I have a series of about seven articles here that are ready to go, but the first thing I need to do is to explain a couple of new wrinkles to my methodology.    The first is something called Team Success Percentage (duh; see title.)

            The Team Success Percentage is built on an old process that I’ve used for many years and probably explained somewhere before, although frankly I don’t know where or when.   Essentially, what we’re going to do is to assign each team a “Success Level” of 5 to 1.   A team that has a great season, wins the World Series or over-achieves relative to previous seasons, is graded a “5”.    A team that has a very disappointing season is graded a “1”.   That’s the first part of the process.   The second part of the process is to track the success of the teams associated with a particular player, and to state that as a percentage.

            To classify the success enjoyed by a particular team, we start with their won-lost records over the previous two seasons.   We use the success over the previous two seasons to establish expected winning percentage for the upcoming season, based on this formula:

            Wins in year (x-2)

            Plus 2 times Wins in year (x-1)

            Plus 162

            All of that divided by

            Wins + Losses in year (x-2)

            Plus 2 times Wins + Losses in year (x-1)

            Plus 324

            At this point the question will be asked whether we use a different figure for the teams before the 162-game schedule was adopted in 1961/62.   The answer is “no”; we use 162 and 324 regardless of how long the schedule is.   Let us take the New York Mets in 1987.   In 1985 the Mets went 98-64.   In 1986 they went 108-54.   As an expectation for 1987, then, we have:

            98

            Plus 216

            Plus 162

            Making a total of 476

            Divided by

            162

            Plus 324

            Plus 324

            Making a total of 810

            The Mets for 1987 thus had an expected winning percentage of .5877, or an expectation of 95.2 wins heading into the season. 

            The expectation for any team is thus based

            40% on their performance in the previous season,

            20% on their performance in the season before that, and

            40% on the tendency of all teams to move toward the center.

            There’s pretty good research behind the method, but we’re not going to get into that.   You can take it for whatever you think it’s worth.    The Mets in 1987 are expected to win 95.2 games if they play 162.

            They did play 162, but they won only 92 games.   That’s an underperformance of 3.2 games, or -3.2 games.

            Next we apply this scale:

            +10 or more                 Success Group 5          Highly Successful Season

            +4 to +9.9999              Success Group 4          Successful Season

            -3.9999 to +3.9999     Success Group 3          Neutral Outcome Season

            -4 to -9.9999               Success Group 2          Disappointing Season

            -10 or less                    Success Group 1          Bad Season

            The 1987 Mets, at -3.2, are in Success Group 3, the “blah” group.   It was a “blah” season.

            There is an exception to the classifications above, which is for World Series teams.   There are two rules that overrule the ones above, which are:

            1)  Any team that wins the World Series is in Group 5, since, if you win the World Series, that’s always a highly successful season, and

            2)  Any team that appears in the World Series is always at least in Group 4, no lower.

            I have used that system or something very much like it for years.   What’s new here is using that system to track the success of a player’s teams across the course of his career.   What we do is, we multiply the “success group” for each player’s team by the sum of his Win Shares and Loss Shares.   We total up these figures for the player’s career, and divide by the career total of Win Shares and Loss Shares.

            This can theoretically result in an average as low as 1.00 or as high as 5.00.  We convert this to a “Team Success Percentage”, then, by subtracting one and dividing by four.   Let’s do Rick Ferrell as a for-instance.

            Rick Ferrell was a 1930s/1940s catcher, good defensively and with a good strikeout/walk ratio, no power and didn’t hit for a great average, who was selected to the Hall of Fame in 1984 in a selection generally considered to be either a joke or a scandal, depending on how far your sense of humor extends.   Anyway, these are Ferrell’s career batting records:

            To that, let’s add the won-lost records of his teams:

            OK, he played for some good teams there and some not-so-hot teams, as you can see.   This chart adds to the chart below the “Success Group” of each team:

            Ferrell played for three teams that had highly successful seasons—the 1934 Red Sox, the 1942 St. Louis Browns, and the 1945 Washington Senators.   On the other hand, he played for five teams that had absolutely miserable seasons.   How do we sum this up?

            This chart adds to the one above the Win Shares and Loss Shares for Ferrell in each season:

            We make a weighted average of the Game Shares times the Team Success Groups, but I should warn you that the Win Shares and Loss Shares are not integers.    In 1929 that’s actually 4.241 Win Shares and 4.925 Loss Shares, so when you add them together and multiply by four, you don’t get 36, but 36.7.

            Anyway, the weighted average works out to 2.85.   Subtract one and divide by four, and you have .463.  The Team Success Percentage of the teams for which Rick Ferrell played was .463.   On average, they had somewhat disappointing seasons.

            The average is intended to be .500, but it isn’t exactly.   It isn’t exactly because of

            1)  The World Series rules, which upgrade some teams but don’t downgrade any teams, and

            2)  Expansion.

            We expect a first-year expansion team to play .400 baseball, so first-year expansion teams, on average, come out about 3.00 like other teams.  But those first-year expansion teams push the rest of the league over expectations, which pushes the system off-center just a little bit.  The average “Success Group” for all teams isn’t 3.00 but 3.074, so the average “Team Success Percentage” for all players isn’t .500 but .519.

            OK, you got that?   It’s just a way of making an objective statement out of a general observation.   The general observation is that Bill Terry played mostly for very good teams.  The specific calculation is that his Team Success Percentage was .684.    The general observation is that Willie McCovey played for more good teams than bad ones.    The specific calculation is that his team success percentage was .582.  These are the Team Success Percentages for a selected list of players:

            Just checking to see whether you were awake.   I believe the .260 figure for Indian Bob Johnson is the lowest figure I have seen for a player who had a long career, but I have actually figured several players with higher percentages than Chipper Jones.   We’ll get to those in the articles that will follow, over the next several days.  

            For the sake of clarity, at no point are we arguing that the Team Success Percentage is a direct measure of the contribution of the player.   1950s third-string catcher Charlie Silvera would have an extremely high Team Success Percentage—perhaps the highest of all time—but nobody would argue that the 1950s Yankees succeeded because of Charlie Silvera.   Nonetheless, it is an accurate observation that Charlie Silvera’s teams were remarkably successful.   That’s all we’re trying to do. ..draw the objective fact that a player was (or wasn’t) on successful teams into the oeuvre of the objective analysis.

And Also

            I mentioned a long time ago that I had a couple of new wrinkles to my method.   The other one is. . .and I think I have explained this earlier, but maybe not. . .the other one is that I’m going to start using the term “Win Share Value”.

            Win Share Value is actually much closer to the “old style” Win Shares than is new Win Shares.    Win Share Value is figured as:

            Win Shares plus (Win Shares minus Loss Shares)/2.

            If your Win Shares and Loss Shares are 15-15, this makes a Win Share Value of 15:

            15 +  (15 – 15) / 2

            If your Win Shares and Loss Shares for a season are 20-10, this makes a Win Share Value of 25:

            20 +  (20 – 10) / 2 =  25

            If your Win Shares and Loss Shares for a season are 15-21, this makes a Win Share Value for the season of 12:

            15 +  (15 – 21) / 2 =  12

            By doing this, we recover the ability to make one-dimensional values that you can just add up to get the “dead weight value” of a group of players or a group of seasons, which was an important asset of the earlier system.   Also, by doing this we recover the ability to state what percentage of a player’s value is in his hitting, what percentage is in his fielding, and what percentage is in his pitching. . .another important advantage of the earlier method.    I will illustrate the use of this method in the articles that follow over the next week.