March 16, 2012
Extra Innings Excerpt
What is the Effect of the Increase in Strikeouts?
Extra Innings: More Baseball Between the Numbers, edited by Steven Goldman, is the sequel to Baseball Prospectus’s 2006 landmark Baseball Between the Numbers, a book that gave many their first taste of state of the art sabermetric thinking in the years after Bill James and Moneyball. BP now returns with a sequel that delves into new areas of the game, such as how to evaluate managers and general managers, the true effects of performance-enhancing drugs, how prospects are recruited and developed in Latin America, and more. The book is now available for pre-order from Amazon and Barnes & Noble and should ship ahead of its official release date of April 3, 2012. Today, we present the first of two excerpts from the book.
You are no doubt aware of the parable of the boiled frog. Toss the unfortunate amphibian into a pot of boiling water and he'll jump out of it. Put him in a pot of water and then slowly raise the temperature to boiling, however, and he'll stick around until he's cooked. Like most parables, it's something you hope people haven't put to the test in the laboratory.
Beyond conjuring up ingenious methods of amphibicide, one of the underlying messages of this particular story is that, if you don't pay attention, a massive change in your environment might be happening. You can see where environmentalists get all worked up about it, but our purpose here is a wee bit more esoteric. It's time to talk about one of the most fundamental changes to the game today, something that has slowly crept up on us, and perhaps on the game as a whole. Quite simply, it's time to notice that we're living in the Age of Strikeouts, and to take note of the effects of this change on which teams win and why, using which players, and how, and what this means for team-level success in the seasons to come.
Flash back to the 1960s, before and then on into the high-mound era that culminated in the “Year of the Pitcher” in 1968. In 1961, the point at which Roger Maris set the single-season home run record with 61, runs teams scored 4.53 runs per game. That dropped to 3.42 runs per game per team in '68. That's a 25 percent drop, where a quarter of runs scored for every team just went away. If you're old enough, you know the names of the pitchers who owned that time: Koufax, Gibson, Drysdale, McLain, even Lonborg. This environmental change helped the pitchers of this era not simply dominate, but generated their statistical impact on record books.
As they did it, one of the things that was part and parcel of their dominance was a spike in the strikeout rates in baseball. In 1955, strikeouts were 11 percent of all plays and 18 percent of all outs. By 1961 those numbers had already moved up to 14 percent of all plays and 22 percent of all outs—steep increases—but by 1968, the number of strikeouts had climbed all the way to 16 percent of all plays (a 45 percent increase across 14 seasons) and 24 percent of all outs (a 25 percent inflation rate).
That's an impressive historic anomaly, one done away with when the mounds were lowered, the strike zone expanded, and (eventually) with the introduction of the designated hitter. It's also a fraction of where strikeout rates are today, because when it comes to getting our results at home plate, we've long since left the Year of the Pitcher in the dust. As of the 2011 season, the rate of strikeouts has reached 19 percent of all plays—nearly a 73 percent increase from Mickey Mantle's heyday—and account for 29 percent of all outs, a 61 percent increase. Both marks are historic highs for baseball as it has been played since 1893, when pitchers had to move out to 60 feet, six inches from 50 feet.
Now, that's obviously a major environmental change, and you can probably guess at a lot of the contributing factors. Beyond the lower mound or fewer at-bats going to pitchers attempting to hit thanks to the DH, if you've a mind to complain you'll bring up more hitters swinging for the fences, or perhaps the endless tedium of an ever-growing number of relievers putting in appearances in games. If you like parsing pitch data, you'll probably jump to the conclusion that the increasing popularity of the splitter in the '80s, the changeup in the '90s, or the cut fastball today has something to do with it. Add in video scouting and the contemporary wealth of data, and you might well guess that it's easier than ever to get a strikeout.
This is no doubt pedantic, but keep in mind what a strikeout actually is. Unless you've got a particularly infamous backstop routinely letting guys reach on the strikeout-plus-passed ball or wild pitch play, it's a high-yield event in terms of guaranteeing an outcome: an out. Set against that, you've got an overall decline in the number of balls in play. We'll talk about what that means for the declining significance of defense on a team at greater length shortly, but what has essentially happened is that teams have traded high-percentage play, settling accounts at home plate with a whiff, for fewer total chances for the defense.
On average, it follows that as strikeouts have gone up, the number of balls in play have gone down over time. In 1968, 73 percent of all pitcher-batter outcomes wound up in play, a figure that with the DH batting in one league instead of both nevertheless moved upwards of 76 percent in the 1980s. Today, the number of balls in play has dropped to 69 percent in the 2011 season. Again, that's without the higher mound, and with the DH batting in only one of the leagues. It may not sound like much of a difference, but this represents a fairly dramatic change. We've moved from relatively low-yield events for both the batter and pitcher—defensible balls in play that may or may not drop for hits—to a growing number of plays in which the defense has next to no impact on whatsoever: home runs and strikeouts.
Broadly speaking, Defensive Efficiency (DE) kept pace by going down. Going back to the 1950s, major league pitchers could count on 70 to 71 percent of balls in play to be converted into outs. In 1977, DE dipped below 70 percent for the first time since 1940, but that was just to 69.8 percent, but in the '80s it started heading down below 70 as often as not, only briefly bouncing back in 1988-1992. But defenses hit a new low in 1993, an expansion season, with a DE of .693.
Since that year, DE has never gotten back over 70 percent, bouncing around in the 68 percent range in the '90s, in the 69s in the early Aughties, dropping back around 68 percent in 2006-2008 (the immediate aftermath of the so-called “Steroid Era”), and only just moving higher than that 1993 game-wide mark in 2011. For all the attention given to defense in recent years, it's nice to see some small result, but for all of the attention focused on the subject since the Rays' big improvement on defense from 2007 to 2008, defenses in general simply aren't turning as many balls in play into outs as they did in 1992.
Put all that together in terms of what's changed in the last 20 years:
Relative adjectives like “bad” and “good” don't mean much here—we can't say all fielders in 1992 were better, but with a thorough command of the obvious we can say they were converting more balls in play into outs. Subsequently, pitchers have helped themselves out by settling accounts at home plate, which lowers the overall importance of defense, especially at the positions with the lowest likelihood of seeing a play, like the outfield corners, but generally every position. Indeed, in today's game, you'll find that some teams have headed in the exact opposite direction from the much-ballyhooed lesson of the Rays, punting defense and winning with strikeout-dependent staffs that don't need a top-notch defense to keep runs off the board.
Keep in mind, whatever the available metrics suggest about a player's value relative to his peers (the way a curve-driven metric like UZR suggests) or in terms of an absolute historical standard (like the Fielding Runs calculated by Clay Davenport or Sean Smith), even a bad big-league defender is among the best players at the position on the planet, indeed, in baseball history. Derek Jeter might be “terrible,” and simultaneously playable. Why? Because strikeouts are such an increasingly large part of the game.
That adaptive realization has already had an impact on today's game. Brian Cashman didn't despair over the fact that Derek Jeter has become a living monument at short. Instead, the Yankees' GM used the power of his bankroll to do something after the 2008 season that minimizes the impact Jeter can have on defense: he signed the two best strikeout starters available on the market: The best power lefty workhorse active in the game today in CC Sabathia and the American League's reigning strikeout leader, A.J. Burnett. Sabathia's combined 2008 strikeout rate with the Indians and Brewers was 24.5 percent; Burnett was only slightly lower (24.1 percent), but he had topped that with a 25.5 percent clip in 2007.
Few teams could afford that sort of expense, but at a time when Derek Jeter's awful defense was already the stuff of sabermetric canards and snickering references to whatever advanced fielding metric you care to place your faith in, Cashman did. The payoffs were immediate. The Yankees saw their team-wide balls-in-play rate in 2008 drop from a below-average 70 percent to an American League-low 66 percent. Their team-wide strikeout rate jumped from 18.5 percent to a league-leading 20.1 percent.
With lightly-regarded defenders up the middle, such as the ancient, indifferent Jorge Posada behind the plate, plus Melky Cabrera's wanderings around center, and Jeter's status as one of the oldest men to ever play short on a contender, the Yankees bludgeoned their way to 103 wins and 915 runs scored, clubbing people to death at the plate while stranding them there on defense.
Come October, the Yankees didn't need depth in the rotation, using just Sabathia, Burnett and old standby Andy Pettitte to start all 15 of their postseason games. When they beat the Phillies in the 2009 World Series, it wasn't especially elegant as Burnett and Pettitte struggled, but they still managed to get 50 strikeouts in 53 innings, avoiding that defense en route to the win. How did that happen? Not least because the relief corps, armed with Joba Chamberlain, Phil Hughes, David Robertson, lefty Damaso Marte, and the great Mariano Rivera got 11 of their 32 outs on Ks in their four wins over the Phillies.
Maybe the 2009 Yankees used strikeouts to compensate for a few weak defenders, and maybe they were just lucky, but what happened the following season, when the Giants won the 2010 World Series, suggests that the results were anything but random. Here again, you have a lot of the same elements. Aubrey Huff had been a semi-regular DH before joining the Giants; he played first base and a lot of outfield, in particular roaming around in right field in his starts against right-handed pitching during the Giants' hottest stretch of the season, during July and August before they picked up Cody Ross. Opposite Huff in left field was another discarded DH, Pat Burrell. The infield was manned by sub-mediocrities, featuring the then-bloated Panda, Pablo Sandoval, at third base, an aging Edgar Renteria at short, and the limited range of Juan Uribe and Freddy Sanchez up the middle.
None of that mattered all that much, because the Giants' pitching staff was busy leading the major leagues in strikeout rate, erasing 21.6 percent of all batters at home plate. They also managed an MLB-low clip for balls in play of just 65 percent, taking pressure of a defense stocked with the geezers for whom Brian Sabean is famous. Who was getting all those strikeouts? Tim Lincecum and Matt Cain and Jonathan Sanchez, certainly, but the pen could call on Santiago Casilla, who struck out a quarter of all batters faced, rubber-armed Sergio Romo who terminated 28 percent of all batters, and closer Brian Wilson inspired batters to fear the beard not by making them wonder if he'd pillaged the Gettysburg hair and makeup department, but by striking out 30 percent of the men who stepped in against him.
The Giants’ power out of the pen was critical in propelling them past the Phillies in the NLCS, where Bruce Bochy's frenetic pitching changes and playing matchup games helped produce 22 strikeouts in 19 1/3 relief innings in 21 individual relief appearances in six games, not to mention a series win. That set the Giants up to do the same thing to the Rangers, only much more handily.
This brings us to the next critical point about the Age of Strikeouts: we're far past the age of individual heroes.
Strikeouts, Strikeouts Everywhere? Isn't Nolan Ryan Retired!?!
You didn't need to watch Tony LaRussa's nightly do-si-do over the last three decades to notice the game-wide change in pitcher usage patterns. In 1969, teams averaged less than two relievers per game. In the '70s, that dropped to an average of fewer than three relievers total between two opposing teams in any individual game. Things picked up during the '80s with the increasing employment of situational relievers and assorted eighth-inning heroes; teams moved past two relievers per game apiece in 1989, then jumped past three relievers per game in 1995 in the wake of the strike, and cresting in 1998 with 3.2 relievers per team per game. Things have come down since, to around 2.9 relievers per game now, but all that means is that we're still sitting through almost three relievers appearing for every team in every game.
So, managers have adapted in monkey-see, monkey-do fashion, aping not just La Russa's usage pattern for Dennis Eckersley, but also ageless lefty Rick Honeycutt and power set-up righties like Eric Plunk. The process accelerated during the offensive boom between 1994 and 2006 as more runs scoring meant more relievers to cope with an increasing workload. They moved past pitching more than a third of every ballgame in 1995, generally staying there... until the last two seasons. We'll return to that in a moment.
Generally speaking, the benefits of using more relievers are clear: it makes it hard for the batter to get used to any one pitcher's stuff or pitching pattern, making the accrual of their season stats less and less an aggregation of 162 batter vs. pitcher matchups, and more a matter of two, maybe three at-bats against the starter, then one against some guy you may have never seen before, and then one more against an ace reliever you see twice a month. Even a batter who manages to make contact is not guaranteed to make hard contact.
As chicken/egg things go, you can wonder whether the present environment makes record-setting strikeout fiends at the plate like Mark Reynolds possible. And while Reynolds' career 33.2 percent strikeout rate tops that of Rob Deer (31.2 percent), but in the context of their eras, Deer's prodigious whiffery at a time when strikeouts were much less common makes his mark the more remarkable feat. It seems quaint to note that Reggie Jackson's career 22.8 percent strikeout rate could generate the all-time record for strikeouts with 2597. That record will fall to Jim Thome in 2013, assuming Thome decides to play; if not, Alex Rodriguez could put Mr. October in the shade towards the end of this decade, presumably while also gunning to knock Barry Bonds down a peg on the all-time home run list. But A-Rod and Thome have had to cope with something that Reggie never had to confront: A cast of thousands of pitchers instead of hundreds.
That change is directly reflected in today's spike in strikeout rates. Back at the start of the era of divisional play, strikeout rates for starting pitchers and relievers were essentially even, right around 5.8 K/9. Forty-three years later, starters have picked up almost a strikeout per nine without pitching as far into the ballgame.
However, the real key to the spike in strikeout rates has been the relievers, as they've reached 7.5 K/9 while pitching a larger portion of every ballgame.
So what about the recent drop in the amount of gametime relievers are pitching per game, what does that mean? Well, think about it: Teams are still carrying seven or eight relievers, but with offense down and starting pitchers going later into ballgames, that means that there is a shrinking amount of playing time going to the same number of guys in the bullpen. Relievers are making almost exactly the same number of appearances today that they did before the downturn in the last two years—but those appearances are growing shorter.
That trend was reflected in the Giants' World Series victory in 2010, when Bochy's aggressive use of his pen was much remarked upon. However, that aggressiveness was simply the paragon of a virtue reflected in game-wide trends for using relievers for increasingly short stretches to exploit every incremental advantage, matchup by matchup.
To The 2011 World Series
Tony La Russa and company had already made a point of risking their interior defense in 2009 by moving outfielder Skip Schumaker to second base to add an OBP source to the lineup in front of Albert Pujols and Matt Holliday. Schumaker was terrible at the keystone and has remained terrible despite his best efforts, but the Cardinals won the NL Central while Schumaker produced a .364 OBP as their most-regular leadoff man.
In 2011, still counting on Schumaker at second, the Cards defied any thoughtful concerns for their defense by adding Lance Berkman to the lineup to play right field. Berkman was 35 years old and seemed older still with his bulk. He hadn't played the outfield much since 2007, or regularly since 2004. Did the Cardinals care? Hell no. They got Berkman relatively cheaply for $8 million, parked him in right, and exerted season-long defensive indifference from the dugout and executive suite. They committed to the equally immobile David Freese at third base. They brought in Ryan Theriot and dared to move him back to short after the Cubs and Dodgers had deemed that unwise.
The penalty? They finished 23rd in the majors in Defensive Efficiency, led the league in scoring, and—with the help of making a concession to defense by adding shortstop Rafael Furcal at the deadline—raced to the wild card. However, they'd gotten to October without much of a strikeout staff, just 22nd with a 17.7 percent punch-out clip. The only above-average starters they had in terms of strikeout rate were Chris Carpenter and Jaime Garcia, but with both shy of striking out batters 20 percent of the time, neither compared to the Yankees' or Giants' wizards of whiffery in postseasons past.
Leave it to LaRussa to find the perfect way to compensate—some might say overcompensate—for the problem and making the term “ball control” relevant to baseball. After squeaking past the Phillies in the League Division Series, he chucked standard operating procedures for staff management. While he had used his starting pitchers normally enough in the LDS, but was forced to deploy Carpenter in the deciding Game 5. Wrong-footed in their rotation even as they advanced, LaRussa simply treated every game situation, and perhaps every batter, as a high-leverage moment that might demand a pitching change. Between the two rounds and the 13 games it took to win them, he used 58 relievers over 54 1/3 innings, getting 43 strikeouts while routinely exploiting platoon matchups. If Bochy's flurry of gutsy pitching changes and reliance on his pen in 2010 seemed inspired, La Russa's frenzied churn of relief help kept the Rangers guessing in a blur of changing matchups capped by Jason Motte's triple-digit heat.
Long-Term Implications of The Age Of K
That might seem obvious, but sabermetric groupthink seems stuck on esoteric notions like the underrated defensive value of left fielders, this at a time when fewer balls are in play, when the rate of how many balls drop safely have also declined, and when runs are nevertheless becoming more scarce. Sometimes it's worth according some measure of respect to old-fashioned mashers at the positions they're traditionally associated with, right and left and first base.
All too often, sabermetrics can be an exercise in documenting the previously observed. For more than a century, perhaps all the way back to the game's earliest days, skippers understood that fewer balls go to right and left field than anywhere else, so they were only too happy to stuff unglovely sluggers in the distant corners of the diamond to reap some offensive benefit. If folks start noticing that teams are winning the World Series with designated hitters in the field, we'll have come full circle from the elaborately exaggerated enthusiasm for the 2008 Rays and their defensive turnaround.
Instead, the lasting lesson from those Rays and life in the Age of Strikeouts is that bullpen assembly and management is more important than ever. As tedious as the endless shuffle of men from the pen might be to witness, and as much as many of us might lament the loss of a 14th or 15th position player and the impact of that on in-game offensive tactics, the recent success of bad defensive teams with deep pens and strong lineups might provide a template that keeps us exactly where we are for years to come.
In the abstract, the shrinking amount of gametime available to relievers as starters pitch deeper into games and run scoring drops ought to trigger a reverse of the industry-wide adoption of the seven- or eight-man pen. But every stathead from Bill James on down has been arguing for a relief reformation of rosters for a good decade now, and after the performances of La Russa and Bochy in the last two World Series, it's becoming something easier to emptily assert than really believe in.