Mauer on wrong side of pitchf/x framing leaderboard

[Mike Sixel]

Get rid of the umps, and use technology. Who cares if they "hit their spot"....I care if they threw a bleeping strike or not. (tom Glavine, I'm looking at you)

[RealityBaseTwinsFan]

I don't follow. It sounds like Mauer was the constant, not the pitchers, necessarily.

 

Part of the numbers was a comparison of how pitches were called with different catcher/pitcher combos. However, since Mauer accounted for over half the sample of the starting pitchers he caught during the period you cannot get a good comparison, especially when Butera accounted for another large chunk of those pitchers sample.

 

 

This has become somewhat mythical I think. Mauer sets up in the middle of the plate because Deduno has poor command, yes, but not because his movement is unpredictable. He's not throwing knuckleballs.

 

This is ironically one of the things pitchf/x makes very clear. In addition to measuring where the ball crosses the plate, it also tells us the verticl and horizontal movement of pitches, even the axis of rotation and RPMs.

 

Whether you call it inconsistent movement/poor command the effect is still the same, Mauer ends up having a lot of glove movement in order to catch him, which means pitches are less likely to be called strikes because the pitcher isn't hitting his spots.

 

Samuel Deduno and Throwing Strikes | FanGraphs Baseball

 

The reason Doumit mentioned fastball command as a key to success for Deduno is because his fastball gives opposing hitters fits. It supposedly cuts and sinks with no predictability. It’s why PITCHf/x categorizes his fastball as both a fastball and a cutter, and it’s also why Doumit says catching Deduno’s fastball is “like catching a 92-mile-an-hour knuckleball.”

 

Also for a fastball that is a large amount of horizontal movement spread (10 inches between highest and lowest value), not to mention it has both left and right cut on it. For example looking at the most recent starts from Gibson, Pelfrey, Diamond, Verlander, and King Felix, all of them had cut in only one direction on their fastball and 4 of them had less than 5 inches of difference between the highest and lowest horizontal numbers.

 

The other thing is, if a guy is wild and constantly throwing out of the zone, and falling behind in counts, then if anything that could make a catcher appear to be a better framer. Because in 2-0 and 3-0 that's when umpires are more likely to call borderline strikes than any other counts. Likewise, a pitcher having good command wouldn't necessarily be given more borderline strikes, particularly if he's getting ahead 01 and 02 all the time.

 

See Glavine, see also Maddux, see also Radke. Reputation also plays into it and while counts makes some difference on borderline strikes I doubt Deduno getting the benefit of the doubt in poor counts is anything more than a red herring given the issue here is clearly fastball command/inconsistent movement.

[Willihammer]

Alright, for sake of argument, let's say Deduno's fastball does move more than most. Still its nowhere near a knuckleball.

 

http://www.brooksbaseball.net/pfxVB/cache/movement.php-pitchSel=285079&game=gid_2013_08_05_tormlb_seamlb_1&batterX=&innings=yyyyyyyyy&sp_type=1&s_type=2.gif

 

http://cdn.fangraphs.com/blogs/wp-content/uploads/2013/06/RA_stroF.gif

 

Doumit was a lousy framer in Pittsburgh, and he's a lousy framer in Minnesota. He's almost certainly worse than Thole, the guy who has to catch an actual knuckleballer.

[Willihammer]

By the way, point of clarification. The strike zone overlay baseballprospectus uses does take into account batter height:

 

I herein define the zone as a rectangle at the front of home plate with the borders set at the average location where umpires call at least 50 percent strikes to batters of that handedness and—for the top and bottom boundaries—of similar height. The exact zone boundaries used are as follows, where “px” is the horizontal location of the pitch crossing the front of home plate, in feet, and “pz” is the height of the pitch at the front of home plate:

 

RHB zone: -1.03 < px < 1.00 and (0.92 + batter_height*0.136) < pz < (2.60 + batter_height*0.136)

 

 

LHB zone: -1.20 < px < 0.81 and (0.35 + batter_height*0.229) < pz < (2.00 + batter_height*0.229)

 

 

http://www.baseballprospectus.com/article.php?articleid=14572

[Kobs]

But there is nothing but faith to prove that pitchf/x is more accurate than the umpires.

 

I'd take a strike sniffing dog over the umpires. Two barks is a strike.

[jorgenswest]

It appears the majority of those on this board approve of the Twins apparent disregard for this data.

 

I am on the other side of the debate than the Twins, but not sure if there is anything else to say. I think they are one of only a few teams that would play Ryan Doumit at catcher. There is volumes to read and replies in a forum setting really can't change or inform anyone's thoughts.

 

I hope the Twins are on the winning side of this debate. I fear that they are not and some organizations are laughing at their roster decisions related to catcher.

[drjim]

It appears the majority of those on this board approve of the Twins apparent disregard for this data.

 

False.

[RealityBaseTwinsFan]

Alright, for sake of argument, let's say Deduno's fastball does move more than most. Still its nowhere near a knuckleball.

 

Which proves.......oh wait its a non sequitur. Also umps expect a KB to be all over the place, not so much with a fastball.

[Kobs]

Taking a look at this, I have a hard time taking these numbers as something remotely player specific. Mauer has often done very well in numbers that attempt to quantify framing in past years. I get the feeling this is more based on the quality and type of pitching more than anything else. Some guys may be good at this, but I can't take the peaks and valleys of a season's numbers as indicative of anything.

 

I also can't help but check out his comparison of the "best and worst" of the week and see a vastly different set of pitch types. There's a massive difference between a pitch that was supposed to be low dipping below the strike zone and an inside pitch that ends up outside.

[Willihammer]

Which proves.......oh wait its a non sequitur. Also umps expect a KB to be all over the place, not so much with a fastball.

 

My point is only that Deduno's fastball is nowhere near erratic as a knuckleball, as Doumit claimed. And Deduno is by far the most erratic pitcher on the staff. In general, the Twins field strikethrowers, so its a stretch to put very much of the blame for Doumit's or Mauer's lousy framing scores on the guys they are catching, I think. This is a reputation the Twins staff as a whole has had for going on 2 decades now.

[Willihammer]

Taking a look at this, I have a hard time taking these numbers as something remotely player specific. Mauer has often done very well in numbers that attempt to quantify framing in past years. I get the feeling this is more based on the quality and type of pitching more than anything else. Some guys may be good at this, but I can't take the peaks and valleys of a season's numbers as indicative of anything.

 

I also can't help but check out his comparison of the "best and worst" of the week and see a vastly different set of pitch types. There's a massive difference between a pitch that was supposed to be low dipping below the strike zone and an inside pitch that ends up outside.

 

Even in a third year of catastrophic starting pitching, the Twins are 9th in walk rate and 3rd in strike percentage, according to pitchf/x. So I find it unlikely that very much, if any of the lousiness in the framing scores are explained by the staff.

[jay]

Even in a third year of catastrophic starting pitching, the Twins are 9th in walk rate and 3rd in strike percentage, according to pitchf/x. So I find it unlikely that very much, if any of the lousiness in the framing scores are explained by the staff.

 

Even without trying to put a specific value on the impact of framing, it seems pretty clear that it exists. If it exists, we (and the Twins) should probably use the data to assess performance in some manner.

 

However, we've seen arguments to doubt the technology, doubt the strike zone, doubt the staff, etc and all have been rebutted. What other reasons can we think of to discredit this data?

[Guest USAFChief]

Even in a third year of catastrophic starting pitching, the Twins are 9th in walk rate and 3rd in strike percentage, according to pitchf/x. So I find it unlikely that very much, if any of the lousiness in the framing scores are explained by the staff.

if a catestrophic staff is third in strike percentage, why the insistance that the catchers are loosing so many strikes?

[Willihammer]

if a catestrophic staff is third in strike percentage, why the insistance that the catchers are loosing so many strikes?

 

According to this, Pitchers who allow under 4 runs/9 get about 1 extra strike per 300 pitches thrown, as compared to guys who allow between 4 and 6 runs per game. So, being a lousy pitcher doesn't appear to be a huge factor.

 

I agree that pitchers can and do have an influence, but I just don't see the Twins as being especially disadvantaged where it seems to matter, certainly not to the point where it absolves Mauer and Doumit of losing all these strike calls. Aside from their bloated ERAs, they're not particularly young, lefthanded, or wild.

[Kobs]

Even in a third year of catastrophic starting pitching, the Twins are 9th in walk rate and 3rd in strike percentage, according to pitchf/x. So I find it unlikely that very much, if any of the lousiness in the framing scores are explained by the staff.

 

Those numbers are irrelevant. The Twins philosophy is to throw strikes, strikes, and more strikes (The Twins being 9th in walk rate is actually pretty damning, since that is the one thing they are trying to outdo the rest of the league in). The fact that the Twins' pitchers are trying to hit the strike zone, and do, most of the time, doesn't mean that they have good control. Most teams don't want to throw as many strikes as the Twins.

 

Logically, this is dependent on how often the ball is delivered away from where the catcher set up. That's even what all of the examples of "bad framing" are in these blog posts.

[drjim]

Attempting to Explain the Unexplainable | FanGraphs Baseball

 

Not sure if people saw this, but I thought it was an interesting way to look at pitch framing, though the article is more about why umpires miss calls.

[Guest USAFChief]

Attempting to Explain the Unexplainable | FanGraphs Baseball

 

Not sure if people saw this, but I thought it was an interesting way to look at pitch framing, though the article is more about why umpires miss calls.

Great find Jim.

 

couple things...

 

1. If these are the 7 worst pitch calls halfway through the season, we don't have much of a problem.

 

2. IMO this points out possible problems with technology as much as possible movement influence on umpires. The pitch to Keppinger is described by pitch f/x as "1.31 inches from the center of the zone," but it's quite possibly not even a strike. The bottom of the zone is "the hollow beneath the kneecap" as the batter is prepared to swing. The freeze frame of that pitch shows the ball at, possibly even below that point on Keppingers front leg. We can't be sure if the ball had even reached the plate yet in that shot, but its already low enough to be a questionable strike at best, and possibly a ball. Yet it's described by technology as only 1.31 inches from the center of the zone.

 

If pitch f/x is as accurate as its manufacturers and promoters claim, how is that possible?

[Brock Beauchamp]

Great find Jim.

 

couple things...

 

1. If these are the 7 worst pitch calls halfway through the season, we don't have much of a problem.

 

2. IMO this points out possible problems with technology as much as possible movement influence on umpires. The pitch to Keppinger is described by pitch f/x as "1.31 inches from the center of the zone," but it's quite possibly not even a strike. The bottom of the zone is "the hollow beneath the kneecap" as the batter is prepared to swing. The freeze frame of that pitch shows the ball at, possibly even below that point on Keppingers front leg. We can't be sure if the ball had even reached the plate yet in that shot, but its already low enough to be a questionable strike at best, and possibly a ball. Yet it's described by technology as only 1.31 inches from the center of the zone.

 

If pitch f/x is as accurate as its manufacturers and promoters claim, how is that possible?

 

I think the biggest problem with "accurately" measuring the strikezone is its ambiguous nature in the first place. What is "the hollow beneath the knee"? Where does it stop and where does it start? What if the player's pants are bunched up and slightly obscure the underlying physiology that is used for the basis of the zone? What happens if the batter loads his stance upon pitch delivery, moving both of his knees? Does the strikezone stay static pre-load or adjust with the body after load?

 

I'm all for Pitch f/x and generally, I think it's pretty good. But given the fact that it's trying to measure something in motion and generally vague in the first place, I'm incredibly skeptical of it being accurate within tenths of an inch.

[Willihammer]

Chief your qualm is with what fangraphs and/or baseballprospects/brooksbaseball have defined as the zone.

 

Pitchf/x's x/y/z coordinates should be within an inch of the actual spot of the ball in timespace. Unless the system in that park has gotten out of calibration. Whether that spot is a strike is up to interpretation.

 

One of the tests MLB did to determine the accuracy was setup a foam plate and a pitching machine, so the foam indented at impact. They compared measurements and were satisfied. The readings are sound. Its the zone that these 3rd party internet people have created that may or may not be questionable.

 

edit: that zone, by the way, is based off the strikes that umpires actually call. That is, the left and right extreme stop at the +/- 50% of called strikes.

[Guest USAFChief]

Chief your qualm is with what fangraphs and/or baseballprospects/brooksbaseball have defined as the zone.

 

Pitchf/x's x/y/z coordinates should be within an inch of the actual spot of the ball in timespace. Unless the system in that park has gotten out of calibration. Whether that spot is a strike is up to interpretation.

 

One of the tests MLB did to determine the accuracy was setup a foam plate and a pitching machine, so the foam indented at impact. They compared measurements and were satisfied. The readings are sound. Its the zone that these 3rd party internet people have created that may or may not be questionable.

Which brings us full circle back to assigning values to catchers pitch framing abilities, no? Aren't the people attempting to define pitch framing using pretty questionable definitions and data?

[Willihammer]

Which brings us full circle back to assigning values to catchers pitch framing abilities, no? Aren't the people attempting to define pitch framing using pretty questionable definitions and data?

 

No, they're using the average zone that umpires actually call.

 

And after what, 100 games, and maybe a turn and a half through all the umpires in MLB, I think we can basically wash away an umpire bias at this point.

[Willihammer]

Its process of elimination. Umpires? Check. Pitchers? I think, check. Batters? They've faced around 1200 at this point, so I say check. Home/away bias? Check. Check check check. The biggest or at least most obvious variable left is catchers.

[jay]

Even in a third year of catastrophic starting pitching, the Twins are 9th in walk rate and 3rd in strike percentage, according to pitchf/x. So I find it unlikely that very much, if any of the lousiness in the framing scores are explained by the staff.

 

I know I've read specific stats that prove what we all see -- umpires are less likely to give you a strike call on the edges when you're ahead in the count. When you're behind in the count (specifically w/ 3 balls), the umpire is more likely to keep the at bat alive with a strike call.

 

If Twins pitchers are consistently ahead in the count and have a high strike %, this would seem to work against the Twins catchers and attempts to assess their framing skills. Does this make sense to anyone else or do I need more coffee?

 

Nothing can defend Ryan Doumit and his consistently wretched results though. He's pretty much the poster child for all the mechanics you can isolate on how NOT to get strike calls.

 

Which brings us full circle back to assigning values to catchers pitch framing abilities, no? Aren't the people attempting to define pitch framing using pretty questionable definitions and data?

 

Assigning values -- questionable. Definitions -- less questionable. Data -- not very questionable.

[jay]

I'm all for Pitch f/x and generally, I think it's pretty good. But given the fact that it's trying to measure something in motion and generally vague in the first place, I'm incredibly skeptical of it being accurate within tenths of an inch.

 

Pitch F/X itself only captures the three dimensional plane the ball traveled on and the ball's movements. There's nothing vague about that or the measurements. The strike zone is applied to those measurements based on an interpreter's definition, which is about the only part of it that you could possibly categorize as vague.

 

In any fashion, variations on that strike zone by the umpire should cancel out across teams and catchers as the sample size grows. If a certain group of umpires call a lot of 'high' strikes, that effect will be shown equally on all catchers over time. So, it's hard to blame umpires for the differences shown across catchers.

[Brock Beauchamp]

Pitch F/X itself only captures the three dimensional plane the ball traveled on and the ball's movements. There's nothing vague about that or the measurements. The strike zone is applied to those measurements based on an interpreter's definition, which is about the only part of it that you could possibly categorize as vague.

 

In any fashion, variations on that strike zone by the umpire should cancel out across teams and catchers as the sample size grows. If a certain group of umpires call a lot of 'high' strikes, that effect will be shown equally on all catchers over time. So, it's hard to blame umpires for the differences shown across catchers.

 

Measuring a ball's location in 3D space is easy. That's not the problem. As I said in my OP, the problem is the vague nature of the strike zone in the first place.

 

Will some of that be negated over large numbers? Sure, and I believe it is... Which is why I believe in using pitch framing numbers but don't believe they're exactly accurate. Much like defensive metrics, they should be used in tandem with the eyeball test and over large periods of time. Even a single season can be tainted by a single pitching staff. If you have 2-3 starters who throw in a particular style that influence the outcome, a single catcher could have up to 50% of their innings influenced by the pitcher. At that point, the metric is on very shaky ground.

[ThePuck]

It's a slippery slope when we only agree with metrics if they pass our eye tests (especailly when everyone's eye tests are different), while giving them less value, or none at all, if they don't. That goes for all metrics. Defensive metrics are done by people who watch way, way, way more games and games of all teams in order to get them. Groups of people doing it as their job to give people the best information available to decide who is good at what and who isn't.

[Brock Beauchamp]

It's a slippery slope when we only agree with metrics if they pass our eye tests (especailly when everyone's eye tests are different), while giving them less value, or none at all, if they don't. That goes for all metrics. Defensive metrics are done by people who watch way, way, way more games and games of all teams in order to get them. Groups of people doing it as their job to give people the best information available to decide who is good at what and who isn't.

 

I agree in principle but it's also sketchy to trust a metric that has multiple variables without any counter-test to validate its accuracy. Sure, metrics like WAR have their value (and I use them quite often) but when that metric lists the world champs as being a mediocre or bad team when they won ~90 games in the regular season, how much stock can we put in that metric without looking for outside confirmation of accuracy?

 

For example, right now the Pirates have the best record in baseball. According to FG, they are 18th in team WAR. Yikes.

[snepp]

For example, right now the Pirates have the best record in baseball. According to FG, they are 18th in team WAR. Yikes.

 

I think you're looking at the "rest of season" projection page.

 

 

Their position players are currently 7th overall in WAR, while the pitching staff is 17th. I don't know where they rank with the two combined, but my California math says that it's higher than 18th.

[Brock Beauchamp]

I think you're looking at the "rest of season" projection page.

 

 

Their position players are currently 7th overall in WAR, while the pitching staff is 17th. I don't know where they rank with the two combined, but my California math says that it's higher than 18th.

 

Nope, I'm on the totals page. FG doesn't really explain the page but as far as I can tell, that's the total for the season per team (and by position).

 

Depth Charts » Team WAR | FanGraphs Baseball

[snepp]

Yeah, those are the rest of season projections. When you look under the Standings menu it's listed as "Team WAR Totals (RoS)".

 

 

You have to go under the "Teams" menu heading then look at the team batting and team pitching sections to see year-to-date WAR totals.