Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Hall of Merit > Discussion
Hall of Merit
— A Look at Baseball's All-Time Best

Friday, June 11, 2004

Spottswood Poles

If anyone would like to pass along some numbers or something for this intro, it’d be great.

Joey Numbaz (Scruff) Posted: June 11, 2004 at 04:42 AM | 86 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

   1. Chris Cobb Posted: June 11, 2004 at 01:41 PM (#671135)
Poles is going to call for a lot of discussion, so I thought I'd get the ball rolling a little early by posting the data I've gathered.

Spotswood Poles Data

Summary of expert rankings

HF% Career Name-pos (born) BJ – MVP - All-Star
52% 09-23 Spotswood Poles-CF (1889) #4 cf - 3 - 3*

Note that his all-star count from Holway is probably too low as Holway did not name an east all-star team in 1909 or 1912, both seasons in which Poles might well have been named.

From Holway

1909 – .263 for Philadelphia Giants no ab listed, but 3 teammates listed with 20 and 22 so I estimate 5-19)
1910 – 14-42 for Phil Giants (his .333 avg. is 4th best in the East); Holway all-star
1910 – 32-89 (.360) in Cuban Play (2nd best in league; Pete Hill best at .365)
1911 -- .268 for NY Lincoln Giants (3 teammates listed with 37, 48, and 50 ab, so I estimate 12-45)
1911 – 22-86 (.258) in Cuban play; HR Johnson wins batting title with .410 clip; Hill is 3rd listed at .284.
1912 – 12-34 (.308) for NY Lincoln Giants and one game for Brooklyn Royal Giants
1912 – 0-4 vs. a Giants/Highlanders ML team
1912 – 55-151 (.364) in Cuban play; wins batting title
1913 – 39-130 (.300) for NY Lincoln Giants; teammates Johnson and Lloyd hit .382 and .325; Holway all star
1913 – 10-15 vs. Pete Alexander and Walter Johnson
1914 -- .481 for NY Lincoln Giants (no ab listed); Holway all star
1914 – 16-52 (.306) for Lincolns touring Cuba
1914 – 13-57 (.228) in Cuban play
1915 -- .255 for NY Lincoln Stars (4 teammates listed with 50, 50, 65, & 79 ab, so I estimate 14-55)
1915 – 5-13 vs. NY Giants and Philadelphia Phillies
1916 -- .185 for NY Lincoln Giants (no ab listed)
1916 – 6-15 in winter ball in Florida
1917 -- .265 for NY Lincoln Giants (no ab listed)
1917 – 4-12 vs. Joe Bush all-star team
1918 – Military Service, earned 5 battle stars and a Purple Heart
1919 -- .310 for AC Bacharach Giants (no ab listed)
1919 – 3-12 vs. a mix of ML competition
1920 – no ba listed in play for NY Lincoln Giants
1921 – no records listed
1922 – no ba listed in play for NY Lincoln Giants
1923 -- .248 for NY Lincoln Giants, but fuller records probably available, as this was first year of Eastern Colored League

Totals
102-340 (.300) vs. Black Teams, 1909-1913, 1915, 1916 winter ball
.297 mean avg. for 10 seasons as regular, 09-17, 19
138-435 (.317) in Cuban Play (including 4 seasons +1914 Lincolns’ tour)
22-56 (.393) vs. Major-League competition

262-831 (.315) grand total

Holway doesn’t give career totals for Poles

From Riley
Poles’ averages vs. all competition for NY Lincoln Giants, 1911-1914
.440, .398, .414, .487
lists him as batting over .400 in his career against all competition
hitting .319 in Cuba (his avg. without the Lincolns’ tour)
hitting .594 against major-league competition

I9s gives him the following projected mle career totals:


AB hits 2B 3B HR BB BA OBP SLG OPS RC
7390 2238 318 137 28 941 .303 .382 .394 776 1149
   2. Jeff M Posted: June 11, 2004 at 01:53 PM (#671140)
Those I9 totals look a lot like Bill Monroe's.
   3. karlmagnus Posted: June 11, 2004 at 02:33 PM (#671173)
Holway gives .268, Riley .440 in 1911 for the NY Lincoln Giants. Is there a reason for this discrepancy, or is the data really that uncertain?

I expected to vote for Poles, but the above data is a lot less impressive than I expected -- as Jeff M said, like Monroe's, who's off my ballot. How can he have hit over .400 against all competition, when the figures we have hover around .300?

Feeling my way VERY tentatively with Negro Leaguers, but Poles' figures above don't knock my socks off as I'd expected them to.
   4. karlmagnus Posted: June 11, 2004 at 02:37 PM (#671181)
Based on the ML equivalents, assuming roughly '10s park factors and environment (which may be a bad assumption) he'd be on my ballot in the 14-15 range.
   5. Michael Bass Posted: June 11, 2004 at 02:49 PM (#671201)
Can someone do a year by year breakdown like has been done with Foster and Hill in the past (Star, really good, above average, average, below average, etc.)

My gut for now is that he's not as high as Hill, more in the Monroe area of my ballot (which is, I think 11th or 12th at the moment), but I'm still kind of a blank slate.
   6. PhillyBooster Posted: June 11, 2004 at 03:11 PM (#671218)
My ballot has two centerfielders currently -- Pike and van Haltren. Is there any reason to believe that Poles was better than either of them, or that I should have three centerfielders on my ballot?

Of the four Negro League newcomers we're considering (Sanchez, Santop, Petway, Poles), I'd rank them pretty much in that order. Right now, Sanchez and Santop look to definitely make my ballot. Petway and Poles, I think I would need some more convincing.
   7. ronw Posted: June 11, 2004 at 03:20 PM (#671235)
Sanchez? I'm not sure who that is. Do you have anything for him?
   8. Chris Cobb Posted: June 11, 2004 at 03:26 PM (#671246)
Sanchez?

I'm guessing PhillyBooster means Jose Mendez.

Holway gives .268, Riley .440 in 1911 for the NY Lincoln Giants. Is there a reason for this discrepancy, or is the data really that uncertain?

There is a reason for it. Holway's numbers are based on data drawn exclusively from competition between black teams, based on extant box scores and other written accounts. Riley's numbers are based on all available data for the season, which includes large numbers of games against white semipro teams. Riley's figure may well be drawn from a contemporary source, in fact. In a few seasons, black teams happened to get pretty detailed coverage in the black media, and one team that seems to have been well-covered is the 1911-1914 NY Lincoln Giants, for which Poles played at this time.

Holway's numbers have a firmer basis in documentary evidence, but for this period they tend to be based on the records of only a few games, so they are subject to the problems of accuracy that accompany small sample sizes. Riley's numbers have a less-well-documented but much larger statistical base.

When I adjust Holway's numbers to get MLEs, I use a multiple that ranges from .85 to .90. When I adjust numbers from Riley that are based on play vs. "all levels of competition", I use a multiple that ranges from .70 to .75 . I also tend to apply a multiple of .95 to the i9s numbers since I think they tend to be a bit generous, but that varies from player to player.
   9. karlmagnus Posted: June 11, 2004 at 03:29 PM (#671250)
Thanks very much; he's solidifying around the bottom of my ballot, probably on it for 1929-33, then off when the glut hits.
   10. PhillyBooster Posted: June 11, 2004 at 05:49 PM (#671522)
Ugh. That's what happens when you're trying to decide whether to write "Santop" or "Mendez" first.

You end up righting "Sanchez".
   11. Chris Cobb Posted: June 11, 2004 at 06:05 PM (#671555)
My ballot has two centerfielders currently -- Pike and van Haltren. Is there any reason to believe that Poles was better than either of them, or that I should have three centerfielders on my ballot?

Well, it's not like we've been electing centerfielders by the bushelful . . . My current interpretation of the evidence doesn't put him above either Pike or Van Haltren, but I think it might be possible to argue that he is better than VH on the basis of his peak, which was clearly strong, but whose height I haven't tried to work out yet. More work needed to try to answer this question . . .
   12. Jeff M Posted: June 13, 2004 at 09:06 PM (#675129)
My ballot has two centerfielders currently -- Pike and van Haltren. Is there any reason to believe that Poles was better than either of them, or that I should have three centerfielders on my ballot?

I agree with Chris Cobb. I don't think there's much evidence Poles was better than Pike, but I could easily see him above Van Haltren. Then again, VH lives at #48 on my ballot, so if Poles is better, it doesn't mean much to me.
   13. dan b Posted: June 16, 2004 at 04:03 PM (#679789)
Is there any reason to put any stock whatsoever in the I9 projections? They look like optomistic conjecture to me.
   14. Chris Cobb Posted: June 17, 2004 at 05:11 AM (#681989)
Is there any reason to put any stock whatsoever in the I9 projections? They look like optomistic conjecture to me.

Of course they are conjecture: any assessment of the Negro-League players, especially prior to 1920, is necessarily partly conjectural, because the statistical record is so fragmentary.

Are they optimistic? What's the evidence? If you want others to take your criticism of i9s seriously, you need to offer us some sort of reasoned analysis here . . .

I happen to agree that they are _probably_ somewhat optimistic, but I still place a good deal of stock in them and find them helpful in evaluating negro league candidates, for several reasons.

1) They have made an effort systematically to translate Negro-League performance into major-league performance, so their numbers can be used to compare Negro-League players to one another. As long as they are consistent in their projection methods, their numbers can be useful in this respect.

2) Because their translations are systematic, they can be also be systematically adjusted by voters, according to their sense of the degree to which they are optimistic or pessimistic. I lean towards seeing them as overestimating offensive value by about 5%. I'm not sure that I am right and they are wrong, but that's where I'm at now in comparing their projections to mine. If anyone wants to discuss why 5%, I'd be happy to do so, but it's too late tonight for me to go into it.

3) Other expert assessments available suggest that they are not _wildly_ optimistic in their projections. In the case of Poles, their data suggests that Poles is a borderline HoM candidate. This agrees with the expert assessment of him in CPPD, in which about half of the experts indicated that he should be elected to the Hall of Fame, and half didn't. It also agrees with Bill James's ranking of him. As the #4 centerfielder, he might be worthy of election, and he might not.
   15. John (You Can Call Me Grandma) Murphy Posted: June 17, 2004 at 05:46 AM (#682033)
As the #4 centerfielder, he might be worthy of election, and he might not.

I'm not really sold on Poles. I'll be surprised if he makes my ballot when he's eligible, but you never know.
   16. KJOK Posted: June 18, 2004 at 02:37 AM (#684182)
Holway gives .268, Riley .440 in 1911 for the NY Lincoln Giants. Is there a reason for this discrepancy, or is the data really that uncertain?

Just wanted to add to Chris' excellent reply that yes, in many cases, the data is VERY uncertain.

Just for example, MacMillan, Riley and Holway all list JH Lloyd as hitting .564 in 1928. However, 40 of the 46 Lincoln Giant boxscores have been researched, and for those 40 games he hit .357, meaning he would have to hit around 1.000 in the missing 6 games to hit .564 - obviously, not likely.
   17. KJOK Posted: June 18, 2004 at 03:28 AM (#684448)
Now, to get into a philosophical discussion that I think is important.

While I'm a big proponent of "perfect" talent distribution across positions at the major league level over time (meaning that, while any given year may have more talent at SS than LF, over time the overall talent of the two positions will be roughly the same because managers will try to get the best 8 players on the field), I do NOT take this position with leagues below the major league level, such as Japanese Leagues, Minor Leagues, Negro Leagues, etc.

In THOSE leagues, the most talented offensive players will "drift" towards the most difficult defensive positions, because the talent distribution is such that these players are often the most talented OFFENSIVE AND DEFENSIVE players on their team.

What this means in terms of the HOM for Negro League players is that positions such as Catcher, Shortstop, and Centerfield will end up with more "great" players playing those positions than play 1B, LF, RF, or even 2B and 3B.

What this means in terms of evaulating players such as Poles is that being the "#5 CF" in the negro leagues would mean that he's a better player than the "#3 LF" in the Negro Leagues, or the #2 1st baseman, etc.
   18. Chris Cobb Posted: June 23, 2004 at 02:28 AM (#693331)
I've been working on negro-league outfielders for a few weeks. Here are a couple of posts that start to pull together the data I've compiled and analyze it a bit. I start with the i9s data because it's easy to work with.

I9s career data, Negro-League outfielders eligible 1929-34, plus Pete Hill
Name        Career   AB   hits 2B  3B  HR  BB    BA   OBP  SLG OPS  RC
Pete Hill   1901-20 9706  2986 519 214 101 1001 .308 .372 .436 809 1592
C Torriente 1913-28 8630  2768 567 188 170 592  .321 .364 .489 853 1511
Jules Thomas 1909-25 8451 2482 451 104 141 621  .294 .342 .422 764 1211
Spot Poles  1909-23 7390  2238 318 137  28 941  .303 .382 .394 776 1149
H McNair    1911-27 7340  2052 400 103  54 794  .280 .350 .384 734  975
C Blackwell 1916-26 5786  1652 202 135  18 618  .286 .354 .376 731  788
B Hall      1913-23 5565  1616 301 107  19 555  .290 .355 .393 748  776
G Shively   1914-24 5209  1509 242 140  11 347  .290 .334 .396 730  696
Jim Lyons   1911-23 5210  1445 222 142  17 353  .277 .323 .384 707  668
P Duncan    1909-20 5704  1516 163  72  11 638  .266 .340 .325 665  656


Comparison group, Major-league outfielders eligible 1928-35, less Cobb & Speaker (career win shares in parentheses to give a period-adjusted assessment)

Z Wheat (380) 09-27 9106  2884 476 172 132 650  .317 .367 .450 817 1497
M Carey (351) 10-29 9363  2665 419 159  70 1040 .285 .361 .386 747 1298
H Hooper(321) 09-25 8785  2466 389 160  75 1136 .281 .368 .387 755 1244
B Veach (265) 12-25 6656  2063 393 147  64 571  .310 .370 .442 712 1077
G Burns (290) 11-25 7241  2077 362 108  41 872  .287 .366 .384 750 1012
K Willi.(202) 15-29 4862  1552 285  77 196 566  .319 .393 .530 923 1018
L Tobin.(179) 14-27 6174  1906 294  99  64 508  .309 .364 .420 784  943
C Milan (266) 07-22 7359  2100 240 105  17 685  .285 .353 .353 706  906
B Jacob.(178) 15-27 5507  1714 328  94  83 355  .311 .357 .450 807  886
R Youngs(206) 17-29 4627  1491 236  93  42 550  .322 .399 .441 840  808
I Meusel(161) 18-27 4900  1521 250  93 106 269  .310 .348 .464 813  792
T Walker(164) 11-23 5067  1423 244  71 118 416  .281 .339 .427 766  730
A Strunk(174) 08-24 4999  1418 213  96  15 573  .284 .359 .374 732  673
B Shott.(161) 09-23 4945  1338 154  65   9 713  .271 .365 .333 698  600


Notes on Park & Period Factors – Lyons and Blackwell played a good deal in St. Louis, in a great hitters’ park. Torriente and Duncan played a good deal in Chicago, in a great pitchers’ park. McNair, Hall, and Blackwell did more of their hitting after 1920; Poles, Thomas, Lyons, Shively, and Duncan did more of their hitting before 1920; Torriente’s career was about evenly divided.


Expert notes
Ranked by James -- Torriente (#2 cf) Poles (#4 cf), Lyons (#5 cf), Blackwell (#9 cf), & McNair (#10 lf).

Support from the CPPD experts -- Torriente (96%), Poles (52%), Lyons (12%) have votes to elect.

Holway all-star awards -- Torriente (9), Shively (7), Thomas (6), and Hall (6) do better in Holway all-star awards than Poles (3), Lyons (2), McNair (2), and Blackwell (2). Poles may be shortchanged because Holway does not name all-stars for some of his early, and best, seasons.
   19. Chris Cobb Posted: June 23, 2004 at 02:37 AM (#693357)
Analysis

The three top Negro-League candidates in this group, by the i9s numbers, are Torriente, Poles, and Thomas. None of the rest are likely to be ballot-worthy, but they are worth a careful look when they become eligible to make sure of that. Here's the analysis tha tI have on each so far. I've given much more attention to Poles, since he is the first eligible.

Christobal Torriente – Pretty clearly a HoMer. Compares favorably with Pete Hill, elected on his first try, and Zack Wheat, who may well also be elected on his first try in 1933. Torriente won’t go on his first (or second) try, given the class of 1934, but he should be a leading candidate during the late 1930s.

Spotswood Poles – By the i9s numbers, he looks like a borderline candidate. By those numbers, he's between his very near contemporaries Veach and Burns offensively, although he's ahead of both defensively as he played a good centerfield to their left field. Poles has .138 RC/PA, Burns .125, Veach .149. Their eqas are .287 and .293 respectively, so by the i9s numbers we might peg Poles at .290, which, interestingly, is the EQA for another player who looks to be a close match for Poles: Fielder Jones. The next centerfielder up in my rankings from F. Jones is George Van Haltren. He wasn't the defender Jones was, but he was a significantly better hitter (.299 eqa). VH and Poles both had 15 seasons as regulars (if one gives Poles full credit for career length), which tops Jones' 13. It looks to me like Poles would land between VH and F. Jones in my rankings, if (1) the i9s numbers are to be taken at face value and (2) Poles' peak is typical in proportion to his career. If the i9s numbers look to be too high, he would drop down to Fielder Jones' level or below. If his peak was very high, he could climb higher in relation to Jones and Van Haltren, neither of whom had an outstanding peak. With the i9s career numbers and a great peak, he'd about match Van Haltren, as Hugh Duffy does in my system on the strength of his peak.

So my initial estimate of Poles' placement for 1929 places him anywhere from 8 (Van Haltren is now at 10) to about 25 (Fielder Jones is at 20).

It remains to examine his seasonal stats more closely to check the i9s estimates and construct an estimate of his peak value.

Jules Thomas – Might have been as good or better than Poles. Haven't had time to study him yet, but he will bear careful consideration in 1931.

Hurley McNair – Probably doesn’t match any of the major-league comparison set who have a chance at making some ballots (a group that ends with Veach and Burns, I think). He was a corner outfielder, so his defensive value probably wouldn't catch him up to Burns and Veach, whom he trails a bit offensively. He came into his own during the 1920s in a good though not great hitter’s park (as far as I know), so there’s no reason to see him as having more value than his RC ranking suggests.

Charlie Blackwell – Similar to McNair in my assessment. More defensive and baserunning value as a centerfielder, but not enough to get him into the ballot picture. Played in a great hitters’ park in a great hitter’s era, for the most part.

George Shively – Better, in my opinion, than I9s stats suggest. Played from 1910-13 on the West Baden Sprudels under C. I. Taylor. Riley and Holway count the Sprudels during these years as a major-team, but i9s doesn’t include them in Shively’s career. The Sprudels weren’t a _good_ team, but they were playing Chi American Giants & other top teams, and Shively had very good seasons in 12 & 13. If we add, say, 240 RC to his I9s total, that moves him to 936. Throw in the fact that he put up these numbers primarily in deadball seasons, was known as one of the fastest men in baseball, and played center field, and he begins to look like a borderline candidate, possibly next in value after Poles and Thomas.

Jimmy Lyons – Shively’s exact contemporary, he has a stronger reputation (as evidenced by his ranking by Bill James and his garnering some support in the CPPD experts poll), but his numbers aren’t as strong. Neither I9s nor the Holway all-star rankings like him as much as Shively, and I9s gives him more seasons than it gives to Rabbit Shively, for less value. I suspect Lyons’ reputation benefits from the fact that he put together a couple of great seasons for the Chicago American Giants during the first two years of the Negro National League, where Shively did his best work for the ABCs prior to the formation of the league. While the ABCs don’t have the glamour of the American Giants, they were covered extensively in the _Indianapolis Freeman_, so Shively’s statistical advantages may be partly attributable to better documentation. When extant numbers and reputation don’t square, I tend to favor the numbers, but Lyons has an argument to rank ahead of Shively, if values reputation over fragmentary data.

Blainey Hall – Shortest career on the board, but the man could hit.

Duncan -- Probably somewhat better than I9s numbers make him out to be because he spent a good deal of time on the Chicago American Giants, playing in an extreme pitchers’ park, but still he doesn't look like a serious candidate.

My current ranking of the negro-league outfielders:

Torriente
Hill
Poles
Thomas
Shively
McNair
Blackwell
Hall
Lyons
Duncan

My current ranking of the best of negro-league outfielders against their immediate major-league contemporaries:

Cobb
Speaker
Wheat
Torriente
Hill
Carey
Poles
Thomas
Burns
Shively
Hooper
Veach
Milan
McNair
Williams
   20. Jeff M Posted: June 23, 2004 at 04:15 AM (#693526)
Since we don't know exactly where league quality lies for the negro leagues, I thought I'd look at some measure of relativity, albeit imperfect.

For the Negro League outfielders posted above, they have an RC/27 of about 5.55 (simply using the RC number posted and AB-H for outs). Poles is at 6.05, or about 10% better than the average guy in that set of players. Obviously that set of players is better than league average, but my point is to measure Poles against other quality players (b/c there ain't much else to go on).

Here are the other three main guys:

Hill (6.39, or about 15% better)
Torriente (6.95, or about 25% better)
Thomas (5.47, or about average)

Poles has an OPS+ of about 99 among this group if you add up the raw stats and calculate OBP and SLG from that -- in other words, it's weighted by PA and BA, respectively. If you don't weight them (i.e., add up the OBP and divide by 10, and do the same for SLG), his OPS+ is about 107.

Here are the other three:

Hill 106 (weighted) and 115 (unweighted)
Torriente 115 (weighted) and 126 (unweighted)
Thomas 94 (weighted) and 96 (unweighted)

Obviously none of that is park adjusted.

For comparison's sake, let's do the same for the major leaguers on the list, except we have park-adjusted numbers for them, so I won't calculate RC/27 or OPS+ myself. I don't know if this will be a valid comparison, because I suspect the two groups are not of equal quality. Most of the major leaguers were quite a bit above average, and my gut tells me the major league group was probably better.

I was going to get the average RC/27 for that group by weighting their career RC/27 by PA, which produced an average career RC/27 of 5.48. But there were two leagues involved, and the league average RC/27 was all over the place, so I thought that might skew the numbers.

Instead, I took their RC/27 relative to league average from the Major League Handbook, weighted it by PAs, and came up with a average relative RC/27 of 1.275 for the group (which is quite good).

Comparing the relative RC/27 for each ML outfielder who made Chris' final list against the average relative RC/27 for that group:

Wheat (1.44, or about 13% better than the group)
Carey (1.22, or about 5% worse)
Hooper (1.24, or about 3% worse)
Veach (1.35, or about 6% better)
Burns (1.29, or about average)
Williams (1.47, or about 15% better)
Milan (1.21, or about 5% worse)

Best on the list was Young, who was about 16% better than the group by this measure, and worst on the list was Strunk, who was about 9% worse than the group by this measure.

Doing the same for OPS+ (using calcs from Baseball Reference, the weighted average in the group is 116.9 and the unweighted average is 117.3, so there's no practical difference):

Wheat 110
Carey 92
Hooper 98
Veach 109
Burns 98
Williams 117
Milan 93

Best on the entire list was Williams, and worst on the entire list was Carey.

Eyeballing this data, and assuming the major leaguers as a group were a little better than the negro leaguers as a group, I'd put Thomas near the bottom of these outfielders before making some adjustments if necessary for length of career. I'd put Poles somewhere in the middle of the pack before making some adjustments if necessary for length of career.

*Final note, this isn't designed to refute Chris' rankings, because I think it is measuring something different. I'm trying to get a sense of where the Negro Leaguers ranked among other good players at the same position at the same time (lacking league average data). Then comparing that to where the major leaguers ranked among other good players at the same position at the same time. And I'm using RC/27 and OPS+ to do it.

Chris is ranking players by comparing the career totals of Negro Leaguers with the Major Leaguers (and I suspect making some other subjective adjustments). I think you have to discount those I9 totals. I've been discounting them 10% (but not in the data above).
   21. Chris Cobb Posted: June 23, 2004 at 04:35 AM (#693547)
Interesting stuff, Jeff! Your in-group analysis is very helpful. My sense, though, is that the transition to the lively ball would need to be dealt with somehow in working with the MLEs i9s provides for the negro-leaguers, so I'd want to make some adjustment for that prior to calculating OPS+ or RC+ within the group. Without that adjustment, I think Poles and Thomas will be underrated by it.

I think you have to discount those I9 totals. I've been discounting them 10% (but not in the data above).

FWIW, for quick comparison purposes I use a 5% discount on the i9s totals. But I try to work out my own estimate of a player's quality independently, then prorate the i9s estimates accordingly to get a sense of what full stats for the player would look like.
   22. sunnyday2 Posted: June 23, 2004 at 03:13 PM (#693826)
The discussion above would seem to support what I have long believed to be the case: At his peak, Spot Poles was every bit the offensive player that Pete Hill was (107-106 advantage Poles for in-group OPS+). And Poles, I believe, was virtually 100% a CFer, while Hill was (what?) 55-65-75% CF and played some LF.

Now that's just the peak, but for a peak voter (like me) that's a big part of the story. Substantially equal.

PRIME--I'll give Hill a slightly longer prime, maybe 10-11 years versus Poles' 9(ish), AND I'll give him a slightly higher level of play for those "shoulder" seasons--that is, non-peak prime years. Advantage Hill.

CAREER--Hill played 25 years but my general sense is that 5 of those were as a player-manager-pinch hitter-reserve, so say 20 years as a regular, to Poles' 14 years regular out of a 15 year career. That's a big edge to Hill for career voters. For me, the career is only about 10-15% of the rating, since most of a player's value is already accounted for in his prime years.

Anyway, add it up, and for a peak voter who had Hill rated #2-3-4-5 (you know who you are) I can't see how Poles could be any lower than 8-10. If Chris is more of a career type voter, then maybe #25 is reasonable (actually it feels a little harsh to me for a guy who had Hill #1).

As for me, I had Hill at #15. As a peak/prime voter, I see Poles as just slightly below Pete. Since the ballot has weakened a bit since then, I would see Poles at #10-15 but definitely on-ballot.


But again, peak voters who had Hill near the top of your ballot, how can Poles not be top 10?
   23. Chris Cobb Posted: June 23, 2004 at 05:10 PM (#694170)
Sunnyday2 -- good to see you back!

Spots Poles conversions

My study of players who played in the Negro League in the 1940s and who then played several full seasons in the majors subsequently indicates that, at that time, Negro-League averages should, on average, be prorated by a multiple of .85-.90 . My experience comparing averages vs. "all competition" to averages vs. black teams for players 1900-1920 suggests that they be converted by a multiple of about .85. Assuming the level of competition among black teams vs. major-league teams remained relatively constant 1910-1950, a prorating multiple between .7 and .75 seems in general to be appropriate when converting a black player's average vs. all competition. One might raise or lower it by a couple percentage points beyond this range, but this is what seems to me to be the "ballpark" for conversions.

Riley shows Poles with a mean avg. of .435 in 1911-1914 vs. all competition.

Holway shows Poles with a mean avg. of .340 in 1911-1914 vs. black teams; this mean average is based on a small number of at bats.

Here are three reasonable conversion options

1) Cautious
Prorate .435 by .7 -->.304
Prorate .340 by .85 --> .289

2) Generous
Prorate .435 by .75 --> .326
Prorate .340 by .90 --> .306

3) Cautious with "Greatness ceiling" modification
Prorate .435 by _difference_ between ML league avg. and avg. needed to convert to league avg. with a prorating of .7 (Need to hit .371 to hit .260 when prorated by .7. Difference between .371 and .260 = .111). This method follows the "productivity ceiling" view of great players in weak leagues.
Anway, prorated this way, .435 --> .324.

Use same "difference" process to prorate Poles' .340 avg. vs. black teams from a .85 multiple --> .294

4) Generous with "Greatness ceiling" modification
Prorate .435 --> .348
Prorate .340 --> .311

5) I9s has Poles at .323 for those 4 years.

Median of these 9 estimates is .311
Mean of these 9 estimates is .314

Reduce Poles in i9s by 5% --> .307
This prorates .435 by .7
This prorates .340 by .9

Reduce Poles in i9s by 10% --> .290
That prorates .435 by .67
That prorates .340 by .85

In the context of his black contemporaries, Poles was one of the top 4 stars in black baseball, the other three being Pop Lloyd, Louis Santop, and Smokey Joe Williams.

Direct comparisons to major-leaguers
If Poles hits .290, 1911-1914, he compares well as a hitter to Fred Snodgrass, 1910-1913. He would have been an above average major-league outfielder, but not an all-star, unless his defense were truly sensational. He would earn appx. 72 batting win shares, for an average of 18 per season.

If Poles hits .307, 1911-1914, he compares well as a hitter to Clyde Milan, 1911-1914. Milan is Poles' match for speed and plate discipline, but Poles probably had a shade more power. He would be among the top 10 players in the AL, but not among the top 3 outfielders, thanks to Cobb, Speaker, and Jackson. He would earn appx. 92 bws, for an avg. of 23 per season

If Poles hits .323, 1911-1914, there's no easy comparison to be made, because there isn't a hitter of his type -- high average, above average plate discipline, average power, great speed who played at quite that level during this period. He would remain below the top offensive echelon of Cobb, Speaker, Collins, Baker, and Jackson, but he would probably have been among the top 10 players in the game. This estimate is shakier than the preceding two, but he would earn appx. 112 bws, for an average of 28 per season.

To complete the picture, I see Poles averaging between 6 and 7 win shares per season during his 1911-1914 offensive peak for his fielding.


Summary Conclusions:

The i9s MLEs are at the high end of reasonable estimates, but they are reasonable. They certainly match his reputation, and they are borne out by his batting averages from Riley. They are not borne out, however, by his batting averages from Holway. This is, for me, the major obstacle in accepting the i9s numbers at face value.

A 10% reduction to the i9s numbers is the low end of reasonable estimates. A cautious translation of Holway's averages matches this treatment of the i9s. This reduction does not match his reputation, however, and it is not borne out by his batting averages from Riley. This, for me, is the major objection to a highly cautious treatment of the i9s MLEs.

My evaluation of these various translations leads me to believe that a 5% reduction of i9s values remains a reasonable practice, though accepting the numbers at face value is also defensible in the case of Poles. I think I will use something between their numbers and a 5% reduction for my final estimate in his case.
   24. Jeff M Posted: June 23, 2004 at 07:41 PM (#694471)
That's good empirical work, Chris. As I mentioned, I use a 10% discount for the i9 numbers. You've convinced me that 10% is the low end of reasonable with respect to the MLEs per se.

But I still feel like a 5% discount needs to be tweaked a bit more to account for some intangible things, like the risks/uncertainty of all the data. Plus, top-to-bottom, the negro leagues at this time were probably not as strong as the major leagues so there ought to be a league quality factor in there too.

That doesn't mean the good Negro League players couldn't play and star in the majors. It just means their numbers might look a lot better playing against weaker competition, on average.

On the one hand, you might say that comparing their performances against major league teams negates that conclusion, and is already factored into the 5%. On the other hand, we aren't talking about a large sampling when we look at stats against major league teams, so it too has an unreliability factor.
   25. Chris Cobb Posted: June 23, 2004 at 09:41 PM (#694616)
But I still feel like a 5% discount needs to be tweaked a bit more to account for some intangible things.

I agree.

top-to-bottom, the negro leagues at this time were probably not as strong as the major leagues so there ought to be a league quality factor in there too.

League quality is the justification for the pro-rating multiple. i9s MLEs are already using a league quality factor -- otherwise they'd have Poles hitting something between .340 and .435 for 1911-1914 instead of .323 . The question is whether they are employing a sufficient league-quality deduction or not.
   26. dan b Posted: June 24, 2004 at 02:58 AM (#695593)
I will use the case for/against Spottswood Poles to express my views on the different resources we have available to evaluate the Negro League candidates.

Riley’s Biographical Encylcopedia is probably the one resource book to have on the Negro Leagues if you only have one. From this book in addition to a good look at Poles’ career and post-career life, we get eyewitness testimonials such as John McGraw listing Poles, John Henry Lloyd, Smokey Joe Williams and Dick Redding as “the 4 black players he would pick for the major leagues.” If our knowledge of the Negro Leagues were limited to the contents of this book, IMO we would elect Poles. Similarly, if our knowledge of 19th century baseball were limited to this kind of information, we would have elected Ed Williamson (named the greatest player of all-time in an 1894 poll) by 1901.

McNeil’s “Cool Papas and Double Duties” makes lousy reading, but his findings are interesting and worthwhile. McNeil’s goal is to identify up to 43 Negro League players for the HOF. He begins by accepting the initial 16 players selected and then seeks out 27 more by polling both former players and experts. He settles on a list of 20 players, one of which is Poles. If our consideration set had been limited to the 36 players he would ultimately enshrine, Poles would be our 2nd black player selected (after Pete Hill.) Grant Johnson and Frank Grant received consideration, but did not make his list.

Holway’s “The Complete Book of Baseball’s Negro Leagues” provides a year by year run down of the pennant races and post season exhibitions with major league competition, together with useful compilations of all-star teams and MVP selections. His All-Century team has a rose colored glasses look at the Negro leagues with 14 of 51 players named having been Negro Leaguers. Holway’s placement of Biz Mackey as 2nd best catcher over Berra and Bench, and Dandridge as the 2nd best 3B over Mathews and Brett are questionable picks. Poles is not one of the 7 OF he chooses for his all-time Negro League all-star team. Holway participated in McNeil’s poll of experts and did not include Poles on his list of 30 players.

IMO, Bill James’ placement of 12 Negro League players in his rankings of the top 100 players of all-time is well reasoned and superior to Holway’s All-Century team. James has Poles as the 4th best CF in the Negro Leagues behind 3 players he ranks as the 4th, 67th and 76th best players of all-time. While justifying his placement of 12 Negro Leaguers in his to 100, he writes “In terms of getting an accurate list, I am more bothered by the exclusions of Bullet Rogan, Spottswood Poles, and Willard Brown than by dropping Harry Heilmann, Ralph Kiner and Early Wynn.” If James is seriously stating that Poles is close to his top 100, then Poles would be at the top of his HoM ballot for 1929 – I don’t see anyone else on the ballot close to a top 100 ranking. At the least, his preference of Poles over Heilmann and Kiner, who rank 16th and 18th at their respective positions, would place him near the top of his ballot.

Earlier on this thread, I took a jab at the i9’s projections for which Chris Cobb rightly chastised me. From i9’s home page “Beginning with the 1901 season, we will move through the first 50 years of modern Major League history at a rate of two to three seasons per year. By the time it is over, sometime in 2019 perhaps, we will have carved out an interesting slice of alternate reality and together shared a significant and important historical exercise.” I’m sorry, but i9’s is 13 guys playing a game of Strat-O-Matic. I enjoy games like Strat and Diamond Mind as much as the next guy, but when I play, I refrain from calling it “significant and important”. If that were the case, get out your history books and make note of the “interesting slice of alternate reality” I just generated by letting Diamond Mind replay the 1960 season:
-Fellow Pirate fans, we can take down our vintage “Beat ‘Em Bucs” bumper stickers and put our Mazeroski bobble heads in a less revered place, because the dramatic home run of October 13, 1960 was hit by Joe Adcock – a 3-run shot in the bottom of the eighth to give the Braves a come from behind 7th game win over the Orioles. (Of course, we wouldn’t have cared anyways since i9’s moved our Bucs to Birmingham in 1901 in their alternate reality.)
-Those of us who consider Mike Schmidt the greatest 3B of all-time might want to take another look at Eddie Mathews after his MVP year where he made a run at the Babe’s hallowed 60, falling short with 59.

This “interesting slice of alternate reality” was generated by a game where all of the players were based on actual performance, not somebody’s guess at what Grant Johnson might have done. Further, with their ambitious plan to play the game until 2019, I expect our project will be long completed before we learn their guesses at what kind of stats Biz Mackey or Willard Brown would have posted, thus limiting their usefulness down the road. ***Old Cliché Alert*** I may be throwing out the baby with the bath water, but I have credibility issues with their project and choose not to take i9’s and their self-proclaimed importance with as many grains of salt as the other sources I have mentioned. Chris Cobb’s analysis of i9’s shows Poles to be a borderline candidate.

Conclusion – James and McNeil would put Poles in the HoM, at or near the top of this year’s ballot. John McGraw might agree, although I am sure he personally feels slighted by our process. Holway says no and i9’s isn’t sure. Maybe with Grant, Johnson and Hill already in, we have our fill of pioneer era Negro Leaguers in the HoM. Having passed on Johnson and Grant, my PhoM will agree with James and McNeil and take Spottswood Poles instead.
   27. KJOK Posted: June 24, 2004 at 03:48 AM (#695628)
Danb:

Nice post. However, I think you are misinterpreting the I9 numbers. The players are "created" to be used in the Stratomatic game, but the numbers are NOT based on Stratomatic results. They are Major League Equivalent projections for the given years, unless I totally misunderstood my email exchange with Dan Levine from over a year ago....
   28. dan b Posted: June 24, 2004 at 04:16 PM (#696077)
Thanks, KJOK. I do recognize that stats need to be created before the game cards can be made and that the numbers we have on the i9’s site are not game results, but the initial projections used to create the game cards. My point is that if the laws of probability allow a Diamond Mind result where Ed Mathews hits 20 HR more than actual, then how can the i9’s home page claim that their results from playing the game will be “a significant and important historical exercise”? Moving on, I am sure that had I taken the time and effort to create the Strat cards that Dan Levine has made, I would want to play the game as well. I just hope I would refrain from expecting significant results.

Note that I am not rejecting the i9’s MLEs, I am just not putting them on equal footing with the other resources available.

Since the website only gives numbers for players who started play prior to 1919, perhaps someone who is communicating with Levine could find out if numbers will be available for the rest of the Negro League players in time for our purposes.
   29. Jeff M Posted: June 24, 2004 at 04:36 PM (#696136)
...then how can the i9’s home page claim that their results from playing the game will be “a significant and important historical exercise”?

Dan, I share your concerns about how much weight to give the I9 numbers that were developed as a backdrop to the Strat playing cards.

But in their defense, I don't think they intend the results to be of historical significance to outsiders. I think they consider playing the game in an integrated way as a "significant and important historical exercise" for the participants. I think the quoted statement from the I9 site is addressed to members who will be playing in the Strat league, rather than to the world. The paragraph before it says:

"Explore. Share. Compete. Learn. Contribute. Grow. These might sound like far-reaching goals for participants in a Strat-O-Matic League, but that is what we hope all of our fellow members will share in doing over the life of this league." (my emphasis in italics)

I wonder if the I9 folks would be surprised that outsiders were even looking at their numbers.

I don't know if you've ever played in a Strat league or similar, but I recommend it. You DO learn a lot about the players...and lots of players you never paid attention to before, which arguably makes it an important and historical exercise for you as a member of the league.

One year I came to love Reb Russell because he NEVER allowed an inherited runner to score. I had never heard of Reb Russell before that. Admittedly, I probably view him better than he actually was because he was a dynamo for me. :)

I agree, however, that the results of such game playing are not themselves of "important or historical" significance.
   30. KJOK Posted: June 24, 2004 at 11:38 PM (#697055)
I'm posting my Spotwood Poles MLE's to the HOM yahoo group. I used the same method I've been using for Pete Hill, etc. as my "new" method is not yet perfected...
   31. Jeff M Posted: June 30, 2004 at 02:37 AM (#706061)
Using KJOK's MLE's, here are the Sim Scores for Poles with a positional adjustment:

CRUZ, JOSE (726)
GRIFFEY, KEN (SR.) (722)
MAGEE, SHERRY (721)
SLAUGHTER, ENOS* (717)
MINOSO, MINNIE (709)
OTIS, AMOS (708)
SHECKARD, JIMMY (705)
CEDENO, CESAR (699)
RAINES, TIM (695)
DAVIS, TOMMY (694)
   32. Jeff M Posted: June 30, 2004 at 04:14 AM (#706158)
Here are Poles' batting WS using the old TangoTiger formula to convert Linear Weights to Win Shares for batting. The formula is:

Batting WS = ((LWTS Runs/10) + (Outs/100))*3

As someone (David?) pointed out, you may want to use something other than 10 as the divisor, but I've left it at 10 b/c that's what I used for Hill before David posted his comment -- you can apply your own discount, as I do. If you have another factor you'd like to use, let me know and I can plug that in pretty quickly.

So here are Poles' batting WS using (1) TangoTiger's converter, (2) the Integrated 9s MLEs and (3) Pete Palmer's LWTS. They are raw (no adjustments for season-length, park factors, league quality, uncertainty of data, etc.):

1909.....10.0
1910.....12.8
1911.....20.3
1912.....24.0
1913.....21.8
1914.....25.5
1915.....21.0
1916.....19.9
1917.....23.6
1919.....15.4
1920.....25.9
1921.....17.6
1922.....11.5
1923.....7.9
Total....257.4

For defense: based on the Integrated 9s MLEs, Poles had 8,331 plate appearances. Assuming approximately 4.1 PAs per game, that gives Poles about 2,032 games played. Assuming all of his games were played at center field (which may or may not be accurate), then at about 8.75 innings per game played, he would have approximately 17,780 innings in CF.

Looking at the WS per 1,000 innings played of Hall of Fame center fielders from the majors, a per 1,000 rate of 4.20 WS for Poles would make him a clear "A" defensive player among those HOFers (comparable to Mays and Dimaggio, not as good as Carey or Speaker and better than Ashburn).

So if you were to give Poles 4.20 WS/1000, you'd get another 74.7 fielding WS to add to his hitting totals, which would put him at about 332 WS.

Just numbers...make adjustments as appropriate for your systems. Or ignore them altogether. :)
   33. Jeff M Posted: June 30, 2004 at 07:01 AM (#706227)
Tried to post this on the general Negro Leagues thread, but it didn't pop up in the sidebar, so I'll put it here too.

There have been some discussions about the methodology of the i9 numbers. You can view the i9 "Projection Notes" here (warning, not a lot of detail):

http://www.i9s.org/i9s.cfm?a=cols&cols_id=6

In particular, here's a quote from that page:

"Note that, from the beginning, our goal is not to claim we somehow know what these players would have done. If, at the end of the day, we can look at the projections and say that we have gotten the shape of their careers correct, that, yes, Pete Hill and HR Johnson were great players, regardless of whether we are correct that Hill hit .322 in 1907, that, at the close of day, Johnson was a legitimate MVP candidate a few times in his career and a borderline HoF candidate, that will be success."
   34. yest Posted: June 30, 2004 at 03:34 PM (#706480)
Spots Poles was called the black Ty Cobb
Had a 610 average in exhibition games against major league competition (25 hits in 41 at-bats)

In the famous 1952 Pittsburgh Courier Poll he made the 4th team

In 1911 he hit 440 for and stole 41 bases in 60 games.
in 1912 he hit 398
in 1913 he hit 414
in 1914 he hit 487

was clocked one time running the 100-yard dash in under ten seconds.

In a game against Grover Cleveland Alexander he rapped thee straight hits against the rangy Hall of Famer

John McGraw said that Poles, John Henry Lloyd, Cannonball Redding and Smoky Joe Williams would be the four black players he would have picked for the major leagues had not the color line been so firmly entrenched.
actor/athlete Paul Robeson believed Spotswood Poles along with Jesse Owens, Joe Louis and Jack Johnson were the greatest black athletes of that time.
   35. Jeff M Posted: June 30, 2004 at 08:00 PM (#707049)
Yest. Source?

I like that: "the rangy Hall of Famer." :)
   36. yest Posted: June 30, 2004 at 08:09 PM (#707080)
to which part it?
   37. Jeff M Posted: June 30, 2004 at 08:39 PM (#707172)
Any. All.
   38. yest Posted: June 30, 2004 at 09:43 PM (#707398)
From thediamondangle.com/sitt/poles

was clocked one time running the 100-yard dash in under ten seconds.
John McGraw said that Poles, John Henry Lloyd, Cannonball Redding and Smoky Joe Williams would be the four black players he would have picked for the major leagues had not the color line been so firmly entrenched.
actor/athlete Paul Robeson believed Spotswood Poles along with Jesse Owens, Joe Louis and Jack Johnson were the greatest black athletes of that time.

From nlbpa.com/poles__spottswood

From geocities.com/Colosseum/Arena/6925/black
In the famous 1952 Pittsburgh Courier Poll he made the 4th team

From thediamondangle.com/sitt/poles.
In a game against Grover Cleveland Alexander he rapped thee straight hits against the rangy Hall of Famer

From the Baseballlibrary.com
Spots Poles was called the black Ty Cobb
Had a 610 average in exhibition games against major league competition (25 hits in 41 at-bats)
In 1911 he hit 440 for and stole 41 bases in 60 games.
in 1912 he hit 398
in 1913 he hit 414
in 1914 he hit 487
   39. Devin has a deep burning passion for fuzzy socks Posted: July 05, 2004 at 04:43 AM (#717169)
Other facts from the NBJHA: James has Poles as the "Best Negro League Player" for 1914 to 1916. He also states that Poles hit .610 (25 for 41) in documented exhibitions against major league players, but I don't know what his source is.
   40. richie allen Posted: July 17, 2004 at 11:26 PM (#739367)
Since i9 seems to have entered the collective conscience here, it's probably worth alerting you to a similar project.

A few of us play a game at stats, inc, which was formerly known as Bill James Classic Baseball. It's a historical simulation, and Stats (formerly in conjunction with James) includes negro league players. Most of the names you've discussed are featured in the game.

Now, the interesting (ish) bit: a website has compiled the results from all the stats leagues, and you can see exactly how Bill James and now the staff at stats inc have "coded" the players:

http://www.fish.com/baseball/league-stats2003/

Stats has normalised data for all major leaguers, which is why the numbers may initially look a little off. Anyway, as we can see fro the CF page at the above link, Poles is "coded" to play first class D, and in leagues has hit a combined .282/.329/.386. This is actually pretty good in the game, but not HOF calibre... similar hitters would be Lyman Bostock and Ginger Beaumont. Now, given Poles' outstanding D in the game he's pretty popular, but nowhere near elite level.

Just thought this might interest you.

cheers
Rich
   41. richie allen Posted: July 17, 2004 at 11:28 PM (#739370)
Further information on the game (which uses Diamond Mind as an engine) can be found at

www.diamondlegends.com
   42. KJOK Posted: July 20, 2004 at 12:00 AM (#742037)
you can see exactly how Bill James and now the staff at stats inc have "coded" the players:

You can see the results of the players, but not how they were actually "coded" I don't believe..
   43. sunnyday2 Posted: November 12, 2004 at 12:53 AM (#962493)
hot topics
   44. sunnyday2 Posted: June 27, 2005 at 11:35 PM (#1435226)
Just received the new National Pastime and there's an article about elite baseball players buried at Arlington National Cemetery. I was surprised that there are only 16 ML players and that the biggest name among them is Lu Blue or maybe Doc Lavan.

However, there are 2 NeLers and they are both arguably bigger names--Spot Poles and Ernest "Boojum" Wilson.

Poles "enlisted in the Army in 1917 at the age of 30 and served in the 369th infantry. Poles received a purple heart in France and left the Army at the rank of sergeant. He returned to his baseball career after the Army but retired inthe early 1920s and went into private business."

Anybody giving WWI credit ought to take a fresh look at Spot.
   45. Dr. Chaleeko Posted: July 29, 2005 at 05:09 PM (#1507736)
SPOTS POLES RECONSIDERED

Since I'm SOBF, I decided I'd take a little time to re-consider him. I decided to rerun his numbers, building OPS+ and WS totals for him.

I used I9s' numbers, shaving 5% off of his rates per the usual I9's discount. Per Chris Cobb's standard, I used the AL in odd years and the NL in even years as the basis for my comparisons.

First I figured his OPS+
YEAR obp slg ops lgobp lgslg obp+ slg+ ops+
-------------------------------------------
1909 331 346 678   318   326 104  106  111
1910 317 336 653   333   344  95   98   93
1911 362 394 756   354   377 102  105  107
1912 405 427 832   345   375 117  114  131
1913 379 379 758   335   346 113  110  123
1914 398 396 794   329   349 121  114  134
1915 370 382 752   340   342 109  112  121
1916 360 380 740   316   344 114  110  124 
1917 382 395 777   326   330 117  120  137
1918 375 388 763   319   337 118  115  133
1919 351 351 702   340   368 103   96   99
1920 408 428 836   334   372 122  115  137
1921 361 370 731   364   419  99   88   88
1922 315 342 658   358   417  88   82   70
1923 293 275 568   357   397  82   69   51
===========================================
     368 381 749   337   361 109  106  115

I believe I9s is using the AL's shortened 1919 schedule, so I adjusted it upward to a full season, and I also credited Poles year missed to war by averaging the surrounding four seasons (only after prorating 1919). Finally, I shortened up his last couple of years where I9s does not. I did this because, frankly, he was so crappy I didn't think he'd get all the PAs they thought he would.

Then I used the OPS+-comp method to match Poles to players in the majors and use their WS rates to generate his BWS. For FWS, I used the method described earlier in this thread. Here's how it came out:
YEAR   BWS   FWS   WS
-----------------------
1909   13.6  3.4  17.1
1910    5.2  4.9  10.1
1911   11.9  5.7  17.6
1912   19.2  5.3  24.5
1913   21.8  5.7  27.5
1914   25.9  6.4  32.2
1915   22.9  6.0  28.8
1916   20.1  5.6  25.6
1917   21.3  6.0  27.3
1918   20.7  5.8  26.5
1919   13.6  5.7  19.3
1920   24.4  5.8  30.2
1921    8.1  5.4  13.6
1922    3.8  3.9   7.7
1923    0.1  1.5   1.6
======================
total 232.7 77.1 309.7

By this analysis, Poles out looking a little less lustrous than before. He's better than Milan, but is he better than Burns, Veach, and company? I've recently downgraded him on my ballot, but I've always had him higher than Averill, so it's worth asking whether I think Poles is still superior after refiguring his numbers.

The short answer is that I now see Poles as behind Averill (with EA getting MiL credit), but it's close. Averill has small advantages in 3 and 5 year peak, and then he opens a little bigger lead in 10-year prime. Poles closes that gap by the end of their respective careers, but it's not really enough to make a difference.

It's possible that when I pHOMed Poles he was the best position player available to me (him or GJB), but his stock has slipped considerably since then in comparison to who's been added to the backlog.

He will not appear on my 1957 ballot and probably won't resurface for some time.
   46. Joey Numbaz (Scruff) Posted: October 19, 2005 at 02:12 AM (#1691894)
Thanks for that summary Doc. I figured I should reconsider him and seeing your quick summary was enough for me to realize why he'd fallen of my map so to speak . . . I don't see him popping back into my consideration set.
   47. sunnyday2 Posted: October 19, 2005 at 03:36 AM (#1692010)
I think we've progressed way beyond using I9 - 5%, haven't we? I mean, everyone agreed early on that I9 had some gigantic extrapolations there. As I recall, Poles hit about .310 in the NeLs with very little power and no walks. I take his 310 WS with the same grain of salt as Cool Papa's 440 or whatever it is.
   48. Gary A Posted: October 19, 2005 at 04:28 AM (#1692063)
As I recall, Poles hit about .310 in the NeLs with very little power and no walks.

I don't know of any published data on Poles's walk tendencies. In 1916, I've got six walks and one home run in eight games (.273/.385/.364); in 1921 three walks and one home run in nine games (.390/.432/.488). So that's 9 walks and 2 homers in 17 games, FWIW.
   49. sunnyday2 Posted: October 19, 2005 at 12:19 PM (#1692266)
Here was what I was remembering from above:

>Totals
>102-340 (.300) vs. Black Teams, 1909-1913, 1915, 1916 winter ball
>.297 mean avg. for 10 seasons as regular, 09-17, 19
>138-435 (.317) in Cuban Play (including 4 seasons +1914 Lincolns’ tour)
>22-56 (.393) vs. Major-League competition

>262-831 (.315) grand total

And that is with his big .400 year included.

I guess I should add, this is acc. to Holway, who had no data e.g. for 1921.

I stand corrected re. his lack of walks or power. I should have said that the data is ambiguous but that we would have to make some assumptions in order for him to be a big hitter.
   50. Dr. Chaleeko Posted: October 19, 2005 at 01:16 PM (#1692314)
Sunny,

I don't think that Poles and candidates of that period were evaluated by any other means than the adjusted I9s method. I could be wrong but I seem to recollect that being the case. That's why I went with the I9s. Maybe Chris C. can tell us if there's a better way for us to work through Poles's career.
   51. John (You Can Call Me Grandma) Murphy Posted: October 19, 2005 at 01:50 PM (#1692358)
I don't think that Poles and candidates of that period were evaluated by any other means than the adjusted I9s method. I could be wrong but I seem to recollect that being the case.

You're correct, Eric.

BTW, Poles doesn't look like he'll ever make my ballot.
   52. Chris Cobb Posted: October 19, 2005 at 03:38 PM (#1692544)
The problem with Poles, as Sunnyday2's reposting of the collected data from Holway shows,
is that there's too little data for the translation and projections methods we've developed for
post-1920 Negro-League data to be applied.

I9s has developed some sort of translation and projection method for these years that seems
plausible, so their projections provided a handy interpretation of the careers of pre-1920
black players. I did a fair amount of cross-checking of their projections against the available
data, and it was that cross-checking that led me to adopt the 5% discount when working
with their numbers.

I think that the actual NeL data suggests a player who is just a bit better than George J. Burns,
and the I9s projections, with a 5% discount, show us a plausible career construction that looks
about like that.

A little better than George J. Burns means, for me, that he is probably just outside the HoM.
I'd guess that Poles will be among the best 45 players not in the HoM when we reach the present.

For me to reconsider my ranking of him, I'd need some new evidence, or a convincing
reinterpretation of the evidence we have.

He has the great reputation, but my policy has been to rank NeL players on interpretations of
the data, rather than on reputations.
   53. Gary A Posted: October 20, 2005 at 04:53 AM (#1693945)
One thing to keep in mind with Poles and other pre-1920 NeL players is that their offensive context was very low, especially in Cuba. I haven't covered Poles's years in Cuba yet nor the regular Cuban League seasons, but here are series averages for various North American NeL and major league tours in Cuba in the 1900s (including Cuban and North American teams):

yr--touring team--ave/oba/slg/runs per game
1904 Cuban X-Giants--.207/.280/.234/4.75
1905 Cuban X-Giants--.210/.267/.222/3.89
1906 Cuban X-Giants--.216/.294/.247/4.29
1907 Philadelphia Giants--.218/.290/.247/3.76
1908 Brooklyn Royal Giants--.197/.275/.221/3.22
1908 Cincinnati Reds--.206/.279/.254/3.46
1909 Detroit Tigers--.218/.283/.271/3.75
1910 Detroit Tigers--.232/.292/.264/3.04
1910 Leland Giants--.213/.284/.240/3.12

That's Cuba. In 103 games among top black teams in the U.S. in 1916, the averages were: .255/.330/.321/4.50.

This isn't anywhere near complete or definitive, but it does help put players like Poles or, say, Pete Hill or Lloyd in context.
   54. Joey Numbaz (Scruff) Posted: October 20, 2005 at 05:37 AM (#1693967)
On the no power thing - don't forget it was the deadball era . . .

Looking forward to the new stats that are going to be released next year . . .
   55. Joey Numbaz (Scruff) Posted: October 20, 2005 at 05:37 AM (#1693970)
Heh, Gary stole my thunder :-)
   56. Brent Posted: October 22, 2005 at 07:11 AM (#1697684)
Spottswood Poles played four winters in the Cuban League: 1910-11, 1912w, 1913w,
and 1914-15. (The winter of 1913-14 no North Americans played in Cuba.) Here is
his record:
Season Team    G   AB   R   H 2B 3B HR SB  Avg  Slg 
10-11  Fe     26   89  20  32  5  0  0  - .360 .416
12w    Fe     23   86  10  22  1  2  0 11 .256 .314
13w    Fe      -  151  40  55  3  8  1 21 .364 .510
14-15  Fe     14   57   9  13  0  2  0  4 .228 .298
Total             383  79 122  9 12  1    .319 .413


Notes:
During these four years the Cuban League consisted of only three teams. Most
of the top black American stars of the time played in Cuba at least some of
those winters.
1910-11: Poles led league in doubles; was second in average and slugging behind
Pete Hill.
1913w: Poles led league in at-bats, runs, hits, triples, and slugging. His team
(Fe) won the pennant.

As Gary has reminded us, offensive context is important, so I’ve calculated the
offensive context from the statistics shown in Figueredo’s Cuban Baseball: A
Statistical History, 1878-1961.
These batting statistics follow Figueredo in
excluding most pitchers. (Apparently pitchers batting records are included if
they also played in the field, so the batting records for Méndez and Pedroso are
included.) In 1910-11 the ball was still quite dead, but it became more lively
in the following seasons:
Season  Lg Avg  Lg Slg
10-11     .224    .270
12w       .255    .314
13w       .276    .336
14-15     .266    .313


I’ve calculated the relative batting average and relative slugging for Poles and
other players over those four seasons by taking the ratio of the player’s own
average to what his average would have been if he had hit at the same rate as the
league each season. Poles’ relative average is 1.24 and his relative slugging is
1.32.

I’ve also calculated a combined average that is analogous to OPS+. (Walks and
on-base percentage are not available to the Cuban League, so it is not possible to
calculate OPS+.) The combined average substitutes batting average for on-base
percentage in the OPS+ formula; I’ll call it BPS+ (for batting plus slugging):

BPS+ = 100*(Avg/lgAvg + Slg/lgSlg – 1).

I believe all Cuban League games were played in the same park, so it is not
necessary to park adjust. Here are the top 10 hitters in the Cuban League over
these four seasons based on this statistic (min. 185 at-bats):
Player            Seasons   AB  Rel Avg  Rel Slg  BPS+   
Grant Johnson           2  199     1.38     1.32   170
Pete Hill               2  205     1.34     1.33   167
John H. Lloyd           3  251     1.32     1.34   167
Carlos Morán            3  325     1.37     1.28   165
Cristóbal Torriente     2  226     1.23     1.37   160
Spottswood Poles        4  383     1.24     1.32   156
Armando Marsans         4  364     1.27     1.28   155
Julián Castillo         3  250     1.06     1.37   143
Rafael Almeida          3  250     1.14     1.12   126
Emilio Palomino         3  241     1.08     0.99   107
 

Morán was a third baseman of Chinese ancestry who played in Cuba from 1900-16.
Marsans, who was white, played for 8 seasons in the majors and also played in
Cuba from 1905-28 (we see from his major league data that he did not draw many
walks, which is not visible from the Cuban data). Castillo was a first baseman
who played from 1901-13; he was probably the top Cuban hitter of the aughts.

Another player I’ll mention is Eustaquio (Bombín) Pedroso. He was a pitcher with
a career similar to that of Méndez, though not quite as good. However, while
Méndez was a pretty good hitter, Pedroso was a great hitter (at least for
the two seasons included in these tallies)—-he hit .329 with a .434 slugging
percentage in 143 at-bats, giving him a BPS+ of 164. He didn't make the above top
ten list because he had too few at-bats.
   57. Chris Cobb Posted: October 22, 2005 at 03:37 PM (#1697810)
Brent, this data is MAGNIFICENT!

It is especially valuable because it includes Marsans, so that we can get rough MLEs from this data.

I've just calculated Marsans' major league BA+ and SLG+ for his time in the NL 1912-14. I dropped
his 1911 seasons because he was a rookie, and I left out his FL play in 1914 because it muddies
the waters of league quality. In 1063 PA, his BA+ was 1.11 and his SA+ was .99. He's only one guy
and we have only 364 AB in the CWL, but his conversion factors are .87/,76. The ba/sa ratio is
exactly of the sort we would expect, and the quality level is AA, which is probable for the CWL at
this fairly early stage in its history. It might be a little high or a little low, but it's plausible.

That gives us MLE BPS+ scores for the major players as follows:

Grant Johnson 120
Pete Hill 118
John Henry Lloyd 117
Carlos Moran 116
Cristobal Torriente 111
Spotswood Poles 108

For Johnson, these are his 39-42 seasons, so a 120 BPS+ makes him look, well, great.

For Hill, these are his 31-34 seasons, so a 118 BPS+ doesn't raise his stature, but it doesn't
show him to be obviously underqualified, either.

For Lloyd, these are his 27-31 seasons, so we would expect better. Lloyd was definitely a
late-bloomer, however, so his peak as a hitter is probably later, and a 117 BS+ is still
highly valuable for an excellent defensive shortstop.

For Moran, this was late in his career. It certainly suggests that his election to the Cuban
Baseball Hall of Fame was well-deserved, and that his reputation in Riley as
an excellent contact hitter without much power is accurate. He would merit further study
from us, if there's more we can know.

For Torriente, these are his 16(!)-19 seasons, so there's nothing here to suggest that he
wouldn't become a great player.

For Poles, there are his 22-25 seasons, so this data does not help his case. Given the
uncertainties of the conversion factor and the fairly small sample size, a 108 BPS+
may not accurately represent his value during these years, and it may not include
his peak. But I think that all three factors -- too low a conversion factor, these values
as unusually low for Poles, and a later peak -- would need to be accepted as being
operative here for this data to be consistent with an HoM-quality career for Poles.
By this data, he looks to be a step behind all of his NeL contemporaries whom we
have elected.
   58. John (You Can Call Me Grandma) Murphy Posted: October 22, 2005 at 03:51 PM (#1697822)
Guys, do you want me to post some player threads for some of the players above?
   59. John (You Can Call Me Grandma) Murphy Posted: October 22, 2005 at 03:57 PM (#1697830)
For Johnson, these are his 39-42 seasons, so a 120 BPS+ makes him look, well, great.

For Hill, these are his 31-34 seasons, so a 118 BPS+ doesn't raise his stature, but it doesn't
show him to be obviously underqualified, either.


This is great stuff, Brent, especially when you factor in the uncertainty of their statistical records that we all faced at the time of their inductions. I don't know if it's 100% validation of their elections, but it's close.
   60. Brent Posted: October 23, 2005 at 02:25 AM (#1698657)
Chris Cobb wrote:

I've just calculated Marsans' major league BA+ and SLG+ for his time in the NL 1912-14...
He's only one guy and we have only 364 AB in the CWL, but his conversion factors are .87/,76...
the quality level is AA, which is probable for the CWL at this fairly early stage in its history.


While a conversion factor of .87/.76 is plausible for the CWL for this period, it seems a little
low to me. The other information that can be used to assess the performance of the CWL is their
performance against major league teams. Looking specifically at the 1910-15 era, I'm aware of
the following series:

1910 - Detroit Tigers go 7-5 against Habana and Almendares.
- World Champion Athletics (minus Collins and Baker) go 4-6.
1911 - Phillies (minus Alexander) go 5-4.
- Pennant winning Giants (fully staffed) go 9-3.
1912 - Athletics (fully staffed) go 10-2.
1913 - Brooklyn Superbas go 10-5.

Some of the Cuban teams' successes in the early years was due to major league teams not
coming fully staffed, but these teams still look better than AA to me.

Thinking about possible reasons that the Marsans comparison might be biased, I thought
of three possible reasons:

- Perhaps his style of play (aggressive base runner whose "power" mostly came from trying
to stretch out an extra base) was more suited to Cuba than to the majors.

- His first 3 Cuban seasons in this 4-year span were with an Almendares team that was
strongly oriented toward pitching and defense. (For example, in 1910-11 Almendares
won the pennant even though their team batting and slugging averages were lower than those
of their main competitor, Habana -- .216 and .263 versus .233 and .293.) In a very small
league, this means that Marsans was not facing as many top pitchers, and conversely, Johnson,
Hill, and Poles were facing tougher pitchers on average.

- With a sample of only 364 AB in the CWL, we shouldn't ignore the possibility that the data
may just be too noisy to reliably infer what we'd like to.

I agree that these data don't help Poles' case. On the other hand, he's close enough to the
other HoMers that if a strong case could be that he was a superb defensive outfielder, he
might remain in the mix.

By the way, Eric Enders wrote a very nice biography of Marsans that was published in SABR's
Deadball Stars of the National League.
   61. Gary A Posted: October 23, 2005 at 02:27 AM (#1698668)
I'm working on reconstructing Cuban League seasons (and exhibition series) from this period from box scores. Nothing's really finished, but I can say that Carlos Moran walked a fair amount, as did Johnson and Hill. Hill hit extremely well against major league competition in 1908-09, as well as in exhibition series matching NeL teams against Cuban teams, 1904-09. (He was pretty much better than anyone, including major league hitters, in these series.)

Btw, let me reiterate my offer to send the 1927/28 Cuban League stats I've compiled to anyone who wants them.
   62. Gary A Posted: October 23, 2005 at 02:31 AM (#1698676)
Also, Rafael Almeida played in the majors at this time, too, though only with 285 ABs.
   63. Brent Posted: October 23, 2005 at 02:43 AM (#1698713)
Guys, do you want me to post some player threads for some of the players above?

I'd say not yet. If I have time, I'd like to look a little more at Pedroso, Castillo, and
Moran, and if I find anything worth posting I'll let you know.

Pedroso had several big years, but as a contemporary and teammate of Mendez, it's pretty
clear that Mendez had a bigger peak. By the 1920s Pedroso was still around but was no
longer a top pitcher.

If we'd looked at Castillo when he first became eligible, he might have been in the mix
among deadball first basemen along with Chance, Konetchy, etc. However, my guess is that
there will be too many questions about Cuban League quality in the aughts to make him a
viable candidate.

Moran is perhaps the most interesting discovery. Neither Figueredo nor The Pride of Havana,
by Roberto Gonzalez Echevarria, focuses much attention on him, but third base is an
underrepresented position and he was probably in his thirties by 1910-15. (Riley doesn't
provide a date of birth, so I don't know his age. I notice Riley says that he batted and threw
left handed, which doesn't seem likely for a third baseman.)

I suppose one possibility would be to set up a thread for miscellaneous Negro League or Latin
American players where we could post information on marginal players who aren't likely to
emerge as major candidates.
   64. John (You Can Call Me Grandma) Murphy Posted: October 23, 2005 at 03:00 AM (#1698781)
Brent, use these: Negro Leagues

Latino Stars
   65. Gary A Posted: October 23, 2005 at 02:45 PM (#1699128)
Using Brent's figures above, Almeida, with 285 Cuban League at bats and 250 major league at bats (1911-13), gets conversion rates of .88/.95.
   66. Gary A Posted: October 23, 2005 at 02:48 PM (#1699131)
(Riley doesn't provide a date of birth, so I don't know his age. I notice Riley says that he batted and threw left handed, which doesn't seem likely for a third baseman.)

Handedness seems to be an area of frequent confusion for Cuban players around this time. Luis Padron, whom we discussed in I think the Beckwith thread, is always said to have been a lefthanded pitcher and batter; but Gadfly commented that he'd seen a photo where he appears to be gripping the ball in his right hand (I think I've seen it, too), and I've recently seen a photo in a 1912 issue of La Lucha where Padron is batting righthanded.
   67. karlmagnus Posted: October 24, 2005 at 12:28 AM (#1700142)
Bear in mind that 1910 photographic technology was eccentric, and photos were frequently printed in mirror-image form, particularly if the newspaper didn't know which way round they should be.
   68. Gary A Posted: October 24, 2005 at 01:02 AM (#1700276)
Yes, of course. The Padron image has his team's name on his jersey ("Havana Park") as well as initials on his cap ("HP"), making it clearly not reversed.
   69. Gary A Posted: October 24, 2005 at 05:36 AM (#1701030)
For Torriente, these are his 16(!)-19 seasons, so there's nothing here to suggest that he wouldn't become a great player.

Torriente was born in either 1893 or 1895 (I haven't seen a specific birth date), and he first played in the Cuban League in the January-March 1913 season. He also played in the 1913/14 and 1914/15 (the first isn't, I believe, covered in Brent's study); so this covers him either from age 17/18 to 19/20, or 19/20 to 21/22.
   70. Gary A Posted: November 07, 2005 at 04:09 PM (#1721515)
Some additional data on Poles: in November/December 1912, the Lincoln Giants played a 13-game series against Habana and Almendares in Cuba. Holway doesn't mention this series, and Figueredo only gives game scores (as always), so this is new information.

G-13 (team 13)
AB-48
H-13
D-0
T-0
HR-0
R-5
W-6
HP-2
SF-0
SH-1
SB-6
AVE-.271 (series ave .254)
OBA-.375 (ser .342)
SLG-.271 (ser .295)

Doesn't add anything to his case. Other players of note in this series:

J.H. Lloyd .348/.444/.413, 13 runs, 8 stolen bases in 13 games
Santop .190/.261/.238 in 8 games (not yet a regular, apparently)
C. Moran .182/.379/.273 in 7 games
J. Castillo .091/.087/.091 in 6 games
Mike Moore .268/.380/.317 in 13 games

Eustaquio Pedroso hit 5 for 8 with one walk in 4 games (.625/.667/.625).
Marsans hit 6 for 12 with a double and a walk in 4 games (.500/.538/.583)
Almeida hit 9 for 18 with 3 doubles in 5 games (.500/.500/.667).
   71. Yoenis Cespedes, Baseball Savant Posted: May 03, 2006 at 01:35 AM (#2001575)
A few more words on Spottswood Poles and some MLEs:

Periodically, someone brings up Bill James' player rankings in the Abstract as a means of comparison to see if there is anyone we are overrating, underrating or overlooking. In the case of Poles, the difference between James and the HOM electorate is huge. In the Abstract, referring to his list of the top 100 players of all-time, James writes:
"I am more bothered by the exclusion of Bullet Rogan, Spottswood Poles, and Willard Brown than by dropping Harry Heilman [sic], Ralph Kiner, and Early Wynn."
This would imply that James has Poles as, at least, one of the top 125 players of all-time, right around the middle of his "personal HOM," so to speak. In positional rankings, Kirby Puckett is the last centerfielder to appear in James' Top 100. Puckett ranks 12th at the position, including Negro Leaguers. James' comments about Poles would place him in the 13-15 range all-time among CFs.

Not one of the 50 HOM voters in the 1975 election put Poles on their ballot.

The only MLE batting Win Shares we have for Poles are in Post #32 from JeffM and Post #45 from Dr. Chaleeko:
Dr. Chaleeko Posted: July 29, 2005 at 01:09 PM (#1507736)
SPOTS POLES RECONSIDERED

Since I'm SOBF, I decided I'd take a little time to re-consider him. I decided to rerun his numbers, building OPS+ and WS totals for him.

I used I9s' numbers, shaving 5% off of his rates per the usual I9's discount. Per Chris Cobb's standard, I used the AL in odd years and the NL in even years as the basis for my comparisons.</b>
The probelm with taking 5 percent off of each of the rate stats is that it does not equal a 5 percent discount overall. From the Willie Davis comment in the Abstract (page 741):
"f you increased each run element by ten percent and froze the batting outs, a player's runs created would increase by almost twenty percent."
Using this reasoning (and, please, someone correct me if I'm wrong), a flat 5 percent discount of Poles' batting rates would imply a 10 percent discount on his Runs Created.

Using the KJOK MLEs posted on the Hall of Merit Yahoo Group, I figured short-form offensive Win Shares for Poles. Notes:

1. I discounted Runs Created by 5 percent.
2. I use Dr. Chaleeko's estimated defensive Win Shares.
3. The 1918 season is an average of the four surrounding seasons.
4. The 1919 season is adjusted upward from 140 to 154 games.
5. Games Played is PA/4.2.
6. AVG/OBP/SLG/OPS+ are set to a .267/.334/.427 league (which is what I use for all hitting rate stats):

SPOTTSWOOD POLES MLEs

Year / G / PA / AVG / OBP / SLG / OPS+ / oWS / dWS / WS

1909: 93 / 389 / .305 / .368 / .470 / 120 / 10.2 / 3.4 / 14
1910: 132 / 554 / .281 / .341 / .450 / 107 / 11.8 / 4.9 / 17
1911: 149 / 624 / .297 / .379 / .507 / 132 / 18.7 / 5.7 / 24
1912: 138 / 579 / .356 / .429 / .548 / 157 / 25.1 / 5.3 / 30
1913: 151 / 635 / .319 / .407 / .497 / 138 / 21.0 / 5.7 / 27
1914: 154 / 705 / .343 / .428 / .520 / 150 / 27.2 / 6.4 / 34
1915: 154 / 663 / .331 / .401 / .506 / 139 / 23.0 / 6.0 / 29
1916: 149 / 625 / .306 / .394 / .507 / 137 / 20.3 / 5.6 / 26
1917: 154 / 665 / .329 / .417 / .525 / 148 / 24.8 / 6.0 / 31
1918: 153 / 646 / .321 / .408 / .514 / 143 / 22.8 / 5.8 / 28
1919: 154 / 649 / .303 / .381 / .474 / 125 / 16.7 / 5.7 / 22
1920: 154 / 646 / .347 / .438 / .550 / 160 / 27.8 / 5.8 / 34
1921: 151 / 594 / .294 / .379 / .467 / 123 / 15.6 / 5.4 / 21
1922: 123 / 518 / .242 / .320 / .409 / _92 / _7.2 / 3.9 / 11
CAREER: 2092 G / 8492 PA / .314 AVG / .395 OBP / .499 SLG / 135 OPS+ / 272.2 oWS / 75.6 dWS / 348 WS
   72. Mark Shirk (jsch) Posted: May 03, 2006 at 02:20 AM (#2001650)
why are games played PA/4.2? Shouldnt' that number be higher?

ALso this is very interesting but I have a few questions.

1. Why are KJOK's MLE's so far off from what else we have about Poles? What is above ranks him near Alejandro Oms. His OPS+ is about 25-30 points higher here.

2. This isn't really a question I guess but Bill James NeL rankings are all over the place. Judy Johnson and Ray Dandridge rank above Jud Wilson and John Beckwith. Bingo DeMoss is his top NeL 2Bman above Frank Grant and Luke Easter is, I believe, the second ranked NeL 1Bman. I relaly love that book and I do trust James' rankings (though at times I disgaree) but he was far from an NeL expert. I trust our numbers and rankings of NeL players more than his.
   73. Yoenis Cespedes, Baseball Savant Posted: May 03, 2006 at 03:39 AM (#2001712)
Games played does not affect my offensive Win Shares calculation in any way. Maybe the PA/G ratio should be higher, I don't know.

I e-mailed Kevin (KJOK) to ask him about his MLEs. He said that the base numbers he used were not discounted or regressed. I corrected for the 5 percent discount in using KJOK's MLEs for sfWS.

I think the standard 5 percent discount method for 19s MLEs significantly underrates the players evaluated. If you use a standard 5 percent discount for all rate statistics, it amplifies into a significantly larger discount in sum. In Dr. Chaleeko's Post #45, that 5 percent discount for OBP and SLG leads to an 8 percent discount of OPS+, 115. If you did just 5 percent off of the OPS+, Poles' mark is 119.

This is a comparison between the raw i9s OPS+, my MLEs and KJOK's MLEs:

1909: 122/120/126
1910: 103/107/114
1911: 118/132/138
1912: 143/157/162
1913: 134/138/143
1914: 147/150/155
1915: 132/139/144
1916: 136/137/142
1917: 149/148/153
1919: 125/125/130
1920: 149/160/165
1921: 119/123/128
1922: 77/92/97
CAREER: 129/134/140

To be honest, I'm not sure why KJOK's MLEs are so bullish on Poles. I assumed that they were a straight conversion from i9s to the 1955-2003 National League context he uses, but I don't think so. I e-mailed him asking if he had any extant data for these conversions beyond i9s. I'll run Win Shares with a 5 percent discount off of the i9s projected Runs Created and see what comes up.

(For what it's worth, the projection I posted previously would put Poles in my rankings almost exactly where he is in Bill James' rankings.)
   74. sunnyday2 Posted: May 03, 2006 at 03:49 AM (#2001715)
There are others who are vastly more qualified to respond to James and j re. Spot Poles (e.g. Chris and Doc) but I'm gonna take a shot anyway.

Poles was an early NeL candidate and the NeL MLEs that Chris, Doc and others have developed were still in flux.

• One of the early sources was the I9's, but there quickly developed a clear consensus that their numbers were very much inflated. The old 5 percent discount from I9 left way too many of the early NeLers in inner circle territory. It strained credibility a lot, and so was even a 5 percent discount from I9 was judged to be incredible.

• KJOK develop some MLEs as well. I don't remember if they were based on I9--in fact, maybe it was K who developed the 5 percent discount from I9. Note than Poles became eligible almost 2 years ago now, and I am going from memory, but in any event K's numbers were also generally agreed to be too high.

My sense, speaking only for me, is that taking another 5 percent off of K's numbers does not return us to reality. I mean, I could say that Willard Brown would easily have hit 60 HR in the MLs. If somebody took 5 percent off of that, would it be real? And if somebody else took another 5 percent off of that, would that be more real? IOW garbage in, garbage out. 19 is garbage from the standpoint of having a basis in reality.

What Chris and Doc and others have done is go back to the actual NeL numbers (as well as MxL and PRL and etc.), normalize them to 154 or 162 game seasons, and then regress those numbers toward the mean (whether the regression might be literal or figurative) and then apply a competition discount. Generally it was agreed that the competition discount should be somewhere in the 5 to 15 percent range, varying through different periods in NeL history. Discounts for the Latin leagues also varies and is a bit higher on average than the NeL discounts (again, if my memory serves me correctly). That is, the NeLs on average were a little bit tougher than the Latin leagues.

In Poles case, what we know is that he hit .300 in the NeLs, .319 in Cuba and over .400 against all competition (obviously weaker competition than the NeLs or Cuba). (OK, being honest, this is what we THINK we know.) Note right up there in the same post the I9 projection of Poles to a .303 BA. Now, we have no idea what that means, except that we interpreted it to mean that I9 believes Poles would hit .303 in the MLs. I don't think we know what kind of run environment they mean. 1909-1923, when Poles played? Or is that an all-time environment? Don't know.

But seriously. Poles hit .300 in the NeLs when the NeLs had not yet gathered all of the best black talent within its sphere, when it was not yet as strong as it would one day be. Yet he would hit .303 in the MLs? At any time in history?

No.

That is too much.

So, we have WS projections for Poles (above) of anywhere from 310 to 348. But I believe that all of these projections spring from I9.

Poles voting totals simply indicate skepticism concerning these projections, and the skepticism IMO is justified.

Unless and until we have more data re. Poles' actual accomplishments on the field of play--and maybe the new ML study will yield something, maybe not--then his actual playing record simply does not justify Poles as a .300 + hitter in the MLs. And even I9 has his SA at .394. James says, no, it's .498?????

Again, no. There is nothing in the record to support anything even remotely like that. If I9 says .394, I say .360. Anecdotally we know he was a slap hitter.

As to James' ratings, I am sorry to say that I agree with j. We, here at the HoM, are capable of looking at the numbers and making up our own minds. James' discussion of his ratings is so anecdotal that I wonder if he really looked at the data that we have looked at. Clearly he did not develop it in the way the Chris and Doc have done. jschmeagol is absolutely right. James' ratings are of no particular value here.

And neither, as I said, is I9, much less (much much much less) any analysis that has him actually exceeding the I9s.
   75. Yoenis Cespedes, Baseball Savant Posted: May 03, 2006 at 04:37 AM (#2001743)
Marc,

The .498 SLG for Poles is compared to a .427 league average. All of the AVG/OBP/SLG numbers I have in my player profiles (MLB and NeL) are converted to a .267/.334/.427 league average.
   76. Yoenis Cespedes, Baseball Savant Posted: May 03, 2006 at 04:46 AM (#2001746)
These are new MLE Win Shares for Poles, taking 5 percent off of his i9s RC:

SPOTTSWOOD POLES
MLE WIN SHARES

1909: 11
1910: 14
1911: 23
1912: 29
1913: 25
1914: 30
1915: 25
1916: 23
1917: 27
1918: 25
1919: 20
1920: 31
1921: 21
1922: 12
TOTAL: 316
   77. Chris Cobb Posted: May 03, 2006 at 05:10 AM (#2001752)
Spotswood Poles

316 win shares

or

348 win shares.

This 10% difference makes all the difference.

The fact that we can estimate a major-league equivalent career value with confidence that our estimate is, say, no more than 15% too high or too low is pretty remarkable, but given that we are making distinctions between top-ballot and off-ballot that turn on less than 5% difference in value between two players, we are forced to rely on guesswork to place Poles in our rankings.

My judgment, particularly considering what we learned from Brent's CWL league data from the 1911-14 period, which looks to have been Poles's peak, is that the low estimate is more probable, and might even be too high.

However, no one has done a comprehensive reassessment of Poles using the data Brent and Gary A. have provided, so I think renewed study of Poles is an excellent idea, and I'm glad James has given the electorate a push in that direction, though his studies, being based off of the same projected data that we were using before, basically revisit the old higher and lower bounds of Poles' value that we considered before.
   78. sunnyday2 Posted: May 03, 2006 at 05:41 AM (#2001762)
I agree. 316 is too much. The best info we have is that he was a .300 hitter in the NeLs.

Also:

>1920 – no ba listed in play for NY Lincoln Giants
>1921 – no records listed
>1922 – no ba listed in play for NY Lincoln Giants

I9 can make a career out of this data if it wants, that is its privilege. I guess we can make 31-21-12 WS out of this too if we want. But deriving 31-21-12 WS from I9 data...like I said, it is made up.

> I think the standard 5 percent discount method for 19s MLEs significantly underrates the players evaluated.

James, if this is so, (IOW if I9 is underrating the NeL players) then we're not just talking Spot Poles here. We need to make room in the HoM for Heavy Johnson and Chaney Smith and another 35 NeLers. We all struggled with this 2-3 years ago, and it doesn't work.
   79. Yoenis Cespedes, Baseball Savant Posted: May 03, 2006 at 06:42 AM (#2001772)
Marc,

I am not talking about the raw I9s data underrating NeL players. I am talking about how some HOM voters have applied a 5% "discount" on AVG/OBP/SLG. Docking all of those rate stats 5% does not lead to a proportional discount of a player's offensive value (runs created). It leads to a much higher discount. Here's an example:

.280 AVG/.400 OBP/.500 SLG in 600 PA (140 H, 300 TB, 100 BB, 500 AB)

The batting line is worth 120 Runs Created. A five-percent discount is 114 RC.

Taking 5 percent off of each rate statistic gives us the following line:

.266 AVG/.380 OBP/.475 SLG in 600.138 PA (134.862 H, 240.825 TB, 93.138 BB, 507 AB)

The batting line is worth 91 Runs Created, a 24 percent discount.
   80. Yoenis Cespedes, Baseball Savant Posted: May 03, 2006 at 06:46 AM (#2001773)
Similarly, if you discount Poles' batting events (H, BB, TB) 5%, you end up with around a 12% discount on Runs Created from the I9s estimates.
   81. Mark Shirk (jsch) Posted: May 03, 2006 at 08:14 PM (#2002113)
The big difference between James' two WS MLE's for Poles are that in teh first Poles five year peak is 34,34,31,30,29 whic is pretty good and would see him ranked just off my ballot with Alejandro Oms, maybe even higher. The second estimate is at 31,30,20,27,25, which isn't as impressive. OFers within 15 WS of 300 for a career with a top five peak that looks like that aren't too rare.

I do want to state that I am only using 5 year peak here for brevity, if you like 7, 6, 3, whatver go ahead. My system puts wieght on each WS above and byond 15 for prime and 25 for peak so that I don't have to make any arbitrary cut-offs in years, intead making an arbitrary but off (sort of) for a baseline.

If we are going to redo Poles I would like to see Bill Monroe redone, he ranks in my top 50, unlike Poles, and has been close to my ballot before. I wish I had the ability to do the redo's but I don't. I also don't have the time these next few weeks with finals and all.
   82. sunnyday2 Posted: May 04, 2006 at 01:33 AM (#2002865)
>The batting line is worth 91 Runs Created, a 24 percent discount.
>if you discount Poles' batting events (H, BB, TB) 5%, you end up with around a 12% discount on Runs Created from the I9s estimates.

My point is it doesn't matter if you discount 5-12-24 percent. In each case you are discounting from imaginary numbers.

Whereas if you start with Poles' actual record (.300 career BA against NeL teams), and if you want to argue that the NeLs were just as tough as the MLs and therefore you want to carry over 100 percent of his value (no discount)--even if you did that, he would be a .300 hitter, most certainly not be a .314 hitter. You're saying the NeLs were 5 percent BETTER than the MLs at that point.

Just forget about I9 and use Poles' real record. That's what I am suggesting. Now you're on planet Earth.
   83. Yoenis Cespedes, Baseball Savant Posted: May 04, 2006 at 02:40 AM (#2003255)
Marc,

I could really do without the unnecessary snark in this discussion. It is disrespectful and detracts from the discussion at hand.

You are getting all exercised over a point that I have previously explained and dealt with. Anyway, if you read Post 75, you would know that the .314/.395/.498 batting line that I previously posted for Poles (which I am backing away from, for the record) is against a league average of .267/.334/.427 because that is the adjustment I make for every hitter (which is something else that I previously posted. An adjustment to a Dead Ball context (.250/.310/.330, let's say) would give us a batting line of .294/.367/.385.

Also, we should consider that Poles played in the Dead Ball Era. The NeLs were playing in a similar offensive context to the MLB, if not worse. One needs only to look GaryA's NeL data at Brent's Cuban Winter League data to see that. This is going to underrate Poles in any conversion method if we fail to adjust for it.
   84. sunnyday2 Posted: May 04, 2006 at 04:16 AM (#2003442)
James,

Sorry. I am not meaning to be snarky, I am only meaning to disagree.
   85. KJOK Posted: September 17, 2011 at 07:19 AM (#3927870)
   86. KJOK Posted: September 17, 2011 at 08:33 PM (#3928214)

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
TedBerg
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 0.7611 seconds
49 querie(s) executed