<
>

Early run differentials can be misleading

4/26/2011
Brett Davis/US Presswire

The Cubs are 10-12, and already have an ugly minus-25 run differential. That's one of the league's worst marks, along with the Pirates (minus-22) and Astros (minus-26). Taken in the broad strokes, it already looks like the Cubs are out of consideration in what was supposed to be a four-team race in the Central.

However, that's the thing about getting too hung up on early-season data, even team-level data: Individual game outcomes have a way of making an oversized impact, fragging any broad initial conclusions you might want to rush into making. Starlin Castro's game-losing three-error inning Monday night contributes to my point, however unintentionally: Over three weeks of baseball, weird stuff is sure to happen.

If the Cubs have a master plan, it certainly didn't involve enduring the chuck-and-duck antics of James Russell and Casey Coleman in almost a quarter of their ballgames. Having made five of the Cubs' 22 starts, Russell and Coleman are filling in for Randy Wells and Andrew Cashner, who are both on the DL. That excuse goes only so far, though; the Cubs aren't even the only team in their division to have lost two of their front five, as the Reds are also dealing with the absence of Johnny Cueto and Homer Bailey. The difference is that the Reds have better depth, especially when it comes to picking a front five; cut the Cubs, and they don't just bleed, they bleed out. Russell and Coleman aren't great choices for a sixth starter on anybody's ballclub.

As far as outsized impacts in the early going, having to employ that pair as rotation regulars has been considerably responsible for the Cubs' big early run differential. Care to guess who was starting when the Cubs lost 11-2 and 12-2? Not that it excuses the offense on those particular days, but there's your run differential right there. Neither Russell or Coleman made it through the third inning, so both times out the games weren't just blowout losses, they were bullpen-taxing embarrassments. The vaunted late-game trio of Carlos Marmol, Kerry Wood and Sean Marshall isn't going to do you much good in that sort of ballgame, but that's the thing -- bad ballgames have a way of getting worse, because you avoid using your good relievers to finish up a game already logged in the loss column.

Absorbing blowout losses is going to have an outsized impact on interpretive metrics, though, leading to extreme -- and incorrect -- conclusions. Last year's Astros make for a nice extreme for the sake of comparison: They finished the year 76-86, with a run differential of minus-118. By the Pythagorean theorem of expected wins and losses, that translated to a record eight games worse. VoilĂ , a conclusion: the 2010 Astros overperformed because they should have won 68 games, and barring a major talent overhaul, should have been damned and doomed to slip even further back in the pack come 2011.

The Astros didn't overperform that much, though, and that's because of the way their season broke down early. In the first two months, rookie skipper Brad Mills' ballclub went 17-34, with a run differential of minus-101, which gives you an expected winning percentage of .285, worse than the 1887 Cleveland Spiders. The 2010 Astros weren't just initially bad, by extrapolating their performance they were supposed to be even worse.

Except that, practically speaking, you can't be worse than the 1887 Cleveland Spiders were. Not even the first-edition Mets could be that bad. The Pirates have never been that bad. But in the first two months, the Astros figured out they needed to turn over almost half of their lineup, because Kaz Matsui and Pedro Feliz were cooked, because J.R. Towles' latest epic flop wasn't worth their time, and because Tommy Manzella got hurt, conveniently enough. Those weren't just bad players they were deleting from a league-worst lineup, as that quartet comprised four of the game's worst hitters in the first two months. That wasn't the only thing that swung the Astros' way; on the pitching side, Wandy Rodriguez rediscovered his curveball and went from one of the game's worst starters to one of its best.

With that sort of in-season reboot, the 2010 Astros went 59-52 the rest of the season. They did that despite a minus-17 run differential, suggesting that they should have won 54 games -- which would still represent a massive improvement from where they had been in the first two months. They even went 32-27 after the trade deadline -- three games better than expected -- despite dealing away Lance Berkman and Roy Oswalt. As far as not doing themselves many favors, they outperformed their expected record despite two season-long challenges: a 29th-ranked defense (via Park-Adjusted Defensive Efficiency) and a 24th-ranked bullpen (per Fair Run Average, which accounts for performance with inherited baserunners).

So, if you're inclined to start drawing conclusions from early season run differentials and what they say about which way a team is automatically going to go in the standings, here's my advice: don't, because major personnel changes make a difference, and Wells and Cashner aren't out for the year. On the other hand, if the Cubs do wind up the season getting two-fifths of their starts from Russell and Coleman, they'll have a very real problem on their hands, one that could convert an ugly early-season run differential into a full-season disaster.

Christina Kahrl helped found Baseball Prospectus in 1996, is a member of the BBWAA and covers baseball for ESPN.com. You can follow her on Twitter.