- FanGraphs Baseball - http://www.fangraphs.com/blogs -

Break Out the Brooms

Four days ago, if I had said that the four game series between the Phillies and the Astros would end in a sweep, I don’t think that anybody would find that claim terribly outlandish. If I had said that it would be the Astros completing that sweep this afternoon, I would have been laughed out of the room. However, behind a 5-1 victory over the Phillies today at Citizen’s Bank Park, the lowly Astros extended their winning streak to five games, including a road sweep over the two time defending National League champions.

Entering the series, the Phillies had a .556 winning percentage and a .529 third order winning percentage – winning percentage based on expected runs scored and allowed and adjusted for strength of schedule – according to Baseball Prospectus’s adjusted standings. The Astros entered the series with a .439 winning percentage and a .405 third order winning percentage. Basically, by either measure, the Phillies have played about 120 points of winning percentage better than the Astros this season.

Let’s take a look at just how unlikely, then that makes this sweep. The method to use here is the log5 method, a method derived by Bill James to estimate the winning percentage when two teams meet. The following formula gives the expected winning percentage for team A facing team B:

According to this method, the Phillies would be expected to win one game 61.5% of the time using raw wins. Using third-order wins – the better method – the Phillies are expected to win one game 62.3% of the time. However, we have to account for the Phillies playing at home. Since the home team wins 54% of the time in MLB, we credit the home team with 20 points of winning percentage and subtract 20 points from the away team. With this adjustment, the Phillies win 65.3% of the time with raw wins and 65.1% of the time with third order wins. Using probability theory, we can then determine the amount of times we would expect the Phillies to win 0, 1, 2, 3 and 4 games in a four game series, as seen in the following chart.

Expecting Philadelphia to sweep the series wasn’t realistic, as the odds of Philadelphia taking four out of four was only 18%. However, they should have won the series – 56% probability of three or four wins – and a series loss should only have happened 11% of the time. Now, we know that some crazy things happened, particularly in the 16 inning game which saw Roy Oswalt make an out in left field. Still, we would only expect a sweep to happen 1% of the time. This one out of one hundred could be disastrous for the Phillies. Over the past week, their playoff odds according to Baseball Prospectus have fallen by 23%, and before this game their odds were roughly 1 in 3. San Francisco is off tonight, but Los Angeles has already won and St. Louis plays the Nationals tonight. The Phillies lost out on a major chance to make a stand in the Wild Card race and capitalize on a three game losing streak in Atlanta. It’s not fair to say that this series will take all of the blame if the Phillies miss the postseason, but there’s no denying that this sweep was disastrous for Philadelphia baseball.