Revisiting the Grady Little Game, 15 Years Later

Fifteen years ago Tuesday, a slight right-hander who wasn’t far removed from being the best pitcher his sport had ever seen wound up and fired for the 123rd time on a late Thursday night in New York. Despite what would happen on the play, the pitch was well-paced and well-placed: 95 mph at a time when 95 still made a person perk up, and either on or just off the inside corner. On a 2-2 count, it was too close to take safely, and trailing by two with one out in the eighth and runners on second and third, the batter opted not to leave his fate up to the umpire. Even the swing didn’t portend trouble: The jammed batter broke his lumber and lifted a looper that was classified as a fly ball but barely left the infield. No AL pitcher who worked exclusively as a starter that season induced a higher percentage of the hitters he faced to strike out or pop up, and either outcome easily could have happened here.

2018 MLB Playoffs

Check out all of The Ringer’s coverage of the baseball postseason

Instead, the ball landed in shallow center between a trio of converging fielders. The Yankees’ Jorge Posada had doubled off of Red Sox starter Pedro Martínez to tie Game 7 of the 2003 ALCS at five runs apiece. Three innings later, the Yankees would walk off on an Aaron Boone homer to extend Boston’s streak of seasons without winning a World Series to 85. And 11 days later, the Red Sox would decline to exercise their option on Grady Little, the manager who had made the fateful decision to let Pedro keep pitching.

Some things in 2018 aren’t so different from 2003: The Red Sox are in the ALCS, Pedro remains a prominent postseason presence, and managers still make maligned decisions about how long to stick with starters in Yankees–Red Sox playoff games. But Pedro is now a TBS analyst, Boone is the manager whose decisions are being savaged, and those crucial managerial junctures in important playoff games are arriving more and more often in the third or fourth inning instead of the eighth. A baby born on October 16, 2003, would be a high-school student today, as I was when I sat—or, during the eighth inning, stood—in the upper deck in right field at the old Yankee Stadium and watched the Red Sox unravel. High school doesn’t feel like it was that long ago, but the Grady game seems like an artifact from an earlier era, one in which managers’ minds and pitching moves bore little resemblance to the way they work today.

With the pennant and the upper hand in a historic rivalry at stake, Martínez had cruised through the game’s first six innings with only one blemish on his line, a leadoff homer by Jason Giambi in the fifth. His first jam arose in the seventh, when he got a groundout and a Posada lineout—hit much harder than the bloop double would be—to start the inning, then allowed another Giambi blast and back-to-back singles to, of all opponents, Enrique Wilson and Karim García. (Wilson, weirdly, was a well-known nemesis of Pedro’s, which is why he started over Boone. The extremely light-hitting utility man boasted a 1.167 OPS against Pedro in 23 plate appearances through the 2003 regular season, but he went 2-for-12 with two singles thereafter, which made much more sense.) With two on and two out in the seventh, righty reliever Mike Timlin was ready to enter.

But Martínez fanned the free-swinging Alfonso Soriano to end the threat and preserve the 4-2 lead. On his way back to the dugout, the starter signaled to the sky, believing his day was done. He had thrown exactly 100 pitches.

David Ortiz homered in the eighth to give the Red Sox some extra insurance, and as Martínez later told Tom Verducci, Little asked him to return to the mound for the eighth. On a TBS segment last week, Pedro implied that he expected to face only one batter, the oft-injured Nick “OBP Jesus” Johnson. Johnson was a lefty, but he was 3-for-7 that season against Sox southpaw Alan Embree, which may have signified something to Little. Pedro could get anyone out—his career platoon split spanned just 29 points of OPS—and whether Little knew it or not, Johnson wasn’t susceptible to lefties. (He retired with one of the largest-ever career reverse splits for a left-handed hitter.) It wasn’t a terrible matchup, and it went Boston’s way, as Johnson popped to shortstop on the seventh pitch he saw. Pedro had gotten his hitter.

Timlin and Embree had been up in the bullpen before the first out. Still, Little didn’t budge, electing to let Martínez face the right-handed Derek Jeter, who doubled (thanks in part to a Trot Nixon misread in right) on an 0-2 fastball that was lower than neck level, where Jason Varitek wanted it.

Pedro’s pitch count stood at 110, with five consecutive switch-hitters or lefties due up. It was an obvious Embree opportunity, but again Little left him in. “You get the feeling that [Embree] will be the pitcher against [Hideki] Matsui one way or the other,” Tim McCarver said in the Fox broadcast booth as Bernie Williams batted. Williams lined a single to center to make it 5-3, which finally brought Little out to the mound—not to pull Pedro, but to ask him whether he had enough left to face Matsui. “A proud man, a proud baseball player, a proud pitcher, never really wants to give up his sword,” Martínez said last week. “I was a wounded warrior, but I wanted to continue to fight.”

Matsui got around on another 0-2 fastball, this one in roughly the same spot as the impending pitch that would yield Posada’s double. Matsui’s double touched down a couple feet fair and bounced into the stands, setting up the equalizer that would finally knock Pedro out of the game. As the crowd roared and the foundations shook, Little trudged out to the mound, head down. He belatedly brought in Embree, who got one out and then passed the baton to Timlin, who held the Yankees scoreless through the ninth.

“This is the most blatant situation for a second-guess in this series, whether to bring Embree in to pitch to Matsui,” McCarver intoned on the telecast after Matsui reached second. That might have been true if much of the audience—including Sox GM Theo Epstein and team owner John Henry—hadn’t wondered what Little was thinking well before his slow hook stabbed Sox fans through their frequently perforated feelings. “I reacted to it exactly the same way that everybody else reacted to it,” remembers Bill James, who was then a rookie Red Sox adviser watching from his home in Kansas. “I make it a point of pride never to second-guess the manager. Not only not publicly, of course, but not privately, either. It does a lot of damage in an organization if people get in the habit of second-guessing the manager. Part of the damage that that inning did was that you couldn’t AVOID second-guessing it.” Although Epstein said the decision to let Little go was “made on a body of work after careful contemplation of the big picture” and that it “did not depend on any one decision in any one postseason game,” it’s unlikely that the Sox would have cut ties with a skipper who’d presided over a combined 194 wins (October included) in his first two seasons without some precipitating incident.

Avoidable as it was, it’s easy to see how that incident developed. As Joe Buck said while Jeter was sauntering to the plate in the eighth, Martínez was “regarded by many as the greatest pitcher in the game today—the most dangerous, the most dominant.” According to a park-adjusted version of Baseball Prospectus’s flagship pitching statistic, Deserved Runs Average, Pedro’s 2001, 1999, and 2000 campaigns are the three best seasons ever on a rate basis by a pitcher with at least 100 innings pitched. In 2003, then, Pedro was coming off three consecutive seasons better than any single season that any other human has ever had, followed by 2002 and 2003 regular seasons in which he recorded ERAs in the 2.20s and placed second and third, respectively, in AL Cy Young voting. When he took the mound in Game 7, Pedro was past his peak but not really past his prime: He was only 31, and according to DRA had been baseball’s fourth-best pitcher that year. Nor was he finished pitching at an elite level: 2004 would be a down year by his unassailable standards, but in 2005, his first season with the Mets, his DRA ranked second in the majors.

Little’s access to an unparalleled ace was part of the problem; the other was a bullpen that recorded the game’s third-highest ERA in 2003. The Red Sox chose to prioritize other parts of the roster over the winter of 2002-03, and the result was a disastrous “closer by committee” that reigned early on more by default than by design. As the summer wore on, the pen improved to the point that it posted baseball’s best park-adjusted FIP in the last month of the regular season, thanks in part to midseason trades for Scott Williamson and Byung-Hyun Kim, the latter of whom Little insisted on anointing as the team’s closer, a role in which Kim had success. Embree, Timlin, and Williamson had combined to allow only one run in 22 2/3 postseason innings prior to ALCS Game 7, but Little’s mind was made up. He’d seen the bullpen blow Boston leads in six of Pedro’s previous starts that season, including his first playoff start, and he wasn’t going to watch it happen again in an even bigger game—especially because, as he claimed in a GQ feature the following April, he wasn’t sure whom he could count on to handle the pressure, having heard some Red Sox players sound psyched out by the club’s so-called curse.

In the wake of Boston’s loss, much was made of Martínez’s .370 batting average allowed between the 106th pitch and the 120th pitch of his starts during the 2003 season. (That seemingly arbitrary range was determined by the data provider STATS, which split pitchers’ performances into 15-pitch increments.) The Red Sox’s front office was well aware of those numbers, as was Little, who chose to disregard them. But Martinez made only 10 regular-season starts in 2003 in which he threw 106 pitches or more, and in those 10 games combined, he coincidentally threw a total of 106 pitches within that 106 to 120 range. (If STATS included ALDS Game 1, in which Martinez threw 130 pitches, that would bring the totals to 11 starts and 121 qualifying pitches.) In retrospect, batting average allowed in a small, single-season sample doesn’t seem like the soundest basis for decision-making.

We have better stats today. We can say, for example, that of the 219 pitchers on record who’ve thrown at least 1,500 total innings and at least 500 when facing hitters for the third time in a game, Martínez posted the 17th-best performance the third time through, relative to his overall line. Roger Clemens, Pedro’s Game 7 opponent, was 7 percent worse the third time through than he was overall; Bob Gibson and Curt Schilling were 10 percent worse; Steve Carlton and Greg Maddux were 11 percent worse; Clayton Kershaw has been 13 percent worse; Randy Johnson was 15 percent worse. Martínez was 3 percent better. In that sense, he was quite successful when he went moderately deep into games.

But when that wondrous/traumatic eighth inning began, Martinez wasn’t facing hitters for the third time in the game. He was seeing them for the fourth time. Pedro didn’t go that deep very often, and when he did, it didn’t go well. Of the 380 pitchers on record who’ve thrown at least 1,500 total innings and at least 150 when facing hitters for the fourth time in a game, Pedro ranks 14th worst relative to his overall line, with a 32 percent penalty in those situations. Granted, heavily penalized Pedro was still pretty good: In an 18-year career, he allowed a .711 OPS his fourth time through the order, which is lower than Embree’s career OPS pitching out of the pen. But he clearly wasn’t the same in that scenario. On a career level, his decline after the 100-pitch point was also slightly steeper than the typical pitcher’s, although that varied in small samples from season to season. (From 2001 to 2003, Pedro was way worse from pitch 101 on than he was overall, but from 2004 to 2006, he was way better.)

So no, newly accessible stats can’t contradict the conventional wisdom: Boston would have been better off with a fresh arm. But the degree to which the loss is typically pinned on Little is probably excessive. “I do think the regular narrative of that game misses something,” says sabermetrician Tom Tippett, who completed his first consulting project for the Red Sox in September 2003 and would go on to serve as Boston’s senior baseball analyst for 13 years. Tippett, who like James was also watching Game 7 from home, was surprised and dismayed when Pedro came back out for the eighth. “I was too nervous to sit, so I paced back and forth in front of the TV, muttering out loud things like, ‘I can’t believe he’s still in there,’ and ‘This isn’t going to end well,’” Tippett recalls. When Martínez blew the lead, Tippett felt as though the sequence had been foreseeable.

But in 2006, NESN asked Tippett to reconstruct the circumstances of Game 7’s eighth inning and simulate the outcome repeatedly using the sophisticated probabilistic model that powered Tippett’s long-running baseball simulation game, Diamond Mind. As he rewatched Game 7 for the first time to prepare for NESN’s “What If” special, Tippett was taken aback by how few major mistakes Martínez seemed to have made in the eighth and how hard he was throwing. Data from Sports Info Solutions—which combines four-seamers and sinkers—confirms that Pedro still had his heat late in Game 7.

“Tom-the-fan-in-real-time was sure Grady was wrong,” Tippett says. “Tom-the-analyst-two-years-later would still have preferred to see Grady go to the bullpen but had to acknowledge that Grady was right to think that Pedro still had plenty in the tank.” Tippett’s simulations suggested that the Red Sox would have had an 82 percent chance to win had they brought in Embree to face Matsui. But the sims also said that the Sox still had a 79 percent to win with Little leaving Pedro in. Granted, Boston’s odds might have been even higher had Embree started the eighth. And the computer couldn’t account for the fact that Little led Pedro to believe that his day would be done sooner; once he’d paced himself for a certain workload and exhausted his reserves to get to his goal, it was probably harder for him to dig deeper. Like Terry Collins in the ninth inning of Game 5 of the 2015 World Series, Little made a mistake by asking Martínez whether he wanted to stay in and expecting a transparent response from an ultra-competitive athlete. But Little’s sin wasn’t directly losing the game; it was making the Red Sox a little less likely to win. “We’d done it through the year, over and over and over,” Little told GQ, referring to trusting his ace. “We got a bad result this time. Most of the time, we got good results.”

Earlier this month, Martínez defended his former manager, saying, “I still believe it wasn’t [Little’s] fault.” But Little’s mode of decision-making wasn’t long for the league, and Epstein was likely somewhat sincere when he said that Little’s ouster wasn’t tied to one event. By his own admission, the manager played “educated hunches,” and he disregarded input from the front office. Earlier in Little’s career, Peter Richmond wrote in GQ, “Grady liked managing … as much as he liked farming. In both, you were pretty much your own man.” But managers aren’t their own men anymore, and Little’s dismissal marked a major step along the path to their current role as implementers of other people’s plans. “This ownership group prefers an increased reliance on thorough and more comprehensive analysis and preparation so that the manager’s decisions are more synchronous with our player-acquisition and development decisions,” team president Larry Lucchino said in October 2003, his buzzword barrage reminding readers that baseball was a business and that the dugout wasn’t at the top of the corporate totem pole. “We seek one unified organizational philosophy.”

Under Little’s successor, Terry Francona, they found it. “If Boston plays 170 or better games next year, then maybe the result of their decision was the right one,” Little said after the Red Sox cast him loose. In 2004, the Sox played 176 games and won a World Series (though not without another eighth-inning meltdown for Martínez). Little, whose media defenders included future basketball sidler Adrian Wojnarowski, later managed the Dodgers for two seasons, and he continues to serve as a special assistant in the Pirates’ front office. But his brand of believing in hunches and defying the data is almost extinct, and as a consequence, so are starts like Pedro’s.

In 2003, there were 231 outings of at least 120 pitches. This year, there have been 12, and Jake Odorizzi’s September 12 start is the lone example since the trade deadline. It’s not inconceivable that we could see another postseason start that long—Justin Verlander threw 124 pitches in a complete-game ALCS win against the Yankees just last year. This year, though, bullpens have pitched a record 48 percent of playoff innings (counting openers as starters). Among postseason starters, only Verlander and Hyun-Jin Ryu have exceeded the 100-pitch threshold, and neither reached 105. Clayton Kershaw is the only starter who’s seen the eighth inning. Most October box scores look more like the Yankees’ side of Game 7, in which no pitcher threw more than three innings. Teams today are smart enough to make starters surrender their swords. But those self-impaling pitchers may have made more memories by fighting for so long.

Thanks to Mark Simon of Sports Info Solutions for research assistance.