What exactly are you trying to measure here?
Is this a "getting out of jams" type stat? Is it a general "how good a pitcher is someone" stat?
Generally, the first thing I look for with a stat is to see how well it actually measures what it's designed to measure. To evaluate that, I need to know what this is attempting to measure.
I guess I'm attempting to coordinate outs that are primarily pitcher-driven as a ratio of the how often that pitcher puts additional hitters on base. If it comes down to runners already on base, it would be trying to answer the question: is this guy more or less likely than some other guy to lock 'em down (Papelbon), clean up the mess (Beckett), hopefully wiggle out of it (Wakefield), or throw gasoline on the fire (Snyder). But I think it also could be used to help measure how well a pitcher can keep guys off the basepaths in the first place.
I personally tend to look at a combination of H/9, K/9, BB/9, and HR/9 when evaluating pitching stats. But I also find K/BB useful to rate command, so I looked into what other rate stats could logically be combined into comparable ratios. Answer -- not much there that could be helpful. Hopefully, though, this one will work somewhat toward rating how well a pitcher contains baserunners from advancing on ball-in-play events (whether his own or inherited).
Now obviously, a GDP is a "ball-in-play event", but sinkerball pitchers are much more likely to get that groundball when they need it (regardless of defense) than an extreme flyball guy. Getting that "extra" one out that doesn't allow advance of a baserunner is particularly important for groundball pitchers who don't have a high strikeout rate, which is why I decided to try adding its base number (rather than the total number of outs generated) to strikeouts. For most guys, I would expect HBP and GDP will be random variance and not particularly important, but for extreme groundball pitchers like Webb, Wang, or even Lopez, it may be a vital component to their success or failure. Also, because infield defense will factor both toward numerator (through GDP) as well as denominator (bleeders that get through), I'm aware this isn't perfect. But I think it might be good enough to merit refining.
And I definitely was thinking of adding IFFO to the numerator to help out pitchers like Wakefield (who seems to get a ton of popups), but decided the formula was too fiddly already to try to make it acount for knuckleballers.
