The A's are still underachieving, says Pythagoras

Way back in June, just 58 games into the season, I wrote about the A's and their Pythagorean record as it related to their actual win-loss ratio. Oakland was a stellar 36–22 at the time, but the simplest metrics in baseball — runs scored and runs allowed — hinted that the A's were even better than their .620 winning clip, hard as it was to believe. They should have been winning at an absolutely ridiculous rate of .730, which would have given them 42 wins through those 58 contests, and 118 over the course of the season.

Here we are on August 19, with the A's — for the first time in what seems like forever — one half game behind the Los Angeles Angels in the American League West. Oakland's winning percentage is .589, good for a 73–51 record 124 games into the season.

Plug in runs scored and runs allowed, and you'll find that Oakland expected winning percentage at this point in the season is .638, which would be good for 79 wins through those 124 games. What do you know? The A's are still underachieving, as compared to their Pythagorean expectation, by six wins.

Of course, a six-win difference is far less substantial through 124 games than it is through 58. The gap is being bridged, which is at least comforting to those of us who have faith in the numbers when analyzed objectively. As they should and almost always do, real-world results are creeping toward expectations set by relatively baseline and very reliable metrics.

Unfortunately, the gap is being bridged more from the high side than the low. That's to be expected — playing .730 baseball was never remotely realistic.  On June 4, the A's should have been winning at a rate of .730 and were winning at .620. Today, on August 19, they should be winning at a rate of .638 and are winning at .589.

Teams' real records rarely deviate from their pythagorean records by more than three games. The A's are obviously a very different team without Yoenis Cespedes, but his absence on the offensive end shouldn't make a huge difference in Oakland's expected scoring figures thanks to the platoon of Stephen Vogt and Jonny Gomes, who could easily become more productive than Cespedes despite being half as flashy and telegenic.

The point remains, though, that the A's are still "underperforming," though they've come much closer in recent weeks to winning as many games as the math says they're supposed to.

As we saw in late June, even series against the Mets aren't gimmes. But Oakland's remaining schedule is interesting, in that they're playing teams that are either very well positioned in playoff races or nowhere near them. The A's play the Angels 10 times between now and season's end, but get seven games against last-place Texas. They play the Mariners six times, but get four chances against the White Sox and three against Philadelphia. If Oakland can break even against Seattle and Los Angeles, the wins will be there for the taking against the lower-caliber opposition.

In all likelihood, the real win percentage and the pythagorean expectation will keep converging, potentially meeting or coming close together somewhere in the area of .610, giving the A's an impressive but hardly history-making 98 or 99 wins. Again, it's not .730, but the post-slump numbers favor Oakland almost as much as the pre-slump numbers did.

The A's will lose plenty against the Mariners and Angels — each of those teams is very, very good. It's the games against the Mets, White Sox, Astros and Rangers that can make or break them, and the first one is tonight, with Scott Kazmir taking on New York's Dillon Gee. First pitch at the Coliseum is at 7:05pm.