Koji Uehara is Trimming the Fat

With no AL or NL West teams in this World Series, we at Midnight Baseball will turn our focus to Red Sox and Cardinals players who used to play out west. Out of all of them, Koji Uehara probably has the highest profile right now. So, with apologies to Carlos, we start with Koji.

Few could have seen such dominance coming this season from Koji Uehara. The Red Sox signed him for one year and $4.25 million, or less than the market value for one win. We can therefore imagine that his suitors were few and miserly. He entered the season behind Joel Hanrahan and Andrew Bailey for the closer’s gig, which doesn’t always go to the best pitcher but shows what the Red Sox thought of him. Bullpens are like totem poles, or pyramids of cheerleaders. They are constructed with a hierarchy, and the people doing the constructing may not have the best criteria behind that hierarchy. The prettiest cheerleader might not be the swellest, the one who deserves the top spot on the pyramid. Coming into the season, Uehara was like a girl who, after many tryouts, finally gets accepted into a cheer squad, but only on the condition that she buy her own uniform, or something. And it’s totally not fair because she has this one awesome, game-changing cheer move, and the baffling thing is, they’ve all seen it before.

The splitter. A rare bird nowadays. Pitchers can send the ball left, right and down, or any combination thereof, but sometimes it seems like one of those directions is prominent for a few years at a time. They fade in and out of favor every few years. A few years ago the cutter was a-buzzing. Sports Illustrated even did a story about it title, “This Is The Game Changer.” There’s been backlash since, notably from the Orioles organization, which has stated that they don’t want their pitching prospects throwing cutters because they think it weakens the arm and saps velocity out of the regular fastball. They may be right, or they just be fucking with the rest of the league. Is Dan Duquette laughing in his office with his underlings, saying “I can’t believe they bought it!”? Is Dan Duquette even the GM of the Orioles still? Is this paragraph going anywhere?

Anyway, on the back of the splitter, Uehara has tossed a historic season for baserunner prevention, as Jonah Keri already described. Perhaps relievers around the league, borderline major leaguers and replacement-level types, will copy the pitch hoping to apea small part of his success. A split-finger renaissance! That’d be cool, but that’s not something we can measure just yet. We can measure how valuable the splitter’s been: the most valuable pitch of its kind in the league, and with the most value per one hundred pitches thrown among pitchers with minimum 50 innings pitched.

Right about now you are probably thinking that Uehara’s great year is the result of some lucky fluctuation–BABIP maybe, or park factors. Uehara’s always had the splitter, ever since he was a suckling babe, and in seasons past he merely very good. What gives?, say you, because you get angry when curious. Well, give me a second, damn it. I can explain.

Uehara’s splitter is good enough that he doesn’t need to instill uncertainty in the hitter to be successful with it. (Look at this page again. Of all breaking pitches, Uehara’s splitter is like eighth-most valuable in the league, even though he’s pitched fewer than 80 innings.) Every pitcher has one or two pitches that just plain aren’t as good as his feature stuff. The reason they keep those pitches around is to give the hitter more things to think about. The more pieces you have in chess, the more moves you can make. Except Uehara is finding out that his splitter works just fine even if the hitter doesn’t have to worry about his cutter and even if he throws it more than the fastball. By cutting out the worst parts of his game, Uehara improved.


w/Rangers means from 7/31/2011 to the Wild Card game in 2012

Don’t worry about the curveball. He’s threw three of them this season, so you can’t even see it on the graph. One of those three went in for a strike. Here it is.

That leaves two main takeaways: 1) Uehara has dropped the cutter this postseason and 2) he’s been increasing usage of his splitter ever since 2011. It has always been his put-away pitch, thrown more in two-strike counts. This postseason, and to a lesser extent this season, he has thrown the splitter as much as he used to when he wanted to strike motherfuckers out. And now, when he has two strikes, the odds of him throwing a splitter are less like a coin flip and more like the odds of getting a prime number when rolling a die. (EMBRACE THE NERD INSIDE.)


You would think the splitter’s value would dilute with all the extra volume but it doesn’t The graph below suggests that in some aspects it is actually getting better results.



Hitters swung at the splitter a little less but when they did they whiffed more, or if they made contact it was more likely to be into the ground, this season compared to last. So far this postseason, hitters are flailing more than ever. Despite these gains, the bottom half of the graph suggests that Uehara’s improvement is tied strongly to BABIP. The colored bars add up to slugging percentage, and their height matches up well with the BABIP line, suggesting that cursed fluctuations are at least partly guilty for that abominable .168 slugging hitters put up against the splitter this season. Even with a normal-ish BABIP of .267 (as it was in 2012), Uehara’s splitter still held hitters to .283 in slugging. And if you take out Jose Lobaton’s definition-of-an-aberration homer, Uehara’s slugging allowed this postseason would be substantially less than the still-very-low .300. There’s really no way to make these numbers look bad.

You can take a look at FanGraphs leaderboards and find that Uehara’s splitter doesn’t have exceptional movement. Neither it nor his fastball have overpowering velocity. Deception, guile and location serve to make his splitter exceptional. Here’s a zone map of his spliiter from the ALCS.

Low low low! Lower than the prices at Costco. (My mom works there. Love you, mom.) Dude hits his spots. That’s either command or control. I can’t remember how the nerds are classifying them these days. Just look at these putaway pitches from 2012–since this is a west-themed blog, I looked at his time with the Rangers. I found the games in which he was called to enter into the most crucial situations. Then I found the most crucial at-bats in those appearances. In a fit of noteworthy luck, those at-bats came against Michael Saunders, Evan Longoria, Mike Trout, and Edwin Encarnacion, three of whom are some of the best hitters in the league and one of whom is left-handed. So we get to see Uehara face hitters of both handednesses. All of these are two-strike pitches, all of the two-strike pitches thrown against these hitters.twostrikepitches


Aside from Edwin’s all-fastball at-bat, Uehara stuck with the splitter. I think the fastball to Trout was meant to back him off some, but I don’t know if Uehara is one of those guys. I haven’t heard that about him, but then again you never hear announcers call an Asian player a bulldog or a firebrand. Announcers are usually middle-aged white guys and they love Texans and big country gentlemen. They don’t like city boys, such as Zitocakes with his yoga, and they respect Asians from a distance, as with Ichiro. There are plenty of cultural barriers between foreign pitchers and the national media: language, customs, training regimens, expectations of privacy, facial hair. We know, at the very least, that Uehara has busted the last one down.


Look at those sideburns. He’s Legolas because he’s as accurate with his splitter as Legolas is with his bow. Get pumped for The Desolation of Smaug, folks.


Clayton Kershaw and the Rise to the Top

Clayton Kershaw starts Game 6 of the NLCS against the Cardinals tonight, his team down three games to two. The Dodgers play on the road, but they’ve got the best pitcher in the world on the mound, so it’s a wash, maybe even advantage Dodgers. I’m not the only one who thinks Kershaw is the best pitcher in baseball. Although recency bias helps people like me favor Kershaw over Justin Verlander and Cliff Lee, it doesn’t necessarily make us wrong. At the very least, sensible opponents of my position would have to grant that Kershaw has been the best pitcher in baseball this season.



By now you’ve noticed the thing between this paragraph and the last. It’s a way to chart Kershaw’s ascent to the Rosin Throne. Kershaw’s played just six years in the majors, and for all possible combinations of those years I’ve provided his WAR and where that WAR has ranked him among all other pitchers in the same time period. In the second chart, I showed only the ranks so I could do the gradient.

I set the midpoint of the gradient at 30, reasoning that the top 30 pitchers in the league could rightly be considered “aces.” Theoretically, an ace is the best starter on a pitching staff, so if you distributed the top 30 pitchers one to each team, they would all be aces. It’s just one definition, and I’m not sure if I even buy into it. (Of course some teams have none and some have two. Besides, the talent pool might not be deep enough for 30 bona fide aces.) But 30 isn’t arbitrary and serves well enough here. All we want is to chart is Kershaw’s rise to the top, and we’re less concerned with his stops in the middle.

Anyway, for the last four years, Kershaw’s been one of the three best pitchers in baseball. Who knows for how many years he’ll be considered the best. Maybe just this year. After all, Tim Lincecum was the best once.



You can find plenty of differences between Lincecum and Kershaw that suggest Kershaw is more likely to sustain his greatness. Lincecum is tiny; Kershaw isn’t. Lincecum’s velocity is closely tied to his springy, elasto-man delivery; Kershaw’s velocity seems more natural. Lincecum was never a free-pass machine, but he never had Kershaw’s current pedigree of avoiding walks. Still, Lincecum’s is a cautionary tale (as if pitchers needed another one): things can go bad, and quickly.

It’s time for me to stop being negative. Kershaw’s rise is truly remarkable. Here’s another guy who came up as a teenager.

felixhalfIn his first six years, King Felix established himself as a front-line starter, but not as one of the best three to five pitchers in the game. I think we can all agree that if Felix can pitch well into his thirties at a high level, he will make the Hall of Fame. What does that say about Kershaw? Could he be one of the best in recent history?

On his Baseball Reference page, Kershaw is graded similar to Pedro Martinez, Roger Clemens and Tom Seaver. Let’s look at their first six seasons (caveat: I started counting at their first season with 40 innings pitched).



No matter how you slice it, Pedro took longer to reach the top than Kershaw did. Seaver posted some monstrous WARs thanks to plenty of 300-inning seasons, and you could call him the best in his fifth year, whereas Kershaw needed six. And Clemens, oh boy Clemens. I don’t think we will ever again see dominance to that degree from a young pitcher. (And I didn’t even get to see it.) Clemens in the late ’80s was unreal, and the closest thing we’ve had since is Clayton Kershaw.


The NLCS Dodgers and Robert Andino, and Other Game 3 Notes

Your NL West champion Los Angeles Dodgers, so far in the NLCS:
85 PA, 14 H, 0 HR, 2 R, 2 RBI, 7 BB, 24 K

Your Seattle Mariners utility infielder Robert Andino, in 2013:
85 PA, 14 H, 0 HR, 5 R, 4 RBI, 7 BB, 27 K

Hell’s bells, Trudy!

It’s that bad!

And it gets worse!

Hanley Ramirez and Andre Ethier might not play, giving eight to ten plate appearances to Nick Punto and Skip Schumaker. Oh, and Cardinals ace slash playoff dynamo Adam Wainwright will start Game 3 in LA tonight. Ahem:

Source: Notgraphs

That curveball has been a playoff pestilence since 2006. Click for schadenfreude.


We’ve covered the Dodgers bench before. A collection of hypothetical Replacement Players would have performed ever-so-slightly better. Los Angeles paid for premium talent around the field, but filled their bench lazily, with creaky veterans. This roster is ill-equipped for injury and facing injury at the second-worst possible time.

A series of ESPN hot zones will help us visualize why the Dodgers are in trouble tonight. All of them show the batter’s batting average over the last two years against right-handed pitchers. (These can be found on each player’s ESPN page.)




You can be sure that, aside from the gains in batting average, Hanley and Ethier would also provide much more power potential than their backups. I don’t want to overdo it with the zone maps, so you can look for yourself at ESPN. It checks out.


Here’s one for fun. Mark Ellis isn’t a bench player, and actually he’s been a serviceable starter this year. He is in many ways a throwback second baseman, with quick hands, a good glove, solid contact rates and not a lot of power. And since Donnie Baseball is a throwback manager, he likes to bat Ellis second, so he can make productive outs and execute the hit-and-run. This is the folly of team baseball. Managers like Mattingly overvalue guys who make outs that, in some contexts, advance baserunners, all because it’s easy to point to that and say, “Aha, a consolation for out-making! All guys make outs, but this guy sacrificed himself for the team.”



That of course flies in the face of everything we now know about lineup optimization. Perhaps tactical errors like this need to be made in the limelight of the postseason for progress to be made. High-profile mistakes will raise discussion and get an owner’s attention.


Hyun-Jin Ryu probably won’t operate with much wiggle room tonight. The 26-year-old “rookie” lefty is a traditional four-pitch pitcher, with a fastball, slider, curve and change. His curveball doesn’t move like Wainwright’s, but that’s like saying he can’t run like Rickey. It’s still a good curveball, and a damn good fourth pitch.

Source: Fangraphs

Ryu’s best weapon might be his changeup. He throws it more than any pitch except the fastball, and for the season it surrendered only three extra-base hits (all home runs, strangely). Looking at just this season, Ryu’s change compares favorably to Wainwright’s curve.

  • Ryu Change: 3 XBH, .168 BAA, .213 SLG, .195 BABIP (Count: 724)
  • Wainwright Curve: 9 XBH, ,172 BAA, .226 SLG, .297 BABIP (Count: 1018)

Against the righty-heavy Cardinals lineup, Ryu’s changeup must be as extraordinary as it has been so far in his young career.

What Went Wrong with Sonny Gray

This article is not meant to assign blame. If anything, credit goes to Justin Verlander, who Verlandered really hard for eight innings. Sonny Gray, for his part, tossed a gritty start without his best stuff, and got killed on the one mistake he made to the world’s best gimpy hitter. (Rearrange those last four words all you like and they’re still true.) Unfortunately for Gray and the A’s, Verlander left no margin for error. All the margin for error was left in Game 4, it turns out. Before we lament another early exit, let’s understand what happened on the mound.

Courtesy of Brooks Baseball, here’s a scatter plot of Gray’s pitches, marked with numbers to show the order in which each pitch occurred in the at-bat in which it was thrown. Using my old friend MS Paint, I boxed all the first-pitch balls I could find with red.


That adds up to 14 1-0 counts. Gray faced 24 Detroit batters in total, so his first-strike percentage was about 42 percent, compared to 59.8 on the year. Most of them were easy takes, especially after Gray established that he lacked his typical command. Tigers hitters got count leverage without much thought on their part. How much of an advantage is that leverage? Well, as a team, Detroit hit .283/.346/.434 this year, but after 1-0 counts, that increases to .292/.399/.453. After 0-1 counts, it drops to .260/.299/.393. Respectively, those three levels of production are equivalent to Starling Marte, Allen Craig and Andrelton Simmons. Gray needed to turn the Detroit lineup into a bunch of glove-first shortstops, and certainly he meant to, but he couldn’t execute that plan. Diamondbacks pitcher Brandon McCarthy offered his hypothesis during the game.

With the pitch data already in our evidence bag, McCarthy’s statement should not be questioned. The man is, after all, an expert on this sort of thing. All I will do is try to back him up with visual evidence. Below are two screenshots, the first from Miguel Cabrera’s homer last night and the second from a Prince Fielder groundout in Game 2.


Both pitches thrown were fastballs, and both fastballs were out over the plate. In the first picture, Gray has an exaggerated lean toward first base, almost like he is in the middle of diving to stop a groundball. In the second, Gray’s back is closer to being parallel with the ground, and he isn’t falling off the mound as much. His left foot is planted more firmly than in the first picture, and his back foot isn’t swinging forward as low and violently. In the first picture, Gray’s back hasn’t rotated as much, evidenced by the increased visibility of the 4 in 54 on his jersey. Finally, for what it’s worth, Gray’s head is much closer to the “o” in “oaklandathletics.com” in the first screenshot. The camera angle may be slightly different, but that wouldn’t make much sense, and even so, the first three things I mentioned, about Gray’s body and not its position in the frame, hold true.

The difference are minute, but you don’t need me to tell you they matter greatly. Cliche alert: the physical mechanics of pitching are as intricate as the workings of a wristwatch. Sonny Gray was just a little bit off last night. His pitch missed a little bit up. Miguel Cabrera’s home run went just a little bit over the fence. The A’s, this year, last year, and in 2006, fell just a little bit short against the Tigers.

The Saga of Dan Straily’s Slider Continues Today

Dan Straily has fewer than 200 innings pitched in the bigs. In the minors, he struck hitters out like hotcakes, but that hasn’t translated to the major league level, for his career K/9 is merely serviceable at 7.33. A 90 mph fastball is far from overpowering, and–in addition to a “Young Dad Moustache“–that’s what Straily has. (Sadly, the moustache is gone now.) Despite a 20% line drive rate this season, hitters are somehow hitting only .266 on balls in play against Straily. He doesn’t induce grounders: of all 96 pitchers with at least 150 IP this season, seven have lower groundball rates. Two of those pitchers are A’s teammates Tommy Milone and A.J. Griffin. With the above information, we can safely surmise the A’s are targeting flyball pitchers for spacious O.co, rangy outfielders tracking down all things fair and infielders running free in the vast fields afoul. That strategy of defense-dependent pitching works when the ball is kept within the defense’s grasp. Straily did a good job of that this season after an abysmal time in 2012; his rate of home runs per flyball was halved from 16.7 to 8.2 percent, the latter figure being slightly better than the league average.

Numbers, numbers, numbers. Anyone with both the internet and inclination can find those, and while they may give you a good portrait of Straily, they say little about Straily’s chances in today’s start. After all, he’s pitching in Comerica Park, subject to a different run-scoring environment, against a team known for it’s above-average power. How will Straily try to neutralize Detroit? Let’s start with their lineup:

  1. Austin Jackson, R
  2. Torii Hunter, R
  3. Miguel Cabrera, R
  4. Prince Fielder, L
  5. Victor Martinez, S
  6. Jhonny Peralta, R
  7. Alex Avila, L
  8. Omar Infante, R
  9. Jose Iglesias, R

Six righties, three lefties, as far as Straily is concerned. Straily has two fastballs, four-seam and two-seam, and just two other pitches, a slider and a changeup. Generally speaking, pitchers throw the slider to same-handed hitters and develop the changeup as an out pitch against opposites. This is true for Straily, but the extent to which he throws his slider to righties is extreme.

Percentage of Pitches thrown that are sliders, 2013 season:

  • Overall – 27.49%
  • vs. RHB – 40.91%
  • vs. RHB w/2 strikes – 54.89%
  • vs. RHB w/full count – 31.82%

When he wants to get righties out, Straily throws the slider twice as much as usual. Even when the count is full and he needs a strike, his slider usage increases. I don’t know Straily’s mind, so I dare not say that the slider is his favorite pitch, but I do dare say that the slider is the most important pitch for Straily tonight. It’s still the top of the first, so here’s your first look at it today:


Click to animate and embiggen.

That’s perfect location, but not at all uncommon for Straily. Here’s Brooks Baseball’s zone profile of Straily’s slider this season, showing how well Straily keeps the ball down and away from right handers.

WOW THAT’S COOL. On Twitter, I will be tracking Straily’s slider. Follow me, internet/no one.

Why the Dodgers will Start Kershaw for Game 4

This was supposed to be a post about Ricky Nolasco and the Dodgers’ bullpen, but right after I started it I found out the Dodgers will start Clayton Kershaw on short rest in Game 4 against the Braves. Don Mattingly’s decision is curious for a couple of reasons. One, if this move works, Kershaw can’t pitch until Game 2 of the NLCS. Two, the only other time Kershaw has pitched on three days’ rest was a relief appearance in his rookie season. Yes, Kershaw is the best pitcher alive, but any number of things from his inexperience pitching on short rest to defensive miscues to gnats from Lake Erie could conspire against him tonight. Then the Dodgers will face elimination two days from now with someone other than the best pitcher alive.

But this would be a shallow article, and I a shallow writer, if all I did was disagree with Don Mattingly. Managers are rightfully secretive about many of their tactical choices, so I’ll do my best to find a good reason why Mattingly would think this is the best decision. It’s not like Nolasco has been bad. In 87 innings for Los Angeles he’s posted an ERA of 3.52 and a FIP of 3.15. Including his time with the Marlins, his 2013 numbers are only slightly worse at 3.70 and 3.34. Unless Nolasco kicked Mattingly’s dog, it’s hard to believe that Mattingly benched him because of something he did wrong.

Nolasco cannot help his own limitations as a pitcher, however. If he could, he probably would have by now, because he’s a good capitalist. Alas, Ricky struggles to pitch deep into games. In his 15 starts for the Dodgers, he’s pitched more than seven innings only three times. More than half the time (eight starts), he hasn’t gotten out of the sixth inning. This flaw doesn’t make Nolasco a bad pitcher, it just keeps him from being one of the very good ones. And tonight, Nolasco’s potential early exit would have exposed the Dodgers’ great weakness.

Here is the Los Angeles bullpen in nerd-table form. There isn’t much to like outside of Kenley Jansen. He is great, and has earned the whispered comparisons to Mariano Rivera with that cutter of his. But everyone else, bleh. Mattingly is understandably wary of finding three serviceable innings there between the end of the fifth and the beginning of the ninth. Especially since Atlanta used this lineup against Zack Greinke, a righty like Nolasco, in Game 2:

  1. Jason Heyward, bats L
  2. Justin Upton, bats R
  3. Freddie Freeman, L
  4. Evan Gattis, R
  5. Brian McCann, L
  6. Chris Johnson, R
  7. Andrelton Simmons, R
  8. Elliot Johnson, S

The six most dangerous hitters are arranged for maximum bullpen hell. The Dodgers named five relievers besides Jansen to their postseason roster. (Chris Capuano, and now Nolasco it seems, are nominal starters who can work long relief.) J.P. Howell and Paco Rodriguez are lefty specialists. Against righties, Ronald Belisario has been a slight disappointment. Chris Withrow has performed well, but it’s harder to trust rookies, isn’t it? Brian Wilson is a wild card, coming off Tommy John surgery but having pitched well in an extremely brief sample. Getting six-to-nine outs from those five pitchers, when platoon advantages are turning into disadvantages after just one batter, sounds like risky business, doesn’t it? Perhaps Mattingly is thinking: if Kershaw demonstrates that he doesn’t have his good stuff, there is at least Nolasco, good for five innings, waiting in the wings.

About Platooning in Oakland, Briefly

In our article two days ago “previewing” the 2013 ALDS, we covered how the Oakland A’s built a successful team: the front office focused on bench depth, finding cheap players who provided good value in limited, a.k.a. special, circumstances. The most common way the A’s specialize is through platooning. You know what platooning is and how it works, so all you need is a green-and-tan table that shows you how well it’s worked for Oakland.



The A’s don’t platoon at shortstop or third; Jed Lowrie and Josh Donaldson are locked in on the left side of the infield. At every other position, though, the A’s platoon to some degree. Above I calculated the combined stats for the primary platooners at catcher, first base, second base and something called OF/DH Monster. I’ll get to that in a bit. Once I calculated the platoons’ combined batting average, OBP and slugging, I looked for similar production from players around the league. Since the plate appearances don’t line up, I looked for matching rate production over similar counting stats. Look and be mildly impressed.

Now, about that OF/DH Monster. The A’s have five outfielders: Coco Crisp, Yoenis Cespedes, Josh Reddick, Seth Smith, and Chris Young. Normally, four of them play in a game–three in the outfield, one at DH. (Sometimes Brandon Moss goes to the outfield, which is why he and Nate Freiman had so many combined PAs.) I added their stats together, all five of them, and divided the totals by four, for the four spots in the lineup they occupy.

They average out to four Alex Gordons! Or four Alejando De Azas, when you account for the speed of Crisp and Young! The A’s wrung four above-average starters out of a group that is three-fifths disappointing. Cespedes and Reddick have been down all year, and Chris Young is hitting just awful. Of course defense plays a huge part in that WAR figure. Mayhaps I’ll look at that later in this ALDS. Let’s hope it lasts.