A guest post: Do the Yankees always collapse in September?

[Editor's note: This is a guest post from longtime IIATMS reader, Professor Longnose. If you'd like to write a guest post, feel free to email me at sgotsulias at gmail dot com. -SG] Aren't You Glad Girardi Rested the Yankee Players All Season So They Could Collapse in September the Way They Always Do??

I don’t know how many times I said that after a loss recently, but I said it often enough that I wondered if I was being unfair. The Yankees certainly had a September to forget this year, and I'm pretty sure I remember them doing it last year, too, but my memory is as short as most fans. Does Girardi always manage them to a limp finish? Fortunately, we have the data to take a look.

Here is a fairly complicated chart doing a fairly simple thing: comparing the Yankees winning percentage from the beginning of the season through August 31 with their winning percentage from September 1 through the end of the regular season for every year of Girardi's managerial reign.

September Under Girardi Table 2

On the left is the Yankee's record--wins, losses, and winning percentage--at the end of August. After that is their place in the standings at that date and the number of games behind they were (negative numbers are games ahead if they're in first place).

Then come their expected chances of making the playoffs at the end of August, per coolstandings.com/Fangraphs. Percent chance of winning the division (DIV), winning the wild card (WC), and making the playoffs in general (POFF) are given. (Unfortunately, I can’t find playoff possibilities for August 31, 2014. I personally estimate their chances at that time as “crappy.”)

After that comes their September-October record. Then there is their final record with games behind and how they did in the playoffs if they made it.

After that, a little spreadsheet prestidigitation. The column “Were they better in Sep-Oct?” is their Sept-Oct winning percentage minus their pre-September winning percentage. If it’s positive, they were better from September on; if it’s negative, they were worse.

Then comes “By how many games?” That’s the previous column multiplied by the number of games they placed from September 1 on. If their winning percentage in September and October was .100 points higher than before that, and they played 30 games, then they won 3 more games in September and October than they would have if they’d played at the same pace as they did before then. As before, negative numbers mean they were that many games worse.

Here are my capsule summaries for each season (from here on in, “September” means “September and October regular season”):

2008: They had almost no chance going into September, and a great month wasn't enough. 2009: Great before September, great in September, great in playoffs. 2010: Lousy September blew the division. 2011: Keeping approximate year's pace in September won the division. 2012: Great September under pressure held the division. 2013: Lousy September finished what little chance they had. 2014: They had no chance going into September and they were mediocre. 2015: The Yankees had a very strong July, going 19-7, accounting for all of the games over .500 they were for the season. They had a 5-game lead over Baltimore and a 6-game lead over Toronto on August 1, when they were 58-45. They went 29-30 after that and blew the division.

So what does it all mean? Let’s do some light data mining, which is not a legitimate way to prove anything but is useful for coming up with hypotheses.

First, the Yankees don’t always have a terrible September under Girardi. It’s an even .500: I count 4 good Septembers and 4 bad ones.

There were 5 years when their possibility of making the playoffs as of September 1 was >90%. They had 3 good Septembers out of 5 in those years. In all 5 years, whether they had a good or bad September, they made the playoffs.

In the other 3 years, their chances were <15%. They had 1 good September and 2 lousy ones. In all three years they didn’t make the playoffs whether they had a good or bad September.

So, basically, September didn’t matter to the Yankees as far as making the playoffs. When they had a great chance as of August 31, they made it; when they didn’t, they went home.

Here’s one that I suspect bugs many Yankee fans: their September record when the division title was on the line. Twice they went into September with great odds of winning the division, once over 90%, once almost 75%. They won both times. But when the odds weren’t so good, when it was a fight, their record wasn’t as good. Three times they had reasonable but not overwhelming chances of winning the division, at 28.5%, 38%, and 51.9%. One of those years they had a great September and won; the other two they played poorly and blew it.

Girardi really didn’t care about the 2010 division title. He admitted that eventually. Even though the rules have changed and a division title is more important now, he’s still managing like he doesn’t care about it. That probably had something to do with their losing in those two years.

Another hypothesis: in 6 of 8 years, their September record was within 2.2 games of expected. That’s really not much of a variation. Basically, they play in September the way they play before September.

A trend: In the 5 years from 2008 to 2012, they had 4 good Septembers and 1 bad one. In the three years since, they’ve had 3 bad ones and no good ones. Maybe—this would be shocking, wouldn’t it?—the team is getting older and running out of gas more now than they did.

What does this tell us? Not much. Once again, I have to admit the data doesn’t say anything definitive about Girardi’s managing, at least at the level of this brief look at it. As usual, grab your crayons and draw your own conclusions.