BBTF's Hall of Merit Discussion :: The New Historical Baseball Abstract

Hall of Merit
— A Look at Baseball's All-Time Best

Friday, April 05, 2002

The New Historical Baseball Abstract

In the extended text (click on discussion) I’m posting Bill James’ thoughts on our first ballot (1906, for players that retired before 1901), according to the NHBA. I think his rankings need to be adjusted for season length, which will make a huge difference for guys like Deacon White and Ezra Sutton.

The top-ranked players eligible for our first election, according to Bill James. The players listed all finished in the top 100. Don’t forget, these rating are heavily skewed towards peak value.

I’ll list career Win Shares (he gives three for each team win), top 3 years, top 5 consecutive, per 162 games.

C-
Buck Ewing 241 WS (27, 23, 21); 98; 29.69
Jack Clements 146 WS (19, 18, 18); 73; 20.44
Doggie Miller 135 WS (24, 17, 15); 65; 16.61
Fred Carroll 109 WS (20, 17, 16); 60; 15.86

James called Fred Carroll the second best young catcher EVER. Second only to Johnny Bench. WOW. I assume he faded . . .

1B-
Cap Anson 381 WS (30, 29, 24); 123; 27.12
Dan Brouthers 355 WS (34, 31, 29); 138; 34.38
Roger Connor 363 WS (36, 32, 30); 145; 29.45
Henry Larkin 177 WS (29, 23, 19); 107; 24.22
Tommy Tucker 176 (30, 20, 17); 96; 16.90

2B-
Bid McPhee 305 WS (27, 23, 21); 107; 23.14
Hardy Richardson 230 WS (32, 25, 23); 111; 27.99
Fred Pfeffer 202 WS (21, 20, 20); 89; 19.60
Yank Robinson 131 WS (24, 21, 20); 101; 21.70
Fred Dunlap 165 WS (38, 17, 17); 100; 27.70

I remember really liking Hardy Richardson in creating my spreadsheet for OW-L. Good to see that validated here.

3B-
Denny Lyons 189 WS (27, 27, 25); 122; 27.31
Ed Williamson 173 WS (21, 20, 19); 87; 23.34
Billy Nash 222 WS (25, 23, 22); 107; 23.32
Arlie Latham 222 WS (25, 24, 24); 102; 22.10
Bill Joyce 155 WS (28, 25, 18); 105; 27.78
Deacon White 190 WS (21, 18, 17); 82; 23.78
George Pinkney 157 WS (29, 22, 22); 110; 21.87
Bill Shindle typo, listed Whitney twice, no numbers
Ezra Sutton 159 WS (28, 22, 21); 98; 24.98
Jerry Denny 140 WS (19, 19, 16); 77; 18.33

Personally, I move Deacon White to the top of this list. He was much older than these guys, so the early part of his career was spent with much shorter seasons. He also caught a bunch, but the numbers do take that into account, although those season were shorter. I listed him as a catcher, and I have him as the top peak value player whose career ended prior to 1906 (as far as I’ve gotten in my spreadsheet). I also had him 3rd in career value, behind only Anson and Jim O’Rourke. He’ll be a very interesting debate early in the voting.

I also had Ezra Sutton ranked as the greatest 3B up that point, ahead of Williamson, Latham, Lyons, Pinckney, Denny. James doesn’t give any credit for 1871-75 (which hurts White also), and again his best years were with shorter seasons. His 24.98 per 162 is pretty high on the list, and he played 18 years. He was as good a 3B as there was before the turn of the century. He’d be behind McGraw on peak, behind Jimmy Collins on career (barely), both of those guys won’t be eligible for a few years.

SS-
Monte Ward 409 WS (51, 51, 31); 188; 36.31 (also pitcher, so high WS for
pre-1893 pitchers)
Jack Glasscock 261 WS (27, 25, 22); 108; 24.36
Ed McKean 221 WS (25, 25, 21); 99; 21.65
Sam Wise 157 WS (22, 19, 17); 84; 21.65

LF-
Jim O’Rourke 305 WS (25, 24, 24); 103; 27.85
Harry Stovey 265 WS (28, 28, 27); 121; 28.88
Tip O’Neill 213 WS (36, 28, 27); 137; 32.73
Charley Jones 160 WS (27, 24, 20); 98; 29.42
Abner Dalrymple 150 WS (25, 23, 18); 86; 25.55

CF-
Pete Browning 225 WS (30, 28, 23); 118; 30.81
George Gore 250 WS (30, 26, 24); 111; 30.91
Mike Griffin 245 WS (30, 23, 23); 109; 26.26
Paul Hines 249 WS (28, 22, 19); 98; 27.23
Bill Lange 139 WS (29, 24, 21); 114; 27.76
Curt Welch 165 WS (26, 24, 22); 106; 24.14
Jake Stenzel 126 WS (28, 24, 24); 111; 26.64
Tom Brown 218 WS (31, 20, 19); 90; 19.77

RF-
King Kelly 278 WS (35, 24, 24); 130; 30.95
Sam Thompson 236 WS (29, 28, 22); 114; 27.17
Mike Tiernan 251 WS (28, 26, 26); 124; 27.54
Oyster Burns 196 WS (28, 26, 25); 113; 26.74
Tommy McCarthy 170 WS (24, 24, 22); 110; 21.60
Orator Shaffer 120 WS (28, 14, 14); 68; 23.08

P-
Amos Rusie 293 WS (56, 41, 40); 205; 29.11 (question on eligibility under token appearance rule (3 G, 2 GS, 22 IP in 1901).
John Clarkson 398 WS (62, 61, 51); 249; 32.69
Old Hoss Radbourn 346 WS (89, 59, 49); 249; 30.19
Tim Keefe 413 WS (70, 47, 42); 236; 30.06
Tony Mullane 401 WS (58, 55, 46); 183; 29.45
Bob Caruthers 338 WS (57, 54, 51); 255; 32.72
Tommy Bond 243 WS (60, 50, 47); 225; 31.50

I really think he underrates the players from the 1870’s and early 1880’s. Just an opinion. I can’t wait until I get the Win Shares book, so I can game adjust the seasonal numbers to get true peaks for the older players. I do see his point that it was inferior competition, just don’t think their accomplishments should be discounted. The older players would have benefitted from the development of the game as well. I strongly discount the “time-machine” method of evaluation.

JoeD has the Imperial March Stuck in His Head Posted: April 05, 2002 at 03:39 PM | 24 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. MattB Posted: April 05, 2002 at 05:04 PM (#509716)

Excluding the National Association completely eviscerates Ross Barnes, who was probably the best player from 1871-1876 (OPS leader in '72, '73, and '76) before becoming a victim of a rule chance.

Barnes is above McPhee at second base on "HOF Standards" and comparable on "HOF Monitor". He looks to be my top choice for pre-1900 second basemen, and wouldn't want him to get lost among the Jamesiana.

2. scruff Posted: April 05, 2002 at 05:42 PM (#509717)

Matt, I'm not so high on Barnes. He was a great player for 6 years, after that he was pretty much irrelevant.

He was the victim of a rule change, but he was the victim of a rule that makes the game what it is now, I don't cut him any slack for that, but I do think he should get credit for what he did before the change. He was 21 in 1871, so it's not like he was 30 and would have had a great career had baseball existed before 1871 (i.e. Joe Start).

I can't bring myself to rank him above someone like McPhee who was good for so long. His peak value is up there with the greatest ever, but his career value isn't too great.

He's a very intersting case, one of the trickiest we'll have to deal with.

3. John (You Can Call Me Grandma) Murphy Posted: April 06, 2002 at 07:57 AM (#509718)

I like Long Levi Meyerle. Not that I necessarily would vote for him, but he was a damn good player for the NA that even some knowledgeable fans don't even know who he was. Does anyone no why he left baseball while he was still productive?

4. John (You Can Call Me Grandma) Murphy Posted: April 06, 2002 at 08:02 AM (#509719)

I also would move Brouthers to the top of the first basemen list (if we are discounting the NA stats for Anson).

5. John (You Can Call Me Grandma) Murphy Posted: April 06, 2002 at 08:10 AM (#509720)

Joe said: (I strongly discount the "time-machine" method of evaluation.)

6. scruff Posted: April 09, 2002 at 04:09 PM (#509722)

John -- I lean towards Cap, because I think we should count the NA.

I'm glad you agree w/my White and Sutton evaluations, I'll be lobbying hard for them, White on the first ballot and Sutton not too long after.

I have no clue about the dimensions of the Jefferson St. Grounds in 1876. I would say that 160 lbs. isn't that small for a player back then though. It's below avg., but not small.

Do you have a Total Baseball handy (I don't right now)? The info might be in there in the ballparks section.

7. scruff Posted: April 09, 2002 at 04:11 PM (#509723)

"Not that I would vote for him in a million years (damn crook)..."

I don't think that should have any impact on his election, unless he was throwing games.

8. John (You Can Call Me Grandma) Murphy Posted: April 09, 2002 at 08:28 PM (#509724)

Scruff: Wasn't Hall part of the Louisville gambling scandal of 1877? I believe he was thrown out of baseball for that.

9. scruff Posted: April 09, 2002 at 08:44 PM (#509725)

Interesting John. I always wondered why his career ended so abruptly at such a young age.

While I've studied the players of that era a bunch, I don't know too much about the stories. It's on my to do list to read up on those things, but it hasn't happened yet. What I'm trying to say is that I've never heard of the Louisville scandal. Did Bill James include it in HBA I, on his gambling history? I've read that, but I don't remember this story. Please provide details if you have any. Thanks!

10. John (You Can Call Me Grandma) Murphy Posted: April 10, 2002 at 04:45 PM (#509726)

On page 18 of NHBA, he lays the whole story out. Hall (who James compares to Hal Chase unfavorably), Jim Devlin, Bill Craver and Al Nichols were all thrown out for betting on games (or "hippodroming).

11. Rob Wood Posted: April 10, 2002 at 06:25 PM (#509727)

Regarding a timeline adjustment, I think James botched it quite badly. He adjusted ALL players in an era by the same amount. This runs counter to the standard relativistic methods employed throughout sabermetrics of comparing players to their own league averages. I have elsewhere described what I think is the proper way to take into account the era. I adjust players relative to their own league averages to reflect the fact that it was easier to stand apart from your contemporaries long ago. This is due to the "sparseness" of quality players, not the lower quality per se.

Using data on the distribution of OPS+ of all regulars, I have estimated the following timeline adjustments (1871 is set to 100 as the baseline): 1871=100, 1880=130, 1890=140, 1900=147, 1910=152, 1920=156, 1930=159, 1940=162, 1950=164, 1960=166, 1970=168, 1980=170, 1990=172, 2000=174.

As an example of how to use these numbers, say that Joe Blow compiled an OPS+ of 150 during the 1870's and John Doe compiled an OPS+ of 140 during the 1970's. To see whose OPS was "better" compared to his league, taking into account the greater ease players 100 years ago had in standing out from the crowd, first find each player's respective timeline adjustment from the above table. Taking the averages of the decade markers, let's say that Joe Blow has a timeline adjustment of 115 and John Doe has a timeline adjustment of 169.

Joe Blow's OPS+ of 150 is then adjusted as follows. Since we adjust only the amount that Joe is above average, we divide 50%/(174/115) to get 33.0%, meaning in this "era sparseness neutral" sense (now making 2000 the "baseline"), Joe's OPS++ is 133.0. Similarly, John Doe becomes 100 + 40%/(174/169) = 138.9. So we see that John Doe is deemed more impressive. The same type of adjustment can be made to any variable that can be expressed relative to the league average, not just OPS.

Final comment. Looking at the timeline adjustment factors, these confirm the fact that play in the 1870's was definitely "primitive", at least as reflected by the distribution of players' performances.

12. scruff Posted: April 10, 2002 at 08:16 PM (#509728)

Thanks, John, I'll read that tonight. I thought I read that whole book, but I guess I missed that.

Intersting post Rob. I'm curious, how did you come up with those numbers? They seem to make intuitive sense, but there's more to the puzzle, such as AA vs. NL in the 1880's, NA 1871-75 vs. NL 1876+, etc. Would you be able to break your numbers out by league and season, maybe to 1 decimal (since you only have 2 "points" difference over a 10-year period)?

13. Rob Wood Posted: April 10, 2002 at 08:44 PM (#509729)

Here's an answer to scruff's question. For each league-season I calculated the standard deviation of OPS+ among regulars. I then divided that by the average OPS+ among regulars. I use this ratio as my measure of "sparseness". So I have this measure for every league-season going back to 1871.

Since this is a lot of data and, for my previous purpose, I was only looking for the general relationship, I then smoothed the data by estimating a parabola. The decade points of the fitted parabola were reported in the previous post.

I would be happy to make the season-by-season data available to the group, but I am not sure what is the best (and easiest) way to make that happen. Does anyone have any ideas? Thanks much.

14. scruff Posted: April 10, 2002 at 08:56 PM (#509730)

Rob -- if you send me the spreadsheet (send to the email in this post, different from others), I can post it to mostlybaseball with a link. Thanks for offering.

I probably won't get to that until next week however, I'm heading to Boston this weekend (leaving tomorrow straight from work) to watch Wells show Pedro how to pitch.

15. scruff Posted: April 15, 2002 at 02:26 AM (#509731)

Thanks for sending the spreadsheet Rob.

A link to it is posted on the front of the Mostly Baseball page, go to the homepage url above.

The data is quite interesting, although I'm not too comfortable explaining it yet. If you'd like to post something of explanation or something Rob, feel free.

16. John (You Can Call Me Grandma) Murphy Posted: April 16, 2002 at 07:15 AM (#509733)

Very impressive Rob. I do have a question though. If we were to use this or any other form of standardization, wouldn't the bad hitters from a high standard deviation era actually look better than a comparable one of today.

If we were to take Joe's brother from the 1870s Moe Blow, he has a OPS+ of 0.50. Now we take John's brother Bo Doe from the 1970s, he has a OPS+ 0.40. If we do the conversions, Moe Blow's adjustment is now 33 (or 33% worse than average), while Bo's is now 35. Our fiend from the 1870s has actually improved! Does this make sense?

Since we are not going to worry about the Bill Bergens or Hot Rod Fords for the Hall, this won't matter. It will matter though when we are averaging below average seasons of great or near-great players. Ty Cobb's rookie season would actually look better now than it did then, which would alter his adjusted career numbers.

I've been working on a method which is based on standardization methods (such as Michael Schell), but with a difference. The difference between the league-leading player and everybody else has to have the same percentage difference between them.

For example, let's take compare 1871 to 1994. Levi Meyerle had a .492 BA, while Tony Gwynn ha "only" a .394. If we were to standardize their records (with the baseline the year 2000), Levi's BA would be .395, while Tony now equals him.

Now, let's compare Gene Kimball's sparkling BA of .191 from 1871 to Scott Servais' robust .195. The formula I would use is (League Leading BA/x BA : x being the other player in question). For Kimball, I would do this: .494/.191=2.59; for Servais, I would now do: .394/.195=2.02. Since Meyerle would be like .395 in 2000, Kimball's BA would now be .153 (.395/2.59). Servais' BA would round out to .196.

I only took one class in statistics in college, so it's quite possible I'm screwing something up. You seem to be more analytically minded than me, so I thought you might be able to see any flaws in my logic Rob.

I can finish up my pre-1900 database by week's end, If you would like Scruff (Joe). I can compile the BA, SA, OBA and OPS of all the players that we are looking at. I have it on Excell.

17. John (You Can Call Me Grandma) Murphy Posted: April 16, 2002 at 07:29 AM (#509734)

The second-to-last line from the second paragraph should have read: Our FRIEND from the 1870s has actually improved!

18. scruff Posted: April 16, 2002 at 08:17 PM (#509735)

Send it along John, that'd be great, the more info the better.

As to your method, I'm not sure I agree with using the league leader. What if one player just has a hell of a year? I think a system that compares players to each other (the ones that stay in the league) and adjusts for age, etc. is pretty good.

I also like comparing pitchers' offense (I still like using RC/27 or XRuns/27 outs adjusted for league or something instead of OPS, but that's minor) to the league average as a method. It shows what ordinary people would do against the standard competition, which is a good way to see how much the league has improved, since ordinary people (or pitchers hitting) would be a good "control" group.

19. scruff Posted: April 16, 2002 at 08:17 PM (#509736)

Actually John, send anything to the e-mail address in this link, not the other.

20. John (You Can Call Me Grandma) Murphy Posted: April 16, 2002 at 08:40 PM (#509737)

Scruff: I'm actually using the standard deviation of all the players for a particular season (similar to Michael Schell's formula in his book). Once I have the SD, then I find out the standardization for the league leader. Then I compare a particular player to the top batsman. I agree it would be idiotic just using the top batter. I should have been more clearer - sorry!

The point is to use the normal standardization formulas (Rob's a good one), THAN use mine for the players other than the number one guy. All the rest of the players have to be in line with the Cobbs or Gwynns, or it won't be historically accurate. If a player was 20% worse than Wagner in 1908, he still should be 20% worse compared to Honus when we standardize his numbers in 2000.

21. John (You Can Call Me Grandma) Murphy Posted: April 16, 2002 at 08:46 PM (#509738)

(I also like comparing pitchers' offense (I still like using RC/27 or XRuns/27 outs adjusted for league or something instead of OPS, but that's minor) to the league average as a method. It shows what ordinary people would do against the standard competition, which is a good way to see how much the league has improved, since ordinary people (or pitchers hitting) would be a good "control" group.)

My concern with using pitcher's offense is: is a decline a result of the overall improved hitting of position players or that pitchers today just don't get the ABs anymore? If pitchers were at bat as many times as Guy Hecker was in the 1880s, would they be as impressive relative to the position players?

Thanks for the e-mail address Scruff!

22. KJOK Posted: April 23, 2002 at 11:04 PM (#509739)

"Does anyone know what the dimensions for the Jefferson Street Grounds for the 1876 Philadelphia Athletics? I'm trying to see if George Hall was helped by short foul lines to propel him to the HR title. He hit all of them out of the park (I believe), but was a very small man that weighed 160 lbs."

Am fairly certain the Jefferson Street Grounds was not enclosed by an outfield fence, so there really are no "dimensions" to speak of...

KJOK

23. John (You Can Call Me Grandma) Murphy Posted: April 24, 2002 at 07:46 AM (#509740)

Thanks for the info KJOK!

24. User unknown in local recipient table (Craig B) Posted: April 29, 2002 at 02:21 PM (#509741)

Not having fences, of course, helps HR totals nearly as much as shotr fences, provided the outfield is mown.

Gappers can roll a long way otherwise.

You must be Registered and Logged In to post comments.

<< Back to main

Friday, April 05, 2002

The New Historical Baseball Abstract

Reader Comments and Retorts

BBTF Partner

Support BBTF

Bookmarks

Hot Topics

Syndicate