You may remember that, before I was so rudely interrupted by making a series of one-day and two-day business trips to every city in North America, I was in the process of explaining how we calculate Win Shares and Loss Shares.   This process began on May 18, 2009, with an announcement that it would be a long and boring process, and I posted boring installments of it several times, most recently on July 6, 2009, when I explained how Win Shares and Loss Shares are figured for catchers (as fielders.) 

            Our next task is to figure defensive (Fielding) Win Shares for first basemen.      I will start this, just to re-establish our position, by quoting the key phrases from the July 6 article about Win Shares and Loss Shares for catchers:

The catcher’s defensive win shares are:

a) the team’s defensive area of responsibility,

b) times .154,

c) times a number close to the team’s defensive winning percentage,

d) times the playing time for this individual catcher,

e) modified by seven other factors.

The seven other factors are:

1. The team’s strikeouts,

2. The team’s walks,

3. The catcher’s errors,

4. The catcher’s assists,

5. The catcher’s passed balls,

6. The catcher’s stolen bases allowed (where data is available), and

7. The catcher’s runners thrown out (where data is available). 

            I see an error in that now;  there’s actually a point (d) that was left out.   In a parallel form, the first baseman’s defensive win shares are:

            a)  the team’s defensive area of responsibility,

            b)  times .099,

            c)  times a number close to the team’s defensive winning percentage,

            d)  times this individual first baseman’s share of the innings at first base,

            e)  modified by four other factors.

            The four other factors are:

            1)  An estimate of the plays made by the team’s first basemen,

            2)  The first baseman’s errors,

            3)  The errors made by the team’s third basemen and shortstops, and

            4)  The “arm rating” of the team’s first basemen.

            To illustrate the process, I’ll compare two first basemen:   Dick Stuart in 1963, and Wes Parker in 1965.  

            Dick Stuart in 1963 led the American League in Home Runs (42) and RBI (118).    He was a notoriously horrible fielder, nicknamed Dr. Strangeglove after a classic movie of the era (Dr. Strangelove).   He did, however, lead all major league first basemen in assists in 1963, with 134.   Wes Parker was the first baseman for the World Championship Los Angeles Dodgers in 1965.   Playing 154 games and batting 542 times, he hit .238 with 8 homers and 51 RBI.   Obviously, he was regarded as a first-class defensive player, or he wouldn’t have been in the lineup with those numbers.   He did not win the Gold Glove in 1965 (Bill White did), but he did later win six consecutive Gold Gloves at first base.  

            Stuart and Parker were joined in a controversy in 1966.   On July 5, 1966—starving for power--the Dodgers signed Dick Stuart, who had been released by the Mets on June 15.    Between then and August 14, Stuart started 20 games for the Dodgers, putting Parker on the bench.    Unhappy about this situation, Parker spoke out about it, saying that if the Dodgers were going to repeat as National League champions, they needed to get him back in the lineup.    Several veteran Dodger players, mostly off the record, told reporters that Parker was right; the Dodgers couldn’t win with Stuart’s glove at first base.

            In that era lineup decisions were normally left entirely in the hands of the manager, but the Dodgers were different; the Dodgers operated like a corporation, getting input from all sources and making a decision based on consensus whenever possible.   After a couple of long and contentious meetings, it was decided to put Parker back in the lineup and Stuart back on the bench. 

            But this wasn’t why I chose Parker and Stuart for these roles; that’s a total coincidence.    I’ll explain later why I’m using Parker and Stuart in this way.  

            OK, the Boston Red Sox in 1963 had 62.88 Defensive Win Shares and Loss Shares.   The Dodgers in 1965 had 71.18 Defensive Win Shares and Loss Shares.   The figure is slightly larger for the Dodgers because the Red Sox allowed more home runs and gave up 100+ more walks.   When the pitchers give up walks and homers, in our system, that takes the ball out of the hands of the defense, as there’s nothing the fielders can do about that.   The Dodgers did offset some of the difference with more strikeouts, but actually, the ’65 Dodgers didn’t have a lot more strikeouts than the ’63 Red Sox did, just 70 more.    The balance of all the factors is that the Dodger defense had more opportunity to have an impact on the game than the Red Sox defense did.  

            Both Stuart and Parker, oddly enough, played 1,376 innings in the field in the season in question.   Based on this, on the number of outs made by each hitter and by the team innings total, Parker is assigned responsibility for 7.78 Defensive “Games”—defensive Win Shares and Loss Shares—while Stuart is assigned responsibility for 7.15.    In that the Dodgers are more of a defense-dependant team than the Red Sox, Parker is assigned a slightly larger area of responsibility.

            But backing up a half-step, the Red Sox had 62.88 Defensive Win Shares and Loss Shares, of which .099 are assigned to the team’s first basemen as an operating base for their Win Shares and Loss Shares.   That’s 6.23 Win Shares for their first basemen, potentially.  For the Dodgers, this is 71.18 times .099, which is 7.05.

            This figure is then modified by the defensive success of the team.   The reason this is done is that the overall defensive records of good defensive teams tend to look very much the same as the overall defensive records of bad defensive teams.    If you put nine good fielders in the field, they’ll make 27 outs a game.   If you put nine horrible fielders in the field, they’ll make 27 outs a game.    If you want your system to recognize the difference between a good fielding team and a poor fielding team, you have to force the difference into the system.   We force it into the system here, by multiplying this “base operating number” by the defensive success percentage of the team.  

            The defensive success percentage of the Dodgers in 1965 was .595; I have explained before how this number was calculated, and I won’t repeat that here.   The defensive success percentage of the Red Sox in 1963 was .503—which is surprisingly good.   At times in the past I have ridiculed the defensive quality of the 1963-1964 Red Sox, suggesting that the ’64 Red Sox were the worst defensive team I ever saw.   The 1964 Red Sox, yes—but the 1963 Red Sox actually were OK in the field.    They fell apart defensively in ’64 due to replacing Chuck Schilling at second base with Felix Mantilla, due to the aging of Frank Malzone, and due to using Carl Yastrzemski out of position in center field.  

            Anyhoo, we modify the player’s base operating number by the team’s defensive success, which produces

            This is a number that represents the operating base of the team’s first basemen.   We modify this by assigning a share of these to the individual first baseman based on how many innings he played.  For Stuart, this is 1,376 out of 1,450.  For Parker, it is 1,376 out of 1,476:

            What we’re saying here exactly is that, if we had no knowledge about Dick Stuart’s defense as an individual, we would assign him credit for 2.97 Win Shares, based on his innings played, the team that he played for and the defensive success of that team. 

            Next we look at the number of plays made by the team’s first basemen—or actually, at an estimate of it.   Now, with Retrosheet, we actually have or could have records of the numbers of plays initiated by each fielder.   I created this system in an environment in which we did not have that information, and I haven’t adjusted to that yet.   I’m still estimating something that we could now count.

            How do we estimate that?   First base plays made are assumed to be the sum of Assists, and unassisted putouts.    The problem is, there is no official record of unassisted putouts by first basemen.   We can estimate it, however, in two ways.   One, which we call UPO-1 (Unassisted Putouts-1), is  the number of putouts by the team’s first basemen, minus .70 times the number of assists by the team’s pitchers, minus .86 times assists by the team’s second basemen, minus .78 times the assists by the team’s third basemen and shortstops, plus .115 times the team’s estimated runners on first base, minus .0575 times the number of outs recorded by the team.  You got that?

            No, of course you don’t.    Let me try to explain.   We are looking for the number of putouts by first basemen which aren’t attributable to assists by the other infielders.   We assume that 86% of the assists by second basemen go 4-3, so we subtract those.   We estimate that 78% of assists by the third baseman are 5-3, so we subtract those.   Those we don’t subtract are the unassisted putouts made by the first basemen. 

            I know that this formula works at some level, and also, I know that it’s not too precise or too reliable.  I am confident that we get higher numbers, by this formula, for first basemen who make more plays 3 unassisted than for first basemen who make fewer.   I am also confident that our estimate is often wrong by some substantial amount.  

            The adjustment for estimated runners on first base (which was explained earlier). ..that adjustment is necessary because, when a team has more runners on base against them, there are more plays made at second base or third base, which means fewer plays going 3-unassisted.    So we figure that formula for every team, and that’s UPO-1.

            UPO-2 is the number of Balls in Play against the team, minus the hits against the team, times .1, minus the assists by the first baseman.   UPO-2 is not an estimate of what the number is for this team, but simply of what it tends to be, on average.   The function of UPO-2 is to stabilize or moderate the data from UPO-1—to draw the numbers toward the norm.   We take two times UPO-1, plus one times UPO-2, divided by 3.    That’s UPO.  

            UPO for the 1965 Dodgers is 216.   For the 1963 Red Sox, it is 170.   To this, we add the ASSISTS of the team’s first basemen.   Dodger first basemen had 101 assists; Red Sox first basemen—remember, Stuart led the majors in assists at first base—had 139.   We add the UPO (Estimated Unassisted Putouts) to the Assists, and we get the Plays Made by the team’s first basemen.   For the Dodgers, this is 317.   For the Red Sox, it is 309.   

            If you are thinking that this is an awfully complicated way to get a not very reliable estimate, the answer is “yes, that’s right.  It is.”   The alternative, at the time I was developing these formulas, was to throw up my arms and say “I can’t figure out how to measure the range of a first baseman.”   I chose not to do that.    I did the best I could with the data that I had.  

            The next step in this marathon process is to compare the estimated plays made by each team’s first basemen (317 for the Dodgers, 309 for the Red Sox) to the Expected Plays made by the team’s first basemen.   The Expected Plays made by the team’s first basemen is

            a)  The league total of plays made by the first basemen,

            b)  Times the number of balls in play against the team,

            c)  Divided by the number of balls in play against the league,

            d)  Adjusted for the left/right bias of the pitching staff.  

            In the American League in 1963, an average of 423.5 innings per team were pitched by left-handed pitchers.    The Red Sox had 233 innings by left-handed pitchers—190 less than average.   Because they had a below-average number of left-handed innings, they probably faced an above-average amount of left-handed hitting, which means that they probably had an above-average number of balls hit in the direction of first base.   For each inning that they were short of left-handed pitching, we increase the expected number of plays at first base by .08.   If you have a hundred extra innings of left-handed pitching, you’ll have 8 fewer balls hit toward the first baseman.

            The 1965 Dodgers, on the other hand, had 963 innings of left-handed pitching, which is the highest total of all time, and had 2,787 balls put into play against left-handed pitching.   They consequently probably had fewer balls hit toward first base.    Making these adjustments, we conclude that the Expected Plays Made by the 1963 Red Sox first basemen are 348, whereas the Expected Plays Made by Dodger first basemen in 1965 are 298.  

            The Dodgers, then are +19 in plays made versus expectation (317 minus 298), whereas the Red Sox are -39 (309 minus 348).  Those include the plays made by ALL of the team’s first basemen, and, of course, neither Parker nor Stuart played 100% of his team’s innings at first base.   Adjusting for that, it’s +18 for Parker, and -37 for Stuart.   For each play made above or below expectation, we credit (or charge) the first baseman with one-fiftieth of a Win Share, or .02.  

            Remember, we had Stuart at 2.97 Win Shares, and Parker at 3.91.   We now amend these figures:

            And this completes Step One of the four adjustments that were outlined at the beginning of the article.

            The next adjustment is easy.   Our next adjustment is for Errors.  Dick Stuart in 1963 made 29 errors in 1,370 plays.   An average American League first baseman would have made 13.54 errors in 1,370 plays, so Stuart was 15.46 errors above expectation.   

            Wes Parker, on the other hand, made 5 errors in 1,534 plays.   An average National League first baseman would have made 15.92 errors in 1,534 plays, so Parker was 10.92 errors below expectation.    For each 16 errors that the player makes above expectation, we charge him with one Win Share—or conversely, credit him with extra Win Shares for not making errors.    For Parker, this is +.68.   For Stuart, it is -.97.  With these adjustments, Stuart goes to 1.26, and Parker to 4.95:

            Our next adjustment is for the errors made by the team’s third basemen and shortstops.   The theory is that, as most errors by third basemen and shortstops are on throws to first base, a good first baseman should be able to reduce this number of errors.    We thus add up the errors made by the team’s third basemen and shortstops, compare that to the league norms (based on the number of plays made by the third basemen, of course), and credit the first baseman with one one-hundredth of a Win Share (.01) for each error prevented or allowed.  

            Both the Dodgers and the Red Sox had a below-expected number of errors by their teams’ third basemen and shortstops.    The Dodgers were under by 3.68 errors—a bonus of .0368, or .04—and the Red Sox were under by 8.47, giving them .0847 or .08.   With these adjustments, Stuart moves up to 1.34, and Parker up to 4.99:

            OK, we’re almost there.   I have one more adjustment.

            Assists by first basemen, as a naked category, are not a reliable quality indicator, because the great majority of assists by first basemen are 3-1 plays, first baseman flips to the pitcher covering.   Some first basemen strongly prefer to flip the ball to the pitcher covering, while some first basemen strongly prefer to make the play themselves, 3-unassisted.   It is not proper accounting to say that a first baseman who has lots of assists has good range, because all it really indicates is that he likes to make the play himself.    Mark Teixeira last year had an absurdly low assists total (49), but we all know that Teixeira is a very mobile first baseman.

            What one could miss, in making this point about assists, is this:  that when a first baseman has an assist that does not go 3-1, it is almost always a skill play (as opposed to a routine play), and it is almost always an important play.   Think about it.   If a first baseman’s assist doesn’t go 3-1, then how does it go?   It’s either 3-2, 3-5, 3-6 or 3-4.   

            When there is a play going 3-2, that means that the first baseman has thrown somebody out at home plate.   That’s a huge play.   That’s a run. 

            When there is a play going 3-5, that means that the first baseman has gunned the ball across the infield and either picked off a runner napping, or thrown out a runner trying to move up a base.   Again, that’s a huge play, when that happens, and it’s a skill play.

            When there is a play going 3-4 or 3-6, that (usually) means that either there has been a double play started by the first baseman (3-6-3 or 3-6-1), or that the first baseman has made the play to second and kept a runner out of scoring position.   These are the most common plays that involve a first baseman’s assist and are not 3-1, but again, that’s a skill play, and that matters.   There are some first basemen who can make that play, keep the double play in order; there are other first basemen who can’t.   Steve Garvey was a fine defensive first baseman and won several Gold Gloves, but he couldn’t throw, and he was afraid to throw; on a ground ball to first he would just take the play at first and let the runner move up, rather than risk throwing the ball away.   

            We attempt to measure this by the “Arm Rating”, again, on a team level.    We look at the team data, for an “Arm Rating” for the team’s first basemen, and then apply that to each first baseman on the team on a pro rata basis.   The Arm Rating consists of two sets of two data points:

            a)  The assists by the team’s first basemen, minus the putouts by their pitchers, and

            b)  The double plays turned by the team’s shortstops, minus the double plays turned by their second basemen. 

            (a) is probably obvious, based on what I said before; the assists that count by first basemen are the ones that aren’t putouts by the pitcher.   The theory of (b) is that, if the first baseman throws well, the team will have more double plays going 3-6-1 or 3-6-3 than 5-4-3—thus, more double plays involving the shortstop than the second baseman.   

            The Red Sox in 1963 had 139 assists by first basemen, and 126 putouts by pitchers, so that’s about 13 assists by first basemen that weren’t 3-1 (understanding, of course, that not all putouts by pitchers are 3-1 plays.)   They had 90 double plays by their second basemen, and 92 by their shortstops, so that’s +2.   +13, +2; their first base Arm Rating is 15.

            The Dodgers had 101 assists by first basemen, but only 57 putouts by pitchers, so that’s +44.    They had 94 double plays by second basemen, only 90 by shortstops, so that’s -4.   +44, -4; their first base Arm Rating is 40.

            The American League average in 1963 was +10, whereas the National League average in 1965 was +16.    The Red Sox were +5 versus the league average (15 minus 10), whereas the Dodgers were +24 (40 minus 16).    Pro rated to the innings, Parker is +22, Stuart is +4.5.

            For each 50 plays made in this way, we credit the player with one Win Share (.02 per play.)   Stuart thus gains .09 Win Shares, and Parker gains .44:

            So Stuart winds up with 1.43 Win Shares, for his defense at first base in 1963, and Parker winds up with 5.43 for his.   There’s a rounding error, of course; it’s actually 1.44 for Stuart, 5.42 for Parker.

            Way back early on, we assigned Stuart the responsibility for 7.15 Defensive Games, and Parker the responsibility for 7.78.   For Stuart 1.44 of these are Wins.   The other 5.71 are losses.   His defensive won-lost record for the season is 1.44 – 5.71, or 1 and 6.  His defensive winning percentage for the season was .201.

            Parker, with 5.42 Win Shares, has 2.36 Loss Shares—a defensive won-lost record for the season of 5 and 2, and a defensive winning percentage of .697. 

            I promised to tell you later why I chose Parker and Stuart for these assignments.   In an effort to demonstrate to you that my first base defensive system does in fact work, I chose 15 major league first basemen from the last half-century—five of them very good defensive players (by reputation), five of them very poor defensive players (by reputation), and five of them middle-of-the-road defensive first basemen, players who didn’t win Gold Gloves, but people didn’t make fun of their defense, either.

            For the “good” defensive first basemen, I chose Vic Power, Keith Hernandez, Wes Parker, Steve Garvey and Mark Teixeira.   For the “bad” defensive first basemen, I chose Stuart, Willie Aikens, Mo Vaughn, Richie Sexson and Jason Giambi.    For the “middle of the road” first basemen, I chose Fred McGriff, Lee May, Norm Cash, Carlos Delgado and Joe Pepitone.     I then figured the Win Shares and Loss Shares—offensive and defensive—for all of those players for each season of their careers.

            Wes Parker in 1965 was 3.06 Win Shares over .500, as a defensive player—the second highest-rated first base season that I found.    These were the top ten:

            1.   Fred McGriff, 1995

            2.  Wes Parker, 1965

            3.  Steve Garvey, 1976

            4.  Vic Power, 1954 (playing center field)

            5.  Wes Parker, 1966

            6.  Keith Hernandez, 1985

            7.  Vic Power, 1962

            8.  Keith Hernandez, 1982

            9.  Steve Garvey, 1975

            10.  Keith Hernandez, 1981

            I didn’t want to use Fred McGriff’s 1995 season to illustrate the process because it’s a fluke….he has a tremendously good rating in that season, very ordinary numbers in all of the surrounding seasons, so I didn’t think he was an appropriate exemplar.    These were the ten worst seasons (as measured by Win Shares minus Loss Shares):

            1.  Dick Stuart, 1964

            2.  Dick Stuart, 1961

            3.  Dick Stuart, 1963

            4.  Jason Giambi, 1997 (playing the outfield)

            5.  Lee May, 1978

            6.  Mo Vaughn, 1997

            7.  Dick Stuart, 1959

            8.  Mo Vaughn, 2000

            9.  Fred McGriff, 1999

            10.  Lee May, 1979

            I actually used Stuart in 1963, rather than 1964, because I mixed up my note. ..I thought 1963 was the worst year.   Not that it matters, and I’m not going to re-edit all of this to show that Stuart was even worse in 1964 than 1963.  

How the 15 Defensive First Basemen Rate

            First basemen are much, much more difficult to rate defensively by the use of traditional defensive statistics than are players at any other position.   While there are many problems with the defensive statistics of second basemen or shortstops or center fielders, there is a general and apparent connection between defensive excellence and defensive statistics, in most cases.    At first base, because there are no easy “range” numbers for first basemen, this is much less true.   There have been more totally absurd defensive ratings given for first basemen, I would venture, than for all other positions combined—rankings arguing that the statistics show that Marv Throneberry was actually a good defensive first baseman, or that Zeke Bonura was—while the defensive excellence of players like Jim Spencer and Dave Bergman, even though they stay in the majors for many years to play late-inning defense at first base, is almost impossible to prove by statistical analysis.

            Let me say this about my defensive ratings for first basemen:  Yes, they are complicated.  Yes, they rely on extremely circuitous logic which could be wrong, in individual cases, for any of a thousand reasons.   Yes, the ratings are wrong sometimes; it is not a system that bats 1.000. 

            But the ratings are right much more often than they are wrong.    This is the essential claim that I make:  that this complicated, convoluted schematic for teasing the defense of first basemen out of a hellish maze of statistical puzzles does in fact work the great majority of the time.   It does reliably give higher ratings for good defensive first basemen than for poor defensive first basemen—not reliably as in “every season”, but reliably over the course of a career.     You can buy that, you can not buy it, but this article is a down payment on an effort to demonstrate that this is true. 

            A couple of additional caveats.   The Defensive Win Shares and Loss Shares system is not set up to give a defensive winning percentage at a particular position.    The player has a defensive assignment, which is based on how many innings he plays in the field and on how many outs he makes, and he has defensive credits that count toward the fulfillment of that assignment.    But the system is not set up so that the credits a player gets for playing first base may be distinguished from those he gets from playing another position—thus, if a player plays first base and other defensive positions, he still has only one defensive won-lost record, not a defensive won-lost record for each position.   I know that many of you would prefer that it worked the other way, but. . .it doesn’t.   I don’t know how to make it work that way.  

            Also, the defensive norms at each position are different.    Let us suppose that you compare a first baseman who hits .275 with 16 homers and an .825 OPS, and a shortstop who hits .270 with 15 homers and an .800 OPS.

            The way that many analytical systems are constructed, we would say that the shortstop who hits .270 with an .800 OPS is a better hitter than the first baseman who hits .275 with an .825 OPS.      He is a better hitter because he is a shortstop.

            OK, I’m not arguing with you, exactly, but. . .that’s not the way this system works.    He is not a better hitter.   A shortstop who has an .800 OPS is NOT a better hitter than a first baseman with an .825 OPS; he just isn’t.    He’s a better fielder.   

            My system is set up to reflect what I think is the reality of the game:  that shortstops and catchers are better fielders than first basemen.   I’m not saying that every shortstop is better defensively than every first baseman.   Wes Parker against Rafael Ramirez, Wes Parker will win in a walk.    But, in general, shortstops and catchers make a larger contribution to the defensive success of the team than do first basemen or left fielders.

            Many systems are set up to “normalize” these differences out of existence, to say that the average-hitting left fielder is an average hitter, that the average-fielding shortstop is an average fielder.   I don’t believe in that.   I believe that first basemen are better hitters than shortstops, and shortstops are better fielders than first basemen.   I think this is obvious.   I think that the mathematical analysis of the game should be set up to reflect what is obviously true, rather than imposing a fictional average on the whole.

            Thus, in my system, shortstops tend to be .600 or .650 fielders, whereas first basemen tend to be .400 or .425 fielders.   You can agree, you can disagree; I believe that that’s the way it should be.

            OK, getting now to my list of fifteen defensive first basemen.    In another article, to be published tomorrow, I’ll compare these 15 players as complete players.  For right now, I’m just talking about their defense.    These are the career defensive won-lost records of the five players that I selected because they are, by reputation, outstanding defensive first basemen:

            Four of the first basemen that I selected because I believed they would show up as excellent defensive did in fact do so.   One—Teixeira—did not. 

            I am disappointed in this, and I understand it to an extent but only to an extent.   I’ll give you four points:

            1)   Teixeira’s defensive won-lost record as a rookie in 2003, when he was playing some third base and some in the outfield, was 1-4.   This poor start is still dragging him down.

            2)   Teixeira—like almost all modern first basemen—sometimes DHs.    He has DH’d 31 games in his career.   The games that a player fills the DH role go into his record as zeroes, since he makes no defensive contribution in those games.

            3)  Teixeira had only 49 assists in 2009, which is an extraordinarily low total.   The National League Gold Glove candidates, Albert Pujols and Adrian Gonzalez, had 185 and 136 assists, respectively.  It is really hard to rescue his defensive evaluation when he just doesn’t have any assists to work with.

            4)   Even with a defensive winning percentage of .454, Teixeira is still above average for a first baseman.

            But these four points are in the nature of whining or pleading.    The reality is, my system does not properly evaluate the defensive contribution of Mark Teixeira, who is shown as an outstanding defensive first baseman by most modern systems, and who I believe is in fact a brilliant defender.    Our system just doesn’t work for him, and I can’t really explain why.  

            Our “bad fielding” first basemen were evaluated as follows:

            Thus, while our system evaluates Richie Sexson as being better defensively than I thought he was and Mark Teixeira not as good, it is still true that all of our “good” defensive first basemen score better than any of our “bad” defensive first basemen, and the overall gap between the groups is quite large.   

            Now, are there any examples of players who were regarded as Gold Glove fielders, but who score badly in our system, or guys who were regarded as bad fielders who score as outstanding?   Ehnn. .. probably are.   I don’t know of any such examples.   I’m leery of running the data for George Scott; he might be a problem for us.   Probably if we ran the data for all first basemen in major league history, we would find some dramatic example of a case where our system totally fails.   I don’t know what it is, and I do not know of any such case.

            In a perfect world, all of our “mid-range” defensive first basemen would rank below Teixeira and above Sexson, but, as the gap between the two is only 48 points (.454 to .406), that clearly is not going to happen.  These are the defensive won-lost records of the players whose defensive reputation was in the middle group:

            So they’re pretty close to fitting neatly in the middle.   All of the “top five” fielders score as better than any of the mid-range fielders, with the sole exceptions that Cash ranks better than Teixeira and equal with Garvey.   All of the “middle five” score as better than any of the “iron fingers” group, with the sole exception that Richie Sexson ranks as better than Lee May.

            Tomorrow, I’ll look at these 15 players as offense/defense packages.