Here you go, David.

Since I have Riley cracked open because I was reading up on Dihigo, here's what information he has on Scales:

Born 8/16/00, B: R T: R POS: 2b
Career 1921-1948
Teams
1921: Pittsburgh, St Louis Giants
1922: St Louis Stars
1923-1924: NY Lincoln Giants
1925-1926: Homestead, Newark
1927-1929: NY Lincoln Giants
1929-1931: Homestead
1932-1934: NY Black Yankees
1935: Homestead
1936: NY Black Yankees
1937: Cuba--Estrellas Orientales
1938: Baltimore Elite Giants
1939: NY Black Yankees
1940: NY Black Yankees, Philadelphia Stars, Baltimore Elite Giants
1941-1944: Baltimore Elite Giants
1945: NY Black Yankees
1946-1948: Baltimore Elite Giants

1921: .208
1922:
1923: .429
1924: .367
1925: .361
1926: .222
1927: .446
1928: .338
1929: .387
1930: .303 (batting 5th)
1931: .393 (batting 5th)
1932:
1933:
1934:
1935: .263
1936:
1937: .295
1938:
1939:
1940: .305
1941: .255
1942: .244 (batting third)
1943: .253
1944: .300
1945: .309
1946:
1947:
1948:

Career: .313

Riley's physical description is like that of a Kirby Puckett or Cupid Childs type body. Scales was college educated and apparently put that to good use learning to position himself for hitters, helping start up the NY Black Yankees in 1932, managing very successfully in both the NgL and esepcially in the carribean, and finally becoming a stockbroker after his baseball career had run its course.

1923 George Scales
St. Louis Stars
G-53 (team 80)
AB-164
H-64
D-10
T-7
HR-11
R-40
RBI-45
W-38
SB-7
AVE-.390
OBA-.495
SLG-.738

Primarily a third baseman this year.

1928 George Scales
Lincoln Giants

Batting
G-46 (team 46)
AB-154
H-55
D-18
T-3
HR-7
R-31
W-22
HP-0
SF-2
SH-1
SB-13
AVE-.357 (NeL east .282)
OBA-.433 (NeL east .333)
SLG-.649 (NeL east .384)

1928 George Scales

Fielding-ss
G-27
DI-232
PO-59
A-90
E-9
DP-11
RF-5.78 (NeL east 5.07)
FPCT-.943 (NeL east .917)

Fielding-2b
G-15
DI-115
PO-32
A-45
E-3
DP-4
RF-6.03 (NeL east 5.39)
FPCT-.963 (NeL east .949)

He also played one game at third base (1 putout in 9 innings), and I don't have fielding stats for one game.

The Lincoln Giants played in the Catholic Protectory Oval, Bronx, which was a very good hitters' park. The pf I have for 1928 is 130, for 33 home games, 18 road games.

A couple of small corrections to his batting stats for '28:

He played in 44 games of 46 games I have box scores for (not 46 of 46).
His at bats are 153 (not 154), making his averages .359/.435/.654.

Don't want to get ahead of myself here, but I find myself excited about this NLer with whom I was only vaguely familiar. Some nice hitting records so far, plus a key defensive position. Looking forward to reading more.

Riley says he was good at getting on base, an observation which seems to be borne out by the statistical evidence presented thus far.

Do we have any idea how good he was defensively?

BTW, thanks for starting the thread, John.

David C. J.,

His nickmane was "Tubby," and Riley suggests that his range wasn't great but that he positioned himself smartly, making up for some of the lack of range.

While we'll need to be careful with park factors for Scales' stats, he definitely fits the profile of NeL stars whom the HoM may like better than history so far has:

very good hitter, adequate but not outstanding (and certainly not graceful) defensively. It includes Beckwith, Scales, Suttles, Wilson.

This group is balanced by the group of players whom history likes better than the HoM: good hitters, speed, outstanding defensive reputation. It includes Bell, Lundy, and Mackey.

I'm not concluding here that Scales should be elected -- I have no idea about that yet -- nor am I concluding that Bell, Lundy, and Mackey shouldn't be (I think Mackey definitely should be elected). But I want to note the pattern in our assessment of NeL players and the way our analysis is relating to received reputations, which I find interesting.

According to a 1973 32-page article in a magazine called Black Sports, he also had a "rifle arm". I'm not sure how to get a copy of the article, though.

Correction -- it is a two page article.

I must admit that I am excited by this guy, Gary's stats really show a good hitter. Though a 130 park factor (probably lower because of sample size) should pull those numbers down some.

I tend to like good hitter and high peak middle infielders (i.e. Jennings and Childs) so Scales may make a surprise entrance on my ballot.

One more question, while Gary's fielding numbers look good rf and f% have their obvious shortcomings. How sure are we that he was better at his position than Wilson or Beckwtih were at theirs?

I must admit that I am excited by this guy,

I am, too (which is funny since I had never heard of him before I posted this thread). Another quality NeLer lost to the vagaries of time.

One more question, while Gary's fielding numbers look good rf and f% have their obvious shortcomings. How sure are we that he was better at his position than Wilson or Beckwtih were at theirs?

Hm. I'd have to look back at the Beckwith and Wilson stats to construct a comparative answer, but it strikes me that an infielder in a small ballpark that turns a certain percentage of fly-ball outs into home runs is likely to benefit in raw range factor, so this data may show Scales as somewhat better than he actually was.

Scales is far enough above average in range factor, however, that I can't see a park effect turning him into an below-average defender, at least at this stage of his career. I'm prepared to consider him as at least an adequate major-league second baseman.

At this moment, I'm curious about his double-play rates. How, at short and at second, does he compare to league averages on double plays per game, or something like that? Gary A., would it be easy for you to answer that question?

Well, let's see: here are his double play rates (per 9 innings) at shortstop and second base, followed by the league averages at those positions:
ss: .427 (.271)
2b: .313 (.251)

Some obvious issues: sample size (which we can't do anything about right now), dp reporting rates, and runners on base.

For Scales's home games (and a good number of his road games), I think we've pretty much got the double plays. Any games in Atlantic City and some games in Philadelphia are suspect, though. Luckily most of the Lincolns' games were at home (33 of 51).

For runners on base (dp opportunities), I don't have anything systematic (though I can whip something up in the next couple of days). However, the Lincolns' on-base percentage allowed was .332 (league average .333); their slugging percentage allowed was .406 (league average .383). They also allowed slightly more stolen bases per game than average (o.97 to 0.91). They start off with a league average number of base runners, but the last two factors would tend to drive down double play opportunities a bit. So it looks like the Lincolns' double play opportunities were a little less than league average.

The team turned .61 double plays per nine innings (league average .49). So both Scales and his team seem (by the limited evidence available) to have turned more double plays than average, with slightly fewer opportunities than average.

If his teams allowed a higher slugging percentage would it likely follow that they had a lower GB/FB ratio? The big thing with DP's (and rf as well) is whether or not a groundball hit to that particular player was common or not. A higher SLG may mean a lower GB/FB ratio, therefore meaning that Scales DP rates are legit.

Yeah, that's what I was thinking, along with the vague idea that a higher SLG would also push along a few baserunners past first or the clear the bases a little more often. I could look do some kind of win shares-type analysis of league and team assists/putouts, too. That would take a little time.

Before we get too carried away with George Scales, we should ask the question, is he the best available Negro League 2B? IOW is he the best NeL 2B not named Frank Grant (of course, Grant never really played in the NeL, per se, so maybe my first question works).

Anyway, I think the answer is no.

Bill Monroe is the only one (eligible today) who was pretty clearly the best, or at worst one of the 2-3 best position players in the NeLs at his peak. The others are not even remotely close to that. Now, of course, Monroe was a bigger fish in a smaller pond. But to me the burden of proof is to show that anybody who came after was as good as Monroe.

Among the best of the rest, I guess that with the glove the pecking order is DeMoss, Allen, then Scales or Hughes.

With the bat: Maybe Scales, then Allen or Hughes, then DeMoss a fairly distant fourth.

Is Hughes eligible, BTW? I forget.

But at the moment I would have to have it:

1. Monroe
2. Allen or Scales--James comps Allen's style to tony Taylor, Frannk White, Manny Trillo, Cookie Rojas, Julie Javier--faint praise indeed, though that is only for style, not quality--his actual numbers comp Scales quite well and I think he is regarded as superior defensively
3. Scales or Allen--Scales he (BJ) likens to Bill Madlock
4. DeMoss--bunter and slap hitter but the best with the glove
5. Hughes--BJ comps him to Sandberg and Larkin

Monroe has been on my ballot in the (distant) past and could get there again. But he has not been on my ballot recently. And Allen has been down in the 50s and 60s. Scales would have to be a lot better than Allen to be as good as Monroe, and would have to be clearly better than Monroe to make my ballot right now.

The evidence suggests neither. Maybe he belongs in the consideration set, and maybe he belongs somewhere around Allen in the 60s. There's pretty much going to have to be a case of mistaken identity (in terms of the above data) for him to make my ballot.

I mean a .313 average even with middling power translates to what?

Is Hughes eligible, BTW? I forget.

Sammy T. will be eligible in 1952.

I mean a .313 average even with middling power translates to what?

That's the question, isn't it? I hope to have an answer on the standard translation scale soon.

Sunny-D,

While I agree that we shouldn't be too quick to lionize Scales while Monroe is still out there, I do think that relying too much on Bill James's rankings is dubious.

We've seen with previous candidates that he's basically read a couple of the same books we may have (namely, Riley and Holway), and his rankings are, therefore, susceptible to many of the same peculiarities and errors that his source material is. And I think he'd be the first to admit this. This issue seems to particularly rear its head at 3B, 2B, and the corner outfield positions. It seems as though expert opinion, the HOF, and statistical analysis are most divergent at these positions which leads to his rankings jumbalaya there.

Again, however, I say that believing that Monroe is probably being underrated by me and the electorate in general.

Erik (Dr. C):

How ironic, everytime I mention Bill James by name somebody objects. Strange.

But anyway, if you read my post in detail you would know I didn't "rely" on James' rankings at all. I have his #5 at #1, and his #1 at #4. I have the guy he comps to Tony Taylor ahead of the guy he comps to Larkin and Sandberg.

My main point--Monroe still seems to have been by far the best of this group, i.e. not only a good glove but one of the best hitters of the day.

By comparison Scales and DeMoss and the others are probably 200 OPS points or more behind the best hitters, and probably 100 OPS points or more behind the best SSs nor anywhere near as valuable defensively as Lundy and Moore and Wells.

Sunny,

You're right I knee-jerked about James. It's strange and interesting, this bash-the-expert-ranking phenomenon. It seems like whenever we talk about a third-party's player rankings that there's a tendency to poo-poo them for their gaps (and this is particularly true for James's approach to ranking NgL and 19th C. players), but when we talk about other things he's done, the usual reverence is in play because we are then using his ideas as part of our toolbox instead of as evidence.

So I guess this is in a sense similar to the warning that Gadfly issued about Chris Cobb's projections: that as discussers (let alone voters), many of us worry about ourselves (or others) taking expert rankings and MLEs as gospel because we know comparatively little about the players involved. [which isn't to suggest that Chris wouldn't advise each of us to make up our own mind, by the way]

Hmmm...given that, I ought to take my own advice and take a long, hard look at Lundy and Redding....

Bill Madlock may be a legit offensive comparable, but right now I am not convinced that Scales awasn't at lesat an average 2B. His DP rates are good, his rf is good (and this despite his teams giving up a high SLG, possibly meaning a low GB/FB ratio), and his F% is decent as well.

Still, I can't really put him above Dobie Moore right now and Moore is 16th. Right now he probably falls into my MI glut of Lundy, Monroe, Doyle, and Sewell, anywhere from 25-40.

It's odd, I want to take yet another look at Monroe, but in my head I can't think of anyone that he is obviously better than (maybe doyle, maybe sewell but that won't get him much higher). In some way I think we as voters are having a hard time adjusting to the fact that we have much stronger eligibles now than 15-20 years ago. Scales would most likely have made a number of ballots had he become eligible when Bill Monroe did. Same with Frank Duncan and Bruce Petway.

A little bit of new Scales data:

1921 George Scales
Pittsburgh Keystones/St. Louis Giants

G-16 (4 of 4 games during known tenure with Pittsburgh; 12 of 12 games during tenure with StL)
AB-61
H-16
D-2
T-2
HR-0
R-8
W-3
HP-0
SH-1
SB-0
AVE-.262 (NNL .267)
OBA-.297 (NNL .329)
SLG-.361 (NNL .366)

PFs for Pittsburgh's Central Park: 73, for 10 home games, 5 road games; for St. Louis's Giants Park, 102 for 46 home games, 30 road games (PF 68 in 1920).

Fielding-3b (all with St. Louis)
G-12
DI-95
PO-10
A-25
E-5
DP-0
RF-3.32 (team 2.66; NNL 3.36)
FPCT-.875 (team .922; NNL .920)

Fielding-1b (all with Pittsburgh)
G-4
DI-32.7
PO-42
A-3
E-2
DP-0
FPCT-.957 (NNL .979)

I've started work on MLEs for George Scales, and I now have to consider park factors . . .

Some questions:

1) The St. Louis Stars were playing in their great hitters' park in 1923, yes?

2) Gary A., do you have any park factors for the Lincoln Giants in the Catholic Protectory Oval? Was it as good a hitters' park as St. Louis, or a bit less extreme? I've only done one season of MLEs for a NY Lincoln Giants player: John Beckwith in 1930, and I used a pf of 110 for that season. Does that seem right, or should it be lowered? raised??

3) Do we know anything about where the NY Black Yankees played? Even late in the 1930s, Holway has Scales with some excellent power numbers when he was playing for the Black Yankees. Were they in the Oval or some other hitters' park then?

Thanks!!

I don't know about the years Scales was there, but in Green Cathedrals Philip Lowry has the 1938 Black Yankees playing at Triborough Stadium, which was located near the Triborough Bridge on Randall's Island. Scales wasn't playing there in '38, and Lowry only has the NY Black Yankees in the Triborough Stadium for 1938, but maybe they were there other years?

Chris, the only pf for the Catholic Protectory Oval is for 1928: 130, for 33 home games, 18 road games. I don't have systematic evidence for other seasons, but my strong impression is that it was a very good hitters' park. I am close to having enough data for 1924 park factors in the east, so maybe I can put those together over the weekend (my main concern is not having all the scores from Harrisburg). Anyway, an adjustment of 110 seems reasonable to me for now.

Yes, the Stars were in the great hitters' park by 1923 (they moved in mid-1922).

At some point in the 1930s the Black Yankees started playing in Yankees Stadium (that in fact may be why they adopted the name). I have a vague memory that in 1934 they used a park somewhere in New Jersey, but I will have to check up on that to make sure.

I'm running a search and I see that the Black Yankees opened their 1943 home schedule at Yankee Stadium.

I'm looking at news stories of Black Yankees games from 1934 to 1945, and so far consistently I am seeing Yankee Stadium mentioned as the home park.

The Black Yankees supposedly played a number of games in the Hinchliffe Football Stadium in Patterson, NJ between 1936 and 1945.

In 1938, I believe Triborough Football Stadium was at least one of their primary fields.

George Scales MLEs

These MLEs are developed by the usual processes. Here are a few specifics of which you should be aware.

1) I set Scales’ MLE seasons as running from 1923 to 1938. I calculated MLEs for 1921 to 1940 to make the regression smooth for his MLE seasons, and I have posted these seasons as well. Scales played through 1946, but I did not see any likelihood of MLE credit for these, his age 41-46 seasons.

2) The park factor for Scales from 1923-30 is 110, as he played in great hitters’ parks throughout this period, first for the St. Louis Stars and then for the NY Lincoln Giants. After that I have set his park factor as 100.

3) There is very little data available for Scales’ seasons from 1932 to 1939 except for 1935, his last year with the Grays. He played with the NY Black Yankees in 4 of those seasons when they were not in a league, so few of their games have been documented. He played in Santa Domingo in 1937, and there is data for that, but there’s no telling what the level of competition was.

Unfortunately, Macmillan 10th and Holway give very different views of Scales’ abilities as a hitter from 1932 through 1936. Macmillan has him falling off a cliff in 1932, posting averages of .217, .250, .146, .248, and .217 for those years, based on at bat totals of 46, 8, 48, 137, and 60, respectively. Holway doesn’t give ab totals, but it looks like his data base isn’t much larger. His averages, however, are significantly higher, more or less in line with a gradual decline from Scales’ prime: .292, .375, .200, .256, .358.

Thus, we have two very different pictures of Scales’ later career. Regression analysis brings them towards one another, but his career totals still differ significantly depending on which set of numbers you accept as more valid. I have therefore done two sets of MLEs. They use the same data 1921-31, diverge from 32-36, and join again 37-40. The MLEs for 1937 are based purely on regression analysis of the surrounding seasons.

4) Playing time for Scales is modeled on the playing time of Frisch and Traynor, contemporaries of Scales playing at his two positions and having a comparable stretch of play as starters. Both averaged just over 140 games per season for their first ten years as regulars, then averaged just over 120 for their remaining 5 years. I’ve varied Scales’ playing time from season to season to hit averages of 142 games and 122 games for these parts of his careers, with variation above and below the mean comparable to these two. I’ve also used them as models for the ratio of PA to games.

5) In the common years of 1921 to 1931, I’ve used Gary A.’s data for 1921, 1923, and 1928 and Holway’s data for 1931. Other years are drawn from Macmillan 10th.

MLEs based on Holway, 1932-36

Year Team   EqG  PA    BB  Hits  TB  BA   OBP  SA
(1921 Stl    80  320   22   81  120 .272 .324 .404)
(1922        60  240   25   52   85 .243 .322 .395)
1923 Stl/NY 141  592   96  154  252 .311 .422 .509
1924 NY LG  152  638   77  167  240 .298 .383 .428
1925 NY*    150  630   79  170  253 .308 .394 .459
1926 NY*    120  504   68  117  167 .267 .366 .383
1927 NY     143  601   77  173  261 .330 .415 .498
1928        137  575   78  150  239 .301 .395 .481
1929        153  643   87  180  262 .324 .415 .471
1930 Homest 148  622   89  157  235 .294 .395 .441
1931        142  596   85  161  237 .315 .413 .464
1932 NY BY  134  536   80  132  193 .289 .395 .424
1933        117  468   70  114  159 .286 .392 .399
1934        128  512   80  104  154 .241 .360 .356
1935 Homest 143  572   90  117  183 .242 .361 .379
1936 NY BY  100  387   58   92  154 .281 .389 .468
1937 S.D.    55  200   32   41   66 .244 .364 .395
1938 Bal     23   60   10   10   14 .207 .336 .277
(1939 NY BY  25   75   12   13   20 .199 .325 .320)
(1940 Bal    25   75   11   15   23 .241 .349 .362)
totals**   1986 8136 1153 2039 3072 .292 .392 .440

*Also played for Homestead in these seasons, but data appears to be almost exclusively from NY.
**Totals do not include 1921-22, 1939-40

MLEs Based on Macmillan 10th, 32-36

Year Team  EqG   PA    BB  Hits  TB  BA   OBP  SA
(1921 Stl    80  320   22   81  120 .272 .324 .404)
(1922        60  240   25   52   85 .243 .322 .395)
1923 Stl/NY 141  592   96  154  252 .311 .422 .509
1924 NY LG  152  638   77  167  240 .298 .383 .428
1925 NY*    150  630   79  170  253 .308 .394 .459
1926 NY*    120  504   68  117  167 .267 .366 .383
1927 NY LG  143  601   77  173  261 .330 .415 .498
1928        137  575   78  150  239 .301 .395 .481
1929        153  643   87  180  262 .324 .415 .471
1930 Homest 148  622   89  155  233 .291 .393 .437
1931        142  596   85  160  236 .313 .411 .461
1932 NY BY  134  536   83  115  169 .255 .370 .373
1933        117  468   72  101  139 .256 .371 .352
1934        128  512   85   78  109 .184 .320 .256
1935 Homest 143  572   93  102  142 .213 .340 .296
1936 NY BY  100  387   63   65  107 .201 .332 .329
1937 S.D.    55  200   33   35   53 .210 .340 .316
1938 Bal     23   60   10   10   13 .192 .326 .253
(1939 NY BY  25   75   12   13   20 .199 .325 .320)
(1940 Bal    25   75   11   15   23 .241 .349 .362)
totals**   1986 8136 1174 1936 2876 .277 .382 .413

I do like middle indielders with offensive skills, but right now I can't justify putting Scales ahead of Dobie Moore. I think some of the initial excitement about him didnt' take into account his home parks, which overall were very easy on the bats. At best I could make him a part of the Sewell, Doyle, Monroe, Lundy backlog in the late 20's and early 30's. Of course some of you will be irate that I just insinuated that Scales was a good as Sewell.

By late 20's early 30's I meant spots on my ballot. High 20's, low 30's would have been a better way to state that.

George Scales

-First you have Year, Team(s), PA.
-Second you have Chris's MLE's
-Third, in parentheses, you have pitchers-removed offense context. MLB for the 20s, then NL
-Fourth, you have AVG+/OBP+/SLG+
-Lastly, is the OPS+

MLEs based on Holway, 1932-36

1921    Stl 320  0.272/0.322/0.404   (0.299/0.357/0.416)    91/ 90/ 97     87
1922    Stl 240  0.243/0.321/0.395   (0.297/0.359/0.415)    82/ 89/ 95     85
1923 Stl/NY 592  0.311/0.422/0.509   (0.292/0.356/0.405)   107/119/126    144
1924  NY/LG 638  0.298/0.382/0.428   (0.294/0.356/0.406)   101/107/105    113
1925    NY* 630  0.308/0.395/0.459   (0.300/0.364/0.425)   103/109/108    117
1926    NY* 504  0.267/0.367/0.383   (0.289/0.355/0.402)    92/103/ 95     99
1927     NY 601  0.330/0.416/0.498   (0.292/0.355/0.406)   113/117/123    140
1928     NY 575  0.301/0.397/0.481   (0.290/0.355/0.412)   104/112/117    128
1929     NY 643  0.324/0.415/0.471   (0.298/0.363/0.432)   109/114/109    123
1930 Homest 622  0.294/0.395/0.441   (0.312/0.370/0.464)    94/107/ 95    102
1931 Homest 596  0.315/0.413/0.464   (0.285/0.344/0.403)   111/120/115    135
1932  NY/BY 536  0.289/0.396/0.424   (0.284/0.337/0.412)   102/117/103    120
1933  NY/BY 468  0.286/0.393/0.399   (0.274/0.327/0.376)   104/120/106    126
1934  NY/BY 512  0.241/0.359/0.356   (0.287/0.342/0.408)    84/105/ 87     92
1935 Homest 572  0.242/0.362/0.379   (0.286/0.341/0.407)    85/106/ 93     99
1936  NY/BY 387  0.281/0.388/0.468   (0.286/0.345/0.400)    98/112/117    129
1937   S.D. 200  0.244/0.365/0.395   (0.280/0.342/0.397)    87/107/ 99    106
1938    Bal  60  0.207/0.333/0.277   (0.275/0.339/0.391)    75/ 98/ 71     69
1939  NY/BY  75  0.199/0.333/0.320   (0.280/0.346/0.401)    71/ 96/ 80     76
1940    Bal  75  0.241/0.347/0.362   (0.272/0.337/0.391)    89/103/ 93     95

MLEs Based on Macmillan 10th, 32-36

1921    Stl 320  0.272/0.322/0.404   (0.299/0.357/0.416)    91/ 90/ 97     87
1922    Stl 240  0.243/0.321/0.395   (0.297/0.359/0.415)    82/ 89/ 95     85
1923 Stl/NY 592  0.311/0.422/0.509   (0.292/0.356/0.405)   107/119/126    144
1924  NY/LG 638  0.298/0.382/0.428   (0.294/0.356/0.406)   101/107/105    113
1925    NY* 630  0.308/0.395/0.459   (0.300/0.364/0.425)   103/109/108    117
1926    NY* 504  0.267/0.367/0.383   (0.289/0.355/0.402)    92/103/ 95     99
1927  NY/LG 601  0.330/0.416/0.498   (0.292/0.355/0.406)   113/117/123    140
1928  NY/LG 575  0.301/0.397/0.481   (0.290/0.355/0.412)   104/112/117    128
1929  NY/LG 643  0.324/0.415/0.471   (0.298/0.363/0.432)   109/114/109    123
1930 Homest 622  0.291/0.392/0.437   (0.312/0.370/0.464)    93/106/ 94    100
1931 Homest 596  0.313/0.411/0.461   (0.285/0.344/0.403)   110/119/114    134
1932  NY/BY 536  0.255/0.369/0.373   (0.284/0.337/0.412)    90/110/ 91    100
1933  NY/BY 468  0.256/0.370/0.352   (0.274/0.327/0.376)    93/113/ 94    107
1934  NY/BY 512  0.184/0.318/0.256   (0.287/0.342/0.408)    64/ 93/ 63     56
1935 Homest 572  0.213/0.341/0.296   (0.286/0.341/0.407)    74/100/ 73     73
1936  NY/BY 387  0.201/0.331/0.329   (0.286/0.345/0.400)    70/ 96/ 82     78
1937   S.D. 200  0.210/0.340/0.316   (0.280/0.342/0.397)    75/ 99/ 80     79
1938    Bal  60  0.192/0.333/0.253   (0.275/0.339/0.391)    70/ 98/ 65     63
1939  NY/BY  75  0.199/0.333/0.320   (0.280/0.346/0.401)    71/ 96/ 80     76
1940    Bal  75  0.241/0.347/0.362   (0.272/0.337/0.391)    89/103/ 93     95

Using 1923-38 data for the George Scales totals.

MLEs based on Holway, 1932-36

Counting stats (+/- 2 for rounding?)
8136 PA
6983 AB
2039 H
3069 TB

Percentages
Scales -- 0.292/0.393/0.439
Context -- (0.291/0.351/0.412)
Plusses -- 100/112/107

OPS+ = 118

MLEs Based on Macmillan 10th, 32-36

Counting stats (+/- 2 for rounding?)
8136 PA
6958 AB
1932 H
2875 TB

Percentages
Scales -- 0.278/0.382/0.413
Context -- (0.291/0.351/0.412)
Plusses -- 95/109/100

OPS+ = 109

OPS+ = 118

That's pretty good for a third baseman of that time. WS projections would help for comparisons sake.

John, he was a 2B.

I know we're not supposed to be looking ahead but at a minimum I would wait until I was sure he was better than Sammy Hughes. I think he might be better than Newt Allen and I'm pretty sure he was better than Bingo DeMoss. But I couldn't possibly vote for any of them until I was sure which was the best of the group...

Not to mention I'd have to be convinced that the best of this group is also better than Bill Monroe, and right now I'm pretty sure he (whomever he is) is not.

Still I'd consider Scales if I could be convinced of those two things: 1) Better than Allen, DeMoss and Hughes, and 2) Better than Monroe.

John, he was a 2B.

I meant 2B. Thanks, Marc.

Phil Lowry lists Hinchcliffe Stadium, Paterson NJ, as the home field of the NNL Black Yankees, 1936-37 and 1939-45. The listing immediately precedes the one for Triborough Stadium, New York NY. That is, it's a New York NY listing because the team represented that city. We'll see how he handles the AL Angels in the third edition.

KJOK may have a complete cross-reference in a computer file.

KJOK may have a complete cross-reference in a computer file.

Probably not, since KJOK is the author of #32.
I did say maybe.

George Scales' positions, according to Gary A.'s fielding info and Macmillan 10th (Holway is generally in agreement, though he lists only one position per season)

1921-23 3b
1923-24 2b
1925 3b
1925-26 ss
1927 3b
1928 SS
1928-32 2b
1933 OF
1934-35 2b (Holway has left field in 1933-34 and 3b in 1935)
1936 3b
1937 ??
1938-39 3b
1940 OF (Holway has left field)

Dropping 1921, 22, 39, 40 that comes to

2b -- 8 seasons
3b -- 4 seasons
ss -- 2 seasons
lf -- 1 season
?? -- 1 season

By Holway's count, it's

2b -- 6 seasons
3b -- 5 seasons
ss -- 2 seasons
lf -- 2 seasons
?? -- 1 season

So, yes, he is a second baseman, but he was a second baseman only for about half of his MLE career.

Thinking aloud about George Scales:

Some quick comps on batting:

If the Macmillan numbers are right, Scales' career value as a hitter is almost a dead ringer for Joe Sewell, +1000 PA: Good OBP, league average power.

Sewell has a big advantage on defense, Scales' peak as a hitter probably a bit higher. Without needing closer analysis, I take Sewell.

Scales' career length is almost a dead ringer for Pie Traynor, and the OPS+ is very similar, though Scales' advantage comes from BB and Traynor's from above average power.

Traynor would edge Scales on defense: I take Traynor, no further analysis needed.

If the Holway numbers are more accurate, Scales has Sewell's OBP and Traynor's power, so he has a clear advantage on offense over both, and we have to start estimating WS to see where exactly Scales lands in the glut of good candidates.

We also, at that point, demand a new look at Dick Lundy and Dobie Moore . . .

As to Allen, DeMoss, Hughes, and Monroe.

Allen, DeMoss, and Hughes were surely great defensive players, but they were slap hitters.

Allen and Hughes slugged about .400 in their NeL careers; that leads to MLE slugging percentages around .340-.360: they look to be Rabbit Maranville/Marty Marion type hitters. Hughes may have been a bit better than that, since most of his career falls in a lower offense period, but he was not an exceptional hitter in any case.

DeMoss's hitting is harder to evaluate because of lack of data and the effect of Schorling Park and Rube-Foster small ball on his stats, but he _looks_ a lot worse than Allen and Hughes, and proper adjustments are only likely to bring him up to their level.

Of the three, Allen seems to me to be the best: he might have a decent career case: I see his MLE career as about 18 seasons, 1924-41. DeMoss just doesn't look to have the hitting, and Hughes doesn't have enough career at 15 MLE seasons, 1932-46, given that he hit no better than Allen.

Bill Monroe is another matter. He has a great defensive reputation and a good hitting reputation. I feel like I have no firm idea of how good he really was: I am less certain about him than about any other NeL player who has been under serious consideration.

So among the NeL second base group, Allen deserves a look to see if his OPS+ breaks 100 and his superior defensive value makes up his hitting shortfall in comparison to Scales. Monroe -- well, do what you will with his reputation.

I'll take even the Macmillan Scales over DeMoss and Hughes.

Anyone know much about Rev Cannaday? Riley's got him listed as a 2B; he's a contemporary of all these guys, and he looks like he could hit a little. Is he a member of the NgL 2B glut, or an also ran?

Well, the Mac 8 lists these career stats for Walter (Rev) Cannady, 1922-39:

G-393
AB-1470
H-443
D-58
T-19
HR-31
SB-34
AVE-.301
SLG-.429

His positions are listed as follows:
1922-P,OF,SS
1923-24 (no stats-Homestead Grays out of the league)
1925-SS
1926-SS
1927-SS/3B
1928-SS
1929-SS,2B,1B
1930-2B
1932-no position given
1933-2B
1934-2B
1935-no stats
1936-no position
1937-3B
1938-3B
1939-SS,3B

Riley's got him as primarily a 2B, then a 3B, SS, OF, 1B, but describes him as an "all-purpose" player as well.

He lists him with these stats and positions
1922: .303
1925: .386, 11 HR in 52 G (SS, hit 4th behind OC)
1926: .317, 7 HR (3B)
1929: .378 (SS)
1930: .370
1931: .235
1932: .309
1936: .328
1937: .385 (through mid season)
1938: .321
1944: .222

The Macmillan is completely wrong about Cannady's 1928 season, anyway. They have him playing shortstop for the Lincoln Gts, with these stats:

G-52
AB-205
H-61
D-4
T-3
HR-2
SB-5
AVE-.298
SLG-.376

In fact, in 1928 Cannady played first base for Hilldale, then moved in August to the Homestead Grays, where he played mostly at second (and a little at first base).

His combined 1928 batting stats:
G-65
AB-252 (4th)
H-84 (3rd)
D-18 (tied for 2nd)
T-5
HR-3
R-56 (4th)
W-16
HP-0
SF-3
SH-2
SB-15 (4th)
AVE-.333
OBA-.369
SLG-.480

Rev Cannady 1928 Fielding--first base
G-58
DI-491.7
PO-587
A-22
E-21 (1st)
DP-23
FPCT-.967 (NeL east .979)

Rev Cannady 1928 Fielding--second base
G-8
DI-57
PO-6
A-18
E-2
DP-0
RF-3.93 (team 2b 4.50; NeL east 5.39)
FPCT-.923 (team 2b .952; NeL east .949)

Sorry--should say Cannady moved to the Grays in September, not August. September 16, to be precise.

Paul,
That's "Hinchliffe" stadium in Paterson, although 9 of 10 every 10 references I see calls incorrectly call it Hinchcliffe (it just sounds better, but the former Paterson mayor was Hinchliffe, alas).

I think it closed in 1997, but it may yet be saved. It's on the New Jersey Register of Historic Places, having opened in 1932.
FYI, Larry Doby is from Paterson and always was fond of his roots. Many Paterson parks and such bear Doby's name.

1921 Walter Cannady
Cleveland Tate Stars

G-33
AB-116
H-39
D-9
T-2
HR-1
R-23
W-10
HP-0
SH-8
SB-5
AVE-.336 (NNL .267)
OBA-.389 (NNL .329)
SLG-.474 (NNL .366)

Raw PF for Cleveland's Tate Field: 125, for 24 home games, 13 road games.

Cannady played 13 games in left field, 13 at shortstop, 3 in center field, and 2 at third base.

He also pitched four games, going 2-1 with a 9.24 TRA in 25.3 innings. (The team was 15-21-1 against NNL opponents.)

Can someone post the Macmillan career totals for Scales, without any adjustment? My old Macmillan encyclopedia is the 7th edition and does not have the info. Thanks.

I'd be curious to see MLEs for Cannaday, I suspect that he and Scales fall very close together and end up in the Lazzeri area of the IF glut.

George Scales' career stats, according to Macmillan 10th:

580 g
2032 ab
627 hits
129 2b
27 3b
66 hr
38 sb
.309 ba
.496 sa

Holway has him at 793-2578, .308, with 68 hr

KJOK may have a complete cross-reference in a computer file.

Probably not, since KJOK is the author of #32.
I did say maybe.

Actually, I do, but I've been trying to verify it's accuracy before publishing it. I'll try to post something by this weekend.....

Also, based on limited boxscores that I've found:

Scales was a cleanup hitter in the 1925, 1927 & 1928 seasons. I haven't been able to find a boxscore in those seasons when Scales DIDN'T bat cleanup.

Scales played SS in 1925, 3B in 1927, and SS in 1928.

Updated George Scales raw numbers:

Macmillan10th...309 BA...496 SA...2032 AB
Holway..........308 BA...??? SA...2578 AB
HOF Study.......321 BA...508 SA...2215 AB

It seems that he got a boost from the study over both Mac and Holway. With Holway MLE's, he had a 118 OPS+. Just eyeballing the new stats would bring him over 120, perhaps up to 125. Plus, his OPS is heavy on the OBP, and so may underrate him.

I think Scales needs to be revisited by everyone.

I thought so too, but was told that his record is all park effects.... The raw data certainly puts him a lot higher up than I have ever had him. I'm with Ron.

>I think Scales needs to be revisited by everyone.

Yeah, like a WS projection! Doc? Chris?

I must defer here to Chris. He did the original projections, and they included his estimations for park effects (and maybe league-quality effects?) IIRC.

I decided to try my hand at estimating Scales' Win shares. Any errors are mine alone and a result of my misunderstanding of Chris' and Doc's methods.

I used Scales' MLE OPS+/OBP/SLG and tried to find players at his position with comparable rates. I used both leagues from 1921-1929 and the NL from 1930 to 1938. I thought that was how the translations were done. Then, I found those players win shares and compared their win shares to games played. I used that rate and multiplied that by Scales' games played. Chris said he set Scales' games played to average to 142 games over his first 10 seasons and 120 games over the next few years before the 2 year decline - 1937 and 1938. I wasn't sure how many games Chris used in each year, so I just used 77 games in 1921, 60 games in 1922, 142 games from 1923-1932, 120 games from 1933 - 1936, 54 games in 1937, and a few games in 1938.
There were some years where there were no comparable players at his position because Scales' OPS+ were too high.
1923: 144 OPS+. No one was within 15 points at 3rd base.
1927: 140 OPS+. No 3rd baseman is close.
1928: 128 OPS+ at SS. No one else is over 114.
1931: 135 OPS+ at 2b. No one in NL over 121.
1932: 120 OPS+ at 2b. Highest in NL is 111.
1936: 129 OPS+ at 3b in half play. No one in NL over 111.
In those cases, I tried to find comparables at the 2nd, 3rd, and SS.

Anyway, this is what I came up with.
1921: 7, at 3b
1922: 6, at 3b
1923: 29, at 3b/2b
1924: 20, at 2b
1925: 23, at SS
1926: 17, at SS
1927: 30, at 3b
1928: 24, at SS
1929: 23, at 2b
1930: 18, at 2b
1931: 25, at 2b
1932: 21, at 2b
1933: 21, at 2b
1934: 12, at 2b or 10, at OF
1935: 15, at 2b or 14, at 3b
1936: 14, at 3b
1937: 5, ?
1938: 1, at 3b

Career total: 310
Best any 3 years/peak: 84
Best 3 straight years (Kelly's peak measure): 77
Best 5 straight years/prime: 120
Best any 7 years (Kelly's prime): 175
I believe all these numbers should be increased because the results of the Hall of Fame study.
Position: Good Question.

Some comprarables at 2b:
Frisch:
Career: 366
James Peak: 96
Kelly Peak: 82
James 5 yr Prime: 135
Kelly Prime: 195

Herman:
Career: 298
James Peak: 90
Kelly Peak: 90
James 5 yr prime: 135
kelly Prime: 182

Lazzeri:
Career: 252
James Peak: 81
Kelly Peak: 76
James 5 yr prime: 115
kelly Prime: 165

Cuccinello:
Career: 203
James Peak: 69
Kelly Peak: 63
James 5 yr prime: 92
kelly Prime: 136

McManus:
Career: 202
James Peak: 62
K Peak: 56
James 5 yr Prime: 92
K Prime: 131

3b:
Beckwith (these may have been updated since I last updated his numbers in my spreadsheet)
Career: 317
K Peak: 88
K Prime: 186

Traynor:
Career: 274
James Peak: 80
K Peak: 77
James 5 yr Prime: 119
K Prime: 167

Lindstrom:
Career: 193
James Peak: 82
K Peak: 77
James 5 yr Prime: 116
K Prime: 153

Comparable SS:
Sewell:
Career: 277
K Peak: 76
K Prime: 174

Bancroft:
Career: 269
K Peak: 84
K Prime: 167



In 1923, if a 2b, his 29 would be the best in AL over Collins' 24, but behind Frisch's 31 in the NL.
In 1923, if a 3b, his 29 would be the best in Al over Kamm's 20, and tied/just ahead of Traynor's 28 in the NL.
In 1925, his 23 at SS would put him in the argument with Sewell's and Wright's 24 for best in majors. Dobie Moore would be best overall though.
In 1927, his 30 at 3b would make him best in majors over Traynor's 26 and Hale's 18.
In 1928, his 24 at SS would make him arguably the best in the majors over Sewell's 23 and Jackson's 22.
In 1931, his 25 at 2b would make him arguably the best in the majors with Bishop's 25 and Cuccinello's 23.
In 1932, his 21 at 2b would put him right behind Herman's 23 in NL.

I tried to be conservative. Also, remember these are numbers from Holway. The numbers from the Hall of Fame Study would give him a significant boost - maybe from a 118 OPS+ to a 125 OPS+. Consequently, his win shares would need to be increased, but I don't know by how much each year. If the increase in OPS+ is worth 2 win shares a year in his years as a regular (1923-1936) that would be an additional 30 win shares. If worth 1.5, then an additional 22. Also, his peak and prime would come very close to Billy Herman, Frankie Frisch, and John Beckwith.

Hope this helps.

Since this is the first time I have done this, please be kind when you point out my mistakes.

If people want to see all of my comparisons year by year, I can clean them up and post them by Wednesday.

In the HoF numbers, Scales HR/PA was about the same as Willard Brown.

In the HoF numbers, Scales HR/PA was about the same as Willard Brown.

Without in any way wanting to dismiss Scales as a serious candidate (and he's been in my top 40-50 for a long time), it's important to bear in mind that he was playing during the highest offense era in NeL history, and he was playing in extreme hitters' parks throughout the 1920s. Willard Brown was one of the top 4 hitters for power in the Negro Leagues in the 1940s (after Gibson, with Leonard and Irvin). That might have been top 5 if Luke Easter had been consistently playing league ball. George Scales was not one of the top 4 power hitters in the Negro Leagues in the 1920s.

I think the WS estimates put together by Muller & Scully look reasonable, btw.

I know it is the end of the school year for many, but I was hoping we could get an update of Scales' MLEs after the Hall of Fame study. From the study's numbers, it looks like the mid-30s trough is not as deep as previously thought. Probably only 1933 and 1934.
Also, he may have been able to play into his 40s. These are his numbers from 1940 from the HOF study: .353 to lg .272, slg: .571 to lg .383. His OBP (h+bb/ab+bb) was .468. He was playing for the Baltimore Elite Giants that year. Does anyone know what their park factor was like?

I am trying to compare the numbers from Mac to the HoF to see how various years need to be adjusted, but I am a neophyte at this.

Also, what were the years when so many players went to Mexico? I don't remember what years that happened nor what thread to check. Thx

I wonder how Scales MLEs will compare to Bus Clarkson's...but more than that, I still wonder if Marvin Williams might beat them both.

I'm pretty darned certain that Clarkson is a better player all around than Marv Williams, based on the available data, my own limited interpretation of their histories, and my latest method for finding MLE figures. Williams played in a lot of lower-level leagues. He played very, very well there. But when you convert that to an MLB environment, a lot of air comes out. But also Williams played in a ton of border leagues and leagues at altitude where runs came cheaply. When you account for that, you see still more erosion.

Take his year in the AZ-TX league where he hit .401/.471 est./.854. He was amazing that year. But the league was averaging 7 r/g! So the NL was scoring only 60% as often as AZ-TX.... Then tack on the fact that the league was Class C (second lowest classification, which I take at .65 of the majors) and had very few big stars or future stars, then you see why his translation into the 4.17 r/g National League comes out as .250/.314/.496. In a 4.5 r/g environment, that's .261/.327/.518. Good, especially for a 2B, not great.

Here then are Clarkson's advantages over Williams as I see them:
1) Played in high-level leagues more often
2) Better MLE numbers
3) Played a tougher set of positions (SS/3B vs. 2B/1B), and I suspect played them about as well as Williams played his (neither has hosanahs regarding their gloves).

As for Scales, Chris will likely need to weigh in on that one, I don't have as strong a working knowledge about the Tubster as I do Williams and Clarkson.

This guy looks pretty good. Why didn't he get more support?

Scales didn't have a good defensive reputation, and certainly would not have played SS in MLB, which would be my guess why he didn't get more support.

Right....although he's not a bad candidate even as a 2B -- I'm not voting for him in 2014 but I would have in say....1950

Please click through for my latest George Scales MLEs.