Friday, June 17, 2011
AL projected records
By Christina Kahrl
An interesting thing to note as we near the midpoint of the season, at least in the American League, is that we're not seeing some of the extreme overperformances as far as teams' expected records that we have in recent seasons. Sure, we've seen some early surprises in terms of the Indians racing out ahead initially, but now that we're 70 games or so into the season, the simple fact is that nobody in the AL is coming close to outperforming their expected record.
The basic method was devised by Bill James more than 30 years ago, a simple Pythagorean method using a team's runs scored which allowed you to project that team's record. You might prefer to employ the subsequent Pythagenpat formula developed by David Smyth to refine the exponents and yield a more precise result, or refer to Clay Davenport's work on adjusted standings that evaluate a team's expected runs scored and allowed (to get to "second-order wins") and then take into account the strength of that ballclub's opponents to get to "third-order wins."
Consider the five teams outperforming their expected record so far this season, referring to the Pythagorean win-loss records at Baseball-Reference.com as well as Clay's projected third-order wins:
Generally speaking, whatever the methodology, the answers are generally the same. The Indians owe a lot to having one of the best marks in the league in one-run games (13-8, when something closer to .500 is what you'd expect), while the the Rays and Orioles get some benefit of the doubt via third-order wins because of a schedule that pits them against the toughest top-to-bottom division in baseball.
In the wild-card era, overperforming your expected record by a couple of games might seem fairly significant, but it's not as significant as we've seen in recent seasons. The Angels enjoyed a three-year run from 2007-2009 of outperforming their expected record on average by more than 10 games per season (via third-order wins). It simultaneously defied the analysts' expectation they'd regress from one year to the next while reflecting the number of virtues that Mike Scioscia cultivated as a manager: a solid rotation and an effectively deployed bullpen, as well as some of the best baserunning to play off a contact-driven offense.
Looking at this year's "overperforming" teams in the league, nobody's close to duplicating the Angels' feats, and no AL club's likely to win 10 or more games than expected at this pace, especially as the Indians' run seems to have shot its bolt. This strikes me as particularly interesting given that scoring is down from just under five runs per game in 2006 (4.97, to be precise) to 4.3 R/G per team today, or close to 16 percent. Less scoring makes for narrower margins, but it also involves less opportunity to fatten up a team's offensive tallies against a bullpen's worst pitchers.
If nobody's much more than three games above their expected records, who have the significant underperformers been? Via simple Pythagorean records, the two teams three games worse than their expected records are the A's and the Yankees; and via third-order wins, the A's (-3.1) and Yankees (-3.6) come out as those two teams doing worst in the league relative to their expected records. It would be hard to imagine two teams with less in common, since they're separated by 150 points in winning percentage.
That said, as offense-deprived as the A's have been, it's probably no surprise to find them enduring a case of their fortunes taking a turn from bad to worse. Being an MLB-worst six games under .500 in one-run games adds insult to the oft-injured Athletics. (Interestingly enough, the Angels and Rangers are both five games under in one-run games, which perhaps helps to explain why the Mariners look fairly good these days.) If the A's were at their expected record (33-37), they'd be in a three-team pack about three games behind the Rangers.
The Yankees' predicament is different, but there are a few different things going on. Beyond a slightly below-average record in one-run games (8-11), they've also managed to pile up a ton of runs in blowout wins. To some extent, that's the benefit of a strong top-to-bottom lineup with little or no mercy in it, but mounding up blowout runs doesn't help you much with season-total metrics. Baseball-Reference's "Simple Rating System" grades the Yankees as the best team in terms of the number of runs they're better than an average team (1.7), better even than the Red Sox -- but it's Boston that has run up an 8-1 record on the Yankees.