Complete Games and Relievers

Tuesday, July 22 2008 @ 07:16 PM EDT

Contributed by: Magpie

There was something very special about the 2007 season, and as far as I can tell, it went almost completely unnoticed. So let me tell you about it.

The Texas Rangers, Florida Marlins, and Washington Nationals got no complete games from their starting pitchers. Not a one. Zero. Bugger-all. And until last year, no major league team had ever gone a season without someone going the distance. It just seems weird that three teams should suddenly break through this previously unbroken barrier together.

(Author's Note: I've had this damn piece in its current almost-finished state looking reproachfully at me for about three months. But I'm tired of working on it! I'm sorry! I know I should say a little bit more about some of the more recent developments - but what the hell. There are guys around here besides myself who were alive during those days. I've done the real heavy lifting! I want to get this off my plate, and wrap up its companion piece as well, and Get On With My Life!)

Anyway - if there's one thing we do know about baseball, it is this: ever since the mound was moved back to 60 feet 6 inches, the work required from starting pitchers has decreased. And decreased. And decreased.

For example. It's been a quarter-century since anyone pitched 300 innings in a season (Steve Carlton, with 304 IP in 1980 was the last. Perhaps for all time, who knows.) The most recently active pitcher to throw 300 innings in a season was Nolan Ryan, whose career ended in 1993. Ryan had worked 300 innings twenty years earlier in 1973 and 1974.

And complete games are vanishing from the earth. And complete games is what I intend to look at - because it's easy. You can assemble a list of complete games for every team and every season. (Whereas sorting out innings pitched by starters and relievers is fiendishly tricky - remember that for a good chunk of the twentieth century, a team's ace starter often doubled as the team's ace reliever back. This practise endured as late as Casey Stengel's Yankees teams of the 1950s.)

The number of complete games has been declining steadily ever since they moved the mound back. Observe:

That's a trend everyone can believe in.

If starting pitchers are doing less, relief pitchers must be doing more. And as it happens, it's also very easy to discover how many times a team has used a relief pitcher. You just add up all the appearances by the pitchers and subtract the number of games. If starters are throwing fewer complete games, we would expect the number of relief pitchers to increase.

With all of this at our disposal, it seemed like a good opportunity to see if we could actually find the tracks in this data of certain well-known and specific developments in the game's history. Some that came to mind were the lively ball, the Babe Ruth Era, World War II, the DH - and we might also look at how the work of certain well known managers compared to what was going on in the game at the time.

And it will give me an excuse to contrive yet more pretty pictures, zooming in specific moments in time.

Prior to 1893, when the mound was moved back, the most appearances by a relief pitcher in any season was 14. This was done by Jack Manning, an outfielder for the 1876 Boston Red Caps. Manning was the regular right fielder, but he also started 20 games on the mound. And clearly, sometimes they brought him in from the outfielder to pitch in the middle of the game. Don't see that too often anymore, although it's how the Blue Jays discovered Dave Stieb. It is obviously very important to understand that until 1891, it was against the rules to make a substitution except in the event of an injury. In the case of the 1876 Red Caps, what seems to have happened is that sometimes Manning and pitcher Joe Borden switched positions in the course of the game.

Everything was different then. When Manning and Borden were swapping positions in mid-game, pitchers were required to throw underhand. Only recently had they been allowed to snap their wrist during the delivery, and just shortly before that were they allowed to take a step during the delivery. There was no pitcher's mound, no rubber - the pitcher threw from a box painted on the field, the size of which altered from year to year.

The game we recognize begins in 1893.

1. In the Beginning... (including the Advent of The Lively Ball) 1893-1920

Contrary to what many people seem to think, the first lively ball era was not the Roaring Twenties. It was a decade earlier. The new ball was introduced in 1911 and the impact was immediate, and substantial. Runs scored in the AL soared from 4573 in 1256 games (3.64 per game) to 5658 runs in 1228 games (4.61 per game). The new National League ball was made by Spalding rather than Shibe, and scoring also increased in the NL although not to the same degree.

The pitchers quickly regained control of the game, largely through the stratagem of defacing the baseball at every opportunity. Scoring soon fell back to the level established in the first decade of the century. The balls were lively, but the players were doing everything they could to deaden them. As you can see, the brief Lively Ball Era of 1911-12 left no impact upon pitcher usage patterns. No one saw it coming, so no one had made any preparations. And it was over - ended by the actions of players - before a strategic response had time to develop.

But something had happened already. After holding steadily in 75-80% range for the first decade of the 60 foot mound, complete games fell to a little over 50% in less than a decade. This period of decline includes the brief Lively Ball era, but was clearly well underway already.

If we give that image an intense scrute, it appears that 1905 is the year when things began to change. And a quick check of the actual numbers tells us that the percentage of CGs fell from 87% in both leagues to 80% in the NL and 79% in the AL from 1904 to 1905. With the exception of the NL's Boston Beaneaters, in 1905 every major league team used more relievers and every major league team's starters threw fewer complete games than they had in 1904. While the trend was almost unanimous, there were a few teams that were leading the way - Cincinnati and Pitssburgh in the NL, Boston, Philadelphia, and Washington in the AL. And New York. Especially New York. The Highlanders were managed by Clark Griffith, who had led the American Association in relief appearances way back in 1891. He was now a 35 year old player-manager, and he called his own number to come out of the pen 18 times in 1905, which was the most relief appearances anyone had ever made in a season. He had is best season in years, going 9-6, 1.68. He started 7 games and finished four of them, but it's very likely that the majority of his 101.2 IP came as a reliever. Griffith also used his other starters out of then pen: Bill Hogg made 22 starts and 17 relief appearances, Jack Powell made 23 starts and 14 relief appearances. He went to his bullpen 76 times, almost twice as often as any team had ever done (previous high was 40, by Brooklyn in 1900.)

In 1912, Griffith took over the Washington Senators. He finished up his playing career there, managed the team through 1920, when he became the majority owner, and ran the franchise until his death in 1955. He had success as a player-manager, and twice he won pennants in Washington after putting a player in charge (Bucky Harris in 1924, Joe Cronin in 1933.) He'd had success as a relief pitcher, and his Washington team would come up with the first great reliever in the game's history.

In the first decade of the twentieth century, Griffith was simply ahead of the curve. After the 88 CGs in 1905, his Highlanders pitched 99, 93, and 90 complete games over the next three years. They were last in the league in this category, but by 1908 the rest of the league had moved much closer to him. Cleveland's pitchers had thrown 140 CGs in 1905, and the other six teams ranged from 117 to 134. By 1908, everyone was in the 102-109 range except Griffith's Highlanders (90) and Hugh Jennings' pennant winning Tigers (119.)

In the National League, two of the league's elite teams were doing some interesting things with relievers. The best team of all, of course, were the mighty Chicago Cubs. Their ace starter doubled up as their ace reliever. Mordecai Brown was one of the very best pitchers in the game - from 1906 through 1910, he won 20 or more games with an ERA below 2.00 every season. He also led the league in saves four years in a row - he made 149 relief appearances in his career and saved 49 games. He was the first man to save more than 10 games in a season, when he had 13 in 1911 (he started 27 games, relieved 26 times - he went 21-11, 2.80).

John McGraw of the Giants also liked to use his best starter to close the game when necessary. Joe McGinnity led the leaue in saves three times during the first decade of the century, and saved 24 games in his career. Christy Mathewson came out of the pen as well, and is recognized as having recorded 28 saves. But McGraw tried some new things as well. In 1906, a young pitcher named George Ferguson worked in 22 games. He only started once, and he pitched fairly well: 2-0, 2.58 with 7 saves. He only had one more year with the Giants, however, before being traded to Boston (where he had three unimpressive years working mostly as a starter.) Doc Crandall was more successful. It's tempting to think that Crandall may have been the game's first true swing man, were it not for the fact that the Red Sox were using Charlie Hall in exactly the same way during the exact same years. But from 1910 through 1912, Crandall started 18, 15, and 10 games and came out of the bullpen 24, 26, and 27 times.He doesn't seem to have made spot starts - he seems to have gone into the rotation for several weeks at a time, and then returned to the bullpen. At any rate, it worked out great - during those three seasons, he went 45-16, 2.88 with 12 saves. Hall was not nearly as effective for Boston, but he did have an outstanding year in 1910 and in 1913 he became the first man ever to make more than 30 relief appearances in a season.

We might take note of one other pitcher - in 1917, Dave Danforth came out of the bullpen 41 times and saved 9 games for the World Champion White Sox. Danforth was an average pitcher at best, and much of this may have been mop-up work. Starters Cicotte, Faber, Williams, and Russell were all used as relievers as well (they combined to make 53 relief appearances and picked up 16 saves.) Danforth may have been picking up Williams in particular, who didn't have a very good season (17-8, 2.97 and he barely completed a quarter of his starts.)

By 1910, CG percentage dropped all the way to 56% in the NL, and by 1913 the AL was down to 51%. In 1916, AL pitchers would barely complete half their starts (632 of 1250, 50.6%), which is the lowest figure for this entire period. At the end of this period, we actually saw an upswing in complete games - in 1917 and 1918.

John Malarkey (now there's a great name!) was the first to match Manning's record of 14 relief appearances in a season. Malarkey, who pitched for the Washington Senators, doesn't seem to have been very good. He went 0-8, 5.99 and may have only been coming in when the game was out of hand (although he did have two saves.)

2. The Clean Ball Era (including the Ruth Revolution): 1921-1950

In 1920, a dirty baseball killed a star player. Baseball responded by banishing the trick pitches developed by defacing the baseball (a grandfather clause was included for designated spitballers) - and for the first time ever, an effort was made to keep clean, fresh balls in play. The clean balls were much easier for the hitters to see - and as an added bonus, they had a lot more life left in them than the bruised and tattered things hitters had been swinging at for the previous ten years. And at the same moment in history, Babe Ruth was having an unprecedented amount of success by trying to hit the ball as hard as he possibly could. All the time. Which was a radically new approach to hitting, and surprisingly successful.

It was during this period when we find that the first team which used on average one reliever per game. That team was the 1922 St Louis Cardinals, and their manager was none other than Branch Rickey, then in his fourth season at the helm. The Cardinals went 85-69, good for third place. Rickey only had three half decent starters: Jeff Pfeffer, Bill Sherdel, and Jesse Haines. The team's 60 CGs was the lowest in the NL, and Rickey leaned heavily on two relievers - Lou North and Clyde Barfoot both made more than 40 appearances out of the pen.

The first AL team was the 1923 Detroit Tigers to use as many as one reliever per game, and you have heard of their manager - a fellow named Ty Cobb. Like Rickey's Cardinals, Cobb's Tigers were using relief pitchers mainly because they didn't have much in the way of starting pitching. The Tigers weren't particularly good; while they went 83-71 and finished second, they were a distant 15 games behind the Yankees. Cobb only had two decent starting pitchers - he didn't settle on any particular relievers to get the team out of trouble, but tried a number of arms. None of whom were much good.

The over all trend line for the thirty years from 1920 to 1949 shows a very slow but steady decline in complete games, from about 56% in 1920 to about 40% thirty years later. The National League is a little ahead of the curve in getting more relievers into the game. In the first great Year of the Hitter, 1930, National League teams used 1200 relievers in 1236 games, and in 1934, for the first time the league averaged 1 reliever per game. It would do so several more times before the American League finally crossed that particular threshold, which didn't happen until 1948. As you can see, immediately after the end of the Second World War, there was an uptick in the number of relievers being used by both leagues. Which seems odd to me - intuitively, you would expect increased emphasis on strategic moves when talent is in short supply, and when was talent in shorter supply than during the war? But no, we see an increased use of relievers after the players returned to action (and after the initial steps to integrate the majors as well.)

But beyond the continuing trend, there are very significant developments that should be noted. This period saw the first great pitcher to be used primarily as a reliever, and eventually the emergence of the relief ace. Fred Marberry joined the Washington Senators in August 1923 - he got into 11 games, 4 as a starter, and pitched very well (4-0, 2.82). In 1924, he made 35 relief appearances and saved 15 games, breaking Miner Brown's 1911 record. (He also made 15 starts). Marberry started once in the World Series, and relieved three times. Bucky Harris did not use his top starters to save the game - Walter Johnson and George Mogridge made no relief appearances at all. Marberry was the first of his kind, partially because of how he was used but mostly because he was by far the best pitcher ever to be used more often as a reliever than as a starter. He pitched in 55 games the next season, all of them in relief, and saved another 15 games; and in 1926, he appeared in 64 games (5 starts) and saved the absolutely unprecedented total of 22 games, a figure which would go unmatched until 1949, when the first generation of true relief aces were finally on the scene. With 52 saves in three years, he was also the all-time leader, and would remain the all-time leader (with 101 career saves) for another twenty years. He slipped a bit in 1927 and 1928, and in 1929 Walter Johnson, his new manager, moved him into the rotation. Over the next three seasons, he started more often than he relieved, and was as good as ever. He was traded to Detroit in 1933, where he had one very good year as a starter. After an ordinary season in 1934, split between starting and relieving, his arm seems to have given out early in 1935.

His numbers may not look like those of a great pitcher - partially it's because he was active during the years of the Great Offensive Explosion, and partially it's because his career was divided between starting and relieving. Lifetime he was 148-88, 3.63, with 101 saves.He started 187 games (and won 94 of them) - but only once did he make 30 starts in a season, and the most games he would ever win a single season would be 19. He relieved 364 times in his career: he had two seasons with more than 50 relief appearances, another season with 46 relief appearances, and three more seasons with at least 30 relief appearances. In those six seasons, when he used primarily as a reliever, he led the league in saves five times.

Marberry was the first great reliever, but his managers had a great deal of trouble fighting the understandable temptation to put him in the rotation. Not so with Grandma Johnny Murphy. Marse Joe McCarthy of the Cubs, Yankees, and Red Sox may have been the greatest and most successful manager of all time - I like to think so. McCarthy is not generally regarded as a particularly innovative figure, but it's hard to find a true predecessor for what he did with Murphy, who had arrived to stay with the Yankees in 1934. He was a finesse pitcher with a good curveball, who made 20 starts, pitched 20 times in relief, and was generally pretty good: 14-10, 3.12. He would start just 20 more games over the rest of his career, but would make three All Star teams coming out of McCarthy's bullpen. He wasn't as good as Marberry, but he had an 11 year run pitching almost exclusively in relief and never had a bad season.

Murphy wasn't exactly a failed starter, but he wasn't an impressive looking pitcher. Hugh Casey wasn't a failed starter either - he'd won 15 games as a rookie while starting 25 times and relieving 15 times - but unlike Murphy, Casey was a hard thrower. He was still starting on occasion in 1940 and 1941 (when he threw the fateful pitch that got away from Mickey Owen in the World Series). By 1942, though, he was pitching almost exclusively out of the pen. He had a very good season in the role, and then went away to war for three years. He came back in 1946 and had the best season of his life, and the following year he saved 18 games, the most ever by an NL pitcher. He was 33 years old by then, and it was his last effective season. He would be released by the Dodgers, Pirates, and Yankees in turn over the next two seasons, and four years later he would commit suicide. He had just three years spent full time in the role of relief ace - but he was impressive as hell, and he was doing it in New York City.

And this was also true of Joe Page, was probably the first true bullpen star. He was a big left hander who'd spent three years moving back and forth between the rotation and the bullpen, with erratic results. He was another hard thrower, whose control came and went. In 1947, Bucky Harris put him in the bullpen and left him there. He had a terrific season: 14-8, 2.48 with 17 saves, and he added a win and a save in the Yankees World Series win over the Dodgers. He had an off year in 1948, but under new manager Casey Stengel he would have one of the greatest seasons any relief pitcher had ever had to that point: 13-8, 2.59 in 60 games, while saving 27 games, far more than anyone had ever done before. The save statistic did not exist, of course, and the save totals of these pitchers has been determined after the fact. What Page's save total does is describe how his manager was using him - it tells us that he was used to close out the game.

That was it for Page, though. He had a poor season in 1950, and the Yankees released him the following May. He was very impressive to watch, and working in New York for a pennant winner he received quite a bit of notice - but the example of Jim Konstanty may have made an even larger impression on the baseball world. Konstanty was a minor league journeyman, who'd had two brief cups of coffee with Cincinnati and Boston before arriving in Philadelphia in late 1948. He was 31 years old by then. He'd also added a new pitch, a palmball. In 1949, he appeared in 53 games, all in relief - he went 9-5, 3.25 with 7 saves. The Phillies finished 3rd, a distant 16 games behind the Dodgers. But in 1950, the Whiz Kids won a dramatic showdown with the Dodgers to claim their first pennant in 35 years. Konstanty appeared in a record 74 games, and went 16-7, 2.66 with an NL record 22 saves. He received the MVP award for his efforts, the first reliever ever to win the honour - one assumes that a perfect storm conspired to make this possible. Both Konstanty and the Phillies had come out of nowhere, and Konstanty had done things no pitcher had ever done before. The understanding was that a relief pitcher had made winning a pennant possible. That it would never have happened without him.

3. The Years of Upheaval (including Franchise Shifts, Expansion, and the DH: 1951-1979

After Page and Konstanty, it became an accepted baseball truth - everybody had to have a relief ace. Al Brazle, who had been swinging back and forth between the rotation and the bullpen for the Cardinals, became a full-time reliever and led the NL in saves in 1952 and 1953.

And then there were the Boston Red Sox. As always, the BoSox fancied themselves contenders for the 1951 AL pennant, and why not? They could certainly score runs - in 1950, they'd scored an astonishing 1027 runs. Only five other teams have scored more than 1000 runs in a season since 1900, and those five all played during the great offensive explosion of the 1930s. Those Red Sox may have had the greatest offense of all time. They'd only had three decent pitchers, however: starters Mel Parnell and Joe Dobson, and swingman Ellis Kinder. The rest of the staff was a disaster - those three were the only men with ERAs below 5.10, and the bullpen was especially frightful. So they traded Dobson to the White Sox for two pitchers, and presumably taking note of the Konstanty phenomenon, made Kinder a full-time reliever. Kinder had also been a minor league journeyman who didn't make it to the majors until he was past 30. In 1949, he'd gone 23-6 as a starter with the Red Sox; in 1950 he was 14-12, 4.26 in 48 games, 23 starts. By 1951, he was already 36 years old, but for the next three years, he was outstanding as a full-time reliever. But the Sox didn't win anything. They didn't have enough starting pitching to keep up with the Yankees (the two pitchers they acquired for Dobson didn't help.).

In 1952, the Dodgers put Joe Black, their latest acquisition from the Negro League, into the bullpen. Black gave them an an outstanding season (15-4, 2.15 with 15 saves). The Dodgers saw Black's future in the rotation, however - he actually made three starts in that year's World Series. It was only good season - at his manager insistence, he would try to add some new pitches to his arsenal the following spring, and lost his command of the ones he'd been throwing so effectively. But in 1952, another rookie became just the second man to pitch in more than 70 games in a season. His name was Hoyt Wilhelm, and he had no worries about his repertoire at all. He was a one-trick pony.

Wilhelm was no ordinary rookie - he was 29 years old, for one thing. His career had been interrupted by World War II, where he'd won a Purple Heart, and he'd been stuck in the Giants farm system for four years. He threw a knuckleball, which wasn't regarded as quite so freakish a thing as it is today. And he was very, very good. As a rookie, he went 15-3, 2.48 - he was second in RoY voting (behind Black), and fourth in MVP voting.

Wilhelm was the first truly great career reliever, but this is the period when that first true generation of great career relievers begins to emerge. It all happened in the National League. In 1953, Roy Face came up with the Pirates. He started 23 games over his first two seasons (out of 83 appearances), but the Pirates concluded that his forkball worked best coming out of the pen. He would pitch another 14 years in Pittsburgh, and in 1959 put together a spectacular 18-1 season. Lindy McDaniel made his debut with the Cardinals in 1955 - he would pitch in 987 games over 21 seasons. Ted Abernathy came to the majors the same season - he would pitch in 681 games over 14 seasons and along the way would become the first man to save more than 30 games in a season. Don McMahon came up halfway through the 1957, and his brilliant work (2-3, 1.54 with 9 saves) helped the Braves win a World Championship. He would pitch in 874 games, all but 2 in relief, over the next 18 years.

There were all kinds of relief pitchers having utterly spectacular seasons, pitchers of all shapes and sizes and types. Luis Arroyo, a 35 year old screwball artist had one sensational (and famous, because he did it with the 1961 Yankees) year of glory. Ron Kline and Mike Marshall also relied on the screwball. Eddie Fisher, who had several outstanding seasons as Wilhelm's teammate with the White Sox, was a knuckleballer like Hoyt. The White Sox acquired yet another knuckleball artist towards the end of the 1960s, Wilbur Wood, who put in several brilliant seasons as a reliever before becoming their ace starter.

Because these guys were so effective, and because the whole business of having an outstanding pitcher on the team who always pitched in relief was a new thing, baseball ventured into some uncharted territory. No one knew how much these pitchers could, or should, be used. There was only one way to find out - the old-fashioned way. If you dunk the woman in the water and she drowns, at least you know she's not a witch. If you work a pitcher so hard that his arm is destroyed... now you know how hard is too hard.

They called Dick Radatz "The Monster" - partially because he was 6'5, 235 but mostly because he scared the living crap out of AL hitters for three totally awesome seasons. In 1962, a 25 year old rookie, he pitched 124 IP in 62 appearances. He won 9, saved 24, posted an ERA of 2.24 and fanned 144 hitters. He was just getting warmed up. The next year he went 15-6, 1.97 with 25 saves, striking out 162 in 132.1 IP. In 1964, the Red Sox sent him to the mound 79 times. For the third year in a row, he averaged exactly two innings per appearance. He was simply amazing, striking out 181 in 162 IP, winning 16 and saving 29. But he simply lost his effectiveness the next year, and it never came back. The Red Sox quickly traded him to Cleveland for Lee Stange and the aging, but still effective, Don McMahon.

By now baseball people could have been beginning to suspect that while relievers as a species were probably inherently unpredicatable, the hard throwers tended to only have two or three good years as relief aces. The trick pitch guys - Face with his forkball, Wilhelm with his knuckler - were still going strong. Lindy McDaniel was a finesse pitcher, and so was Clem Labine who was the Dodgers top reliever for the rest of the decade after Black had flamed out. Ted Abernathy was a submariner. The Giants had made little Stu Miller a full time reliever towards the end of the 1950s, and his changeup had made him an All Star. Don McMahon was a hard thrower who had lasted, but he was the exception. But only twice in his long career did McMahon pitch more than 100 innings in a season - in 1964 when he was already 34 years old, and three years later in 1967, when he worked his career high 109.1 IP.

But it would take another ten years for anyone to actually show signs of figuring this out. Meanwhile, Radatz had gone the way of Joe Page and Hugh Casey. Ryne Duren, another flamethrower, only had two good years.

Bill James, who has tread this ground on several occasions (see both editions of the Bill James Historical Baseball Abstract), identified five patterns in the usage of the relief ace since the specialist first truly emerged in the 1930s. He went on to try to develop an optimal usage pattern - which he figured would work out to be about 69 games and 113 IP per season, a workload he asserted would not destroy anybody:

In all honesty, I can't see one iota of evidence that a workload of 70 games, 130 innings is dangerous to a reliever. Goose Gossage had several seasons working that way, and he threw 95 until he was 40.

I'm not so sure about that. Gossage did have three seasons with 133 to 141.2 IP (1975, 1977, 1978); he pitched 70 games once (1977). But from 1979 forward, when he was still just 27 years old, his workload was cut back significantly. Only once during the rest of his long career did he work more than 100 IP (102.1 with San Diego in 1984.) I think that's why he had some gas left when he was 40. I suspect that the only pitchers who can deal with that workload are the ones who don't throw as hard as Gossage, or Radatz, or Page.

The overall trends follow the same well-established pattern. But the advent of the DH in the American League sticks out like a sore thumb, does it not? The DH was instituted largely because offense in the AL had reached a depressingly feeble 3.47 runs per game in 1972 - the additional bat in the lineup goosed that figure up to 4.28 in 1973. And, freed of the duty of automatically pulling the starting pitcher for a pinch hitter, complete games in the AL went up. Al managers had been tracking NL managers quite closely in this regard - in 1972, AL pitchers completed 502 of  1858 games (27.0%); NL starters completed 507 of 1860 (27.3%). But in 1973, AL piitchers completed 614 games, while complete games continued to fall in the NL (down to 447), and this trend would hold for the next few years. AL starters completed 650 games in 1974, the most in either league since 1921. Expansion, and the longer schedule helped that out a bit - still, AL pitchers finished one of every three starts (33.4%), and that hadn't happened since 1954. And as the complete games went up, the use of relievers went down. In 1976, AL starters completed 30% of their games, and AL relievers made 2366 appearances. In the NL, starters completed 23% of the game, and the bullpen was called upon to make 3120 appearances.

Forty years earlier, it had been a couple of respectable teams that had been the first to cross the one-reliever-per-game barrier. In the 1960s, the first teams to average two relievers per game would appear. However, without exception these were very, very bad teams.  The first of all was 1964 Kansas City Athletics. They were indeed terrible, going 57-105 and they had the worst pitching in the major leagues. In a league that posted a 4.06 ERA, the A's team ERA was 5.13 Their best starter was Orlando Pena, who went 12-14, 4.43 (ERA+ of 87) and completed just 5 of his 32 starts. They did have a few decent relievers, however - Wes Stock pitched very, very well in fact. Moe Drabowsky after washing out yet again as a starter (he went 3-11, 5.82 in 21 starts) was actually adequate in the bullpen. Ted Bowsfield was a fairly servicebale left-hander. And John Wyatt was both effective and resilient. Wyatt appeared in 81 games in 1964 - he was the first pitcher to crack the 80 game barrier, and the first pitcher to appear in more than half of his team's games in the 20th century.

At this point in history, only a handful of pitchers had ever appeared in even 70 games (beginning with Konstanty in 1950) and the fact that Wyatt was able to do so and maintain his effectiveness did seem to encourage the others. He held his record for exactly one year - the very next season Eddie Fisher of the White Sox established a new AL standard by pitching in 82 games, while Ted Abernathy raised the major league bar to 84. Wilbur Wood would appear in 86 games in 1968, and in the final year of the decade Wayne Granger would cross the 90 game barrier.  From this time forward, 70 and 80 appearance seasons have been as common as the grass.

Bill James has suggested that there were three decisive moments in the development of the modern relief ace. The first is the definition of the Save rule, which dates to 1969. This new statistic had very little impact on how managers actually used their relievers for the next two decades or so. What the rule did was describe how managers tended to use their ace relievers - a key component of their job was to save the game. (Today, of course, almost every manager is a slave to the Save - the demands of their players, who actually do need the stat when it comes to negotiate their next contract, now drives the game strategy.)

The second moment was the Bruce Sutter phenomenon. Through August 1 1977, Sutter pitched at roughly the same level Mariano Rivera reaches in the post-season. He had appeared in 47 games, and worked 85 IP. He was 5-1, 1.06 with 24 saves. There were still two months left, though, and the Cubs had to shut him down for most of August with a sore shoulder. He returned near the end of the month and was fine, if not otherworldly for the rest of the season. Next season, the same thing happened. Through July 31 1978, Sutter was absolutely untouchable - in 41 appearances, he was 6-3, 1.79 with 17 saves, and had fanned 72 hitters in 65.1 IP. But over the last two months, he went 1-7, 5.94. In 1979, he held up through August. The conclusion drawn by manager Herman Franks was that the demands on Sutter needed to be reduced.

James' third moment is the emergence of the first pitcher whose specific job was to fill that setup role. That pitcher was Ron Davis of the Yankees. Simply having a relief ace was no longer enough. Multiple relievers were required.

This had been in motion for some time. The great teams of the 1970s - the Reds, the A's, the Orioles - generally had deep and talented bullpens, rather thana pen with one star and a bunch of innings-eaters. Rollie Fingers generally led the A's in saves, but he always had excellent help from Darold Knowles or Bob Locker or Paul Lindblad. Anderson's Cincinnati bullpens always ran deep. Wayne Granger was used almost like a modern ace when Anderson first arrived, but Clay Carroll pitched just as often and just as well. Pedro Borbon arrived in 1972, Rawly Eastwick and Will McEnaney a few years after that. The burden was spread amongst them all. Weaver's starters were so good that he didn't have quite the same requirements from his pen - nevertheless, he generally had at least three relievers getting steady work - the closest thing to an ace was probably Eddie Watt, but its hard to see how he was all that more important to the team than Pete Richert, and third man Dick Hall was used often and effectively. They would eventually be replaced by a revolving door of effective relievers - Grant Jackson, Bob Reynolds, Dyar Miller, Tippy Martinez, Don Stanhouse - Stanhouse was the closest any of them came to fitting the idea of a relief ace, and this was as much because he seemed to have the personality for the job.

4. The Modern World (including LOOGYs Gone Wild, More Expansion, and Better Living Through Chemistry): 1980-2007

(Author's Note: This is about where I run out of steam!)

The lesson of Bruce Sutter was that two good relievers were better than one, even one great one. With remarkable speed, almost every major league team identified this role on the staff - the man who would "set up" the "closer."

The first team to average three relievers per game were the 1995 Rockies. They were the only team of the last millenium to do that. But since the new century dawned, no fewer than 35 teams have averaged 3 relief pitchers per game.

Something that emerges very clearly from the Data Table posted just below is the increased use of secondary relievers. The relief ace has completely vanished from the picture. Since Bobby Thigpen's 57 save season in 1990, when he led the AL with 77 appearances, the only closers to lead the league have been Rod Beck in 1998 and Billy Koch in 2002. This trend got up a head of steam in the 1980s: we saw Mark Eichhorn lead the league in appearances while setting up for Tom Henke, and Chuck Crim lead the league twice while setting up for Dan Plesac.

Some Scattered Observations

It's in this most recent era that the trend of using ever more and more relievers have taken off. Its not entirely because of the still continuing reduction of demands on starters. That has continued, of course. But

One of the things driving the use of more and more relievers is the realization that there are some pitchers who are of marginal or no real use when pitching more than a couple of innings, but who are useful when they just pitch one inning at a time.
- Bill James

Consider, for example, Mariano Rivera - there are those who have suggested that an argument can be that Mariano Rivera is the most effective pitcher who has ever lived. The gist of the argument, as our old friend Craig Burley put it. is that "statistically, he's harder to score runs off of than any other pitcher, ever...And that's a pitcher's job - to make it hard to score runs off him."

I don't think that's the pitcher's job, so I can't come to the same conclusion (you'll note if you check the link that Craig made it very clear that he was talking about a pitcher's "quality," not his "value.") But value is what greatness means to me. A pitcher's job is not to make it hard for the other team to score runs - a pitcher's job is to help his own team win. Obviously, being difficult to score on is a pretty big component of that job. But a pitcher helps his team more if he extends that difficulty over many innings. Doing that makes him more valuable to his team. Which, as far as I'm concerned, makes him better. But that's another discussion altogether.

Anyway, Rivera hasn't done that. He was a starting pitcher, and quite a good one actually, in the minors until he was 25 years old. When he came up to the Yankees in May 1995, he went into the rotation... and didn't do all that great. He went 3-2, 5.40 in 8 starts - one of them was brilliant (8 scoreless innings against Chicago), a couple were pretty good, and the rest... well, not so much. He was pulled from the rotation, made two more spot starts over the rest of the season (allowing 9 ER in 10 IP) and hasn't started a game since. And as well as he's pitched as a reliever since then, why would he? For what it's worth - which probably isn't all that much, as Rivera has surely been a better pitcher than he was when in the minors - Rivera he has given up fewer hits and homers to AL hitters, while striking out more of them, than he did when facing minor league hitters.

Rivera may well have developed into an effective starter - at this point, we'll never know - but we may be forgiven for doubting if he would have been as effective as he has been as a reliever. Like many great relievers, but not that many great starters, Rivera is essentially a one-pitch pitcher. It's a fabulous pitch - a cut fastball that he throws around 95 mph with control. But it's unlikely that he could throw his pitch that hard for more than two or three innings at a time at most - and even if he could, one-pitch pitchers rarely succeed as starters anyway, when they have to take their one pitch through the lineup three times or more. (Knuckleballers are the exception, but the knuckleball has the happy tendency, for its practitioners, of generally not repeating itself in any predictable fashion.)

Because the requirements of the two jobs are so different, I don't think it's really possible to compare starters to relievers in terms of "quality." I think it's a bit like comparing the defensive abilities of shortstops and left fielders. Was Barry Bonds a better defensive player than, say, Derek Jeter? In one strictly limited sense, yes - because Bonds was a great defensive player at his peak and Jeter was not.. But the fact that Jeter could actually play shortstop, even if he couldn't play it anywhere near as well as Bonds played leftfield, made him more valuable defensively than Bonds. The greatest defensive player in the game can not be a left fielder, practically by definition. I think the same this also applies to relief pitchers. And pinch hitters, while we're at it.

Anyway, here is the fundamental insight underlying it all: there are actually lots and lots of pitchers with clear limitations and shortcomings who can still help your team, sometimes a very great deal, if their usage is strictly rationed. This has possibly been the dominant strategic development of the past quarter century. The great lesson has been that there is no shortage of useful pitchers.