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

Monday, July 25, 2005

Willard Brown

Willard Brown

Eligible in 1958.

John (You Can Call Me Grandma) Murphy Posted: July 25, 2005 at 01:51 PM | 114 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.

Page 2 of 2 pages  < 1 2
   101. Joey Numbaz (Scruff) Posted: February 24, 2006 at 06:15 AM (#1873411)
"He didn't walk once in 67 PA in MLB, but managed an equivalent of 48 BB in 610 PA in his other play?"

Actually that would be 48 BB in 543 PA in his other play, as I assume the 67 PA include his MLB numbers.
   102. Dr. Chaleeko Posted: February 24, 2006 at 05:41 PM (#1873750)
Joe,

The reason is that I'm using a wider sample than only his summer play. I also include winter league play in my MLEs. Brown was spectacular in Puerto Rico that year, just monstrous, a translated line of .388/.446/.743, which raises his overall assessment from where you would expect based solely on his summer play.

As for the walks, they are determined by his actual 1947 MLB walk rate in combination with the walk rate observed in the seasons for which we have data. For reasons to do with the fragmented statistical record of a lot of NgL hitters, I express walk rate as BB/H, this could be causing distoritions in his walk rates, though theoretically I don't think it should except perhaps at the extremes.
   103. Mark Shirk (jsch) Posted: February 24, 2006 at 05:59 PM (#1873778)
Wouldn't using BB/H show big swings in BB's since AVG, and thus hit totals, can be so dependant on luck and experience such big swings from season to season?
   104. DL from MN Posted: February 24, 2006 at 06:45 PM (#1873851)
> translated line of .388/.446/.743

I have a lot of trouble with these translations. The sample sizes are really small, the park factors are unknown, the quality of play is a guess. How many actual plate appearances (in ballgames) did Willard Brown have in his career? I've seen extrapolated plate appearances for his career but I don't have a good sense of how much extrapolation is being done or how much regression. I would agree he had a terrific month in Puerto Rico but a lot of people have put up good numbers for a month in AA.
   105. sunnyday2 Posted: February 24, 2006 at 07:36 PM (#1873967)
DL, there have been loooooong discussions of these issues over the years--more so, certainly, back when the NeLers were still coming on the ballot year after year. I'm hardly the expert, but as a user of the data for how many years now:

The sample sizes aren't that small, at least not on a career basis. We can fairly easily compare the NeLers among themselves, except the pre-1920 guys. Then, yes, it's a bit of guesswork as to where the NeLers fit in with their ML counterparts...

Which brings us to the competition level (quality of play). Pretty much everybody agreed that it is reasonable to think that "some" NeLers 1920-1950 were HoF caliber players. "Some," of course, isn't terribly helpful. But eye-balling some of the early conversions, we were finding that there were NO HoMers among the NeLers, and other early conversions suggested that the 10 best players of the 1930s were ALL NeLers. Neither of these extremes seemed palatable.

So the conversions eventually found a level. But their true meaning is subject to much qualitative analysis. The only mistakes would be 1) to think that there should be an easy way, and 2) to just give up and disregard the rest of the NeL backlog.

The truth is that Dickie Pearce was a much harder case than Willard Brown and, hey, nobody died.
   106. Dr. Chaleeko Posted: February 24, 2006 at 07:53 PM (#1874011)
jschmeagol,

I've wondered the same thing. I think at the extremes it might, but that in the main it probably shouldn't because hits as a percentage of PAs aren't all that different over most seasons. 28% versus 32%. The tradeoff I'm making is that most seasons will fall within the normative range, but for those careers where I have virtually no other data, I can almost almost find hits.

As Sunny just said, it's not an exact science.... With any luck the big upcoming NgL $250,000 study book will help us somewhat with these problems.
   107. KJOK Posted: February 24, 2006 at 08:44 PM (#1874101)
We somewhat know the park in this case also, since it's the same park the Kansas City A's played in.

It played as a slight hitters park in the AL, but that was after some Charly O. moving in of the fences. As far as Negro League parks go, it was HUGE, and not really conducive to power hitting (350 down the lines, 408 to left center, 450 to CF...
   108. Al Peterson Posted: February 24, 2006 at 09:10 PM (#1874139)
Dr. C,

On the latest MLEs for Willard Brown what was the basis for the half seasons in 1938 and 1943? I can't find in the thread mentions of injury and I see earlier projections for those years show full-time play.
   109. Chris Cobb Posted: February 24, 2006 at 09:36 PM (#1874174)
DL wrote:

I have a lot of trouble with these translations.

The sample sizes are really small,


This varies from season to season. In 1947 we have, relatively speaking, an abundance of data. Brown had 234 AB in Puerto Rico 1947-48 (a long month, I guess). With another 58 major-league at bats and 211 AB in the NAL, that's pretty close to a full season for 1947: 503 at bats. Now for Brown in the 1930s, we're sometimes looking at fewer than 50 recorded at bats, but for the rest of his NeL career we're dealing with 100-300 recorded ab per season.

the park factors are unknown,

See KJOK above, but generally true in this case, though enough data is available to calculate rough park factors for some NeL parks in the 1920s.

the quality of play is a guess.

It's a highly educated guess for the NeL and for at least the high minor leagues. The conversions that Dr. Chaleeko and I use for the Negro Leagues, which some experts view as setting the quality of play too low, are based on a study of all the players who had significant performance records in both the NeL and the majors. They are not just guesses. The PRWL level of play is more of a guess, and I haven't worked with that data myself, but if you want to see translations for Brown that rely only on NeL and MxL data for which the conversion rates are statistically derived, you can look at my earlier translations for Brown upthread.

How many actual plate appearances (in ballgames) did Willard Brown have in his career?

There's no way of knowing how many actual plate appearances he had, since a lot of his play in the Negro Leagues did not get documented. If you are asking how many _recorded_ plate appearances he had, I can't give an exact figure of the top of my head, but it's got to be 6000+. He had almost 2000 recorded in Puerto Rico, about 2200 recorded between the Negro Leagues, Mexico, and the majors, and he played about 6 years in the minors after 1950, and I'm sure we've got another 2000 ab there. For an NeL player, that's a very full statistical record.

I've seen extrapolated plate appearances for his career but I don't have a good sense of how much extrapolation is being done or how much regression.

Plate appearances are determined by extrapolation on a season-by-season basis from estimates of the number of games Brown would have played in the majors in a given year (determined by usual progression of usage patterns due to aging, his documented level of play, his documented injury history [in Brown's case his history of injury-free play], and his record of military service) and the usual ratio of plate appearances to games played.

In the MLEs that I calculate, I use regression to a five-season mean to project a player's recorded performance into his extrapolated plate appearances. In the MLEs that Dr. Chaleeko calculates, no regression is used. So if you are more comfortable with translations that include regression, you might take a look at my figures upthread. They don't include PRWL data or the recent additional data provided for Brown, but they are not all that different from the good Doctor's. They certainly suggest that his numbers are in the ballpark for Brown.

I would agree he had a terrific month in Puerto Rico but a lot of people have put up good numbers for a month in AA.

As mentioned above, it was 234 at bats, and those at bats need to be considered in the context of an entire career of data for Brown.
   110. Dr. Chaleeko Posted: February 24, 2006 at 09:41 PM (#1874178)
Al,

In 1938 the discrepancy comes in because of the availability of new information (new to me, that is). I had underestimated how many games the Monarchs played in 1938 originally, so that effected my projection. In addiiton I added in the CWL data since the first projection, and Brown played in a smaller percentage of his team's games than he did in the NAL, where he was a little over half.

In 1943, I'm not sure what the cause was, but he played about half of KC's games, so that's where the current projection comes from.
   111. Dr. Chaleeko Posted: February 24, 2006 at 10:11 PM (#1874233)
I have a lot of trouble with these translations.

Me too, and I make them! We all do, as Sunnyday said, there's risk in them, but what else is there?

I think one of the interesting things about the MLEs is that although Chris and I are operating on the same general theoretical basis, we use somewhat different datasets and somewhat different means to arrive at very similar conclusions. With two people working through the numbers and getting a similar result, this should at least offer some reassurance to the skeptics.

The counter argument to that last sentence might go: But what if the theoretical model behind our work is so widely influencing us that its scope is much wider than any deviation between our computational and data-management strategies?

It's a fair question, but I still think that's not all that likely. I'm incorporating a lot of PAs that Chris isn't (thousands in Brown's case), and he's using advanced mathematics while I'm feverishly finger-reading my multiplication tables. Yet despite these major differences in handling the sample, the extra data in my caluclations ultimately buttresses the conclusions of the advanced math---and vise versa. It may not fully agree, but it's very rare that Chris and I have been way apart on a guy.

Thinking about it that way makes me feel more comfortable that we're drawing the best picture of each guy that we can from the currently available data. I could be wrong, of course (it's happened 100 times on this thread!), but being wrong is a data point for improving the MLE system.
   112. Mike Webber Posted: April 15, 2007 at 10:59 PM (#2335802)
Willard Brown story by Joe Posnanski

Interesting piece by my favorite columnist about HOM honoree Willard Brown.

"The late Buck O’Neil used to tell the story of Sonny Brown in classrooms. And then he would ask them, “What’s the lesson here, children?” Invariably, they weren’t sure.


“What’s the lesson of Willard Brown?” Buck O’Neil would ask. And he would answer the question himself.

“The lesson,” he said, “is that it wasn’t easy.”"
   113. David Concepcion de la Desviacion Estandar (Dan R) Posted: September 21, 2008 at 04:55 PM (#2949257)
Here's Brown in my WARP. Methodological notes:

1. I used Dr. Chaleeko's MLE's for playing time, positions played, and batting rates, and Chris Cobb's Fielding WS for fielding rates.
2. I completely fabricated the baserunning wins based on a guy with a strong but not fantastic reputation for speed.
3. '44 and '45 are filled in using my war credit equations.


Year SFrac BWAA BRWAA FWAA Replc WARP
1935  0.27  0.6   0.1 
-0.3  -0.8  1.1
1936  0.80  0.9   0.2 
-0.5  -2.3  3.0
1937  0.96  4.0   0.4 
-1.0  -2.9  6.3
1938  0.42 
-0.1   0.1  0.2  -0.4  0.6
1939  0.79  1.9   0.2  0.6  
-0.8  3.5
1940  0.82  0.0   0.2  0.0  
-0.8  1.1
1941  0.70  4.0   0.2 
-0.1  -0.7  4.8
1942  0.99  4.5   0.2  0.1  
-0.8  5.7
1943  0.47  2.4   0.1  0.1  
-0.5  3.0
1944  0.80  3.8   0.2  0.0  
-0.7  4.7
1945  0.80  3.8   0.2  0.0  
-0.8  4.8
1946  0.98  4.9   0.2 
-0.2  -0.8  5.7
1947  0.94  5.3   0.2 
-0.7  -1.0  5.7
1948  0.87  2.8   0.1 
-0.1  -1.0  3.8
1949  0.87  2.3   0.1  0.3  
-0.7  3.4
1950  0.58  0.0   0.1  0.1  
-0.5  0.6
1951  0.85  1.6   0.0 
-0.3  -0.8  2.1
1952  0.60 
-0.8   0.0 -0.4  -0.5 -0.6
TOTL 13.51 42.1   2.7 
-2.3 -16.8 59.3
TXBR 12.91 42.8   2.7 
-1.9 -16.3 59.9
AVRG  1.00  3.1   0.2 
-0.2  -1.2  4.4 


3-year peak: 17.7
7-year prime: 37.6
Career: 59.9
Salary: $154,664,663, just a tiny nudge over the borderline, but probably deserving of election. Worse than Wynn, a bit worse than Dawson, comparable to the unelected Leach, superior to the elected Averill and Roush.
   114. Joey Numbaz (Scruff) Posted: September 21, 2008 at 06:03 PM (#2949284)
You are the man Dan, thanks!
Page 2 of 2 pages  < 1 2

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Vegas Watch
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 0.3504 seconds
49 querie(s) executed