelphia Stars)
Alex Pompez (Cuban Stars / NY Cubans)
Frank Forbes (NY Cubans; also a player w/ Lincoln Gts & Penn Red Caps)

--Journalists--
Sam Lacy (Baltimore Afro-American)
Wendell Smith (Pittsburgh Courier)

--Major League Player--
Eppie Barnes (Pittsburgh Pirates)

Joe Reichler and Dick Young were “ad hoc” members who apparently didn’t vote.
   73. Mike Webber Posted: September 01, 2008 at 11:26 PM (#2925258)
Why did Cooperstown's original Negro Leagues Cmte close up shop after electing only nine, and why not Stearnes among the first nine?


I believe there was one Negro Leaguer for each of the nine positions, though that makes DiHigo the second baseman. You will occasionally see him referred to as a 2b/Pitcher, and I think that is why.
   74. Chris Cobb Posted: September 01, 2008 at 11:28 PM (#2925260)
I wish I had the time (of course, I DO, but like everone else I choose to work on other things...) to analyze your framework. I hope others make efforts to sanity-check your studies.

Certainly, the framework should be analyzed, but I think we're past the sanity-check stage as far as the MLEs. Not that sanity checks aren't important, but the changes in the system now amount to fine-tuning: the basic methods and conversion factors are the same ones that I have been using since the MLE project began, so there's nothing major here that hasn't been scrutinized lots of times. There could still be elements that are wrong, but we would need more data to identify errors and correct them.

Basically, I just want to make it clear that, when you see these new numbers, they are not very different from the old numbers, and there's not a new process that has to be vetted. Rather, it's an old process that has been refined.

I never finished old MLEs for Turkey Stearnes, but I got far enough to have regressed, career batting averages and slugging averages for him, based mainly on the data from the Macmillan 8th edition encyclopedia. That "original system" Stearnes projection was .325/.537. The "new system" Stearnes projection is .317/.558.

Some of that difference may be a result of the underlying data, since the new projection uses the HoF data and all the complete seasonal data from Gary A. Some of it may result from the change in applying the slugging conversion rate to ISO rather than slugging percentage. Some of it may result from having more accurate league context information from the NeL.

But whatever the sources of the differences, the two projections are still showing us quite similar players.
   75. Paul Wendt Posted: September 02, 2008 at 04:31 PM (#2925935)
73. Mike Webber Posted: September 01, 2008 at 07:26 PM (#2925258)

> Why did Cooperstown's original Negro Leagues Cmte close up shop after electing only nine, and why not Stearnes among the first nine?

I believe there was one Negro Leaguer for each of the nine positions, though that makes DiHigo the second baseman. You will occasionally see him referred to as a 2b/Pitcher, and I think that is why.</i>

Ha! I wrote "Why ... electing only nine, and why precisely one at each position, and why not Stearnes among three outfielders?" but I checked my facts and corrected it as you see.

The three outfielders were Irvin, Bell, and Charleston in that order. They closed up shop with the middle infielders together, Dihigo and Lloyd. (The regular Veterans Committee elected Rube Foster as a manager.)
   76. Paul Wendt Posted: September 06, 2008 at 06:17 AM (#2930975)
After reading NYT coverage of the annual announcements:

>>
I believe there was one Negro Leaguer for each of the nine positions, though that makes DiHigo the second baseman. You will occasionally see him referred to as a 2b/Pitcher, and I think that is why.
<<
Yes, that probably explains any 2b-p listings for Dihigo. No, I don't believe there was any program to elect one at each position.

Probably they tried to elect living players. Here is the list, year by year.
bold : living
underline : committee member (Irvin, thruout; Johnson, ??)

Paige
Gibson, Leonard
Irvin
Bell
Johnson
Charleston
Dihigo, Lloyd

The first five years, 1971-75, there was a formal announcement, reception, press conference hosted by Bowie Kuhn, sometimes on the day of committee meeting. The elected players were present, one each year; Paige and Johnson (at least) with their wives. It appears that NYT learned the outcome only at that event, whether or not it was the night of the committee meeting.

So there was some advance planning, and it appears to me there was some secrecy.

The NYT article actually mentions a time lag and advance notice to the player only once.
1972-02-04
Yesterday ""Bell was announced as a member of the Hall of Fame by its special committee on black baseball."
. . . Even when he found out last week that he had joined Paige, Josh Gibson, Buck Leonard, and Monte Irvin at Cooperstown for their accomplishments in the black leagues, he didn't get excited." [That is the "Cool Papa" theme.]

The NYT coverage in 1973 and 1975 does not mention that Irvin and Johnson were committee members.
   77. Tiboreau Posted: February 03, 2009 at 04:21 AM (#3066212)
Does Stearnes's reputation support a placement alongside this magnitude of immortals?

I was messin' around today, creating a list of the Top 25 position players in baseball history. Using solely Dan's WARP, Turkey Stearnes makes my list at #21 (excluding non-HoF eligible ballplayers: Barry Bonds). Looking to compare my rankings to the first expert who popped into my head, Bill James, I broke out the NBJHBA and listed his top 25 HoF eligible position players. James ranks Turkey Stearnes, based solely on reputation, 19th, 2 spots ahead of my own ranking of Stearnes--once again, based on Dan Rosenheck's interpretation of Chris Cobb's MLEs--and 2 spots ahead of his ranking of John Henry Lloyd. Figured this would be one more data point, along with Ted Knorr's comparison of Stearnes to Musial, concerning Turkey Stearnes's reputation. . . .
   78. Paul Wendt Posted: February 03, 2009 at 04:38 PM (#3066464)
James ranks Turkey Stearnes, based solely on reputation, 19th . . . 2 spots ahead of his ranking of John Henry Lloyd.

Stearnes based solely on reputation and Lloyd based solely on reputation?

Seriously,
1.
The next Negro Leagues player in his rankings is Mule Suttles #43. That unusual call may be informative regarding how Bill James gauged reputation.

2.
Here are the Bill James rankings of major league players other than pitchers. Bold marks the centerfielders, five in the top ten.
Ruth
Wagner
Mays
Cobb
Mantle

Williams
Musial

Speaker
Aaron
DiMaggio

Gehrig
Morgan
Bonds [productive at press time]
Collins
Schmidt
Hornsby
Robinson, Frank
Henderson
Ott

Foxx

Overall Foxx is #29. Make way for four major league pitchers, Satchel Paige, and four other Negro Leaguers. He puts Charleston just above Cobb (#4 overall), Gibson above Musial (at 9-10), Stearnes and Lloyd above and below Henderson (overall #25, #27).

Are there nearby points of reference at the same fielding positions? Only for Charleston. The next CF is Duke Snider #50(!). The next shortstops are Vaughn #39 and Ripken #48. The next catchers are Berra #41 and Bench #44.

My analytical speculation regarding James's reference points now seems too flimsy and unimportant to say. Instead let me out with a display of the primary fielding positions of his high-ranking players.

Numerals are the overall ranks.
Bold now represents the Negro Leaguers.
* is Mark McGwire #31 which James revised to about 70 two years later.
'OF' is other field, left or right

P : 8 17 19 20 23 36 38
C : 9
1b: 14 29 *
2b: 15 18 22 32 35 40
ss: 2 27 39
3b: 21 30 34
CF: 3 4 5 6 11 13 25 (next, 50)
OF: 1 7 10 12 16 24 26 28 33 37

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Ray (RDP)
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.1012 seconds
49 querie(s) executed