22.  A Hall of Fame Standard

                For several years I have used 100 more Win Shares than Loss Shares as a standard dividing the Hall of Famers from the non-Hall of Famers.    I will note that this standard also works here; 100 more Good Games than Poor Games very nicely divides the Hall of Famers from the pretenders:

                Everybody above the +100 line is in the Hall of Fame except for Clemens (steroid issue), Schilling (183; just one year on the ballot), Mussina (151; just two years on the ballot), Kevin Brown (141), David Cone (140), Messersmith (105), Tiant (105), Jack Morris (101) and pitchers who are not yet on the ballot.   Below 100, nobody is in the Hall of Fame except two pitchers for whom we don’t have full-career data, but who certainly would be over 100 if we had complete career data.

23.  Run Support and the Cy Young Award

                How often does it happen that the pitcher who leads his league in offensive support will win the Cy Young Award?

                It’s happened 7 times in history. . ..well, probably less than that, but let me explain.   For the purposes of this study, I considered a pitcher to be eligible to lead the league in run support if he made 28 starts.    Normally, figuring things like who led the league in offensive support, we use 20 starts to mark eligibility, or even 15, or sometimes 25.

                But a) if you consider pitchers eligible at 20 starts, then someone making 20 to 27 starts will almost always lead the league, since the random effects are much stronger at 20 starts than at 28, and b) pitchers making 20 to 27 starts never win the Cy Young Award.    This, then, would guarantee that the count of pitchers leading the league in offensive support and winning the Cy Young Award would be close to zero, and it would also ensure that the second count which is the punch line of this bit would also be close to zero, so we wouldn’t learn anything from doing that.   By using 28 starts we exclude from the study the 1981 and 1994 strike-shortened seasons, but that’s OK; I don’t care much about them.

                The 7 pitchers who have led their leagues in offensive support and also won the Cy Young Award are:

                1)  Early Wynn, 1959 (5.19 Runs Per Game.)   Early Wynn led the American League in offensive support average, but not the majors; the Cy Young Award in 1959 went to one major league pitcher, not one pitcher in each league.    Wynn was 39 years old and fat, but a really good hitter; he hit .244 that year with two homers, an OPS over .700.

                2)  Vern Law, 1960 (5.18).  Like Wynn, Law led his league in offensive support, but not the majors, and won the Cy Young Award for the majors.   Law, again, was a very good hitter, although he didn’t have a big year with the bat in 1960.   But he hit over .200 every year from 1961 to 1966, and hit .300 twice in that span.  The major league leader in offensive support in 1960 was:  Early Wynn.

                3)  Whitey Ford, 1961 (5.67). Led the majors.

                4)  Don Drysdale, 1962 (5.66).   Led the majors.   Drysdale, of course, was a power hitter, hit 7 homers in a season twice in his career, hit 5 homers in 1961, but didn’t hit any in 1962, when he averaged .198.   But the ’62 Dodgers had a tremendous offense.

                Interrupting the narrative for a moment. . .four straight years (1959-1962), the pitcher who won the Cy Young Award has led his league in offensive support.     It has happened only three times since then.

                Sandy Koufax started winning the award in ’63.   Koufax was the first pitcher of the Cy Young era who was so good that he made offensive support irrelevant.   Koufax led the league in everything. .. .strikeouts, ERA, starts, innings, wins, winning percentage.   He didn’t need the help of a good offense to dominate the Cy Young voting.   His offensive support in his three Cy Young seasons varied from 4.02 runs per start to 4.30.

                5)  Denny McLain, 1968 American League (5.05). 

                6)  Roger Clemens, 1986 American League (6.09).

                7)  Randy Johnson, 2002 National League (5.54).  

                OK, seven pitchers have led their league in Offensive Support and also won the Cy Young Award, but how many pitchers would you guess have led their league in fewest opposition runs per start, and won the award?    More than 7, or fewer?

                Seven pitchers have led the league in offensive support and won the Cy Young Award—but 37 pitchers have led the league in lowest runs per game for the opposition, and won the award.   Of the 7 pitchers who benefitted from the most offensive support and won the Cy Young Award, three also led the league in fewest runs allowed per game:

                In Vern Law’s starts in 1960 the Pirates outscored their opposition 176 to 114, or 5.18 to 3.35, and won 26 of the 34 games, leading the league in both categories.   You may note that this is almost the same ratio that Drysdale had in 1964 (171 to 116), when Drysdale’s teams lost 21 of his 40 starts. 

                In Whitey Ford’s starts in 1961 the Yankees outscored their opponents 221 to 124, or 5.67 runs per game to 3.18, and the Yankees went 34-5 in those games.

                In Randy Johnson’s starts in 2002 the Diamondbacks scored 194 runs and allowed 99, or 5.54 to 2.83.   The Diamondbacks went 29-6 in those games.  

                So only four pitchers in history have won the Cy Young Award while leading the league in offensive support without also leading the league in fewest runs allowed per start.

24.   Ferguson Jenkins, 1974

                Ferguson Jenkins of the Texas Rangers in 1974 made 23 starts against teams with .500 or better records, and was 18-3 against them.  I mentioned earlier that Gaylord Perry in ’72 also won 18 games against teams with .500 or better records, but he and Jenkins are the only pitchers who did, and 18-3 is quite a bit better than 18-11   It is easily the best record in my data against opponents with .500 or better records, and it doesn’t get less impressive if you look more carefully at the record against individual opponents.   The Rangers were in a division pennant race with the Oakland A’s, eventual World Champions; Jenkins started 5 times against them and was 5-0 with a 0.60 ERA.  The Rangers were second in the division, the Twins third, and Jenkins was 5-1 against the Twins. . .10-1 against the two key division opponents.

                Ferguson Jenkins and Catfish Hunter in 1974 were both 25-12 with exceptional control and very good ERAs; think what you want to about it, but given the assumptions of that era, it was clear and certain that either Jenkins or Catfish would win the Cy Young Award;  both pitchers 25-12, each pitcher made 41starts.    Jenkins ERA was 2.82, Hunter 2.49; Catfish walked only 46 men in 318 innings, but Jenkins walked only 45 in 328 innings, and whupped Catfish in strikeouts, 225 to 143.  I would guess the three main reasons that Catfish won the award, in order of importance, were:

                1)  Catfish had been a Cy Young contender for several seasons but had not won the Award, and there was a certain built-up sympathy for him, a feeling that it was his turn,

                2)  Catfish’ ERA was a little lower, and

                3)  The A’s won the division.

                Having said that, I suspect that Jenkins would have won the award had the Cy Young voters known that Jenkins was 18-3 against first-division teams; again, just a guess, but it was a very close vote, and I suspect that that piece of information would have been enough to tip the scales, had it been part of the debate. 

                As to who deserved the award. . .Gaylord Perry.  Really interesting five-man contest. Jenkins’ 18-3 won-lost record against first-division teams is tremendously impressive, but

                a)  you have to beat the bad teams, too, which Jenkins failed to do, going 7-9 against them, and

                b)  the nugget has the weakness of won-lost records generally:  that it represents not merely how the pitcher pitched, but how the team played behind him.

                One tends to assume that the park favored Hunter more than Jenkins, thus that the Park Adjustment would favor Jenkins rather than Hunter, but actually it doesn’t; Arlington Stadium in that season had a Park Factor of 91, whereas the Coliseum was at 97.  If you factor in the park, Catfish moves further ahead in the ERA comparison.  

                Back to Gaylord.   The American League that year had about seven pitchers who had what we might consider legitimate Cy Young seasons--five that we’ll compare, but a couple that we won’t even get to.   Ferguson Jenkins was 23 runs better than an average pitcher, park adjusted; Catfish was 39 runs better—and Gaylord was 42 better, with a 2.51 ERA in a park that was tougher than either Jenkins’ or Hunters’.   Gaylord made "only" 37 starts, four fewer than the two top contenders, but whereas Hunter was 27-11 in Good Starts/Bad Starts and Jenkins was 29-11, Gaylord was 29-6.  In margin above .500 (5.00), Jenkins is +67, Hunter +68, and Gaylord is +95.

                In Baseball Reference-WAR Gaylord shows not only as the deserving Cy Young Winner, but actually as the league’s legitimate MVP, just ahead of Luis Tiant.  Tiant received no first-place votes for Cy Young but was 11^th in MVP voting.    Nolan Ryan won 22 games and struck out 367 batters.  Gaylord had a 15-game winning streak.  Jenkins, Hunter, Gaylord, El Tiante and Nolan Ryan. . ..the Cy Young Vote has them Hunter-Jenkins-Ryan-Perry-Tiant; the MVP vote has them Jenkins-Hunter-Tiant-Ryan-Perry, Baseball-Reference WAR has them Perry-Tiant-Jenkins-Hunter-Ryan, and my method has them Perry-Ryan-Tiant-Hunter-Jenkins.   

                Should mention. . .Bob Purkey in 1962 (23-5) was 16-2 against .500 or better teams, numerous other pitchers also won 16 games against .500 or better teams, and two other pitchers won 17 against .500 or better teams, both of them in the National League in 1969: Tom Seaver, 17-6, and Phil Niekro, 17-10.

                Two pitchers in my data won 20 games against sub-.500 opponents:  Don Newcombe in 1956, and Bob Welch in 1990.

25.  Nolan Ryan and the Big Unit

Nolan Ryan in 1973 struck out 254 batters against .500 or better opposition; Randy Johnson in 1999 struck out 254 against sub-.500 opposition.   Actually. . .and I don’t know what to make of this, except it is interesting. . .the top four strikeout totals ever against .500 or better teams are Nolan Ryan in 1973, Nolan Ryan in 1976, Nolan Ryan in 1974 and Nolan Ryan in 1972, whereas the top four totals against sub-.500 teams are Randy Johnson in 1999, Randy Johnson in 2002, Randy Johnson in 2001, and Randy Johnson in 1997.   The most strikeouts Johnson ever had against winning teams was 174, in 1991, and the most strikeouts Ryan ever had against losing teams was 174, in 1978.

26.  Pedro vs. Koufax, or

29 vs. 40

                It is almost impossible for 29 to be better than 40.   Let us suppose we are talking about the impact of 29 home runs, versus the impact of some greater number of home runs.   If the greater number of home runs is 30, then it is not at all unlikely that the 29 home runs had more impact on the games in which they were hit it in than the 30; I might guess that there would be a 45 to 47% chance that the 29 would in fact be more valuable than the 30.    29 against 31, 32. . .sure; it is very possible that the 29 have more impact than the 31 or 32.

                When you get to about 35, 36 home runs in the "greater number" category, this becomes substantially less likely.     If you compared a group of players who hit 29 home runs each in a season and a group of players who hit 36 home runs in a season, and you asked "how many runs did these players drive in with their 29 or 36 home runs?", you WOULD find some cases in which the player who hit 29 home runs drove in more runs (with his homers) than the player who hit 36—some cases.    But you would find a great many more cases in which the player with 36 homers drove in more runs (with the homers) than the player who hit 29.  Probably 85, 90% of the time, the player with 36 homers is going to have more RBI on homers.   If you look at impact on the games played in, same thing; the guy with 36 homers is USUALLY going to have had much more impact on the games than the guy with 29.

                29 against 40. . .well, at this point it’s very hard for the greater number not to win.   It is very hard for 29 to beat 40.   It IS possible; it WOULD sometimes happen, one to three percent of the time, perhaps.    29 would beat 40 sometimes if you took the highest-impact 29 you could find, and compared it to random, undescribed "40s".    One of the guys who hit 40 homers would only have driven in 48 runs with the 40 homers, while one of the guys who hit 29 homers would have driven in 55 runs with the 29 homers. 

                But if you pre-select the 40s, so that you’re only dealing with the high-impact 40s, the BEST of the 40-homer seasons. . .then it just becomes totally impossible for the 29s to beat the 40s.   If you look at twenty guys who hit 40 homers in a season, one of them will have driven in 68, 70 runs with the 40 homers, and there is just no way that the player with 29 homers is going to be able to beat that.   29 can’t beat 40 if it’s a strong 40.   That make sense?

Comparing Sandy Koufax in the mid-1960s to Pedro Martinez in 1999-2000, Pedro was not only more dominant than Sandy was, he was far more dominant than Sandy was, in part because it was the steroid era, and the steroid era magnified the extent to which a good pitcher could dominate.    But Pedro in his best years, 1999 and 2000, made only 29 starts each season.     Koufax made 40 starts in 1963, 41 in 1965, 41 in 1966.   It is very hard for 29 starts to have more impact on a pennant race than 40.    If you compare Pedro to a randomly selected 40-start pitcher—Joe Coleman in 1973, 40 starts with a 3.43 ERA—sure, Pedro had more impact in 29 starts than Coleman did in 40.  

                But Koufax, while he wasn’t dominating at the same level as Pedro or Randy Johnson, was still the best of the 40-start pitchers.  So Koufax almost has to rank ahead of Pedro in terms of his impact on the teams, his impact on the pennant races, I believe, because 29 just can’t beat 40 if it’s a strong 40.   You just can’t have as much impact on the pennant race, making 29 starts a season, as Sandy Koufax did making 40 or 41 starts a season.   It just almost can’t be done. 

                In terms of the quality of his starts, Pedro is off the charts.   Pedro’s starts in 2000 had an average level of 9.07, easily the highest average in the data.    There are 3,344 pitcher/seasons of 20 or more starts in my data.   The highest average game levels are 1. Pedro Martinez, 2000 (9.07), 2. Pedro Martinez, 1999 (8.66), 3. Bob Gibson, 1968 (8.53), and 4. Pedro Martinez, 1997 (8.47).   Koufax’ highest level was 8.05 (1966), which ranks 12^th on the list.   

                The average Game Level for all starts is 5.00, but the average game level for pitchers making 20 or more starts is 5.33, with a standard deviation of 0.92.   Koufax, then, is just short of three standard deviations above the norm, while Pedro—and only Pedro—is four standard deviations above the norm.   He is a whole standard deviation ahead of Koufax, and it’s not any kind of fluke, because he is up near that level in several seasons

                But still. . .29 just can’t beat 40, if it’s a strong 40.  It can’t really happen.   So Koufax in his best seasons is still the number one starting pitcher of the expansion era.

27.  Park Effects and Cy Young Voting

                Which has a larger "improper influence" on Cy Young voting:   Offensive Support, or the advantage of pitching in a pitcher’s park?

                Being the contrarian that I am. … OK, contrarian is a nice word for it.   Being the troll that I am, I kind of wanted to argue that the park a pitcher pitches in, which lowers his ERA, actually has a larger impact on Cy Young voting than his offensive support.   In the first half of the data—that is, in the data from 1956 to 1982—you actually can make a very good argument that that is true.    From 1956 to 1982, three-fourths of Cy Young Awards (33 out of 44) went to pitchers who worked in parks with Park Factors lower than 100, and many, many of those went to pitchers in extreme pitcher’s parks.   Through 1982, Cy Young Awards going to pitchers in parks with Park Factors of 90 or below outnumbered 20 to 4 those going to pitchers in parks with Park Factors of 110 or above.

                But in 1983, interestingly enough, this suddenly ceased to be true.   At all.   Since 1983 most Cy Young Award winners have actually worked in hitter’s parks, and pitchers winning Cy Young Awards in parks above 110 outnumber those in parks below 90, five to three.

28.   Quality of Competition in a Season

                I argued earlier that the quality of competition faced is not a legitimate variable distinguishing one pitcher from another over the course of a career.   But in a season?

                Well, in a season it is closer.   John Buzhardt in 1961 went 6-18 with a 4.50 ERA; he later gave the White Sox three pretty good seasons.  In 1961 he was pitching for a terrible team, but also, the average won-lost record of the teams that he started against was .550.   That’s hard to do; you figure the best team in the league is a .600 team; you’ve got to get halfway there to face .550 opposition.   Buzhardt started 22 times against teams with winning records, 5 times against teams with losing records, started six times against the league champion Cincinnati Reds,  and never started at all against the worst team in the league (his own) or the second-worst team, the Cubs.

                Esmil Rogers of Toronto almost tied him last year (2013); Rogers opposition winning percentage was .547.  On the other end, Brad Radke in 2002 had an opposition winning percentage of .428, and Mike Garcia in 1954 of .438.  Everybody else is .447 or higher.

                The standard deviation of winning percentage in a career, 100 or more starts, is (I reported earlier) .007.   In a season (20 starts) it is .013.  The lowest opposition winning percentage by a Cy Young Winner is .469, by Johan Santana in 2004; the highest is .518, by Pedro Martinez in 2000.

                As to what that means in terms of runs. . . .Santana in 2004 made 34 starts.   If the quality of his opposition was .469, rather than .500, that gives him a "head start" of 1.054 wins, which is essentially equivalent to 11 runs, so Santana started the computations 11 runs ahead in the Cy Young contest.    Pedro, facing .518 competition in 2000, was working with a handicap of 5 runs.

                There is another little bias like that that we don’t look at but should, which is starts at home and on the road.   We tend to assume that evens out, which it does, generally.   But Nolan in 1991 made 20 starts at home, 7 on the road, and in 1978 he made 21 starts at home, 10 on the road.    Those are the two most imbalanced home/road splits in the data, followed by:

                And on the other end of the list:

                Nieves in ’87 was 14-8 despite making two-thirds of his starts on the road.    Nieves had a 6.22 ERA at home, 4.23 on the road, so obviously they were avoiding pitching him in Milwaukee, whereas Ryan was no doubt being used to try to bolster attendance.  

                Career difference of home to road starts, highest:

                Missing 13 career starts for Whitey Ford; Casey liked to use him at Yankee Stadium because it was a left-handers’ park.   Why they were using Dave Burba as an attendance draw, I don’t know.    More career starts at home than on the road: