Word. Time to part ways with Randy. The loss of bad attitude alone should buoy things a bit.
 Eric Martin | 2:21 PM

Joel, I have to say, this analysis doesn't make much sense. The numbers crunching is typically granular, but what it adds up to -- particularly RJ's walk & hit rate with runners on -- is that the guy is unusually, um, "unlucky" with runners on. If we're arguing that Johnson's ERA should be more like a 4.00 than a 5.00, that still doesn't get you number-one-starter status, especially if we're being asked to discount his abysmal recent habit of pitching himself into extra-bases trouble.

Look at the guy's last three seasons: http://www.baseball-reference.com/j/johnsra05.shtml
His innings-pitched declines; his strikeout total falls; his ERA and walk total rises. This is to be expected in a declining pitcher. And he's still good, if not great. But the question is, after sinking so much into RJ and getting less than expected, why not try to cut our losses?
 Spencer Ackerman | 4:00 PM

Because if a pitcher yields an unusually high number of hits per balls hit into play during one season, you very, very commonly (almost without exception, really) see that pitcher regress to the mean and improve a lot the year after. (See Bill James on Voros McCracken.)

And (ahem) "Year two-thouu-sand! Clap, clap, clap-clap clap!" Schilling, Dice-K (never thrown in MLB but c'mon), Paps (500 ERA+ last season, park adjusted), Wake and Beckett is nobody's definition of "iffy."
 Unknown | 12:24 PM

"Because if a pitcher yields an unusually high number of hits per balls hit into play during one season, you very, very commonly (almost without exception, really) see that pitcher regress to the mean and improve a lot the year after."

Jack, I've just gotta wonder how well that rule holds for pitchers in their early to mid-40s. At a certain point in a player's career, regression towards his career mean has to be less likely than regression, period.
 low-tech cyclist | 1:12 PM

Well said, Rufus, however I must offer a clarification: Pitchers of Johnson's age do get worse, but the way pitchers of any age tend to get worse is by losing their power and control---walking more batters, striking out fewer, serving up more meat pitches that get hit hard. You can measure this decline in BB, K and HR, and aging pitchers typically see regress in one or more, and especially strikeouts. Groundball pitchers don't typically become flyball pitchers in mid-career, so the ratio of hits per balls hit into play usually reflects quality of defense and chance variations, not the quality of the pitcher.

And in fact when you identify pitchers who have hits/balls-in-play ratios that differ from the rest of their teams' pitchers, it's typically a fluke thing that does not remain constant the next year.

Look at Johnson between 1998 and 2002 (2003 was a shortened season for Johnson): Each season featured more than 300 Ks, 20-30 HR, and about 75 BB. Then see the last two seasons: 211/32/47 in 2005 and 172/28/60 in 2006. What doesn't change is total number of hits: 194 in 2006 (5.01 ERA) and 197 in 2002 (2.32 ERA). Johnson's age is showing up in fewer strikeouts, not in balls in play suddenly turning into hits more often---if anything, he's getting more fielding outs because he can't rely on the strikeout as much.

There's no question that Johnson is a worse pitcher than he was five years ago. But it's likely---very likely, in fact---that this particular aspect of his pitching (yielding more hits with runners on than you'd expect) is a function of poor luck. His extension during his windup is visibly worse, so you might be able to argue that he has a problem pitching out of the stretch that explains the variance. But it's a really tenuous correlation, since you would expect that to hurt him in the full windup just as much, and you don't.

Okay, end pedantic rant.
 Unknown | 2:38 PM