Jump to content

Providing independent coverage of the Minnesota Twins.
Subscribe to Twins Daily Email

The Forums

Article: Escobar Steps Up

Minnesota Twins Talk Today, 07:42 AM
The Twins won on Friday night in Detroit. Later in the night, Angels and the Rangers lost, so the Twins have a 3 1/2 game lead on the 2nd...

Article: MIN 10, DET 4: Twins Blow Out Tigers Bullpen

Minnesota Twins Talk Today, 07:27 AM
One moment it looked like the Twins were going to blow a quality start from Ervin Santana, the next moment they were blowing out the Tige...

Possible items to include in the October 1 game thread op...

Minnesota Twins Talk Today, 07:20 AM
I am asking for TD posters to contact me via private message. I have the honor of opening the game thread for game 162, Sunday, Octo...

With all these firings, Who's replacing them?

Twins Minor League Talk Today, 07:08 AM
We have been hearing about a lot of makeover going on throughout the Twins minor league system from scouts to MiLB coaches and staff to t...

Jake Reed - 2017

Adopt A Prospect 2017 Today, 07:02 AM
Jake Reed:      He didn't win the over/under wins contest last year to win free Harry's Razors....      Pos...

Photo

Pitch Framing and Twins Pitchers

Posted by JohnFoley , 20 March 2017 · 1,754 views

jason castro pitch framing analytics jose berrios front office
On Wednesday, November 30, 2016 the Twins announced the signing of free agent catcher Jason Castro to a 3-year, $24.5MM contract. It was a move that was widely attributed to the members of the Twins’ new front office comfort with advanced analytics. Jason Castro is widely regarded as very good defensive catcher due in large part to his ability to frame pitches and steal strikes for his pitchers. In 2016 Castro ranked third in all of baseball in Baseball Prospectus’ Framing Runs statistic, with +16.3. Kurt Suzuki, the Twins primary catcher in 2016, ranked 92nd at -6.8. Suzuki’s main backup, Juan Centeno, ranked 97th with -9.7.

Castro is a roughly average offensive catcher. He put together a 88 wRC+ in 2016, which ranked 17th among catchers with at least 250 PAs, via Fangraphs. For reference, the league average wRC+ for catchers in 2016 was 87. But, he got a $24.5MM contract primarily because of his framing and the Twins are expecting him to make an impact on their pitching staff.

So where might the Twins pitchers benefit from better framing? Let’s look at the Twins pitchers (that are still with the organization in 2017) that threw at least 50 innings in 2016, sorted by innings pitched:

Table 1   Twins 50  IP


Using this list of pitchers, we can utilize Fangraphs' excellent heatmaps tool to explore each pitcher’s distribution of pitches around the strike zone. For example, here is Kyle Gibson’s 2016 pitch% heatmap, which displays the percentage of pitches thrown to each particular segment in and around the strike zone (shown from the pitcher’s perspective). The rulebook defined strike zone is outlined in black.

Gibson Pitch% Heat Map


There are not many surprises here, as we can see Gibson most often pitches down in the zone, and to his arm side. This is likely driven in large part to the high number of 2-seam sinking fastballs he throws (27.2% of total pitches in 2016, per PITCHf/x data available on Fangraphs).

What this data also lets us do, is explore each pitcher’s propensity for pitching to the edges of the strike zone. Let’s assume much of the benefit of pitch framing occurs at the edges of the strike zone, where pitches are less definitively a ball or a strike to the eyes of the umpire. By focusing on the edges of the zone we can identify which Twins pitchers might benefit most from better framing.

For this analysis, I focused explicitly on the strike zone segments just inside and just outside the rulebook strike zone, which are the areas between the gold lines in the graphic below:

Gibson Total Edge Pitch%


Using the pitch data in these sections, I calculated a metric for each Twins pitcher labeled "Total Edge%". These data points are summarized in the table below and show us the percentage of pitches thrown on the edge, or just off the edge of the strike zone, by each Twins pitcher in 2016:

Table 2   Twins Total Edge%


What we can see is the Twins starting pitchers seemed to pitch toward the edges of the strike zone more than the league average and more than their reliever teammates in 2016, with the exception of Brandon Kintzler. Ervin Santana is approximately at league average, which was 44.7%. Kyle Gibson is significantly above, at almost 49%. Jose Berrios, Phil Hughes, and Hector Santiago are all up around 47%. So, as a starting point, we can assert that Gibson, Berrios, Hughes, and Santiago are the primary candidates to benefit from better framing.

But how do they fare in getting called strikes around the edges of the zone?

Using the same heatmaps tool, we are also able to visualize each pitcher’s called strike percentage (cStrike%), in each segment of the strike zone. Here is Gibson’s for 2016:

Gibson Total Edge cStrike%


As we would expect, pitches located in the middle of the zone are nearly always called a strike, evidenced by the bright red boxes and rates at or near 100%. As before, our interest is just on and just off the edge of the strike zone, which I again outline in gold. Here, we see more variation, with the called strike percentage ranging from as high as 88% in the zone to Gibson’s arm side, to as low as 27% inside the zone up and to his glove side. We also see, pitches just off the strike zone are called strikes at a much lower percentage than pitches just in the zone, as you would expect. But, we need a reference point. How do the Twins compare against the rest of baseball?

Using this data, I calculated two additional metrics, labelled as "In-Zone Edge cStrike%" and "Out-Zone Edge cStrike%", which delineate the called strike percentage on the edge and in the zone, and on the edge and out of the zone. Focusing on these strike zone segments, I calculated the called strike percentage for each Twins pitcher. Also included are the MLB averages for each metric.

Twins In Zone Edge cStrike%


What we see above, is that 6 of the 10 Twins pitchers to throw 50 innings last season had a lower than league average called strike rate on pitches on the edge and inside the legal strike zone. Ryan Pressly and Jose Berrios appear to be the most impacted, with called strike rates significantly less than the league average of 64.9%, at 52.8% and 57.5% respectively.

But what about just off the edge?

Twins Out Zone Edge cStrike%


When we focus on the segments just off the strike zone we see this same trend play out, but even more significantly. The visual above shows that 8 of the 10 Twins hurlers had lower than league average called strike rates on pitches just off the strike zone. This indicates that they were not getting many strikes stolen in their favor. In most cases for the Twins, the difference from league average is quite significant. Berrios, Michael Tonkin, Pressly, Taylor Rogers, and Santiago each have rates right around half the league average of 10.4%. The net result, when we add up the In-Zone and Out-Zone Edge cStrike% for Total Edge cStrike%, is that 7 of the 10 Twins pitchers studied had called strikes rates around the edges of the strike zone that were decidedly less than league average.

Now, this probably isn’t all that surprising intuitively. We know the Twins as a whole did not pitch well last year (29th in ERA, 27th in FIP, per Fangraphs), and we know the Twins catchers did not rate well as pitch framers. Kurt Suzuki and Juan Centeno combined to catch nearly 86% of the Twins defensive innings last season. But for as bad as the team pitched, it is also clear the pitchers were not getting much help from their catchers.

But how many pitches are we talking about here? If we assume a league average called strike rate on the edges of the strike zone (which was 36.1% in 2016) for the Twins, we can estimate an additional number of pitches that would be called strikes. This is what we find:

Table 3   Estimated Called Strike delta


By this analysis, it seems that Jose Berrios, Ryan Pressly, and Ervin Santana would benefit the most from better pitch framing, with each gaining roughly 20 additional called strikes over the course of the season.

But how much does a pitch being called a ball, instead of a strike, actually matter?

Let’s look at the major league batting average by count in a plate appearance. The data in the table below is from a 2014 Grantland article written by Joe Lemire, and calculates the batting average for plate appearances ending on specific counts. For example, the batting average on plate appearances ending on the 0-1 pitch is .321. The data fluctuates slightly year to year, but in any given season, you’ll find a table that generally looks like this:

Table 4   Batting Average By count


By this measure, the value of a strike, depending on the count is quite significant. In a 1-1 count, for example, if the next pitch is called a strike, making the count 1-2, the batter’s expected batting average drops from .319 to .164. Similarly, if the pitch is a ball, making the count 2-1, the batter’s expected average increases to .327. That’s a .163 swing in expected batting average.

Others have approached this differently by trying to calculate the expected outcomes by the result of the at bat that reaches each count. So, for example, what is the expected outcome for all plate appearances that reached an 0-1 count, regardless of whether it was the 0-1 pitch that the outcome of the plate appearance was created? Different approaches aside, we find a similar result according to a revisit of the idea by Matt Hartzell published on RO Baseball in 2016:

Chart 1   Batting Average By count



Chart 2   OBP By count


While the differences here are not quite as steep as before, we still see the swings matter. Batting average after a 1-2 count is .178, where after a 2-1 count it is .247. That’s still a .069 swing in batting average. We also have added on-base percentage, and see the trend holds. OBP after a 2-1 count in 2016 was .383, versus just .229 after a 1-2 count.

So, all of this helps us show the Twins have a pitch framing problem and pitch framing matters because getting more pitches called strikes leads to less runners on base.

But can Jason Castro fix it?

To try to find out, let’s look at the Houston Astros, Castro’s former employer. Using the same methodology as with the Twins pitchers, I again calculated the cStrike% on the edges of the strike zone for the all Astros pitchers that threw more than 50 IP in 2016. What we find is pretty telling:

Astros Total Edge cStrike%


Of the 12 Astros to throw more than 50 IP, only one, Michael Feliz, had a lower than league average called strike rate around the edge of the strike zone. But even he was roughly league average (36.06% compared to league average of 36.11%). The rest of the pitchers studied were above league average, and in most cases, quite comfortably so. Six of them are clustered close together right around 41.0%.

Now, to be fair, not all of this is directly attributable to Castro. These are different, and arguably, better pitchers. And Castro didn’t catch every pitch thrown (he caught 61.9% of the Astros defensive innings in 2016). But the difference is stark and by this rough measure, it seems Jason Castro will make a positive impact for the Twins pitchers.

To the Twins credit, they recognized they had a weakness, and they used the free agent market to acquire a player they hope can help address it.

  • Parker Hageman, jorgenswest, nicksaviking and 2 others like this



Photo
jorgenswest
Mar 20 2017 07:51 PM
Great work!
I know Castro hasn't caught all ST innings, by far, but, it seems like the Twins have pitched better this spring. Excited to see what this season brings!
Photo
tarheeltwinsfan
Mar 21 2017 06:34 AM

Thanks for all the time and effort you took to write this interesting article. I think you are correct. Also, Buxton/Kepler/Rosario in the OF will significantly help the Twins' pitchers by catching more fly balls. Now about the IF...Regarding the SS's defense...the Twins need help there. 3B is a work in progress, but improving. 2B is league average and 1B is above league average. How is Gordon's defense at SS? Will he remain at SS or move to 2B?

Photo
Parker Hageman
Mar 21 2017 07:45 AM

Great work John.

Photo
Willihammer
Mar 21 2017 11:01 AM
Awesome job. Really hopeful about Castro's defense.
Great article, this is very encouraging! It doesn't take a lot to make a big difference when it comes to pitch framing. It will be really interesting to see if Castro can put most of the Twins pitchers Strike % above the MLB average line as he did with Houston.

Thank you!

 

If you haven't seen it yet, Parker also wrote a great piece a few days ago that includes quotes from Castro about his approach and mindset to framing: http://twinsdaily.co...t-framing-r5420

 

I hadn't read it when I posted, but it's obviously very well related.

    • nicksaviking likes this