SoxProspects News
|
|
|
|
Legal
Forum Ground Rules
The views expressed by the members of this Forum do not necessarily reflect the views of SoxProspects, LLC.
© 2003-2024 SoxProspects, LLC
|
|
|
|
|
Forum Home | Search | My Profile | Messages | Members | Help |
Welcome Guest. Please Login or Register.
Vazquez Pitch Framing Log
|
Post by WindyCityRedSox169 on Nov 23, 2014 21:54:27 GMT -5
Here's the thing though, a starting catcher is involved in almost every pitch thrown by his team for the whole year. It only takes a few extra strikes stolen per game to give you an extra win or two over the course of the 120 games or so that you get out of a starting catcher. We've got the data with pitch/fx to show that those extra strikes *are* there - to doubt that there's a difference in called strikes of that magnitude between catchers is to doubt pitch/fx, really. I'm a little sympathetic to the arguent that teams would be valuing it more if it really were this valuable, but I've yet to see a convincing argument about why the current models overestimate the effect. But isn't that Chris's point? Yes a win or two, which is valuable, however talking about Christian Vazquez as a 6 WAR player? That is where the meh comes from. Yes we have all of this data that shows these extra strikes are there however do we really know to what extent those extra strikes do for a team? If Vazquez steals strike one or two and the next player gets a hit, stealing a strike really doesn't do anything for a team. Stealing strike three or stealing an early strike that allows a pitcher to get to strike three is valuable certainly. However I don't know if there is a great way of being able to determine that.
|
|
|
Post by rjp313jr on Nov 23, 2014 22:29:31 GMT -5
It's impossible to truly know which pitch a catcher "stole" a strike on and even if you could it's impossible, to know that strikes effect on the outcome of the game. Even if the pitcher stole strike three to end the game, you don't know said team wouldn't have won without that.
Not everything is quantifiable in terms of wins and losses.
|
|
|
Post by jimed14 on Nov 24, 2014 9:14:14 GMT -5
Maybe it's my perspective as someone who's worked a lot as an umpire, combined with that perspective being at a level well below pro ball and thus perhaps not really in tune with what's going on there, but to me, even just thinking critically about the subject shows how laughable the idea of pitch framing being worth multiple wins is. How many pitches per game can a catcher really "steal"? The pitch needs to be borderline enough to go either way. And then, how many of those stolen pitches really wind up contributing to the result of an at-bat? This is hard to quantify, but stealing the first pitch, which is then followed by a home run on the second pitch, does nothing. There's just no way, in my mind, that pitch framing can be as valuable a skill as the ability to contribute in the batter's box, to execute pitches, or to turn batted balls into outs. Perhaps the difference between elite receivers and abysmal receivers is enough to be significant, but I find it hard to believe that the state of average receiving in the majors is so bad that even an elite receiver can be worth multiple wins for that skill alone, versus the combined skills of receiving, game-calling, controlling the running game, and so forth. To me that's like saying a hitter's ability to lay off of pitches outside the zone is worth multiple wins, or a pitcher's ability to throw first pitch strikes is worth multiple wins. Important part of the picture? Yup. A skill important enough to significantly modify your valuation of the player on its own? Meh. To me, it seems very obvious as I always remember those borderline pitches that show up as strikes on the k-zone replay that would extend innings by 20 pitches because a ball was called. They are so vivid in my memory. They save runs and bullpen usage. But then again, I don't know how much you can give it to the catcher and how much you can give it to the pitchers' reputations and the umpires. I'm sure the catcher who caught Greg Maddux or Roy Halliday had better framing stats because of that.
|
|
|
Post by rjp313jr on Nov 24, 2014 10:01:12 GMT -5
That's the point. You can't tell if it's umpire error or a number of other things. They should just go with an automated strike zone. The technology is there to be completely accurate. Have the home plate up there to deal with everything but balls and strikes.
|
|
|
Post by jimed14 on Nov 24, 2014 10:06:42 GMT -5
That's the point. You can't tell if it's umpire error or a number of other things. They should just go with an automated strike zone. The technology is there to be completely accurate. Have the home plate up there to deal with everything but balls and strikes. I'd hate to lose the advantage now. And we also would lose the advantage of not having many LHB who get screwed on about 4 inches on the outside corner.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,933
|
Post by ericmvan on Dec 27, 2014 7:15:22 GMT -5
I don't want to get on your case specifically, but if you have b-ref bookmarked, you can answer that question in less than 30 seconds. The answer is that Ross caught [Lester] 11 times and Saltalamacchia, a below-average pitch framer, caught him 22. Well he his bWAR was driven by great BWAR in OAK, which had, by the numbers I've seen below average pitch framing, so there seems to be a bit of a disconnect here. Replying to this from the Porcello thread, and putting it where it belongs ... In Oakland, he had really good pitch-framing numbers with bad pitch-framing catchers. Was that luck, or is that because BP and Dan Brooks haven't figured out how to properly calculate the pitcher's contribution? This is one of many reasons why I've started taking half of the measured pitch-framing effect as a conservative estimate / adjustment. One of the things that can be done with the data they provide is calculate each pitcher's framing effect, annually, relative to that expected based on how well his catchers framed everyone else. There should probably be no year-to-year correlation. If there is, then some of the measured pitcher / catcher framing is the pitcher's contribution. However, in that data, the sum of a catcher's numbers broken down by pitcher do not equal their total numbers, so it's clear that they're already doing some such juggling. It's unclear just what they're doing (or it is, at least, until I take a closer look at it.) One thing they definitely do not do, and need to do, is directly measure how frameable a pitcher is. Imagine: Pitcher A throws everything precisely on the de facto black, where pitches have a historical 50% chance of being called a strike. Pitcher B throws half his pitches such that half of those taken are really obvious strikes, and half are pitches not even Sandoval would swing at. An average-framing catcher will have an average pitch-framing score with both catchers. A great pitch-framing catcher will be off the charts with pitcher A and still average with pitcher B. A bad pitch-framer will have terrible framing numbers with pitcher A and be average with pitcher B. Realistically, given two equally good pitch-framing catchers, the one who works with a staff that uses the edges of the zone more will end up with better pitch-framing numbers. Who should get credit for that difference? It's actually not the product of either the pitcher or the catcher alone, but the fact that they work together and are a good team. If you had a pitching staff that leaned towards pitcher B -- a bunch of guys who challenged guys in the zone but were wild -- you'd be wise, once pitch-framing gets valued correctly, to grab a catcher who was bad at it but could really hit (e.g., Jaso), because he wouldn't lose much value handling that staff. This is a lot like matching pitchers to their parks based on GB/FB tendencies: none of these skills exist in a vacuum, they're all contextual. I don't think I'm going to survive another year without trying to figure this all out, so I hope to spend a couple of days crunching all the data sometimes in the second half of January.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,933
|
Post by ericmvan on Dec 27, 2014 7:54:43 GMT -5
It's impossible to truly know which pitch a catcher "stole" a strike on and even if you could it's impossible, to know that strikes effect on the outcome of the game. Even if the pitcher stole strike three to end the game, you don't know said team wouldn't have won without that. Not everything is quantifiable in terms of wins and losses. That everything on a baseball field is in theory quantifiable in runs (and hence wins) is the fundamental concept of all of sabermetrics. You don't need to be psychic and know what would have happened if a guy makes a great catch, or a batter singles instead of striking out, or anything else. We always know, on average, what the run value of anything that happens is. For instance, going from 1st to 3rd on a single with 0 outs was worth, on average, an extra .282 runs last year, versus stopping at 2nd. Getting thrown out costs you .970 runs. If the runner on 1st advances to 2nd on the throw to 3rd, it's only .806. Now, there's a further level where you might want to tweak these values based on a myriad of other factors. That's what makes sabermerics interesting. We have similar run values for any called ball or strike in any count. We know how much better guys hit after getting ahead 1-0 versus behind 0-1, and we can use that to calculate the run value of getting the strike call rather the ball call. The pitch/fx data gives us the precise location of every pitch. So "all" we have to do to accurately measure pitch-framing is the following: -- Create an optimum set of zones or buckets in order to lump the data together. You could divide the strike zone into 9 areas, or 900. Neither would be smart. But somewhere in between is a set of divisions that works. (There are also statistical methods that don't involve creating zones or buckets with fixed borders; I have no idea whether they could be used for this data.) -- Now you can calculate, for each bucket (little strike zone segment), the probability that a pitch taken in that bucket will be called a strike, given the handedness of the batter, the count, and the inning (and maybe the score differential) -- all factors known to influence the size and shape of the strike zone in general. You do this simply by counting. -- If the buckets are very small, you probably want to smooth the data to remove randomness. -- Now, for a given pitch, it's either taken as a strike or a ball, and we know how often that happens on average. If that pitch is a strike 60% of the time, and it's called a strike, then the act of framing has added 40% of the run value of a strike in that count. -- Now, here's the challenging part. How much credit should be given to the umpire, the batter, the pitcher, and the catcher? If it's Greg Maddux in his prime pitching to JBJ (transported back in time), with a generous ump behind the plate, the actual probability of a strike is probably much higher than 60%, meaning the catcher should not get much credit at all for framing that pitch. If it's some rookie pitching to an elite hitter with a reputation for having a good eye, and a stingy ump behind the plate, the catcher deserves a lot more credit, because the actual probability of a strike in that count is much lower than 60%. There are statistical techniques, most notably ANOVA, for separating out the contribution of pitcher, catcher, batter, and umpire. However, with ANOVA, any pitcher who worked with just one catcher in a season presents a problem; all those PA need to be taken out of the initial data analysis (I' imagine there must be techniques for estimating their values). I don't think that BP and Dan Brooks are using ANOVA, and that may be a reason. No matter what technique you use, it's easy to adjust for the batter and ump, because they are constantly changing. Each ump works with huge lists of pitchers and catchers, ditto for batters. But each pitcher is only working with a couple of catchers, and each catcher is only working with 12-30 pitchers. It becomes much harder to properly divide up the credit and blame between them, and, as I explained in the last post, part of the total effect is actually something they share by dint of working together. Another confounding factor: the strike zone shifts each season. Do you redefine your probabilities accordingly? If you do, you immediately lose all that interesting information about how the umps are changing the zone from year to year! The alternative, however, is to redo the overall historical probabilities after each year, which is somewhat of a nightmare. I'll have to think about how to best do this, but off the top of my head, you could use the first year of pitch/fx data as your historical baseline, and then for each subsequent season, you calculate all your framing data twice: once with those baseline probabilities, and once with data and probabilities from that year. That would give you a measure of how the changing strike zone was impacting each catcher's value. I have a lot of confidence that the pitch-framing run values measured by BP (and by StatCorner, which agree quite well) represent real framing. At present I have a lot less confidence that any given catcher's numbers represent his true contribution, after proper adjustment for the pitchers he was working with. There is certainly reason to believe that working with a staff that works the edges of the zone a lot may be exaggerating the catcher-framing effect for some catchers, i.e., that a Vazquez who is measured at 3.4 wins per 120 games might only be worth 2.0 wins with the average pitching staff; the other 1.4 wins of his real, measurable value could be the result of getting an unusually high amount of pitches on the black compared to the average catcher playing as much. That's real value, but it's contextual value, like the extra value Bill Mueller had playing half his games in Fenway that was nowhere in sight when he was playing his home games in Pac Bell / AT&T. And with catchers, it's even more contextual, in that it can more easily change from year to year. Turn over the pitching staff, replacing guys who work the black with guys who pound the zone hard, and suddenly, your elite-framing catcher is losing value! It's as if you moved the fences in and changed the value of your flyball pitchers. Teams appear to be valuing catchers as if all this were the case, i.e., discounting the pitch-framing measured by BP by about half ... and yet it can't be true that every catcher works with an especially frameable staff! Half of them would be working with a staff that was extra frameable, driving their measured numbers further from zero, while half would be working with staffs that were less frameable than average, driving their numbers closer to zero. Whether those catchers and staffs can be identified using the numbers BP is providing is what I hope to try to answer next month.
|
|
|
Post by thelavarnwayguy on Dec 28, 2014 20:18:18 GMT -5
From Eric: "At present I have a lot less confidence that any given catcher's numbers represent his true contribution, after proper adjustment for the pitchers he was working with. There is certainly reason to believe that working with a staff that works the edges of the zone a lot may be exaggerating the catcher-framing effect for some catchers, i.e., that a Vazquez who is measured at 3.4 wins per 120 games might only be worth 2.0 wins with the average pitching staff; the other 1.4 wins of his real, measurable value could be the result of getting an unusually high amount of pitches on the black compared to the average catcher playing as much."
This makes a hell of a lot more sense to me than the ridiculous pitch framing values I hear bandied about for Vasquez. I haven't researched this much but isn't it clear that there are a whole lot of variables to this process beyond the catcher's inherent pitch framing ability? I'm curious how Vasquez does framing low pitches, with all the ground ball pitchers we now have on staff. If low pitches in particular are a strength for him, some of these off season moves make even more sense.
|
|
|
Post by dewey1972 on Dec 29, 2014 22:08:26 GMT -5
A little data confirming that the pitcher plays a dynamic role in the interaction between balls and strikes in this article by Neil Weinberg at Fangraphs today. Quick summary: Matt Garza pitched to Wellington Castillo and other bad pitch-framers in 2013. Then he pitched to Matt Lucroy (one of the best) last year. In 2014, with Lucroy as his catcher, he pitched differently, throwing a smaller percentage of his low pitches (those in the bottom third of the strike zone or lower) in the strike zone and higher percentages farther away from the strike zone. This would make sense, as Garza would expect that Lucroy would be able to get some of those pitches out of the zone called strikes. Of course it doesn't prove causation, but it's a good sample (over 800 "low" pitches each year) and it fits what we'd expect, that pitchers would feel they had more room for error with good pitch framers.
|
|
|
Post by philsbosoxfan on Jan 4, 2015 2:23:16 GMT -5
From the Staff Twitter Feed: Mark Simon @msimonespn Defender of the Day: Red Sox C Christian Vazquez was impressive at pitch framing in a small sample of work pic.twitter.com/19e612WRpU Retweeted by SoxProspects.com
|
|
|
Post by philsbosoxfan on Jan 5, 2015 22:43:04 GMT -5
Ivy Envy ?@ivyenvy 2h2 hours ago This week's episode of the podcast - "Pitch Framing with Harry Pavlidis" @harrypav ow.ly/GQxZG #Cubs
|
|
|
Post by philsbosoxfan on Jan 9, 2015 3:48:23 GMT -5
One factor which I've never seen mentioned that to me, seems like it would have a lot of impact, is a catcher's ability to alter where in the pocket he catches the ball. If you think about the size of the pocket, a glove catching the same pitch could be a few inches differently positioned based on where the ball actually lands in the pocket. It seems logical that no movement or less movement would have a huge effect on an umpire standing behind him.
|
|
|
Post by Chris Hatfield on Feb 5, 2015 10:23:09 GMT -5
New article up on BP that, I believe, is free: www.baseballprospectus.com/article.php?articleid=25514Discusses a new approach to pitch framing from the Brooksbaseball guys. I'll let others analyze what exactly they're doing, but some notes: - CVazquez, still good at framing - Blake Swihart, also good at framing - David Ross and Ryan Hanigan, yup, good at framing. Tek too. - Jarrod Saltalamacchia, not good at framing - Mike Piazza.... good at framing? ??
|
|
|
Post by philsbosoxfan on Feb 5, 2015 11:39:41 GMT -5
New article up on BP that, I believe, is free: www.baseballprospectus.com/article.php?articleid=25514Discusses a new approach to pitch framing from the Brooksbaseball guys. I'll let others analyze what exactly they're doing, but some notes: - CVazquez, still good at framing - Blake Swihart, also good at framing - David Ross and Ryan Hanigan, yup, good at framing. Tek too. - Jarrod Saltalamacchia, not good at framing - Mike Piazza.... good at framing? ?? Good stuff. From the comment section: I'll have to read that about 10X but on the the top framing careers, is there a metric which shows top framing runs "amount of opportunity" (innings caught) for instance, if yadier molina has had more opportunities than Brian McCann or Lucroy on a ratio basis, would they be better?
Using raw CSAA (per frame chance) our top five are J. Molina .028 J. Lucroy .022 C. Vazquez .021 C. Stewart .021 R. Rivera .020
Yadi checks in at .013, McCann .011. Scioscia, btw, checks in at a robust .017 in the pre-golden era of framing _and_ at the tail end of his career. Very impressive.
BTW, for a quick rule of thumb, each .01 of CSAA per chance roughly translates to a single Win Above Average over the course of a full season
|
|
|
Post by Chris Hatfield on Feb 5, 2015 12:27:01 GMT -5
Re: the discussion from a few months ago, I can buy two wins over the course of a season.
|
|
|
Post by greatscottcooper on Feb 5, 2015 12:38:07 GMT -5
What I find fascinating isn't so much the strike calls out of the zone, but how much more than average he's getting strike calls IN the zone.
|
|
|
Post by ramireja on Feb 5, 2015 12:47:49 GMT -5
Beautiful stuff thanks for sharing.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,933
|
Post by ericmvan on Feb 5, 2015 13:08:03 GMT -5
Re: the discussion from a few months ago, I can buy two wins over the course of a season. They've saved me a lot of trouble, analyzing the old data set! I think I'll wait another year before diving back in, as I think it likely they'll tweak what they've currently got. That Swihart was tied for 3rd among all AA and AAA catchers is great news.
|
|
nomar
Veteran
Posts: 10,825
Member is Online
|
Post by nomar on Feb 5, 2015 18:09:26 GMT -5
Re: the discussion from a few months ago, I can buy two wins over the course of a season. They've saved me a lot of trouble, analyzing the old data set! I think I'll wait another year before diving back in, as I think it likely they'll tweak what they've currently got. That Swihart was tied for 3rd among all AA and AAA catchers is great news.Agreed. If he's really that good, that's huge.
|
|
|
Post by philsbosoxfan on Feb 5, 2015 19:46:40 GMT -5
Also agree about Swihart and he appears to be framing at about the same rate as Vazquez.
Also interesting to me, I realize this isn't intended to be a comprehensive list in recent history but, I'm surprized that two catchers, generally considered to be defensive extremes, aren't listed anyplace. Ivan Rodriguez and Jim Sundburg.
|
|
|
Post by mgoetze on Feb 6, 2015 7:18:01 GMT -5
Re: the discussion from a few months ago, I can buy two wins over the course of a season. The main article rates Vazquez 2014 at +31 runs per 7000 framing opportunities. 31 runs is more than 2 wins nowadays.
|
|
|
Post by Chris Hatfield on Feb 6, 2015 8:35:45 GMT -5
Re: the discussion from a few months ago, I can buy two wins over the course of a season. The main article rates Vazquez 2014 at +31 runs per 7000 framing opportunities. 31 runs is more than 2 wins nowadays. I was responding to the information in the comment directly above mine.
|
|
|
Post by mgoetze on Feb 6, 2015 10:08:56 GMT -5
The main article rates Vazquez 2014 at +31 runs per 7000 framing opportunities. 31 runs is more than 2 wins nowadays. I was responding to the information in the comment directly above mine. Yes but ... why? I mean, sure, 0.007 is roughly the same as 0.01, but when there are more exact numbers staring you in the face...
|
|
|
Post by philsbosoxfan on Feb 6, 2015 11:37:43 GMT -5
The numbers at in the comment were generated by the authors. If both sets of numbers are true then that could mean that 7000 framing chances equates to 162 games which is more than a full season for any catcher. Maybe the 2.1 represents the derived average for a feature catcher (which would seem low)?
It is odd though and this doesn't appear to be a typo and using 7000 as just an arbitreary number would also be odd. Maybe somebody with a subscription could ask.
|
|
|
Post by mgoetze on Feb 6, 2015 13:30:10 GMT -5
I'm pretty sure it's just that when they said 0.01 CSAA = 1 win is a quick rule of thumb that roughly translate, they did indeed mean it's a quick rule of thumb that only gives a very approximate answer...
I'm somewhat sure that 7000 framing chances is supposed to be ~120 games, that would be about 58 chances per game, which seems about right.
|
|
|