September 24, 2010
The Half Giants
On August 22, the Rockies languished 11 games behind the first-place Padres in the National League West. Tommy Bennett kept the faith despite the team’s predicament, but aside from his beard (whom I’ve anthropomorphized since our first encounter), few observers shared his belief in a rosy denouement in Denver. With a double-digit deficit in the loss column and just 39 games to play, the Rockies appeared to have gotten themselves into another nice mile-high mess.
Of course, we know what happened next: fast-forward less than four weeks, and the standings look quite a bit different: on September 18, the Rockies trailed the Padres by a single game. In retrospect, maybe we should have seen this coming. After all, the Rockies have made a habit of reeling off late-season victories in the last few years. In 2007, they famously won at a .613 clip to force a 163rdgame and earn a spot in the NLDS after a .500 first half. In 2008, the team finished with a losing record, but after going 39-57 before the All-Star break, played at a .530 pace in the second half. And finally, last season’s Rockies finished the first half six games over .500, but played .608 baseball the rest of the way. Earlier this season, the club was even criticized for its apparent unconcern, ostensibly arising from the confidence imparted by its players’ past second-half successes. Case closed, right? The Rockies are a second-half club.
Well, not exactly. Despite their recent winning ways, the Rockies have mustered only a .524 winning percentage in the second half, after turning in a .557 first half; we can’t very well affix the “second-half team” label to a group that won more frequently before the ides of July. Granted, since that August 22 low point, the Rockies have gone 20-9, including a 10-game winning streak from September 3-12 (which partially coincided with a 10-game Padres losing streak. Maybe the Rockies' second-half mojo merely took a little longer to kick in this season, though we can’t blame the departure of Brad Hawpe, a decidedly first-half kind of guy during the Rockies’ ’07 and ’09 second-half surges.
Or maybe it was never there to begin with. As a species, we’re wired for pattern recognition, and three consecutive seasons of elevated second-half performance seems meaningful. But should it? The Rockies aren’t the first team to garner this sort of split-season reputation; in fact, the Athletics earned a similar notoriety earlier this decade. From 1999-2006, the A’s upped their performance in the second half, with improvements of at least 130 points of winning percentage in 2001, 2002, and 2006. Of course, considering the amount of turnover associated with even the most successful teams, it shouldn’t come as a surprise that the ’99 and ’06 A’s had little in common, at least on the field: only the presence of green-and-gold uniforms and Eric Chavez tied them together. Without subscribing to the belief that GM Billy Beane consciously constructed his clubs of that era to succeed late in the season, it’s difficult to pinpoint a cause for such an extended run of second-half success.
The recent Rockies and A’s are not the only—or even the most noteworthy—examples of teams that excelled during one half of the year for a series of seasons. The following table lists the teams whose first and second halves were separated by at least 100 points of winning percentage (in the same direction) for a minimum of three consecutive seasons:
The Mets entry here isn’t entirely legitimate, since it includes the 1994-95 strike. If we lower the differential threshold from .100 to .080, our list increases to 11—and this decade’s Tigers' streak gains a fifth season. As Rob McQuown observed in the process of collecting this data (with assistance from Eric Seidman), Tigers fans have had relatively little to cheer about after the All-Star break for quite some time:
The Tigers haven’t improved in a second half since 2000. Is that meaningful or predictive? Probably not, but it can’t have been much fun for Tigers fans. Rob also checked into two teams whose second-half fortunes he suspected might be adversely affected by weather and scheduling. Since 1974:
Are these second-half declines the result of Arlington’s heat and the Cubs’ preponderance of day games? It’s not out of the question, but as we’ll soon see, other factors could be responsible.
The existence of “second-half teams” might not seem completely unreasonable; after all, furnish one club with enough second-half players, and it’s bound to outperform its early-season record later on in the year. The only flaw in that logic is that “second-half players” don’t seem to exist. A few might be lurking around the league—perhaps we won’t have conclusive proof that Mark Teixeira isn’t an inherently slow starter until his career performance in April surpasses the Titanic’s—but not in sufficient numbers to produce any significant league-wide correlation between first- and second-half splits. In light of the scant evidence to suggest that such players roam major league clubhouses, a team full of split personalities would likely be harder to find than a natural hair on William Shatner’shead.
Let’s take a look at the individuals responsible for the Rockies’ blistering second halves in 2007. That season’s turnaround owed much to the improved plate performances of Matt Holliday, Garrett Atkins, and Troy Tulowitzki; on the mound, Josh Fogg and Aaron Cook (before injury struck) turned in better post-break production. In 2009, Tulowitzki again caught fire after May, but Holliday had gone off to St. Louis (via Oakland), Fogg had been consigned to the bullpen, Atkins’ second half was nearly as subpar as his first, and Cook’s fortunes took a turn for the worse. Even though the results were similar, the cast responsible for the turnaround was composed largely of different characters.
If we can’t credit the individual players involved with a tendency toward improved performance in one particular half, and the likelihood of intelligent second-half design seems low, how can we account for the existence of these seemingly perennially skewed teams? We could credit someone in the clubhouse with incredible powers of motivational speaking in closed-door meetings—Yorvit Torrealba, let’s say—but it might be wiser to accuse the same culprit we sometimes blame for sky-high BABIPs, abysmal records in one-run games, and unusually poor run support: chance.
A simple Monte Carlo simulation reveals how drastically records can fluctuate as a result of nothing but random variation. Let’s take the example of a true-talent .500 team. If this perfectly average squad could play a million games in a league full of .500 opponents without its players succumbing to age or injury, it would win 50 percent of them. However, over a smaller sample, that .500 club could put forth the same average performance, but see the results of a much more or less talented team. In 10,000 trials, using a mean of .500 and a half-season’s standard deviation in winning percentage of .028, I got a maximum value of .608, and a minimum value of .390.
If we simulate first and second halves for a thousand .500 teams, using a coin-flip model, we find a wide range of outcomes: the fictional team that showed the most improvement during its imagined second half won 19 more games in its second 81 games than it had in its first 81, while the reverse was true of the team that experienced the greatest decline. Of course, our theoretical true-talent .500 teams don’t ever face easier schedules in one half, and they don’t make any theoretical acquisitions at the trading deadline (the Yankees, perhaps the team most willing to add payroll mid-season, have averaged a winning percentage 48 points higher in the second half than the first half over the last decade, though that may have nothing to do with their willingness to deal, and the weakened state of the sellers whom they play post-deadline).
Actually, it wouldn’t be unreasonable to expect the Rockies to show some tendency to excel in the second half, given Coors Field’s other anomalous effects—perhaps the Rockies’ new acquisitions or veterans returning from offseasons spent elsewhere play better after taking some time to acclimatize. Since 1993 (excluding 1994), and without any adjustments for roster composition or strength of schedule, the Rockies have added an average of 21 points of winning percentage in the second half.
The Rockies’ latest run, spearheaded by Tulowitzki and Carlos Gonzalez, has put them in contention as we prepare to enter the season’s final week, but their chances of making Rocktober last more than a few days remain under 10 percent. Still, as Padres supporter Marc Normandin told me earlier, no NL West fan who’s followed the last few seasons will be entirely comfortable until the Rockies flatline—and even then, a Zombieland-style double-tap might be in order. Fortunately for the Rockies’ rivals, the team’s former undead acts probably aren’t predictive of their present or future performance; despite the Rockies’ ballpark sponsor, it won’t take a silver bullet to kill them, only the usual blend of pitching, hitting, and defense.