A Tale Of Two Pitchers: The Francisco Liriano Enigma

When the Chicago White Sox sent shortstop Eduardo Escobar and left-hander Pedro Hernandez to the Twins for Francisco Liriano, most agreed that the biggest risk for Chicago was not the loss of either young player, but the mere allocation of meaningful starts and innings to the game’s most combustible southpaw. If a player was going to make Kenny Williams regret the trade, it would be the arrival, and not the departures.

It struck me as odd that the masses responded in this way, not because a Liriano implosion isn’t a distinct possibility, but because you don’t see that sort of reaction very often. For example, when the Reds acquired Jon Broxton for two interesting pitching prospects, most everyone agreed it was a steep price—overly steep for many, including myself. However, I didn’t see much of "Well, what if Broxton totally screws up the 20 innings he gets the rest of the way?" Or "Brandon League’s having a down year and is coming out of Safeco; he might be a downgrade." Or "Hunter Pence’s OPS isn’t much above Nate Schierholtz’s; maybe he wont be much better." Those are all possibilities, but veteran additions’ imperfections are usually swept aside at trade time, as fanbases anoint their saviors and marketing departments are more than happy to cash in on the frenzy. But Liriano can apparently leave such a bad taste in one’s mouth that he doesn’t get the benefit of the doubt of mediocrities like Broxton and League.*

*Of course, by no means am I saying the above opinions are anything close to unanimous. Just some general impressions I get from post-trade Tweets and articles that I found interesting and/or surprising.

And yet, when Liriano strikes out 15 batters, as he did on July 13, nobody reacts in total shock. There are few pitchers who can rack up that many whiffs in an outing and leave fans saying merely "Well, he was on tonight." Verlander, Kershaw, Strasburg, Hernandez…are there any other names you’d put above Liriano as more likely to post gaudy numbers in the ever-revered "K" column?

Of course, I’m not oblivious as to why regard for Liriano embodies such a paradox. The start after he struck out 15, Liriano struck out 10. The start after that, he didn’t get out of the third inning. Liriano’s career has been a roller coaster, with peaks and valleys, both connected and disconnected to injuries. One month, he’ll be the best pitcher in the AL not named Verlander; the next, he’ll be Jonathan Sanchez.

At this point, I think we’re all resigned to the notion that Liriano will continue to madden for the next decade. But it’s not often discussed what drives him to these extremes. Can data tell us anything? Let’s find out.

These are both charts of Liriano’s pitch locations over a multi-start stretch. One is of a stretch in which he struck out at least five batters every time out, never allowed more than four earned runs, and only twice failed to escape the sixth inning. The other is of a stretch that saw him allow at least four earned runs every start, never make it out of the sixth inning, and ultimately get demoted to the bullpen. Which is which?

There’s a fair bit of consistency here. We can see three things are clear:

1) He avoids the upper third of the strike zone.

2) He works the slider mostly to his glove side.

3) He works the changeup mostly to his arm side.

That’s a nice place to start analytically. For such an enigmatic pitcher, Liriano isn’t all that complicated on a stuff level. He’s got a moving fastball, a slider, and a changeup, and he generally tries to work the fastball down, the slider down and glove side, and the changeup down and arm side.

Let’s further examine Liriano’s stretch of bad pitching (his first six starts of the season) vs. his stretch of good pitching (his first ten starts following his re-insertion on May 30) (Yes, that means the second graph, with far more pitches, is the good Liriano). First, a comparison of the basic stats:

Good: 10 GS, 63 1/3 IP, 38 H, 28 BB, 75 K, 2.84 ERA, 2.70 FIP
Bad: 6 GS, 26 2/3 IP, 37 H, 19 BB, 21 K, 9.45 ERA, 6.84 FIP

So, uh, quite the difference there. I think that this more or less captures Liriano at his best and worst. One could pick out his best starts and worst starts individually, but it would make the data much more tedious to compile without a whole lot of extra gain. Next, let’s look at his pitches:

Fastball (good): 50.7% usage, 93.0 mph, 9.21" armside run, 6.85" "rise"*
Fastball (bad): 56.8% usage, 91.9 mph, 9.89" armside run, 7.89" "rise"

*In PITCHf/x, vertical movement is positive not because a ball rises, but because its backspin resists gravity enough to make the ball sink less than a spinless ball with the same velocity would. An average four-seamer has about nine inches of "rise," so Liriano is actually throwing something of a sinker—in fact, his movement trajectory is a fairly standard heavy running two-seamer.

91.9 mph when he’s bad; 93.0 mph when he’s good? Man, this guy sure is trying to prove the Duffy Rule.

Also note the increased usage and inch less of sink on the fastball; the former is more important than the latter, however.

Slider (good): 34.3% usage, 86.1 mph, 0.67" "armside run,"* 2.85" "rise"
Slider (bad): 26.8% usage, 83.4 mph, 0.47" "armside run," 1.99" "rise"

*Obviously, Liriano’s slider doesn’t actually break to his arm side. Note the ~9" difference from his fastball’s H-Mov to his slider’s. It’s quite common for guys with natural fastball run to have breaking pitches plot as having little horizontal movement; similarly, guys with natural cut often see their breaking pitches show up as having huge horizontal sweep.

There are huge differences here. Liriano clearly tightened up his slider, adding nearly 3 mph to it while making it slightly shorter. He also increased his usage of the pitch significantly.

Changeup (good): 15.1% usage, 85.5 mph, 9.77" armside run, 5.73" "rise"
Changeup (bad): 16.5% usage, 84.1 mph, 10.6" armside run, 5.96" "rise"

The changeup also came in faster in the good period, which is slightly paradoxical on face, but less so when you realize that its gain in speed basically mirrored the fastball’s.

So overall, Liriano threw more sliders, fewer fastballs, and slightly fewer changeups, while adding a bit over 1 mph to his fastball and changeup and about double that to his slider. The movements of the pitches were all mostly stable.

So that’s the difference in the arsenal itself; how about the results?

Fastball (good): 55.4% strikes, 19.3% called, 5.9% swinging
Fastball (bad): 55.3% strikes, 18.6% called, 5.3% swinging

Similar numbers of strikes overall, but good strikes—called and swinging—both slightly higher in the good period.

Slider (good): 69.1% strikes, 13.4% called, 25.7% swinging
Slider (bad): 63.7% strikes, 12.2% called, 17.8% swinging

And here we go. Not only did the rejuvenated Liriano lean more heavily on his slider, he also derived far superior results from the pitch. That explains a heck of a lot.

It’s worth noting that the "bad" slider is far from bad: all of those numbers are quite good for a breaking pitch. However, Liriano’s strike percentage on his fastball is putrid even with his turnaround, so his slider needs to be more than excellent to overcome its deficiency.

Changeup (good): 53.6% strikes, 4.6% called, 25.8% swinging
Changeup (bad): 50.6% strikes, 8.4% called, 13.3% swinging

This one doesn’t seem so self-explanatory. After all, little about the changeup changed, but suddenly its swinging strike rate doubled.

Clearly, though, Liriano lives and dies with his offspeed offerings. His fastball, for all its velocity, did him no favors in April and didn’t pick up much in June; it was the slider and changeup that took quantum leaps forward. It’s worth noting that the fastball actually plays reasonably to lefties, against whom he completely eschews the changeup—in the good period, the fastball went for 61.9% strikes and 9.7% swinging strikes against his fellow southpaws. Then again, that only magnifies the weakness to right-handers (53.5% strikes, 4.8% swinging, again in the good period).

If Liriano falls behind a righthanded batter and backs himself into a corner where he has to turn to the fastball, then, it’s clear he would unravel quickly. On 2-0 counts to righties in the good period, he went to the fastball almost exclusively, and 20 of 42 fastballs went for balls. Two were swung on and missed, two were fouled off, and the other eighteen were put in play. That’s 38 out of 42 outcomes that were suboptimal.

Of course, the fastball is like no other pitch; it can’t be totally ditched. If a slider or changeup isn’t working, you can junk it, but not a fastball. You have to throw a fastball, unless you throw a knuckleball, and even Tim Wakefield threw some "fastballs" in there from time to time. Furthermore, I fear for Liriano’s elbow enough already without the notion of him turning to Michael Wuertz levels of sliders.

So how do you throw a fastball when you shouldn’t be throwing a fastball? That’s a damn good question; I’m glad you asked. Wait, am I talking to myself now? Another damn good question; I’m glad you asked!

In all seriousness, though, I’m not sure there’s a great answer. But I will make an interesting, out-of-the-box suggestion: Why not use some combination of leverage and batter skill to decide?

We know Liriano shouldn’t throw many fastballs to righties—the predictability of going slider/change to them is a downside, but they are just so much more effective than the fastball that I doubt the two cancel out. So why not go exclusively slider/change (with maybe the occasional first-pitch heater to keep them honest) to the best couple of righties in an opposing lineup, and make up the "fastball quota" to weaker batters or lower-leverage situations? For example, when Liriano faces his old team, perhaps he should pepper Josh Willingham and Trevor Plouffe with junk but challenge Jamey Carroll and Brian Dozier. And if he’s got a five-run lead in the sixth, he could throw a few more heaters to save his arm and also give a different look the third time through the order.

It’s worth thinking about, I think, though pitchers don’t usually extreme-ize* their approaches in such a fashion. Then again, Liriano is, well, quite extreme himself.

*-Ify, perhaps?

In any case, Liriano is now under the tutelage of well-respected pitching coach Don Cooper, not to mention the fact that he’s escaped from an organization that preaches precisely the opposite approach to his (Of course, in exchange for that, he’s gone from a cavern to a bandbox, but that’s the cost of these things). It will be quite interesting to see how, if at all, Cooper tinkers with Liriano down the stretch.

It’s tough to say if the changes that drove Liriano’s ace impression in June and July will stick—every version of him seems so convincingly permanent that none of them convince anyone of anything anymore. What does seem clear is that any sustenance of Liriano’s success will depend on the increased usage, velocity, and strike generation of his slider first. There’s also the matter of the changeup’s improvement, which is likely a combination of small-sample swings and playing better off the other two pitches (more velocity on the fastball, more usage of the similar-velocity slider). One could say it’s the slider that makes Liriano able to succeed, but it’s the changeup that makes him dominant. And the fastball…well the fastball is the kind of thing that gets you kicked out of a rotation that includes Cole De freakin’ Vries.

Cole De Vries, man! I saw that guy in 2010, in Double-A. He was the swingman. And he had a 5.80 ERA! And two years later, he had pitched his way into a rotation of other plucky journeymen, and one of the most dominant arms of the past half-decade had pitched his way out of it. Oh, baseball. I could never have imagined Cole De Vries would throw a major league pitch, but here he is, with a 3.81 ERA, no less (We’ll shove the 5.37 FIP under the rug, right? Hey, he pitches to contact!*).

*Snark aside, seriously, you have to like the low walk rate De Vries has. Pitching to contact is great when a guy actually, you know, does it, rather than when it's used as an excuse for why a guy isn't striking anyone out. Also, hey, kudos to an NDFA for making the majors, not to mention two years after even I wrote him off. Good for him.

**I'm really overdoing the footnotes.

***Do these even count as footnotes? I'm not even sure what the technical term is. Ahh, the speed of technology, 'tis ever increasing.

****Did I really just go from talking about my lame asides to some rambling musing about technology? Well, that's a damn good question, I'm glad you asked! Wait, I'm glad I asked, right? WELL, THAT'S A...

Wow, that was random. But that’s just what Cole De Vries does to ya, right? Okay, no, I’ve just been at this too long and need sleep. And how many players can have the word "freakin’" inserted into such awesome places in their names?

Yeah, I really need sleep.

To sum up, though, it does seem that the statlines of Francisco Liriano’s starts are not generated on a cosmic dartboard, nor are they tremendously affected by meteorological trivialities. There are clear differences between Liriano on a roll and Liriano in a rut, and they can be found on multiple levels of his pitching data, from the process to the results. More than the outcome of his starts, it’ll be interesting to see where he goes from here. Will he fall apart like Oliver Perez and Jonathan Sanchez? Will he put it together again like Oliver Perez and…2010 Francisco Liriano?

Whatever the case may be, we’ll all be surprised. And yet, we won’t be surprised at all. Ahh, well, that’s just part of the beauty of the game.