September 29, 2010

Baltimore Regional

Cleveland Regional

St. Louis Regional

Los Angeles Regional

BROOKS STAGGERS ON

            In the first matchup of the Brooks Robinson Invitational Tournament between two Hall of Famers, Brooks Robinson himself used a 5-4 margin in the “Team Success” category to defeat Jimmy Collins, 63-62, and move into his tournament’s Elite Eight.

            Both Robinson and Collins played for good teams in their careers, but Robinson played for better teams, in a longer career.

            The first sentence of Jimmy Collins’ Wikipedia entry reads, “James Joseph Collins (January 16, 1870 – March 6, 1943) was a Major League Baseball player at the turn of the 20th century who was widely regarded as being the best third baseman prior to Brooks Robinson.”  I believe that that’s accurate, with one quibble.  The mantle of being baseball’s greatest defensive third baseman passed not from Collins to Robinson, but from Collins to Pie Traynor, and from Traynor to Robinson.    It was not that Collins was rejected from that role, so much as that he was forgotten, and, as Traynor was fresher in memory, he stepped into that position.

            By the statistical method used here—which is a purely objective method, without reliance on any subjective judgments—Collins ranks first among these 66 third basemen in terms of the quality of his defensive performance.   We credit him with a “Fielding Won-Lost” record of 103 wins, 23 losses, an .818 percentage.  In other words, his defensive reputation appears, as best we are able to measure it, to be fully justified by the facts.

            We credit Collins with an overall won-lost record of 263-143, which is above the minimum standard of a Hall of Famer.    Collins’ career winning percentage is higher than Robinson’s, but Robinson’s marginal win contribution (above Collins) is .454—above the replacement level by any reckoning.  It is my view that Collins is a legitimate Hall of Fame selection, despite a relatively short career by Hall of Fame standards.

Jimmy Collins—Career Won and Lost Contributions

Ron Santo 94, Sal Bando 85

            It is funny the stupid stuff you remember 45 years later, but I remember the day the Kansas City A’s drafted Sal Bando (June 8, 1965); I remember telling my father that I had a good feeling about Bando because he sounded just like Ron Santo.    Right-handed hitting third baseman, slow, good power. . .even his name sounds the same, the same structure and rhythm.

            Bando was a college teammate of Rick Monday, who was the first player taken in that first draft, whereas Bando was a sixth-round pick—but for reasons I never understood, Bando from the get-go was the other player the A’s were excited about getting.  He immediately received much more attention than the A’s 2^nd, 3^rd, 4^th and 5^th round draft picks (Joe Keough, Bob Stinson, Pete Koegel, and Richard Suggs.)

            In his second year as a regular, 1969, Bando had a season that looks like a Ron Santo season—31 homers, .281, 113 RBI, plus 111 walks.   Bando didn’t hit at that level after that one season, but then, Santo played in a hitter’s park; Bando played in perhaps the worst hitter’s park in baseball in that era, the Oakland Coliseum.    In 1971 Bando hit .291 with 15 homers, 52 RBI on the road—Ron Santo numbers—but hit just .248 with 9 homers, 42 RBI at home.   In 1973 he hit .322 with 15 homers, 53 RBI on the road, but hit just .251 in his home park.   In his career he drove in 546 runs on the road, 496 at home.

            In two seasons (1969 and 1973) Bando played at something close to an MVP level, although he finished just 15^th and 6^th in the MVP voting.    Although Bando did have the advantage of playing on three straight World Championship teams, his star recognition was limited by his very low batting averages, playing in an era when batting average was still the king of batting stats.

            Bando was the captain and leader of the A’s championship teams.   He was a consistent performer, with six straight “20-win” seasons (27-5, 22-8, 22-8, 22-11, 27-6, and 20-11).   His team success percentage, .794, was the highest of any player in this tournament.   His offensive context, 3.89 runs per game, was the fourth-lowest in the tournament.

            By Win Shares and Loss Shares, Bando ranks just a hair behind Jimmy Collins (263-148 vs. 263-143), but I have decided to rank Bando ahead of Collins.   The difference between them in losses is too small to be convincing as to which player was better, and Bando has two uncounted advantages:

            1)  That he played 70 years later, against what must be thought to be better competition, and

            2)  That he was the leader of a great team, and that his teams were always successful to a very unusual extent.

            Collins also played for great teams, but Collins played for

            5 teams having highly successful seasons,

            1 having a moderately successful season,

            2 having neutral seasons,

            3 playing for teams having disappointing seasons, and

            4 teams having bad seasons.

            Bando played for

            7 teams having highly successful seasons,

            5 teams having moderately successful seasons,

            2 teams having neutral seasons,

            1 team having a disappointing season, and

            1 team having a bad season.

            So it’s not the same; Bando has a 12-2 ratio between good teams and bad teams; Collins has a 6-7 ratio.  And the one team with which Bando had a bad season, the 1967 A’s, he was in the minors most of the year.   I feel that Bando was a better player, and that’s the way I’m going to rate them. 

            Ron Santo’s marginal win contribution, above Bando, was .732. 

Sal Bando—Career Won and Lost Contributions

            As I have run out of ways to rate players in the wrap up, I’m going to respond to a couple of comments by readers.

            One reader suggested that, in evaluating a player’s fielding and batting contributions, I should use a higher “de facto replacement level” for fielding than for batting.   I’ve thought about this, and I’m not sure I’ve got the answer right, but. . .you don’t replace a player as “separates”, to borrow a phrase I have picked up from watching Project Runway with my wife.   You replace the player as a whole.  It is his overall performance that decides whether he keeps his job, not his performance in one role or the other.   If a player’s defensive performance drops below a tolerable level, he may be moved to a less demanding defensive position—left field, first base or DH—but he is not actually replaced.   I think that we have to use a unitary replacement level.

            Another reader argued that it is improper to use “awards” as a separate category in evaluating players, since the player is given the award for the same accomplishments that are already “scored” by the statistical evaluation.

            Well. . .yes and no.    The reason that we should pay attention to awards is to show respect for the opinions of others.   We don’t know everything; we can’t actually measure every phase of the game, at least yet.   When the point is reached at which we have measured everything, then we have nothing more to contribute to the baseball community, and at that point we should shut up and go away. 

            My good friend Alan Schwarz wrote a book tracing the history of baseball and numbers, in which he (I believe) argued that the history of sabermetrics actually goes back a long way, that there have always been people doing stuff like this.   Well. . .he’s wrong. 

            Look, there have always been people messing around with baseball statistics, using them to rate players, etc.    That stuff isn’t sabermetrics; it’s just farting around with numbers.

            What makes sabermetrics sabermetrics is that we push the envelope of what is known, and create new knowledge and original understanding by so doing.   When you take the things that you know and use them to evaluate players, that is using knowledge.   It isn’t creating knowledge, and it isn’t sabermetrics.   What I’m doing here. . .rating the players. . .this isn’t sabermetrics.   It doesn’t have anything to do with sabermetrics.

            The difference between sabermetrics and ratings is the difference between what you know and what you don’t know.    Ratings are knowledge-driven, and focus on what we know.   Sabermetrics is question-driven, and focuses on what we do not know.   All of the value is in focusing on what we don’t know, not in insisting of the importance of what we do.   By focusing on what we don’t know, what we can’t measure. . .by pushing our ignorance into sharply pointed questions, we acquire the ability to do research, thus expanding our range of understanding.  It’s a slow process and it’s frustrating, but that’s the justification for what we do; that’s what makes it matter—not that we know so much, but that we know so little.

            Respecting the opinions of others is central to the effort to focus more clearly on what we don’t understand.   If two players look alike in the statistics, but one of them has won six Gold Glove awards and made eight All-Star teams and the other was named Player of the Week for June 6-12, 1994, we need to try to understand that.   What is it that we’re not seeing in the statistics?

            You can, of course, say that the awards were decided as they were because the people who gave the awards were morons and they didn’t know what they were doing.   Sometimes that’s true.   Sometimes it’s arrogant.   The people who gave the awards may have known something that we don’t know, something we haven’t measured.    The creation of knowledge is essentially a process of outreach into ignorance, and we have to value the places where the knowledge that we are seeking might be hiding.