BILL JAMES ONLINE

Vagabonds and Homebodies

January 13, 2022
                               Vagabonds and Homebodies

 

Introduction

 

        Comparing two players of reasonably equal Hall of Fame credentials, one of whom moves from team to team and the other of whom stays put for most of his career, the player who is easily identified with one team is not only more likely to be elected to the Hall of Fame, but MUCH more likely to be elected to the Hall of Fame.  A player who hopscotches from team to team may be reducing his Hall of Fame election chances by 50% or more by doing so.

            Last week on "Hey, Bill"  a reader (Phil Dellio) suggested that the knockaround, move-around, get-out-of-town nature of Gary Sheffield’s career might be impacting his Hall of Fame voting performance.   Actually, he was making a slightly different point, beyond that one, but anyway. . .that seems credible.  I may have suggested the same thing myself some time in the past, not sure, but it seems reasonably possible, so I responded that I would try to figure out how to study the issue. 

            I have now done that study.  Staying with one team for a longer period of time either directly results in better Hall of Fame chances for the player, or is allied with some other trait, some skill not identified and adjusted for in this method, which results in a quite significant increase in the Cooperstown chances of any player with a less than impeccable Hall of Fame resume.   Or, frankly, even a player WITH an impeccable Hall of Fame resume.  

 

GENERAL OUTLINE OF THE STUDY

            The breakthrough that made it possible to do this study was that I realized that I needed some way to measure the extent to which a player was identified with one team.   Without that, all you really have is one-team players and multi-team players, or something along those lines.   If all you have is one-team players and "not one-team players", you can’t do the study at all, because (1) there aren’t enough one-team players to work with, and (2) a grossly disproportionate percentage of one-team players are superstars. 

            You can divide players into two-team players, three-team players, etc., or you can divide them into groups of players who played 90% with one team, 80% with one team, 70% with one team, etc.   Those options are awkward and don’t really work.  John Smoltz, Gil Hodges, Juan Marichal, Norm Cash, Willie Mays, Hank Aaron, Warren Spahn, Hank Greenberg and Kevin Youkilis are actually NOT one-team players, but we automatically associate them with one team.   There has to be a way to measure the extent to which a player is ALMOST or essentially a one-team player, versus the extent to which he is essentially a hired gun. 

            Having reached the point of realizing that we needed a way to measure that, the solution was obvious.  I created what we could call a supporting methodology, which is "one team identification percentage."  I don’t want to get into a digression about how exactly that was done before I get to the main point, so I’ll explain in detail how that was done in Appendix 1, and present some output data from it in a second Appendix.  When I did this, by the way, I learned that Gary Sheffield just missed having the MOST scattershot career in baseball history for a player with 150 Win Shares—and had EASILY the most broken-up career for a player of his ability.  Sheffield has a one-team identification percentage of 14.4%.  Only one player in history is lower than that.   You won’t guess who it is, but I know that some of you will want to try, so I’ll leave that hanging for now, and then in one of the appendices I’ll tell you who the one player was whose career was even more broken up than Sheffield’s. 

            That problem solved, I estimated each player’s chance of selection to the Hall of Fame based on his career Win Shares.   I have done that before, of course, but having done it before I knew how to do it better than I had before, and with more recent data, so this time I did it somewhat better than I had before.  Again, I don’t want to get into a methodological digression at this point in the article, so I’ll explain how that was done in Appendix 3, and report on conclusions from that process in Appendices 4 and 5. 

            So at this point I had (a) a method to estimate each player’s probability of Hall of Fame selection, based on his career Win Shares, and (b) a method to state the degree to which each player had either a jump-from-team-to-team career, or a stay-with-one-or-two-teams mostly career.  A Gary Sheffield career, or a Cal Ripken career. 

            I then studied this data in two different ways.   To start with an example that runs counter to the general conclusion of the study, let’s take Lou Whitaker. 

            Lou Whitaker is credited with 351 Win Shares in his career.

            A player with 351 career Win Shares has a 70.4% chance of being selected to the Hall of Fame.  I’ll explain in an Appendix how this was calculated, but in simple terms, there are seven players between 347 and 352—Duke Snider (352), Max Carey (351), Lou Whitaker (351), Lou Brock (349), Barry Larkin (347) and Dwight Evans (347).  Five of the seven are in the Hall of Fame, and 5 out of 7 is 71%, so. . .that’s generally how we get there, only working with much larger groups of players.   

            Lou Whitaker played only for one team, the Tigers, so his "one team percentage" is 100%.

            We enter this, then, like this:

Player

1 team Pct

Expectation of Selection

Actual Selection

Input

Whitaker

100

0.704

0

Negative .704

 

            The fact that Whitaker has not been selected to the Hall of Fame is a Negative .704 marker for the group of players who played 100% with one team, if that makes sense. 

            But then we can add two contemporary and similarly qualified players, Ryne Sandberg and Barry Larkin.  They had 346 and 347 Win Shares, and both are 100% one-team players, and both went right in.  (Sandberg played very briefly for the Phillies, but earned no Win Shares there, so in terms of value he is 100% a Cub.)

 

Player

1 team Pct

Expectation of Selection

Actual Selection

Input

Whitaker

100

0.704

0

Negative  .704

Larkin

100

0.677

1

Positive     .324

Sandberg

100

0.671

1

Positive    .329

 

            Since no player’s Hall of Fame chances are either zero or 100%, and every player either is in the Hall of Fame (1) or is not (0), each player is either a positive or a negative contributor to the group that he represents. 

            In the study group. . . .I can see now that I’ll have to add an Appendix to explain who all is in the Study Group.   Appendix 6.   In the study there are 117 players who are 100% one-team players, a group including Terry Moore, Bobby Higginson, Ron Guidry, Bobby Feller and Roberto Clemente.    And Al Bumbry; Al Bumbry played one year for San Diego, but earned no Win Shares. 

            In this group of three players (Whitaker, Larkin and Sandberg) we have an expectation of 2.05 Hall of Fame selections, with an actual selection total of 2.00.  They’re on target.  But if we take all of those 117 players and figure the Hall of Fame likelihood of each one, based on his Career Win Shares, we have an expectation of 41 Hall of Fame selections (40.91).   In fact, 59 of them are in the Hall of Fame, 18 more than expected, or 44% more than expected.   Thus, if we take all of the one-team players in history, they have 44% more Hall of Fame selections within the group than would be expected based on those players’ performance numbers.  Lou Whitaker, Todd Helton and Stan Hack were left out when they could be in, but Bill Mazeroski, Travis Jackson, Bob Lemon, Phil Rizzuto and Earle Combs—all one-team players—are all in when they could have been left out.  On balance, the group over-achieves in Hall of Fame recognition compared to what they achieved on the field.   

            Then, if we take all of the players who are not QUITE one-team players but pretty close to it, players who are clearly and undeniably associated with one team, they ALSO over-achieve.   Gil Hodges, Frank Chance, John Smoltz, Hal Newhouser, Lou Boudreau, Tony Lazzeri, Earl Averill, Lloyd Waner, Nellie Fox, Chief Bender, Joe Sewell, Stan Coveleski, Jack Morris, Freddie Lindstrom and others all have less-than-50% chances of making the Hall of Fame, based on their career Win Shares, but all have one-team identification percentages over 85%, and are all in the Hall of Fame. 

            If we take all players in history with one-team identification percentages over 90%, they have an expectation of 43.5 Hall of Fame Selections, but actually have 63, making them +19.5 or +45%.

            If we take all players in history with one-team identification percentages over 80%, they have an expectation of 119 Hall of Fame selections, but actually have 143, or +24.

            If we take all players in history with one-team identification percentages over 70%, they have an expectation of 155 Hall of Fame selections, but actually have 180, or +25.

 

            Suppose that we start from the other end of the scale, from the Gary Sheffield end of the scale.  In the data given above, we started aggregating the data from the top end of the scale, from the Stan Musial/Kent Hrbek/Dave Concepcion/Kirby Puckett end of the scale.  This time, we’ll start from the Gary Sheffield/Royce Clayton/Doyle Alexander/Claudell Washington/Fred McGriff/Roberto Alomar/Dave Kingman end of the scale.

            Starting on that end, all players in history who have a one-team identification percentage less than 25% have an expectation of 10 Hall of Fame selections (9.76).  They actually have only three—Roberto Alomar, Deacon White and Dan Brouthers.   Players in this group who could be Hall of Famers but aren’t include Fred McGriff, Tony Mullane, Bill Madlock, Don Baylor, Rusty Staub and others.  And Sheffield, of course.

 All players who have a one-team identification percentages less than 33% (one third) have an expectation of 30 Hall of Fame selections (30.27).  They actually have only 20.   Players with a one-team identification percentage less than 40% have a expectation of 59 Hall of Fame selections.  They actually have only 39.  Players with a one-team identification less than 50% have an expectation of 96 Hall of Fame selections.  They actually have only 72.

            I told you that I had two lines of approach to analyze this data.  This was the first one.  To close off this line:  If you sort all players in the study by their one team identification percentage, you can draw a line at any point on that list—at ANY point, without exception—and the players above that line will have more Hall of Fame selections than predicted by their Win Shares, and the players below that line will have fewer. 

            Now, the other study; the other study is better, so I saved it for last.  This one doesn’t rely on the Hall of Fame expectation for any one player.  That doesn’t come into play here. 

            The other study sorted players by their one-team identification percentage.  This one sorts players by their Career Win Shares.  I sorted them into 10 groups.  All players in the study had at least 150 career Win Shares, while the #1 player, Babe Ruth, had 756.   So the ten groups were:

           

Group 10

400 or more Win Shares

51 players

Group 9

350 to 399 Win Shares

49 players

Group 8

320 to 349 Win Shares

47 players

Group 7

300 to 319 Win Shares

46 players

Group 6

275 to 299 Win Shares

69 players

Group 5

250 to 274 Win Shares

79 players

Group 4

225 to 249 Win Shares

123 players

Group 3

200 to 224 Win Shares

144 players

Group 2

175 to 199 Win Shares

187 players

Group 10

150 to 174 Win Shares

244 players

 

            Of the 51 players in Group 10, 47 are in the Hall of Fame, or 92%.   Of the 244 players in group 1, only 2 are in the Hall of Fame, or less than 1%.   This is the full chart of that data:

 

Group 10

400 or more Win Shares

51 players

47

Hall of Famers

92.2%

Group 9

350 to 399 Win Shares

49 players

42

Hall of Famers

85.7%

Group 8

320 to 349 Win Shares

47 players

28

Hall of Famers

59.6%

Group 7

300 to 319 Win Shares

46 players

25

Hall of Famers

54.3%

Group 6

275 to 299 Win Shares

69 players

23

Hall of Famers

33.3%

Group 5

250 to 274 Win Shares

79 players

26

Hall of Famers

32.9%

Group 4

225 to 249 Win Shares

123 players

19

Hall of Famers

15.4%

Group 3

200 to 224 Win Shares

144 players

13

Hall of Famers

9.0%

Group 2

175 to 199 Win Shares

187 players

11

Hall of Famers

5.9%

Group 10

150 to 174 Win Shares

244 players

2

Hall of Famers

0.8%

 

            In general, the data makes perfect sense, right?  The more Win Shares a player earns, the more likely he is to go into the Hall of Fame.  The only anomaly is that players with 275-299 Win Shares have not done meaningfully better in Hall of Fame selection than players with 250 to 274.   I wouldn’t worry about the anomaly; it’s just data. 

            OK, having established those groups, I then divided each group into a top half and a bottom half based on their one-team identification percentages—in other words, 10 High and 10 Low, 9 High and 9 Low, 8 High and 8 Low, 7 High and 7 Low, etc.    Group 6 is players with 275 to 299 career Win Shares—that is, good Hall of Fame candidates, but most of them actually NOT in the Hall of Fame.   There are 79 of those.   Group 6 High is players who had 275 to 299 career Win Shares and had a high one-team identification percentage.   This Group includes Bob Feller, Jim Rice, Bobby Doerr, Kirby Puckett and Bill Terry, and also Amos Otis, but I won’t mention that because I don’t want you to think I am stumping to put Amos Otis in the Hall of Fame, I mean, just because I named my dog after him.   Anyway, Group 6 Low is players who had equivalent careers, 275 to 299 Win Shares, but did so moving more rapidly from town to town.   This group includes Heinie Manush, Jack Quinn, Lave Cross, Kenny Lofton and Omar Vizquel. 

            There are 69 players total in Group 6, thus 34 players in Group 6 High, and 34 players in Group 6 Low.  (In Groups with an odd number of members, which is most of them, I eliminated the player who was in the center of the group in terms of his one-team percentage, so that the other two groups would be equal in size.)

            Of the 34 players in Group 6 High, 15 are now in the Hall of Fame, or 44%. 

            Of the 34 players in Group 6 Low, only 7 are in the Hall of Fame, or 21%.

            Among players who had 275 to 299 career Win Shares, a player is more than twice as likely to be selected to the Hall of Fame if he spent most of his career with one team than if he moved from team to team in his career.  The data for this group is reasonably typical of the study.   In all 10 groups except the bottom one, the percentage of Hall of Fame selections is notably higher among players who spent more time with one team.  Group 10 has only two Hall of Famers (Bruce Sutter and 19th century player Tommy McCarthy), both of whom have relatively low one-team percentages, but still, that’s two players, so it doesn’t really mean much.   Here is the whole chart:

 

            Group 10 High         25 players, all in the Hall of Fame            100%

            Group 10 Low         25 players, 21 in the Hall of Fame               84%

 

            Group 9 High           24 players, 23 in the Hall of Fame            96%

            Group 9 Low            24 players, 18 in the Hall of Fame            75%

 

            The exception in Group 9 High is Lou Whitaker.  He is the only player with 350 or more Win Share who had a high one-team percentage but is NOT in the Hall of Fame. 

 

            Group 8 High           23 players, 16 in the Hall of Fame            70%

            Group 8 Low            23 players, 11 in the Hall of Fame            48%

 

            Group 7 High           23 players, 13 in the Hall of Fame            57%

            Group 7 Low            23 players, 12 in the Hall of Fame            52%

 

            Group 6 High           34 players, 15 in the Hall of Fame            44%

            Group 6 Low            34 players, 7 in the Hall of Fame            21%

 

            Group 5 High           39 players, 17 in the Hall of Fame            44%

            Group 5 Low            39 players, 9 in the Hall of Fame            23%

 

            Group 4 High           61 players, 16 in the Hall of Fame            26%

            Group 4 Low            61 players, 3 in the Hall of Fame              5%

 

            Group 3 High           72 players, 9 in the Hall of Fame            12.5%

            Group 3 Low            72 players, 4 in the Hall of Fame              6%

 

            Group 2 High           93 players, 8 in the Hall of Fame              9%

            Group 2 Low            93 players, 3 in the Hall of Fame               3%

 

            Group 1 High           122 players, none in the Hall of Fame     0%

            Group 1 Low            122 players, 2 in the Hall of Fame               2%

 

            And actually, in the bottom group, the two players who are in the Hall of Fame, Sutter and McCarthy, and actually very near the middle of the group in terms of one-team percentage rather than actually being "low".   They both have one-team percentages just under 50%. 

            Anyway. . . make what you will of it, but I was fairly astonished by the extent to which staying with one team for an extended period improves a player’s Hall of Fame selection chances.  I expected that the numbers would be higher for the stay-at-home group, but I had NO idea that the impact of that was as large as it appears to be.   I thought it would be hard to find in the data.   Many, many different things play into the Hall of Fame selection process.  The effect of this in most cases is to scramble the data so that patterns are hard to identify.  Stated another way, one can usually identify the MAJOR influences on outcomes by a study like this, but you can almost never identify the secondary inputs.  Something like a player being a coach post-playing career, or a player playing a big city as opposed to a small city, or a player winning Gold Gloves.. . that stuff is hard to identify, even though it may be significant, because the churning of the data caused by competing major influences will hide the minor influences. 

            But this, based on this study, has to be considered a major factor in the voting.  If a player has less than 300 career Win Shares, his chance of being selected to the Hall of Fame is consistently more than twice as great if he played most of his career for one team than if he split his career among three or four different teams.  It is more than twice as great in almost every sub-group once you get under 300 Win Shares.  

            Over 300 Win Shares, there is still a very significant advantage for the player who plays most of his career with one team, although it doesn’t measure as 2-to-1 because, well, it can’t.  If 80% of a group of players is in the Hall of Fame, the number has to be at least 60% in each part of the study, so it is mathematically impossible to have a 2-to-1 ratio. 

            The exact point at which a player becomes more likely than not to go into the Hall of Fame is 306 Win Shares.  Harold Baines, with 307 Win Shares, is in.   Al Oliver, about the same kind of a player but with 305, is not in.   Of course that’s not really a clear pattern, it’s just a way of thinking about it. 

 

            We should be careful in interpreting our result.  I have suggested here that playing most of his career with one team causes a player to do better in the Hall of Fame selection process.  That’s not necessarily true.  It could be that the players who stay longer in one city are more likely to go into the Hall of Fame because they are actually better players in some manner that is too subtle for the study; they have the same Win Shares, but one of them is actually better than the other.   Has more WAR or something, more World Series titles or better team-building skills.  Maybe these guys who move around just wear out their welcome after a year or two. 

            I would be skeptical of that because the differences are so large.  If we had a bunch of 39% to 35% advantages, 23% to 20%,  7% to 5%. . . if that was the data, I’d say "sure, maybe Win Shares is missing something in its evaluation of the careers."  When the ratio is more than 2 to 1—consistently—it would be harder to explain how that could happen for some "outside the box" reason.  But we don’t necessarily know what is cause and what is effect. 

 

 

Appendix 1

How the One-Team Identification Percentage is Calculated

 

I used the technique, which I have used many times before, of measuring by the sum of the squares divided by the square of the sum.   

            Suppose that a player has 10 Win Shares (or 10 WAR, or 10 games played, or 10 RBI, or 10 homers; it doesn’t much matter.)  Suppose he has 10, and all 10 are with one team.   Then his "one team percentage" is 100%. 

            (10 ^ 2)  / (10 ^ 2) = 100 / 100 = 1.000

            Suppose that he plays for two teams and has five Win Shares for each team; then his "one team percentage" is 50%:

            [(5 ^ 2) + (5 ^ 2)] / (10 ^ 2)  =  (25 + 25) / 100 = 50/100 = .500

 

            Suppose that he plays for three teams, and has four Win Shares for each team; then is "one team percentage" is 33.33%:

            [(4 ^ 2) + (4 ^ 2) + (4 ^ 2)] / (12 ^ 2) =

            (16 + 16 + 16) / 144 = 48/144 = .33333333

 

            It’s an easy, reliable method of measuring the extent to which a group of things is divided into sub-groups, vs. the extent to which they are all in the same group.  So then, John Smoltz played for three teams, but is 99.3% identified with one team, Gil Hodges played for two teams but in 98.5% a Dodger, Juan Marichal played for three teams but is 97.7% a Giant, etc. 

            Oh, there is one little problem with the study that I should acknowledge here.  In my Win Shares file, when a player plays for two teams in one season, that is listed just as "2 teams". . . .or 3 teams, or whatever.   Because of that, those seasons are counted as being not for either team, but for a "different" team, which is identified just as "2 teams".   That’s why Gary Sheffield figures at 14.4%; he was traded in mid-season a couple of times.   If I had the data for each split season, he wouldn’t be at 14.4%; he’d be at 14.9% or something, some number a little bit different.  It’s a little glitch in the data.  There’s always something.

            The one player in history who has a lower one-team identification than Gary Sheffield was 1890s catcher Duke Farrell.  Look him up.

 

Appendix 2

Some Results from the Method Described in Appendix 1

            As I think I mentioned, there are 117 players in the study who earned 150 or more Win Shares in their careers, all for one team.  This includes Brad Radke, Vern Law, Mike Scioscia, Terry Puhl, Bob Allison, Ryan Howard and 59 Hall of Famers. 

            60-some other players are very close to 100%, 90% or higher; these include Ty Cobb, Duke Snider and Eddie Mathews, but also Paul Blair, Johnny Logan, Hank Bauer, Bob Forsch, Eric Karros, Willie Mays, Hank Aaron, Norm Cash and Bob Friend. 

            The average one-team identification percentage is 58%.   It does not vary greatly between great players and lesser players.   Great Players sometimes stay with one team for 12 years, then go to another for 6; Lesser players make it 8 and 4, but the average percentage stays around 58% regardless of what caliber of player you are discussing.

            Carlton Fisk is almost exactly 50%, 182 Win Shares with the Red Sox, 186 with the White.   One team percentage of 50.006.   Another guy like that was Miller Huggins, 110 and 112, and Wally Moon (87 Win Shares with the Cardinals, 88 with the Dodgers, never played for anybody else.)   Jimmy Foxx, Vida Blue and Joe Medwick are around 50%, but with a different pattern, playing mostly for one team but played some with a third or fourth or fifth team. 

            Babe Ruth had 756 Career Win Shares, 574 with the Yankees, 180 with the Red Sox, two with the Braves.  That figures as a one-team percentage of 63.3%.  He had 76% of his value with the Yankees, and you might assume that, if he had 76% of his Win Shares with one team, he would have a one-team percentage of 76%, but that’s not QUITE the way the method works.  It works that way if you split it evenly between two teams or between three teams, but it doesn’t work exactly that way if the splits are not even. 

            Suppose that we used the rule that a player’s one-team identification percentage is simply the percentage of his games or the percentage of his value which is accumulated with one team—76%, for Babe Ruth.  But suppose that you had two players who each had 200 Win Shares in their career, but one player played for only two teams, like Carlton Fisk or Wally Moon, but had 100 Win Shares with each team.  The other player played for six teams, but had 100 Win Shares with one team, 30 with a second team, 25 with a third, 20 with a fourth, 15 with a fifth, and 10 with a sixth team. 

            Player A  100 + 100                                    =200

            Player B  100 + 30 + 25 + 20 + 15 + 10  =  200

 

            Obviously these two players are not the same in the extent to which their careers are broken up into separate parts—but if you used the "Babe Ruth is 76%" alternate, they would measure as being exactly the same, both 50% one-team players.  That’s why we don’t do it that way.

            The lowest one-team identification percentage for a Hall of Fame player is 19%, by Dan Brouthers.   Robby Alomar is second on that list, with 23%.  Other Hall of Famers with low one-team identification percentages include Deacon White (23%), Dave Bancroft (25%), Rick Ferrell (26%), Dennis Eckersley (26%), Goose Gossage (26%), Heine Manush (26%), Gaylord Perry (27%) and Hoyt Wilhelm (27%). 

            Bert Blyleven, mentioned earlier in this discussion, is at 29%, 13th lowest for a Hall of Famer.  Cy Young is at 30%.  Nolan Ryan is at 33%, Randy Johnson at 37%, Reggie Jackson at 38%. 

 

 

 

 

Appendix 3

How a Player’s Hall of Fame Expectation Was Figured

 

            Lou Whitaker has 351 Career Win Shares.  One can easily observe that most players with 340 or more Win Shares go into the Hall of Fame, but how exactly did I determine that the figure of Whitaker should be 70.4%?

            There are 49 players in history with 350 to 399 Win Shares, of whom 42 are in the Hall of Fame, or 85.7%.   We COULD say that, since Whitaker is in this group, his expectation is 86%.

            But that’s not exactly right—actually it is not all that close to being right—because Whitaker is almost at the very bottom of his group.  The percentage for players between 320 and 349 Win Shares is only 59.6%.   The center of the 350-399 group is 374.5, while the center of the 320-349 group is 334.5, so Whitaker is actually quite a bit closer to the center of the 320-349 group than he is the center of the 350-399. 

            I assumed that the percentage of Hall of Fame players for any group of players represents the appropriate figure for the CENTER of that group, and then I interpolated between the centers of the groups.  The center for Group 3 is 334.5, and the percentage for Group 3 is 59.574, so 59.574 is the assumed percentage for a player with 334.5 Win Shares, although there is no such thing as a half of a Win Share. 

            The Center for Group 2 is 374.5 and the percentage is 85.714.   The distance between the centers of the two groups is 40 Win Shares.  Whitaker covers 16.5 of those 40.   If you divide 16.5 by 40 and multiply that by the difference between 59.574 and 85.714,  then add the result to 59.574, you get 70.357, I hope.  So I credited Whitaker with a 70.357% percent chance of Hall of Fame selection. 

            Since Whitaker hasn’t been selected, he scores at negative .70357.   But if you do this for all 1,039 players within the study, the aggregate plus or negative is zero.  Actually, it isn’t EXACTLY zero; it is some very small number, less than .50, because that’s just the way these things are; there’s always some little thing that keeps everything from balancing out exactly.

            Probably this.   The chart numbers jump suddenly very when a player crosses 300 Win Shares.   Because of that large jump, if I used the procedure outlined above (in re Whitaker), there would be a "fault line" at 309.5, 309.5 being the center of Group 4.   Using the procedure outlined above the numbers would jump suddenly forward between 309 and 310.  It is unrealistic to believe that a player’s chance of making the Hall of Fame increases dramatically when he earns his 310th Win Share, so I didn’t want the chart to say that it does. 

            So in interpolating the data in that range of the chart, from 320 down to 275, I used a different method, to smooth out the curve in that area.   Occasionally theoretical data is better than real life data. 

 

Appendix 4

The Hall of Fame Expectations for Each Level of Win Shares

 

            One might assume that a player who has a very large number of Career Win Shares would be an automatic Hall of Fame selection, thus have a Hall of Fame Expectation of 100%.   You can see, on reflection, that this is not true.  Barry Bonds and Roger Clemens, perhaps the greatest pitcher and the greatest hitter of my lifetime, are not in the Hall of Fame, and Pete Rose, the All-Time leader in hits, is not in the Hall of Fame.   There are other things that could keep a very great player from being recognized.  If there was an OJ Simpson type candidate in baseball, he would not be elected.  In the current environment, a player who expressed clear racist hostility would not be elected.  It’s not 100% automatic, no matter what a player has done on the field. 

            How close does a player get to 100%, by his performance on the field?  The data says 92.2%, so that’s the number that I went with, not saying it will hold in the future.  My rule was than any player with 425 Career Win Shares or more has a 92.2% chance of being selected to the Hall of Fame. 

            By this research, there is a Hall of Fame probability associated with each number of career Win Shares.  These are fun or useful numbers to know.   They are relevant—not determinative, but helpful—in any Hall of Fame discussion.  Kyle Seager just retired with 205 career Win Shares.  What are the Hall of Fame chances of a player with 205?   8.1%.   Kyle Seager is not a Hall of Famer, but there are Hall of Famers who had less.  Albert Pujols is at 494, so he is quasi-automatic.  He’s automatic unless he does something bad.  Joey Votto is at 333.  What is the probability for a player with 333?   59.3%.   More likely than not, but not a lock.  Johnny Mize got to 338, despite missing three seasons due to World War II, but then had to wait almost 30 years to be welcomed into the Hall of Fame.   He was bitter about it. 

            This chart gives the Hall of Fame probability for a player with any number of career Win Shares from 150 to 425.

 

Win Shares

Probability of Selection to Hall of Fame

+0

+1

+2

 

+3

+4

+5

+6

 

+7

+8

+9

420

91.6

91.7

91.8

 

92.0

92.1

92.2

 

 

 

 

 

410

90.3

90.4

90.5

 

90.7

90.8

90.9

91.1

 

91.2

91.3

91.5

400

89.0

89.1

89.3

 

89.4

89.5

89.6

89.8

 

89.9

90.0

90.2

390

87.7

87.8

88.0

 

88.1

88.2

88.4

88.5

 

88.6

88.7

88.9

380

86.4

86.6

86.7

 

86.8

86.9

87.1

87.2

 

87.3

87.5

87.6

370

82.8

83.4

84.1

 

84.7

85.4

85.7

85.9

 

86.0

86.2

86.1

360

76.2

76.9

77.5

 

78.2

78.8

79.5

80.2

 

80.8

81.5

82.1

350

69.7

70.4

71.0

 

71.7

72.3

73.0

73.6

 

74.3

74.9

75.5

340

63.2

63.8

64.5

 

65.1

65.8

66.4

67.1

 

67.7

68.4

69.1

330

58.6

58.8

59.1

 

59.3

59.5

60.2

60.8

 

61.2

61.9

62.5

320

56.5

56.8

57.0

 

57.2

57.4

57.6

57.8

 

58.0

58.2

58.4

310

53.6

54.7

54.9

 

55.1

55.3

55.5

55.7

 

55.9

56.1

56.3

300

45.4

45.9

46.5

 

47.2

48.2

49.2

50.2

 

51.1

52.0

52.8

290

36.4

37.4

38.4

 

39.5

40.5

41.6

42.7

 

43.7

44.2

44.8

280

33.2

33.2

33.2

 

33.3

33.3

33.3

33.3

 

33.3

34.3

35.3

270

33.0

33.0

33.1

 

33.1

33.1

33.1

33.1

 

33.2

33.2

33.2

260

31.5

32.2

32.9

 

32.9

32.9

33.0

33.0

 

33.0

33.0

33.0

250

24.5

25.2

25.9

 

26.6

27.3

28.0

28.7

 

29.4

30.1

30.8

240

17.5

18.2

18.9

 

19.6

20.3

21.0

21.7

 

22.4

23.1

23.8

230

13.7

13.9

14.2

 

14.4

14.7

14.9

15.2

 

15.4

16.1

16.8

220

11.0

11.3

11.6

 

11.8

12.1

12.4

12.6

 

12.9

13.2

13.5

210

8.8

9.1

9.4

 

9.6

9.9

10.2

10.4

 

10.7

11.0

11.3

200

7.5

7.6

7.8

 

7.9

8.1

8.2

8.3

 

8.4

8.6

8.7

190

6.4

6.5

6.6

 

6.8

6.9

7.0

7.1

 

7.2

7.3

7.4

180

4.5

4.7

4.9

 

5.1

5.3

5.5

5.7

 

5.9

6.0

6.1

170

2.4

2.6

2.8

 

3.0

3.2

3.5

3.7

 

3.9

4.1

4.3

160

0.8

0.8

0.9

 

1.0

1.2

1.4

1.6

 

1.8

2.0

2.2

150

0.5

0.5

0.5

 

0.5

0.6

0.6

0.6

 

0.7

0.7

0.7

 

 

Appendix 5

Hall of Fame Candidates

 

            This study rests on three key numbers for each player:

1)     His career Win Shares,

2)     The probability that a player with that number of Win Shares is in the Hall of Fame, and

3)     His one-team Identification Percentage.

 

This chart gives those three pieces of data for all players within the study who had 200 or more Win Shares, but are not in the Hall of Fame:

 

First

Last

Win Shares

HOF Probability

One Team %

 

First

Last

Win Shares

HOF Probability

One Team %

Barry

Bonds

704

92.2%

59.2%

 

Frank

Schulte

239

16.8%

77.9%

Roger

Clemens

437

92.2%

39.0%

 

Roger

Peckinpaugh

239

16.8%

48.7%

Gary

Sheffield

430

92.2%

14.4%

 

Will

White

239

16.8%

53.8%

Manny

Ramirez

408

90.0%

37.7%

 

Cupid

Childs

238

16.1%

34.6%

Tony

Mullane

399

88.9%

21.2%

 

Harry

Davis

238

16.1%

75.1%

Bill

Dahlen

394

88.2%

33.6%

 

Bobo

Newsom

237

15.4%

26.9%

Rafael

Palmeiro

393

88.1%

48.3%

 

George

Hendrick

237

15.4%

30.2%

Darrell

Evans

363

78.2%

33.7%

 

Wally

Moses

237

15.4%

43.5%

Rusty

Staub

358

74.9%

23.8%

 

Phil

Cavarretta

237

15.4%

94.3%

Bobby

Abreu

356

73.6%

41.6%

 

Willie

Wilson

237

15.4%

85.8%

Sherry

Magee

354

72.3%

61.9%

 

Edgar

Renteria

236

15.2%

27.7%

Lou

Whitaker

351

70.4%

100.0%

 

Joe

Adcock

236

15.2%

53.1%

Dwight

Evans

347

67.7%

94.4%

 

Shawn

Green

236

15.2%

37.1%

George

Van Haltren

344

65.8%

35.4%

 

Bob

Watson

236

15.2%

66.9%

Mark

McGwire

343

65.1%

49.4%

 

Doc

White

235

14.9%

70.1%

Fred

McGriff

342

64.5%

18.7%

 

Cy

Williams

235

14.9%

62.4%

Dick

Allen

342

64.5%

44.4%

 

Mel

Harder

234

14.7%

100.0%

Jimmy

Sheckard

339

62.5%

37.7%

 

Prince

Fielder

234

14.7%

52.0%

Jeff

Kent

339

62.5%

30.8%

 

Del

Ennis

233

14.4%

85.7%

Bob

Caruthers

337

61.2%

36.2%

 

Willie

Horton

233

14.4%

74.0%

Jim

McCormick

334

59.5%

41.4%

 

David

Justice

233

14.4%

42.2%

Will

Clark

331

58.8%

51.4%

 

Ben

Chapman

233

14.4%

35.1%

Bobby

Grich

329

58.4%

50.6%

 

Derrek

Lee

233

14.4%

41.9%

Tommy

Leach

328

58.2%

57.7%

 

Charlie

Hough

233

14.4%

45.0%

Dave

Parker

327

58.0%

44.8%

 

Juan

Gonzalez

233

14.4%

73.4%

Reggie

Smith

325

57.6%

35.9%

 

Dutch

Leonard

233

14.4%

37.0%

Jason

Giambi

325

57.6%

43.9%

 

Jesse

Tannehill

233

14.4%

47.9%

Willie

Davis

322

57.0%

71.2%

 

Dennis

Martinez

233

14.4%

36.1%

Vada

Pinson

321

56.8%

59.5%

 

Wes

Ferrell

233

14.4%

44.1%

Sammy

Sosa

321

56.8%

80.4%

 

Tim

Salmon

232

14.2%

100.0%

Graig

Nettles

321

56.8%

46.4%

 

Donie

Bush

232

14.2%

90.9%

Todd

Helton

318

56.1%

100.0%

 

Babe

Herman

232

14.2%

40.5%

Luis

Gonzalez

318

56.1%

38.6%

 

Andy

Van Slyke

231

13.9%

59.0%

Stan

Hack

316

55.7%

100.0%

 

Gene

Tenace

231

13.9%

44.6%

Jimmy

Ryan

316

55.7%

71.6%

 

Ray

Durham

231

13.9%

40.9%

Jack

Clark

316

55.7%

34.1%

 

Chuck

Knoblauch

231

13.9%

57.2%

Norm

Cash

315

55.5%

97.5%

 

Roy

Sievers

231

13.9%

36.7%

Lance

Berkman

313

55.1%

70.3%

 

Paul

Derringer

231

13.9%

54.1%

Jose

Cruz

313

55.1%

77.7%

 

B. J.

Surhoff

231

13.9%

39.7%

Bernie

Williams

312

54.9%

100.0%

 

Jack

Fournier

231

13.9%

31.3%

Willie

Randolph

312

54.9%

65.8%

 

George

Uhle

231

13.9%

56.4%

Keith

Hernandez

311

54.7%

43.8%

 

Kent

Hrbek

230

13.6%

100.0%

Johnny

Damon

307

50.2%

24.5%

 

Garret

Anderson

230

13.6%

94.1%

Al

Oliver

305

49.2%

42.9%

 

Ryan

Klesko

230

13.6%

48.3%

Jim

Wynn

305

49.2%

61.2%

 

Hardy

Richardson

230

13.6%

32.2%

Scott

Rolen

304

48.2%

29.5%

 

Ginger

Beaumont

229

13.4%

59.5%

Carlos

Delgado

303

47.2%

53.7%

 

Gene

Woodling

228

13.1%

29.0%

Ken

Singleton

302

46.3%

59.5%

 

Dizzy

Trout

228

13.1%

93.2%

Bobby

Bonds

302

46.3%

41.8%

 

Dode

Paskert

227

12.9%

42.2%

John

Olerud

302

46.3%

30.3%

 

Stuffy

McInnis

227

12.9%

41.8%

Jim

Edmonds

301

45.4%

49.0%

 

Ray

Lankford

227

12.9%

83.6%

Buddy

Bell

301

45.4%

34.8%

 

Al

Dark

226

12.6%

38.5%

Brian

Downing

298

44.4%

60.2%

 

Bill

Buckner

226

12.6%

34.3%

Frank

Howard

297

43.6%

52.8%

 

Tommy

Bridges

225

12.4%

100.0%

Steve

Finley

297

43.6%

25.1%

 

Urban

Shocker

225

12.4%

57.5%

Cesar

Cedeno

296

42.7%

74.3%

 

Lee

May

225

12.4%

33.8%

Mickey

Vernon

296

42.7%

60.8%

 

Larry

Jackson

225

12.4%

37.0%

Brett

Butler

295

41.8%

26.6%

 

Pete

Browning

225

12.4%

39.7%

Dale

Murphy

294

41.0%

82.3%

 

Jamie

Moyer

225

12.4%

33.9%

Mark

Grace

294

41.0%

84.4%

 

Dick

Groat

225

12.4%

48.6%

Kid

Gleason

294

41.0%

30.4%

 

Mickey

Lolich

224

12.1%

89.7%

Dave

Foutz

292

39.4%

40.6%

 

Rafael

Furcal

224

12.1%

42.9%

Fielder

Jones

290

37.1%

58.0%

 

Eric

Davis

224

12.1%

57.4%

George

Burns

290

37.1%

70.4%

 

Dolph

Camilli

224

12.1%

51.3%

Larry

Doyle

289

35.6%

76.3%

 

Willie

McGee

224

12.1%

51.1%

Tommy

John

289

35.6%

27.0%

 

Bill

Nicholson

223

11.9%

82.9%

Miguel

Tejada

288

33.5%

33.6%

 

Bob

Shawkey

223

11.9%

78.2%

Jack

Powell

287

33.3%

33.2%

 

Roger

Maris

223

11.9%

47.1%

Bob

Elliott

287

33.3%

44.4%

 

Topsy

Hartsel

223

11.9%

73.3%

Ed

Konetchy

287

33.3%

32.0%

 

Harvey

Kuenn

223

11.9%

50.4%

Bob

Johnson

287

33.3%

60.8%

 

Todd

Zeile

223

11.9%

23.5%

Brian

Giles

287

33.3%

33.7%

 

Charlie

Root

223

11.9%

100.0%

Kenny

Lofton

287

33.3%

44.0%

 

Billy

Nash

222

11.6%

67.2%

Toby

Harrah

287

33.3%

50.8%

 

Miller

Huggins

222

11.6%

50.0%

Jack

Quinn

287

33.3%

20.4%

 

Darrell

Porter

222

11.6%

30.8%

Amos

Otis

286

33.3%

98.6%

 

Freddie

Fitzsimmons

222

11.6%

62.3%

Jack

Stivetts

285

33.3%

52.5%

 

Ruben

Sierra

222

11.6%

41.3%

Chili

Davis

285

33.3%

31.7%

 

Bill

Hutchinson

221

11.3%

99.1%

Mike

Smith

285

33.3%

37.0%

 

Arlie

Latham

221

11.3%

48.2%

Sal

Bando

283

33.3%

70.0%

 

Chris

Chambliss

221

11.3%

33.7%

Charlie

Buffinton

283

33.3%

34.7%

 

Ed

McKean

221

11.3%

39.5%

Boog

Powell

282

33.2%

81.5%

 

Curt

Flood

221

11.3%

100.0%

Omar

Vizquel

282

33.2%

41.3%

 

Dom

DiMaggio

220

11.1%

100.0%

Ron

Cey

280

33.2%

67.2%

 

Mark

Buehrle

220

11.1%

66.4%

Fred

Lynn

280

33.2%

37.8%

 

Lon

Warneke

220

11.1%

48.0%

Julio

Franco

280

33.2%

32.0%

 

Andy

Pafko

220

11.1%

42.3%

Tony

Fernandez

280

33.2%

46.7%

 

Nomar

Garciaparra

219

10.8%

62.7%

Bert

Campaneris

280

33.2%

78.3%

 

Tim

Hudson

219

10.8%

46.1%

Steve

Garvey

279

33.2%

67.9%

 

Doc

Cramer

219

10.8%

30.4%

Ken

Boyer

279

33.2%

77.9%

 

Vic

Wertz

219

10.8%

29.6%

Lave

Cross

278

33.2%

22.8%

 

Richie

Hebner

219

10.8%

47.1%

Dixie

Walker

278

33.2%

53.4%

 

Charlie

Keller

218

10.6%

95.5%

Torii

Hunter

277

33.2%

42.1%

 

Art

Fletcher

218

10.6%

79.2%

Moises

Alou

277

33.2%

22.0%

 

Tino

Martinez

218

10.6%

41.2%

Andruw

Jones

276

33.1%

80.4%

 

Kirk

Gibson

218

10.6%

56.8%

Jim

Whitney

275

33.1%

56.9%

 

Tom

Brown

218

10.6%

14.4%

Robin

Ventura

274

33.1%

46.2%

 

Larry

French

218

10.6%

42.3%

Rocky

Colavito

273

33.1%

39.7%

 

Carl

Furillo

217

10.3%

100.0%

Adonis

Terry

273

33.1%

26.1%

 

Placido

Polanco

217

10.3%

27.1%

Heinie

Groh

272

33.1%

61.2%

 

Jeff

Heath

217

10.3%

51.6%

Aramis

Ramirez

272

33.1%

39.5%

 

Harlond

Clift

216

10.1%

77.3%

Jose

Canseco

272

33.1%

37.8%

 

Pete

Runnels

216

10.1%

46.2%

Cy

Seymour

272

33.1%

35.6%

 

George

Scott

216

10.1%

48.2%

Joe

Judge

270

33.0%

97.1%

 

Hooks

Dauss

215

9.8%

100.0%

Dave

Concepcion

269

33.0%

100.0%

 

Danny

Murphy

215

9.8%

92.8%

George

Foster

269

33.0%

55.0%

 

Tom

Daly

215

9.8%

62.3%

Ron

Fairly

269

33.0%

38.5%

 

Kip

Selbach

215

9.8%

20.9%

Tony

Phillips

268

33.0%

30.8%

 

Brady

Anderson

214

9.5%

97.2%

Bill

Freehan

267

33.0%

100.0%

 

Heinie

Zimmerman

214

9.5%

51.3%

Eddie

Yost

267

33.0%

68.2%

 

Tommy

Corcoran

214

9.5%

37.3%

Bobby

Bonilla

267

33.0%

30.7%

 

Rudy

York

214

9.5%

72.1%

Clyde

Milan

266

33.0%

100.0%

 

Chuck

Finley

213

9.3%

76.4%

Wilbur

Cooper

266

33.0%

88.6%

 

Frank

Dwyer

213

9.3%

50.2%

Herman

Long

265

33.0%

81.4%

 

Tip

O'Neill

213

9.3%

70.2%

Vern

Stephens

265

33.0%

49.1%

 

Raul

Ibanez

213

9.3%

38.8%

Chet

Lemon

265

33.0%

50.6%

 

Sam

Leever

212

9.0%

100.0%

Harry

Stovey

265

33.0%

43.7%

 

Theodore

Breitenstein

212

9.0%

49.8%

Bobby

Veach

265

33.0%

88.5%

 

Mike

Hargrove

212

9.0%

42.4%

Don

Mattingly

263

32.9%

100.0%

 

Jose

Cardenal

212

9.0%

29.3%

Augie

Galan

263

32.9%

40.2%

 

Frank

White

211

8.9%

100.0%

Silver

King

263

32.9%

38.0%

 

Edgardo

Alfonzo

211

8.9%

68.7%

Don

Baylor

262

32.9%

25.0%

 

Orel

Hershiser

210

8.8%

59.4%

Jim

Fregosi

261

32.2%

73.7%

 

Curt

Simmons

210

8.8%

44.5%

Jack

Glasscock

261

32.2%

22.5%

 

Bob

Boone

210

8.8%

42.7%

Ellis

Burks

260

31.5%

22.8%

 

David

Wells

210

8.8%

23.1%

Roy

Thomas

260

31.5%

80.9%

 

Milt

Pappas

210

8.8%

34.2%

Paul

O'Neill

259

30.8%

55.9%

 

Dave

Stieb

210

8.8%

100.0%

Ken

Griffey

259

30.8%

47.8%

 

Eddie

Rommel

209

8.7%

100.0%

Guy

Hecker

259

30.8%

48.1%

 

Rico

Carty

209

8.7%

43.8%

Jorge

Posada

258

30.1%

100.0%

 

Garry

Templeton

209

8.7%

49.1%

Larry

Gardner

258

30.1%

48.5%

 

John

Anderson

209

8.7%

19.8%

Gus

Weyhing

258

30.1%

32.8%

 

Sherm

Lollar

209

8.7%

73.8%

Rick

Monday

258

30.1%

35.3%

 

Bill

White

209

8.7%

53.9%

Buddy

Myer

258

30.1%

76.9%

 

Eddie

Joost

209

8.7%

58.4%

Bucky

Walters

258

30.1%

60.0%

 

Tommy

Henrich

208

8.5%

100.0%

Gary

Matthews

257

29.4%

26.6%

 

Brickyard

Kennedy

208

8.5%

89.1%

Carl

Mays

256

28.7%

27.3%

 

Ed

Morris

208

8.5%

35.2%

Luis

Tiant

256

28.7%

42.5%

 

Lonny

Frey

208

8.5%

55.9%

George

Mullin

255

28.0%

85.5%

 

Reggie

Sanders

208

8.5%

29.6%

William

Hoy

254

27.3%

15.6%

 

Harry

Steinfeldt

208

8.5%

49.7%

Paul

Konerko

254

27.3%

99.2%

 

Bob

Friend

207

8.4%

98.1%

Maury

Wills

253

26.6%

63.8%

 

Devon

White

207

8.4%

26.5%

Wally

Joyner

253

26.6%

35.0%

 

Tommy

Davis

207

8.4%

34.3%

Curt

Schilling

252

25.9%

30.8%

 

Jimmy

Williams

207

8.4%

50.7%

Darryl

Strawberry

252

25.9%

62.2%

 

Bret

Boone

207

8.4%

36.7%

Dick

Bartell

252

25.9%

29.4%

 

Deacon

Phillippe

206

8.3%

80.2%

Andres

Galarraga

252

25.9%

30.5%

 

J. D.

Drew

206

8.3%

29.3%

Mike

Tiernan

251

25.2%

100.0%

 

Kenny

Rogers

206

8.3%

42.7%

George

Gore

250

24.5%

46.8%

 

Chris

Speier

206

8.3%

44.4%

Fred

Tenney

249

23.8%

80.0%

 

Ed

Reulbach

206

8.3%

58.8%

Carlos

Lee

249

23.8%

32.4%

 

Rico

Petrocelli

205

8.1%

100.0%

Gary

Gaetti

249

23.8%

38.4%

 

Ron

Gant

205

8.1%

32.8%

Paul

Hines

249

23.8%

42.2%

 

Doug

DeCinces

205

8.1%

45.2%

Billy

Pierce

248

23.1%

81.7%

 

Jim

Perry

205

8.1%

43.8%

Lance

Parrish

248

23.1%

47.8%

 

David

Cone

205

8.1%

24.3%

Tim

Wallach

248

23.1%

74.2%

 

Tom

Zachary

205

8.1%

27.8%

Jim

Gilliam

247

22.4%

100.0%

 

Claude

Ritchey

205

8.1%

47.1%

Greg

Luzinski

247

22.4%

62.0%

 

Hippo

Vaughn

205

8.1%

73.6%

Marquis

Grissom

247

22.4%

25.1%

 

Bobby

Thomson

205

8.1%

52.1%

Pedro

Guerrero

246

21.7%

55.8%

 

Ryan

Howard

204

8.0%

100.0%

Mike

Griffin

245

21.0%

53.1%

 

Tim

McCarver

204

8.0%

52.0%

Magglio

Ordonez

245

21.0%

50.9%

 

Pink

Hawley

204

8.0%

29.9%

Dusty

Baker

245

21.0%

43.5%

 

Murry

Dickson

204

8.0%

30.7%

Jay

Bell

245

21.0%

44.7%

 

Tommy

Harper

204

8.0%

28.2%

Jimmy

Dykes

245

21.0%

67.0%

 

Bob

Allison

203

7.9%

100.0%

Wally

Schang

245

21.0%

25.7%

 

Elston

Howard

203

7.9%

90.5%

S Sam

Jones

245

21.0%

23.0%

 

Wally

Pipp

203

7.9%

72.1%

Jason

Kendall

245

21.0%

49.7%

 

Tony

Cuccinello

203

7.9%

26.8%

Carney

Lansford

244

20.3%

48.1%

 

Garry

Maddox

203

7.9%

50.3%

Babe

Adams

243

19.6%

100.0%

 

A. J.

Pierzynski

203

7.9%

31.2%

Tommy

Bond

243

19.6%

58.5%

 

Ted

Kluszewski

203

7.9%

78.1%

Joe

Kuhel

243

19.6%

50.2%

 

Earl

Whitehill

203

7.9%

51.2%

Al

Orth

243

19.6%

35.7%

 

Ken

Williams

202

7.8%

81.0%

Albert

Belle

243

19.6%

45.3%

 

Fred

Pfeffer

202

7.8%

53.1%

Mike

Cameron

243

19.6%

20.9%

 

Terry

Pendleton

202

7.8%

36.8%

Ken

Caminiti

242

18.9%

47.5%

 

Marty

McManus

202

7.8%

35.2%

Del

Pratt

242

18.9%

34.8%

 

Gavvy

Cravath

202

7.8%

87.0%

Kevin

Brown

242

18.9%

23.6%

 

Vida

Blue

202

7.8%

50.2%

Bill

Madlock

242

18.9%

23.7%

 

Frank

McCormick

202

7.8%

75.1%

Wally

Berger

241

18.2%

68.3%

 

Bill

Donovan

202

7.8%

63.3%

Felipe

Alou

241

18.2%

37.0%

 

Claude

Osteen

201

7.6%

56.7%

Frank

Tanana

241

18.2%

33.2%

 

Willie

Kamm

201

7.6%

54.5%

Alfonso

Soriano

241

18.2%

26.6%

 

Luis

Castillo

201

7.6%

50.7%

Matt

Williams

241

18.2%

51.1%

 

Lenny

Dykstra

201

7.6%

48.4%

Cecil

Cooper

241

18.2%

72.9%

 

John

Titus

201

7.6%

73.5%

Johnny

Callison

241

18.2%

76.0%

 

Patsy

Donovan

201

7.6%

36.8%

Dick

McAuliffe

241

18.2%

95.1%

 

Bill

Dinneen

200

7.5%

33.4%

Dolf

Luque

241

18.2%

72.8%

 

Mo

Vaughn

200

7.5%

59.3%

Davey

Lopes

240

17.5%

58.5%

 

George

Burns

200

7.5%

26.4%

Rick

Reuschel

240

17.5%

45.9%

 

Jim

Sundberg

200

7.5%

63.4%

Jerry

Koosman

240

17.5%

49.7%

 

 

 

 

 

 

Joe

Carter

240

17.5%

40.7%

 

 

 

 

 

 

 

 

Appendix 6

How I did some little things

 

            John Dewan and Sports Info Solutions after the 2019 season provided me with a list of each player’s Win Shares in each season in major league history.  The list has not been updated since then, but that’s not a problem because we are studying the effect of playing for multiple teams (or for a single team) on Hall of Fame selections, and anything that has happened in the majors since 2019 is not relevant to the Hall of Fame anyway.  The list is 91,000 and some lines long.

            I sorted the list by (1) the player’s name, and (2) the team, and made totals of the number of Win Shares by each player for each franchise. If the franchise moved, I counted that as the same franchise, and if a player played for a team, left and came back to the same team, that is still counted as the same franchise.   I used codes like "1" for the Angels, "4" for the Baltimore Orioles (St. Louis Browns, 1901 Milwaukee), etc.   As I mentioned before any two-team player is coded just as "99" for that season, so that functions as an independent franchise.  A flaw in the study. 

            Then I made one-line summaries of each players’ Win Shares with different teams, T1 (team one), T2 (team two), etc.  The player who posted value for the most teams was Matt Stairs, who had value for 11 different franchises.  I think he played for more teams than that, but Montreal and Washington are the same franchise, and maybe he had zero value for one of them, I don’t know. 

            Anyway, with only one line per player, that cut the number of lines in the data to something like 19,000.   Then I eliminated everybody who had less than 150 career Win Shares, since players with less than 150 career Win Shares are not viable Hall of Fame candidates, although I think there may be one player somewhere who stumbled drunkenly into the Hall of Fame with less than 150.  Anyway, that cut the number of lines in the study to something more like 1,300. 

            Then I started eliminating players who were not eligible for the Hall of Fame.  99% of those players are not eligible because they have played in the last five years, but I also eliminated anyone who was not eligible for the Hall of Fame because he was banned from baseball, like Pete Rose or Shoeless Joe.  Their data would pollute the study, rather than inform it. 

            I also eliminated the managers who were good players and could have been elected as players or as managers, you’re not sure, like Joe Torre, Red Schoendienst, John McGraw, etc.   A manager elected with 250 Win Shares.. . what does that tell you?   Eventually we got down to 1,039 eligible players in the study.

            There is probably a player or two who snuck into the study with 150 wins in 9 seasons or less, not technically eligible for the Hall of Fame, although that didn’t stop them from electing Addie Joss.  I should have eliminated anybody like that, but I didn’t think about it until it was too late, and I’m not sure how I would have found all of those, anyway.  

            Thanks for reading. 

 

 

 

             

 

 

 
 

COMMENTS (20 Comments, most recent shown first)

Guy123
Guy's suggestion is intriguing, that much of the difference could be era. My gut would be surprised if that were the entire effect, but half seems quite plausible.

I agree that the effect Bill found is much too large to be explained entirely by the era in which a player played. Still, I think it plays some role. I would also guess that within each WS cohort, the HOF players have (on average) more career WS and a stronger peak. Ideally, all of that would be controlled for to isolate the single-team effect. More simply, it would be interesting just to repeat the analysis separately for players who retired, say, before and after 1960 (and remove players who are still on the ballot or who are obviously blocked by perceived PED use).
an hour ago
 
110phil
It seems like, roughly speaking, a "high" in one group gives you the equivalent HOF chance as a "low" in the next higher group. So a "Group 7 High" is roughly the same as a "Group 8 Low." This is not as obvious when looking at the percentages as it is after converting to an odds ratio as Tango suggested in comment 2.

For what it's worth, here are the odds ratios in those pairs:

24 vs 5.3 (9 high vs 10 low)
2.3 vs 3.0 (8 high vs. 9 low)
1.3 vs 0.9 (7 high vs. 8 low)
0.8 vs 1.1 (6 high vs. 7 low)
0.8 vs 0.3 (5 high vs. 6 low)
0.3 vs 0.3 (4 high vs 5 low)
0.1 vs 0.05 (3 high vs. 4 low)
0.1 vs 0.06 (2 high vs. 3 low)

The difference between categories seems to be 25 WS, so maybe you could say that the difference between "high" and "low," all else being equal, is 25 WS.

Guy's suggestion is intriguing, that much of the difference could be era. My gut would be surprised if that were the entire effect, but half seems quite plausible.
3:36 PM Jan 18th
 
rtayatay
Not only might someone who moves around a lot be perceived as having personality issues, someone who stays on one team might be perceived as loyal. I also think there is something to the way our minds like to put things into neat little boxes. My favorite example is Gil Hodges. I think Hodges probably has the highest percentage of his career value summed up in one decade on one team. It's easy to say 'Gil Hodges, 50's, Dodgers'. Compare Hodges to Mark Teixeira. Teixeira has 266 win shares and 50.6 WAR, Hodges 263 WS and 43.9 WAR. Very similar in terms of value. What comes to mind when you think of Teixeira? It's a mess of teams and seasons.
2:53 PM Jan 16th
 
CharlesSaeger
We haven't finished inducting Modern Era players yet. Cut off anyone who debuted after 1990 or so, maybe even 1980.
1:20 PM Jan 15th
 
shinsplint
jgf704, I thought that 40.0% for 2001-2010 was curiously low too. But there were 2 expansions in the 1990s adding 4 teams. So many of the players affected by those expansions (by being traded to or drafted by one of those teams) retired after 2000.
12:04 PM Jan 15th
 
hotstatrat
Speculating as to why Vagobonds do so much poorly than Homebodies in Hall of Fame balloting.

One reason as I believe Bill and/or others have already suggested is that players, who have a hidden negative such as some toxic affect on his teammates, are more likely to be passed along to other teams. Sportswriters who vote on the Hall of Fame may be at least vaguely aware of who those players are.

The other reason Bill or others sited is that, perhaps, a player's greatness is more difficult to see when it is divided by several teams.

Not mentioned yet (?) is that team loyalty is valued by baseball writers. Players who jump from team to team lose the love of their fans no doubt to some degree and the BBWAA factor that in when voting for the Hall of Fame. There is likely some pressure on baseball writers to honor homebodies from their fans more so than vagabonds.​
11:33 AM Jan 15th
 
jgf704
Adding onto what Guy123 wrote, and shinsplint's data, it does seem to me there are 2 different effects in concert:

* modern era players are more vagabondy (I expect the 1961-70 data point is an expansion-induced outlier; 2001-2010 is relatively low, and I don't have an explanation for that)

* modern era player Hall of Fame standards are higher
11:31 AM Jan 15th
 
shinsplint
I thought I'd research the one-team identification percentage (OTIP) by decade since 1901. I used Bill's technique, except I used games played instead of Win Shares in the formula. I found the average OTIP of all players who retired in that decade, and had at least a 9 year difference between the year of their debut and the year they retired. The columns are decade, number of players, OTIP.

1901-1910 142 44.6%
1911-1920 185 55.6
1921-1930 259 59.1
1931-1940 227 55.7
1941-1950 302 59.2
1951-1960 249 58.5
1961-1970 281 47.0
1971-1980 356 51.3
1981-1990 436 50.1
1991-2000 465 47.2
2001-2010 603 40.0
2011-2020 724 45.8

As you can see, the vagabond effect started to get much more pronounced starting around the 1960s, or perhaps even earlier in the 1950s since that's when the bulk of the career for many who retired in the 1960s was. In fact, the average player who retired in the 1960s, using my criteria, started his career in 1953 and retired in 1965. As you can see the 2001-2010 decade was the best one for Samsonite, with 40.0 OTIP. In fact, the player with the lowest OTIP ever using my criteria retired in 2009. It's Paul Bako, who played with 11 teams in 12 years. No wonder he never made the Hall of Fame! Well, that and his 62 OPS+.

Guy123 found that marginal HOF players by WAR who never made the HOF had an average debut year much earlier than those that made the HOF. I assume that's true for marginal HOF players by Win Shares as well. If so, then combined with my research, players who played most of their career before 1960 or so both 1) had a greater chance of being in the HOF if they had marginal WAR or Win Shares, and 2) had a higher average OTIP. Therefore, one could say that either Bill's theory of being a vagabond is deleterious to one's HOF chances is true, or that more marginal players were chosen in baseball's earlier days and it just so happened there were more one-team players then.

Another factor was alluded to earlier by abiggoof. Given "X" number of Win Shares, a lesser player will obtain that number in a greater number of years than a better player. Would it be fair to say that this lesser player will likely be traded more often, as lesser players are considered more expendable? If so, perhaps the vagabond factor is not as big a reason for getting shoulder shrugs from the HOF than is the fact that the player is not that good to begin. I have some evidence this may be true. I researched all the players on Bill's list in Group 6. These non Hall of Famers have an average career of 17.1 years. I assiduously dug up all the HOF players in Group 6 by using the site query for Win Shares trying players one by one. I found 24, even though Bill says there are 22. Not sure which is right, but the average career for these HOF players is 15.6 years--1.5 years shorter than the non-HOF players. So, maybe that's a factor as well? Probably less of a factor than the other factor I mention above.

10:16 AM Jan 15th
 
FrankD
Very interesting article, a lot of food for thought. Just an aside, Ruth is in this vain two players: Ruth the Pitcher (Boston) and Ruth the Hitter (Yanks). Or maybe three - the old, fat Ruth (Braves) kind a like the fat Elvis. You (Bill) have said this many times: a player who moves from team to team a lot is harder by fans to perceive. I would have thought this effect would be more prevalent in the past but I don't see your data showing this. Finally, there may be a perception or rationality that a player who changes teams a lot has something else (maybe hidden) that is a negative. Look at Hornsby.​
7:59 PM Jan 14th
 
llozada
Hey Bill,
Really interesting study. Have you thought about adding this consideration to your HOF Monitor? Maybe turning the team-identification into points is not that easy, but maybe something multiply the score by a factor? Something like: Joe Mauer shows up with a score of 92, but he is 100% identified with one team, that adds 10%, 5%, 15%?
Thanks-
2:46 PM Jan 14th
 
chrisbodig
As you alluded to, you first referenced the "one-team" benefit many years ago, prominently enough that I've had it on my brain for a long time.

I thought about this just last month when Tony Oliva and Gil Hodges were elected to the Hall while Dick Allen just missed. I had always thought of Hodges as a two-team candidate (Dodgers and Mets, as a manager) but that's still a one-city thing.

On the current ballot, I'd say this analysis helps to explain one of the many reasons why Jeff Kent (an obvious Hall of Famer to me) has been overlooked. Kent has some other issues: relatively low WAR and being a late-bloomer. I'm sure you mentioned this factor once as well but it seems to me that players who create most of their HOF value late in their career often struggle as well.

Most Hall of Famers imprint their immortality on our brains early in their careers. They're young, they're exciting, and they make All-Star teams. This was the case with Oliva, who was a Rookie of the Year, two-time batting champion, MVP contender in his first two seasons, and an All-Star in his first eight campaigns. And, of course, had a longtime PR campaign on his behalf from his lone ballclub.

Anyway, this study gives me hope for the future Cooperstown candidacy of Dustin Pedroia.
12:52 PM Jan 14th
 
Guy123
Following up, it does seem likely that some of what Bill's study found is an era effect rather than a one-team effect. His appendix lists non-HOF players in the study but not the HOFers, so I can't precisely replicate his study. But using WAR, players with 55-69 career WAR have about a 50% chance of making the HOF (higher for those closer to 69, lower for those close to 55). Of the 65 eligible players in that range, 35 are in the HOF and 30 are not (54% in).

The two pools are similar in average career WAR (HOF 63, non 60), plate appearances (9248, 9021), and OPS+ (126, 125). But there is one enormous difference between them: the years in which they played. The average HOF debuted in 1956 and played his last season in 1974; the non-HOF players, on average, debuted in 1975 and finished in 1992. That's an astonishingly large difference: the typical non-HOF player began his career a year after the average HOFer retired!

I can't replicate Bill's one-team metric, so I don't know how much higher the one-team identification rate was in earlier periods, but it must have been somewhat higher.
9:37 AM Jan 14th
 
Guy123
The HOF admission standard has been considerably higher in recent decades, and it seems possible that this explains some of the pattern identified in this study. It would be interesting to take each Win Share group and divide it into "old" and "new," based on retirement year, instead of low and high single-team identification. The difference in HOF admission rate might be nearly as high. And I would also expect a considerable overlap between the "old" and "one team" groups.
11:34 PM Jan 13th
 
Manushfan
Ahhhh Mr Manush is name checked a couple times here...yeah he ping ponged around some, has a Tigers hat for his plaque but was with the Senators a bit longer.
6:01 PM Jan 13th
 
Drewcannon3
This would be a really interesting finding. Like you, though, I’m a little suspicious because it’s so stark - could it be possible that timing is really what’s driving this? By that I mean that the eras when players were most heavily switching teams were pre-reserve clause and since about 1990, and those are also the least-represented time periods in the Hall of Fame. (19th century because it wasn’t really baseball, 1990s onward because of steroids and lack of time to go through the full election process.)
5:14 PM Jan 13th
 
sayhey
Thanks, Bill---honoured (I'm Phil Dellio).

I wonder if it's worth keeping closers out of the study. I don't know: when I think of Fingers, Gossage, Lee Smith, Mike Marshall, Sutter (3 teams in twelve seasons), and others, so many of them moved around a lot---I wonder if writers take the vagabond factor as a given with closers.

Countering that, though, there's Rivera and Hoffman. So maybe, ultimately, they're no different than anyone else.
4:04 PM Jan 13th
 
tangotiger
Bill:

Could you aggregate on pitchers and non-pitcher,s to show HOF prob and actual HOF? As you know, one of my contentions is that there aren't enough WSh given to pitchers (and SP specifically). It would be interesting to see if WSh see a bias with the voters in this regard.

You could even do it by "primary" position (Dawson a CF, because that's where his value is), to see if the HOF is biased by position.
2:33 PM Jan 13th
 
abiggoof
Great stuff. One thought: is there a correlation between how long it took players with comparable Win Shares to get them and how many teams they are at least moderately identified with? Or both factors and Hall selection?

I don’t mean 5 WS tacked on at the end with three or four teams because they stuck around extra years, but itinerant careers that kept going and have (usually) less overall value than players with similar playing time, or weaker peaks. Your Group 6 low examples strike me as having elongated or disjointed careers vs. Group 6 high.

DiMaggio had 387 in an abbreviated time, Staub 358 in forever, so there’s similar output, but not excellence. I think this leads to McGriff, Sheffield, Reggie Sanders, Blyleven, John, Lofton and others getting devalued in the public eye because they are seen as lesser, so-called accumulators like Moyer and Vizquel unless they have that ring or magic moment associated with them.
2:32 PM Jan 13th
 
tangotiger
Bill:

You can get around the 100% limit by turning your numbers into an odds ratio. For example, for group 8, you have 48% for the team-switchers. 48% is 48/52 or 0.92 HOF to non-HOF.

If you multiply that number by 3, you get 2.77, meaning 2.77 HOF to non-HOF, or a percentage of 73% (2.77 divided by 2.77+1).

If you do that across the board, you'll see that X3 works out pretty well.
2:21 PM Jan 13th
 
Gfletch
Funny thing for me to see Roberto Alomar identified as having a "scattershot" team identification. For me, Alomar is like 90% a Blue Jay, with a 10% afterthought memory that he came from the Padres. But that's just a personal fan's reaction (I loved the Blue Jays up until the mid 1990s). But of course I don't have a vote for the HoF, anyway.

Fun article, thanks.
1:58 PM Jan 13th
 
 
© 2011 Be Jolly, Inc. All Rights Reserved.