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

Sunday, August 03, 2008

Election Results: Greatest Third Basemen? Schmidt! Mathews! Brett! Boggs!

By unanimous opinion, Phillie legend Mike Schmidt was picked as the greatest third baseman by our electorate.

With 93% of all possible points, Braves slugger Eddie Mathews made a strong showing himself to finish in second place.

Kansas City great George Brett narrowly bested Red Sox and Yankee star Wade Boggs 86% to 84% for 3rd and 4th place respectively. They were the last candidates to finish with at least 75% support.

RK  Player           PTS  Bal   1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18
-----------------------------------------------------------------------------------------
-----------------------------------------------------------------------------------------
 1  Mike Schmidt     414   23  23                                                   
 2  Eddie Mathews    387   23     22           1                                    
 3  George Brett     357   23      1 13  7  1  1                                    
 4  Wade Boggs       347   23         6 13  4                                       
 5  Frank Baker      305   23         4  1  5  4  7     2                           
 6  Jud Wilson       279   23               8  6  3  2     1  2     1               
 7  Ron Santo        271   23            1  3  5  6  4  2     1        1            
 8  Paul Molitor     236   23               1  4  4  1  5  4     2  1  1            
 9  Dick Allen       215   23                  1  3  6  2  6  2        1  1        1
10  Heinie Groh      174   23                        2  3  3  5  3  4  1     1  1   
------------------------------------------------------------------------------------
11  Jimmy Collins    165   23            1     1     2  1  1  4  3  4     2  3  1   
12  Ezra Sutton      157   23               1        2  2  2  1  5     3  4  3      
13  Brooks Robinson  156   23                        2  2  2  2  2  6  3  3  1      
14  Darrell Evans    134   23                        1  3     2  2  3  4  3  3  2   
15  John Beckwith    129   23                        1     2  2  2  2  6  6  1     1
16  Stan Hack        122   23                           1  2  2  4  2  2  3  3  3  1
17  Graig Nettles     43   23                                             1  3 11  8
18  Ken Boyer         42   23                                          1     5  5 12
Ballots Cast: 23

Thanks to OCF and Ron for their help!

John (You Can Call Me Grandma) Murphy Posted: August 03, 2008 at 11:34 PM | 46 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. John (You Can Call Me Grandma) Murphy Posted: August 04, 2008 at 02:02 AM (#2888943)
hot topics
   2. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 04, 2008 at 02:09 AM (#2888951)
Wow, Jud Wilson finished 6th even though his highest individual vote placement was 7th.
   3. OCF Posted: August 04, 2008 at 02:10 AM (#2888952)
No, John posted the chart with some padding. Wilson's highest vote placement was 5th, and he got 8 of those.
   4. OCF Posted: August 04, 2008 at 02:14 AM (#2888957)
The average consensus score was 84. Here are the individuals

sunnyday2: 90
Devin McCullen: 88
Tiboreau: 88
Chris Cobb: 88
Rafael Bellylard: 87
Mark Donelson: 87
Dan R: 87
Howie Menckel: 87
ronw: 87
andrew siegel: 86
Rick A: 86 (median)
OCF: 85
DL from MN: 84
mulder & scully: 84
whoisalhedges: 83
Esteban Rivera: 82
Sean Gilman: 82
John Murphy: 80
Joe Dimino: 80
AJM: 79
Rusty Priske: 79
TomH: 78
bjhanke: 77
   5. OCF Posted: August 04, 2008 at 02:24 AM (#2888965)
Oops, I misplaced the median - actually, I'm the median score.
   6. John (You Can Call Me Grandma) Murphy Posted: August 04, 2008 at 02:26 AM (#2888966)
No, John posted the chart with some padding.


Actually, the ballot counter screwed up again. At least it doesn't look like it did last election.
   7. John (You Can Call Me Grandma) Murphy Posted: August 04, 2008 at 02:29 AM (#2888970)
BTW, can somebody explain Jimmy Collins over Ezra Sutton? Sutton destroys him in career length and his peak was comparable to Collins, too.
   8. AJMcCringleberry Posted: August 04, 2008 at 03:04 AM (#2889007)
I was right, I am the only one who likes Boyer.
   9. sunnyday2 Posted: August 04, 2008 at 03:08 AM (#2889018)
OK, I was last on the SS consensus score, now 1st on 3B. What's with that?
   10. OCF Posted: August 04, 2008 at 03:23 AM (#2889035)
How do we compare so far to the 1999 fan-voted "All-Century Team"? So far we've worked through the infield positions. The All-Century Team had two at each infield positions, except in the instances in which a "panel of experts" intervened and added an extra player. Position by position:

Catcher:

ACT: 1. Bench. 2. Berra.
HoM: 1. Gibson. 2. Bench.

And we had Berra third, and Gibson did get significant fan support. In this case, our only real disagreement with the fans is on the evaluation of the Negro Leagues.

First base:

ACT: 1. Gehrig. 2. McGwire.
HoM: 1. Gehrig. 2. Foxx.

We had McGwire 11th.

Second base:

ACT: 1. Robinson. 2. Hornsby.
HoM: 1. Collins. 2. Hornsby.

We had Robinson 6th.

Third base:

ACT: 1. Schmidt. 2. Robinson.
HoM: 1. Schmidt. 2. Mathews.

We had Robinson 12th. He clearly had no business being on the ACT. And Mathews may be the most quickly forgotten superstar of them all.

Shortstop:

ACT: 1. Ripken. 2. Banks. X. Wagner.
HoM: 1. Wagner. 2. Lloyd. 3. Ripken.

We had Banks 11th, and with some of our more vocal members arguing that that was too high.

The ACT lumped all three outfield positions, and elected more of them than they did infielders. Here's the ACT outfielders - what changes will we have?

1. Ruth
2. Aaron
3. Williams
4. Mays
5. DiMaggio
6. Mantle
7. Cobb
8. Griffey
9. Rose
X. Musial

OK, Griffey is ineligible for us - but we wouldn't put him there anyway.
   11. Tiboreau Posted: August 04, 2008 at 04:29 AM (#2889100)
I was right, I am the only one who likes Boyer.

No, BP's WARP likes him, too. . . .
   12. Paul Wendt Posted: August 04, 2008 at 05:27 AM (#2889142)
Nettles and Boyer seem to be outliers.
I wonder whether Tommy Leach and Pie Traynor would finish last, if included.
or Ray Dandridge.

The ACT lumped all three outfield positions, and elected more of them than they did infielders. Here's the ACT outfielders - what changes will we have?
The All-Century ranking puts five centerfielders at #4-8. We would need a runoff for comparability. With the tendency for blackball stars to play center, we would probably get a strong showing in the center also, even with Joe Jackson eligible in place of Ken Griffey.
   13. Joey Numbaz (Scruff) Posted: August 04, 2008 at 11:18 AM (#2889209)
Was my ballot counted? I voted Sutton 5th . . .
   14. bjhanke Posted: August 04, 2008 at 11:20 AM (#2889212)
I don't mind having my Beckwith vote declared "unconstitutional," because I simply couldn't get a handle on him. Lack of time is lack of time, and I was away from home and on vacation. My off-consensus votes for Baker and Collins are things I am much more sure of, but if they seem just too wrong, you can discard them, too. A quick note for Grandma - my methods of approach to very early baseball lead me to give Sutton credit for about 17 real seasons. That and Collins' revolution of his position are why I have Ezra so far behind Collins. I do have reasons for this - although it's possible no one will accept them - but I need sleep and time to get the post written. Oh, also, I apologize to everyone for getting Deacon White wrong back in catcherland. When I applied the ISO test I used on Collins to White, it was immediately obvious that I was wrong about the Deacon. He was not a slap hitter in the context of his time. I wish I had discovered the tool earlier. - Brock
   15. Joey Numbaz (Scruff) Posted: August 04, 2008 at 11:21 AM (#2889214)
Oh wait . . . the heading is off, I'll fix it.
   16. Joey Numbaz (Scruff) Posted: August 04, 2008 at 11:26 AM (#2889216)
I just can't fathom Hack over Nettles by such an easy margin, wow.
   17. Paul Wendt Posted: August 04, 2008 at 11:28 AM (#2889217)
Sutton enjoys a 5th-place vote from one of his best friends, just enough to float him over the sunken good ship Brooks.
A small crowd votes Brooks #13 with Ezra just above or just below.

The report uses columns 3-20 for eighteen.
   18. Joey Numbaz (Scruff) Posted: August 04, 2008 at 11:30 AM (#2889218)
Who had Eddie Mathews 6th, wow!
   19. Paul Wendt Posted: August 04, 2008 at 11:54 AM (#2889223)
16. Joe Dimino Posted: August 04, 2008 at 07:26 AM (#2889216)
I just can't fathom Hack over Nettles by such an easy margin, wow.

Partly that is the same old career dimension.
. . . Joe, you bumped Santo above Evans else you would have been the best friend of Ezra, Darrell, and Graig.
Maybe "pennants added" now amounts to the lowest replacement level in the electorate.

The report shows a comfort zone for Beckwith at #16-17 ahead of Nettles, Boyer and "one or two others" but no agreement and probably discomfort about which one or two others.
   20. Joey Numbaz (Scruff) Posted: August 04, 2008 at 12:17 PM (#2889237)
Yeah, that makes sense Paul. With Dick Allen finishing 9th, it sure seems like the peak guys are the ones that stuck around for this exercise.
   21. TomH Posted: August 04, 2008 at 12:29 PM (#2889254)
True, Joe.

If we announced to the world that we have Brooks Robinson ranked as the 13th best 3B ever.. well, I would not wish to be the guy at the microphone. Especially in my neighborhood!
   22. John (You Can Call Me Grandma) Murphy Posted: August 04, 2008 at 01:23 PM (#2889288)
I wonder whether Tommy Leach and Pie Traynor would finish last, if included.
or Ray Dandridge.


I would have had Traynor above, at the very least, Boyer and Nettles, FWIW.

If we announced to the world that we have Brooks Robinson ranked as the 13th best 3B ever.. well, I would not wish to be the guy at the microphone. Especially in my neighborhood!


I knew a guy would claimed Thurman Munson was vastly superior to Johnny Bench. After awhile, I just didn't get involved in that conversation with him any more. Of course, his opinion made more sense than another person I knew who though first base was the most difficult position, which made Don Mattingly the greatest player of his era.

Yes, both fans were Yankees. :-)
   23. John (You Can Call Me Grandma) Murphy Posted: August 04, 2008 at 01:24 PM (#2889289)
Oh wait . . . the heading is off, I'll fix it.


Thanks, Joe!
   24. Sean Gilman Posted: August 04, 2008 at 09:35 PM (#2889759)
Who had Eddie Mathews 6th, wow!


That was me. I had no idea everyone else thought it was so obvious.
   25. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 04, 2008 at 09:40 PM (#2889761)
I challenged Sean's vote asking for an explanation, but none was forthcoming.
   26. Sean Gilman Posted: August 04, 2008 at 09:55 PM (#2889777)
You did on the last day of voting, five days after I posted my ballot. If you're going to challenge votes, especially from people with an inordinately long voting history, you might want to give them more than a few hours to respond before whining about how responses to your questions aren't immediately forthcoming.

I use WARP1, with Brett and Boggs ahead due to quality of competition and Santo and Baker ahead on peak (Best 5 years and Best 7 consecutive years). Mathews, has I noted on my ballot, has the second or third most career value, but I felt the other two factors were enough to move the other four guys ahead of him.
   27. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 04, 2008 at 10:04 PM (#2889785)
I hadn't checked the thread in the interim. I wasn't whining, just asking to hear where you were coming from.
   28. Sean Gilman Posted: August 04, 2008 at 10:13 PM (#2889796)
Sorry if that was too testy. I'm still recovering from a long, weird weekend.
   29. OCF Posted: August 04, 2008 at 10:26 PM (#2889817)
BTW, can somebody explain Jimmy Collins over Ezra Sutton? Sutton destroys him in career length and his peak was comparable to Collins, too.

With his 4th place vote for Collins and 15th place vote for Sutton, bjhanke accounts for all of that vote difference all by himself. Take away Brock's vote and Sutton would have been 11th with 153 points with Groh and Collins tied for 12th with 150 points, with that cluster well behind Allen (203) and well ahead of Evans and Beckwith (129, 128).

Other major Collins over Sutton voters: Esteban Rivera (6, 14), whoisalhedges (8, 15), Howie Menckel (9, 15), Sean Gilman (8, 14), balanced by the (mostly milder) Sutton over Collins voters: Joe Dimino (13, 5), John Murphy (13, 9), Rafael Bellylard (16, 12), Rick A (11, 8), andrew siegel (11, 8), Chris Cobb (15, 12), OCF (15, 12). Well, I guess Brock and Joe nearly cancel each other out there.

Very few people had Boggs and Brett more than one vote apart. Even Sean, with his idiosyncratic Mathews ranking, just moved Boggs and Brett up one notch, together. Hence Boggs versus Brett turned out to be pretty much a straight head-to-head vote between those two, which Brett won but which was far from unanimous. Many voters indicated in their comments that they thought Boggs versus Brett was a close call.
   30. Howie Menckel Posted: August 05, 2008 at 01:34 AM (#2890131)
I have Collins as a historically, staggeringly great 3B defensively.
Ironic, because I am a defense-heavy guy less than most voters, but this was an extreme example to me.
Sutton has that weird trough; I'm not a 'consecutive seasons' voter, but he's an odd case.
   31. NJ in DC Posted: August 07, 2008 at 09:22 PM (#2894566)
Where could A-Rod potentially end up on this list?
   32. OCF Posted: August 07, 2008 at 09:32 PM (#2894576)
A-Rod is still on the SS list and will probably stay on the SS list. On that list he seems reasonably likely to be 2nd overall. See posts #27 and 28 here.
   33. NJ in DC Posted: August 07, 2008 at 09:35 PM (#2894578)
Thanks.
   34. bjhanke Posted: August 13, 2008 at 04:33 AM (#2900806)
This is the first post of what turned into an essay on FSEs for very early players (and high-SD leagues overall) and the limits of said FSEs. What started as an essay comparing Ezra Sutton to Jimmy Collins and Home Run Baker has turned into an essay. It has been shortened considerably by Don Malcolm turning me on to Baseball-Reference.com. Best $29 I ever spent on baseball research materials.

A couple of things you should know first. For those of you with real math backgrounds, I think a bit differently than you are trained to think. I was really an engineer in college. My degree says Mathematics, but that's just because Vanderbilt U. didn't allow the Applied Math department in the engineering school to offer a major. So I was trained to use math as just one of many tools, especially including Plausbility Tests, which are ad hoc comparisons designed to keep raw math results under engineering control. In general, what they teach engineers is that it's very important to ask all (or most) of the good questions, let the mathematicians answer them, and then plausibility or reality test the math results to make sure the question didn't lead the mathematician astray. Not his fault you asked a bad question. I think like that by personality, not just training. So just a fair warning: this essay may drive you hard-formula guys nuts. But then, so do my ballots, apparently.

The second thing you guys should all know, and which you would know if I had not found the Hall of Merit just as it was starting this balloting, is that I did a lot of work on the 19th century in the early and mid 1990s. Two reasons. First, a magazine called Gravengood's commissioned me to write a series of essays on the history of baseball, decade by decade. For each decade, I promised to make up an all-decade team and an all-time-to-date team. That is, for the 1920s, I had to make an all-1920s team and an all-1871-1929 team. The only part of this that got published before the mag folded was the 1800s part, so I did that work. The other time I looked at the 19th c. was when the Historical Abstract came out and I realized immediately that the 19th c. rankings were all too low. The first thing I figured out to do was what I think this group calls FSEs (full season equivalencies, right?). This produced results that were too high. At the time, I was trying to correct, not avoid, Bill's timeline, so I realized that the timeline was actually too shallow when dealing with the 19th. Bill just didn't FSE first, or something.

Anyway, what all that means is that I did a lot of 19th c. analysis with tools that would be considered hopeless now, in the days of databases. Basically, I had to hand-copy spreadsheets of whatever sample was small enough for me to be willing to type in. When I looked at things like how many seasons a 19th c. third baseman could actually play after FSE, I know I missed people. Hell, I missed people in the 1920s. But, thinking like an engineer, that didn't bother me as much as it might. I was used to plausibility tests. Entire detailed databases and curve-fitting algorithms are not my idea of necessities. Plausibility Testing is.

As I said, I really should have written this stuff up long since. Some of the comments I'm getting indicate that the group here is not sure how I actually think and work. That would not be true if I had had a couple of months to settle in before filling out a catcher ballot. Anyway, enough of the confessional. On to the work as it stands now, related to Ezra Sutton, Jimmy Collins and Home Run Baker, and with the help of BB-Ref's available sorts.

Sigh. Baseball Reference. I ran a list of career games played at any position on all players who played at least 50% of their games at third base (dropping the percent didn't do much except pick up Tommy Leach; raising it dropped off Deacon White and then Ezra Sutton himself) between 1871 and 1930. That ought to catch anyone who could even remotely be considered a contemporary of Sutton, Collins, or Home Run Baker. Who was the leader? Lave Cross, at 2275 games played. Cross is early enough that FSE would raise that some; this is the actual games played. Second is Larry Gardner, at 1923 games, or 342 games fewer than Cross. That's a huge drop. If you drop 342 games off of Gardner, you get Jimmy Austin, at #8. That means that Cross is himself the outlier (am I using this term correctly?). His games played are way ahead of anyone else's.

And then there's Ezra Sutton FSE. 2643 games played FSE. 368 more than Lave Cross. Yes, that's right. The drop from Sutton FSE to Cross actual is more than from Cross actual to Gardner actual, which is the same as from Gardner to the #8 guy among the actuals. Do you really think this is true? That the greatest credit for number of games played by any third baseman through 1930 should go to a player from the 1870s when the infields were trash and fielders didn't wear gloves. Really? And he's an enormous outlier on top of a real outlier? This is what I mean by failing a plausibility test.

But wait; it gets worse. Deacon White is out there. I ran the FSE on White. 2938 games played FSE. Almost 300 more than Ezra Sutton. So it's not just Ezra. White FSE has 563 more games played than Cross does actual. Surely you have given up buying this by now. It's just not plausible. Not reasonable in any way. White's FSE is a quarter more games played than the leader in actuals, who is himself an outlier. I'm sorry. Unadjusted FSEs are just not reliable when dealing with very early players, including third basemen. Oh, yeah, for those of you who want to complain that Deacon White was a catcher. I ran the same list for 50% catchers. The leader in actuals? Wally Schang, at 1812 games played. You FSE Deacon White and you get a 62 percent leap over Wally Schang. Just. Not. Reasonable. The real lesson here is that you simply cannot run blind FSEs and trust the results. You have to make a percentage adjustment, or you have to apply a cap of some sort.

Well, if I haven't convinced you by now. I'm not going to, so let's get on to the next two questions: What causes blind FSEs to be so unreliable and what would be reliable? That's the next post.
   35. bjhanke Posted: August 13, 2008 at 04:34 AM (#2900807)
This is the second post of a three-post set. If you haven't read post #1, please do that or this one ain't gonna make any sense. It should be right above this one on the thread list.

Back in the 1990s, I figured at least two reasons why blind FSEs would not be reliable and then stopped compiling the list, as the important thing is to know that there are more than one reason. The first reason is wear and tear on the body. I've never placed much credence in the concept that a player has a certain number of games his body can stand to play, because there are no athletes who don't start baseball until they're 30 and then last in the bigs until they're 50. But there really does seem to be a factor there of some sort that results in a cap on how much you can FSE a player. And yes, it is related to the era. When I ran the 50% third base sort for all time. Brooks Robinson rises to the top and Lave Cross is in 11th place. It was easier to play third later than earlier.

The second reason is more mathematical, but harder to explain, as it's not directly related to FSEs. It's related to how many years an early player could actually play at what was in his time a major league level. Note that this is about years, not games. What I realized first about the very early guys is that they start too early and end too late to compare to later guys without some sort of adjustment. Too many 20-year-olds. Too many 40-year-olds. Why?

What I think is involved is what I call the Standard Deviation / Percentage Dissonance (SD/% for short). Again, this is something from the 1990s that I can't define yet for lack of curve fitting skills and a real database. The idea is this: Replacement Rate, which we commonly express as a low winning percentage (your choice), can also, and maybe better, be defined by a number (or fraction) of standard deviations below mean performance. I would guess about one SD, but that's nothing more than a guess based on a few observations. Then there is what I call the Starting Line, which is the Replacement Rate concept applied to starters. This is the line below which you are still a major league ballplayer, but not a starter. That's the Starting Line and it, too, seems to be expressible as some fraction of a standard deviation below the mean.

So far, pretty easy. But here's the rest: I took a look, as best I could, at the rate at which players gain quality when young and lose it when old. These rates do not appear to be expressible in terms of SD. They appear to be expressible in terms of a percentage of the player's eventual mean performance. I'm not sure that "percentage of mean" is the exact formula, because I don't have the database or curve fitting skills needed to do that, but I'm pretty sure that player gain and loss move unrelated to SD. Let me use percentage here to save space; if anyone knows the actual formula, please post up. I will note that I am very sure that the rate of increase is faster than the rate of decline. I just don't have a hard math formula to offer. Percent of mean is my best guess.

What happens, then, in a high-SD league? Well, if player gain and loss is not related to SD, regardless of what the actual formula is, then a player in a high-SD league will reach Replacement Rate and the Starting Line earlier than a player in a low-SD league. And when he declines, he will last longer before he reaches the two rates and eases into retirement. This is, in my opinion, what happened to players like Cap Anson, Deacon McGuire, Orator Jim O'Rourke, and also Ezra Sutton. They may have been playing in conditions that wore their bodies down more than later conditions would have, but they also could play at an earlier age and last until an older age than modern players of the same mean quality before reaching the Starting Line and Replacement Rate.

And when you take careers that have too many seasons in them, and then FSE, what do you get? Too many games, which is exactly what we do get when we do blind FSEs. That's where SD/% relates to FSEs being too large.

OK, that's two reasons out of what may be many, why blind FSEs don't work in the 19th century. So what would be reliable? Well, of course, it would depend on the year. Each year (more or less), the SD gets a little lower as the game matures. So the dissonance between SD and the formula (% or whatever) narrows. You need a separate cap (or reduction formula) for each year. It also turns out that you need a separate cap for at least some positions, most notably catcher. Remember the 62% jump between Schang actual and White FSE? That's much worse than third base. But third base is bad compared to some other spots.

Here's what I use, because it passes the Plausibility Test. I take the (now, due to BB-Ref, complete) list of who played how many games at the position during the player's time period, with a little extension because a player's contemporaries don't all start and stop in the same season. So I used 1871-1930 to cover Suton, Collins and Baker. Then I take the top result and throw it out, which is standard statistical practice. At third base, for Sutton, that leaves Larry Gardner's 1923 games. I will FSE Sutton out to that number of games, but no higher, and I'm not even sure that's right. It's just the furthest I will go at all. So, how many seasons does that represent? Well, Sutton played 18 seasons. Gardner's 1923 is 72.8% of Sutton's FSE of 2643 games. So I credit Sutton with 72.8% of 18 full seasons, or about 13-14 seasons. That's for comparison, remember, only to EARLY third basemen. If you're talking all time, well, the second most games played by a primary third baseman all time is George Brett's 2707. That's more than Sutton FSE, much less Sutton actual. So what you have is Sutton FSE looking very high but reasonable on an all-time list, but being revealed as far too high when you restrict the list to players before 1930. Well, which is really his peer group? Larry Gardner or George Brett? It's Gardner, and you can make a case that even Larry is too late. You just have to make large playing time adjustments when comparing early third basemen to later ones. Third was a tough, tough spot to play.

So, in terms of my third base ballot, what I did was consider Ezra Sutton to have a good long career, but not a complete explosion of Rose-like durability. That dropped him well below those of you who took his FSE at face value, or even his actual 18 seasons - some of them less than 30 games long - at face value. In particular, it keeps him from just lapping the career-length field of Jimmy Collins and Home Run Baker.

Which leads me to the next post, which is about Collins' and Baker's career lengths. One hint: there's a reason I haven't revealed the third-place guy on the 1871-1930 list yet.
   36. bjhanke Posted: August 13, 2008 at 04:34 AM (#2900808)
This is the third and last post of a set. If you haven't read the first two yet, please do that, or this one won't make any sense. They should be right above this one on the thread list.

I left you with the unnamed third man on the games-played list of career third basemen between 1871 and 1930. You've no doubt guessed it. It's Jimmy Collins, at 1725, behind Lave Cross (2275) and Larry Gardner (1923).

That's right. It's fact. Jimmy Collins did NOT have a short career. He did not have a medium career. He did not have a long career. He had a VERY long career for third basemen of his era. This is one reason why I have him ranked so highly on my ballot. His times aren't comparable to the times of George Brett or even Brooks Robinson. Within his time, which I defined as a sixty (60) year span, he has the #3 career length at his position. And Gardner isn't really a contemporary; he can better be described as Collins' successor at third for the Red Sox. All time? Collins drops to 40th, which is probably why people think he had a short career. But in his own time? Third. Of course, that means a huge adjustment in perception, which led to my ranking.

Now, I did say above that I would credit Ezra Sutton with Larry Gardner's career length, which is one place longer than Collins'. So what made me rank Collins so much higher on the ballot?

1. Defense. I agree that Ezra Sutton was an excellent third baseman. He played too many games at shortstop to think anything else. And I do think that his ranking in Win Shares (B+) is too low. Bill, who does not count the National Association, considers Sutton to have started his career in 1876, at the age of 25. Defense is a young man's game. I might even give Ezra an A, not just an A-. But Jimmy Collins is Jimmy Collins. He gets a Win Shares A+, and I'm with Bill on that.

2. Sutton's offensive advantage doesn't really exist. His career OPS+ is 119. That's good, but Home Run Baker is at 135, just to give you a standard for comparison. Deacon White is at 127 (and I do regret ever having called him a slap hitter). Collins is the last in this group, at 113, but compared to Sutton, that's only 6 points. I also have suspicions about Collins' ballpark adjustments. I think that his ballparks favored home runs, which he didn't hit many of, but suppressed triples, which were his feature stat. I know nothing about Sutton's adjustments, so I can't do anything but take them as they are. I have no seasons of individual stat splits for Ezra. His ranking might change if I ever get some.

3. Jimmy Collins revolutionized the way third basemen play the position, and that revolution has held up right to date. You occasionally hear about some guy who is supposed to have revolutionized his position. Uusally by some fan trying to get a borderline case into the Hall of Fame. When you check it out, the claim almost never stands up. But it does for Collins. All of his contemporaries agree that he did this. He invented bunt charging and "cutting down the cone," at the very least. We still play the position the way Jimmy Collins did. The most compelling support for me? John McGraw. McGraw, as you all probably know, was famous for claiming that every invention in baseball through about 1920 was made by his teams. And he was the playing third baseman on the Baltimore Orioles of Collins' day. They are close contemporaries. Yet McGraw made no attempt to claim the credit for himself. He freely gave it to Collins. That's a clincher, at least to me. Yet another version of Plausibility Testing.

OK, so much for Jimmy Collins. Now, why do I have Home Run Baker ranked so highly when he has a short career punctuated with two dropout seasons?

1. Offense. Running through the same time frame as above (1871-1930), and putting a minimum of 1000 games in to keep out flukes, Home Run Baker has the second-highest OPS+, at 135, behind only Denny Lyons, who retired at age 32. OPS+ is a rate stat. Retiring early is a minus, because it shaves off the decline phase. Baker retired at 37, after 2 years of OPS+ below 100. Otherwise, he would end up higher than Lyons. BTW, just FYI, Ezra Sutton does rank 6th at this stat, Heinie Groh ranks 7th, Jimmy Collins 9th, and Larry Gardner 11th. Sutton belongs in the group somewhere, but I just can't take his FSEs seriously. Also, I ran the same OPS+ sort, but for 1871-2008. Baker ends up 7th, a fraction behind George Brett. Above him are Brett and Lyons, Al Rosen, who just made the 1000-game cutoff, and Schmidt, Mathews, and Chipper Jones. That is, Home Run Baker is, taking the whole career, the best-hitting third baseman before Eddie Mathews. And, as you all know by now, I do give extra credit for dominating something for a long time. Baker led third base offense for about 80 years.

2. Playing time credit. We all (or most of us) give it for wars. But one of Baker's dropout years was because of the death of his wife and the need to set up a plan to care for his children. I really want to give him credit for some of that year. It counts, to me as being "beyond the player's control." This may be my personal oddity, but mine it is.

3. Baker does not have a short career. His career length, on the scale listed above, is 9th. That's right. Two dropout years or no, Home Run Baker has the 9th-most games actually played by any third baseman between 1871 and 1930. That's not a short career. That's a LONG career. Not a really long career like Collins', but a long one. And so, I apply no deduction at all for it. It's not Baker's fault that third base got easier to play over time. He played when he did. And within the constraints of his time, he has a long career, not a short one.

Well, that's it. I imagine you're all sick and tired of my opinion of FSEs by now. But I really needed to say it, because my odd voting patterns depend largely upon it, and I hadn't had time to establish how I thought before this voting got started and I needed to focus on votes. My timing was just lousy. Sorry about that. But like I said, this needed saying. Orator Jim is coming up in center field, and the same thing is going to happen, I'll just bet.

- Brock
   37. flournoy Posted: August 13, 2008 at 04:46 AM (#2900815)
Where will Chipper Jones land on this list? (If he retired today, or if he finished his career however you expect him to.) I think he hangs with Mathews right now, and by the time he's done, he might give Schmidt a run for his money.
   38. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 13, 2008 at 06:31 AM (#2900843)
I have Chipper 5th as of the end of 2007, behind Schmidt/Mathews/Boggs/Brett. This year gets him pretty close to the 80's B's, and he has a real shot to be #2-#4 by the time he's done. He's still well behind Mathews for me, though.
   39. TomH Posted: August 13, 2008 at 12:07 PM (#2900892)
Thanks, Brock! As a math guy cleverly disguised with an engineering degree, I understand your persepctive. We should all get to chew on this for a while; it cuts to the core of ranking 1870s/80s players against MLBers since then.
   40. Paul Wendt Posted: August 13, 2008 at 06:05 PM (#2901419)
It seems crucial to understand some purposes and concepts. Details posponed.

1.

All,
That 2643 games which Brock calls "Ezra Sutton FSE" is the career games played for Sutton according to the "Neutralize Stats" feature at baseball-reference, using the default 162-game context.

2.

Brock,
Essentially, what are you trying to do with Full Seasons Equivalent and Plausibility? Is it fair to say that you are seeking a point estimate of how much Sutton would have played in the majors if he became a regular thirdbaseman in 1910? That would be place him entirely within the 154-game era and contemporary to Baker and Gardner.

one detail: What is the source for 18 seasons? From me the number should be 16.3# for 1870-88 or 17.2# for 1871-88, or the same with more severe rounding.

3.

All,
I think of full seasons as a measure of playing time that abstracts from games. As for the familiar player batting line, pitching line, fielding lines; as for win shares and WARP, the fundamental scope is one player-team-season. The fundamental range is 0 to 1. As for the player lines and ratings, it may be summed over multiple player-team-seasons, such as and most famously all of the team-seasons in one player's mlb career.

For concrete example here are the full seasons equivalent fielding games at thirdbase for 1876, top ten only
Fielding Thirdbase, NL 1876, fse
1.000 Schafer, Boston
1.000 Ferguson, Hartford
1.000 Anson, Chicago
1.000 Nichols, New York
0.984 Battin, St Louis
0.971 Hague, Louisville
0.816 Meyerle, Philadelphia
0.707 Foley, Cincinnati
0.369 Booth, Cincinnati
0.133 Sutton, Philadelphia

It may be no surprise that the 9th and 10th men in this eight-team league were the leading substitutes for the 8th and 7th men. The Athletics made four fielding switches involving 3B but the full seasons for Meyerle and Sutton sum to less than one because the team relied (enough) on four substitutes for Meyerle. Cincinnati made nine fielding switches and relied enough on Booth as the leading sub that the full seasons for Foley and Booth sum to more than one.
   41. Paul Wendt Posted: August 13, 2008 at 06:10 PM (#2901432)
For Sutton's mlb career 1871-1888 the sum of those "annual" numbers is 11.84, his career full seasons fielding 3B.
11.84 = . . . + 0.133 + . . .
   42. David Concepcion de la Desviacion Estandar (Dan R) Posted: August 13, 2008 at 07:06 PM (#2901589)
This is similar to my SFrac, except that it can reach a bit higher than 1 in cases of very high seasonal PA totals.
   43. bjhanke Posted: August 14, 2008 at 04:00 AM (#2902497)
First, to everyoe, thanks for good responses with no flames! So nice to be in with his group of people.

Second, I am in Indianapolis right now, working at GenCon, so don't have much time to post up. Details starting Monday.

Third, quick for Paul. Thanks for your post, which is helpful to start with and promises more. You are correct on point #1. Right out of BB-Ref. Re #2, I use FSEs to get a start on evaluating an early player (nothing else provides so good a start), and Plausibility testing to keep FSEs from running amok. Using "playing time equal to that of the #2 in-the-time-period's actual games played" is a specific method of plausibility testing. The source for 18 seasons is the simplest one possible: calendar seasons played = 1871-1888 = 18 seasons. Of course, they vary in length and his percentage of each one played varies. No adjusting here, just calendar years. I did count them right, didn't I? It would be so embarrassing if I hadn't.
   44. Paul Wendt Posted: August 14, 2008 at 06:56 PM (#2903188)
some details
"1" "2" and "3" refer to the numbering above. By accident they appear once each here.

1a.
All,
That 2643 games which Brock calls "Ezra Sutton FSE" is the career games played for Sutton according to the "Neutralize Stats" feature at baseball-reference, using the default 162-game context.


Multiply that by 154/162 and you have the "neutral" career games played in 154-game context --without registration and sign-in. For Sutton that is 2512 games.


3a.
It may be no surprise that the 9th and 10th men in this eight-team league were the leading substitutes for the 8th and 7th men. The Athletics made four fielding switches involving 3B but the full seasons for Meyerle and Sutton sum to less than one because the team relied (enough) on four substitutes for Meyerle. Cincinnati made nine fielding switches and relied enough on Booth as the leading sub that the full seasons for Foley and Booth sum to more than one.

Full seasons abstracts from games but I use official games data to calculate it, fielding games at one position (which at the pitcher position is identical to pitching games) or batting games. The measure may be slightly improved for recent seasons by using fielding innings or by using plate appearances and games by batting position. I use fielding and batting games because they are readily available for all of major league history.

Designated hitter is a "fielding position" so the games at all fielding positions may be used together with batting games to underestimate "pinch games". My methods recognize only those pinch-hitter and pinch-runner appearances that are not balanced by position switches for the same player-team-season; that is, only insofar as the sum of fielding games at all positions including DH is less than the sum of batting games.


4.
"playing time equal to that of the #2 in-the-time-period's actual games played" is a specific method of plausibility testing.

Yes.

a.
It is extremely sensitive to the demarcations. For example Tommy Leach played almost 10% more games than Larry Gardner, so counting him as a thirdbaseman significantly relaxes the cap on Sutton's career games played. For another example, the window range 60 years after debut (1870-1930 for Sutton) is just enough to make Brooks Robinson a reference point for Pie Traynor (1921) or Jimmie Dykes (1918), not enough to bring in Nettles and Evans, and not enough to make Robinson a reference point for Larry Gardner (1912).

For the outfielders, if the method is applied to outfielders generally, 1870-1930 is just long enough to make both Cobb and Speaker reference points. That is fairly loose, albeit binding for O'Rourke and perhaps Hines. For 1870-1920 the leaders in games played are Crawford and Clarke(?). Clarke played about 15 full seasons, Speaker about 18, so 1920 in place of 1930 makes a much tighter cap.

b.
Consider the earliest time period, 1870 to 1930 by Brock's current definition. So long as the window is broad enough to include the full careers of a generation that played mainly at 140+ games, the method almost rules out that any early player may be judged to have the greatest longevity at his position. Anson gets credited with a longer career than Beckley only by literally playing more official league games.
   45. bjhanke Posted: August 15, 2008 at 01:25 AM (#2903858)
Paul notes, "Consider the earliest time period, 1870 to 1930 by Brock's current definition. So long as the window is broad enough to include the full careers of a generation that played mainly at 140+ games, the method almost rules out that any early player may be judged to have the greatest longevity at his position. Anson gets credited with a longer career than Beckley only by literally playing more official league games."

Absolutely, and yes, it's one of the weaknesses of plausibility testing. One thing you cannot do with plausibility testing is to just flatten everyone from hish SD leagues. That would not be fair - nor plausible. It took me quite a while to figure out who to work down and whose FSEs to trust more (I ended up trusting Anson's more than anyone else's, which may be why my vote for him does not look out of consensus), and why. Please remember, plausibility testing is about plausibility, not hard fact, and certainly not a be all and end all. It's just a means of controlling hard math estimates like FSE. And the more questions you ask, the better your estimated results will be, as long as the questions are pertinent and realistic. Plausibilty testing is all about asking lots of questions, but I was already up to three posts....

On, yeah. One more thing for Paul. You are right about 60 years being a lot of time frame. I was trying to compare Sutton and Collins and Baker, and needed enough years to include the front end of Ezra and the back end of Frank's contemporaries. For Ezra alone, I would probably have run out to maybe 1905 or so, which would include at least most of Cross, but not Gardner. There just ain't no hard rules in plausibility. Or, to quote my own teaching mantra for high schoolers, "Statistics can be defined as that branch of mathematics that deals with things that are not as of yet subject to mathematical proof. As soon as you can prove something, you are out of the field of statistics."
   46. bookbook Posted: August 15, 2008 at 11:38 PM (#2904880)
Edgar Martinez could have competed in this field. Too bad.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
JE (Jason Epstein)
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

Syndicate

Demarini, Easton and TPX Baseball Bats

 

 

 

 

Page rendered in 0.8458 seconds
49 querie(s) executed