The graph in question:
Toward the end of the season, Barry Zito wasn't giving up fewer runs, but he was pitching better -- allowing fewer walks, and getting more swing-throughs. What the graph purports to show is that his curveball was getting better and better as the season went along. When it slurves, Zito can't control it. But when it's his normal 12-to-6 (or 11-to-5, really) curve, he commands it better.
Better curveball command means all sorts of good things for Zito. It's a get-ahead pitch, and it's a strikeout pitch. But perhaps most importantly, it makes his fastball play better. And that allows things like this:
On the previous pitch, Holliday just tipped a big curve to stay alive. He had to be wary of it for the next pitch, and that's why he couldn't catch up to the blistering 86-m.p.h. heat from Zito.
Through three innings, Zito has his curve and his command. He'd like a lead, too, I'm sure.