I haven't even watched today's game, yet. I have mlb.tv and no cable, so I have to wait about two hours for the replay to be available on demand. But I saw the results. Liriano actually appears to have pitched well.

An oxymoronic "superfically deep" analysis suggests his performance was "really" better: 9 of 22 batters faced struck out, 7 swinging, only 2 walks. Sure, he wasn't inducing groundballs like vintage Liriano, but beggars and choosers and all that.

So I took a look at the pitch f/x data on brooks baseball to see what, if anything, he did differently.

Continue reading this post »

Liriano pitched wonderfully again.In his dominant Seattle start the Pitch F/X data suggested he was throwing the ball with wicked movement (as against pedestrian movement during the dark days of April/Early May). His success yesterday appears to have relied at least as much on location: he simply avoided throwing meatballs. His Pitch F/X movement data suggests he had did have pretty good moving stuff, although he did not have "Mariners-start-type" movement (as a certain "analyst" would put it), but his velocity numbers were actually down across the board relative to earlier this year.

Continue reading this post »

This Liriano thing just gets more interesting and more weird. During a lull last night in the parade of punk rock and or roll records I generally have working while watching Twins games on the good ol' Roku, I heard something that recalled the questionably funny Arsenio Hall Show segment: "Things That Make You Say 'Hmmm'".

During the 1st inning, the organization's primary propaganda mouthpieces DicknBert engaged in what at first appeared to be some unmitigated walking back of the narrative they'd been spewing for most of April and early May: Liriano had succeeded against Seattle, they said, because he was now pitching "his way", not the "Twins way". They actually said "The Twins Way" while admitting it hadn't worked. I was stunned. The Twins organization FALLIBLE!?!? Unpossible!

As my post discussing his Pitch F/X numbers vs. Seattle on the 17th stated, it was abundantly clear to me that Liriano had thrown differently against Seattle than he'd been throwing since the first start of the season, when he had lively stuff with a ton of sink but wasn't given borderline calls, walked 5 and was scorned "in puhlic, if you wheeel" by every live body the Twins could set to the task. On the 17th Pitch F/X said his primary fastball was a two-seamer with fantastic sink, his slider had crazy left to right and downward movement, and his change was dropping off the table. The results were certainly sexy.

So now he was to be left to his own devices? Huzzah! Or Yay. Or something like that.

Then Bert (I think) proceeded to extrapolate, saying something like "Yeah, Liriano, he's gonna go with that four seamer, not worry about having to throw two different fastballs." Wha-wha-wha-WHAT!? Last year when he was, y'know, kinda good and stuff, he threw one fastball 42.7% of the time and the other 6.6% of the time. The problem is, Pitch F/X called 2010's (overwhelmingly) primary fastball a TWO seamer, just like it did in start 1 this year, just like it did against Seattle on the 17th.

As I alluded to in my previous fanpost, there are problems with going with the Pitch F/X software's identification of a pitch, especially historically pre-2010 between two and four seam fastballs. My analysis consequently just decided to say "What is the movement like on the fastball he's throwing most of the time, regardless of what Pitch F/X calls that pitch?" But the thing is, the reason the software calls his supposed two-seamer a "two-seamer" is because it moves like one: tons of run-back, good effective sink (i.e. less rise v. a backspin-less pitch). That's just not how a four-seam fastball usually behaves. So it's reasonable to suppose that it at least might be, in fact, a two-seam fastball, right?

Back to DicknBert. They had made a huge deal back in the "Twins Way Is The Right Way" chapter of Liriano's 2011 campaign about him dropping the two seam fastball.  Again: when he was playing the unsuccessful but Good Soldier, the organization's mouthpieces were (1) lauding him for not moving on the rubber (since scrapped, pitch f/x appears to show he was moving again vs. Arizona as he had when dominating the Mariners) and (2) they were pimping that he was throwing only the easier to control four seam fastball. Wait, what? So he was now pitching "his way", but part of "his way" suddenly included the Randerson/Gardenhire directive to scrap the two-seamer, despite the evidence that the two-seamer is what he had used when dominating Seattle and all last season?

Needless to say, I wondered if this was just face-saving talk for Bert and the Twins. Maybe the pitch f/x data would show that Liriano was still throwing the two-seam fastball he used last week. That must be it!

The data is in, and it's perplexing and intriguing. According to Pitch F/X, he indeed threw 49 four-seamers and only 9 two-seamers against the Diamondbacks. We know he didn't dominate, although he pitched pretty well and generated a good number of swings-and-misses. (BTW, he got zero B.S. strike calls and had 3-4 strikes called balls, per Pitch F/X.) Did he actually change what he was throwing, and what did whatever he throw look like according to Pitch F/X?

Velocity numbers were steady with the Seattle game with the exception of his change being exceptionally slow (which is obviously great if he maintains deception).

Taking a look at the movement numbers, there's a pretty short story to tell, pregnant with possible implications. His primary fastball vs. Arizona ("four-seamer" per both pitch f/x and DickNBert) had ordinary Vertical-Move (i.e. no effective rise or sink) for a four-seam fastball but a lot of horizontal movement. Indeed, it ran-back more than his (overwhelmingly primary) two-seam fastball did in some 2010 starts, and a lot more than MLB four-seamers generally do. Weird. Kind of a hybrid, but this isn't a good thing: think about a two-seam running fastball with zero effective sink. That's what last night's fastball looked like, and that's what his primary fastball looked like in April.

His change-up, too, had a bunch of run on it, but it went from falling off the table to sitting flat, sinking less than last year and a whole lot less than the great pitch he had working against Seattle. It wasn't as craptacularly flat as it was through most of April, but still: not good movement wise. The velocity change/improved separation could easily make up for this if he disguises the pitch, though.

Liriano's slider had pretty decent left-to-right break, although nothing like it had v. Seattle (especially once three "sliders" that were probably fastballs were removed from the Seattle samples, as discussed). It's downward movement was right back in line with the mediocrity it's had for most of 2011. The insanely nasty pitch he had in Seattle was gone, but the improved (vs. this dark days of this season) boring action on right-handers should generate weak contact and sawed off bats, if not as many swing-over whiffs.

So what to make of all this? The Pitch F/X data pretty much says "he pitched more like he was pitching earlier this year than like he did against Seattle or last year, albeit with a little more left-to-right action on his slider and a little more velocity separation on his change than he's had in 2011." So did he just "have great stuff" against Seattle and go back to his "average stuff" against Arizona? Maybe. (Seriously: maybe.) Three other possibilities occur.

First: Pitch F/X calibrations may vary from ballpark to ballpark. Perhaps in Seattle he was throwing the four-seamer and "the same" sort of change and slider and the cameras and software were just determining that everything had downward movement it didn't really have, and reading all the horizontal movement as shifted slightly to the right. That would explain why his primary "two-seam" fastball and change-up didn't have quite the run-back against the Mariners they often do (although the difference was not huge).

The problem is, I looked at Felix Hernandez's pitch f/x data. He's right-handed, so if the cameras and software in Safeco are biased to read more movement as more rightward/less leftward than it really is, his primary two-seamer should show awesome run and his breaking stuff should break right to left less than expected. And all his stuff should evidence the same vertical drop Liriano had.

Hernandez's slider indeed had less right-to-left movement than it's had in most (but not all) starts over the past year, but it was also very flat, having a lot less sink than it does when "normally nasty". Similarly, his curve had less sink than it's had, while breaking left only a shade less than average (and within a normal range). His two-seamer/sinker had pretty good sink, although within his normal range. Its run-back was ordinary for him and for a two-seam fastball in general.

In sum, maybe the Safeco Pitch F/X cameras "gave" Liriano's slider an extra inch of left-to-right break, but he had a ton of sink whereas Hernandez had his second flatest slider of the year, so I don't think Safeco's Pitch F/X cameras/software (systematically or "for one night only") are faulty. The only possible savior for this hypothesis is that the cameras/software were somehow working differently for right and left handed pitchers. Glen Perkins had insane sink on his two sliders and very good sink on his two-seamer (SSS!), so I suppose this is possible, but so did right-hander Brandon League, so I wouldn't bet on it.

(BTW, Hudson's Pitch F/X numbers are the picture of game-to-game steadiness, so Arizona's data would seem sound, as well.)

Second possibility: Liriano changed what he threw, deliberately as directed. This is the "Kreminology" hypothesis, and it would hold that DicknBert talking about how Liriano was going to be throwing exclusively the four-seamer meant something. They weren't speculating idly. Lip-service to Liriano pitching "his way" aside, the Twins had told him to scrap the two-seamer he threw so effectively against Seattle and they meant it. He did and his primary fastball lost five inches (i.e. tons) of effective sink. Throwing the four-seamer (or perhaps "backing off" to try to throw deeper into the ballgame) somehow "spilled over" mechanically and affected the slider and change-up, which flattened out.

Third: There is an armslot or other "big" delivery mechanics issue at play. Against Seattle, he did something differently mechanically (armslot, stride length, whatever) that put a lot less pure backspin on everything, and his pitches accordingly dropped off the table. DickNBert may be correct and he may have been throwing the four seam fastball v. Seattle, too (but with a result that looked exactly like a two-seam sinking fastball "should" look), but whatever he's "supposed" to be throwing he needs to figure out how to throw it with the action he had last week.

I'm very curious whether anybody noticed anything different in his delivery. I'm curious whether the fastball he was throwing that had 8+ inches of run-back was somehow a four seam fastball. I'd like to think DicknBert were just babbling - recall when Bert declared Liriano had just started to move around on the rubber months after he'd started to last season - and that the Twins wouldn't try to change what Liriano did in Seattle. Can he succeed with just the improved horizontal movement on his slider (which, by the way, was the only standout feature of his no-hitter) and change-up separation?

To conclude with a little visual aid, here's his break chart versus Seattle followed by versus Arizona, with gravity. Notice how much more sink his pitches had in Seattle.

via pitchfx.texasleaguers.com

And vs. Arizona

via pitchfx.texasleaguers.com

Well that certainly was exciting. More exciting still, the pitch f/x data suggest Liriano was indeed throwing very differently last night than he has this season (with the very intriguing exception of game one: see conclusion). Indeed, he appears per pitch f/x data to have had stuff resembling what he had last year, if not even a little better.

To be clear: the velocity is about the same as it's been throughout 2011. It's the movement that's n'other class.

The H-move and V-move measurements pitch f/x provide can be a little counter-intuitive, so I figured some words about them might be necessary. Skip to "after the jump" if you understand pitch f/x numbers in general.

(By the way, I mainly used the graphs at the bottom of Liriano's fangraphs pitch f/x page for this, although I also used the single game charts to I.D. the extremes of awesome to which his change-up got at times last night.)

Basically, H-move and V-move measure the break on the pitched ball relative to a ball punched out of a theoretical pneumatic cannon (or whatever) with no spin whatsoever, thrown from the same release point at the same initial angle. The last bit is important to internalize if you're going to believe the data, as the ball is always thrown from a couple of feet off the side of the rubber starting across the pitcher's body, angled downwards. Forgetting that and imagining a ball thrown literally straight from the middle of the pitcher's mound to home plate, parallel to the ground, makes the data impossible to believe/understand.

A fastball's "h-movement" number is always back towards the pitcher's throwing hand (think about the "sideways component" of a normal 45 degree arm angle and the backspin of the seams coming off the fingertips and it makes sense). A two seam fastball gets more h-movement (run-back) than a four seamer: good running two-seamers often move 10 inches or more "backwards". For example, in 2010 Liriano averaged 9.3 inches of run-back with what was then his primary fastball, the two-seamer.

This next bit might be weird to accept: hard breaking balls (e.g. sliders) often don't actually break sideways in the way they are perceived to, but actually simple run-back toward the pitching-hand side a lot less. Good ones break, although some pitchers' high H-break numbers are offset by relative flatness (V-movement, see below). Last year, Liriano's slider had half an inch of true break on average, although in many games he had 2 inches or more of break working.

Changeups move horizontally like two-seamers, running back to the pitching hand side quite a lot.

Vertical movement is at least as important as horizontal movement. Four-seam fastballs (and, consequently, any fastball the pitch f/x software identifies, rightly or not, as a 4 seam fastball - you cannot always trust pitch f/x's classifications of fastballs, especially in years past) have a ton of "rise" - 10 inches or more is common - relative to the theoretical spinless ball. A great rising fastball is believed to be what allows dominant closers to consistently post high pop-up rates.

Although a "sinker" is generally just a two-seam fastball, two-seamers don't actually sink relative to a theoretical spinless pitch projected at the same initial angle: they just rise a lot less. (Of course, two-seamers do actually sink relative to the earth, thanks to gravity. They would just hit the ground a little sooner if not for the backspin on them.) For example, Nick Blackburn's world-beating (ahem) sinker has about 5 1/2 inches of rise on it relative to a spinless baseball, but the effect of this in conjunction with gravity and the extra h-movement (run-back or tail) on his two-seamer is that it dives down and back relative to where a "regular" four seam fastball would be, thus it "sinks".

Change-ups have a lot less rise than four-seamers: generally a V-Movement number of an inch or two less than a two-seam fastball. Again, though, they don't actually break downwards compared to a spinless pitch (i.e. they don't have top-spin), despite perception. Unless they're really, really wicked (hint, hint).

Sliders have very little rise on them, even less than change-ups, and this is what makes them appear to drop off so dramatically. But they also don't always actually net out to significantly negative V-Movement number due the difficulty of putting topspin on a ball thrown that hard. Typically. Even Liriano averaged out to 0.1 inches of rise during his dominant 2010, although he had negative V-Movement (true sink) in 17 of his starts.

So what about Liriano last night?

Continue reading this post »

Fangraphs has posted the first UZR numbers for the 2011 season. I think I've seen every inning this year and had remarked after his shocking diving catch th'other day that I had yet to see Delmon botch a play this season, so I'm actually not surprised his number is laughably good so far.

Did I say laughably? SHAME ON ME!

Ultimate Zone Rating is admittedly a statistic that need a large sample to be reliable, but in this case, I totally, definitely, certainly think we can safely make a joyous exception. Delmon Young, Twins fans, hasn't just turned the proverbial corner defensively, he's demolished the building around which he was walking, or running, or driving (or more likely gliding effortlessly past en route to his latest webgem) or whatever. THIS is the real deal. It must be. It has to be. No, it is. Surely the first (Bert says "Rawlings") Platinum Glove in MLB history will be awarded to the Crimson Ghosted one this fall.

Delmon Young's UZR/150, a number at which some may scoff after a mere two weeks but which is in indisputable fact so rock solid in this particular, exceptional, iconoclastic case you can take it to the bank and withdraw it next year to find you've turned a 300% profit and can buy that new deluxe panini-maker you've been eyeing up and down all winter: 37.8 defensive runs above average. Money babies. Money. When his bat clicks, we're talking about 10-15 WAR! And it's all for the love of the love of Daisy.

Of course, the unparalleled gazelle-like LF work of Carlos Lee does just shave him for the league lead, but c'mon: small sample size there, obviously.

Other Twins UZR/150s to date after the jump, if'n you care.

Continue reading this post »

Well, here's a couple looks at what Vanover did to Liriano and the Twins today. Two looks, from the ump's perspective. Basically, he refused to call a low strike. The Yank Me pitchers didn't throw the ball there, and thus weren't hurt. The Twins did a great job of "keeping the ball down" (where have I heard that phrase... hmm... tip of the tongue...) but Vanover evidently felt the Yankees should only have strikes called on indubitably homerable pitches. You can't ask for all those balls to be strikes, but some of them?

In addition, a look at the second picture shows he had a few cracktacularly bad "ball" calls (green triangles/squares pretty much smack in the middle of the zone) against both teams.

The Twins seem to have this problem every year when they play the Yankoffs and Bahston on the road. You can say he should have adjusted, but it remains that Liriano was making great pitches over and over and getting virtually none of them called. There are like 10 low-edge pitches Liriano could easily have had called strikes that weren't, a handful more just a couple inches off the corner, and only one strike call in his favor (belt high just off the plate).

BTW, clicking on these should make 'em bigger. Used non-normalized for the second because the triangles piled up on top of each other too much in the normalized version.

via www.brooksbaseball.net

via www.brooksbaseball.net

It's spring training.

They're the Houston Astros.

But still.

Charlie Morton: 6 IP, 21 BF, 5 K (3 Swinging), 0 BB, 7 GO, 3 AO, 4 H, 0 R.

Could've been worse.

Sign of things to come/corners turned/true talent, or utterly meaningless March irrelevance? Something in between?

A mere distraction, because we all just know the Bucs could have had at least half a Royals-esque farm system in exchange for world beating all star Nate McClouth if Huntington wasn't so diligently carrying out Nutting's orders to continue losing so as to make millions off of revenue sharing? *cough*

So I ended up thinking a lot about and looking up a lot about and typing a lot about Ben Revere as an OF in the fanpost thread re: UZR. I figure I'd throw it up here and see what people think. It may seem like a no-brainer to stick him in left given his less-than-stellar arm, but poking around and looking at just how many defensive runs there are to be had in the various positions, I'm not so sure.

Here's what I wrote:

There is of course one throw that is longer from RF than from LF: the throw to 3rd. Having borderline guys take third on your right fielder is not good, so nearly every team with a noodle armed outfielder puts him in LF.

But where are most balls in the air that hang up for any length of time hit? Right field. This is a neat look at why.

OK, now if the Twins call Revere up, we can pretty safely assume the guy will bring extremely good range to any outfield position. If the "other" corner is still Delmon or Cuddyer and the "range" one of them brings to the table, there may well be more to be gained through putting Revere's range in RF (versus LF) than lost due to his below average arm. Taking away a bloop single altogether (of which most are hit to right field, since most hitters are right-handed [see the "neat look" link above) is just SO much more valuable than preventing a guy from going from first to third.

The same goes re: putting Revere in CF: taking away an extra base hit in the alley or over the head or a sinking linedrive would-be single is worth many extra bases taken by baserunners.

It comes down to how much of each circumstance happens, and how many more extra outs Revere would convert with his range in right (or center) field than in left field. Based on the fact that fielding metrics show a much wider spread of likely true talent for range than arm (Here is every RF with more than 300 innings, click "Arm" and "RngR" over on the right side to sort, maybe throw out the top couple on either end to simulate regression and eyeball the distribution), there's little doubt which it's more important to have in the absolute sense. (Answer: Range.)

And clicking over to left fielders the spreads aren't as wide as in right, which tells me there are indeed a lot less opportunities to help or do damage there, both with range and arm.

As was pointed out, Target Field has a big left field, and this might skew things a bit in favor of a rangey left fielder. However, it also seemed (standard ANECDOTE warning!) there were a ton of fliners hit to the right field alley last year for extra bases that Kubel or Cuddyer were only a couple steps from getting to.

In any case, the Twins will very probably put Revere in left if he comes up in 2012 and it'll all be moot, so we should be thankful Target Field is built such that his range will have more utility than it might in the average LF.

And this will be good. Delmon was -11 runs UZR/150 last year. The 32 year old version of Juan Pierre was 13.4 UZR/150 in LF last year (and I've never heard Revere's arm is as bad as that, although it's worth noting that Pierre's arm only actually cost him 2 runs). Assume Revere plays 2010 Pierre-like defense, which seems safe, and that's 24 runs, 2.4 wins/150 games. Damn girl.

OTOH, if he plays right and posts Jay Bruce's 2010 range of 19 runs saved per 150 games (he's faster than Bruce, though; indeed I believe he's faster than every single right fielder on the list, so he could be better) and his arm costs him 7 runs over such a span (this would be the same as the absolute worst rate for a right fielder in MLB), that's +12 runs UZR/150 vs. Cuddyer's -18.4 UZR/150 and Kubel's -16.8 UZR/150. 30 runs/3games! Even more improvement than replacing Delmon! It's possible that a healthy Cuddyer (albeit a year older) might be more like -12/150 or so, so it'd be a similar swap to the LF move.

But you know, looking at the CF talent spread and assuming Revere's speed is what it looks like it is, I think I like Span in right and Revere in centerfield, arm and all (with Cuddyer/Kubel in left). Click on ARM on the right side of this list of center fielders with 300 innings and you can see arm just ain't the big deal it's cracked up to be in centerfield (even allowing for some selection bias) but range can get you a ton of runs. There's a limit to how much help Revere's presumably extreme range gives you in RF compared to CF, where the XBH-denying potential is so much greater. You move the underperforming RFer to LF and you minimize his damage (by minimizing his chances, as discussed above), you get what should be Bruce-like range in RF and improve your CF defense, too.

Of course, these numbers are probably just bunk, etc...

This was actually pretty interesting to look at. I think I'll copy it in a fanpost.

EDIT: some additions from the other thread worth (I hope) adding:

Right Fielders save the most runs coming in to snag would-be bloop singles in front of them or tracking balls hit mid-distance down the line. Those are the balls that hang up, hit weakly on late or poking swings by right handed bats. Range makes a huge difference in terms of how many are outs and how many are singles (or doubles down the line). Truly DRIVEN balls tend to go out of the park/off the wall/be truly uncatchable (and hit to the pull field, which is LF, more often than not). The deep-playing RF alley in Target Field means a better Right Fielder might be able to save a bit there vs. say, the dome, too.

I was, as I tried to make clear, quite struck by how little CF arm seems to matter, for all the fuss people make regarding Revere’s arm not playing.

I mentioned selection bias above regarding this. That is, the idea that maybe arms APPEAR not to matter because all MLB center fielders have good arms, so nobody's terribly far from the really existing average. I said that I thought selection bias wasn't that important.

Here's a bit more explanation: the spread of arm factors in CF is narrow. If you look at 2010 CFs with 300+ innings and throw out the top outlier on each end (simulating regression to the mean and much better representing the group as a whole, in any case), the range is only +3.9 to -3.5 runs saved or given over the course of an entire season, which tells me center fielders don’t make enough throws where the outcome is in much doubt for it to count for nearly as much as people suppse. (RF, same formula, -4.8 to +6.2.)

Now, when he was a fulltime CF, Juan Pierre’s yearly arm factors averaged out to all of 5 runs below average. Remember, the selection bias idea would depend on all those CFs having strong arms AND that arm having real value, but the value being hidden by the relatively equivalent strength of everybody's arm. But Juan Pierre has a PATHETIC arm (I watched a lot of White Sox Baseball last year. Oh god is it bad), yet he was only thereby costing his team 5 runs (while posting range-run numbers from slightly to FAR above average). If CF arm mattered as much as some people think and there's a selection bias towards "cannons", it's not possible that Pierre's arm numbers would have been so inoffensive.

Bernie Williams in his twilight years was brought up as an example of what a bad arm in CF can cost a team. His last season arm-runs in center field, extrapolated as if he'd played as much as the busiest CFer in baseball last year (1350 innings) would have been worth about 8 runs below average. Let's say Revere's arm is THAT bad. (Which it really probably isn't.)  If his range in 1350 innings is +18 – which he has the speed for it to be – IT DOESN’T FREAKING MATTER. He'd still be a great center fielder.

And again, I don’t think Revere’s arm is last-year Bernie Wililams or Juan Pierre bad.

Let’s suppose Gardenhire’s love-of-grit means Tolbert keeps his job (and actually manages to be decent when called upon) and Hughes goes to AAA. Let’s suppose Hughes tears it up. Let’s further suppose that while Casilla is actually decent and Nishioka’s pretty good, (I raise this because I think this is firmly in the realm of the possible) Michael Cuddyer spends the first half of the season (say, until mid or late July) sucking as much as he did last year (when he posted an atrocious .4 WAR in SIX HUNDRED SEVENTY FIVE PAs).

Is there ANY chance the Twins cut bait outright on Cuddyer, give Kubel the RF starts v. RHP and call up Hughes as full time RH DH/RF (rather than utility guy)?

In other words, is Hughes's only hope of regular playing time on a team already effecitvely carrying 2 DHs a Casilla fail, and how deep does the organization’s love of Michael Cuddyer’s declining sub-averageness and its inability to identify such go? Is there another organization in baseball (other than KC w/Jason Kendall) who would have so constantly and insistently touted the virtues of a guy who was by every objective measure a pretty bad MLB player? (team MVP barf)

EDIT: Since I'm not sure I made it clear enough, I am talking about a mid-season scenario, NOT the Opening Day Roster.