- STAFF PICKS
- GIFTS + GIFT CARDS
- SELL BOOKS
- FIND A STORE
Used Trade Paper
Ships in 1 to 3 days
More copies of this ISBN
Whatever Happened to the Hall of Fameby Bill James
DRYSDALE AND PAPPAS
Don Drysdale won 209 games in his career, and lost 166.
Milt Pappas, Drysdale's contemporary, posted an almost identical won-lost record, 209-164.
Don Drysdale was selected to the Hall of Fame, by the BBWAA, in his tenth year of eligibility.
Milt Pappas received so little support (five votes from 324 ballots) that he was dropped from the ballot after one turn.
Pappas, an outspoken fellow, was very unhappy about this, and has been known to compare his own record to Drysdale's in an occasional interview, so this comparison between them is well known:
It is truly a travesty to have Don Drysdale elected to baseball's Hall of Fame. When Milt Pappas was not listed on the Hall of Fame ballot, his protests were met with howls of derision. He won only 209 games and had a winning percentage of only 560, they said. Not good enough for the Hall.
There are many distinctions which can be drawn between Drysdale and Pappas, and many similarities. Presenting first the full record.
Drysdale was more of a power pitcher than Pappas, recording 758 more strikeouts. He had better earned run averages and better ERA components (control, opposition batting average). Drysdale pitched for more championship teams and more teams that were in the pennant race. He was a better hitter, and he established a well-known record in 1968, since broken by Orel Hershiser.
Almost all of those issues have another side to them (if he pitched for better teams, shouldn't he be expected to have posted a better won-lost record?), and I'll discuss all of them later, in Chapter 31. I wanted to deal here with what I think is really the key difference between them, which is the issue of consistency versus peak performance.
One of the most dependable patterns in Hall of Fame voting, both in the BBWAA vote and from the Veterans Committee, is that players who have big seasons are much more likely to be selected than are players of equal overall accomplishment, but greater consistency. Drysdale, who had 27 points in the Black Ink Test, had many more big seasons than Pappas, who had 5.
I could give countless examples to demonstrate this. It is an overstatement, but not much of one, to say that every marginal Hall of Famer in history had some big seasons with eye-catching numbers, while every marginal player who isn't in the Hall of Fame didn't have those big seasons. I'll give you four examples:
1. Jack Chesbro and Jesse Tannehill. Chesbro won 199 games in his career, lost 131. In his best season, 1903, he threw a wild pitch on the last day of the season that cost his team the pennant, and that made him one of the most famous big-game goats of the first 25 years of this century. He was selected to the Hall of Fame in 1946.
Jesse Tannehill, who was born in the same year as Chesbro (1874), was Chesbro's teammate for much of his career, first in Pittsburgh (1899-1902) and later with the New York Yankees, then called the Highlanders (1903). His career record (197-116) is very similar to Chesbro's (199-131), but distinctly better — yet he is not in the Hall of Fame.
In fact, the four pitchers who were the Pirates' rotation when they won the National League in 1901 (Chesbro, Tannehill, Sam Leever and Deacon Phillippe) all had extremely similar career records.
Chesbro had probably the poorest career record of the four, Sam Leever the best, yet Chesbro is the only one who is in the Hall of Fame. Why?
You all know the answer. He had the big year.
2. Dazzy Vance and Lon Warneke. Dazzy Vance, a National League pitcher of the 1920s and 1930s, won 197 games in his career, lost 140. He's in the Hall of Fame.
Lon Warneke, a National League pitcher who came along a few years later and also threw very hard, won 193 games and lost only 121.
Warneke's record is a little better, but in 1955, when both pitchers were eligible, Vance drew 205 votes and was elected. Warneke didn't draw a vote. Why?
Vance had some monster years. He went 28-6 in 1924, led the league in ERA at 2.16 and struck out more men than any other two pitchers in the major leagues. He also went 22-9 in 1925, 22-10 in 1928 and in 1930, at the age of 39, led the National League in ERA.
Warneke had some good years, too — 22-6, 22-10, 20-13 — but just not quite at the same level. Big years get you in the Hall of Fame.
3. Roger Maris and Bob Allison. Bob Allison's career totals are very similar to Roger Maris's.
Neither Maris nor Allison is in the Hall of Fame yet, but Maris drew strong support, peaking at 176 to 184 votes each year from 1986 to 1988.
Bob Allison was the most feared baserunner of his time. He played center field when he first came up, played it well — yet his vote total peaked at zero. Why?
You know the answer. Maris had the big year.
4. Ron Guidry and Sandy Koufax. You may be surprised to learn that Ron Guidry's career record is comparable to Sandy Koufax's.
Koufax's ERA is a little better, but half of that difference is created by league ERAs.
Koufax, of course, flew into the Hall of Fame because he was the dominant pitcher of his time, 1963-1966. Guidry had only one year of comparable dominance, and certainly cannot expect to sweep into the Hall of Fame as Koufax did.
Another example, which I won't use in the same way because I'm not sure what the Hall of Fame voters are going to do, is Don Sutton and Steve Carlton. Don Sutton, believe it or not, has a career record that is substantially similar to Steve Carlton's.
Steve Carlton had big seasons, and because of this he found the Hall of Fame door as open as Madonna's...uh, arms. He was elected on the first ballot when he became eligible in 1993. Sutton, who plodded along at an annual pace of 17 or 18 wins, will still provoke an argument despite his overwhelming credentials. Carlton scored 69 points on the Black Ink Test; Sutton scored at 8.
Bob Carroll, whose writing I enjoy very much, commented on this phenomenon in his 1985 article for the National Pastime ("For the Hall of Fame: Twelve Good Men"). "Consistency may be the hobgoblin of little minds," wrote Carroll, "but it can also make certain ballplayers nigh unto invisible. Indian Bob Johnson never had one of those super seasons that make everyone sit up and whistle. While phenoms came, collected their MVP trophies, and faded, he just kept plodding along hitting .300, with a couple dozen homers and a hundred ribbies year after year...like a guy punching a time clock."
His specific point has merit. Johnson's career numbers, in the context of their time, are probably better than Hack Wilson's, but Wilson had the big year. Carroll, at least, is rational on the subject; some guys get really carried away with it.
Anyway, to state first the argument that this is an injustice...think of the player's career record as if it was a season's record. Suppose that Ken Griffey, Jr., winds up this season with 37 homers, 118 RBI and a .326 average, but Juan Gonzalez posts even better numbers. Some people will argue that Griffey should win the MVP Award because he's a better defensive outfielder, some people will argue that Griffey should win because the Mariners had a better year, somebody might argue that he should win because he's a better baserunner and hit better in the clutch. But would anybody argue that he should be given extra credit because he hit .437 in July? It's obvious, isn't it, that if he has better numbers in July but poorer numbers overall, he must have done worse some other time?
The seasons of a career, we might argue, are like the pieces of a season — individually interesting, but not fundamentally relevant to value.
Now let's look at it from the other side. You know the old saying about a statistician...if you have one foot in a block of ice and the other in a fire, a statistician will tell you that on average you're comfortable. This problem is like that: It is dangerous to ignore fluctuations in performance, and assume that the aggregate total reflects the impact of the elements.
But in this case it is held against the player if he stays at a comfortable temperature. The players who put one foot in a block of ice and the other in the fire get extra credit for it. Is that fair?
Because pitchers are less consistent than hitters, this is a more common problem in evaluating pitchers than it is in evaluating hitters.
The analogy that seasons are to careers as months are to seasons, it seems to me, has one major flaw, which is that it doesn't account for pennant races. A pennant is a real thing, an object in itself; if you win it, it's forever. If a pitcher goes 7-0 in June and puts his team into first place, that's just June; if he then goes 1-5 in July and the team slips to fourth place, the end result is pretty much the same as if the hot streak had never existed.
This is not true of seasons. If a pitcher goes 24-5 in one year and leads his team to a World Championship, that flag is going to fly forever. If he goes 5-17 the next year, they don't take the flag down.
On the other hand, we cannot assume that the pitcher who has a big year and then a bad year has had a positive impact on his team's chances in the pennant race. Bill Singer went 20-12 with the Dodgers in 1969 and 20-14 with California in 1973, but in between went 10-17 and 6-16. He never played with a championship team. If he had just been a nice, consistent 14-14 every year, his teams would have won at least one pennant and possibly two. The 1969 and 1973 teams were so weak that they weren't going to win no matter what Singer did, but the 1971 Dodgers, for whom he went 10-17 with a 4.17 ERA, missed the National League West championship by one game. There's no question but that a decent year by Singer would have put them over the top. The 1972 team, for whom he went 6-16, missed by a wider margin, but with a staff led by Claude Osteen (20-11) and Don Sutton (19-9), it's a safe statement that they could have won if they had gotten better performance out of the back end of their starting rotation.
So which is better: a consistent starting pitcher like Milt Pappas, or a hot-and-cold pitcher who has some big seasons, like Drysdale?
I designed a simple study to test this issue: Is there a difference between the two types of career profiles in terms of their effect on the pennant chances of their teams? Suppose that you had an average team, as good as the next team, might win the pennant once in a while by dumb luck but not very often. Suppose that you added to that team a Milt Pappas/Don Sutton type pitcher — consistently good over a long period of time, but never brilliant or overpowering. This pitcher has very few twenty-win seasons or seasons as a Cy Young candidate. How many pennants would he win for them?
Suppose that you took him away and added a pitcher of the exact same overall quality, but with a different profile — a shorter career, a slower beginning, but a more brilliant middle phase. We'll call this a Don Drysdale/Steve Carlton type of pitcher. How many pennants would he win for them? Would there be a difference between the two? Would the Drysdale/Carlton type pitcher push his team to more championships when he had the Cy Young-type seasons? Or would the steady, consistent Pappas/Sutton type pitcher help his team to just as many championships, or even more, but without your necessarily being able to identify which seasons those were?
Well, as you probably have guessed if you know me, I wrote a computer program to do exactly that. To begin with, I wrote a program to represent a .500 team playing season after season, 162 games per year. As I'm sure you know, a .500 team winning and losing games at random for a year won't often go 81-81, nor even necessarily stay close to 81-81; they will routinely win anywhere from 71 to 91 games, and occasionally will win (or lose) as many as a hundred or even more, just by luck.
What I wanted to see here is how often they would win a "pennant," represented in the initial study as winning 93 games or more. I ran that team, the "control" team, through 3,000 simulated seasons.
Then I added to the team first a pitcher representing a Milt Pappas/Don Sutton type career. This pitcher pitched for 20 years, started out as about a .500 pitcher (actually .470) and rose gradually to a peak of .655, then descended back to .490. Then he retired, and, this being a computer, started over again, again and again for 3,000 simulated seasons. He had an overall expected winning percentage of .584.
In the third phase, I added to the team a pitcher with a Drysdale/Koufax/Carlton type of career — shorter, but more brilliant in the middle. This pitcher started out as a .400 pitcher, rose to a peak of .700, then faded back to .500, posted an expected winning percentage of about .380 in his 16th and final season, and retired. This team then was without a key pitcher for four simulated seasons, and in the 21st season they started over, with the pitcher back at .400. He also had an overall expected winning percentage of .584, with the same number of expected decisions in his career.
A second difference between the two pitchers, in addition to the fact that the Pappas/Sutton type pitcher reached a peak of .655 and the Drysdale/Carlton type pitcher reached a peak of .700, was that the Drysdale/Carlton pitcher had more decisions per season at the peak of his career, so that in a typical career he would have as many decisions as the Pappas/Sutton type, but would get them over with more quickly. The Pappas/Sutton type pitcher would start out with 6 to 18 decisions in his first year (this also randomly determined), then would increase his workload in the middle of his career, winding up with somewhere around 30 decisions in his peak years. The Drysdale/Carlton type pitcher would start out in a similar range — 7 to 17 decisions as a rookie — then would increase until, at his peak, he would have 30 to 39 decisions per season, as Drysdale and Carlton did.
In a reasonably typical 20-year cycle, the two pitchers would produce won-lost records like this.
The aggregate totals, as you can see, are the same. The first pitcher probably wouldn't be a Hall of Famer; the second pitcher certainly would. In a typical 20-year cycle or "career," the Pappas/Sutton-type pitcher might or might not win 20 games once. In the 3,000-year simulation, 150 cycles, this pitcher won twenty games in a season 73 times, or 49 times per career. This is almost the same as the actual data for Pappas and Sutton: Sutton won twenty games once, Pappas didn't. The other pitcher, the Drysdale/Carlton model, won twenty games in a season 600 times in 150 cycles, or 4.00 times per career — again consistent with the actual data for those two pitchers (Drysdale won twenty games twice, Carlton six times.)
Now, I really had no idea what this study would find. What I was interested in was whether one pitcher's team would win the pennant more often than the other pitcher's team. I would probably have guessed, had I made a guess, that there would be no real difference, that the Pappas/Sutton pitcher would yield as many pennants in a career as the Drysdale/Carlton type.
The result was almost shocking. In 3,000 simulated seasons, the .500 team with no pitcher added won its pennant 88 times, or 2.9 percent of the seasons.
The team with the Pappas/Sutton type pitcher increased this to 159 pennants, or 5.3 percent of the seasons.
The team with the Drysdale/Carlton-type pitcher increased it to 218 pennants, or 7.3 percent of the seasons. Sixteen of these pennants were accounted for in "out years" after the pitcher was retired; while he was active, his team won the pennant 202 times in 2,400 seasons, or 8.4 percent.
Let's put that in chart form.
The incremental benefit of the Pappas/Sutton-type pitcher was 71 pennants in 3,000 years. Even though his overall record was the same, and the overall record of his team was the same, the incremental benefit of the Drysdale/Carlton-type pitcher was 130 pennants in 2,400 years.
I decided to repeat the study, with two changes. First, because the number of pennants involved is small enough to be subject to random fluctuations, I decided to extend the study to 10,000 simulated seasons, rather than stopping at 3,000. Second, I put in a randomizer so that a team would not automatically win the pennant if they won 93 games, but rather might win it with anywhere from 90 to 96 wins. Less than 90 wins, no pennant; above 96, automatic pennant. This increased the number of "pennant" seasons.
The data from the second study is substantially the same as the first study — but even more dramatic.
The incremental value of the Pappas/Sutton-type pitcher was 231 pennants in 10,000 years. The incremental value of the Drysdale/Carlton-type pitcher was 420 pennants in 8,000 years — a huge difference.
I modified the assumptions, and did a third version of the study. I made five changes for the third version:
1. Rather than assuming that all the teams would be .500 teams when the key pitcher wasn't on the mound, I randomized the quality of the team. In some years, the win expectation of the team without the key pitcher might be as high as .610; in other years, it might be as low as .410.
2. I also improved the average quality of the team, from .500 to .510.
3. I reduced the quality of the "added" pitchers from an expected winning percentage of .584 to an expected winning percentage of .560.
4. I made the contrast between the two pitchers less sharp. The Pappas/Sutton-type pitcher was increased from 30 decisions in his peak years to 32; the Drysdale/Carlton type was decreased from a maximum of 39 decisions to 38, and also the quality difference between them in their best years was reduced very slightly.
5. I gave the Drysdale/Carlton-type pitcher a 17th season, and gave both pitchers a few more career decisions. I changed the expected career won-lost record for the prototype from 218-155 to 240-189.
These changes were intended to improve the realism of the study, make the conditions studied more like actual teams. The effect of the changes was to make it much, much more likely that the team could win the pennant without the help of this pitcher, and thus much more likely for the "Bill Singer problem" to occur, wherein the key pitcher might actually prevent the team from winning the pennant if he had an off season.
In the third version of the study, the difference between the two pitchers in terms of pennant impact was sharply reduced.
In this study, which does replicate real-life conditions better than the others, the incremental value of the Pappas/Sutton type-pitcher is 180 pennants in 10,000 years; the incremental value of the Drysdale/Carlton type is 203 pennants. The advantage decreases, but it doesn't go away.
So it is very clearly true, based on these studies, that big years mean something. The historic bias in favor of pitchers who have big seasons appears to be justified — and Pappas' argument that he should be treated equally with Drysdale, because his record is almost the same, should be rejected.
Copyright © 1994, 1995 by Bill James
What Our Readers Are Saying
Other books you might like