You can look back, Satch. It's only your impending HoM induction gaining on you.

Satchel Paige Data

From Riley

Born July 7, 1906

Teams: Birmingham Black Barons (1927-30), Baltimore Black Sox (1930), Cleveland Cubs (1931), Pittsburgh Crawfords (1931-37), Bismarck semi-pro team 1935, Santa Domingo 1937, Mexican League 1938, Kansas City Monarchs (1935-36, 1939-48, 1950), Cleveland Indians (1948-49), St. Louis Browns (1951-53). Appeared also with NY Black Yankees and Memphis Red Sox (1943), and the Philadelphia Stars (1946, 1950)

Seasonal Data from Holway

1927 8-3, .738 wp 2nd in league. 4th in team dec. 2.0 WAT
1928 13-4, 5th in wins, .765 wp 4th, 3.06 TRA 2nd, 112 K 2nd. 3rd in team dec. 6.2 WAT
1929 11-11, 194 K 1st. 1st in team dec. (tie). 4.3 WAT
1930 9-4, 3.07 TRA 3rd. 4th in team dec. 3.4 WAT
1931 6-3 for Clv., 1-2 for Pgh. 2.0 WAT Clv, -0.8 WAT Pgh, 1.2 WAT total
1932 21-9, 1st in wins, .700 wp 4th, 2.79 TRA 2nd, 113 K 1st. 1st in team dec. 4.5 WAT
1933 7-9, 3.50 TRA 5th, 69 K 1st. 2nd in team dec., -3.5 WAT.
1934 20-5, 2nd in wins, .800 wp 3rd, 1.84 TRA 1st, 134 K 1st. GSA. 1st in team dec., 2.1 WAT
1935 In Bismarck
1936 11-3, 1st in wins, .786 wp 1st, 3.15 TRA 4th, 68 K 1st, GSA, all-star. 1st in team dec., 3.5 WAT
1937 8-2 in Santa Domingo
1938 0-1 in Mexico in 3 games (arm trouble develops)
1939 pitching for KC Monarchs “B” Team (minor-league equivalent)
19-3 in Puerto Rican winter league (1st in wins, 3rd in TRA, 1st in K)
1940 2-0. Not in top 5 dec., 0.4 WAT
1941 7-0, 2nd in wins, 1st in wp, 2.11 TRA 3rd, 1st in K. 2nd in team dec., 1.8 WAT
1942 7-5, 3rd in wins, 1.95 TRA 3rd (in league & on team), 42 K 1st. 1st in team dec. –2.7 WAT
1943 9-15 for KC, 1-0 for Mem., 3rd in wins, 71 K 1st. 1st in team dec., -5.1 WAT
1944 5-6. 1st in team dec., 0.2 WAT
1945 5-9. 1st in team dec., -3.8 WAT
1946 5-1. 1.32 TRA 1st in league, 27 K 3rd. 4th in team dec., 0.6 WAT
1947 1-1. Not in top 5 dec., -0.6 WAT
1948-49 in majors
1950 No data
1951-53 in majors

Career Totals
147-88, .626 adding numbers above
147-92, .630 according to Holway (147-92 is a .615 wp, of course)
14.5 WAT (was 23.7 WAT from 1927-1936)

A more complex career record than the legend would suggest. Not that it matters that much, I suppose. He has enough value from 1927 to 1937 to merit induction. I'd like to have a more exact account of how Satchel was being used and against whom in the 1940s. His w-l record from those years is not impressive, and his WAT are poor, but he was leading his team in decisions. Was he pitching too much because he was the gate attraction? Was he facing the best pitcher on his opponents' teams every time? Or was Satchel's image just far better than his actual talent during these years?

Or was Satchel's image just far better than his actual talent during these years?

Given how well he pitched in MLB in his 40s, I'm inclinded to think he was at worst slightly worse than his reputation.

He did pitch very well (3.52 DERA!), but as a relief pitcher. Would he have been as successful if he'd been asked to throw 200-250 innings? His record in KC certainly suggests that he was throwing _lots_ of innings 1943-46 (things seemed to change in 1947), and that could well have diminished his effectiveness. So there's a big question in how to project MLEs for him for the period when he appears, at least, not to be any longer a lights-out starter, but in which he appears to be seeing very heavy use.

I'm pretty sure that simply using the approach to pitcher conversions that I have been using will not reflect well on Satchel at all during this period. I'd like to get a better sense of how his usage affected his record and his performance during these years.

1928 Satchel Paige
NNL Birmingham Black Barons

W-11
L-4
SV-4 (*1st in league)
TRA-2.93 (2nd)
G-24
GS-14
CG-10
IP-126
H-103
HR-2
W-21
K-112 (3rd)
HB-3
SB-6
DP-4
HR/9-0.14 (4th; NNL 0.40)
W/9-1.50 (7th; NNL 2.47)
K/9-8.00 (*-1st; NNL 3.97)

Paige wasn't overworked (the league leader in innings pitched had 230); the Black Barons distributed innings fairly equitably, with five pitchers throwing from 85 to 133, and two others just under 50.

Despite its reputation, the Birmingham park (Rickwood Field) played as a more or less neutral park in 1928.

one of the best pitchers ever.

& today (8/9) is the anniversary of his induction into Cooperstown.

If you spot Satch among all NeL pitchers, he is clearly not #1, but perhaps #2. Nevertheless, he is overrated because his reputationn would have him at #1 by a huge margin. Unlike Cool Papa Bell, he is a HoMer anyway (and yes, Angerfaced, possibly one of the top 10 pitchers ever) but he is probably overrated to the same extent.

Nothing like as good as Smokey Joe Williams, but unless Chris' conversions are FULL of surprises, he has to be a pretty easy HOM'er. Unlike Bell, we have his late ML career to look at, and with an ERA+ of 124, all after 40, it shows that even in his 40s he was among the best in the league.

A dip in his 30s wouldn't be unprecedented, though, so I would suggest that we do NOT go through huge contortions in projections to eliminate it, since it reduces the credibility of MLE projections for other guys who may need the solidity (though in practice I guess we've enshrined almost all the borderline cases by now.)
It's like adding walks for Willard Brown -- yes, he might well have added them in the majors, but he'd have been a subtly different player if he did, and you shouldn't just add nonexistent value.

The big thing I'd like to know is whether he faced disproportinate numbers of winning teams. In other words, what were his MOWP+s like?

The big thing I'd like to know is whether he faced disproportinate numbers of winning teams. In other words, what were his MOWP+s like?

From what I understand of how the NeL schedule worked, esp. in the 1940s (barnstorm during the week, meet for the big games against other league teams on the weekends), I doubt that Satchel's MOWP would be all that unusual. Gadfly has mentioned that big matchups between the top pitchers were frequently arranged. That makes economic sense--it's a way to hype a game, draw a crowd. It would be fairly easy to arrange with the barnstorm/league schedule the teams were playing. It fits the prevalance of stories about marquee pitching match-ups in the informal history of the leagues. So if Satchel was facing tougher opposition, it wouldn't be from facing tougher teams, but from pitching against every team's ace.

FWIW: in 1928 Paige wasn't yet a legend, but a brief look at his record shows 18 of his 24 pitching appearances against over-.500 teams.

Here are a few random thoughts on Satchel Paige:

1) When I first read his autobiography (Maybe I’ll Pitch Forever, “as told to David Lipman”) some 40 years ago, it was an important milestone in my adolescent life. I believe it was the first “adult” book that I read, and it not only opened my young eyes to the history of the Negro Leagues, but it also provided a gentle introduction to the great social issues of the time—segregation, integration, and civil rights. I loved the wit with which Satchel told his story.

Just a few months ago I ran across the book on the shelf of a local bookstore and decided to re-read it. I was curious to see if the book would hold up to age and the passage of time and retain its youthful charms. I’m pleased to say that I very much enjoyed re-reading it. With experience, I now can see that Paige was at times selfish, manipulative, and a braggart, but with his home-spun humor and charisma he continues to be a most likeable character. His story truly is a fascinating one.

2) How and when did Satchel Paige the pitcher become Satchel Paige the legend? It’s interesting the extent to which he remains a legend, known by countless people who’ve never heard of Carl Hubbell or Lefty Grove. When I purchased his autobiography a few months ago, the young woman at the register, who couldn’t have been older than 25, smiled with recognition and commented, “Oh, you’re reading about Satchel Paige.”

Actually, Paige’s autobiography suggests that for the white press, the legend may have started with a 1936 winter game in California against “Dick Bartell’s All Stars” that for some reason set off a flurry of articles about Paige. Interestingly, Paige actually lost that game (2 to 1). But by 1936 Paige clearly had already been a legend for several years within the black community. Can one of our Negro Leagues experts shed some light on the origins of the legend?

3) Chris’s statistical summary doesn’t mention Paige’s post-1953 record. Paige barnstormed in 1954-55, but in 1956 he was back playing for Bill Veeck again, this time for Miami in the International League. Assuming that the statistics given in his autobiography are accurate, it was an excellent season—he went 11-4 with a league-leading 1.86 ERA. This record suggests to me that he could have been a successful major league pitcher at the age of 50. The next season he went 10-8 and again led the league with a 2.42 ERA. He also pitched for Miami in 1958, though no statistics are given—apparently the season was marred by salary disputes. In 1959-61 he was back to barnstorming, but at the end of 1961 he pitched 25 innings for Portland in the PCL, with 19 strikeouts and a 2.88 ERA. He continued to barnstorm from 1962 through about 1967. And of course he pitched those 3 scoreless innings for Kansas City in 1965 at the age of 59.

4) During Paige’s major league career, I wonder why he wasn’t used as a Sunday starter, in the manner of Lyons and several other older pitchers of that era. In 1948, he actually did quite well as a starter—Cleveland won 5 of his 7 starts, and he pitched 3 complete games, including 2 shutouts. By the time he reached the majors he was relying on a mixture of pitches and deliveries. (Neyer and James describe his repertoire as “everything, including the kitchen sink.”) It seems to me that this type of pitcher tends to be more successful in a starting role than in relief. Also, Paige was always able to draw fans to the park, so why didn’t a team like the Browns, which didn’t have much else going for them, take advantage of his box office appeal?

If not for the legend of Satchel Paige (that Brent was talking about), The Bingo Long Traveling All-Stars and Motor Kings wouldn't have worked as well as it did as a movie. Mind you, the movie wasn't about Paige, but it was about the legend. (And the James Earl Jones character tapped into the legend of Josh Gibson.)

Here's my take on Paige:

If we run his numbers through the methods we have developed, he might well rank below the wonderful Mize. But that exercise misses the point. This isn't the Hall of What Did Your Do in the Offficial Negro Leagues Converted into White Major League Numbers.

Paige had a bizarre, wild, wonderful career of team-jumping, league-jumping, showboating, barnstorming, and following the almighty dollar b/c/ that seemed to be the best option open to him. If he had been allowed to pitch in the big leagues, he would have taken his wonderful stuff, his incredible competitiveness, and his amazing ability to bounce back from injuries and been a rotation anchor for 15 or 20 years. (He also probably would have worn out his welcome in a couple of cities and pitched for three or four teams, but so what?)

The available evidence both statistical and anecdotal suggests that, when sitting still and focusing on the fortunes of one team at a time, Paige pitched at at least the level of a Maddux or Seaver or Spahn or Matthewson, though possibly not at the level of Johnson, Grove, or Clemens. I see no reason why he wouldn't have continued to pitch at that level for 15 or 18 years if given a chance to get rich doing it for the Yankees or Indians.

On that basis, he ranks somewhere between the 4th and 10th best pitcher of All-Time and will lead my ballot.

andrew,

With all due respect, I don't get what you are saying at all. I mean, Satch may have been the 4th to 10th best pitcher of all-time, but...

You say this isn't the Hall of What Might Have Been and then you go on to say that what might have been is that Paige might have been better under different circumstances.

Then you say that "when sitting still, etc." Paige was better than when jumping around. Maybe you have some info that I don't have, but certainly the data in post #2 above shows no such thing. His best years appear to have been with the Crawfords, (maybe he should wear a Crawfords hat ) but here is his record with the one period of 3 consecutive seasons that he spent with Crawfords:

21-9
7-9
20-5

He also had 3 straight years in Birmingham in 1927-29:

8-3
13-4
11-11

Then he was "just" 32 when he went to KC and apparently had 4 straight years before bouncing back and forth a bit 1943-48, and didn't do much in KC. Does anybody know why he was on the "B" team in 1940? Maybe that doesn't mean the 2nd string, maybe it just means they had 2 teams out on the road?

And then you mention that what might have been is he might have pitched at a very high level for 15 to 18 years. Nothing I said above is meant to imply that he didn't pitch at a high level for 15 years, anyway. I still don't see him as better than Ray Brown for their NeL years, though he (or Brown) may still be the #2 NeL pitcher of all.

And of course it is impressive that on top of a pretty stellar NeL record he pitched well in the MLs (well) after his 40th birthday. And of course he will have an elect-me spot on my ballot. And I do believe you rate people based on their good to great years, not their bad years.

I'm just not sure yet whether he goes above or below Mize.

Sunnyday:

We don't really disagree signficiantly.

My real point was simply to suggest that, while Negro League stats translated into MLE's are a crucial tool for evaluating excluded players, they are just a tool.

It seems to me that such stats will by definition underrate guys like Satchel who skipped teams, sat out seasons, goofed off, showed off, put more effort into exhibitions than league games under certain circumstances, etc.

I don't hold it against Satchel that his career was a weird patchwork b/c/ I have a hard time believing that he would have done anything other than taken a spot in a major league rotation and stayed there for 15-18 years if he had been allowed to pitch in the white majors. So, I take an estimation of his ability (developed through the stats and all the other more subjective evidence) and slot him in based on that, not worrying so much how many decisions he had with the Monarchs or the Crawfords in any given year. Does that get me kicked out of the sabermetric union?

andrew, I agreewith your 1st, 2nd and 4th paragraph. As to your third, I give you John Beckwith.

Does anybody know why he was on the "B" team in 1940? Maybe that doesn't mean the 2nd string, maybe it just means they had 2 teams out on the road?

This is one of the central stories in Paige's "legend"--he hurt his arm badly in Mexico in 1938, and was unable to pitch at all at the beginning of 1939. So he was signed by the Monarchs to lend his name to (and play first base for) their "B" travelling team, a sort of minor-league barnstorming outfit managed at that point by former Monarchs' third baseman Newt Joseph.

In the course of the 1939 and 1940 seasons he started pitching again, still for the Monarchs' B team, developing a more diverse collection of pitches, including his first serious attempt at a curve ball. In winter 1939 he played in Puerto Rico and went 19-3. He got into a couple of Monarchs' games in 1940, and finally made it back to the black big leagues in 1941.

A question about walk rates:

In Gary's post #5 we have a very good BB/9 (well below league average, among the league leaders) for one of Paige's early year, when he was very much a power pitcher.

His major league years came a decade after his major injury, when he, as Gary said, featured "a more diverse collection of pitches." In 48-49, league walk rates were insanely high (it was the Age of the Eddies) and Paige's rates were well below league average. When he returned in '51, league walk rates were lower and his was higher - so he was worse than league rate in '51, about even in '52, and better in ''53.

At the peak of his career, in the early '30's, what was his control like?

Paige was, of course, famous for his control. The NBJHBA lists him as "best control pitcher" for the Negro Leagues.

As Gary observed in # 19, Paige was really two different pitchers, before and after his injury -- or maybe even three. Neyer and James splits his career into 3 periods:

1. 1926-38, when he relied mostly on his fastball, with an occasional curve. Neyer and James write, "Paige was famous for three things: his amazing control, his fastball, and his near near-complete lack of any pitch aside from the fastball. He threw his curve very rarely..."

2. 1939-49, when his pitches were "1. sneaky fastball 2. slow curve 3. change" Neyer and James add, "though he eventually recovered [from his injury] his fastball was never again what it had been. He began to throw other pitches, and he also mixed his deliveries, throwing submarine, sidearm, and overhand." Neyer and James don't mention his "hesitation pitch" -- some type of delayed delivery that Paige had apparently used effectively in the Negro Leagues. However, when he used it in the majors, the league president (Will Harridge) declared it illegal, in a ruling that was controversial.

1951-65: pitch selection is described as "everything, including the kitchen sink." Neyer and James add, "By the mid 1950s, Paige was throwing virtually every pitch from virtually every angle. He still threw both of his fastballs, his curve, and his change of pace ('nothing ball'), but he also fooled around with a screwball, a sidearm/submarine fastball ('whipsy-dipsy-do'), a slider, a kuckleball that he called his 'bat-dodger', a blooper, and anything else he happened to think of on a given day."

The following is Paige's Negro League career according to Holway's book *Josh and Satch.* It's clearly based on a less complete set of data than the *complete Book*, but I thought the details about walks, strikeouts, complete games, TRA, and so on would help to fill in gaps. (Note the small differences between this and my compilation of the 1928 season above.)


Yr W-L TRA G GS CG IP H BB SO
27 8-3 3.27 20 9 6 93 63 19 80
28 12-4 3.07 26 16 10 120 107 19 112
29 11-11 5.28 31 20 15 196 191 39 184
30 11-4 2.84 18 13 12 120 92 15 86
31 5-5 4.47 12 6 5 60 36 4 23
32 14-8 3.79 29 23 19 181 92 13 109
33 5-7 3.68 13 12 10 95 39 10 57
34 13-3 1.99 20 17 15 154 85 21 97
35 0-0 0.00 2 2 0 7 0 0 10
36 7-2 2.70 9 9 9 70 54 11 59
37 1-2 3.46 3 3 2 26 22 6 11
40 1-1 4.50 2 2 2 12 10 ? 15
41 7-1 2.21 13 11 3 67 38 6 61
42 8-5 2.88 20 18 6 100 68 12 78
43 5-9 4.26 24 20 4 88 80 16 54
44 5-5 1.28 13 ?? ? 78 47 8 70
45 3-5 ??? 13 7 1 38 22 2 23
46 5-1 ??? 9 9 1 68 65 12 48
47 1-1 1.64 2 2 2 11 5 ? ??
50 1-2 5.88 8 ? ? 26 28 ? ??

Also, from *The Negro Leagues Book*, Paige's record at Miami of the International League in 1956-58 (ages 49-51):

Yr W-L ERA G GS CG IP H BB SO
56 11-4 1.86 27 ?? ?? 111 101 28 79
57 10-8 2.42 40 ?? ?? 119 98 11 76
58 10-10 2.95 28 ?? ?? 110 94 15 40


Gary - is "TRA" restricted in any way to just the innings pitched by the pitcher in question, or is it just all runs scored in those games, no matter who pitched what?

I'm looking at that 1932 line, and what strikes me is that a WHIP of 0.58 just doesn't match a RA of 3.79.

He's given for that year as 29 G, 23 GS, 19 CG, 181 IP. But 19 complete games (if 9 innings) are already 171 IP, leaving only 10 innings to account for 4 starts and 6 relief appearences.

Note also that the all-time major-league single-season record for WHIP is 0.74. In fact, those Holway numbers include a four-year stretch, 490 IP, with an overall WHIP of 0.61, ever year of it better than that major league record. And yet the TRA's for those years are reported as 4.47, 3.79, 3.68, 1.99.

Can you explain? And how much should we trust these numbers or are some of them too good to be true?

For the one year you provide your own data, you show slightly more innings in slightly fewer games, and a slightly lower TRA than Holway. If the differences lie in the same direction in the 1931-1934 stretch, that would help, but the differences would have to larger to allay my worries about whether the numbers make sense.

Looking at Paige's hits allowed, you're right about them, of course. I didn't notice because I was merely transcribing Holway's numbers, not really thinking about them.

The problem may be a particularly stupid custom of some of the early Negro League statistical compilations--that of just adding up whatever information you can find, without regard to the relationship between statistical categories. That is, you might know that a pitcher threw a complete game victory and gave up four runs, but you may not know how many hits he gave up or how many strikeouts he had. So you add in the win and the innings, but nothing for strikeouts and hits. If this happens several times in a season, your strikeouts and hits/i.p. get thrown off pretty seriously.

Looking at the *Josh and Satch* book, it turns out Holway gives a list of Paige's appearances in 1934. Now, of course the information there doesn't add up to the statistical line he's got in the back of the book for 1934 (what else is new?), and in light of later research I'm pretty sure it's highly incomplete. But aside from that--he's got three games for which he gives innings pitched, but nothing, a blank, for hits allowed. For the 15 games that definitely have the hits allowed, Holway has 76 hits in 124 innings (including Paige's July 4 no-hitter over the Grays). (He also lists one game as a loss, but with no innings or other data--if he does this very often, that might explain some of the strange innings-games ratios.)

Of course, the same goes for walks and strikeouts. He lists w's and k's for only 10 games, with 20 walks and 95 strikeouts in 87 innings. Since box scores that list walks will usually also list strikeouts, you can assume that the K/W ratios are about right.

This would give for Paige's 1934 season the following rate stats, which to me seem much more reasonable:
H/9 inn. 5.52
W/9 2.07
K/9 9.83
WHIP 0.84

Paige really did have a remarkable season that year. I actually have worked on 1934, but lost all of it to a hard drive meltdown some years ago. I still have the box scores, and could dig them out to check on this one year, I suppose.

Holway's 1928 matches my compilation so closely because NNL box scores for the 1920s are generally very good--there's little room for ambiguity. But in the 1930s (particularly in the east), things start to get sketchier.

Btw, Paige's *real* strikeout rates are extraordinary for the Negro Leagues, the best of any pitcher in the post-1920 era (and probably pre-1920 as well). In terms of that, and also K/W ratio, he leaves Rogan, Foster, and Brown in the dust.

Thinking about an analogy for the young Satchel Paige, the young Tom Seaver comes to mind. Both relied mostly on fastballs, but threw a couple varieties of them to keep batters guessing. They threw a lot of strikeouts and didn't give up a lot of walks. Seaver, of course, pitched during a period when strikeouts were much more common, so Paige made a much bigger impression.

PITCHERS OVER 40:
RSAA                           RSAA      IP       N_W      N_L    
1    Hoyt Wilhelm                100    744         62       26   
2    Jack Quinn                   95   1205.1       87       64   
3    Cy Young                     44    883.2       60       39   
4    Nolan Ryan                   43   1060         61       52   
5    Satchel Paige                40    476         35       24   
6    Ted Lyons                    35    223         18        7   
7    Roger Clemens                32    214.1       15        7   
8    Red Faber                    29    545         36       29   
9    Dutch Leonard                28    286         19       12   
T10  Doug Jones                   18    262.2       14       12   
T10  Connie Marrero               18    396         23       20   

If I had to condense what I knew about how good a pitcher Satchel Paige was into a few bites, it might be:

NeL data: 23.7 WAT from 1927-1936, probably faced key opponents pitchers often
Did likely not have the peak or prime of a few other NeL pitchers
MLB data: 124 ERA+ from age 41 on
Career length, approx 30 years!
Great fastball, superb control, very good at changing speeds
Smart and determined. Came up huge in some big games

--
capsule: overrated on rep, maybe like Christy Mathewson. And like the Big Six, still one of the top 12 pitching careers ever.

I've gone through some old threads to gather up and re-post a few items that seem particularly relevant to evaluating Paige:

Chris Cobb posted the following item as # 17 on the Reevaluating Negro League Pitchers thread:

Posted by Chris Cobb on May 16, 2005 at 11:15 AM (#1340164)
Here's another data point: Black and Gray Ink Totals, derived from Holway's League Leaders lists

Caveats: Completeness of lists varies by era -- sometimes top 1-2 listed, usually top 5, sometimes K leaders or TRA leaders not included at all. Data selected by Holway favors certain types of pitchers

Calculation Method:
Pitchers 4 pts for wins, TRA/ERA, K. 3 pts for winning%.

Pitcher Totals – Black Ink top 20

1. Satchel Paige 50
2. Bill Foster 49
3. Dick Redding 44
4. Joe Williams 35
4. Ray Brown 35
6. Hilton Smith 30
7. Bill Byrd 27
8. Joe Rogan 23
8. Nip Winters 23
10. Gentry Jessup 20
…
Pitcher Totals Gray Ink Top 20

1. Joe Williams 130
2. Ray Brown 122
3. Satchel Paige 114
4. Bill Byrd 113
5. Joe Rogan 102
6. Bill Foster 97
7. Hilton Smith 84
8. Dick Redding 78
9. Leon Day 72
10. Ted Trent 68
…


Gary A posted the following item as # 18 on the Chet Brewer thread:

Posted by Gary A on May 18, 2005 at 07:49 PM (#1346974)
In the California Winter League, Brewer's record over 13 seasons was 43-13.

Comparing some of the top NeL pitchers in this league:

Chet Brewer 43-13
Andy Cooper 22-6
Reuben Currie 26-19
Bill Foster 24-1 (!)
Robert Griffith 20-2
Satchel Paige 56-7
Andrew Porter 23-6
Bullet Rogan 42-14
Jim Willis 41-10

From 1931 to 1936, Paige was 50-2 (!!); his record is brought down by a 6-5 record from 1943-47.

Not all pitchers have complete strikeout and innings pitched records; of these, the only ones whose records are complete are:

Currie, 162 K, 94 W, 402 IP
Griffith, 228 K, 29+ W, 214 IP
Paige, 770 K, 138+ W (2 seasons with 121 IP are missing walks), 572 IP
Rogan, 351 K, 189 W, 516 IP


Dr. Chaleeko posted the following item as # 58 in the Reevaluating Negro League Pitchers thread:

Posted by Dr. Chaleeko on May 20, 2005 at 04:55 PM (#1350600)
These leader boards are a bit preliminary because I've still have got a few identity questions to resolve and some data-neatening to do (they look pretty rough, I know), but I think I'm close enough along that it won't make a tremendous difference. Also, I don't know if I'll finish it by Monday, so I figured better to post now. If/when any of these change, I'll repost.

Remember, everything is based on Holway's data except where Gary has kindly supplied new data. That means only U.S. league and regional data is supplied for the years 1920-1948. (I may return to the pre-league years in the near future, no promises). This also includes data for the 1931 Negro Southern League because a) Holway supplied complete W-L data and b) a lot of top players were in the league that year. Holway's published career data often conflicts with itself (as Chris and Gary have recently pointed out), and I'm basing % of team decisions strictly on the sum of individual pitchers' wins and losses.

Top 25ish totals are shown for each leaderboard.

PRELIMINARY CAREER LEADER-BOARD LISTS

WINS
brown, ray …………………158
paige, satchel ………154
bell, william …………145
foster, bill ……………144
rogan, bullet …………144
byrd, bill …………………133
cooper, andy ……………130
winters, nip ……………127
holland, bill …………118
mcdonald, webster 117
…
LOSSES
holland, bill ……………105
mcdonald, webster …96
paige, satchel …………93
byrd, bill ……………………84
yokely, laymon …………80
…
DECISIONS
paige, satchel ………247
holland, bill …………223
byrd, bill …………………217
brown, ray …………………214
foster, bill ……………214
…
WAT
brown, ray …………………32.75639
cooper, andy ……………21.75161
salmon, harry …………21.63823
byrd, bill …………………18.98093
foster, bill ……………18.9751
winters, nip ……………18.1168
paige, satchel ………17.75274
mchenry, henry ………16.07841
rogan, bullet …………13.7111
rector, connie ………13.63975
…


Gadfly posted the following item as # 68 on the Reevaluating Negro League Pitchers thread:

Posted by Gadfly on May 22, 2005 at 06:31 PM (#1355004)
I second that Doc, this is great stuff and good work.

One thing to remember when comparing Negro League pitchers is that star Negro League pitchers very often were 'featured' against one another. In other words, their teams would publicize match-ups between the best pitchers to build the gate.

The probably best known example of this is when they actually stopped the Championship Series one year just to promote a Satchel Paige-Slim Jones match-up.

The obvious point of this being that, if a star pitcher pitches mostly against other star pitchers, his winning percentage (WP) will not truly represent his talent.

[Of course, in the Majors, star pitcher match-ups, because of pitching rotations and a much lesser need for publicity, are much rarer. However, a good example is the famous Walter Johnson-Joe Wood match-up in 1912.]

When I first started studying the Negro Leagues, I was surprised to find that Satchel Paige's WP did not seem all that great. But then I slowly came to realize that Paige was pretty much just pitching against the other team's aces.

Basically, Paige was pitching .650 ball against .650 pitchers from 1932 on.

There is a book about the Homestead Grays (In the shadow of the Senators) which details how Paige pitched again and again against the Homestead Grays one year. He pitched fantastically, drew enormous crowds, but couldn't buy a victory.

Of course, the Grays were the Negro League's answer to that generation's New York Yankees.

This 'star pitcher' effect obviously hurts Paige more than anyone, but is also noticable in quite a number of other pitcher's careers (Slim Jones, Willie Foster, Rogan, Joe Williams, Mendez, etc.)

Of course, pre-1920 star caliber pitchers have an additional type of 'star pitcher' problem. Basically their teams were only playing the cream of the other Negro League teams.

Mendez is a good example of this. In 1911, Mendez pitched 38 games total for the barnstorming Cuban Stars. He won 33, lost 3, and 2 games ended tied. He threw a total of 14 shutouts.

Against colored teams, Mendez pitched 14 games, winning 12 and losing just 1 (His no decision was the famous 12 inning 2-2 tie versus Rube Foster). Mendez shut out 3 Negro teams.

In those 14 games, Mendez piched one game against the Cuban Giants, one game against the Philadelphia Giants, and one game against the All Cubans.

The other 11 games were all against the absolute cream of the Negro teams (American Giants 5, Brooklyn Royal Giants 4, Lincoln Giants 2). With the quality of his competition taken into consideration, Mendez' record of 12-1 actually understates how good he was (which seems, now that I'm writing it, kind of unbelievable).

Of course, this simply makes deducing greatness even harder; but it is an effect to be aware of.

Minor note on Paige:

In 1938, Paige hurt his arm in Venezuela in late July or early August (before the injury, Paige was pitching very well though in bad luck because his team could not hit at all). He went to Mexico in August and continued to pitch with his arm sore, generally ineffectively but with one memorable game against Dihigo.

In 1939 & 1940, Paige pitched for and was the star attraction for the 'Satchel Paige All-Stars' which was described above as the Monarch's "B" team and makes it sound like Paige was not good enough for the Monarchs themselves.

The truth is a little different. When Paige returned to the States in 1939, the NNL Newark Eagles held his reserve rights under the NNL-NAL agreement, having reportedly purchased them from the Crawfords.

Because Paige's arm was injured, the Eagles did not bother sending Paige a contract for 1939. J.L. Wilkinson, the Monarchs' owner, signed the injured Paige to formed the Satchel Paige All-Stars to showcase Paige and capitalize on his name.

Also, Wilkinson, who was one of the smartest (and nicest) men in the Negro Leagues, assigned the Monarchs long-time trainer (Frank Floyd) to travel with the All-Stars and work on Paige's arm. Paige, to his credit, would not forget Wilkinson's kindness and stayed loyal, or as loyal as Paige could be, to the Monarch's owner.

Paige's arm came back in early 1939 and, traveling with his All-Stars, he once again became the biggest thing behind the color line. However, he could not play for the Monarchs main team without violating the Negro League Agreement.

So Wilkinson and Abe Saperstein, the booking agent for the Negro American League (NAL) booked Paige's team all over the country, playing both NAL and NNL teams plus the strong independent teams (the Ethiopian Clowns, House of David, Bushwicks, etc.) across the country.

And they made a truck load of money doing it.

The Newark Eagles' owners, Abe and Effa Manley, went ballistic at this, claiming that Paige (and the huge gates he was drawing) belonged to them and Wilkinson and Saperstein were violating the Agreement between the two Negro Leagues.

[The situation became borderline bizarre when the always randy Paige evidently proposed that he would report to the Eagles if the attractive Effa Manly would become his girlfriend.]

This situation eventually got so bad that the two Negro Leagues got to the brink of going to war with one another. However, in the end, the Eagles received some players for Paige from the NAL (one of whom was Bus Clarkson) in late 1940 and it was settled peacefully.

[Also, though it was not stated, it is pretty obvious that Wilkinson agreed to feature Paige against the NNL teams often.]

Basically, this was the biggest story in the Negro Leagues for most of 1939 and 1940. Paige was, as he had been from about 1934, the most valuable and important player (i.e. money) in the Negro Leagues.

In late 1940 and through 1941, with the dispute over, Satchel Paige began pitching more for the Monarchs. However Paige was always, from 1941 on, used as a 'feature pitcher' to draw a gate. In other words, any team could hire him to goose the gate.

And Wilkinson eventually reformed the second team (he had broken it up in 1940 when the dispute was settled), again named it the Satchel Paige All-Stars, and toured it around, making bundles of money (especially from 1946 to 1948).

[Wilkinson, a smart man, realized that Paige was far more valuable playing for other teams than the Monarchs. The Monarchs, with or without Paige, drew very well. But Paige, just by his name, could draw huge gates to teams that had no chance of drawing like the Monarchs.]

Basically, from 1941 to 1948 when he went to the Majors, Paige would play for 1) the Monarchs, especially when they had a big series; 2) the Satchel Paige All-Stars, anywhere and any place a buck could be made; and 3) any team willing to pay Wilkinson and Paige enough to feature him (including many NAL and NNL teams like the piss poor New York Black Yankees).

If I get the chance, I'll post Paige's published 1944 and 1945 NAL statistics. Paige did not play much for the Monarchs in either season for the above reasons, but was clearly still the best pitcher in the NAL by his stats.

Thanks Gadfly! I had just written up and was going to post my take on Paige's career, but I see I'll have to go back and revise the section on 1939-46.

Since it’s unclear whether our (overworked) Chris Cobb will be able to compile MLEs for Paige, I thought I’d try to systematically walk through Paige’s qualifications using less precise “back of an envelope” methods. Maybe without Chris we can’t determine exactly what his DERA+ or win shares would have been, but I think we should be able to figure out roughly where he stacks up against some of the other HoM pitchers, which should give us a pretty good idea of where his career should be ranked.

I’ll compare him just to our live-ball era HoM pitchers (thereby avoiding numerous difficulties that would arise in comparing pitchers from different periods). That’s a group of nine pitchers: Brown, Coveleski, Faber, Willie Foster, Grove, Hubbell, Lyons, Rogan, and Vance. (I suppose we could also include Dihigo in the comparison set, but I won’t.)

Let’s set Grove aside for the moment and split the other eight into two groups: peak/prime pitchers, who were elected primarily based on 9 to 11 years of strong performance, and career pitchers, for whom career length and sustained good quality were keys to their election. I’ll classify Faber and Lyons as career pitchers and the other six as peak/prime pitchers (though I recognize that Hubbell and Brown could fit in either category, depending on where one draws the line).

For the six peak/prime pitchers, I’ll compare their performance during their best 9-11 seasons to Paige’s performance during the first 11 years of his career (before his arm injury), 1927-37.

Let’s start with a comparison to the three Negro League pitchers—Brown, Willie Foster, and Rogan, since their NeL statistics can be directly compared to those of Paige. I will base my comparisons on the pitchers’ Negro League W-L records and wins above team (WAT). Without systematic or reliable ERA data, these are the most accurate and comprehensive information available. (Although I know Chris’s MLEs incorporate other factors such as defensive support and run support, my understanding is that they are largely driven by W-L records relative to team.)

Of course we should always keep in mind the usual caveats associated with W-L and WAT:
- They are affected by differences between pitchers in run support, quality of opponents, and various other factors discussed on Chris J’s website.
- They can be quite unreliable, especially for single seasons. There are many example of good pitchers with poor W-L records (or mediocre pitchers with good W-L records). Examples are Mel Harder’s 1933 season (15-17, -1.1 WAT, 158 ERA+), and George Uhle’s 1923 season (26-16, +4.8 WAT, 105 ERA+).
- Although W-L and WAT are more reliable over longer periods (5-10 years), the effects of extraneous factors do not always go away even over a full career, as Chris J has demonstrated.
- Because WAT is centered at zero and can be negative, career WAT statistics can be tricky to interpret. Several seasons of below-average performance can bring down a pitcher’s career WAT, even though those seasons may still have had value.

Based on their actual (Holway) NeL W-L records relative to team (usually listed by Chris Cobb near the top of each pitcher’s thread), I see the peak/prime periods for the four NeL pitchers lining up in the following order:

Name.. Years.. W.. L. .Pct TPct WAT
Brown. 1934-44 125 39 .762 .547 36.0
Paige. 1927-36 107 53 .669 .521 23.7
Foster 1925-34 132 60 .688 .600 16.9
Rogan. 1920-28 143 57 .715 .647 13.6*

* Rogan’s record reflects a correction by Gary A to the Holway statistics for 1923; see # 38 on the Reevaluating Negro League Pitchers thread.
Note - TPct is what the pitcher’s winning percentage would have been if his percentage for each season had been the same as the average of the other pitchers on his team.

So, based solely on W-L and WAT during their peak/prime, Paige appears to be significantly behind Brown and somewhat ahead of Foster and Rogan. These rankings, taken together with the MLEs that Chris has derived for Brown and Foster, then allow us to compare Paige to the MLB peak/prime pitchers, since Foster’s MLEs were very comparable to Coveleski and Vance, and Brown’s MLEs were close to Hubbell’s. So based just on their peak/prime years, I see these pitchers lining up in the following order:
1. Hubbell
2. Brown
3. Paige
4. Vance
5. Foster
6. Coveleski
7. Rogan

Of course all of these pitchers also had value outside of their peak/prime seasons, but Paige certainly has many more meritorious seasons outside his 11-year peak than any of others. While this analysis suggests that based on a peak/prime ranking, Paige is certainly HoM worthy, his ranking does seem to lie quite a bit below the myth. However, there are two factors suggesting that this analysis based on W-L records may understate Paige’s peak:

1. This analysis is based only on NeL play, so it misses two of Paige’s seasons that contributed mightily to the legend—his 1935 season in North Dakota (when he dominated the first national semipro championship tournament in Wichita), and his 1937 season in the Dominican Republic (where he won the critical final game and went 8-2 record against stiff competition consisting of top NeL and Latin American players).

2. As has already been discussed on this thread, the other elements of Paige’s record—his high strikeout rates, good K/W ratios, and good WHIP statistics—suggest he may have actually been better than his W-L record. Gadfly has suggested a possible explanation—that Paige’s opposing pitchers were usually the other elite pitchers in the league, which would have greatly hurt Paige’s run support.

Turning to career value, it seems the task is to come up with (a) a rough estimate of Paige’s MLE career innings pitched and (b) a rough assessment of the quality of his pitching outside of his peak/prime.

For innings pitched, I’ll give two scenarios. Based on his rank relative to team in number of decisions (and new information from Gadfly on 1938-40), a conservative set of assumptions is the following:
1927-31 – 190 IP/yr
1932-37 – 240 IP/yr
1938-39 – 100 IP/yr (injury in 1938)
1940-44 – 220 IP/yr
1945-46 – 160 IP/yr
1947 NeL – 70 IP
1947-53 MLB – 476 IP
Total – 4556 IP

An less conservative set of assumptions would add 20-40 innings per year during his first 11 seasons, giving him about 300 additional innings. One could also add 110 innings per year for 1950 and 1954-58, when Paige was either barnstorming or pitching (extremely well) in Triple AAA. These assumptions would give him a total of well over 5000 innings. (Of course, counting exhibition games and winter leagues, Paige probably actually pitched well over twice that many innings during his 40-year career.)

The quality of his work outside his peak? 1940-46 is a mixed bag; some seasons (1940-41, 1946) appear to have been excellent, a couple of others (1943, 1945) appear to have been pretty bad. The other two seasons are 1942, when his WAT is negative despite an excellent TRA, and 1944, when his W-L record is slightly above average. And of course Gadfly has just informed us that Paige actually had a good year in 1945 if his performance for other teams is counted.

Of course for 1947-53 we have his actual MLB data (ERA+ of 164, 131, 92, 128, 119). Adding in the 1940-46 NeL period, I’d characterize his work from 1940-53 as consisting of the following: two or three great seasons (ERA+ above 140), possibly two or three poor seasons (ERA+ below 95), and the rest pretty good (ERA+ between 95-140). Of course, thanks to Gadfly, there are now question marks about those “poor” seasons; moreover, most pitchers with long careers have at least a couple of poor seasons.

Based on career length, peak, and quality of non-peak seasons, I think Paige’s career value clearly belongs ahead of Lyons and Faber. Based on his peak value and what he did outside his peak years, I think Paige belongs ahead of all the peak/prime candidates (with the possible exception that an extreme peak-oriented voter might place him behind Hubbell).

That leaves Lefty Grove. Based on the available data, I can’t see Paige going ahead of Grove. Grove’s peak appears to have been much higher and his career (especially if you include minor league credit) was almost as long. I suppose an extreme career-oriented voter who gave Paige credit for 5,000+ innings and adjusted his peak for high-quality opposing pitchers might have Paige ahead, but I think Grove pretty clearly remains the best live-ball pitcher to date.

So I have Paige ahead of all the other live-ball HoM pitchers and as one of the top 50 players in baseball history. That may be a step down from his reputation and from the Bill James ranking (# 17 player and # 2 pitcher of all time), but it still will be good for the # 1 spot on my ballot this year, ahead of Johnny Mize.

Thanks, Brent, for that analysis of Paige. As it happens, I have been working on projections for him, which I now post.

It will be interesting to compare this analysis to Brent's. I think our conclusions, from what I can see from skimming his post, are pretty similar.

Satchel Paige MLEs

All these are based on Holway data posted above except for

1935 Bismarck. Paige’s recorded record for that season is 30-2, 330.7 IP, team appx. 85-22, .794 . Conversion factor applied: .55 (semi-pro norm)
1937 S.D. 8-2, team 18-13, as reported by gadfly. Conversion factor: .85 (NeL norm)
1947 average of 1946 and 1948. Available data too small for meaningful projection.
1948-53 Major League Data

For all NeL seasons a .85 conversion factor was used, except for 1933 and 1942, in which the conversion was based on TRA. See notes below for explanation of both projections.

Prime Years
Year  Team   IP    snW   snL DERA  DERA+  CWS
1927  Bir    130   9.6   5.7 3.46  130
1928  Bir    232  19.9   7.4 2.74  164
1929  Bir    277  16.8  15.8 4.37  103
1930  Bir    203  15.3   8.6 3.36  134
1931  Cl/Pgh 240  14.1  14.1 4.50  100
1932  Pgh    273  18.8  13.3 3.78  119
1933  Pgh    185  12.1   9.7 4.03  112
1934  Pgh    277  20.2  12.4 3.52  128
1935  Bis    262  20.1  10.7 3.28  137
1936  Pgh    310  24.3  12.2 3.19  141
1937  S.D.   289  23.5  10.4 3.00  150
11yrs*      2678 193.1 121.9 3.56  126    238**

*Note:  Career totals for wins and losses do not match seasonal totals exactly, because of distortions in the pythagorean projections in extreme seasons. 

**See notes for explanation of method used to calculate career win shares.   

1938-39 Injury years; pitching in “minors

Kansas City Years
Year  Team    IP   snW   snL DERA  DERA+  CWS
1940  KC      77   6.4   2.7 2.94  153
1941  KC     223  18.4   7.7 2.94  153
1942  KC     281  15.5  17.6 4.66   94
1943  KC#    293  11.1  23.4 6.52   69
1944  KC     220  11.3  14.6 5.11   88
1945  KC     206   6.9  17.3 7.11   63
1946  KC     117   8.6   5.2 3.46  130
1947  KC      95   6.5   4.7 3.80  118
8 yrs.      1512  80.9  97.0 4.93   91     57 

#See notes below on 1943-1945 projections

Majors Years
Year  Team    IP  W      L  DERA  DERA+   CWS
1948  Cle   72.7  6      1  3.24  139
1949  Cle   83.0  4      7  3.48  129
1950 No Data
1951  Stl   62.0  3      4  4.71   96
1952  Stl  138.0 12     10  3.37  134
1953  Stl  117.3  3      9  3.31  136
5 yrs.     476.0 28     31  3.52  128     42

Total
4666 ip, 302-250, 4.00 DERA, 113 DERA+,  336 win shares

Prime Comparisons
Paige, 1927-37
2678 ip 193.1-121.9 snw-L 3.56 DERA 126 DERA+ 238 ws

Carl Hubbell, 1928-38
2833.7 ip, 205-112 w-l, 3.52 DERA 128+, 259 ws

Wes Ferrell, career
2623 ip, 193-128 w-l, 3.90 DERA 115 DERA+, 233 ws

Career Comparisons

Paige, 1927-37, 40-49, 51-53
4666 ip, 302-250, 4.00 DERA, 113 DERA+, 336 ws

Lyons 1923-42, 46
4161 ip, 260-230, 3.93 DERA, 115 DERA+, 312 ws

Ruffing, 1924-42, 45-47
4344 ip, 273-225, 4.16 DERA, 108 DERA+, 322 ws

Rixey, 1913-17, 1919-33
4494.7 ip, 266-251, 4.06 DERA, 111 DERA+, 315 ws 

Notes.

I. There are three ways in which these projections probably underestimate Paige’s value.

1) It makes no account of differential usage patterns. Gary A. has shown that in 1928 Paige faced a disproportionate number of good teams, and gadfly has noted that the NeL tended to feature ace vs. ace matchups regularly, meaning that Paige typically faced the best pitcher on the opposing team.

2) It does not compensate for the rise in competition levels in the early 1930s. Paige’s career shows the typical early 1930s trough that we’ve observed on all the NeL players. This hurts him more than any other pitcher I’ve studied because these seasons came in the heart of his peak. During these years of stiffest competition, he was clearly the best pitcher in the league, overall, but these show up as among his weaker seasons.

1933 provides an excellent illustration

The five best pitchers in the league that year, by Holway’s TRA, are

Ray Brown 2.87 TRA
Sam Streeter 3.04 TRA
Bill Foster 3.08 TRA
Leroy Matlock 3.23 TRA
Satchel Paige 3.50 TRA

That’s a pretty good set of pitchers.

Here are their win-based MLE DERA+ projections, as I have them

Ray Brown 90
Sam Streeter (no data)
Bill Foster 89
Leroy Matlock 121
Satchel Paige 67

Obviously these projections, with the possible exception of Matlock’s, are totally bogus. These were great pitchers, as the rest of their careers show, and as their TRA data for this season shows also. But that excellence didn’t translate into wins at anything like the rate we would expect for any of these pitchers except Matlock. I think we are seeing the results of a combination of very high competition levels and a good deal of ace match-ups.

You may notice that Paige’s DERA for 1933 is not listed as 67. Because this number was obviously inconsistent with his actual performance, I calculated his DERA differently for this season, pegging it to Leroy Matlock’s, based on the ratio between their recorded TRAs for the season.

I made a similar adjustment to 1942, where Paige was among the league leaders in TRA but several wins below his team. I calculated average MLE DERA for the three top starters on the team (all in the top 5 in TRA in the league), and then found individual DERAs based on the ratios of the three pitchers’ TRAs.

3) The estimates straightforwardly project both Paige’s usage and his DERA for the WWII years. For these seasons, Paige is consistently several wins below an undistinguished team, but he always leads his team in decisions by a large margin. Perhaps ace vs. ace matchups account for some of this strange combination, but it may be also that Paige was being significantly overworked in these years, because he was such a tremendous drawing card. If he had been used in a way that maximized his effectiveness, I think his DERA during these years would have been consistent with his career norms, since it returned to those levels once he began to be used more sparingly in 1946 and in his ML career. As a result, Paige is projected as throwing a huge number of bad innings 1943-45, which hurts his career DERA tremendously, though it pads his innings totals.

II. On win shares

I have not calculated seasonal win share totals for Paige. The results swing so widely without regression that I thought it would probably be unhelpful to have them.

The career win-share totals are generated by multiplying Paige’s innings pitched by .058 to get the value of an average pitcher, according to the win shares system. Then 3 win shares are added or subtracted for each win Paige is above an average pitcher during that stretch of his career. Then that total is multiplied by .9 to scale the result to Bill James win shares: when this method is applied to pitchers with major-league records, it generally produces a result that is 10% higher than the total produced by the win shares system.

P.S.

As should be obvious, this analysis was developed without reference to gadfly's notes on Paige's later career. If Paige gets two more good seasons added to his resume for the combination of 1938-1940, that would raise his career IP up towards 5000 and move his career value far ahead of the Lyons, Rixey, Ruffing set.

Having read Gadfly's post and Brent's posts, I certainly concur with Brent's conclusion that Paige ranks ahead of all other pitchers 1920-1950 with the exception of Grove. He might have been better than Grove, or he might not.

Among NeL pitchers, he ranks ahead of all others except Joe Williams. He might have been better than Williams, and he might not.

Anyway, he's an obvious #1 this year in my view. Mize is underrated, but he's not better than Paige.

chris and gadfly,

where can i get actual stats for the ND years for guys like Trouppe? I haven't seen any stats on the site Brent has kindly pointed out, but I may not have spotted them.

where can i get actual stats for the ND years for guys like Trouppe? I haven't seen any stats on the site Brent has kindly pointed out, but I may not have spotted them.

I dug the stats for Paige's season out of the prose account of the 1935 season. Numbers get mentioned there, but not systematically, so I dont' know if it has anything on Trouppe. That site must have a source for its data, but I don't know what it is. Gadfly, when he checks in, will probably be much more help!

Doc-
There are some team stats for Bismarck in the Double Duty biography by Kyle McNary. If I remember correctly they include Trouppe.

Chris, thank you very much for the analysis.

Chris Cobb wrote:

I have not calculated seasonal win share totals for Paige. The results swing so widely without regression that I thought it would probably be unhelpful to have them.

As they say, fools rush in where angels dare to tread. To try to get some idea of what seasonal win shares with regression might look like, I first calculated WS according to the formula that Chris used for the career totals, then applied a moving average to dampen out the swings. I also constrained them so that the totals for the two intervals matched the estimates that Chris has provided. These are not intended to serve as serious estimates for individual seasons--rather they are just intended to give us a rough impression of the contours of Paige's career.

Year WS
1927 12
1928 24
1929 20
1930 18
1931 15
1932 21
1933 14
1934 25
1935 25
1936 32
1937 32
1940. 7
1941 17
1942 12
1943. 3
1944. 5
1945. 2
1946. 6
1947. 5


I know I'm talkng to the wrong group of folks here, but does anyone think that Satchel Paige's career WS would have looked anything like the pattern calculated by Brent (quite accurately from the available statistical record) if he had been white or if the majors had been color-blind. For a guy like Satchel, I just don't really care what his cumulative negro league stats are. Or rather I don't care beyond their anecdotal value in determining his true level of ability (and perhaps his durability). I think every negro leaguer needs a different methodology to best make him whole for the lost opportunities of baseball segregation. For Satch, I say the only fair thing is to figure out how good he was (Bill James says Johnson and Grove, but the spectrum here seems to be from Seaver/Mathewson to maybe Lyons/Ferell), to make adjustments based on anything particularly out of the ordinary in his durability or aging pattern, and then to guesstimate a full major league career on the basis of those considerations. Not exactly hard science, but how else can you be fair to this guy?

I was able to dig up some of my old research on Paige's 1934 season. For games with box scores, I have this:

W-12
L-1
TRA-2.36
G-17
GS-14
CG-13
SHO-4 (including a no-hitter)
IP-126
H-83
R-33
BB-24
K-135
BB/9-1.71
K/9-9.64

Unfortunately, I don't have home runs allowed. This includes the championship game of the Denver Post tournament, when Paige pitched the House of David to a 2-1 victory over Chet Brewer and the Kansas City Monarchs (themselves fortified by Turkey Stearnes, Sam Bankhead, and other stars), striking out 12.

He had games of 17 and 14 strikeouts, and struck out 12 three times. He only once walked as many as 4 in a game, and walked 3 only three times.

Btw, the Holway stats I posted in #22 above were aligned perfectly when first posted (using the "code" tag)--but now they're screwed up.

I know I'm talkng to the wrong group of folks here, but does anyone think that Satchel Paige's career WS would have looked anything like the pattern calculated by Brent (quite accurately from the available statistical record) if he had been white or if the majors had been color-blind.

I think there's a good chance that they would have looked quite a bit like this for much of Paige's career. The statistics do not suggest that he was pitching an outrageous number of innings 1927-1934: he was generally on teams with strong pitching staffs that seemed to spread the work around. Birmingham had Paige, Salmon, Streeter, and Pittsburgh had Paige, Matlock, Hunter, Streeter, and William Bell in various combinations.

There are two big caveats to this.

First, Paige's peak during the early 1930s was probably higher. Brent's win shares are based on MLEs that are based on a conversion factor that is definitely too low for that era. Not being in a position yet to estimate accurately a more correct factor, I think it better to stay consistent, but Paige's 1932-34 ws almost certainly should be higher. He's probably missing 10-15 win shares here.

Second, we have two very different views of Paige's performance in 1939-40 and 1942-45. On 1939-1940, one story suggests that Paige was rehabbing his arm in these seasons and not up to NeL pitching. Gadfly's more informed narrative indicates that it was a combination of contractual issues and financial opportunities that kept Paige out of the NeL during these seasons (we know he was totally dominating in the 1939 PWL, btw). On 1942-45, Holway's stats suggest he was not particularly effective and possibly overworked. Gadfly's narrative suggests that he was still a top-notch pitcher whose Holway stats are misleading because they don't accurately reflect the full scope of his performance or its circumstances--play for the Monarchs in "big series," star appearances for other NeL teams, and barnstorming.

If Gadfly is correct (as is probable, though I'd like to see more supporting stats), then I think Paige probably earned 60 more win shares in these 6 seasons than Brent's numbers, based on my MLEs indicate.

Brent, the saying is "fools rush in where angels FEAR to tread." It's much more duanting that way.

But thanks for rushing in or treading anyway. And thanks to Chris, too.

Brent didn't add up his WS--they total 295. Chris says maybe another 60, that of course equals 355.

Even at 355 I think the notion that Paige's reputation is a bit overblown holds true. Add to that Brent's seasonal peaks of 32-32-25/128. Maybe some of Chris' additional 60 would raise the peak, but it sounds like maybe not much. Those numbers basically get you:

Eddie Plank 360/31-29-29/133 (total differential = 10)
Gaylord Perry 367/39-30-26/134 (24)
Phil Niekro 375/30-28-28/118 (33)
Steve Carlton 367/40-29-26/111 (35)
Fergie Jenkins 323/37-26-26/135 (39)

All of whom are rated between #15 and #34 by James. And knowing that Satch's peak is somewhat depressed by the MLE methodology and perhaps especially in his case, who would doubt that he is better than all these nominal comps? Not I. And none of them is a contemporary, rather all pitched at times when racking up these kinds of numbers appears for whatever reason to have been easier. His closest comps from the lively ball/pre-expansion eta (1920-196) would be:

Robin Roberts 339/35-32-31/153 (50)
Red Ruffing 322/27-25-24/116 (58)
Warren Spahn 411/32-31-28/120 (60)
Eppa Rixey 315/26-26-24/118 (63)
Ted Lyons 311/30-26-23/110 (72)
Early Wynn, Red Faber and Burleigh Grimes (all exactly 77)

IOW, nobody close though if you took a + and - approach Paige is +11 versus Hubbell and +32 versus Feller (unadjusted for WWII) but -95 versus Grove.

Brent didn't add up his WS--they total 295. Chris says maybe another 60, that of course equals 355.

Remember, Brent is only covering Paige's NeL career. To this total, 42 win shares from major-league play should be added, which gets Paige up to 397.

Oops. 397, eh?

And I probably should have boosted his peak with the 10-15 that Chris posits for 1932-34. So now we're talkin'

Paige 397/32-32-30/138

Now his differential from Grove is 49, or if we take into consideration the + and -, it would be -37

As for Eddie Plank, it is 47 and +47
Robin Roberts 77 or +39
Warren Spahn 35 or +7
Tom Seaver 16 or +8

This is of course based on a best-best-best-case scenario. If you're any little bit less enamored than this, post #42 might still be a better comp. Your choice. I'm not sure yet.

Gary A wrote:

I was able to dig up some of my old research on Paige's 1934 season. For games with box scores, I have this:

W-12
L-1
TRA-2.36
G-17
GS-14
CG-13
SHO-4 (including a no-hitter)
IP-126
H-83
R-33
BB-24
K-135
BB/9-1.71
K/9-9.64


These are very impressive statistics - 9.6 K/9 and 5.6 K/BB. By comparison, the 1934 MLB leaders were Paul Dean with 5.8 K/9 and Hubbell with 3.2 K/BB.

Gary, would you be able to identify the opposing pitchers in these games (without going to too much effort)? If you could, it might help us assess the importance of Gadfly's hypothesis that Paige was often or usually matched against other star pitchers.

Brent, do you have a WS breakdown of Paige's time with the Crawfords and Monarchs (an innings breakdown would be good, too)? I'm trying to decide on which cap to give him for his plaque. Thanks!

I don't have a breakdown; Paige moved so much from team to team that it would be hard to figure. His peak performance came mostly with the Crawfords, but he spent more seasons with the Monarchs. Personally, I'd be inclined to give him a KC cap.

Paige spent nine seasons with the Monarchs as opposed to 4.5 with the Crawfords. His IP with the Monarchs were probably not quite twice his IP with the Crawfords, but close.

If Gadfly's view of his performance during the early 1940s is correct, Paige would also have the most win shares with the Monarchs. Brent's WS based on my MLEs have him at 92 win shares with the Crawfords, 57 with the Monarchs.

I estimate above that, if Gadfly is right, the 57 estimate is about 60 short, bringing Paige to 117 win shares in the Monarchs years. I also posit that his peak is 10-15 short, bringing him to 102-107 for the Crawfords years.

I don't know how you judge what cap to give, but all the career measures point to the Monarchs, though the peak points to Pittsburgh.

Paige certainly got on with the Monarchs better than with the Crawfords, given that he left the team late in the 1934 and 1936 seasons and left it entirely for 1935 and 1937.

My inclinations are similar to Brent's, though if Paige doesn't get a Crawfords cap, that great team will almost certainly not be represented in the HoM. But then, it lasted only 5 years, so it had a pretty small window of greatness when compared to NeL teams like the Chi Am Giants, the Homestead Grays, and the KC Monarchs.

I estimate above that, if Gadfly is right, the 57 estimate is about 60 short, bringing Paige to 117 win shares in the Monarchs years. I also posit that his peak is 10-15 short, bringing him to 102-107 for the Crawfords years.

Then it has to be a Pittsburgh cap. Almost the same amount of WS in almost half the innings is the clincher for me.

BTW, seasons, at bats, or innings pitched just don't cut it for me on their own. Quality of performance does. It's similar to stating what was Ernie Bank's principle position - he started many more games at first, but he had far more value at short.

BTW, thanks for the help, guys!

some stuff i gathered on Satchel
From geocities.com/Colosseum/Arena/6925/black
1932 Pittsburgh Crawford’s also traveled around the country during the season and went 99-36 on the year;
(Their roster included Satchel Paige, Judy Johnson, Oscar Charleston, Cool Papa Bell and Josh Gibson.
Satchel Paige went 32-7 on the year and Gibson batted .380 and hit 34 HRs)
Notes: During subsequent barnstorming years, these stats were recorded: Josh Gibson hit 68 HRs (.467 Avg) in 1933, 69 HRs in 1934 and 84 HRs in 1936. Satchel Paige went 32-7 in 1932, 31-4 in 1933, 10-1 in 1934 and 24-3 in 1936.
Baseball Almanac
Satchel Paige
Joe DiMaggio said "After I got that hit off Satchel (Paige), I knew I was ready for the big leagues."
and "The best and fastest pitcher I've ever faced."
Bob Feller said “Paige was the best pitcher I ever saw."
Bill Veeck said "The best righthander baseball has ever known."
Ted Williams said "Satch was the greatest pitcher in baseball."
Dizzy Dean quotes on Satchel Paige
"I know who's the best pitcher I ever see and it's old Satchel Paige, that big lanky colored boy. My fastball looks like a change of pace alongside that little pistol bullet ole Satchel shoots up to the plate." - Sport (1969)
"If Satch and I were pitching on the same team, we'd clinch the pennant by the fourth of July and go fishing until World Series time."
“He's a better pitcher then I ever hope to be”


ESPN Classic Sports Century
Satchel Paige
For example, he claimed to have started 29 games in one month for a white semipro team in North Dakota.
Pitching for the Crawfords (off and on) from 1932-37, Paige, according to one source, went 23-7 in 1932, then won 31 of 35 decisions in 1933, including 21 straight wins and 62 consecutive scoreless innings. Paige himself claimed to have won 104 of 105 games in 1934.
In 1934 and 1935, Paige opposed baseball's best pitcher, Dizzy Dean, in six exhibition games, winning four.

HALL OF FAME BIOS

Bio
Leroy "Satchel" Paige was a legendary storyteller and one of the most entertaining pitchers in baseball history. A tall, lanky fireballer, he was arguably the Negro leagues' hardest thrower, most colorful character and greatest gate attraction. In the 1930s, the well-traveled pitcher barnstormed around the continent, baffling hitters with creatively named pitches such as the "Bat Dodger" and "Hesitation Pitch." In 1948 he was sold to Cleveland on his 42nd birthday, becoming the oldest player to make his major league debut, while helping the Indians win the pennant.

Quote
"He made his living by throwing the ball to a spot over the plate the size of a matchbook."
— Cool Papa Bell

Did You Know... that on August 20, 1948, a 42-year-old Satchel Paige pitched the Indians to a 1-0 victory over the White Sox in front of 78,382 fans, a night-game attendance record that still stands?
his HALL OF FAME PLAQUE reads
LEROY ROBERT PAIGE
"SATCHEL"
NEGRO LEAGUES 1926-1947
CLEVELAND A.L. 1948-1949
ST. LOUIS A.L. 1951-1953
KANSAS CITY A.L. 1965
PAIGE WAS ONE OF THE GREATEST STARS
TO PLAY IN THE NEGRO BASEBALL LEAGUES.
THRILLED MILLIONS OF PEOPLE AND WON
HUNDREDS OF GAMES. STRUCK OUT 21 MAJOR
LEAGUERS IN AN EXHIBITION GAME. HELPED
PITCH CLEVELAND INDIANS TO THE 1948
PENNANT IN HIS FIRST BIG LEAGUE YEAR
AT AGE 42. HIS PITCHING WAS A LEGEND
AMONG MAJOR LEAGUE HITTERS.

Gary, would you be able to identify the opposing pitchers in these games (without going to too much effort)? If you could, it might help us assess the importance of Gadfly's hypothesis that Paige was often or usually matched against other star pitchers.

Here are Paige's 14 starts in 1934 included in my count above, with opposing starters in parentheses:

Vs. Philadelphia Stars: 7 (Jones 3, Holmes 2, McDonald 1, Ellis 1).

Vs. Chicago American Giants: 4 (Trent 2, Cornelius 1, Powell 1; Bill Foster may have been injured that year, as I only have 4 starts for him before the playoffs)

Vs. Homestead Grays 1 (Dula)

Vs. Bacharach Giants 1 (Farrell)

Vs. K.C. Monarchs 1 (Brewer); Paige started for the House of David team in the Denver Post tournament championship game.

He also started (and won) a July 1 game against Nashville, with Robert Griffith as the opposing pitcher; I hadn't included this as part of the above stats because I don't have walks or strikeouts for the game. It was a 7-inning shutout.

The Philadelphia Stars, btw, were the NNL champions that year, and the Chicago American Giants were the 2nd half winners.

Thanks Gary. So it does look like there is some evidence that he was being used more against the better teams.

Brent #31
. . . the other elements of Paige’s record -—his high strikeout rates, good K/W ratios, and good WHIP statistics—- suggest he may have actually been better than his W-L record. Gadfly has suggested a possible explanation—that Paige’s opposing pitchers were usually the other elite pitchers in the league, which would have greatly hurt Paige’s run support.

There will always be crucial gaps in NeL playing records. Has anyone focused on completing game logs as published by retrosheet on mlb team webpages: date, site, teams, final score in runs, and starting pitchers? Is all of that data within the scope of the HOF sponsored data gathering?


1. Hubbell
2. Brown
3. Paige
4. Vance
5. Foster
6. Coveleski
7. Rogan

Note that number 4 to 7 were elected with some misgivings.


sunnyday2 #44
Paige 397/32-32-30/138

Now his differential from Grove is 49, or if we take into consideration the + and -, it would be -37

As for Eddie Plank, it is 47 and +47
Robin Roberts 77 or +39
Warren Spahn 35 or +7
Tom Seaver 16 or +8

There are big differences in peaks, Roberts > Paige > Spahn. Only Plank is similar in career:peak pattern, and Paige is roughly Plank plus 10%. (For those who have not internalized the win shares analysis of Eddie Plank, this says more about him than about Paige.) In pitching qualities, not quantities, Seaver is a better match.

I have a few comments to make on Satchel Paige.

After Paige’s HoM election, some of the Negro league data from the “Hall of Fame” study were released, first in Hogan’s Shades of Glory for the players who were already elected to the HoF by 2005, and then in a couple of pdf files showing statistics for the players appearing on the ballot for the 2006 Negro Leagues Committee election. (Unfortunately, three years later we're still waiting for the release of the data for the rest of the players.)

I’ve analyzed these data for several players and, fortunately, the HoF data have not significantly changed the overall picture that had been assembled on the various HoM threads. (This is, of course, a tribute to the careful analysis and compilation of MLEs that was mostly done by Chris Cobb and Eric Chalek, with very helpful assists from Gary Ashwill, KJOK, Gadfly, and others.) There is one exception, however. In the case of Satchel Paige, however, based on the HoF data from Hogan I’m now convinced that some of the analysis shown on this thread is faulty, and that we seriously underestimated his performance, especially for 1940-47.

In our earlier analysis, most of us agreed that Paige experienced a very good peak during 1930-37, though it wasn’t clear that Paige’s best seasons were better than those of, for example, Ray Brown. In # 32, however, I described Paige’s 1940-46 seasons as “a mixed bag,” and Chris (# 33) estimated his MLE DERA+ for 1940-47 as 91, that is 9 percent worse than the major league average. I see a very different picture coming from the HoF data.

From 1930-37, the HoF data show Paige as an excellent pitcher—W-L record of 45-22, TRA = 3.02, K/9 = 6.8, BB/9 = 1.5, H/9 = 6.6, K/BB = 4.5, and WHIP = .899. His W-L record, although excellent, seems a little below what it should be given his other statistics, but that may be a reflection of the reported tendency to pair him up against other top pitchers. Turning to 1940-47, based on the analysis that appeared in this thread I expected his statistics to be a lot worse, but that wasn’t the case. His W-L record was worse—29-24—but the rest of his record was comparable to, or in some respects better than, the 1930-37 period, with TRA = 3.15, K/9 = 7.7, BB/9 = 1.3, H/9 = 6.7, K/BB = 5.8, and WHIP = .897. If you leave out 1943 and 45, which were a couple of “off” seasons, his record is phenomenal (TRA = 2.43, BB/9 = .9, K/BB = 8.3, WHIP = .791). Why was his W-L record so out of line with the rest of his statistics?

One thing I discovered is an unusual usage pattern. For 1941-47, the HoF data show him pitching 94 G, 81 GS, 22 CG, but only 466.3 IP. I took his total IP and subtracted 9 innings for each CG and 1 inning for each relief appearance. That left about 255.3 innings that were spread across his remaining 59 starts, which works out to about 4.3 innings per start (excluding his CG). What seems to have been going on was that in order to capitalize on his fame while still saving his arm, they would frequently advertise him as the starting pitcher and let him pitch 3 or 4 innings, then lift him for a RP. That seems like it would have been a good strategy to take advantage of his drawing ability and to allow him to pitch more games, but I can also certainly see how it must have hurt his W-L record. For a few really important games, though, they would allow him to go the distance.

I remember thinking it was anomalous that our NeLg projections for Paige for 1940-47 were so weak, yet his record on the 48-49 Indians was so good (ERA+ of 164 and 138). Now I think that his record for most of the decade must have approached that level, with the exceptions being 1943 and 45.

To conclude, Satchel Paige truly had a remarkable career. His K/BB rates are almost otherworldly, especially in relation to other pitchers of the era. Instead of 8 years of excellence followed by a few good years here and there, I’m now seeing him as consistently excellent for nearly 25 years, with only a handful of less than stellar seasons. Paige clearly has the best statistics of any NeLg pitcher over the 1920-47 period covered by the HoF data, and Williams would have had to have had a truly amazing pre-1920 record to knock down Paige from the # 1 NeLg position overall. In my mind Satchel Paige has regained most of the reputation that he lost in our earlier analysis on this thread. If I had based his ranking in the 1924-58 group on what I knew 4 years ago, he’d probably have ranked fourth. Now he’ll rank second, and I think it’s a valid question whether he shouldn’t rank ahead of Grove.

I’ll end with a few complaints about the HoF data. (1) The data would be much more useful for analysis if league averages were available. (They provided league averages for batting average and slugging, but not for any of the pitcher categories.) (2) Data on home runs allowed aren’t included, which is a pretty basic omission. (3) Although the HoF data show ERA as well as TRA, I’m treating it with caution; I figure if Gary A isn’t able to report ERA, there must be some pretty significant gaps in the box score data, and I’m afraid the HoF may be reporting the results of some kind of model rather than actual data. (4) I’d love to have some basic team-level defense data (like DER and fielding percentage). My sense is that NeLg teams were like the major leagues of 1890-1910—lots more errors and big differences between teams in fielding performance. Getting a handle on that would be very helpful for evaluating NeLg pitching records. (Oh, and individual-level fielding data would be nice too!)

Wasn't Hilton Smith known as Paige's caddy, finishing those games that Satch started in order to get the name on the marquee?

Yes that is the conventional summary of Hilton Smith, and that is why some label him a relief pitcher.

IIRC here a few years ago someone generalized that Paige often pitched 3 innings.

(For all I know Paige may be credited with some wins in 3-inning starts.)

Wikipedia may be the other source.
Hilton Smith at wikipedia
In late 1936 Smith signed with the Kansas City Monarchs. From 1937 until his retirement in 1948 Smith was a star pitcher on the Monarchs, possessing what may have been the best curveball in baseball history; but he was overshadowed by his more flamboyant teammate Satchel Paige. Often Paige would pitch the first three innings of a game, leaving Smith to pitch the remaining six; Paige would nonetheless receive credit for the win. Also, unlike Paige, Smith was a very good hitter. Still, Smith was credited with 20 or more wins in each of his 12 seasons with the Kansas City Monarchs.

> unlike Paige, Smith was a very good hitter

I've been reading the new Larry Tye biography _Satchel_ and he mentions that Satch could hold his own with the bat (.357 with a .393 SLG in 1930 for Birmingham) when he was younger but ended up with a .218 lifetime NgL average. Not a great hitter but not below average for a pitcher and understandably much better when he was younger. He would play OF on off days and 1B during his rehab. Some other things I've come across reading the book and comparing it to the thread (I'm only halfway through the book).

Comes up with a blazing fastball and no control, his control develops over his first 2 seasons with Chatanooga (1926-27) until he could hit his spot whenever he wanted. Still mainly a fastball pitcher. The curveball develops in the early 1930s and when Dizzy Dean goes on and on in 1934 about how Satchel didn't have a curveball, Satch strikes out Diz on 3 curves in a barnstorming game saying "How's that for a guy who doesn't have a curveball?". Not surprisingly, Tye figures 1934 is his best season. Tye actually has some quotes about Satchel saying he used the curveball too much in the high elevation of Mexico where it wasn't breaking as well which caused the arm injury. He did pitch hurt against Dihigo. Tye believes the arm injury is a torn rotator cuff based on all the symptomatic evidence and he is nursed back to health by the Monarch's trainer Frank Floyd using close to the same non-surgical treatment they use for a rotator cuff injury today. In late summer 1939 he says to his catcher "You'd better be ready, because I'm ready today." By all accounts he is back to his pre-injury self with the added bonus of having picked up off-speed junk pitches during his rehab stint with the Satchel Paige All-Stars (Monarchs B Team). In the Puerto Rican winter leagues of 1939 he's back to being an ace pitcher again. There is a side mention in the book of Bus Clarkson being the best hitter in the Winter League when Satchel intentionally walks him with the bases loaded to give up the only run of that particular game. 208 K in 205 IP for the Witches in Puerto Rico. The injury basically takes 1938 to winter of 1939 out of his record.

That's as far as I've read, I'll fill in more as I read more. I do get a picture of someone who was pitching pretty much anywhere anyone with a big game and cash could get him to show up. He was throwing year-round for the highest bidder. If they needed a win he often threw a complete game, if they just needed a gate attraction he pitched 3 innings.

Good book, full of other tidbits on other players like Chet Brewer (have we worked him up?) and Willie Foster. Easy read with plenty of great stories and some investigation that takes the air out of them. Compiled stats are in the appendix for the Negro Leagues, East-West All-Star games, North Dakota (total 35-2, 301.2 IP, 440/27 K/BB, 1.79 TRA), California Winter League, Cuba, Mexico, Puerto Rico and the Miami Marlins. No stats playing for Trujillo though.

More thoughts from the book. During the 1940s Satchel was more gate attraction than top performer but he was still dominant in short stints. I think he would have been a dominant reliever post-injury, he pitched great 3 innings at a time.

It's true that Paige often started games in which he pitched only 3 innings, but that was an economic decision to allow him to pitch more often and take advantage of his enormous drawing power at the gate. For big games, however, he could still go the distance -- during the period 1940-47, 23 of his 82 starts in official Negro league games (28%) were complete games. As I discussed in # 55 above, the strikeout, walk, and TRA/ERA statistics don't show a drop-off from the 1930s to the 1940s. His W-L record did drop off, but that was undoubtedly hurt by all of the short starts where he didn't pitch enough innings to be eligible for a win. I don't think the evidence supports the idea that he was no longer a "top performer" in the 1940s.

I agree. Like I said above, he pitched to what the situation required. I'll just say it trended more toward gate attraction as the 1940s went on. He was still obviously a quality pitcher. Does anyone want selected statlines copied in from those listed above? I'm going to return the book to the library soon.

Hey, everyone,

Please find my latest MLE and a somewhat extensive mini/tangent essay on Satch Here. A link to the fully articulated MLE method, with Paige as an example, is in the first graf of the article.

Astounding work, Doctor.