Rich Hill and Rethinking the Perfect

Rich Hill keeps forcing us to rethink everything.

He has us rethinking how to pitch, finding success even while employing just a fastball and curve — and frequently utilizing the latter as his primary pitch.

Has has us rethinking when it’s appropriate to give up on an arm. Perhaps never, in his case. Hill’s return to the majors — and to a three-year, $48-million deal — began in independent ball two years ago.

And there was Hill in the ninth inning on Wednesday, forcing us to rethink what it is to be perfect. Three outs away from retiring the first 27 batters of the night, he also faced the prospect of having to go at least 10 innings to secure a perfect game in Pittsburgh.

The question of the perfect game ended in the ninth. A ground ball by the Pirates’ 25th batter of the night, Jordy Mercer, was misplayed by Dodgers third baseman Logan Forsythe.

Such a cruel way for Rich Hill to lose a perfect game! First time in @MLB history a PG bid has ever ended in 9th inning on an error.

— SABR (@sabr) August 24, 2017

We should all have some capacity for sympathy regarding Forsythe, as we all know what it is to screw something up.

Because the game remained scoreless after nine innings, and because Hill had yet to throw a 100 pitches, he took his no-hit bid into the 10th.

Josh Harrison promptly ended the no-hitter and the game, lifting a fly ball just over the shortest portion of left-field wall at PNC Park — and over the outstretched arm of Curtis Granderson — to give the Pirates a walk-off win.

#Pirates second baseman Josh Harrison celebrates his walk-off homer to break up the no-hitter against the Dodgers tonight at PNC Park. pic.twitter.com/plIo44PXPS

— Christopher Horner (@Hornerfoto1) August 24, 2017

Pedro Martinez was the last pitcher to take a no-hitter into extra innings, back in 1995. At the New York Times, Tyler Kepner notes that the last pitcher to throw nine no-hit innings and lose was Mark Gardner of the Expos at the Dodgers on June 26, 1991.

At some level, the result represented a sort of revenge for the Pirates. Fifty-eight years ago, left-hander Harvey Haddix took a perfect game into extra innings against the Milwaukee Braves, only to have it end on this sequence: leadoff error, sacrifice bunt, intentional walk, walk-off double. Only Haddix and Martinez have lost perfect game bids in extra innings. Hill was nearly a third.

Of course, technology and new ways of understanding batted-ball quality could make such brushes with history less frequent. It could also have us rethinking what constitutes “perfect.”

For instance, if we had robot umps, Hill’s perfect game would never have made it beyond the eighth, and Forsythe would be less inclined to feel a sense of shame.

Consider the location of the seventh and final pitch of that at-bat:

For those who support robot umps but who were also rooting for Hill on Wednesday night, that called third strike on Sean Rodriguez represents a collision of interests. Robot umps have no heart, no situational awareness, and no understanding of historical context (unless artificial intelligence and machine learning can develop empathy). We might see fewer perfect games and no-hitters with robot umps. The call on Rodriguez is an exhibit of how bias can emerge when officiating a potentially historic game. An umpire will favor one type of error over another in that situation.

There’s something else, too, about Wednesday night and the notion of “perfection.” In the Statcast era, we have a better understanding of the batted-ball quality and the rate at which certain balls in play become hits. Just as we’re losing a sense of what defines a position — as Russell Carleton suggested at Baseball Prospectus this week — how might our understanding change the way we grade pitching performances?

Wrote Carleton:

“I once again find myself returning to the fact that baseball seems to lack certain words, or that its vocabulary lags behind its reality.”

Hill was excellent on Wednesday. But he also allowed an exit velocity of 105.7 mph on a batted ball by Starling Marte, a sharp grounder with a hit probability of 56%.

Not so perfect after all! (Although, it should be noted: Marte’s was the only ball put in play all evening that featured a hit probability above 50%, according to Baseball Savant.)

Chase Utley also needed to supply a moment of heroism in the eighth to temporary keep perfect game bid alive.

Not so perfect after all! (Although, I should add: Josh Bell’s liner only had a 46% hit probability.)

What do you call a no-hitter through nine innings that ultimately ends in a loss? What do you call a dominant outing in which a pitcher loses but throws nine-plus efficient outings and allows only one ball in play with a hit probability above 50%? A Rich Hill?

The walk-off homer Hill allowed traveled only 347 feet and featured a hit probability of just 16%. It was simply well placed, beyond the only shallow part of left field in the asymmetrical park.

I mean, just look at that thing in LF. By a whisker. pic.twitter.com/VI345LqMYK

— Mike Petriello (@mike_petriello) August 24, 2017

So the walk-off hit would have been a fly out in some parks. But Marte’s ground ball would have been a hit over half the time. And if a robot were calling balls and strikes, Hill’s perfect game would have never advanced to the ninth.

If we indulged every possible exception, it would be impossible for a pitcher to author a truly “perfect” game. So even while many of us benefit from and enjoy the deeper understanding of batted-ball and pitch quality provided by Statcast, an appreciation both of tradition and drama dictates that we ought to preserve our definition of what constitutes a “perfect” game and no-hitter. The perfect game has clear, unbiased criteria: a runner does not reach base. While runners may reach base on poor batted-ball quality, reaching and not reaching base is a clear, binary distinction.

Here’s what we know for sure: Rich Hill continues to be a great story, baseball is weird, baseball is crazy, and Wednesday night provided us with some good theater even if we lack a label the performance.