January 14, 2004
The Class of 2004
Analyzing the Pitchers
[Note: The research for this piece, and much of the writing, was done prior to the Hall of Fame voting results being announced.]
The Baseball Writers of America's standards on what constitute a Hall of Fame pitcher are in a curious spot now, both when it comes to starters and relievers. Spoiled by a group of contemporaries who won 300 games from the mid-'60s to the mid-'80s (Tom Seaver, Steve Carlton, Gaylord Perry, Don Sutton, Nolan Ryan, Phil Niekro), the writers haven't elected a non-300-winning starter since Fergie Jenkins in 1991. That Perry, Sutton and Niekro took a combined 13 ballots to reach the Hall while Ryan waltzed in on his first ballot with the all-time highest percentage of votes is even more puzzling. Apparently what impresses the BBWAA can be summarized as "Just Wins, Baby"--which is bad news for every active pitcher this side of Roger Clemens and Greg Maddux.
Of the 59 enshrined pitchers with major-league experience, only two of them--Hoyt Wilhelm and Rollie Fingers--are in Cooperstown for what they accomplished as relievers. While the standards for starters are somewhat easy to discern (if lately a bit unrealistic), the growing number of quality relievers on the ballot, the continuous evolution of the relief role, and the paucity of standards to measure them by present some interesting challenges to voters.
If there's an area in which performance analysis has struggled mightily against mainstream baseball thought, it's in hammering home the concept that the pitcher doesn't have as much control over the outcome of ballgames--as reflected in his Won-Loss totals--or even individual at-bats--hits on balls in play--as he's generally given credit for. Good run support and good defense can make big winners of mediocre pitchers on good teams, and .500 pitchers of good hurlers on mediocre teams. As such, it's important to examine the things over which a pitcher has control and account for those he does not. Once again, the Davenport system rides to the rescue.
The Davenport system adjusts for the same factors--park effects, league environment, and era--for pitchers as it does for hitters. It also adjusts for the level of defense behind a pitcher, and converts to the currencies of Pitching Runs Above Replacement (PRAR) and Wins Above Replacement Player (WARP3). In examining these pitchers, we'll also use Pitching Runs Above Average (PRAA) as well as PRAR. Why both? They make reasonable proxies for "peak" and "career." A pitcher with many PRAA but fewer PRAR likely had a high peak and a short career, while one with the same number of PRAA but more PRAR likely had a longer career. While durability should not be confused with excellence, league average has value, as anybody who's ever watched a team's fifth starter get pummeled knows. Like the hitters, we'll see how career, peak, and weighted WARP3 totals shake down with regards to the standards set by the Hall's enshrined pitchers, with our stated goal being to add those which elevate those standards.
Fifteen pitchers appeared on the 2004 Hall of Fame ballot, seven holdovers and eight newcomers. Of the holdovers, the three relievers--Bruce Sutter (53.6%), Lee Smith (42.3%) and Rich Gossage (42.1%) had received considerably more support than the four starters--Bert Blyleven (29.2%), Tommy John (23.4%), Jack Morris (22.8%), and Fernando Valenzuela (6.3%)--leading into 2004. Of the newcomers, a man who had success on both ends, Dennis Eckersley, made it in with ease. Other starters who appeared on the ballot for the first time in 2004 were Danny Darwin, Doug Drabek, Jimmy Key, Dennis Martinez, Dave Stieb, and Bob Tewksbury. Randy Myers was the sole reliever making his ballot debut.
W L SV IP ERA ERA+ AS CY 3C HOFS HOFM Blyleven 287 250 0 4970 3.31 118 2 0 0 50.0 113.5 Darwin 171 182 32 3017 3.84 106 0 0 1 20.0 28.0 Drabek 155 134 0 2535 3.73 101 1 1 1 21.0 35.0 John 288 231 4 4710 3.34 111 4 0 0 44.0 100.0 Key 186 117 10 2592 3.51 122 4 0 2 33.0 66.0 Martinez 245 193 8 4000 3.70 106 4 0 2 37.0 67.0 Morris 254 186 0 3824 3.90 105 5 0 3 39.0 108.5 Stieb 176 137 3 2895 3.44 122 7 0 1 27.0 55.0 Tewksbury 110 102 1 1807 3.92 104 1 0 0 10.0 11.0 Valenzuela 173 153 2 2930 3.54 103 6 1 2 25.0 66.5 PRAA PRAR WARP3 PEAK WPWT PKPCT Blyleven 311 1408 135.8 45.6 90.7 33.6 Darwin 86 804 79.4 27.0 53.2 34.0 Drabek 52 634 57.5 33.0 45.3 57.4 John 78 1154 112.1 30.6 71.4 27.3 Key 195 817 87.4 37.7 62.6 43.1 Martinez 38 939 94.5 34.7 64.6 36.7 Morris 27 916 90.2 38.8 64.5 43.0 Stieb 176 823 89.2 45.4 67.3 50.9 Tewksbury 21 444 42.6 22.6 32.6 53.1 Valenzuela 33 689 68.9 36.0 52.5 52.2 --------------------------------------------------- AVG HOF SP 239 1002 97.0 44.9 70.9 46.3 AVG HOF P 234 977 94.8 43.6 69.2 46.0A quick rundown of the less obvious abbreviations: ERA+ is the ratio of the park-adjusted league ERA to that of the pitcher, with 100 being average and higher numbers better. AS is All-Star appearances, CY is Cy Young awards, and 3C are for leading the league in any of the pitcher's "triple crown" categories of wins, ERA, and strikeouts. HOFM and HOFS refer to Bill James' two Hall of Fame systems, as explained in my companion piece on the hitters. WARP3 is the era-, schedule-, and difficulty-adjusted version of WARP; PEAK is the player's five best consecutive WARP3 totals; WPWT is the average of the career and peak figures--we're double-crediting a player's best seasons here to let the cream rise to the top--and PKPCT is the percentage of a player's career value resulting from peak.
Bob Tewksbury was a soft-tossing slop artist who lasted through parts of 13 years in the majors. It's tempting, when poring over his career, to wonder if he was actually a lefty; seldom do pitchers last with so little apparently going for them besides handedness. Though he struck out only 4.0 per nine innings, Tewk got by with good control (1.5 walks per nine), decent double-play support (0.8 per nine), and keeping the ball in the park (0.7 HR per nine). He had his best years with the Cardinals in the early Nineties, a four-season run where he won 53 ballgames with a 116 ERA+ and placed third in the '92 Cy Young voting. A handy mid-rotation guy, but not worth much thought here. From the Davenport standpoint, there isn't even a remotely similar pitcher who's fooled the Hall voters.
A teammate of Tewksbury's when they were rookies with the Yankees, Doug Drabek was the ace of a Pirates staff that won three straight NL East titles to start the Nineties. He won a Cy Young in '90 for his 22-6, 2.76 ERA season. Despite a 2-5 record, he was excellent in the postseason for the Pirates, with a 2.05 ERA in 48.1 innings. Like Barry Bonds, he left the Bucs for the bucks after 1992. He signed a four-year deal with the Astros, and his career started its downward trajectory--an 18-loss season with a league-average ERA in '93, a good but strike-shortened '94, and then two years which proved that even Astro pitchers get bombed. He lasted only two more years in the bigs, with ever-diminishing returns, finally retiring when his ERA was mistaken for a Boeing model ("Ever flown a 7.29?"). From the Davenport standpoint, his good years cancel out his bad ones. But his WARP numbers are similar to those of Hall of Famer Chief Bender, who went 212-127 with a 2.46 ERA and 112 ERA+ from 1903 to 1917 (with a token inning in '25):
PRAA PRAR WARP3 PEAK WPWT PKPCT Drabek 52 634 57.5 33.0 45.3 57.4 Bender 92 550 58.6 30.1 44.4 51.4On the surface it's hard to believe that park, league, era, and defensive factors could deflate Bender's record to the point that it resemble's Drabek's, but that's the point of such an analysis. The Chief was 15 games above .500 with a 98 ERA+ over his first four years. He had a considerably better defense behind him, according to the Davenport cards; his Defense-adjusted ERA (DERA, not to be confused with Voros McCracken's Defense-Independent ERA) is 0.18 runs higher than his Normalized Run Average (insert Heston joke here), while Drabek is 0.12 runs lower. Bender loses a lot in the adjustments for era, making the two pretty similar in their overall value, at least regarding the regular season. The Chief does get a bit of extra juice for being part of three World Champions and five pennant winners and for pitching well in the World Series, 6-4 with a 2.44 ERA and 9 complete games in 10 starts. But less that separates the two than a quick perusal of their traditional stats would indicate.
Danny Darwin, a.k.a. "The Bonham Bullet" and "The Rich Man's Mike Morgan," was a useful swingman for eight teams over 21 big-league seasons. He won in double-digits seven times, but never more than 15 games, and he lost in double-digits 11 times. He's one of only eight men to win 170 games in the bigs but have a losing record for his career. Despite that losing record, he was an above-average pitcher from the vantage of ERA+ about two-thirds of the time, and he even won an ERA title in 1990. Perhaps the most surprising thing about his career is that he never made a postseason appearance, though he was a handy late-season pickup for a couple of division winners (the 1986 Astros and 1997 Giants). He's earned his right to be on the ballot, and by the Davenport measurements he's better than a few Hall of Fame pitchers, but that's as far as his case goes.
Fernando Valenzuela had one of the most memorable entries into big-league stardom. On the strength of a mysterious screwball and a mid-motion skyward gaze, he did it all for the Dodgers in strike-torn 1981: an 8-0, 0.50 ERA start, a World Championship over the Yankees, Rookie of the Year and Cy Young honors, and an extra 20,000 fans every time he pitched. But he became just another great arm sacrificed at the altar of Tommy Lasorda. Fernando averaged 266 innings a year for the next six years, and though he held up under the strain, he was merely good, rather than great. A shoulder injury prevented him from being a part of the Dodgers' World Series victory in 1988, and after that, he was never the same pitcher. He wandered the majors for a few years, enjoying a nice valedictory season with the Padres in 1996. He was still pitching in the Mexican Winter League as recently as last January, a national hero and a beloved figure in baseball history. All of the intangibles of Fernandomania can't shore up the conclusion via the Davenport numbers that he comes up considerably short not only of the Hall of Fame standards, but also the of the level of many of the other starters here.
A finesse lefty who didn't strike many batters out, Jimmy Key spent the biggest chunk of his career as the No. 2 starter for the Blue Jays behind Dave Stieb. That tandem was a big reason for the team's rise to respectability in the mid-'80s and to power in the early '90s. He spent nine years in Toronto, eight as a starter, and in that role averaged 14 wins and 204 innings a year while posting a 3.38 ERA and 123 ERA+. He left the Jays as a free agent after their 1992 championship for the bright lights and big bucks of the Yankees, and for awhile the pinstripes got their money's worth: he was 35-10 with an 141 ERA+ over his first two years in the Big Apple, and finished second to David Cone in the '94 Cy Young vote. He lost most of '95 to a torn rotator cuff, gutted through '96 a diminished pitcher, but had a strong postseason which included winning the World Series clincher. For that, in the Bronx, his name will forever be spoken in reverent tones. He had one more strong season for the Orioles before his arm finally gave out again.
From the Davenport view, Key would be a below-average choice for the Hall, but hardly without parallel. His WARP3 numbers are comparable to two Veteran's Committee selections, Jim Bunning and Eppa Rixey, and he's got the highest PRAA of the three. Bunning went 224-184 with a 3.27 ERA and 114 ERA+ from 1955-1971, while Rixey, wen 266-251 with a 3.15 ERA and a 115 ERA+ from 1912-1933.
PRAA PRAR WARP3 PEAK WPWT PKPCT Bunning 161 1020 86.8 39.3 63.1 45.3 Key 195 817 87.4 37.7 62.6 43.1 Rixey 119 927 83.1 33.3 58.2 40.1There are two dozen Hall of Fame pitchers (including the two relievers) with lower WPWT than Key, though many of them, such as Sandy Koufax and Dizzy Dean, had much higher peaks and shorter careers. This isn't to say that Key belongs--stronger candidates are on this ballot--just that he's better than some of the guys in the Hall, and his career was nothing to sneeze at.
Dave Stieb was the Blue Jays' right-hand man during Key's career, and without much argument, he's the greatest pitcher in franchise history, with 174 wins. He joined the Jays midway through their third season (1979) and became their workhorse ace, averaging 275 innings a year from 1982-1985, after which arm woes took him down to a 200-inning a year pitcher. He was a member of the Jays' '92 team that became World Champs, but back troubles finished him in August of that year, and he fizzled out with the White Sox the following year. Such near-misses seems to typify his career; his close calls with no-hitters (four times, including two consecutive starts, he took them into the ninth, but gave up hits; he finally got his no-no in 1990) are legendary. A strange but somewhat successful comeback in 1998 delayed his appearance on the ballot, but it doesn't change the fact that he's got no shot. By the Davenport numbers he's slightly below average in weighted WARP3, but he's got an excellent comp in Juan Marichal, who was 243-142 with a 2.89 ERA and a 122 ERA+ from 1960-1975:
PRAA PRAR WARP3 PEAK WPWT PKPCT Marichal 192 976 89.4 47.1 68.3 52.7 Stieb 176 823 89.2 45.4 67.3 50.9The Marichal-Stieb comparison isn't a bad one historically. Stieb spent a lot of his peak years in other pitchers' shadows (Clemens, Bret Saberhagen), just as Marichal did (Koufax, Bob Gibson), and as such, never received quite the acclaim he deserved.
Dennis Martinez, "El Presidente," saw some remarkable ups and downs in his 23-year big league career. The first Nicaraguan in major league history, Martinez was a starter on the 1979 AL champion Orioles, but by the time the O's won the World Series in 1983, his career was in the dumps. His 7-16, 5.53 ERA was so poor he was left off of the postseason roster, but he used that experience as the impetus to begin addressing his alcoholism. It didn't change his career overnight--arm troubles continued to hamper him for a few years--but Martinez turned it around when he signed with the Montreal Expos as a free agent in 1987 (he had been traded to Montreal during the '86 season). Starting at age 32, he reeled off seven straight good-to-excellent years for Les Expos, going 97-66 with a 2.96 ERA and 126 ERA+ in that span. He spent three more years as a good starter for the Cleveland Indians as they turned the tide on decades of mediocrity, and as a member of the Atlanta Braves bullpen in 1998, passed the aforementioned Marichal to become the all-time leader in victories for a Latin American-born pitcher.
From the Davenport standpoint, he's a bit below average for a Hall of Famer, But like Stieb, he's got a no-doubt Hall of Famer with comparable WARP3 totals, in this case Whitey Ford, who went 236-106 with a 2.75 ERA and 132 ERA+ from 1950-1967:
PRAA PRAR WARP3 PEAK WPWT PKPCT Ford 238 994 95.5 36.5 66.0 38.2 Martinez 38 939 94.5 34.7 64.6 36.7Obviously, there are some important differences between the two. For one thing, while their PRAR are comparable, Ford has 200 more PRAA for his career. For another, Whitey's usage patterns were greatly affected by pitching for the Yankees, who would often shut him down in September to keep him from racking up gaudy win totals (lest they have to pay him more) and to preserve his arm for the World Series, where he was the master, holding many important Series pitching records. Martinez was a pretty fair postseason pitcher (2-2, 2.47 ERA in 43 innings) in his own right, but he never had as much opportunity or success as Ford. El Presidente wouldn't shame the Hall of Fame, but again, there are stronger candidates on this ballot.
Jack Morris is perceived as one of those stronger candidates. Like fellow candidate Alan Trammell, Morris was part of the home-grown nucleus which anchored the Tigers in their fine 1984 championship season, and he was the gritty ace on two other World Champions for two other teams. He racked up some high win totals over the course of his 18 seasons, notching 20 or more victories 3 times, topping 17 eight times, and in double-digits 14 times. In addition to his high win totals, he put up some stellar performances in the postseason (7-4, 3.80), most notably a 10-inning 1-0 shutout in Game Seven of the '91 Series. Morris has acquired a larger than life reputation based primarily on that performance, and for awhile it seemed like it might carry him to the Hall of Fame. But his career ERA and ERA+ are nothing to write home about, and they especially took a hit during the last two years of his career, raising his overall ERA from 3.73 to 3.90. Supporters have dismissed the high ERAs, claiming that Black Jack "pitched to the score." But Joe Sheehan, building on work done by Greg Spira, did an exhaustive bit of research via Retrosheet, poring over Morris' career inning by inning to examine that claim. His conclusion: "I can find no pattern in when Jack Morris allowed runs. If he pitched to the score--and I don't doubt that he changed his approach--the practice didn't show up in his performance record." Morris' record is more a product of good run support than it is any special strategy.
Davenport-wise, Morris would be a below-average Hall of Famer, one who's in the same cluster as the elected Ford and Bunning, as well as candidates Martinez and Key. His closest analogue among Hall of Famers for WARP pattern is Red Faber, who went 254-213 with a 3.15 ERA and 119 ERA+ from 1914-1933, all with the White Sox:
PRAA PRAR WARP3 PEAK WPWT PKPCT Faber 209 975 90.5 39.4 65.0 43.5 Morris 27 916 90.2 38.8 64.5 43.0Like Ford to Martinez, Faber is well ahead of Morris (182 runs) in PRAA despite similar PRAR totals--Black Jack is only 27 runs above average for his career. Faber's got his World Series glory, too, going 3-1 with a 2.33 ERA in 1917, including a decisive complete game victory (interestingly, Faber was a member of the ill-fated 1919 White Sox, but he never appeared in a game, nor was he implicated in the scandal which followed). Morris's weighted WARP total is almost identical to Martinez, though his higher peak changes the comparison somewhat. Still, he looks garden-variety among this bunch, yet another sign that he may not be as special as his supporters like to believe. Get back, Jack.
A decent but sub-.500 finesse pitcher for nine years with the Indians and White Sox, Tommy John enjoyed modest success as a Dodger, lifting his career record to 124-106 before an elbow injury ended his career prematurely in 1974... Except it didn't. At the hands of Dr. Frank Jobe, John underwent an unprecedented reconstructive surgery and then a similarly unprecedented rehabilitation. Against long odds, he returned to the Dodger rotation in 1976 and then the following year began a stretch which saw him win 80 games over four seasons--the first two for the Dodgers, the latter two for the Yankees--and appear in three World Series over five (alas, never on the winning side). Within that streak he was a three-time All Star and the Cy Young runner-up in both leagues, and after that he still had eight seasons left as a league-average pitcher. Already reliant on groundballs rather than strikeouts, he was seen as the prototype of a certain breed of successful lefties, which Bill James called "The Tommy John family of pitchers." In total, John pitched fourteen seasons AFTER the surgery, retiring at age 46--long enough to put him in some very select company. He finished with 288 wins (24th all time) and 4710 innings (18th), with virtually everybody above him either in the Hall or headed there. Had it not been for the year and a half he missed, he might have achieved the magic 300 wins which guarantees enshrinement.
The Davenport system doesn't deal with ifs, but does give us an interesting view of John. His weighted WARP total is above average for a Hall of Fame pitcher, thanks to a career total that would place him in the top 20. But his peak is extremely low; only five Hall of Fame pitchers are lower, and even the two relievers are slightly ahead of him in that department. In fact, John has the lowest percentage of value from peak of any of the 74 pitchers in my spreadsheet (the next five: Steve Carlton, Don Sutton, Warren Spahn, Dennis Eckersley, Nolan Ryan). Looking over his year-by-year stats, John had a lot of seasons, such as his 17-win Dodgers campaign in 1978, which were perceived as excellent but which come out only three to five wins above replacement. Only once was he more than seven wins above replacement. Maintaining that level for five seasons isn't particularly noteworthy among this crowd: 47 Hall of Famers and five candidates can claim that. But among that group (a PEAK of 35.0 or less), he's got no Hall of Famer besides Hoyt Wilhelm who's anywhere near him in career WARP.
John is the strangest specimen we've encountered in this study, and my first impulse (and my own personal preference) would be to say, "the system says 'IN'." But looking again, I'm not so sure. He's only 78 runs above average for his career--exactly three per year, not terribly impressive. Only four Hall of Famers are that low; two of them, Rube Marquard and Burleigh Grimes, are actually below average! I don't think he raises the Hall's standards enough to justify a vote.
Which brings us finally to Bert Blyleven, the stathead's choice among Hall-eligible starters, and quite possibly the best player not in the Hall of Fame. Blyleven was a power pitcher with a devastating curveball (or two) who reached the majors at age 19 with the Twins. Though they won the AL West in 1970, Blyleven spent most of his six and a half seasons in Minnesota toiling for mediocre clubs, racking up innings and decisions galore: 16-15, 17-17, 20-17, etc. Through that stretch he was also posting ERAs 25-50 percent better than league average and striking out about 230 guys a year. Contract issues hastened his exit, he was traded to Texas, and a year and a half later, arrived in Pittsburgh. For the Pirates he remained a front-line starter, albeit one with a considerably smaller workload; Chuck Tanner loved his deep bullpen, and Blyleven didn't. Nonetheless, Blyleven sparkled in the 1979 postseason, going 2-0, with a 1.42 ERA to help the Pirates win the title. After the 1980 season, he was traded to Cleveland, and after a good 1981, an elbow injury cost him nearly all of 1982. He returned to his usual workhorse self after the injury, and was traded back to Minnesota in mid-1985, where it was like he never left: 17-16, 17-14, high strikeout totals, good ERAs, and homers galore (a record 50 in 1986). In 1987, along with Frank Viola, he sparked the 85-win Twins to an unlikely World Championship, shining in the postseason (3-1, 3.42 ERA). He left Minny for California after 1988 and had one of his best seasons (17-5, 140 ERA+) and then sandwiched two mediocre years around one completely missed with rotator cuff surgery.
Hall of Fame voters perform all kinds of gymnastics in attempting to justify why Blyleven doesn't get their vote, most of them fixated on his relatively unimpressive winning percentage (.534), his 250 losses, a win total just shy of 300, and his failure to win a Cy Young award. But his career totals place him in some pretty elite company: 5th all-time in strikeouts (only Ryan, Carlton, Clemens, and Randy Johnson are ahead), 9th in games started, 9th in shutouts, 13th in innings, and 25th in wins*, with virtually everybody around him on those lists either in the Hall of Fame or headed there. The Davenport numbers tell a similar story about Blyleven. Only 10 Hall of Famers have higher PRAA than Blyleven (he's tied with Ed Walsh at 311), and only seven have higher PRAR. A mere five Hall of Famers have higher career WARP3 totals (Walter Johnson, Cy Young, Warren Spahn, Tom Seaver). And while 21 Hall of Famers have higher peaks, only eight have higher weighted scores. Welcome to flavor country.
One of the traditional complaints against Blyleven is that he didn't win any Cy Young awards, and that he didn't win 300 games while a whole bunch of his contemporaries did. Well, here's how Bert compares to his enshrined contemporaries, ranked by weighted score:
PRAA PRAR WARP3 PEAK WPWT PKPCT Seaver 421 1463 142.9 50.2 96.6 35.1 Blyleven 311 1408 135.8 45.6 90.7 33.6 Perry 255 1434 133.4 47.7 90.6 35.8 Ryan 263 1488 131.1 42.2 86.7 32.2 Niekro 209 1385 130.0 42.2 86.1 32.5 Carlton 222 1357 123.8 38.0 80.9 33.6 Jenkins 236 1234 115.7 43.0 79.4 37.2 Palmer 230 1120 108.9 46.4 77.7 42.6 Sutton 170 1354 117.3 36.3 76.8 30.9 Hunter 38 836 76.0 42.0 59.0 55.3One of these pitchers is not like the others, but it isn't Blyleven, it's Catfish Hunter, a pitcher who supposedly "pitched to the score" and thus had some high ERAs, not to mention a relatively short career. Blyleven is second among this group in WARP and PRAA, fourth in PEAK, and second in WPWT. At worst, by these measures, he's the fourth most valuable pitcher in this group. If that's not a Hall of Famer, I don't know what is. There isn't a player on the 2004 Hall of Fame ballot who deserves a vote more than Blyleven.
Relievers are in a very strange spot these days. Late-model closers of the one-inning variety are exalted by the media despite having jobs which are no more difficult than those of NFL placekickers. Their fireman predecessors, who often pitched two or three innings at a time, get almost no love from the Hall of Fame voters. That contradiction is a direct response to the changed usage pattern which has the team's best reliever coming in only in situations (usually the ninth inning) where he can receive a statistical cookie, a save, rather than in a tie ballgame in the seventh, when the game may be on the line but the save rule doesn't apply. Thus the 80-inning/40-save closer has come to be held in higher esteem that a 110-inning/25-save stopper.
This changed usage pattern makes relying solely on save numbers ill-advised when considering the value of a reliever--in the end, it's runs which are important. The Davenport system is well-suited to handle the comparisons we need to make among relievers across eras, and it can offer us guidance in gauging the overall value of these relievers compared to other positions. Unequivocally, the Davenport numbers tell us that it's nearly impossible for the best late-model relievers to be more valuable than the best everyday players or starting pitchers. While yearly WARP3 totals above 10 are common for elite players, the best closers approach eoght only in a Mariano Rivera/Eric Gagne-caliber year, and those don't come around too often.
So if the best closers can't approach the values of other players, do they have any business being enshrined? A negative answer to that question would be overly harsh. With increasingly specialized roles, we've come across performers capable of posting near-perfect seasons. These players should have the ability to receive their just rewards for a job well done over a long and stellar career. And based on our survey of position players, the Hall clearly has a precedent which we could apply: catchers. Due to the strenuous physical demands of playing the position (and, admittedly, to a few poor choices by the Veterans' Committee), the average enshrined catcher has about 82% of the value of the other enshrined hitters, yet nobody is suggesting we shouldn't induct catchers that meet the standards. We simply need to be realistic about what constitutes an appropriate standard for a reliever: is it 90% of a starter's value, 80%, 70%, 60%?
There's a body of research which has shown that good relievers have a quantifiably greater effect on the outcome of a ballgame. Research using play-by-play data and a Win Expectancy Matrix has shown that the results of the plate appearances against them are magnified by some factor, which is called the Leverage Index. A starting pitcher will have a Leverage Index very near 1.0, but an ace reliever might have one approaching 2.0, meaning that the batters he faced were twice as important to the outcome of a ballgame.
In his own exploration of the Hallworthiness of three pitchers still on the ballot, a researcher named "TangoTiger" has suggested multiplying the components of a reliever's line by his Leverage Index and then comparing him to a starter. I'm going to suggest we do something similar but not quite as drastic here. First of all, we don't have uniform play-by-play data to calculate Leverage Indexes for each reliever on the ballot. But we do have LIs for three of the top relievers on the ballot, and they range from 1.7 to 1.9.
I'm not going to suggest that we practically double every pitcher's WARP3 totals here and measure his place against enshrined starters. Rather, I propose examining the conclusions we reach if we set a standard for relievers that's 70% of a starter's value. That's the equivalent of applying a 1.43 Leverage Index to a player's WARP3 line, but to avoid confusion I'm just going to keep reminding us of the 70% standard.
PRAA PRAR WARP3 PEAK WPWT PKPCT Eckersley 277 1128 115.3 36.3 75.8 31.5 Gossage 238 757 80.8 34.1 57.5 42.2 Myers 140 449 48.7 22.1 35.4 45.4 Smith 229 664 71.9 29.8 50.9 41.4 Sutter 148 471 50.6 28.0 39.3 55.3 --------------------------------------------------- AVG HOF SP 239 1002 97.0 44.9 70.9 46.3 70% STD RP 67.9 31.4 49.7 46.3Randy Myers was an excellent left-handed reliever who came up with the Mets during the Davey Johnson era, finally sticking in the big leagues in 1987. He was the closer for their division-winning team in 1988, then was traded to Cincinnati for John Franco, and the deal paid immediate dividends for the Reds. Along with Norm Charlton and Rob Dibble, Myers became one of the notorious Nasty Boys. Myers saved 31 games and posted a 2.08 ERA and a scoreless seven appearances in the postseason, helping the Reds to a World Championship for which he shared co-MVP honors with Dibble. The Reds tried to make him a starter midway through the next season, with mixed results; Myers posted a 3.45 ERA in relief but went only 2-6, and was traded to San Diego over the winter. Myers spent the next seven years wandering the majors as a one-inning closer and a soldier of fortune. He racked up some high save totals in the process, leading the league three times, with a high of 53 in 1993. In a cautionary tale on the dangers of waiver-wire shenanigans, the Padres put in a claim to block Myers from being traded by the Blue Jays to the Braves. He pitched poorly and then tore his rotator cuff the following spring, never to pitch again, but the Padres were left to pay out $13.6 million.
As good as he was, Myers didn't pitch a lot of innings, topping 100 only the year he started, topping 80 only two other times, and averaging 1.2 innings per appearance over the course of his career. His Davenport numbers aren't anywhere near our admittedly abitrary cutoffs for peak or career WARP3, and there are better candidates on this ballot.
Bruce Sutter holds a historic spot in the evolution of the relief pitcher, but he holds an even more important one in the evolution of pitching in general. Sutter came up with the Cubs in 1976, and by the next season he was lights out, pitching 107 innings with a 1.34 ERA and 129 strikeouts while posting 31 saves. Credit for his success was due largely to his mastery of the split-fingered fastball, a pitch which was unfamiliar to big league hitters. Sutter didn't invent the splitter, but he was the first successful practitioner of a weapon that's made many a pitcher wealthy. But the innovations around Sutter didn't stop there. Prompted by his ace reliever's tendency to wear down as the season went on, in 1979 Cubs manager Herman Franks decided to limit Sutter's usage mainly to close games when his team was ahead--in other words, save situations. Sutter tied the National League record with 37 and won the Cy Young, thanks to a 2.22 ERA/101 inning season.
After five stellar seasons in Chicago, he was traded to the Cardinals, where he posted three strong years as well as his first subpar one. He was an instrumental piece of the 1982 world champions, saving 36 games in another 100-inning season and notching a win and two saves in the World Series. He pitched a high of 122.2 innings in '84 and set an NL record with 45 saves while posting a 1.54 ERA. Coincidentally or not, that was his last effective season. Lured by Ted Turner's cable riches, he left for the Braves via free agency after 1984. But in Atlanta his shoulder began to break down, and he was never the same pitcher again. He pitched 152 innings of 4.55 ERA ball for Ted's $10 million, and was done at 35 years old.
The traditional case for Sutter is that in addition to being attached to two notable innovations, he was one of the few relievers to win a Cy Young, a six-time All-Star who threw a lot more innings than today's closers. Excluding the strike year of 1981, he averaged 104 innings a year from 1977-1984. Despite two years in the vicinity of 8 WARP3, the Davenport numbers don't put him near our 70% standard, even at his peak. That's a bit surprising given his dominance of NL hitters, but it aptly illustrates the limited impact of even a 100-inning-a-year role and the difficulty of maintaining that level. Unless he's given an extraordinary amount of credit for the pitch which he didn't invent, Sutter's claim on the Hall of Fame isn't too strong. No vote for him here.
The physically intimidating Lee Smith stepped into the large shoes vacated by Sutter in Chicago and did a very credible job in six seasons as their 100-inning per year closer. From 1983-1987, he finished in the top five in saves every season, leading the league once. He was traded to Boston after 1987 and continued to post high-quality seasons, though his innings and save totals dipped a bit. Traded again to the Cardinals, he flourished, topping Sutter's NL save record and recording 160 saves in parts of four seasons--taking over the all-time lead in that category--before having to pack his bags again. Through five more stops, the innings began to take a toll on his body, and his managers limited his usage to about 50 frames a year, one inning at a time, to keep him effective. He spent his last two seasons in a setup role, with diminishing returns, finally hanging it up in 1998.
From a traditional standpoint, Smith's case starts with his status as the all-time saves leader, his seven All-Star selections, and an amazing string of consistency which followed him to virtually every stop on his 18-year ride. Until his final 22-inning season, his adjusted ERA was always better than league-average--32 percent better for his career. On the down side, his teams never went farther than an LCS, and he got bombed in his brief postseason appearances, blowing two ballgames in best-of-fives. His Davenport numbers are right around our arbitrary 70% standard for relievers, slightly higher on career WARP3, slightly lower on peak. He had four outstanding seasons and several more good ones, but his late-career numbers don't add much to his case due to the low inning totals. But on the whole, he's almost right at the Hall average for PRAA (without any 70% adjustment), and considering his relatively low number of innings, that's pretty impressive. His is by no means an open and shut case, and if we move the bar to 75%, our verdict on him would shift. But I think a vote for him under this system is reasonable--more reasonable than one for Tommy John, at least.
Roughly speaking, Rich "Goose" Gossage set the standard for relievers for a decade, pitched in the majors for another decade, and ten years later is still held up as a yardstick for dominant relievers. From 1975-1985, excepting a year-long failed experiment as a starter, Gossage blew hitters in both leagues away, helping his teams to three pennants, making nine All-Star squads, and keeping his ERA well under 3.00 every single year. He came up with the White Sox, emerging as a dominant force in 1975, his fourth season, when he threw 141.2 innings with a 1.84 ERA and a league-leading 26 saves. The Sox tried to make him a starter, but after a dissapointing season, they traded him to Pittsburgh, where he had an even better year with a 1.62 ERA. That prompted Yankee owner George Steinbrenner to throw big bucks at him--6 years, $2.75 million--despite the fact he already held the reigning Cy Young winner, Sparky Lyle. But with his 100 MPH heat, Gossage usurped Lyle's role as the Yankee stopper. He was brilliant in his six seasons with the Yanks, posting a 2.10 ERA--83% better than the adjusted league average--saving 25 games per year, striking out about a batter per inning, and averaging 86 innings annually despite a Bronx Zoo-brawl related injury in '79 and the strike in '81. He left for San Diego via free agency after 1983, and the move paid immediate dividends, as they reached the World Series in '84. He had his first bad season in 1986, and by '87 had lost his go-to status in the Padre pen. Traded to the Cubs after that season, he began the familiar trudge of the past-prime reliever, not quite settling in a setup role and making five more stops (including a cameo with the Yanks) and spending 1990 in Japan. He topped 50 innings only once in that stretch, mostly due to injuries, but he held his own when he did pitch.
Gossage's case as a Hall of Famer is a reasonable one on the traditional merits; that decade of dominance resonating in the public mind thanks in part to a lot of postseason exposure (19 games, 31.1 innings, 2.87 ERA). Based on the number of innings thrown and his better-than-average ERA, a reasonable case can be made for him as the second-best reliever ever, behind Hoyt Wilhelm. His Davenport numbers are just as strong. Gossage's two best years are off the charts--above 9 WARP3, and by peak, career, and weighted numbers he's better than many of the starters in the Hall, and his PRAA is right at the Hall average. Furthermore, he compares favorably with the two enshrined relievers, with the highest peak among them by a healthy margin:
PRAA PRAR WARP3 PEAK WPWT PKPCT Wilhelm 269 888 91.2 30.7 61.0 33.7 Gossage 238 757 80.8 34.1 57.5 42.2 Fingers 165 692 74.2 31.1 52.7 41.9Gossage would be above our standards bar even if we set it at 80% instead of 70%. His is as good a case for a Hall of Fame reliever as we've come across on this ballot.
Dennis Eckersley holds another historic spot in the evolution of relievers: the model one-inning closer. But he took a strange path to get to that point. Eckersley came up with the Indians in 1975 as a greasy-haired, 20-year-old flamethrower. He pitched well for three mediocre Cleveland teams before being traded to the Red Sox in 1978. In the year of Bucky F. Dent, Eck went 20-8 for the Sox, and put up a 2.99 ERA in 268 innings. He had one more fine season for the Sox, followed by four mediocre to lousy ones which wore out his welcome. Early in 1984, he was traded to the Cubs (for Bill Buckner--ouch!) and helped them to their first postseason appearance in 39 years. But with the Cubs one win away from a World Series berth, Eckersley got bombed in his LCS start, and the team lost three straight games and the series. He had another good season with Chicago in '85 before alcohol troubles began to take their toll, and he entered a rehab program following the '86 season.
Eckerlsey was traded to the A's in 1987 and placed in the bullpen, where he began to resurrect his career. He was a jack of all trades that season, pitching middle and late relief, spot starting, and closing, and putting up a 3.03 ERA in 116 innings. But it was the next season which made his legend, as manager Tony LaRussa developed a pattern of using Eckersley in the ninth inning of close games which the A's led. The model closer racked up 45 saves in only 73 innings of work and posted a 2.35 ERA as the A's won the pennant. Although he gave up a memorable home run in the World Series and Oakland was upset by the Dodgers, Eckersley and the A's had found a winning formula. He reeled off four fantastic years of high saves (as high as 51), low ERAs (as low as 0.61) and astonishing strikeout and walk totals (55 to 3, 73 to 4), all without topping 80 innings. With his help, the A's won three more divisions and two pennants during his time as closer, and in 1992, Eck won both the Cy Young and the MVP. He and the A's began to slide in 1993, but Eckersley maintained his hold on the closer role for three more years before following LaRussa to St. Louis. He fared better in his two NL seasons than he had in the AL, and finally closed his career as a setup man in Boston at age 43.
From a traditional standpoint, Eckersley is as unique a candidate as the Hall of Fame has seen for quite some time. Taken individually, neither his time as a starter nor his renaissance as a reliever would rate much more than a passing nod from voters. But taken together, he's got a 24-year career with totals approaching 200 wins and 400 saves, a combination no Hall of Famer can approach. He's 3rd all-time in the latter category, with 390. And you can forget any 70% or 80% standards when it comes to Eck. The Davenport numbers put him well above the average Hall of Fame starting pitcher in PRAA, PRAR, weighted and career WARP3, though his peak is somewhat below. Interestingly enough, that peak score comes from his first five years in the majors, when he was a starter; his prime as a closer rates at 29.8 WARP3--dead even with Lee Smith. The difference is that Smith was throwing 105 innings a year during his prime, as compared to an average of 71 for Eckersley. One way or another, he's a "yes" vote.
Altogether that's four 'Yes' votes--Blyleven, Smith, Gossage, and Eckersley--coupled with the four hitters from our previous examination. While a "small Hall" advocate may cringe, our ballot still contains fewer than the maximum 10 names, and the choices behind it are ones which, from an advanced statistical perspective, would elevate the standards of the Hall. Skeptics may puzzle at a vote for three relievers when only two such pitchers are enshrined, but this trio would certainly go a long way in establishing standards by which future candidates can be judged.
The creator of the Futility Infielder web site, Jay Jaffe is a graphic designer and freelance writer living in New York City. He hasn't been above replacement level since Little League, but he can be reached at firstname.lastname@example.org.