HONORS VS. ACCOMPLISHMENTS
I. General Introduction
What I am essentially trying to do here is to compare a player’s ACCOMPLISHMENTS to his HONORS. What I have done is to:
(1) Identify a list of 310 players, most of whom have been listed by someone or other as "overrated" or "underrated", but about 75-100 of whom are just there to normalize the data.
(2) Created an "Honors Score" for each player, summarizing all of the significant honors received by that player into one number,
(3) Estimated the EXPECTED honors for each player, based on Career Win Shares, B-R WAR, and the honors given to the most comparable players, and
(4) Identified in this way the "most underrated" and "most overrated" players in the data, based on the comparison of their actual and expected honors.
As a young writer, I spent several years avoiding the terms "overrated" and "underrated". People would ask "What terms do you use?", but it wasn’t the WORDS that I was avoiding; it was the concepts. The problem with "overrated" and "underrated" is "Where are they rated?" One person might say that Alan Trammell is underrated, by which he might mean that Alan Trammell had not been given some honor that he should have been given, and the next expert might say that Alan Trammell was overrated, by which he might mean that Alan Trammell was not as great as Barry Larkin, maybe. Since there was no way of knowing where Alan Trammell was "rated" overall by everybody, the entire discussion was meaningless, thus not an appropriate thing for me to invest my time and effort in. Not only "no way of knowing", but also simply no such thing; the ratings of players change every day, and they change dramatically sometimes years after the player has retired. There is, I would say, no such thing as where players were generally rated.
Over the years I abandoned my initial position of "just not talking about this." I moved on from that for two reasons. First, so much of the sporting public’s discussion is about who is overrated and who is underrated that being cut off from that is awkward and limiting. If you want to contribute to the public debate, you can’t use your own language and your own list of topics. You have to respond to what others are talking about.
And second, over the decades, a much clearer sense of who is rated where has evolved. A sense of who is rated where has evolved in large part because of modern metrics, which tend to place guardrails on the discussions as to who was better than who. It has also evolved because of the popularity and success of web sites like Baseball Reference and Fangraphs, which provide places you can go to easily to check out the modern values assigned to Omar Vizquel, Alan Trammell, Ozzie Smith, Dave Concepcion, Larry Bowa and anybody who even looks like Larry Bowa. It has also evolved because of the overwhelming popularity of fantasy sports and their derivatives, fantasy sports with gambling involved, which create a "need to know" about the relative values of active players. And it has evolved because of the gradual addition to the public records of information such as player salaries, which are a form of rankings.
Because of these changes, the arguments I made against joining in this argument by doing research about it are no longer compelling. So here I am.
Today’s article is how-it-works. I will outline the process of the research, how the conclusions I have come to were reached. In tomorrow’s work, I will name and discuss the players who have been given more awards than would be expected based on their accomplishments. On Wednesday I will take a break for a short article about something else. On Thursday, I will name and discuss those who have been under-valued, underrated or discounted. And at the end of the series, I will explain in more depth the details of the system, to the extent that I can stand to do that.
I have spent about four months working mostly on the articles that I will be publishing this week, so I hope that you will look at them, I hope that you will find some value in them, and I hope that you will get something out of them. If you don’t, at least I did. Doing this research has changed, improved and refined my own thinking on a good number of related and vaguely-related subjects. Thanks for reading.
II. General Explanation of the Method
II A. Honors
The full methodology that underlies my conclusions is complicated, and if I were to pause now to explain all of that to you, I would lose 90% of my audience before I got to the conclusions.
But let us suppose that each of the 310 players within the study has two numbers, an "Honors Score" which summarizes the weight of all the honors received by the player in his career, and an "Expected Honors Score", which is the weight of the honors we would have EXPECTED the player to have, based on the weight of his accomplishments. We can suppose that, because in fact each player does have those two numbers. For Rocky Colavito, the two numbers are 168 and 158, which means that the Honors given to Colavito were essentially what we would expect, given his performance. There is no indication there that he was either overrated or underrated. Larry Walker for his career was 429 over 288, so he was given more honors than one would expect for a player of his accomplishments; we don’t HAVE to use the word overrated, because that brings up side issues. Walker’s opposite number would be Kenny Lofton, who is at 139 to 265. Walker earned 308 Win Shares and 72.7 WAR; Lofton had 287 Win Shares and 68.4 WAR. Walker appears to be slightly better overall than Lofton, just slightly, and you can agree with that or disagree, but we would not expect the "recognition ratio" to be 3 to 1. 429 to 139 is a little more than 3 to 1. Walker is significantly over-valued in the public eye, Colavito placed about where he should be, Lofton significantly under-valued. If you leave the Hall of Fame out of it, Walker still beats Lofton in the Trophy Case, 269 to 139, or almost two to one.
That’s where we’re headed; we want to get numbers like that for every player in our study. Let us take, as a working group, Willie Mays, Gary Carter, Sammy Sosa and Dick Stuart. The awards that were included in this study are:
(1) Hall of Fame selection or performance in Hall of Fame voting,
(2) All Star selections,
(3) Selection to post-season All Star teams (The Sporting News or Associated Press),
(4) Rookie of the Year Awards,
(5) Gold Gloves, and
(6) World Series MVP Award.
And there is a 7th consideration above, which I’ll get to in a moment. The points awarded in three of those six categories are "layered"; that is, being selected to the Hall of Fame in your first year of eligibility is given more points than being elected to the Hall of Fame 40 years later. Getting to 35% in the voting (but not being selected) receives more points than topping out at 6%, being a starter in the All Star game receives more points than being a substitute or being named but not going. Winning a Gold Glove at shortstop receives more points than winning a Gold Glove at first base. Altogether there are 29 levels of award points:
10 based on the Hall of Fame or Hall of Fame voting,
2 based on mid-season All Star teams,
1 based on post-season All Star teams,
10 based on MVP awards or MVP voting,
1 based on Rookie of the Year awards,
3 based on Gold Gloves,
1 based on the World Series MVP, and
1 more that we’ll get to later.
OK, we’ll talk about it now. I also included batting championships. That was the toughest choice in the process of setting this up. In general, what I was trying to do here was to compare a player’s AWARDS to his ACCOMPLISHMENTS. An argument can certainly be made that winning the "batting championship" is something that a player does on his own, an ACCOMPLISHMENT, and thus that it should be treated as an accomplishment, not as an award. An award is something you vote on. No one votes you the batting championship.
But the problem is that, while in a pure sense it is an accomplishment, a batting championship also has many of the characteristics of an award, and in ways that almost no other accomplishment does. My judgment, thinking about it as deeply as I can, is that winning the "batting championship" has more the characteristics of an honor than the characteristics of an accomplishment, and I decided to give some light weight in the "honors" category to the "batting champions."
Here’s how it works. Willie Mays was a first-ballot Hall of Fame selection. That is the highest award you can possibly win in baseball, so we will credit that at 250 points:
|
Player
|
HOF
|
|
Willie Mays
|
250
|
Gary Carter was not a first-ballot Hall of Fame selection, but was selected to the Hall of Fame within 20 years of his last major league game, and for that he is credited with 160 points. Sammy Sosa is not in the Hall of Fame, but did receive 18.5% support in Hall of Fame voting. For that, he gets 25 points. Dick Stuart was never on a Hall of Fame ballot:
|
Player
|
HOF
|
|
Willie Mays
|
250
|
|
Gary Carter
|
160
|
|
Sammy Sosa
|
25
|
|
Dick Stuart
|
0
|
Willie Mays started an All Star game 18 times and played in the game off the bench another 6 times, for which he is credited with 318 additional "honors" points. You get 15 points for being an All Star if you start the game, 8 points if you don’t. Mays gets 270 + 48. Gary Carter appeared in 11 All Star Games, 7 Starts, so that is 137 points. Sammy Sosa appeared in 7 Games, 5 starts, and Dick Stuart appeared in 2 All Star games as a pinch hitter:
|
Player
|
HOF
|
A* G
|
|
Total
|
|
Willie Mays
|
250
|
318
|
|
568
|
|
Gary Carter
|
160
|
137
|
|
297
|
|
Sammy Sosa
|
25
|
91
|
|
116
|
|
Dick Stuart
|
0
|
16
|
|
16
|
Next is the post-season All-Star teams. The Sporting News and the Associated Press for many decades have chosen post-season All-Star teams. I gave eight points for each season that the player made one of those teams; in other words, if you made both of those post-season teams in 1988, that’s eight points but it isn’t 16. Willie Mays made one of those teams or the other 11 times in his career, Gary Carter 7 times, Sammy Sosa 3 times and Dick Stuart once:
|
Player
|
HOF
|
A* G
|
PS AS
|
|
Total
|
|
Willie Mays
|
250
|
318
|
88
|
|
656
|
|
Gary Carter
|
160
|
137
|
56
|
|
353
|
|
Sammy Sosa
|
25
|
91
|
24
|
|
140
|
|
Dick Stuart
|
0
|
16
|
8
|
|
24
|
MVP voting. If a player won the MVP Award, I counted that as 60 points worth of star recognition. A Hall of Fame selection is 100 to 250 points, depending on how long it takes you to be selected, so an MVP Award counts as about one-third as much recognition as the Hall of Fame. I gave
60 points for winning an MVP Award
40 points for finishing second in the voting
25 points for finishing third
15 points for finishing fourth
10 points for finishing fifth
5 points for finishing sixth
4 points for seventh
3 points for eighth
2 points for ninth, and
1 point for finishing tenth in MVP voting.
Willie Mays won two MVP Awards, which counts for 120 points. He also finished second in the voting twice, third twice and fourth twice. He finished fifth only once, but finished sixth three times. That makes a total of 305 points worth of recognition for MVP voting performance (120-80-50-30-10-15). Gary Carter never won an MVP Award, but finished second once, third once, and sixth twice. Sammy Sosa did better than Carter in MVP voting, winning one award and finishing in the top 10 in the voting seven times. Dick Stuart never finished in the top 10 in MVP voting:
|
Player
|
HOF
|
A* G
|
PS AS
|
MVP
|
|
Total
|
|
Willie Mays
|
250
|
318
|
88
|
305
|
|
961
|
|
Gary Carter
|
160
|
137
|
56
|
75
|
|
428
|
|
Sammy Sosa
|
25
|
91
|
24
|
112
|
|
252
|
|
Dick Stuart
|
0
|
16
|
8
|
0
|
|
24
|
Willie Mays was the only one of these four to win the Rookie of the Year Award, which is 20 points, and the only one to win a batting championship, which is 6 points:
|
Player
|
HOF
|
A* G
|
PS AS
|
MVP
|
Rookie
|
Champ
|
|
Total
|
|
Willie Mays
|
250
|
318
|
88
|
305
|
20
|
6
|
|
987
|
|
Gary Carter
|
160
|
137
|
56
|
75
|
0
|
0
|
|
428
|
|
Sammy Sosa
|
25
|
91
|
24
|
112
|
0
|
0
|
|
252
|
|
Dick Stuart
|
0
|
16
|
8
|
0
|
0
|
0
|
|
24
|
For Gold Gloves, I counted 15 points for a Gold Glove at shortstop, catcher or second base, 10 points for a Gold Glove at first base, and 12 points for a Gold Glove at third base or in the outfield. Willie Mays won 12 Gold Gloves, so that’s 144 points. Gary Carter won three Gold Gloves (45), and Sosa and Stuart didn’t win any. That makes the final totals for AHP, All Honors Points, as follows:
|
Player
|
HOF
|
A* G
|
PS AS
|
MVP
|
Rookie
|
Champ
|
Gold G
|
|
Total
|
|
Willie Mays
|
250
|
318
|
88
|
305
|
20
|
6
|
144
|
|
1131
|
|
Gary Carter
|
160
|
137
|
56
|
75
|
0
|
0
|
45
|
|
473
|
|
Sammy Sosa
|
25
|
91
|
24
|
112
|
0
|
0
|
0
|
|
252
|
|
Dick Stuart
|
0
|
16
|
8
|
0
|
0
|
0
|
0
|
|
24
|
Willie Mays’ "Honors Score" is the highest of any player in our study, presumably the highest of any player in history. Gary Carter is actually tied in the Honors Score with his Hall of Fame teammate, Andre Dawson.
So that is how we summarize the values of all of the honors given to a player over the course of his career. Now we will deal with the EXPECTED Honors for a player of his accomplishments.
II B. Expected Honors
To estimate the number of honors we would have expected each player to receive, we ask three questions:
1) How many Honors Points would we have expected this player to receive, given his Career Win Shares?
2) How many Honors Points would we have expected this player to receive, given his career WAR as presented by Baseball Reference? And
3) How many Honors Points were received by the players most comparable to him?
I try to answer those three questions, and then the average of the three is the player’s Expected Honors Points.
Willie Mays had 642 Career Win Shares. We would expect a fella with 642 Win Shares to have 855 Honors Points, based on the distribution pattern of Win Shares and the distribution of Honors Summary Points.
Players with less than 150 Career Win Shares do win Awards. They make All Star teams sometimes, show up in MVP voting, and they win Gold Gloves. They win batting championships and are named Rookie of the Year. However, Award points increase slowly in the range of 1 to 150 Win Shares, because players with less than 150 career Win Shares very rarely win MVP Awards, although it has happened, and almost never go into the Hall of Fame.
In the range of 150 to 250 Career Win Shares, Honors Points per Win Share increase quite significantly. Players in that range of performance make a lot more All Star teams and quite often win MVP Awards. I haven’t done a count of how many MVP Awards have been won by players with 150 to 250 Career Win Shares, but my guess would be that it would be around 50%.
In the range of 250 to 350 Career Win Shares, Honors Points increase at a very rapid rate. Players in that range of ability are all or almost all considered Hall of Fame candidates, and many of them are selected. At 250 Win Shares, you MIGHT get into the Hall of Fame, but probably not. At 350 Win Shares, you’re an odds-on favorite to get in. The density of MVP Awards (per player) increases. They’re regular All Stars, regulars on the Gold Glove teams.
Above 350 Career Win Shares, and especially above 425 Career Win Shares, the rate of increase in expected Honors Points per Win Shares slows steadily, essentially because there are just no more honors left to be won. After you’re in the Hall of Fame, there are no more awards to be earned. The players in the range above 425 get into the Hall of Fame sooner and win even more MVP Awards, and tend to make the All Star team every year even after they are no longer good players. But the baseball community just runs out of awards to give them, unless you start counting Silver Slugger Awards and Player-of-the-Week Awards as significant honors. And doing that would be very problematic, because those awards do not have a long history, so including them would screw up the scales.
Anyway, Willie Mays had 642 Win Shares, which creates an expectation of 855 Award Points. Gary Carter had 337 Win Shares, which creates an expectation of 337 Award Points; just coincidence that the lines cross at that point. In that area of performance, one extra Win Share will result in about 3.4 Extra Honors points, as that’s the high-impact area. Sammy Sosa had 321 Award Points, which would predict about 283 Honors Points, and Dick Stuart had 114 Win Shares, which creates an expectation of 30 Award Points. Honors Points/Award Points. Interchangeable terms.
|
Player
|
Win Sh
|
|
Willie Mays
|
855
|
|
Gary Carter
|
337
|
|
Sammy Sosa
|
283
|
|
Dick Stuart
|
30
|
Willie Mays is credited with 156.1 WAR by Baseball Reference, the second-highest total in the study. 156.1 WAR creates an expectation of 1,512 Honors Points. Gary Carter, with 70.2 WAR, has an expectation of 387 Honors Points based on his WAR. Sammy Sosa, with 58.6 WAR, has an expectation of 245 Honors Points based on that. And Dick Stuart, with 7.8 WAR, has an expectation of 16 Honors Points based on that.
|
Player
|
Win Sh
|
WAR
|
|
Average
|
|
Willie Mays
|
855
|
1512
|
|
1183.5
|
|
Gary Carter
|
337
|
387
|
|
362
|
|
Sammy Sosa
|
283
|
245
|
|
264
|
|
Dick Stuart
|
30
|
16
|
|
23
|
The third element of the Expected Honors for the player is the honors given to the most-comparable players. In discussing the results over the next few days I will mostly use the comparable players element of the process to explain and defend the results.
In some cases the other players identified as comparable to a given player would not ordinarily strike you as similar players. The key is "in what respect are they similar?" The similarity system used here is based on five things: the player’s year of birth, his career games played, his defensive position, his career Win Shares, and his career WAR. A good example would be Dale Murphy and Brett Butler. We would not ordinarily think of Dale Murphy and Brett Butler as being similar players, other than in that they both came out of the Atlanta system at about the same time. Murphy’s presence on the team caused the Braves to trade Butler, but they’re not similar players in obvious and important respects.
But in regard to what we are studying here, they are about as similar as any two players could be. Murphy was born in 1956; Butler in 1957. Murphy played 2,180 career games; Butler, just 2% more. Both men were center fielders. Murphy is credited with 294 Win Shares; Butler, with 295. Murphy is credited with 46.5 WAR; Butler, with 49.7. When you match the weight of their accomplishments, rather than the style or method of their accomplishments, they’re one of the best matches in the study. (They’re actually the third-best comparison in the study. The only two which are better and Chuck Hinton and Lee Maye, and Ron Cey and Toby Harrah.)
But while this system CAN put forward two very different types of players as being similar in the weight of their accomplishments, it usually does not. 70, 80% of the time, you look at who is scored as being most similar to who, and you say "Sure, I can see that." Keith Hernandez and Will Clark. Gary Gaetti and Tim Wallach. Bert Campaneris and Luis Aparicio. Joe Medwick and Indian Bob Johnson. Vic Wertz and Roy Sievers. Ferris Fain and Joe Cunningham. Jerry Lynch and Wes Covington. Jeff Bagwell and Frank Thomas. Lou Whitaker and Alan Trammell. You can draw your own conclusions about that.
In the normal case, I would identify the 25 best comps (most comparable players) for each player, and score the similarity. The system would (usually) ignore those players whose similarity scored at less than 900, and then, in assessing the expected Honors Points for a player, weight each comparable player according to how strong the similarity was.
Willie Mays here is a special, one-of-a-kind problem. The most-comparable players were Hank Aaron, Stan Musial, Mickey Mantle, Barry Bonds and Frank Robinson. The problem is that everyone in the study received fewer Honors Points than Willie Mays, so no matter who you compare him to, the comparison element of the study is going to say that his expected Honors Points are less than those he actually received. In that case I had to tweak the system by adjusting the Honors Points of the comparable players to their Win Shares and WAR. Frank Robinson was a very great player, but he wasn’t Willie Mays.
Rather than getting lost in the weeds here, let’s just do Gary Carter, a simpler case. The comparable players for Gary Carter were Johnny Bench (757 Honors Points), Carlton Fisk (419), Ivan Rodriguez (686), Ted Simmons (267) and, for some reason, Alan Trammell (318). Giving different weights to all of those and also considering those further down the list, I concluded that the players most comparable to Carter came in at an average of 490 Honors Points, 490.174. Sammy Sosa came in at 244 (Vladimir Guerrero, Bobby Abreu, Ichiro Suzuki, John Olerud and Johnny Damon), and Dick Stuart at 26 (Wes Covington, Joe Cunningham, Jerry Lynch, Gus Zernial and Lee Maye. Most of Stuart’s best comps were actually not in the 310-person study group, because there are a very large number of players with accomplishments similar to Stuart’s, and it would take forever to include them all in the study.)
So now we have three estimates of the expected Honors Points for each of these four men:
|
Player
|
Win Sh
|
WAR
|
Comps
|
|
Average
|
|
Willie Mays
|
855
|
1512
|
1099
|
|
1155
|
|
Gary Carter
|
337
|
387
|
490
|
|
405
|
|
Sammy Sosa
|
283
|
245
|
244
|
|
257
|
|
Dick Stuart
|
30
|
16
|
26
|
|
24
|
Except that catchers are a special case. You will notice above that Gary Carter’s "comparable players" estimate (490) is much higher than his Expected Honors by either Win Shares or WAR. This is true for most catchers. Catchers, particularly outstanding catchers, quite often win awards exceeding their WAR-level. They have won many MVP Awards when they were nowhere near leading the league in Win Shares or WAR. In 1955 the league leaders in Win Shares were the Dodger center fielder, Duke Snider (36 Win Shares) and the Yankee center fielder, Mickey Mantle (41 Win Shares.) The Most Valuable Players were the Dodger catcher, Roy Campanella (29 Win Shares) and the Yankee catcher, Yogi Berra (24 Win Shares.) In 1963 the Yankee center fielder, Tom Tresh, tied for the league lead in Win Shares with another outfielder, Carl Yastrzemski. The MVP Award went to the Yankee catcher, Elston Howard. In 1972 the National League leader in Win Shares was Steve Carlton (40). The elected MVP was Johnny Bench.
In WAR, the same thing; neither Campanella nor Yogi Berra led the league in WAR or in WAR by a position player in ANY of the three seasons (six total) in which they won the MVP Award. The standards are different for catchers, I think for good reasons and legitimate reasons, but for good reasons or bad, the standards are different. Catchers have different ratios of Honors to Win Shares or Honors to WAR than other players do. If we’re going to pin down each players Expected Honors with a reasonable amount of credibility, we have to adjust for that. The ratios for catchers are about 10% higher, so I adjusted the expectations for catchers upward by 10%, so that "expectation" would better match "reality". These are the Expected Honors for these four players:
|
Player
|
Win Sh
|
WAR
|
Comps
|
Adjustment
|
|
EXPECTED
|
|
Willie Mays
|
855
|
1512
|
1099
|
|
|
1155
|
|
Gary Carter
|
337
|
387
|
490
|
40
|
|
445
|
|
Sammy Sosa
|
283
|
245
|
244
|
|
|
257
|
|
Dick Stuart
|
30
|
16
|
26
|
|
|
24
|
So now we have both Honors Points and Expected Honors Points for these four players:
|
Player
|
Honors
|
Expected
|
|
Willie Mays
|
1131
|
1155
|
|
Gary Carter
|
473
|
445
|
|
Sammy Sosa
|
252
|
257
|
|
Dick Stuart
|
24
|
24
|
In these four cases, the Honors given to the player essentially match what we would expect them to be. They’re basically the same.
This is more the exception than the rule. I chose these four players to illustrate how it was supposed to work; the player performs at Level X, he gets awards at Level X. But I knew going in that this could not possibly be the general rule. There are 80 cases in the study in which the player essentially matches his expected honors with real honors, and they match in the aggregate; that is, if you take all players who should have 300 Honors Points, they will average 300 Honors Points, and if you take all players who should have about 170 Honors Points, they will have about 170 Honors Points on average.
But in individual cases, the data is all over the map. I told you earlier about Larry Walker and Kenny Lofton, two players of roughly similar accomplishments, but one of whom had received three times as much recognition as the other. That’s not an extreme case at all. That’s a fairly routine case.
I divided the 310 players from the study into 9 classes of over-awarded and under-awarded players:
(1) Those who were horribly over-honored relative to their accomplishments,
(2) Those who were seriously over-honored, but not by as much as the first group,
(3) Those who were significantly over-honored,
(4) Those who were slightly over-honored, maybe a little bit over-honored,
(5) Those whose honors matched expectations, like the four players from our sample study,
(6) Those who were slightly underrated, under-honored,
(7) Those who were significantly shorted when the trophies were handed out,
(8) Those who were seriously overlooked in terms of honors, and
(9) Those who were horribly under-honored.
Walker and Lofton were not in the "horrible examples" ranges. They were in the third groups—the third group from the top (Walker) and the third group from the bottom (Lofton.) They were both in the middle of those groups. There were many players who were more over-honored than Walker, and many who were more overlooked than Lofton. In tomorrow’s article, I’ll start telling you who those players were, and I’ll try as best I can to understand why that happened in each case.
III. More Complete Explanation of the Honors Points
This is a longer version of an earlier section of this article, edited down earlier to enable us to get more quickly to the heart of the issues. This is still not a full explanation of the method; it is just a larger down payment than the earlier quick summary. I will try to explain most of what I have done detail by detail at the end of the week, probably won’t get everything done because it would take thousands of words to explain things that honestly I think most of you would accept on faith if I didn’t explain them at all. I’ll explain them as best I can on Friday, or if I don’t get it done by Friday, whenever I get it done.
But in general, this is how the research was done. Explaining the path of the arrow here, we are working toward a point such that, at the end of this research, we will be able to say that Rocky Colavito in his career received honors and awards totaling up to 168 points. His EXPECTED honors, based on his performance, are 158 points. 168 and 158 are about the same, so Rocky Colavito in his career was neither overrated nor underrated; he was appropriately honored for his accomplishments, neither more than he should have been nor less.
Which was a point of public dispute 60 years ago; Sport magazine published an article entitled "Has Rocky Colavito been overrated?" That article may have been my first introduction to the overrated/underrated terms for any player. I was probably 11, 12 years old at the time, there wasn’t any sports talk radio; that might be the first time I had heard it. The answer is "No"; Rocky Colavito was neither overrated nor underrated; he was rated about right. It took me 62 years to get from there to here. Anyway, Larry Walker for his career was 429 over 288, so he was given more honors than one would expect for a player of his accomplishments; we don’t HAVE to use the word overrated, because that brings up side issues, but perhaps his career doesn’t quite match with all of the recognition that he has received, although certainly he was a very good player. A lot of players are worse than that; Walker I have classified as "significantly" over-honored, which is the third category down from the top, behind "seriously over-honored" and "horribly over-honored." The third category up from the bottom is "seriously under-honored" (or under-recognized, or underrated). In that category, Walker’s opposite number would be Kenny Lofton, who is at 139 to 265. Walker earned 308 Win Shares and 72.7 WAR; Lofton had 287 Win Shares and 68.4 WAR. Walker appears to be slightly better overall than Lofton, just slightly, and you can agree with that or disagree, but we would not expect the "recognition ratio" to be 3 to 1. 429 to 139 is a little more than 3 to 1. Walker is significantly over-valued in the public eye, Colavito placed about where he should be, Lofton significantly under-valued.
That’s where we’re headed; we want to get numbers like that for every player in our study. Let us take, as a working group, Willie Mays, Gary Carter, Sammy Sosa and Dick Stuart.
The first question we have to ask is "Who should be included in the study?" I started by just writing down a list of players that I thought of as underrated or overrated, had about 25 of those. Then I googled lists of underrated players. There are dozens of published lists of underrated players. I googled "players who should be in the Hall of Fame but aren’t", and the opposite, made my own list of guys who are in the Hall of Fame but maybe shouldn’t be. I did a few Twitter polls, asking people to choose this guy or that one, just to see if that would document that some players were more highly valued than others. I added some "neutral" players that no one would much think of as overrated or underrated, and I identified comparable players to those already on the list. In other words, since Dale Murphy was already on the list, I looked up players comparable to Dale Murphy, and put some of them on the list as well. Since what we are essentially asking is whether Dale Murphy was given more awards or fewer than comparable players, we need a list of comparable players for everybody. Eventually, I had 310 players in the study, which is enough that you have kind-of comparable players included for everybody, although you could always, of course, add more.
I did not include Babe Ruth or Ty Cobb or Rogers Hornsby, because the universe of awards that we deal with was radically different in that era. Honus Wagner never won the MVP Award, because there wasn’t any such award. Ted Williams drove in 145 runs as a rookie in 1939, but didn’t win the Rookie of the Year Award because there wasn’t any Rookie of the Year Award until 1947. The major awards that we are dealing with were created between 1931 and 1957, and there is an evolutionary space in there in which some of the awards exist and others don’t. That is a much smaller problem than you may assume, but the first player chronologically who is within the study is Hack Wilson, born 1900. And that’s really too early; I mostly tried to avoid players born before 1920, although I included a few of them. Everything is a balancing act; so much pure information, so much pollution. I could not include Negro League players, because the data to do that just simply does not exist.
The second question(s) we have to ask is/are "Where is this player rated? What is the general perceived value of this player’s career?" Our operating assumption is that the awards and honors that the player has received are indicative of his perceived value. Working on that assumption, we then give the player "Honors Points" based on the major awards that he has received.
The awards that were included in this study are:
(1) Hall of Fame selection or performance in Hall of Fame voting,
(2) All Star selections,
(3) Selection to post-season All Star teams (The Sporting News or Associated Press),
(4) Rookie of the Year Awards,
(5) Gold Gloves, and
(6) World Series MVP Award.
There was also a space in the form for Cy Young Awards, but it would have taken me another month or two months to include pitchers. Sometime while I was working on this Tom Tango observed in a Twitter Post that almost all of the debate about overrated and underrated players was about position players, not pitchers, and I decided that relieved me of the responsibility to figure data for all those damned pitchers; thanks, Tom. And there is a 7th consideration above, which I’ll get to in a moment.
The points awarded in three of those six categories are "layered"; that is, being selected to the Hall of Fame in your first year of eligibility is given more points than being elected to the Hall of Fame 40 years later. Getting to 35% in the voting (but not being selected) receives more points than topping out at 6%, being a starter in the All Star game receives more points than being a substitute or being named but not going, and winning a Gold Glove at shortstop receives more points than winning a Gold Glove at first base. Being second in the MVP voting receives more points than being fourth. Altogether there are 29 levels of award points:
10 based on the Hall of Fame or Hall of Fame voting,
2 based on mid-season All Star teams,
1 based on post-season All Star teams,
10 based on MVP awards or MVP voting,
1 based on Rookie of the Year awards,
3 based on Gold Gloves,
1 based on the World Series MVP, and
1 more that we’ll get to later.
OK, we’ll talk about it now. You talked me into it. I also included batting championships. That was the toughest choice in the process of setting this up—much, much tougher than, for example, establishing the relative values of one award to another. In general, what I was trying to do here was to compare a player’s AWARDS to his ACCOMPLISHMENTS. An argument can certainly be made that winning the "batting championship" is something that a player does on his own, an ACCOMPLISHMENT, and thus that it should be treated as an accomplishment, not as an award. An award is something you vote on. No one votes you the batting championship.
But the problem is that, while in a pure sense it is an accomplishment, a batting championship also has many of the characteristics of an award, and in ways that almost no other accomplishment does. For many years players were put on magazine covers because they won the "batting" championship. That still happens, not as often. Lists of "batting champions" going back 150 years are easy to find, the results widely known. Trophies are given based on it.
My judgment, thinking about it as deeply as I can, that winning the "batting championship" has more the characteristics of an honor than the characteristics of an accomplishment. I discounted the batting championship heavily based on this concern. Finishing second in the MVP voting counts for 30 points; winning the batting championship counts for only 6. Without the "dual nature" concern, I would have weighted the batting championship at something like 35, 40 points, maybe higher. I knocked it down to six because it is part award, part accomplishment, but I just didn’t think it was right to ignore it.
So here’s how it works. Willie Mays was a first-ballot Hall of Fame selection. That is the highest award you can possibly win in baseball, so we will credit that at 250 points:
|
Player
|
HOF
|
|
Willie Mays
|
250
|
Gary Carter was not a first-ballot Hall of Fame selection, but was selected to the Hall of Fame within 20 years of his last major league game, and for that he is credited with 160 points:
|
Player
|
HOF
|
|
Willie Mays
|
250
|
|
Gary Carter
|
160
|
Sammy Sosa is not in the Hall of Fame, but did receive 18.5% support in Hall of Fame voting. For that, he gets 25 points:
|
Player
|
HOF
|
|
Willie Mays
|
250
|
|
Gary Carter
|
160
|
|
Sammy Sosa
|
25
|
Dick Stuart was never on a Hall of Fame ballot:
|
Player
|
HOF
|
|
Willie Mays
|
250
|
|
Gary Carter
|
160
|
|
Sammy Sosa
|
25
|
|
Dick Stuart
|
0
|
Willie Mays started an All Star game 18 times and played in the game off the bench another 6 times, for which he is credited with 318 additional "honors" points:
|
Player
|
HOF
|
A* G
|
|
Total
|
|
Willie Mays
|
250
|
318
|
|
568
|
|
Gary Carter
|
160
|
|
|
160
|
|
Sammy Sosa
|
25
|
|
|
25
|
|
Dick Stuart
|
0
|
|
|
0
|
You get 15 points for being an All Star if you start the game, 8 points if you don’t. Mays gets 270+48. Gary Carter appeared in 11 All Star Games, 7 Starts, so that is 137 points. Sammy Sosa appeared in 7 Games, 5 starts, and Dick Stuart appeared in 2 All Star games as a pinch hitter:
|
Player
|
HOF
|
A* G
|
|
Total
|
|
Willie Mays
|
250
|
318
|
|
568
|
|
Gary Carter
|
160
|
137
|
|
297
|
|
Sammy Sosa
|
25
|
91
|
|
116
|
|
Dick Stuart
|
0
|
16
|
|
16
|
Next is the post-season All-Star teams. The Sporting News and the Associated Press for many decades have chosen post-season All-Star teams, which are somewhat redundant of mid-season All Star teams, but more different than the same. Those teams represent a full season’s work rather than a half-seasons, they only honor 9, 10 or 11 players per league, some years only 9 or 10 players for BOTH leagues, and they almost never pass out semi-gratuitous inclusions to superstars based on their past history. They honor you based on what you did this year, period. The voting panels are different; they’re not selected by fans or players or managers; they’re selected by media. They’re more different than the same.
I gave eight points for each season that the player made one of those teams; in other words, if you made both of those post-season teams in 1988, that’s eight points but it isn’t 16. Willie Mays made one of those teams or the other 11 times in his career, Gary Carter 7 times, Sammy Sosa 3 times and Dick Stuart once:
|
Player
|
HOF
|
A* G
|
PS AS
|
|
Total
|
|
Willie Mays
|
250
|
318
|
88
|
|
656
|
|
Gary Carter
|
160
|
137
|
56
|
|
353
|
|
Sammy Sosa
|
25
|
91
|
24
|
|
140
|
|
Dick Stuart
|
0
|
16
|
8
|
|
24
|
MVP voting. If a player won the MVP Award, I counted that as 60 points worth of star recognition. A Hall of Fame selection is 100 to 250 points, depending on how long it takes you to be selected, so an MVP Award, generally speaking, counts as about one-third as much recognition as the Hall of Fame. But ANY significant attention in MVP voting is evidence of recognition as an outstanding player. I gave
60 points for winning an MVP Award
40 points for finishing second in the voting
25 points for finishing third
15 points for finishing fourth
10 points for finishing fifth
5 points for finishing sixth
4 points for seventh
3 points for eighth
2 points for ninth, and
1 point for finishing tenth in MVP voting.
Willie Mays won two MVP Awards, which counts for 120 points. He also finished second in the voting twice, third twice and fourth twice. He finished fifth only once, but finished sixth three times. That makes a total of 305 points worth of recognition for MVP voting performance (120-80-50-30-10-15). Gary Carter never won an MVP Award, but finished second once, third once, and sixth twice. Sammy Sosa did better than Carter in MVP voting, winning one award and finishing in the top 10 in the voting seven times. Dick Stuart was mentioned in MVP voting three different seasons, but never finished in the top 10 in the voting:
|
Player
|
HOF
|
A* G
|
PS AS
|
MVP
|
|
Total
|
|
Willie Mays
|
250
|
318
|
88
|
305
|
|
961
|
|
Gary Carter
|
160
|
137
|
56
|
75
|
|
428
|
|
Sammy Sosa
|
25
|
91
|
24
|
112
|
|
252
|
|
Dick Stuart
|
0
|
16
|
8
|
0
|
|
24
|
Willie Mays was the only one of these four to win the Rookie of the Year Award, which is 20 points, and the only one to win a batting championship, which is 6 points:
|
Player
|
HOF
|
A* G
|
PS AS
|
MVP
|
Rookie
|
Champ
|
|
Total
|
|
Willie Mays
|
250
|
318
|
88
|
305
|
20
|
6
|
|
987
|
|
Gary Carter
|
160
|
137
|
56
|
75
|
0
|
0
|
|
428
|
|
Sammy Sosa
|
25
|
91
|
24
|
112
|
0
|
0
|
|
252
|
|
Dick Stuart
|
0
|
16
|
8
|
0
|
0
|
0
|
|
24
|
When you have been a writer as long as I have been you can hear what every reader is thinking as he or she reads each line. What some of you are thinking right now, the dumber ones who have to have everything explained to them, is "Sammy Sosa didn’t lead the league in batting average, but he did lead the league twice in home runs and twice in RBI. If you’re going to count one, shouldn’t you count the other? Isn’t it really the same thing?"
"Leading the league" is the same thing. What is different is the recognition level attached to it. That’s COMPLETELY different.
This should not be a difficult concept, that two things may be the same in one sense but completely different in another sense. It’s a common thing. But analytical discussions are heavily populated by literal-minded people who always want to say that if two things are the same in one sense, they must be the same in EVERY sense. I shouldn’t be too hard on you. Sabermetrics IS about simplifying things, making simplified thought-models so that complicated things become clear. That is a large part of what we do. It gets tiresome when you have to remind people that some things have been left out of the simplified model. It seems like you shouldn’t have to explain that to people, but what it really is is that people need to give themselves permission to do that. Schools teach you, society sometimes teaches us, to think in very linear ways. You have to follow the rules. That is how lawyers are trained to think, I believe—that definitions are rigid, not flexible. This is the definition of X; when you think about X, you must use this definition of X. To be a sophisticated thinker, you have to give yourself permission to use a broader, more encompassing definition of X, even if (and especially if) X is being used in a context which seems at first blush to require a limited usage.
The points aren’t here because the player led the league in batting. That’s an accomplishment. What we’re talking about here is recognition. The recognition attached to leading the league in batting, historically, has been vastly greater than the recognition attached to leading the league in any other category. It starts with the language used; he is the batting champion. The simple two-word reference to it, which we still use, implies that it covers everything that a batter does. Throughout most of baseball history, the batting champion has been assumed to be the league’s best hitter. Up until 1950, at least, many more sports fans would have known who won the batting championship two years ago than would have known who won the MVP Award. The recognition attached to it was immense. Until 1975, at least, probably more fans could recite the list of batting champions over the last 20 years than knew who was in the Hall of Fame. It was a huge deal.
I have discounted the batting championship recognition value from what it could be, possibly 35 or 70 points. . .I have discounted that to 6 points because it is an accomplishment as well as a source of disproportionate recognition, and I have discounted it because this isn’t 1975, and the way we think about things has changed. But it would not be appropriate, in an accounting of player’s honors and awards, to completely ignore the honor and celebrity which has historically been given to winning the batting championship.
OK, so that wraps it up here except for Gold Gloves and World Series MVPs, and none of these four guys was a World Series MVP. For Gold Gloves, I counted 15 points for a Gold Glove at shortstop, catcher or second base, 10 points for a Gold Glove at first base, and 12 points for a Gold Glove at third base or in the outfield. Willie Mays won 12 Gold Gloves, so that’s 144 points. Gary Carter won three Gold Gloves (45), and Sosa and Stuart didn’t win any. That makes the final totals for AHP, All Honors Points, as follows:
|
Player
|
HOF
|
A* G
|
PS AS
|
MVP
|
Rookie
|
Champ
|
Gold G
|
|
Total
|
|
Willie Mays
|
250
|
318
|
88
|
305
|
20
|
6
|
144
|
|
1131
|
|
Gary Carter
|
160
|
137
|
56
|
75
|
0
|
0
|
45
|
|
473
|
|
Sammy Sosa
|
25
|
91
|
24
|
112
|
0
|
0
|
0
|
|
252
|
|
Dick Stuart
|
0
|
16
|
8
|
0
|
0
|
0
|
0
|
|
24
|
Willie Mays’ "Honors Score" is the highest of any player in our study, presumably the highest of any player in history. Gary Carter is actually tied in the Honors Score with his Hall of Fame teammate, Andre Dawson, who is also at 473.
IV. The Comparability System
In some cases the other players identified as comparable to a given player would not ordinarily strike you as similar players. The key is "in what respect are they similar?" The similarity system used here is based on five things: the player’s year of birth, his career games played, his defensive position, his career Win Shares, and his career WAR.
I didn’t explain at the time why these five elements were used to identify comparable players. Win Shares and WAR. . . .I assume you understand why these were used to identify players of comparable value. I assume you understand why defensive positions were included, so that shortstops are not considered best comps for left fielders or elephants. I’ll explain in just a moment how that was done, assuming that you get the "why".
The Year of Birth was included so that players would tend to be compared to contemporary players. The mechanics of award selection change over time. The Hall of Fame constantly revises its processes. New Awards are added, the leagues expand, the ways in which players are used on the field and the ways in which players are evaluated by the media evolve over time. ONE of the relevant criteria for establishing expectations for any player is his contemporaries. It is not the ONLY relevant criteria, as people very often say that is, but it is something that should be included. Year of Birth was used to pin the player to his contemporaries.
And Games Played was used because the ratio of Win Shares to Games Played or WAR to Games Played is very much an element in the extent to which a player is recognized as a star. Joe Gordon and Ralph Kiner are in the Hall of Fame with 242 Win Shares, Tony Oliva with 245, George Kell with 229. Vada Pinson is NOT in the Hall of Fame with 321 Win Shares, Dwight Evans not in with 347, and Rusty Staub not in with 348. Part of that is a time-frame difference, but the biggest element is that Pinson, Evans and Staub were long-term accumulators, picking up 8 or 10 or 15 Win Shares a year, whereas Oliva, Kiner and Gordon had shorter, more brilliant careers, and thus received many more honors per unit of Win Shares or WAR. Pinson, Evans and Staub played 2469, 2687 and 2951 career games; Oliva, Kiner and Gordon played 1676, 1566 and 1472. It makes a difference. If Aaron Judge or Shohei Ohtani retires with 250 Career Win Shares, they will go into the Hall of Fame easily; if Elvis Andrus lasts until he gets 300, it won’t be enough.
The system was set up so that each of those five elements has almost precisely the same impact on who is seen as comparable to who. When you compare two players at random, you get (on average) a similarity score of 750, with this system. The sum of the penalties for dissimilarity is 250 points; 1000, minus 250, is 750. The system was created so that the average in each of the five categories is 50 points.
To use Kenny Lofton for illustration, the most-comparable player to Kenny Lofton in this system was Jim Edmonds. Lofton was born in 1967, Edmonds in 1970, which creates a small 10-point separation penalty between them. They are contemporary players. They were both center fielders, so no separation penalty there. Lofton played 2103 major league games, Edmonds 2011; there is no penalty if the difference is less than 100 games, so no separation penalty for that. Lofton had 287 career Win Shares, Edmonds 301, so there is an 8-point separation penalty for that. Lofton had 68.4 WAR, Edmonds 60.4, so there is a 20-point separation penalty for that difference. Adding it up and saving the decimals, the differences between them add up to 38.8 points, so the similarity is scored at 961.2.
To focus on the "average" player in regard to Lofton, that would be Tino Martinez. Martinez and Lofton were both born in the same year, 1967, so there is no separation penalty for that, and they were within 100 of one another in career games played, so there is no separation penalty for that. But Lofton was a center fielder and Martinez a first baseman, which is a 108-point separation penalty. Lofton had 287 Win Shares, Martinez 218, which is a 41-point penalty, and Lofton tops Martinez in WAR 68.4 to 29.0—which I think is a completely inaccurate and unrealistic measurement of the value difference between them—but in any case that creates a penalty of 101 points in the similarity system. That adds up to 250 points of separation (108 + 41 + 101), which pegs the similarity between them at 750, which is an average number for two players picked out of a hat.
The 10 players most-similar to Lofton, within this 310-person group of overrated and underrated candidates, are Jim Edmonds, Scott Rolen, Andruw Jones, Bernie Williams, Larry Walker, Willie Randolph, Brett Butler, Tony Phillips and Dale Murphy. While you might think that Lofton has been underrated, and our system agrees that he has been substantially under-honored, the fact is that four of the ten players with similar credentials have received even fewer honors than has Lofton. Bernie Williams, Willie Randolph, Brett Butler and Tony Phillips were all MORE neglected in award voting than was Lofton.
V. Similar Defensive Positions
We want in general to compare catchers to catchers and second basemen to second basemen, but without being so rigid about that that we fail to recognize that Alan Trammell is comparable to Lou Whitaker. I start that process by assigning "weights" to each position, as follows:
Catcher 100 points
Shortstop 70 points
Second Base 60 points
Center Field 55 points
Third Base 50 points
Right Field 35 points
Left Field 20 points
First Base 8 points
DH   Nothing
This is not about assigning VALUES; this is about measuring DISTANCE. What we are concerned about here is the distance BETWEEN a shortstop and a second baseman, not with how much more valuable in general shortstops are than second basemen. You will notice that catchers are way off on an island by themselves, because catching does not generally interact with the rest of the defensive spectrum, and catchers are not generally comparable to other players. Shortstops are sort of like second basemen (a 10-point separation) or third basemen (20 points.) Center Fielders are like second and third basemen, in that center field is a "high skill demand" position. If you have a .350 hitter who HAS to play somewhere, he plays first base or DH or maybe left field; you don’t force him into the lineup at third base or center field.
Once we have THOSE values—and I have used this system or ones similar to it hundreds of times over the years—but once we have THOSE values, then we still face the question of "how much weight do we give to each point of separation on this chart?"—just as we face the question of how much weight do we give to one year of separation in the year of birth, for example. You remember that I was trying to give exactly equal weight, overall, to each of the five categories of similarity, 50 points each. That number turned out to be 2.3. The "average distance between two players" on the chart above was 22 points, so if we weight each of those at 2.3, then, on average, the "defensive position separation" between two players is 50 points.
We’ll do the over-honored tomorrow. You’re not going to agree with a lot of it. There is going to be player after player at the top of the over-honored list that you’re going to say "JOHN D. ROCKEFELLER! He’s not over-rated! He’s a legitimate Hall of Famer!" If you don’t, the next person will. But perhaps I’ll be able to explain as I go along, and occasionally even convince you.