ESPN End of Season Ownership: 22.6%

It may not look like it on the surface, but the first and second halves are a stark contrast for James Paxton, who pitched like an ace down the stretch after scuffling in the first half.

Consider:

First half - 3.91 ERA, 8.4 K/9

Second half - 3.72 ERA, 8.9 K/9

That’s it? Wait, so what gives? A key difference is what opposing batters were doing to Paxton down the stretch.

In eight starts before the All-Star break, Paxton was beat up to the tune of a .317/.356/.432 slash line. In 12 starts after the break, opposing batters hit just .253/.278/.388. That’s a huge reason why his WHIP went from an unreasonable 1.59 in the first half to a solid 1.12 in the second. The improvement of his command was also noticeable, as he bumped his K/BB rate from a steady 3.2 in the first half to an otherworldly 7.2 in the second.

The change wasn’t due to velocity, as his numbers stayed relatively stable. Paxton also didn’t find more grounders as the season went on, which can frequently move guys in a positive direction. Instead, his ground-ball rate dropped by nearly 3 percent from the first to second half (49.7 percent to 46.9).

There are two spots where Paxton made a shift. He dialed back on his solid four-seam fastball (routinely 96-97 mph on average) as well as his changeup and cutter and poured all of his effort into his curveball. The southpaw threw that pitch nearly 20 percent of the time in the second half after throwing it just 5 percent in the first half.

Moving away from the changeup was a good move, as it was getting pummeled (.563 slugging against in the first half) and it got worse after the midsummer classic with an unthinkable .929 mark. Still, it was a leap of faith in the curve, which had allowed a slugging percentage of .546 in the first half. After the break, opposing batters slugged just .227 against it -- a huge reason for Paxton’s second-half improvement.

The other spot where Paxton took a jump was going from a pop-up rate of 2.3 percent to 12.1 percent. In other words, he went from the absolute bottom of the barrel -- Michael Pineda was the worst qualified starter in 2016 with a 3.8 percent pop-up rate -- to being a fringe top-10 guy as far as inducing popups.What’s unclear is how exactly Paxton did it. All five of his tracked pitch types from Brooks Baseball shows an increase in fly balls induced from the first half to the second half, but none can be isolated as the reason why he had such a sizable jump.