Inside the new system that wants to revolutionize how we look at pitches

MLB columnist
Yahoo Sports

On Aug. 4, 2010, with two outs in the bottom of the sixth inning, the Boston Red Sox sent veteran utilityman Bill Hall to the plate as a pinch hitter. Hall quickly faced a 0-2 count, and on the fourth pitch of the at-bat, he grounded the ball weakly to second base to end the inning. This looked like just another dog-days plate appearance, that final pitch every bit as ordinary as the name of the man who threw it: Joe Smith.

Joe Smith has an interesting role in the Quality of Pitch (QOP) metric. (USAT)
Joe Smith has an interesting role in the Quality of Pitch (QOP) metric. (USAT)

At the time, Smith’s career wasn’t terribly distinguished, either. He was 26 years old, a sidearming right-handed reliever who overwhelmed right-handed batters like Hall but struggled enough that his ERA going into that game was 5.24. Of all the players to throw the single finest pitch of the last seven years, Smith was far from the likeliest candidate. And to the naked eye, it looked like little more than a regular 93-mph fastball in the upper-right quadrant of the strike zone.

Scroll to continue with content
Ad

“Nobody really understood how good that pitch was,” Jarvis Greiner said. He is 24 years old and wants to change baseball, and like many idealists before him, and many more to come, Greiner is convinced he has the tool to do it. On Thursday, at the Society of American Baseball Research Analytics Conference in Phoenix, Greiner will introduce his Quality of Pitch (QOP) metric, an attempt to leverage the massive swell of data available from baseball’s PITCHf/x system and calculate a single number that quantifies just how good a single pitch was.

[Yahoo Sports Fantasy Baseball: Sign up and join a league today!]

For more than five years, Greiner and Dr. Jason Wilson, his professor at Biola University about a half-hour outside of Los Angeles, have attempted to strip away every bit of context from a pitch and grade it simply on what it does. First they focused on just curveballs, using a camera-tracking system that captured pitchers from Biola throwing them and modeling the data to subjective grades of each pitch given by the school’s coaches. Then they honed their current model with the data of more than 5 million major league pitches thrown since 2008 and logged by PITCHf/x. And now Greiner believes he can compare fastball to curveball to slider to splitter to cutter to changeup and say which is the best.

No. 1 of more than 5 million is the Joe Smith fastball. The pitch stretched the limits of Greiner’s model, which currently rates pitch from -10 to +10, with anything 5.00 or better considered a quality pitch. Smith’s fastball scored a 10.031, one of a few pitches with a QOP greater than 10.

“It was a perfect combination of all three factors: movement, velocity and location,” Greiner said. “He’s a sidearmer, so he comes from way off the map, to the outside corner. To the regular fan, it’s like, OK, [93] mph, that’s cool.”

To Greiner, it reinforced why he started researching the subject during an undergraduate statistics class with Wilson: Our eyes often lie and make the extraordinary look positively ordinary. Greiner pitched at Biola until his sophomore year, when a labrum injury ended his career. His father, Wayne, was visiting Biola from Edmonton, Alberta, where Greiner now lives, and they sat in a hot tub at a Holiday Inn, discussing how great it would be to find an all-encapsulating number for pitches.

“It always had bothered me how subjective the art of pitching analysis was,” Greiner said. “There’s really no way to compare pitchers’ breaking balls to each other. That guy has a good curveball. What does that mean? The only way we really had to measure was you had to throw hard. That’s the only way we identify talent. It’s very medieval. Just because you’re throwing hard doesn’t mean you’re going to be successful.”

While some components of the model have changed since the initial study on curveballs, the thrust remains the same. Greiner uses five variables: pitch velocity, pitch location, total break (horizontal and vertical), breaking point (how late a pitch moves) and rise of pitch (maximum height). Some guidelines for a great pitch: low rise, late break, significant total break, high velocity and, most important, location on the corners.

[Baseball is back! Check out Yahoo For Spring Training for the best spring training pics.]

“I was surprised how heavily location factored in,” Greiner said. “You go into this thinking [movement] is the most important thing to quantify. ...It surprised me how much the model improved when we increased [location]. It ends up being a heavy factor into how good every pitch is.”

QOP might give us a better understanding of why Mark Buehrle is so effective. (USAT)
QOP might give us a better understanding of why Mark Buehrle is so effective. (USAT)

Greiner points to Mark Buehrle, the Toronto Blue Jays’ 83-mph-throwing starting pitcher, as someone who benefits from the location of his pitches. Buehrle also gets by changing speed, a factor QOP does not consider and one of its potential weaknesses. Pitch sequencing – or how a pitcher approaches a hitter or multiple hitters by linking different types of pitches in an intentional order – is believed to be very important, and the QOP system ignores it, considering its dependence on context inconsistent with the model.

Similarly, Greiner has not measured the result of every pitch against QOP to see if there is a correlation. He believes in the sanctity of his model’s ability not just to exist in a theoretical world but one where a 9.00 QOP pitch gets hit well on an exceedingly rare basis. That is indeed the sort of thing that could change baseball, allowing analysts to dig deeper and perhaps find one secret that makes for a great pitch alongside the multifaceted secret sauce Greiner believes the QOP system can provide.

He sees a day where Doppler radar systems like TrackMan can capture data from high school games or showcase events and give scouts an objective measure of a pitch’s effectiveness that says far more than a radar gun ever can. A fringe benefit: With high speeds and max-effort pitching being two of the culprits in the rash of Tommy John surgeries, pitchers learning harder does not necessarily equal better could help curb the culture of velocity that pervades youth baseball.

If the model holds, Greiner sees QOP alongside another three-letter acronym, MPH, on scoreboards and TV broadcasts everywhere. For the next year, Greiner plans on making tiny tweaks to the model based on how it compares to the subjective grading of particular pitches by baseball coaches and others. This certainly opens the model up to criticism, considering the model hews to subjective analysis, which itself is open to interpretation. Accordingly, Greiner wants to find the very best and brightest evaluating minds to help streamline the model even more, hoping for a buy-in from a major league team that will lend its scouts to assist.

This isn't the first attempt at a system like QOP. More than half a decade ago, MLB.com, which runs the PITCHf/x system, tried its own single-pitch quantification system called Nasty Factor. Using velocity, spin, sequencing and result – some of the variables Greiner considers outside the pitcher’s control – Nasty Factor started with what engineers felt was a strong degree of accuracy. Then the game slowly evolved, and while the model remains in use, it’s not at nearly the confidence level it once was, prompting a tweaking of it in the near future.

Now it has a companion, one with a success story: Joe Smith. Since 2010, the year of his super pitch, he has been one of the finest relievers in baseball. Among the 120 relief pitchers with more than 150 innings pitched, Smith’s 2.25 ERA is seventh best. Smith’s average QOP in 2010 was below the median – about 4.15. His outliers, however, showed the potential for more, potential he has more than fulfilled.

When Greiner presents QOP on Thursday, he’ll see his baby go through its greatest test yet. Some of the smartest people in baseball will assess it, question it and wonder just how well it holds up over the long haul.

“If it can really do what I say it can do, it’s going to have implications for a lot of people,” Greiner said. “We’re going to be under the microscope. I know we are.”

And he’s good with that. More than five years of time and energy led him to this place, and all he wants to do is be Joe Smith in the sixth inning against Bill Hall. Nothing flashy. Nothing eye-popping. Just really, quantifiably good.

More MLB coverage from Yahoo Sports:

What to Read Next