Perhaps the first lick of old baseball wisdom that I ever encountered was that championship teams are strong up the middle.   This was 1960, 1961; the fifties were still echoing in the wind, and it was still the general practice to look first, in selecting the MVP, at the catchers, the center fielders, the shortstops and the second basemen.   The first principle of baseball analysis, I suppose, was that baseball is 75% pitching, but the second was that championship teams are strong up the middle.

            With the development of Win Shares, it occurs to me that I am now in a stronger position to evaluate this theory.    I had Matthew Namee construct a spreadsheet, Win Shares by Teams, which has all of the Win Shares in history stored in a meaningful pattern—for example, the Win Shares earned by Jason Giambi in 2003 (28) are stored on the row set aside for the New York Yankees, 2003, and in the column set aside for the regular first basemen.    The Win Shares earned by Mike Piazza in 1993 (31) are stored on the row set aside for the Los Angeles Dodgers, 1993, and in the column reserved for the regular catcher, and the Win Shares earned by Wade Boggs in 1983 (34) are stored on the row marked for the Boston Red Sox, 1983, and in the column which is indicated for the team’s regular third baseman.  

            This spreadsheet makes it possible to test in a fairly simple, direct, and convincing manner whether it is or is not true that championship teams are strong up the middle.  First, however, I would like to ask you to do something.   I would like to ask you to set this aside for a moment, close your eyes, and think as deeply as you can about this question:  is it true that championship teams are strong up the middle?

            I have done studies of this issue before, I am sure—not as good as this study—and I have expressed my opinion before.   That opinion was that I thought it was probably true that championship teams were strong up the middle.   I was saying this, as much as anything, to avoid the argument.   The assertion requires the student, requires the person defending the assertion, to form a mental image of all of the championship teams within his knowledge, form a general impression of their strength at catcher, at second base, at short and in center, form a general impression of their overall strength, relate the general impression of the strength up the middle to the general impression of the overall strength, and then mentally weigh that against the comparable ratio on non-championship teams.   Meaning no disrespect to whatever sportswriter or manager first came up with this idea, this seems to me to be asking a great deal of the human mind.   Human beings are really not very good at this kind of mental exercise, which requires us to draw together a wide range of memories and associations, and balance one against another.   That’s why I ask you to try to think deeply about the issue; if you actually try to think it through, you should realize how nearly impossible it is to do this.

            Using this spreadsheet, Win Shares by Teams, I sorted out three groups of teams.   Group One was championship teams, Group Two was average teams, and Group Three was bad teams.   I used all major leagues, 1900-2003, including the Federal League and the strike-shortened 1981 and 1994 seasons.   Each league was represented by one team in each group—thus, there were 209 teams in each group, two per season from 1901 to 2003, plus the National League in 1900, plus the two seasons of the Federal League.

            For the “championship team”, I used

            a) whatever team represented the league in the World Series, if there was such a team (which there was for 200 of the 209 leagues), or

            b) whichever team had the highest winning percentage (for the other nine leagues).

            For the “average” teams, I used

            a) whichever team had a winning percentage closest to .500,

            b) with ties broken by the ratio of runs scored to runs allowed.   If two or three teams were equally close to .500, I used whichever team was closest to scoring exactly as many runs as they allowed. 

            For the “bad” teams,

            a) whichever team had the worst winning percentage,

            b) with ties broken by the worst ratio of runs scored to runs allowed. 

            The “championship” teams in the study had an overall winning percentage of .622 (101-61 in a 162-games season).   The “average” teams had an overall winning percentage of .500, and the “bad” teams had an overall winning percentage of .359 (58 wins in a 162-game schedule.)   The “bad” teams were further from .500 than good teams because, since 1969, the “championship” team has sometimes not been the team with the best won-lost record. 

            We are then in a position to compare the 209 regular first basemen on championship teams to 209 first basemen on average teams, representing the same leagues, and 209 first basemen on lousy teams, representing the same leagues.   If it is true that championship teams are strong up the middle, then it should be true that the advantage of the championship teams is larger at the “up the middle” positions than it is at the “corner” positions (first and third base, left and right field). 

            The conclusion of our study is that it appears to be unquestionably true that championship teams are strong up the middle.    Unless there is something wrong with my study, unless there is some hidden bias in the Win Shares system toward up the middle players on championship teams, then it is clearly true that championship teams are strong up the middle.  

            The 209 regular catchers on championship teams, in this study, earned a total of 3,424 Win Shares.   The regular catchers on average teams earned 2,417 Win Shares, and the regular catchers on bad teams earned 1,614 Win Shares.   The ratio of good teams to average teams was 1.42 to 1; average teams to bad teams had a ratio of 1.50 to 1, and good teams to bad teams, a ratio of 2.12 to 1.

            This chart summarizes the parallel data for all the eight regular positions:

            The ratios are obviously tied to the defensive spectrum.  The ratio between good and bad teams, at first base, is 1.66 to 1; at shortstop, it is 2.04 to 1. Both the ratios and the actual differences between the good teams and the bad teams are markedly higher at the “up-the-middle” positions than at the corner positions.  Sorting the chart above:

            The championship teams are 98% better than the bad teams at the “up the middle” positions, 73% better at the corner positions.   The largest values are at the “corner” positions, the positions played by Lou Gehrig, Albert Pujols, Ted Williams and Barry Bonds.  But the largest differences are at the up-the-middle positions.

            Although only one-seventh of the teams had Designated Hitters, the ratio between good and bad teams at DH, as we would expect, is even flatter than the ratio at first base:

            While at pitcher, the difference between good and bad teams is more notable in pitching depth than in front-line starting quality:

            However, this effect is not to be taken as reliable, due to the fact that the pitchers are sorted by quality after the fact. . .the man who starts the season as the team’s third starting pitcher often winds up the year as the number one starting pitcher, etc.   If we sorted infielders after the season as “first infielder”, “second infielder”, etc., best to worst, we would probably get a similar “depth effect” among infielders.  Also, the relief ace data cannot be compared directly to the data at other positions, since the time line is different for relief aces than for players at other positions.  Many of the early teams didn’t have relief aces, and the overall ratio of wins by good teams to wins by bad teams has changed significantly over time.

            In a sense. . .returning to the main topic here, which is strength up the middle. . .in a sense, it is more true that bad teams are weak up the middle than that good teams are strong up the middle.   Good teams are strong up the middle, but they are equally strong—even stronger, perhaps, at the corner positions.   But as you go down in the standings, the weakness becomes more apparent at the “up the middle” positions than it is at the corner positions.

            The strongest team in the study, up the middle, was the 1956 Yankees.  Their catcher, Yogi Berra, hit .298 with 30 homers, 105 RBI, while their center fielder, Mickey Mantle, won the Triple Crown and the Most Valuable Player Awards.   But the Yankee double play combination, which often was not good in those years, was also pretty decent in 1956.   Shortstop Gil McDougald hit .311 with 13 homers and a .405 on-base percentage, while second baseman Billy Martin hit .264 with 9 homers.    All four were also good defensive players, and the ’56 Yankees led the league in double plays by a wide margin.   While the second-place Cleveland Indians turned 130 double plays, the Yankees turned 214.  

            These are the ten strongest teams in the study, up the middle:

1.  1956 Yankees (C—Berra, 2B—Martin, SS—McDougald, CF—Mantle)

2.  1909 Pirates (C-G. Gibson, 2B-Dots Miller, SS—H. Wagner, CF—Tommy Leach)

2 (tie).  1953 Dodgers (C—Campanella, 2B—Jr. Gilliam, SS—Reese, CF—Snider)

2 (tie).  1949 Dodgers (C—Campanella, 2B—J. Robinson, SS—Reese, CF—Snider)

5.  1950 Yankees (C—Berra, 2B—Jer. Coleman, SS—Rizzuto, CF—DiMaggio)

6.  1961 Yankees (C—E. Howard, 2B—Richardson, SS—Kubek, CF—Mantle)

7.  1975 Reds (C—Bench, 2B—Morgan, SS—Concepcion, CF—Geronimo)

8.  1957 Yankees (C—Berra, 2B—Richardson, SS-McDougald, CF—Mantle)

9.  1972 Reds (C—Bench, 2B—Morgan, SS—Concepcion, CF—Bobby Tolan)

9 (tie) 1952 Dodgers (C—Campanella, 2B—J. Robinson, SS—Reese, CF—Snider)

            The 1949 and 1952 Dodgers had four Hall of Famers up the middle. 

            The ’56 Yankees are also the “most imbalanced” team of all time in terms of the comparison between their strength up the middle and their strength at the corner positions.   Up the middle they were formidable; at the corners they were unimpressive.    Their first baseman, Bill Skowron, was a good player--.308 with 23 homers--but on the other hand, he was a first baseman who did not hit as well as the catcher or the center fielder on the team.   The third baseman, Andy Carey, was a glove man who hit .237, the right fielder, Hank Bauer, was a power hitter who hit .241, and the left fielders were a collection of out-of-position guys like Elston Howard and Norm Siebern, both of whom became outstanding players, but about six years later and not playing left field. 

            The other end of that spectrum is interesting.   Of the three teams in history which were weakest up the middle compared to their strength at the corners, two were managed by the same man, Frankie Frisch.   Frisch was one of those guys who liked to say about his shortstop that he didn’t care what he hit as long as he did the job in the field, which is a real useful theory if you’re trying to get the crap beat out of you.   Frisch’s up-the-middle guys were such weak hitters that they dragged the team down.   His 1937 Cardinals had fantastic players in left field (Joe Medwick, who won the triple crown, also leading the league in hits, doubles and runs scored) and Johnny Mize (who was second in the league in batting average, on base percentage, slugging percentage, doubles and total bases, third in RBI, fourth in home runs.)    Up the middle, however, he had one good player (Terry Moore) and three glove guys who didn’t hit.    Given the league’s two best hitters, they finished fourth.    Frisch’s ‘44 Pirates were the same—very strong at the corners, but so punchless through the middle that they fell short overall. 

            Of the “average” teams in our study—the 209 teams that finished at or near .500—not a single one was of anything like comparable strength up the middle to these ten teams listed before.   The top five teams above are not only the top five in this study, but also the top five of all time; there is no non-championship team which was equally strong up the middle.

            Using the standard of 60 Win Shares at the four up the middle positions to indicate strength up the middle, 84% of the championship teams in this study were strong up the middle, whereas only 43% of the average teams were strong up the middle, and only 2% of the bad teams were strong up the middle.   This fact, however, is potentially misleading, because there is a very similar breakdown if you use 60 Win Shares at the four corner positions to represent strength at the corners.    By that standard, 92% of the championship teams were strong at the corners, as opposed to 64% of the average teams and 14% of the bad teams.   

            Good teams are strong everywhere; we all know this.   However, we are asking the simple question:  is it true that good teams are strong up the middle?  Certainly it is true that good teams are strong up the middle, and this appears to be distinctly more true than saying that good teams are strong at the corners, or even that good teams have strong pitching staffs.    One could romanticize this fact inappropriately, and use it as an excuse to vote for Marcus Giles as the 2004 NL MVP, even if Scott Rolen or Albert Pujols is actually a better player.   It’s not a universal, sweeping truth; it is not a magic bullet explanation of why some teams win and others lose.   But the best answer to the question is:  Yes.   Championship teams are strong up the middle.