15.  Your Author is a Slow Learner

              To this point in this process I have been rating players based on the information which could have been known at that time.   The best catchers in 1947 are evaluated based on modern analysis of the facts as they were known at the end of the 1947 season.    I say "to this point in the process", but actually I was doing that well beyond the point in the process that we have reached together.  I generated complete rankings based on that process, and was merely explaining and defending my conclusions.  

              Well down the road from where we are now, in stuff that I have written but not yet published, I found some rankings that were hard to defend.   This began to trouble me.  I began to justify this approach in my head, but found it difficult to do.   Why exactly are we saying that, in evaluating the performance of players from 1947, what those players did in 1946 is relevant to our understanding of how good the really were, but what happened in 1948 is not?           

              The breaking point came for me when I was working on the first basemen of the 1980s, and I needed to explain why, in 1985, Don Mattingly ranked second.  Mattingly drove in 145 runs in 1985.  He hit .324, which was actually well below his average in 1984 or 1986.  He ranked second behind Eddie Murray, which is hardly a disgrace; Eddie was also a great player, but still, Mattingly had 32 Win Shares in 1985, to Murray’s 28.   Why is Murray ahead, in evaluating where those two players were at that moment?

              He is ahead, I told myself, because at that time we could not have known for certain that Mattingly was not having a Norm Cash season, a season like Walt Dropo had in 1950 or a season like Mark Trumbo had in 2016, a season above his real level of ability.

              Well, yes, responded the wiser Evil Bill, but this is 2017.   The issue in 2017 is not what we could have known for certain in 1985; it is what we can know for certain in 2017 about 1985.   Mattingly was not in any sense a fluke.  He was a tremendous player.   We are discussing the careers that players have had, more than we are discussing careers which they are in the middle of having.

              I was using the data through 1985 to evaluate the players of 1985, I realized, because I am in the habit of doing so.   The problem of best use of the backward-looking data is one with which I have long wrestled.   In the winter of 1985, I had to rate the players of 1985 based on what they had done through 1985, in order to produce the 1986 Baseball Abstract.   One month from now I will be on MLB television, and I will be ranking the players of 2018, based on the information available through 2017.   I look at the problem that way because I have long been in the habit of looking at it that way.

              But it isn’t right; for this purpose it isn’t right.   For what we are doing right now, 1946 is no more relevant to 1947 than 1948 is.  That one should consider what players have done over a period of seasons to determine who is actually the best player is a valid approach, but that we are trapped in the position of being only able to look at it from one direction is not true.  

              Unfortunately, by the time I reached this realization, I had written these articles up through first base, 1984.   I had wasted a good many man-hours of work.  This is painful, but it is also in the nature of research, that in order to move forward toward better research methods, you have to admit that what you have done in the past was not perfectly done.   It is not a good policy to change directions in the middle of the trip, but much more of the trip is ahead of us than is behind us, and I have decided that I need to do so. 

16.  The New, New Method

              OK, here’s what I changed to.   I changed to evaluating players based on what they had done over a five-year period, weighted 1-2-6-2-1.   In other words, where a player ranks after the 1952 season will be based 8.3% on what he did in 1950, one-sixth on what he did in 1951, one-half on what he did in 1952, one-sixth on what he did in 1953, and one-twelfth (or 8.3%) on what he did in 1954.   This preserves the policy that one-half of the rating is based on that season’s performance and one-half on the surrounding seasons, but changes it so that the "surrounding seasons" includes the seasons ahead as well as the seasons behind. 

              I figure the player’s Win Shares per season based on that weighting system, and also his games played per season based on that weighting system.   I preserved the rule that a player’s Win Share value after the season cannot be less than 75% of his Win Shares in that season, so that rookies are not excessively discriminated against; however, with this forward-and-backward looking method, this rule is less necessary and has less impact.  I then adjusted the player’s value weighting, as I explained before, with some limited credit for missed playing time, so that a player who plays well but misses two months with an injury gets some compensation for that when we are asking how good a player he actually was.  

              Everything is the same, except that I changed the weighting for seasons from 1-2-3, or 1-2-3-4 for veteran players, to 1-2-6-2-1.   It is clear to me, having redone the rankings, that this adjustment is appropriate, and that it does lead to better lists of the top players for seasons from the past. 

              We’ll re-do 1948 for illustration.  In 1948 Yogi Berra hit .305 and drove in 98 runs.   At the time, we might have thought that this might be a fluke season, and we might not have granted Yogi the top of the rankings out of the concern that it could be a one-season fluke.   But knowing what we now know, it is apparent that that season was in no way a fluke. 

17.  Yogi and Campanella

              In 1948 Yogi Berra and Roy Campanella emerged as the best catchers in baseball.   They were not the best by default, like Phil Masi and Bill Salkeld; they were tremendous players.   Yogi won the American League’s Most Valuable Player Award in 1951, 1954 and 1955, and Campanella won the NL Award in 1951, 1953 and 1955.   Campanella was one of the greatest defensive catchers of all time and had better hitting numbers than Yogi in his best seasons, but in our system Yogi rates a hair ahead of Campanella every year because:

              a) Yogi was much more consistent with the bat than Campanella was, having good years every year while Campanella fought injuries every other year, and

              b) some of Campanella’s advantage in the numbers is due to the parks they played in.   Ebbetts Field was a better park for a hitter than Yankee Stadium.  In 1953 the Park Run Index in Yankee Stadium was 78; in Ebbetts Field it was 107.   The difference wasn’t usually that large, but there were more runs to work with in Ebbetts Field, thus a lower ratio of actual value to runs created. 

              Yogi was the best catcher in baseball for more than a decade, but choosing between Yogi and Campy, there wasn’t a bad choice.   They were both great players, both deserving of a #1 ranking:

              One thing I love about doing this is watching the players rotate onto and off of the lists.   Yogi gets onto the list in 1948 with Walker Cooper and Jim Hegan; gradually Hegan and Cooper are replaced by Seminick and Sherm Lollar, gradually Seminick is replaced by Ed Bailey and Del Crandall and Gus Triandos.   The list never or very rarely changes completely; it carries forward names for a few years or occasionally for a long time, so that you could easily make a shared-the-list chain connecting Mike Grady to Gus Triandos.   That is something we love about baseball which is impossible to explain to a person who isn’t a fan, watching the generations share the field and then replace one another.  

18.  Fourth Stage Explanation of the Process

              Yogi Berra by 1960 was a shadow of his former self, his defensive skills mostly gone and his still-potent bat limited to a smaller number of games, but he was still the #1 catcher in baseball because nobody else had yet arisen to take his place.   People who didn’t know baseball and didn’t care about baseball still knew Yogi Berra—and not merely for the malapropisms which later dominated his image, but as a complex and lovable figure.  He had transcended baseball to become part of the culture. 

              1960 was the last year I have him rated #1, and that may be one year too long, but it is a fortunate time because we need to take a break there and explain about expansion.  

              When there were 16 teams we credited the top catchers with a total of 24 points—10 for first place, 7 for second, 4 for third, 2 for fourth, 1 for fifth.   That’s 1.50 points per team.  In order to be fair to the players of later generations we need to maintain that average of 1.50 points per team, while continuing to recognize one player for each three teams, rounded down, and giving no player more than 10 points for his rank in a season. 

              In 1961, when there are 18 teams, that means 27 points.   We will award points for 1961 as 10-7-4-3-2-1; that is, 10 points for ranking first, 7 for ranking 2^nd, 4 for ranking 3^rd, 3 for ranking 4^th, 2 for 5^th, and 1 for 6^th.   It’s a total of 27 points, or 1.500 per team. 

              From 1962 to 1968, when there were 20 teams, that means 30 points.  We will award points in that era as 10-8-5-4-2-1 (10 for first, 8 for 2^nd, 5 for 3^rd, 4 for 4^th, 2 for 5^th, 1 for 6^th.)

              From 1969 to 1975, when there were 24 teams, that means 36 points.  We will award points in that era as 10-8-6-4-3-2-1-1 (10 for 1^st, 8 for 2^nd, 6 for 3^rd, 4 for 4^th, 3 for 5^th, 2 for 6^th, 1 for 7^th, 1 for 8^th.)

              From 1976 to 1992, when there were 26 teams, that means 39 points.  We will award points in that era as 10-8-6-5-4-3-2-1-1 (10 for 1^st, 8 for 2^nd, 6 for 3^rd, 5 for 4^th, 3 for 5^th, 2 for 6^th, 1 for 7^th, 1 for 8^th.)

              From 1993 to 1997, when there were 28 teams, that means 42 points.  We will award points in that era as 10-8-7-6-4-3-2-1-1 (10 for 1^st, 8^th for 2^nd, 7 for 3^rd, 6 for 4^th, 4 for 5^th, 3 for 6^th, 2 for 7^th, 1 for 8^th, 1 for 9^th.)

              From 1998 to the present, when there are 30 teams, that means 45 points.  We will award points in that era as 10-8-7-6-5-4-2-1-1-1 (10 for 1^st, 8 for 2^nd, 7 for 3^rd, 6 for 4^th, 5 for 5^th, 4 for 6^th, 2 for 7^th, 1 for 8^th, 1 for 9^th, 1 for 10^th.)

              Also in 1961 and from 1961 to the present, since teams began playing 162 games, we began figuring "injury credits" or replacement games based on 162 games, rather than 154.   That is, if a player had an established games played level of 140.00 in 1960, we would project his established Win Shares Level into 147 games, giving him half credit for the games that he is missing, whereas if he has the same level in 1961, we would project his established Win Shares Level into 151 games. 

19.  The Elston Howard Years

              Yogi Berra had several quite famous backups.   In this era teams carried three catchers.   Yogi’s two backups for several years were Charlie Silvera, who was probably a quality catcher but who rarely got on the field, and Ralph Houk, who was the kind of guy that teams liked to have on the roster and who rarely had to do anything as a player.   Yogi also ran off four or five young catchers who went to other teams and became top-rank catchers, among these Sherm Lollar and Gus Triandos. 

              Yogi’s last backup was Elston Howard.   Bill Dickey, who had trained Yogi to be his replacement, was still on hand and trained Elston as well, and then also the Yankees hired Jim Hegan to tutor Howard, so Howard was exceptionally well mentored by Berra, Dickey and Hegan.   Being only four years younger than Yogi, a little less than four years younger than Yogi, he didn’t get into the lineup as a regular until he was 32 years old, but when he did he was the best catcher in baseball, and he was a true number one.  He was a great player from 1961 to 1964:

              We’re listing the top six here because there are 18 teams in 1961 and 20 teams beginning in 1962.   Divide that by three, round down.    In the earlier version of these rankings, which I had prepared for posting but did not post, I had John Romano ranked as the #1 catcher of 1962.   Romano was a tremendous player for a couple of years, but I’m glad I made the switch. 

20.  The Torre-Freehan Era

              Joe Torre was in the Spud Davis-Rudy York-Ted Simmons-Mike Piazza mold as a catcher, a great hitter but not a complete defensive catcher.  For a few years Torre, McCarver and Bill Freehan were the best catchers in baseball:

              Joe Torre has now been elected to the Hall of Fame in part or in the whole based on his accomplishments as a manager.  It was always my opinion that Torre should have been selected as a player. 

              Freehan, like Elston Howard, was not a great player for a long enough period of time to be a Hall of Famer, but he was a true #1, rather than a default #1, for a few years.   

21.   The Johnny Bench Years

              Johnny Bench was the best catcher in baseball from 1969 to 1976, arguably longer than that:

The odd thing is that Johnny Bench had a very good year in 1977, hitting .275 with 31 homers, 109 RBI, but nonetheless slipped from first to fourth in the catcher rankings.  There are two reasons for this:

              1)  Bench had a poor year in 1976 after he had surgery to remove a potentially cancerous growth; the surgeons had to cut through some shoulder muscles, and it effected his play.   Prior to 1976 Bench had a wide lead over the other catchers, but the off season in 1976 gave away almost all of his lead, and left him in a more vulnerable position.

              2)  While Bench had a good year in 1977, both Ted Simmons and Carlton Fisk had much better years, both hitting well over .300 with power.

(This article was supposed to be up the first thing this morning; apparently I screwed up the posting process.   Sorry.)