May 19, 2011
Prospectus Hit and Run
Foreseeing a Brave New World
Since the calendar turned to May, the Braves have been the hottest team in baseball, winning 12 out of 16 games after going just 13-15 through the end of April. The combination of that streak and the Phillies' recent stumbles—including a series loss to Atlanta this past weekend—was enough to bump the Braves to the top of this week's NL Hit List, Baseball Prospectus' weekly power rankings. That's fairly unusual for a team that sits in third place in its division, a game and a half out of first. Just what in the name of Larry Wayne Jones is going on?
At its most basic—which is already pretty complex—the Hit List is based on an objective formula that averages a team's actual winning percentage and three projected winning percentages from our Adjusted Standings. These, in turn, are derived from runs scored and runs allowed via a variant of the Pythagorean Winning Percentage formula:
All of that data can be found in our daily Adjusted Standings, which now publishes the Hit List Factor as well as an Adjusted Hit List factor that accounts for the difference in quality between the two leagues. Also in the mix this early in the season is each team's PECOTA-projected winning percentage, though starting next week that will no longer be part of the equation.
The reason for incorporating all this data into the Hit List is that run scoring and run prevention give us a better indication of a team's strength going forward than their actual record, and using all four percentages is a way to correct for teams that are over- or underperforming relative to their true talent level. Momentum is ignored; studies show that recent history based upon a handful of days or weeks is less predictive of a team's future performance than season-to-date indicators, and that the use of multiple projected winning percentages is more predictive than relying upon just one.
Checking in on the Braves (25-19 through Tuesday), their .619 first-order winning percentage, based on 183 runs scored and 139 allowed, is 51 points (2.3 wins) above their actual record. That winning percentage and their +44 run differential are NL bests, 20 points and 10 runs better than the Phillies—a good clue that they belong atop the rankings. Such discrepancies are often tied to performances in one-run games and blowouts; indeed, the Braves are just 6-7 in the former, but 8-2 in games decided by five or more runs, outscoring their opponents by 41 runs (65-24) in such affairs. Even there, they're falling shy of their runaway MLB-best .861 expected winning percentage—115 points better than the next-best team—in such games. In other words, they're not converting as many of those runs to wins as you'd expect, though over time we'd expect them to.
The Braves' .651 second-order winning percentage is 83 points (3.7 wins) above their actual record. They've scored three runs more than expected based upon their hits, walks, total bases, and other events, but they've allowed 13 more than expected. Such discrepancies are often tied to better- or worse-than-expected performances in higher-leverage situations; scouring the team's splits—where, admittedly, cherry-picking opportunities abound—we find, for example, that the difference between the pitching staff's performance with nobody on base and with runners on is much greater than the major league average:
With men on base, the Braves' staff has yielded an OPS that's 49 points higher than with the bases empty, a discrepancy more than three times the major league average. Looking at it another way, while they're 119 points of OPS better than average with the bases empty, they're "only" 83 points better than average with men on base, though it's also true that they've faced fewer situations with men on base than average.
Looking at the staff's performance broken into low, medium, and high-leverage situations, which account for inning and score margin, it turns out that the Braves are 85 points of OPS better than average in high-leverage situations, 75 points better in medium-leverage situations, and 148 points better in low-leverage situations. In other words, a disproportionate amount of the good work they're doing isn't translating to wins and losses, because it's happening in games that are essentially decided.
As for the Braves' third-order percentage of .702, 134 points (5.9 wins) above their actual record, the opponents they've faced have accumulated an above-average .264 True Average but only a .263 True Average allowed, so their already-widening run differential from the second-order adjustments widens a bit further. Throw in PECOTA's high preseason expectations (a .538 projected winning percentage) and a 19-point adjustment downwards based on NL's interleague inferiority relative to the AL, and their overall Adjusted Hit List Factor comes out to .597, 22 points ahead of the Phillies.
All of which is our longwinded way of saying that all things considered, the Braves appear to be something closer to a .600 ballclub than a .568 one, while the Phillies, who are winning at a .610 clip, are closer to a .575 club.
Extrapolating from the two teams' current performances, the implication is that over time we expect them to swap places in the NL East standings. Of course, each team's performance in those areas is itself made up of numerous small-sample showings by individual players likely to produce closer to their career norms (Dan Uggla probably won't hit .196/.270/.375 all season long, and Jair Jurrjens won't post a 1.66 ERA) or be displaced by other players (Freddy Freeman won't survive hitting .226/.321/.358, and soon enough the Phillies' second base fill-ins—"hitting" .230/.282/.288—will be replaced by Chase Utley, now rehabbing and hopefully back by month's end). Still, this is our best estimate of current team quality, and it shows the Braves as the better team.
A version of this story originally appeared on ESPN Insider .