Previous entries:

Notes on 2003:

It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of Light, it was the season of Darkness, it was the spring of hope, it was the winter of despair, we had everything before us, we had nothing before us, we were all going direct to Heaven, we were all going direct the other way—in short, the period was so far like the present period, that some of its noisiest authorities insisted on its being received, for good or evil, in the superlative degree of comparison only.

—Charles "Don't Call Me Jason" Dickens

This past season was an extreme year for closers. There were very positive signs balanced against very negative ones.

On one hand, many established closers lost their jobs: Mike Williams, Jose Mesa, Scott Williamson, Armando Benitez, and Billy Koch. Still other closers were being converted to starters more and more often (in 2003, Danny Graves and Byung-Hyun Kim, at least for some time) since Derek Lowe found success in the starting rotation two years ago. The Red Sox’s attempt at “bullpen by committee” ended disastrously. Five “closers” had ERAs greater than their league average in 2003.

On the other hand, Dennis Eckersley became the third reliever in the Hall of. A slew of closers had sub-2.00 ERAs in 2003. Eric Gagne had arguably the greatest season ever for a reliever and won the NL Cy Young, the first such award given to a reliever in eleven seasons. Two of the top relievers in the game (Gagne and John Smoltz) were converted starters.

Here’s what I said about his season when he was awarded the NL Cy Young:

We shall nobly save, or meanly lose, the last best, hope of earth.

—Abraham "Nunez" Lincoln

Think not so much of what thou hast not as of what thou hast: but of the things which thou hast select the best... At the same time, however, take care that thou dost not through being so pleased with them accustom thyself to overvalue them, so as to be disturbed if ever thou shouldst not have them.

—Marcus "Giles" Aurelius Antoninus from his Meditations

He pitched magnificently. In 77 games and 80+ innings he had recorded well over 50 saves and had an ERA well under 2.00. Who cared if his winning percentage was just .400? The Cy Young voters look at his stats and gave the Cy Young to…Bob Welch.

Oh, sorry. I wasn't speaking of Eric Gagne, who won the NL Cy Young the other day. It was Bobby Thigpen's great 1990 season that landed him in fourth place in both Win Shares (behind Clemens, Finley and Stewart and fractions ahead of Dennis Eckersley) and the Cy Young vote (behind Welch, Clemens, and Stewart but again ahead of Eckersley).

Thigpen and his one great season were the precursor to Gagne. Thigpen never had Gagne's stuff but he was used in a similar fashion and produced similar results. Leading up to Bobby Thigpen's record save year in 1990, baseball was pushing the closer limits as much as possible. In 1988, the major-league average closer recorded 25 saves for the first time. Every non-strike year since then, the average major-league closer has exceeded the 25-save threshold. As the Eighties ended teams used their closers in more and more a similar fashion. The standard deviation for team closers' save totals (6.96) dropped to its lowest since 1976 (not counting the strike year of 1982) and the standard deviation as a percentage of the average saves (27.38) was at an all-time low (just 25.41%).

This complacency was broken wide open in 1990 when, even though save totals fell (the average closer had 26.81), the ways that closers were used changed dramatically. The standard deviation of saves per closer nearly doubled and was an all-time high (11.11, which has been exceeded three times since, twice in the last two years). I have already offered my theory on this in my relievers study in the Nineties section. To summarize, the brief offensive onslaught of 1985-87 had subsided and the strategy of using one of your best pitchers as a closer who you hold off on using until late in the game to prevent a late-inning comeback by your opponent fell out of favor.

Tony LaRussa started using Dennis Eckersley in what I have termed the post-modern closer mold, i.e., fewer innings, shorter outings, and more saves. Oakland also had a great supporting cast in the bullpen so limiting Eckersley's innings didn't hurt them. This culminated with his 1992 MVP- and Cy Young-winning season, the second 50-save year for a closer.

Meanwhile, Thigpen's career feel apart: after his record 57-save season he only recorded 53 total saves for his career. Even though by 1992 Eckersley was used less in the post-modern role than he had been in the previous few (80 innings in 69 appearances), the post-modern closer took root. This was accelerated as most of the next decade was spent in an expansion-induced offensive explosion. Closers were held back until the ninth for fear that a lead would be lost in a final walk-off at-bat.

However, as the offensive onslaught slackened, the battle for the preferred usage of the closer role heated up like a debate between Heat Miser and Cold Miser. Since 2000, the varied use of relievers has gone through the roof. The standard deviation of saves among all closers hit 11% of the average for the second time since 1990. It's been over 11.25% each of the last two years. Meanwhile, the average number of saves climbed to an all-time high of 32.93 in 2002 only to fall by over four saves (to 28.00) this year. Eric Gagne led all closers with 55 saves, but this year also witnessed a team save leader—dare I say closer—who recorded only 5 saves (Franklyn German and/or Chris Mears of Detroit) and four other AL teams that didn't have a reliever record more than 16 saves (Boston, Seattle, Toronto, and the White Sox). Five saves for a team leader is the least that has been recorded since 1975 in a non-strike year.

By the way, here's a quicky comparison of Gagne's 2003 season with Thigpen and Eckersley's 1990 seasons:

Name     W-L  G SV  IP   H BB   K  ERA Adj WHIP  K:BB K/9IP Win Shares

Thigpen  4-6 77 57 88.2 60 32  70 1.83 210 1.04  2.19  7.11 20

Gagne    2-3 77 55 82.1 37 20 137 1.20 335 0.69  6.85 14.98 25

Eck      4-2 63 48 73.1 41  4  73 0.61 606 0.61 18.25  8.96 20

They all had great years, but only Gagne won the Cy Young.

OK, so where does that leave us? Oh, yeah, I don't think Gagne deserved the award this year (I would have tabbed Mark Prior or Jason Schmidt myself) and I'll explain why.

First, let me explain from an historical point of view. Just as Bill James said in 1991 of Bobby Thigpen's save record, "Obviously, Thigpen's record was brought about in part because of a change in the way that relief pitchers are used-a generalized change, operating th[r]oughout baseball," so too is Gagne's stats brought about by a recent return to that strategy.

Baseball took a slight evolutionary detour after the great success of Dennis Eckersley. This Neanderthaloid evolutionary dead end took a decade during perhaps the game's greatest offensive era, the Mets-ozoic Era, which also encompassed two rounds of expansion.

Perhaps with all the expansion, bullpens got too watered down. It became very difficult to build a bullpen as effective as the Eck-era A's pen. There have been very good pens of late, but they seem to be more capturing lightning in a jar than crafting a pen by design.

Perhaps the trend played itself out. I mean, how long can you employ one of your best pitchers for just 50 innings a year? Byung-Hyun Kim moved back to the rotation and then back. Former closer Danny Graves went 4-15 in 2003, his first year as a full-time starter. And of course Derek Lowe successfully moved back to starting last year.

Also, teams started to realize that reliance on a player just because he has recorded 30 or so saves in the past is ludicrous. Big name closers like Mike Williams, Jose Mesa, Scott Williamson, Armando Benitez, and Billy Koch lost closer jobs on their teams or were traded to other teams and no longer closed. Journeymen like Rocky Biddle, Joe Borowski, Tom Gordon, Tim Worrell, Lance Carter, and Rod Beck inherited closer jobs and for the most part performed well.

Needless to say, the job of closer has gone through a reevaluation period the last couple of seasons. In Gagne, the Dodgers appointed a hard-throwing yet highly unsuccessful young starter as their closer in 2002. It revitalized his career to say the least. However, I see Gagne as a throwback to the Thigpen line. It's not that he hasn't pitched well, very well. I just think that his performance looks all the better for the shambles that closer role has become on many clubs, as well as his calling Dodgers Stadium home (though his 2003 home-road splits show no bias to Dodgers Stadium). He happens to be in the right role at the right time.

To illustrate, there have been 87 closers all time with at least 20 saves and an ERA under 2.00, starting with Ellis Kinder in 1953. 53 of them have done it while pitching at least 80 innings.

Well, one can argue that 20 saves is not 55. However, I think that the number of saves a closer records to a very large degree depends on his era and his usage. The Reds' Ted Abernathy recorded a 1.27 ERA in 1967 with 28 saves and a 6-3 record in 106.1 innings pitched. I know it was a pitcher's era, but his league-adjusted ERA is about the same as Gagne's (295 to 335) and he did it in so many more innings. In 1967, 28 saves were enough to lead the majors. So of course Abernathy got a ton of Cy Young votes, right? Uh, no. He didn't receive a one. Mike McCormick, Jim Bunning, and Fergie Jenkins were the only men to receive any votes that year. Abernathy's season was worthy of 24 Win Shares, one fewer than Gagne this year and fractions behind league leader Bunning.

You say you want someone more recent. How about John Wetteland's 1993 season with the Expos? He was 9-3 with 43 saves in 85.1 innings over 70 appearances. He struck out 113 and walked just 28. His ERA was 1.37 and his adjusted ERA was 304. That's a pretty close match to Gagne. He was awarded 21 Win Shares, but was fourth behind league-leading Greg Maddux and Jose Rijo.

I should also point out that the Dodger bullpen this year featured two other pitchers with ERAs under 2.00, Paul Quantrill (1.75 ERA in 77.1 innings) and Guillermo Mota (1.97 in 105 innings). I have said that the Dodgers are the anti-Rockies. Pitchers go to the Dodgers and resurrect their careers (e.g. Hideo Nomo's second stint, Wilson Alvarez, Kevin Gross, and Odalis Perez , at least in 2002). Then when they leave they fall flat on their faces (e.g., Hideo Nomo after his first stint, Kevin Gross, Chan Ho Park, Ismael Valdes, Ramon Martinez, and Tim Belcher ). Not that pitching for the Dodgers bars a player from winning a Cy Young, but it should make us a bit leery just like we were of Vinny Castilla's and Dante Bichette's stats in Colorado.

Well, what about Gagne's historic 14.98 strikeouts per nine innings figure? Bill Wagner was the man he beat out (14.95 K/9 IP in 1999), and Wagner finished fourth in the Cy Young vote that year.

The last stat that people will throw out is Win Shares. Gagne led NL pitchers with 25 Win Shares, three more than Mark Prior, Jason Schmidt, and Livan Hernandez. Did anyone notice that Rheal Cormier finished 19th in the NL in pitching Win Shares, ahead of every other Phillie? Cormier had a very good season (1.70 ERA, 8-0 record, 54 hits and 25 walks in 84.2 innings/65 appearances). However, given that the Phils had four men win 14 games and three pitch over 200 innings, few would pick Cormier as their best pitching asset.

The reason is that Win Shares for relief pitchers is inherently problematic. I don't want to take anything away from James: Win Shares is probably the crowning moment of the career of baseball's greatest analyst since Henry Chadwick. However, trying to assign Win Shares to relievers is like trying to hit a moving target. The landscape of relief pitching has changed dramatically since James published Win Shares just two years ago. How can a standard formula be applied to all relievers throughout baseball history, especially when it takes a half dozen just to figure fielding Win Shares for third basemen throughout baseball history?

Given that relief pitching Win Shares were derived basically via a compromise in James' formulae, who's to say they are accurate for the 2003 season? Individual pitching Win Shares are derived from assigning claim points to a team's staff via a set of criteria and then meting out Win Shares appropriately. The criteria are runs allowed (the largest factor); wins, losses, and saves; save-equivalent innings; and batting. The claim point formula for wins/losses is ((W*3)-L+Sv) / 3 (p. 35). That seems pretty straightforward. Of course, one could argue that this formula really doesn’t measure anything, but at least it's straightforward enough. My problem is with Save Equivalent or Crucial innings. That formula is to multiply saves by three, cap the result at 90% of actual innings pitched, and finally add one for each hold (Hello, Rheal Cormier). The save-equivalent innings are then multiplied by the pitcher's component ERA added to a constant (.56) minus the team cutoff to get the claim points. OK, why not? But why 90%? Why not 80%? Or perhaps why not an era-specific percentage? Why multiple by three? Why not 2 or 4? Why not 2 in 2003 and 4 in 1967? To say that a save is a save no matter the era is problematic at best.

Win Shares is a valuable tool, but it's just that, a tool. Sometimes it proves useful; sometimes not. And given that the formulae for starting pitchers' and relief pitchers' Win Shares differ a great deal, it becomes dangerous to use Win Shares as the be all and end all for ranking all pitchers.

I think Win Shares is the shakiest ground from which to build one's argument for Gagne's Cy Young legitimacy. Gagne had a fine season, but I cannot accept an argument that his 82.1 innings were superior to Prior's 211.1 or Schmidt's 207.2 even if they came in save opportunities. And what of save opportunities? James showed in his New Historical Baseball Abstract that a closer is best used in games in which his team leads by one, the score is tied, or possibly if his team trails by one. Only one of those three would even be considered a save opportunity. Using a closer to hold a three-run lead in the ninth is mere overkill.

So, next we will take a look at Gagne's game log to determine if his appearances were indeed that crucial to his team's success to merit winning the award. Here is a table of Gagne's appearances and the situation when he entered the game. A Dodger lead is represented by the number of runs they led by at the time. If the Dodgers trailed, then the number of runs they trailed by is represented by a negative number:

Situation  -3+ -2  -1 Tied  1   2  3+

# Games:    2   0   1  13  24  11  26

Gagne seemed to be used in almost all of the Dodgers extra inning games (16). All of his five decisions came after he entered a tie ballgame. He gave up four runs in a tied game on May 12, one on June 23, and one on July 2, all for losses. He gave a run in a tie game on August 20, but the Dodgers came back to win the game. His second win he garnered when the Dodgers broke a tie ballgame in extra innings that when he was pitching. His other three appearances in which he relinquished a run were a) he gave up two runs when the Dodgers already trailed by more than 3, b) he gave up a run when he was provided a two-run lead, and c) he gave a run with a 3-run lead.

Basically, Gagne made 37 or 38 significant appearances out of 77. His 37 appearances when staked to two or more runs (26 with 3 or more) are the 19-yard field goal for closers. A competent one should be able to hold that lead. Yes, he did not blow any and yes, he only gave up two runs in those 37 games. However, if he were an average reliever, how many of those could he have blown? The Dodgers' team ERA was 3.16. An average Dodger pitcher would not have blown more than a handful of those games.

So it comes down to 13 tie ballgames, 24 with a one-run Dodger lead, and the one the Dodgers trailed by one run, i.e., the close games in which Gagne pitched. He "blew" four of those 13 tie ballgames, but none of the one-run leads. So basically Gagne was given the award for 33 or 34 ballgames in which he pitched mostly one inning. Prior started 30 games and Schmidt 29 and each went significantly longer than one or two innings in those ballgames. Finally, the Dodgers were 26-23 in one-run ballgames, so how significant was Gagne's performance anyway?

Gagne had a stellar year, but the role of the closer is still too marginalized to merit winning a Cy Young award.

Since writing that, I have heard various arguments in support of Gagne's Cy Young crown. The best I have heard was by my friend Chris who said that if you have to give the award to a closer once in a while, then this was the year.

There was no one dominant starter in the NL. Prior and Schmidt were great but neither started more than 30 games (Prior did not have a start between July 11 and August 5, after colliding with Marcus Giles; Schmidt missed a few turns at around the same time). Had they each had three or four extra starts as a typical Cy Young-level starter usually would, I bet that the Cy Young would have been a dogfight between the two probable twenty-win starters. Without those few extra starts, they become a 17- and an 18-game winner with great stats: Rick Reuschel for a new generation. How many Cy Youngs did Big Daddy win anyway?

Gagne's 2003 season was arguably among the best ever for a reliever. Even so, as I showed above, he probably only had thirty meaningful innings all year. Compare that to thirty meaningful starts of five or more innings for the best starters. If you can accept that thirty innings for a closer are more valuable than thirty starts for a starter, then Gagne's your man.

I don’t share that opinion, but I admit that Gagne’s is one of the greatest seasons ever for a reliever. To prove the point, here’s a table of all the relievers who have posted an ERA of 1.50 or lower while appearing in at least 30 games (sorry, I had to split it in two):

OK, so there were 52 such pitchers, three of whom, including Gagne, pitched this year. Actually Gagne’s record is only slightly better than Smoltz’s, and that was due mainly to the time Smoltz lost and to Gagne’s ungodly strikeouts per nine innings. However, note Smoltz’s strikeout-to-walk ratio superiority over Gagne. If there’s one knock you can make on Gagne’s 2003 season, it’s the walks. He didn’t walk a lot of men (20), but it was enough to affect his K-to-BB ratio a bit, even though it’s double the average for what amounts to 52 of the greatest seasons ever in relief pitching.

For the record, Gagne has the best strikeouts per nine innings, 14.98. Second is Rob Nen in 2000 at 12.55. The lowest ERA on the list is Dennis Eckersley in 1990 at 0.61 (Gagne’s 1.20 was 14th). He also recorded the best walks plus hits divided by innings pitched (WHIP) that year also at 0.61 (Gagne was second with 0.66). Eckersley in 1990 also recorded the best strikeout-to-walk ratio on the list, an ungodly 18.25, almost twice as much as second-place total, 9.13, recorded by John Smoltz in 2003 (Gagne’s 6.85 is third). Three men recorded a zero home run per nine innings ratio (Rob Murphy in 1986,
Frank Williams in 1986, and Bob Miller in 1971; Gagne is 14th). Finally, Dale Murray (1974) is the only man on the list to average over two innings an appearance (2.18). Gagne is 41st in this category. The most Win Shares? John Hiller’s 31 in 1973 (Gagne is third).

So, clearly Gagne’s 2003 campaign is among the greatest ever for a reliever. I prefer Eckesley’s 1990 season for the post-modern closers, Bruce Sutter’s for the modern closers, and John Hiller’s 1973 for ye olde tyme closers. But Gagne was better than a number of closers who won either a Cy Young or an MVP or both.

Dee Average (Or What's Happenin' Now?)

But extreme positions are not succeeded by moderate ones but rather by equally extreme, but inverted, positions.

—Friedrich "Fat Freddie" Nietzsche

This season there were seven closers with sub-2.00 ERAs: Gagne, Smoltz, Mariano Rivera, Bill Wagner, Shigetoshi Hasegawa, Danny Kolb, and Rod Beck. Now, I consider the team save leader to be the closer for these analyses. However, many may say that the first four on the list were the only true closers. The last three recorded 19 saves on average, not a lot in this era.

Yet these seven sub-2.00 "closers" were the most in the majors in fourteen seasons. In 1989, Roger McDowell, Jeff Montgomery, Dennis Eckersley, Jay Howell, Bill Landrum, Gregg Olson, Mark Davis, Tom Henke, and Jeff Russell all had ERAs under 2.00. That also happens to be the last year before 2003 that a reliever (Davis) won the NL Cy Young.

These nine men represent the all-time high for sub-2.00 closers in a season. (1908 matched it when sub-2.00 starting pitchers like Cy Young, Christy Mathewson, Addie Joss, Ed Walsh, Three-Finger Brown, and Rube Waddell also acted as closers for their teams.) The seven this year matches the strike-shortened 1981 season for second (1981 sub-2.00 closers: Rich Gossage, Rollie Fingers, Kevin Saucier, Dan Quisenberry, Rick Camp, Joe Sambito, and Woodie Fryman). By the way, there were only three last year (Gagne, Percival, and Julio) and one in 2000 (Nen).

But at the opposite end of the spectrum, there were five closers with ERAs greater than the league average (Rocky Biddle in Montreal, Jose Mesa in Philly, Mike Williams in Pittsburgh, Jose Jimenez in Colorado, and both Franklyn German and Chris Mears in Detroit). The last time there were that many was 1999 at the height of offensive surge from which baseball has just gotten itself disentangled in the last couple of years. And in 1999, one closer (Dave Veres in Colorado) actually had a park-adjusted ERA 12% better than the league average and another (Boston's Tim Wakefield) was actually the team co-leader in saves. There was just one team save leader with an ERA worse than average in 2002 (Hideki Irabu in Texas).

The all-time was 10 in 1992, right at the cusp of the offensive gluttony of the last decade. I have to list those ten because they read like a list of similar pitchers to Roberto Hernandez: Alejandro Pena, Jeff Reardon, Bobby Thigpen, Mike Henneman, Roger McDowell, Doug Henry, Anthony Young, Mitch Williams, Randy Myers, and Mike Schooler. Thigpen's 4.74 ERA was the highest on the list (as compared to Mesa's 6.52, Williams' 6.27, and German's 6.04 in 2003). Henneman's 3.96 seems paltry in comparison. (Then again, the only other time 10 team save leaders eclipsed the league average ERA was 1969 with Frank Lizzy's 3.63 ERA, Phil Regan's 3.70, Rollie Fingers' 3.70, and four other sub-4.00 ERAs on the list. The highest was John Boozer's 4.28.)