A Quick Glance at Pitch-Framing and Command Extremes

Thursday afternoon, I took a quick glance at 2014 team-by-team pitch-framing projections. This afternoon, I’m taking a quick glance at something else within the pitch-framing field. Longer glances are, of course, superior to quicker glances, but I take quick glances for three primary reasons. One, I don’t have the time, really, to dedicate to longer, research-paper-level glances. Two, I don’t have the mathematical chops to really get into stuff in depth. And three, quick glances make for good starting points, and they usually end up being fairly accurate. If you can get to X in an hour, and if you can only get to 1.1X in ten hours, how valuable are the extra nine hours? Extremely valuable, in science. Less valuable, in casual baseball analysis.

For some years, we’ve had pitch-framing information, for catchers. We’ve been able to tell how many strikes they gain or cost, and we’ve been able to assign run values. A major complication, however, lies in trying to separate catchers from pitchers. It’s the pitchers, after all, who’re throwing the pitches getting caught, and it stands to reason different pitchers might be differently challenging to receive. This is far from a new idea, but it’s an idea worthy of further exploration.

From prior investigations, it seems like the key to good receiving is having a quiet body. If the catcher doesn’t move very much, then all else being equal, he’s got a good shot of getting a strike call. If the catcher does move a lot, a strike is less likely, even if the pitch being caught ends up within the zone borders. People have focused on catcher wrists, catcher heads, and catcher anticipation, among other factors.

But it follows that a catcher won’t have to move very much if his pitcher hits his spot. If his pitcher is more wild, the location will be more unpredictable, and the catcher will have to physically react. Two pitches that end up in the same place aren’t created equal, because it matters where the catcher was positioned, previously. Just thinking about it, it’s easy to see how a chunk of framing data could actually be related to pitcher and pitching-staff command.

What can we do about that? Unfortunately the public doesn’t have access to quality command data. My understanding is it is out there, in private circles, but we have to settle for proxies. Thankfully, the proxies should seldom mislead. It would be almost impossible for us to rank active pitchers by their command, but we can have a pretty good sense of who’s strong in the area, and who’s relatively weak. We can at least probably identify the extremes.

So the task I gave myself was identifying groups at either command extreme, and then looking at their pitch-framing numbers. My proxy for command was simple walk rate, because it seems like it should do just fine. I decided on groups of 25 — 25 pitchers with good command, and 25 pitchers with bad command. I was interested in 2013 pitch-framing data, so, operating under the assumption that command is pretty stable, I put my groups together using 2012 walk rates since 2013 walk rates would be influenced by framing results. I selected from pitchers who threw at least 50 innings in both 2012 and 2013. I also made the particular inclusion of Mariano Rivera, since he was hurt in 2012, but since he obviously has phenomenal command.

The good-command group includes guys like Rivera, Cliff Lee, Sergio Romo, Kris Medlen, and Bartolo Colon. The bad-command group includes guys like Garrett Richards, Matt Moore, Brandon League, Samuel Deduno, and Tim Lincecum. I acknowledge the groups aren’t perfect, and players change from season to season, but this is what I wound up with.

Framing data comes from StatCorner. zTkB% refers to the percent of called pitches within the PITCHf/x strike zone called balls. oTkS% refers to the percent of called pitches outside the PITCHf/x strike zone called strikes. So how did the groups differ, on average, in this past season? No corrections have been made or attempted for catcher identities. Another assumption of mine is that that effect should wash out.

zTkB%

  • Good-command group: 13.3%
  • Bad-command group: 14.9%

oTkS%

  • Good-command group: 8.4%
  • Bad-command group: 5.9%

There’s a difference, with the good-command group being better than average in both categories, and with the bad-command group being worse than average in both categories. The differences appear relatively small, but then most strikes are obvious strikes, and most balls are obvious balls, so the differences are magnified if you look only at pitches closer to the borders. Based on these numbers and these numbers only, pitchers with better command do end up with more strikes, beyond what you’d expect just throwing more strikes in general. And keep in mind that bad-command pitchers might be behind in the count more often, and in those cases the strike zone has been proven to expand.

On average, by StatCorner, about half of all pitches are thrown in the zone. About 36% of pitches in the zone are taken, and about 73% of pitches out of the zone are taken. Based on those averages, per 100 called pitches, the difference between these two groups was 2.2 strikes. Per 100 pitches overall, the difference was about 1.2 strikes. So you can think of it as being about a strike a start. And a strike, they say, is worth something in the area of 0.13 runs. With framing, it’s always about little advantages and disadvantages piling up over the course of one or six months.

The data is preliminary, and it compares only two groups at either command extreme. Most pitchers would fall in between either group, so most pitchers would experience a smaller effect. But it does seem like an effect probably exists, as you’d think it would. Framing isn’t all about the catcher. It’s also about a pitcher’s ability to throw a pitch as the catcher anticipates, which lets the catcher maximize his own receiving ability. A good framer will always be a good framer, but a good framer will look better catching Mariano Rivera than he will catching Brandon League, through basically no fault of his own. The less predictable a pitcher is, the less a catcher will be able to do.




Print This Post



Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.


10 Responses to “A Quick Glance at Pitch-Framing and Command Extremes”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. AK7007 says:

    Did you do anything to see if the differences were actually significant? The oTkS% difference of around 2.5% sounds pretty huge, but what is league average, and what std dev does it have? The ZTkB% seems smaller in comparison, but I don’t really know.

    Vote -1 Vote +1

  2. lefty says:

    How much of the zTkB difference do you think is owed entirely to catchers “crossing over,” or, setting up on the corner and then reaching into the other corner to catch a pitch that’s “wild in the zone?” My hunch is, almost all of it.

    Vote -1 Vote +1

  3. DrBGiantsfan says:

    Thank you. Very nice work and write up!

    Vote -1 Vote +1

  4. It’s not all about “command” either. It’s about the type of pitch. For example R.A. Dickey, by the BB rate he would be deemed good command, but his knuckleball is very very difficult to frame and he loses lots of strikes because of it. Splitters such as Uehara’s may be harder to frame as well due to the difference in vertical location where it crosses the zone and where it’s caught. It breaks away from the plate. Other pitches intended to break away from the zone will be subject as well. On the other side there’s pitches that break towards the zone (around the plate) and these are easier to frame. I have also found slight correlation to velocity and framing, a softer fastball is easier to catch and easier to frame then a 100mph fastball.

    Vote -1 Vote +1

    • LHPSU says:

      But won’t you expect that Dickey also gets some strikes called his way from knuckleballs that dance away from the strike zone at the last second?

      Vote -1 Vote +1

  5. tz says:

    Cool stuff Jeff.

    I’d love to see a similar analysis grouped by “good framers” vs. “bad framers”, to see if perhaps a Lucroy-level framer helps even pitchers with poor command, or (more likely) if a Doumit-level framer just hurts all pitchers across the board.

    Vote -1 Vote +1

  6. channelclemente says:

    Maybe I missed it, but is there a way to weight the value of a pitch as to count or consequence?

    Vote -1 Vote +1

  7. chance1 says:

    Help please.
    Does anybody know how find outcomes for batters facing lefties with men on base or RISP? Splits only allows you to sort by lefty/righty without being able to sort with base runners. Sorry that this wasn’t related to the topic.

    Vote -1 Vote +1

  8. Frank rizzo says:

    I understand the idea of just using data…but you can’t act like umpires don’t give the benefit of doubt to pitchers proven to have better command. That has to be taken into account.

    Vote -1 Vote +1

  9. Dave says:

    Wonder how much of Cole Hamels early 2013 performance can be explained by a lack of Ruiz?

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Current ye@r *