I have decided to change the way that I figure season scores for pitchers.   The old formula was:

            10 * Wins Minus 5* Losses

            + Runs Saved Versus a 5.00 ERA

            + (2 strikeouts – 3 walks) / 3

            + 3 * Saves

            There were two problems with this formula, one of which I’ll get to in a moment.   The big problem that has bothered me is that I thought the “Innings and ERA” component should be larger relative to the Won-Lost record.

            When I was a kid, 40 or 50 years ago, pitchers were evaluated essentially by their won-lost records, and everything else was peripheral.   The baseball community at that time sincerely believed that the runs scored for the pitchers on a team evened out over the course of a season—which we now realize to be false—and that some pitchers had an ability to pitch to the level of the game, winning games if need be 2-1 or 8-7.  This is either

            a) Untrue, or

            b) So difficult to prove to be true that it must be assumed to be a small factor, if such an effect exists at all.

            In the winter of 1989-1990 the Royals signed Storm Davis to a big-dollar contract as a free agent; he had gone 19-7 the previous year, but with a worse-than-league ERA and a strikeout/walk ratio of 91-68.   “We’re not looking for strikeouts or a better ERA,” they explained.   “We’re looking for wins.”

            Pitchers had been evaluated basically by won-lost records for 100 years, and I sort of assumed that it would always be that way.    In looking for a way to rank seasons based on how people would ordinarily evaluate them, I looked first at the won-lost record.

            But that always bothered me, and, after introducing the system a couple of years ago, I began to question whether, in the modern world, people still think the same way to the same extent.   Wins and Losses are still the most-often cited pitcher’s statistic, by far, but they no longer dominate the discussion the way they once did.  I decided that I would be more comfortable with a system that shifted more of the burden to innings pitched and ERA.

            The second problem was strikeouts and walks.   When I published the article last week about strong seasons (The Strong Seasons Index) one of the reader comments about it was that the Season Scores method over-valued walks.   Well. . .the method certainly does not over-value walks relative to strikeouts; in all likelihood it over-values strikeouts relative to walks.   But on the general issue, suppose that we compare Tom Seaver in 1975 (22-9 record, 2.38 ERA, 243-88 strikeout/walk ratio) with Jim Palmer in 1973 (22-9 record, 2.40 ERA, 158-113 strikeout/walk ratio).   Their won-lost records are the same; their ERAs are essentially the same.   Should Seaver be considered to have had a better year because he had more strikeouts and fewer walks?

            Yes, of course he should.  When you consider more information, you get a more reliable result.  Some people would look at it and say, “If the results are the same, what difference does it make how it was accomplished?”    But the “results”—the won-lost record and the ERA—are team results, not individual results.   Palmer had behind him one of the greatest defensive teams in the history of baseball.   Certainly his record is a tribute to his pitching, but the two pitchers should not be considered joined at the hip, because they pitched on different teams.  Even if they had pitched on the same team, they would have pitched on different days, facing different opponents and different opposition starting pitchers.

            But at the same time, when it was argued that the system over-valued walks, I couldn’t simply assume that that was false, as much as I might prefer to.   When I studied the issue, I concluded that the Season Scores method did in fact very substantially over-value strikeouts and walks.  I would put it this way:  that strikeouts and walks are intended to be like the icing on a cake—not an integral component, but a thin layer put on at the end of the process.    The points I had assigned to strikeouts and walks were already small compared to the points assigned based on other factors, but were they small enough?

            Studying the issue systematically. . .no, they were not small enough.   In some cases strikeouts and walks assumed an exaggerated importance.   I decided to fix that at the same time.

            OK, I wanted a system that scored pitchers generally in the same way as the existing system, along the same scale, but which avoided these two faults (1—over-reliance on won-lost records vs. innings and ERA, 2—over-emphasis of strikeouts and walks.)   This is the new system:

            1.  Figure a “crude leverage index” (CLI) for each pitcher by the formula

                        Innings Pitched + 3 * Saves

                        ----------------------------------

                              &nb​sp;     Innings Pitched

            2.  Figure for each pitcher his “Fractional Innings” (FIP), which is three times innings pitched.

            3.   Award Points based on the following formula:

                        ( FIP * .425 – R – ER ) * CLI

            4.   Plus 8 * Wins – 5 * Losses

            5.   Plus (2 * Strikeouts – 3 * Walks) / 15

            6.   Plus Saves

            The formula, in practice, is no more complicated than the previous formula.   We have done five things:

            1.  Essentially doubled the points awarded for Innings Pitched and ERA,

            2.  Made an allowance for unearned runs as well,

            3.  Reduced by 20% the points awarded for Wins,

            4.  Reduced by 80% the points awarded for Strikeouts and Walks,

            5.  Made the points awarded for “Saves” dependent on having a good ERA as well as the saves themselves.

            Actually, we have slightly more than doubled the points awarded for Innings Pitched and ERA.    For most pitchers, this awards about the same number of points as the previous system.  The first Cy Young Award winner was Don Newcombe in 1956; he was 27-7, pitched 268 innings with a 3.06 ERA, struck out 139 and walked 46.  He scores at 340 by the old system, and 340 by the new system.   By the old system he got 235 points for his won-lost record, 58 points for his runs saved vs. a 5.00 ERA, and 47 points for his strikeout/walk data.   By the new system he gets 181 points for his won-lost record, 150 points for his runs saved, and 9 points for his strikeouts and walks.

            Most pitchers historically come out about the same, but obviously this does change some of the scores.   Tom Seaver in 1975, by the old system, got:

• 175 points for his won-lost record (22-9)
• 82 points for his runs saved vs. a 5.00 ERA, (280.1 innings with a 2.38 ERA), and
• 74 points for his strikeouts and walks (243-88).

            A total of 331 points.  In the new system he gets:

• 131 points for his won-lost record,
• 202 points for his runs saved vs. a high ERA, but only
• 15 points for his strikeouts and walks.

            A total of 348; he actually improves a little in the new ranking, since his innings pitched and ERA are very impressive.

            But Jim Palmer in 1973, by the old system, got:

• 175 points for his won-lost record (22-9),
• 86 points for his runs saved vs. a 5.00 ERA (296.1 innings with a 2.40 ERA), and
• 3 points for having one save, but
• LOST 8 points for his strikeouts and walks (158-113). 

            A total of 256 points.   In the new system he gets:

• 131 points for his won-lost record,
• 215 points for his runs saved vs. a high ERA,
• 1 point for his save, but
• Loses 2 points for his strikeouts and walks.

            Which gives him a new total of 345.  Palmer (1973) trailed Seaver (1975) in the old system, 331 – 256.   In the new system they are almost even, 348 – 345.  

            Actually, Storm Davis doesn’t get hurt that bad in the new accounting; he drops from 160 to 158.  The people who get moved down in the new ranking are the guys with sensational strikeout/walk ratios, like Pedro Martinez in 1999, who drops from 451 to 365.

            Which, you know. . .I’m sure there’s a Pedro Martinez fan reading who is irritated by this.   365 is still a huge number; Pedro’s 1999 season, when he was 23-4 with a 2.07 ERA, still ranks as the second-best season in the American League since Guidry in 1978.  But by the old system, Pedro in 1999 ranked ahead of Guidry in 1978 (451-427).   That wasn’t right.  Guidry in 1978 had more wins (25-23), fewer losses (3-4) and pitched 60 more innings (273-213) with a better ERA (1.74 to 2.07).  All of that was being overpowered, in the previous method, by Pedro’s fantastic strikeout and walk data (313-37).

            Bucky Walters in 1939 was the NL MVP, going 27-11 with a 2.29 ERA.   A teammate, Paul Derringer, was also 25-7 with a 2.93 ERA.  Derringer had a much better strikeout/walk ratio (128-35 vs. 137-109), so in the old system we had Derringer ahead of Walters, 335-294; in the new system we have Walters ahead 385-346.    It’s a debatable point and I can see it either way, but you know, you always prefer to have the MVP in first place if you can.

            The biggest surprise of this re-scoring process is that we substantially gained accuracy as a predictor of Cy Young races.   I went into this process assuming that whatever we did to de-emphasize won-lost records would cost us in terms of predicting the Cy Young race, because the Cy Young race rests so heavily on the won-lost record.  In fact, though, we had a net gain of I think five Cy Young “matches” by revising the system. ..a large pickup, and it was actually larger than the count, because many of the Cy Young Award winners who didn’t finish first in their precincts still moved up in the race.  This chart summarizes some of the more interesting Cy Young competitions of the past:

1960 (One Race—All Best Pitchers in National League)

            The 1960 Cy Young race is interesting because there are a large number of pitchers with very comparable credentials.   In the previous system I had Lindy McDaniel, who made a couple of starts for St. Louis early in the season but then moved to the bullpen and had a career year in relief, as the #1 pitcher in the league, and Bob Friend as the #1 starting pitcher in the league, although his teammate Vern Law won the Cy Young Award.   I believe this is also the only race ever in which a starting pitcher won although another starting pitcher (Ernie Broglio) had both a better won-lost record (21-9 vs. 20-9) and a better ERA (2.75-3.08).   I had tried to find a way to get Vern Law in front of Broglio, Friend, and McDaniel all at the same time (not to mention Spahn and Drysdale), but had never been able to do it.   But in this revision—paying no attention at all to that issue—it worked out with Law in first place.

1964—One Race

            In 1964 Sandy Koufax went out with an injury in August, leaving the Cy Young race as a mad scramble eventually won by Dean Chance.   In the previous system I had Koufax as still the best pitcher, followed by Bunning and Marichal.   In this revision Chance, because of his 1.65 ERA, moves to the front of the pack.

1967 American League

            1967 was the first year that there was a Cy Young Award given in each league.    Here we have moved in the opposite direction.   Our old system agreed with the voters that Jim Lonborg was the league’s best pitcher.    The new system prefers Joel Horlen of Chicago.   The voters may have been right. . ..Fenway Park in 1967 had a park run index of 133, where Chicago was at 81.  

            Including Dean Chance here to make the point that he wasn’t a one-year wonder; he was a quality pitcher for about eight years.

1967 National League

            In the old system we had the Cy Young Winner, Mike McCormick, well down the list.   In the revision Jim Bunning is still in first, but by a margin of only one point over McCormick.

1970 American League

            Jim Perry won the 1970 American League award from among a group of pitchers with very comparable credentials.   It was a widely split vote in which the fourth-place finisher—Cuellar—actually had as many first-place votes as the winner, Jim Perry (8 each.)  We had Cuellar as the best pitcher before, Palmer now, can’t get to Perry any way—but Perry does now rank as the #1 pitcher in the American League in 1960, ten years earlier.  Palmer, McNally and Cuellar all pitched for Baltimore, illustrating a point I made earlier, which was that the Baltimore defense in that era was so good that they tended to make whoever was on the mound look something like a Cy Young pitcher.

1971 National League

            Ferguson Jenkins won the Award.   Our previous system agreed with that by a margin of 394-385; our new system disagrees, by a somewhat wider margin.

            Leonard Koppett wrote an excellent analysis of this contest in The Sporting News which helped me to organize my thinking about sabermetric issues.  I was 22 years old at the time; it was an important article to me.   Koppett argued that the voters had it wrong, and that Seaver should have won the award.  However, Koppett dismissed or diminished Park Effects, about which of course there was little or no information available at that time.   Seaver pitched in Shea Stadium, with a park factor of 89; Jenkins, in Wrigley Field, with a park factor of 122.   Also, Jenkins hit 6 homers that year and drove in 20 runs.   If you make park adjustments and give Jenkins credit for his own batting it’s a much closer contest than Koppett thought it was.

1973 American League

            In the American League in 1973, the first season of the Designated Hitter Rule, there were twelve 20-game winners—none of whom was tremendously outstanding, although Nolan Ryan struck out a major league record 383 batters and threw two no-hitters.  I had Blyleven as the best of the group in the previous system, have changed now to Jim Palmer.

1974 American League

            Catfish Hunter, with Oakland, and Ferguson Jenkins, with Texas, both went 25-12 in 1974, and Catfish won the award in a close vote.  In the old system we had Ferguson Jenkins far ahead, because of his strikeouts; now we have Catfish ahead because of his ERA.    Texas in 1974 actually had a significantly lower park run index than did Oakland. 

1976 National League

            Randy Jones of San Diego was a soft-tossing lefty who got a lot of ground balls.   He should have been called the Little Unit.   He led the National League in GDPs induced in 1975, 1976 and 1978.   In 1976 he threw 25 complete games and won the Cy Young Award.  Before, we had Jerry Koosman as the best pitcher in the league, Jones fourth.    Now we have pretty much matched the voting, which went Jones-Koosman-Sutton-Carlton.   Koosman did get seven first-place votes.

1979 American League

            In the old system we had the Cy Young winner, Mike Flanagan, ranked first.   Now we have him ranked behind Jim Kern, a reliever.   But Jim Kern had one heck of a season, and I don’t feel bad about the ranking.   Flanagan pitched 123 more innings and gave up 66 more runs—a marginal ERA of 4.83.    And this doesn’t get into the issue of leveraged innings.

1979 National League

            In 1979 no National League starter had a dominant year except J. R. Richard, who struck out 300+ batters and led the league in ERA.    The Cy Young voters gave the award to Bruce Sutter.   As the chart above demonstrates, we are unable to get to that conclusion, either with the old system or the new one, although we have moved Richard down a little bit and Sutter up a little bit.

            In defense of the voters, the top two National League starters (Joe Niekro and J. R. Richard) both pitched for Houston, and the Astrodome was by far the best pitcher’s park in the league that year.    Sutter pitched in Chicago, which was the best park for a hitter, the most difficult park for a pitcher.   Consideration of this factor would certainly narrow the gap between Richard and Sutter.

            The positioning of J. R. Richard vs. Joe Niekro, on the other hand, seems more difficult to reconcile.  They were teammates.   Richard pitched more innings with a better ERA, had almost three times as many strikeouts and fewer walks.   But Niekro finished ahead of him in the Cy Young voting, because Niekro had better offensive support and was credited with 20 wins.   The Cy Young vote went Sutter-Niekro-Richard. 

1980 American League

            We didn’t get this one before, and we don’t get it now.  

1981 American League

            In the strike-shortened 1981 season Rollie Fingers won the Cy Young Award as well as the MVP.    I included it here to demonstrate that our system very often DOES mesh relievers with starters in a way that is consistent with the Cy Young voting.

Brief Essay

            In last week’s article I observed that there was a real shortage of top-rank starting pitchers in the 1920s which was not an illusion created by higher ERAs in that era.  I will also note now that there is an equally serious shortage of high-quality starting pitchers in the early 1980s.    In the 1970s there were truly magnificent starting pitchers around like Tom Seaver, Jim Palmer, Steve Carlton and Ferguson Jenkins, plus guys like Catfish Hunter, Nolan Ryan, Luis Tiant, Gaylord Perry and Don Sutton.  In 1982 Jim Palmer and Steve Carlton, although in their late 30s, were still arguably the best starting pitchers in their leagues—as Pete Alexander and Walter Johnson had been in the 1920s—mostly because nobody had stepped forward to replace them.    In the early 1980s we had Cy Young Awards won by a couple of relievers, plus guys like Lamar Hoyt, John Denny and Pete Vuckovich, who really were not Cy Young-caliber pitchers.   This changed in the mid-1980s with the emergence of Clemens, Dwight Gooden, Orel Hershiser and Bret Saberhagen.

1984 American League

            In 1984 Willie Hernandez, like Fingers in the same league three years earlier, won not only the Cy Young Award but also the MVP.

1987 National League

            Occasionally there just is no Cy Young pitcher in a league.   The 1987 National League Voting was widely split and the award went to Bedrosian, who was a poor selection but not much worse than any other. 

National League, 1989

            Our new system matches the voters.   Mark Davis won the award.

1992 American League

            And again in 1992.   Eckersley won the Cy Young and the MVP awards—the last pitcher to win the latter.  

1996 American League

            Here again, our new system matches the Cy Young voting, whereas the Cy Young winner (Hentgen) was third in the league in the old system.

1997 National League

            This one we’ve lost.  We had it Martinez over Maddux earlier; now have Maddux ahead of Pedro.   Pedro won the award.   You can’t go wrong with either pitcher. 

1998 National League

            Tom Glavine won the NL Cy Young Award in 1998; the voting went Glavine, Hoffman, Brown, Maddux.   Our new system doesn’t get Glavine to the top, but reduces the margin between Brown and Glavine from 77 points to 17. 

2001 National League

            This was the first of two seasons in which the best two pitchers in baseball were teammates in Arizona, Randy Johnson winning both Cy Young Awards.   Both Johnson and Schilling are marked way down in the revised system, but both remain about 75 points ahead of any other major league pitcher in 2002.   

2002 American League

            Barry Zito won the American League Cy Young Award.   Our system now agrees with this, although it sharply disagreed yesterday.  

2002 National League

            The second of the Johnson/Schilling shootouts.

2003 National League

            Here again, our system now rates the reliever vs. the starters in the same way as the voters.   Gagne won the award.   This is useful, because it allows us to predict, with some accuracy, what a reliever must do to win the Cy Young Award.    He had 40 saves and a 1.80 ERA but he didn’t win; how much more did he need to do?   We can figure an answer to that question with some degree of accuracy, obviously less than 100%.  

2005 American League

            I don’t care what anybody says; giving the Cy Young Award to Bartolo Colon was a freaking joke.  But it proves that twenty wins is still a magic number to many voters.

2006 National League

            2006 was another of those seasons in which nobody in the National League was exactly the re-incarnation of Walter Johnson, but somebody had to win the award anyway.   Our system before would have given it to Smoltz.  The voters gave it to Brandon Webb, and, because of the de-emphasis on strikeouts, our system now agrees.

2008 American League

2008 National League

            I think I said in an earlier article that we had Brandon Webb last year ahead of Tim Lincecum, but this (I think) was a calculation error on my part.  In fact, Lincecum was always ahead. 

            These are the ten best pitcher’s seasons since 1950, as scored by the new system:

            I refer to the Season Scores method sometimes in various research projects.  From now on, I’ll be referring to this new system, albeit by the same name as the old system.   It just replaces or updates the old system.

The Strong Seasons Index

(Revisited)

            So, what does this new Season Scores method do to the stuff I published just last week about starting pitchers who ranked consistently among the best pitchers in their leagues?

            I assume that those of you who read my stuff know and accept that my reputation is not based on being “right”; rather, my reputation is based on my commitment to a process of developing and refining methods to analyze those issues that we debate.   Having a process to approach those questions stipulates that we don’t understand them; we don’t have the answers.   We’re working on it.

            At the same time, one does not normally have to go to the public and say “You know, that stuff I said last week?  Forget it; I’ve got a better answer now.”   It is not difficult to do that when the previous answer was given twenty years ago; it is not easy to do that when the previous answer was given just days ago.   You feel like a fool.   I feel like a fool.  Ideally, one would have a well-enough organized work process to re-evaluate the Season Scores method just before using it in a large, comprehensive study of starting pitchers, rather than just after.  

            But. . .these things happen.   OK, so how does the revision of the system affect the conclusions published last week about starting pitchers?    I re-calculated the “points” in that system—awarded for being among the best starting pitchers in the league in any season—and re-evaluated the conclusions advocated at that time:

1)  Cy Young still ranks as the most consistently effective pitcher in baseball history.    

Still true.  The leaderboard published last week was

1.  Cy Young

2.  Roger Clemens

3.  Randy Johnson

4.  Walter Johnson

5.  Greg Maddux

In the revised study it goes:

1.  Cy Young

2.  Roger Clemens

3.  Walter Johnson

4.  Greg Maddux

5.  Randy Johnson

The Big Unit, deriving some of his strength from very high strikeout totals, is hurt somewhat by the revision of season scores to de-emphasize strikeouts.  But he is still one of the greatest starting pitchers in the history of baseball.

2)  The method draws a relatively bright line of separation between Hall of Famers and non Hall of Famers.

Still true, although there was some blurring.  The “Hall of Fame point level” drops from 47 to 44, but it’s still a pretty clear separation.  

3)  Excluding active and recently retired pitchers, the only Hall of Fame-qualified pitchers who are not in are Blyleven, Guidry and Billy Pierce. 

            Still essentially true.   Blyleven and Guidry both slip backward significantly in the revision, but Blyleven remains almost 30 points over the Hall of Fame line, and Guidry 15 points over.   Neither is a close call; Billy Pierce is still just over the Hall of Fame line.

            Seven other pitchers did slip marginally over the Hall of Fame line, however.   Four of those are guys who actually are in the Hall of Fame:  Stan Coveleski, Burleigh Grimes, Bob Lemon, Vic Willis.   All of those were “high gray area” selections in last week’s study.  In the revision, they’re over the Hall of Fame line.   Lefty Gomez, who was just over the line in the previous article, still is.

            However, three other pitchers who are not in the Hall of Fame also slipped over the Hall of Fame line.   Those are Vida Blue, Carl Mays, and 19^th century pitcher Jim McCormick. 

            4)  Roy Oswalt is still WELL over the Hall of Fame line, in terms of the number of seasons as one of the best pitchers in the league and the rank of those seasons. 

            5)  Other active pitchers who are still over the historic Hall of Fame line in terms of strong seasons:  Randy Johnson, Pedro Martinez, Tom Glavine, John Smoltz, Johan Santana, Roy and Andy Pettitte.   Pettitte is, as he was before, right on the line.   The others are clearly above it. 

            6)  Recently retired pitchers who are still over the historic Hall of Fame line:  Roger Clemens, Greg Maddux, Mike Mussina, Kevin Brown, Curt Schilling, Jack Morris, David Cone, and Dwight Gooden joined now by Dave Stieb, Bret Saberhagen and Orel Hershiser.   Jack Morris—already over the line--moved up a few points.  

The poster who pointed out that the season scores method over-valued walks suggested that Saberhagen ranked as well as he did because of his control rate. .   But actually, the revision—strongly de-emphasizing strikeouts and walks—improved Saberhagen’s position in the study.   Before, he scored at 45 points, with the Hall of Fame line at 47.   Now Saberhagen is at 47 with the Hall of Fame line at 44. 

            Before, we had Saberhagen in 1985 a fraction of a point behind Ron Guidry, both pitchers scoring at 293.   The re-scoring puts Saberhagen back ahead, 290-289; that was another pickup for us on matching the Cy Young Awards.    It’s a trivial point, of course, but Saberhagen picks up that point and another one somewhere else.

            Saberhagen in 1989 was the best starting pitcher in baseball by almost 100 points, finishing 23-6 with a 2.16 ERA.   That actually ranks as a minor historic season—below the level of Guidry, Carlton, Pedro Martinez and Doc Gooden in their great seasons, but just below it.   He won the Cy Young Award in 1985.   In 1987 he was 15-3 at the All-Star break, although overuse led to a second-half collapse.   In 1994 he pitched at a Cy Young level, although Maddux was better and the strike shortened the season and cut into the numbers.   In 1991 and 1998 he just missed the list of the league’s best starting pitchers.   I’m not suggesting that he should be in the Hall of Fame with 167 career wins, but. . .there’s a lot of quality seasons there.

            A few pitchers do, of course, move well up or well down in the revision.   The one pitcher most helped by the revision is Tom Glavine, who gains 26 points, moving from 66 points to 92, from 31^st place all-time to 15^th.    Jim Palmer moved into the top spot in his league not only in 1973 (the 22-9 season that we discussed early in the article) but also in 1970, 1975, 1976, 1977 and 1982, and moved up in other seasons.   Palmer moved from 82 points to 99, and from 18^th on the all-time list to 11^th. 

            If the revision helps Glavine, it has to help Warren Spahn.  Spahn moved from 105 to 118 and from 13^th place into a tie for 6^th.    Mike Mussina, with his wonderful strikeout/walk ratios, is hurt by the revision, dropping from 113 to 99, and Pedro Martinez has the same totals; he also dropped from 113 to 99.    Schilling drops from 86 to 71, from the 16^th spot all-time to 23^rd.  

            The majority of pitchers, of course, don’t really move at all; Carl Hubbell from 79 to 80, Bob Gibson from 59 to 61, Whitey Ford from 68 to 72, Juan Marichal from 75 to 73, Addie Joss from 32 to 35, Ted Lyons from 33 to 34.   Kid Nichols stays at 97 points.    It shuffles the list, but hardly challenges us to re-think history.   Jesse Haines goes from 16 points to 19, while Rube Marquard drops from 24 to 21 in an effort to challenge Haines for the position as least-qualified Hall of Fame starting pitcher. 

            But a few of the rankings changes seem inexplicable at first.   One of the pitchers most helped by the revision is Bob Feller, who moves from 60 points to 77, and, like Jim Palmer, gains several years ranking as the best pitcher in his league.   If the revision helps Feller it also has to help Nolan Ryan, right?   Who is more like Feller than Nolan Ryan?

            But in fact Nolan Ryan is the pitcher most HURT by the revision, dropping from 70 points to 46, down to the border of “qualified for the Hall of Fame”.   De-emphasizing his strikeouts, Ryan’s career is long and notable, but actually contains not a huge number of seasons ranking among the best pitchers in the league. . .enough to put him in the Hall of Fame, but just enough.    It is unfair to describe him as a historical curiosity, but it is more accurate to describe him as a historical curiosity than as a first-rank Hall of Famer on the level of Clemens, Maddux, Randy Johnson, Christy Mathewson and Pete Alexander.

            Feller moves up because de-emphasizing strikeouts is more or less incidental to him.   De-emphasizing his walks puts the focus back on his won-lost records and runs saved, which are outstanding.    Ryan moves down because de-emphasizing his walks is more or less incidental to him.   De-emphasizing his strikeouts puts the focus back on his won-lost records, which aren’t that good.  

            In the original article David Wells scored as over the Hall of Fame line, but I expressed skepticism about this conclusion.   In the revision he dropped from 54 points to 36, from just over the Hall of Fame line to well below it.    Jim Bunning drops from 75 points to 63—still very comfortably over the Hall of Fame line.