Parts I and II.

The Best of Halves, The Worst of Halves

One other thing about the Braves, as well as the Cards, they have a winning percentage well over .700 in the second half (.706 and .750 as of Monday, respectively). Here are the teams that sustained a .700+ winning percentage for an entire half season:

By the way, the Cardinals fell from fourth (.765) and the Braves from ninth (.735) with losses yesterday (two by the Braves).

Now, for the losers: the Diamondbacks (.220) and Brewers (.294) have been horrific in the second half. Let's look at the all-time worst halves:

The D-Backs are fifth all-time, right behind the expansion-year Mets and well ahead of last year's Tigers, both halves. It's too bad that they were just bad and not abysmally so in the first half or we would have one of the worst teams of all time. They do project to 113 losses which is pretty bad though. Maybe they can try a bit harder, lose their last 23 games, and tie the Mets for second in the most losses by a team ever. Still at 113, they will come in tied for sixth. Here are the leaders:

Pythagorean Pathos

The Red Sox are now just two games behind the Yankees, something that seemed extremely unlikely for large segments of the season. It seemed that they were 10-1/2 games back for months. Many statheads (including me) pointed to their Pythagorean expected winning percentage and said that the Red Sox caught catch the Yankees.

So does a Pythagorean winning percentage that outdistances the actual winning percentage in the first half of a season augur improvement in the second half of the season?

Here are the 25 teams that had expected winning percentages at least 75 points higher that their actual winning percentage in the first half of the given season. For each, the improvement in their (actual) second half winning percentage is listed:

The Red Sox actually don’t even make the list. Their expected winning percentage was just 24 points better than actual in the first half. However, a bigger issue in the division was that the Yankees were well outperforming expectations.

Actually, there were a fair number of teams outdistancing expectations in the first half and many have returned to earth. Here's what I said about it a week before the break:

Keep in mind that the Reds also exceed their expected winning percentage by 86 points. They own a 44-38 (.537) record but by their run differential one would expect them to have won only 45.1% of their games. So surely the difference is caused by the smaller sample of games. Over a full season things will even out, but will the run differential change to fit the record or will the record change to more closely fit the run differential?...

That said, teams like the Yankees, Reds, Twins, Giants, etc. who have been outdistancing run-differential expectations may get their comeuppance in the second half.

So hat has happened in the second half to the teams that have outdistanced their first half expectations? Here are the worst offenders with the winning percentage difference for the second half:

The Yankees and Reds both appear on the list and both have had severe declines in the second half.

Actually, you'll notice on both lists that there are a good number of teams that defy their Pythagorean expectations in the second half. I checked the correlation coefficient for the two and found it was just .356, not all that convincing. Yes, teams generally do as their expected winning percentage dictates, but it's not a very strong correlation.

Mo'…Larry…The Cheese

Now that the Pythagorean formula is let us down as a predictor of future success, let's take a look at how much a strong finish affects a team in the next season. We are always told that the local nine finished strong last season, their midseason callups looked good, and it all bodes well for the current season. It's practically a mantra in Philadelphia.

Is their momentum from season to season? Does a team that finishes strong one year improve in the next? By the same token, do aging teams that decline in the second half continue to decline in the next season? Let's take a look. Here are the teams that improved the most in the second half and what they did the next season:

Now here are the ones with the worst declines in the second half:

Again neither table appears to express a correlation between second-half success carrying into the next season. But let's run the numbers for every season.

I got a correlation coefficient of -0.0134, meaning that there is a slight trend in the opposite direction. Not only is there no evidence of momentum from season to season, if anything the trend tends to reverse ever so slightly.

To be continued…