A quick post-mortem (literally) on Dickey’s start last night…remember from last post how sitting consistently around 79 mph was one of the keys to his Cy? Here’s a graph of his velocity (all types of pitches) from one of his stellar starts – a 12K one-hitter against the Rays as an example of how he’s throwing when at his best. Except for fastballs and at the very end of the game, Dickey threw over 80% of his knucklers in the 78-80 mph range back then:
And now here’s the same thing for him last night:
Not so much. It’s hard to tell, but it almost looks like he’s gone back to the ‘slower’ version of his knuckler that he for the most part ditched last season. Dickey also threw more fastballs than usual, including two that left the yard off the bat of Will Middlebrooks. The first time, Dickey seemed frustrated after allowing 4 consecutive hits to start the game, and put it on a tee for him on the first pitch — perfectly down the middle, straight as a board, letter high, 84 mph. Surprise value kinda goes out the window when you throw a meatball like that.
The second time, Dickey fell behind 3-0 on knuckleballs and then threw three fastballs in a row: 1) Called strike, 2) swinging strike 3) giant bomb on exactly the same pitch as in the first inning. I kind of get it if Dickey wasn’t feeling the knuckleball anyway, but seems like an odd call. Middlebrooks, again, was not so surprised.
Incidentally, I was in the first row behind the VIP section for that debacle (the best seat I have ever had and most likely ever will) and Dickey’s knuckle looked awesome as he warmed up between innings. He threw a half-dozen wicked darters for strikes that made me gasp, but once the cameras were on seemed to have trouble finding the plate with anything other than the slower, tumbling variety.
To make a long story short, as soon as R.A. Dickey was acquired this offseason I set off to write the definitive work on why he was so great last year and what to expect in this one, and completely and utterly failed. It turns out quantifying and/or predicting the evolution of that damn pitch is just about as hard as hitting, calling, or catching it…I mean, let’s face it — the man has discovered a pitch that nobody has ever thrown before which makes him great at an age that nobody has ever become great. There is simply no comparison, no career arc, no regression to the freaking mean. And if you saw any of his shutouts last year, it was clearly not a sample size or once-around-the-league-with-a-new-pitch type of thing.
Ah well…not that anybody made any solid predictions as to where his numbers were going to wind up this year on the spectrum running from a) pretty decent as he has been for most of his late career through b) Cy Young winner as he was last season, all the way over to c) best-ever-what-the-heck-is-even-going-on-who-gave-this-guy-the-cheat-codes-to-baseball that he was for about a month in June.
And I did make some pretty graphs as I floundered about. Here’s a nice one, about the speed of his knucklers last year (red) compared to his previous seasons (blue):
Averages can be misleading…Dickey actually threw fewer knucklers over 80 mph last season than in the previous two — but he used to throw two slightly different knucklers, one in the low 70′s, and one in the high, and then started to favour his harder one heavily last year. Here it also looks like he managed to bunch up his fast knuckler at the top of his range (without overthrowing it) much more often as well.
Annnnd this next one might shed some light on why that is the case:
There’s probably a better way to show this than bar graphs, but before last year, Dickey could only get his slower knuckle over the plate for strikes consistently. 50% is kind of a magic number for me in terms of first pitch strikes, fastball strike %, etc…anything below it is either trouble or a chase pitch. The fact that Dickey was able last year to throw his 74-78 mph knucklers in the zone more often than not has to have been a huge factor in his instant success — it makes his strikeout pitch the same as his get-me-over pitch. And even when it wasn’t over the plate it was getting more strikes — it’s no big surprise that hitters miss Dickey’s knuckler more the faster it is — but they also swing at it more often.
|Speed||In Zone %||Swing||Miss %||Chase %|
|Under 75 mph||53||25||40||15|
Now here’s where I got kind of lost…Dickey throws harder knucklers on 0-2, 1-2 counts, and although side-to-side movement on a knuckler is basically random, his harder ones do tend to stay ‘up’ more. Here is a chart of movement, with pfx from side to side and pfz the vertical ‘rise’ on his knucklers compared to gravity.
|Under 75 mph||-1.03||-0.26|
So while the movement on his pitches is random, you could say that R.A. has some control over his knuckler in that he locates with softer ones and then puts hitters away with his hard ones — but last year, he just threw them all hard and in the zone and was incredibly effective as a result.
Also, Dickey’s knucklers that veer up and away from right-handed batters are by far his most effective – not entirely surprising, since that is the the only pitch movement that is totally unique, and not replicating the general movement of either a curve/slider (down and away), changeup/2-seamer (down and it) or 4-seamer (‘up’ and in). The only other pitch that sort of does that is Mariano Rivera’s cutter, or the Doc’s at his best — they travel almost exactly side to side instead of like a semi-slider.
|R.A. Dickey Knucklers by direction|
|Dominant Movement||In Zone %||Miss %||Swing %||Strike %||Hit %|
|Up left (into RHB)||0.50||0.33||0.30||0.43||0.0600|
|Up right (away from RHB)||0.47||0.49||0.29||0.49||0.0427|
Hitters swing through them almost twice as often as ones that sink down and in, and get fewer hits when they do make contact. Especially from the television perspective, it’s a strange thing to get used to — his best pitches are the ones that stay up and over the plate, which for most other pitches is a sign of a mistake. Coming off a start where he didn’t have the greatest control it will be interesting to see if he goes back to his slower knuckler to try to get more strikes until he can consistently locate the hard one that won him the Cy.
Seeing that I don’t even have the time to wrap up the rather interesting stuff that I’ve been working on for months, I think this blog is just going to become a depository for completely ridiculous statistical statements this season, as was its original intention. So here weeee goooooo! Arash Madani, are you freaking kidding me?
Anthony Gose’s ascension to the big-leagues was fast. His progress in the show –15 stolen bases and a batting average of .340 with balls in play through 56 games with Toronto last season — was encouraging.
As I’ve ranted, BABIP is interesting for pitchers, and almost completely useless and/or misleading for everyone else. Having a high BABIP is equally likely due to a low contact rate (or luck — HR hurt it, too). If you take huge cuts and strike out instead of making weak contact on every 2 strike count, congrats! Your BABIP goes up! (Isn’t this just a terrific modern stat?) Obviously Madani trying to imply that this means Gose’s speed means he often reaches base when he does put the ball in play, but the real reason is that if Gose’s AB were extrapolated over an entire season, he would have struck out the 5th most times in major-league history (with only Adam Dunn and Mark Reynolds above him). Ummm, yeah…that’s highly encouraging for a powerless slap hitter with great speed…totally…
At first I was excited to learn that Romero is using pitch f/x information to diagnose his 2012 woes, but that passed quickly to outrage at the fact that it was provided in the form of printouts from Brooksbaseball by Brandon freaking Morrow. It continues to flabbergast me that teams show basically no interest in applying modern technology and analysis to get the most out of their multi-million-dollar investments on the field.
Romero is obviously open to and interesting in using this type of information, and the pitch f/x system takes an incredibly detailed picture of every pitch he throws. Yet, it takes a combination of internet hobbyists and a fellow pitcher taking the initiative for such stunningly useful and freely-available information to make its way to Romero, so he can be “amazed” by it. Le sigh. With apologies to Archimedes: give me a laptop and a place to stand in the dugout, and I will change your WAR!
Because yeah…in my not-so-humble opinion, Morrow botched the analysis (although obviously a standing ovation for him doing anything at all). When you look at more of Romero’s career than just the last two seasons, there is just no way to come to the conclusion that the key to Romero’s struggles is a lack of sinkers. He threw considerably more sinkers to RHB in 2012 (19.3%) than he did in 2010 (14%), and almost as many (16.1 compared to 17.5) to LHB. It was a mix that worked for him fine two (and three) years ago, not something that changed before the debacle that was Romero’s 2012 season.
***Note*** It’s up for debate/kind of arbitrary where your pitch-classifying algorithm draws the line between fastball and 2-seamer/sinker, as there is overlap in terms of movement as well as velocity. The way I split them, it’s a much more modest overall drop than Brooks’ numbers in terms of 2-seamers anyway, from 25% to 18% between 2012 and 2011. Here are a couple of graphs of pitch f/x ‘movement’ to give you an idea of the blurring between the two offerings, and because my contract stipulates I have to include at least one (x,y) chart per article:
The real reason for there appearing to be a sudden drop is that Romero’s sinker use soared in 2011, most likely to compensate for his AWOL changeup. As I argued at some statistical length a while back, everything else about Romero’s repetoire pales in comparison to the fact that his change now moves like a completely different pitch, with almost 6 (!!) inches more drop than average. But, with that transformation it has turned from being his primary offspeed pitch (and a great one at that) to hittable trash that he is quite rightly completely unwilling to throw to lefties.
Unless Ricky is trying to redefine himself as a true sinkerballer, the 26% of 2-seamers he threw in 2011 is very high, not some kind of career norm that he needs to return to in order to find his old levels of success. The one good thing about Romero’s sinker comments was that he seemed to be willing to throw it for strikes and let it be put in play. The percentage of Romero’s 2-seamers that were located in the zone fell to an ludicrous 34% last season, so concentrating on pounding the zone with it in spring training can’t hurt. Romero certainly should be getting more strikes and grounders by throwing it to contact instead of as a chase pitch — but he’ll never be the Ricky of old without his changeup.
I probably would have done more of this if I wasn’t completely statistically gassed by the time the fact that Josh Johnson even HAD a curveball rolled around. But as pointed out in the comment, any excitement over a new pitch has to be tempered by the fact that it won’t be a novelty for long. Only time will really tell, but here’s a look at the evolution of JJ’s new pitch as the year progressed:
1a) Movement – Side-to-side (pfx)
Pfx is where the ball ended up along the horizontal plane compared to where it would have without spin, in inches. Clearly, there is a lot of variation from one pitch to the next but also a downward trend, which means the pitch moved less from right to left as the season went on. Also, there was less spread, which is most likely a sign of increased consistency from JJ. His first 200 curveballs had a standard deviation of 1.93, and his last 230 were at 1.79 (to try to get at what that means in baseball terms would be stupid).
Also, see that patch right over the 200 mark where every single curve is below 4? That corresponds to July 4-18, 3 starts where JJ got shelled in an otherwise excellent stretch.
1b) Movement — Up-down (pfz)
I took the liberty of splitting these pitches up by start to address the elephant in the room — right in the middle of those three crappy starts mentioned above, Josh Johnson’s curveball dropped an extra half foot for one start (immediately after the All-star game, incidentally). Then for the rest of the season, the ‘drop’ on his curve steadily decreased. Experimenting with a new grip? Broken nail? Craft your own narrative…
1c) Movement — Bend (break_length)
Looks like an increasing amount of ‘bend’ early in the season, then a shaky patch around the all-star break, and a downward trend at the end of the season. The single start highlighted in red is his August 25 start, which was his one other blowup of the year after the break. For some reason leading into that start his curveball started to flatten out, and then was atrocious as the Dodgers pounded him for 10 hits and six runs through three innings.
I’m not trying to imply that it was all due to his curveball, but I find it really cool that there are usually completely obvious reasons for a sucky start rather than all the hocus-pocus about ‘not being sharp’, or ‘leaving a few pitches up’ we hear. I mean, of COURSE he’s not going to spill the beans on national TV after the game, but you know JJ was thinking ‘what the heck is wrong with my curveball, we gotta get on that’ after that game. Someday, every pitcher will get post-game reports like this, as umpires do (did?) with QUESTEC.
2a) Results — Miss %
What I did here was plot for every swing whether the result was a hit or miss, and then run a rolling average (series=15). So apparently the league started to make more contact against JJ’s curve after a certain point
2b) Results — Hits
Same thing, with hits/balls put in play. Sample size is starting to become an issue, but that’s a very clear spike in the middle of the season.
2c) Results — In Zone %
2d) Results — Chase out of zone
Just including these to be complete…nothing really pops out as conclusive for me.
- Johnson tightened up his Curve into more of a pure 12-6 as the season went on.
- He was getting more and more drop on it until the all-star break, when something went horribly wrong. From then on, that trend reversed.
- It was getting more and more ‘loopy’ until late in the season, when again, after a terrible start, things started moving in the opposite direction.
- The league was more and more able to make contact with his curve as the season went on.
- This lead to more and more hits coming against it until the midway mark, at which point, yep, things reversed dramatically.
Lumping all this together, I would say that Johnson was working on making his curve more and more nasty as the league was catching up to it, and then somewhere around the mid-way point (possibly spurred on by it falling apart) wisely switched his focus to making it more controllable so he could place it more effectively. But that might be just because I know that’s what happens to every rookie pitcher and it makes sense that it would for a veteran with a new toy…
3ai) Random — average # of Strikes in count when throwing curveball
Just a cute, not-particularly-scientific way of seeing if he was throwing it more often later in the count, i.e. a punchout pitch, as the year went on. Yep. With one crazy stretch where he gave up on it. Doing the same thing for this slider suggests that he went back to using his slider later in the count after the break:
3aii) SLIDER — average # of Strikes in count when throwing slider
Annnnnd I’ll just bury at the very bottom of another long post a semi-apology to Bot Elliott after he clarified the Johnson #3 comment during in his online chat with: “Right now they have a lot of No. 3s or maybe Morrow is a No. 2 and Johnson could be if his shoulder allows him to add missing MPH”. That is without a doubt the most informed reason for saying that, although I still think it’s quite the bombshell to drop without any context or attempt to back it up.
Update: Ok, I take that back. Random comments from mystery men at the winter meetings is just not compelling journalism. YOU’RE IN THE HALL OF FAME, BOB. Have an opinion! Say something! Don’t just pass on this scuttlebutt like it’s this incredible inside information but you just…can’t…give up your source. Putting together that Lawrie runs and Gibby is a red light did not exactly require Deepthroat in the parkinglot, y’know?
May as well start this monster of a post off with a bang and slag a Hall-of-Fame baseball writer: has anyone else noticed that Bob Elliott constantly spices up his articles with highly-controversial statements from ‘veteran evaluators’ who sound a lot more like they’re coming from an old Blue Jays fan, or maybe, you know, himself, than they do anything that a real scout would say? From today’s Sun:
“I like the trade, I like Buehrle and Johnson, but they’re a couple of No. 3s for me,” said one veteran evaluator. If you’re going for it — and it appears by everything they’ve done to date, that they are — they need someone else.”
“They really don’t have a dominant No. 1, a Dave Stieb, Jack Morris, Juan Guzman, Pat Hentgen or Roy Halladay.”
While I’m also of the mind that the Jays’ rotation has not been solidified this offseason as much as most people think, calling Josh Johnson a third starter (whatever that means, anyway) demands an explanation. The only way that statement makes any sense at all is if you think he now has diminished stuff due to his arm injuries, know nothing about him except his ERA last year, or are making argument that a #1 who is injured half the time is the same thing as a #3.
Because after labouring through the first two months of last season (really his first six starts — remember that number!) Johnson essentially returned to his career numbers, which made him the clear #1 the Blue Jays were counting on when they made the trade (although he might not be quite at the dizzying heights of *snort*Jack Morris during his time with the Jays, or *chortle* Juan Guzman). With such a history of arm problems, Johnson could equally well boom or best — but there’s no reason to think he has settled into Buehrle-style solid mediocrity.
Anyway, we just have to trust that the Jays are still haunted by the name Mike Sirotka and have done their due diligence up the wazoo about his physical state…but let’s look at what pitch f/x has to say about Johnson’s 2013 and his prospects going forward. To begin, here’s how the man himself saw his 2012 season:
“I’d say maybe the last 15 starts, I felt so much better than before,” Johnson said. “I was kind of fighting myself, fighting my body, trying to do this or that, maybe trying to find a little bit more velocity. But once I relaxed and just trusted myself, it just kind of came out.
“I don’t know if I was throwing any harder or anything like that. But you could tell with the depth that I had on my slider and my curve ball, and the location of my fastball got much, much better.”
It’s the “trying to find a little bit more velocity” that pops out for me here, because if we isolate his four-seam fastball, that is indeed the case. Trust me for one second here — basically every healthy mature pitcher sees his velocity subtly but steadily increase throughout the season. But Johnson’s average fastball velocity flew up over the first couple of months, and then declined for the rest of the season as he became more effective.
[Edit: Ok, that's a little Mickey Mouse so here's pitch-by-pitch instead of game-by-game, with a rolling average of 5 in black that drops about 3mph through the heart of the season]
Things start to look really scary when you compare his velocity to his previous years. Even at his peak in 2012, Johnson was only back up to where he was in 2011 before he got injured, and then went in the wrong direction (the start of 2009 is missing in the following graph because pitch f/x wasn’t entirely set up yet).
If that doesn’t scare you, it should. As soon as a power pitcher known for his 95 mph heater and a killer slider starts talking about ‘hitting his spots’ and not being concerned about his velocity…run. Anyone remember B.J. Ryan? Seem Tim Lincecum lately? You can talk heart and location and pitching all day but strikeouts matter. And unless you’re a control pitcher to begin with, losing 2-3 mph on your fastball without anything to compensate for it is obviously going to affect your K rate in some way.
But wait…all is not lost, because for some reason last season Johnson’s K numbers DID steadily progress to the point where they were right in line with his old self.
So what gives? Did he find exceptional control at the expense of velocity? I don’t think so. Here’s how often his fastball was thrown in the zone as the season went on:
Obviously being in the zone is not the final word on fastball control, so let’s try another stat that I made up — the % of pitches that were “paint”, i.e. within a few (in this case, 2) inches of the border of the strike zone. I know, I know…I’m lumping in 3-0 fastballs with pitches that actually matter, but it’s not like Johnson became some control artist on the outer half when he started backing off on his velocity — if anything, other that a few crazy games late in the season, he was throwing less paint than when he was trying to throw his fastball harder.
[Edit: I take that back. If you take a rolling average of paint pitch-by-pitch, it does show a minor improvement in his ability to paint over the season, with low stretches in April and July when he was getting shelled.]
Enough charts, time for a table! In pitch f/x terms, “Breaklength” is the maximum distance a pitch deviates from a straight line from hand to glove. I think of it as the size of the ‘bend’ in a fastball, or the ‘loop’ in a Curve. ‘Pfz’ is how much a pitch resists the pull of gravity, and correlates to the ‘explosiveness’, ‘hop’, ‘giddy-up’, or whatever you want to call how much a 4-seamer seems to ‘rise’.
|Josh Johnson’s 4-Seam Fastball|
|Year||MPH||Breaklength||Pfz||In Zone %||Miss %||On Fringe%|
Again, scary stuff. Johnson has lost an inch and a half of fastball ‘hop’ and batters are swinging through his fastball half as often as they did a couple of years ago. You can make the argument that he is moving towards a more accurate, moving fastball, but that does sound a lot more like a decent number three than the top-of-the-rotation stalwart the Jays need.
Ok, onto the next part of his self-evaluation — did his breaking stuff show improved ‘depth’ and pick up the K-slack?
First let’s look at his signature pitch, the slider. Unfortunately, results against a single off-speed pitch can be highly deceiving because so much is based on context. If you have lost fastball control, it might look like your slider has fallen apart, because you never get to throw it in 2-strike counts where a hitter has to defend the plate. Or, if your fastball velocity is down, guys don’t have to cheat to catch up to it so they don’t bite early on your slider.
But anyway, the graphs below are my case that although he did manage to locate his slider where it belongs out of the zone more often, it wasn’t any more effective in terms of swings-and-misses. Opposing batters got far fewer hits off it, but I think that equates to fewer hung sliders, and not ‘increased depth’ that would help with punchouts. Swing and miss rates correlate directly to K% (without the fluctuation, basically), and for Johnson’s slider they actually declined in the second half last year.
Compared to previous seasons, here’s how his slider matches up:
|Josh Johnson’s Slider|
|Year||MPH||Breaklength||PFX||PFZ||Miss %||In Zone|
The PFX/PFZ numbers are interesting — he has traded about a half-inch of sideways movement for a half-inch of drop. As I manage to work into every single post these days, that seems to be a bad thing. Sliders seem to be more effective when they cut straight sideways instead of sinking and bending, more like a cutter than a curveball.
And that seems to have an effect on how he is using it — as a chase pitch rather than a hard-to-hit potential strike. While the overall miss % has risen, that’s only because the number of them thrown in the strike zone has plummeted (@#$@#$ context!). Taking only sliders that were thrown for strikes, Johnson’s Miss % has fallen sharply: 27%/26%/14%/21%. While he is using it in a more conventionally recommended way, it is not as fast, not as hard, not as nasty. And he seems to know it.
I hope you’re ready for some good news, because I sure as hell was at this point. Since Johnson only really started throwing it last year, there’s not a lot to compare Johnson’s curve to, but here we go:
|Josh Johnson’s Curve|
|Year||MPH||Breaklength||PFX||PFZ||Miss %||In Zone|
A 42% miss % on a curve is pretty darn good for a third pitch. Even in 2011, just having it lead Chipper Jones to remark: “Oh, OK, he’s got a curveball now…I hope he doesn’t throw that too many more times.”
But now the kicker — here’s Johnson’s K’s per game in 2012 broken down by pitch type:
See the flatline over the first six starts on his curve, after which point it was just as important a strikeout pitch as his slider? Yeah, having an 6.61 ERA and a .359 BAVG against might lead you to trying something new…now take a look at the % of curveballs Johnson threw as the season progressed:
AHA! So THAT’S how he managed to keep his K’s up when all of his other pitches seem to be less effective. I know these graphs are ugly and unscientific, but there’s more to the story than him throwing his curveball 14% of the time or whatever it was over the entire year. What started as a completely non-existent offering was making up more than 20% of his pitches by the end of the year. That’s ridiculously high — Buehrle has a nice floater, and he threw it 11% of the time last year. Gio Gonzalez uses it as his only breaking pitch and he’s only at 21%.
Josh Johnson used to be able to get away with throwing two plus-plus pitches and a show-me change. After a series of arm problems, he’s just not that dominating power pitcher any more. Half a dozen starts into last season, that became obvious to him and/or his coaches, and he started using a pitch in strikeout situations that he had only tinkered with in the past. It worked amazingly well for him, and he started to get more and more confidence with it as the season went on until he was just as effective as he used to be when his repertoire was much more nasty, but predictable. It saved his year, and possibly his career — and what we get to see in Toronto is the first complete season of him with a full repertoire. Excited again?
POSTSCRIPT: Ok, I’ve been really sick the last few days so I didn’t even read the rest of the quotes from Johnson where he talked to the media about his curve, hence why I’m so excited to have ‘stumbled’ upon the fact that it was a revelation last year. It sounds like John Buck might have been the one to recognize it as a plus offering and prod him into throwing it more and more as the season went on:
This was my first full season to pitch with it and to throw it that much. I was learning the whole year. Good thing I had John Buck back there because he helped me out tremendously. Whenever I was in doubt he would put it down, kind of gave me that re-assurance that this is the right pitch, let’s throw it. So I could (learn) how to throw it and when, where to throw it, things like that.