Baseball Savant’s New Framing Data And Pittsburgh Pirates Catchers


Baseball Savant has released catcher framing data for years in the Trackman era of 2015-present.  How does the metric align with other popular framing metrics and where do Pittsburgh Pirates catchers rank?

In the last week, Baseball Savant has released their new catcher framing model, and the site has also published graphics displaying a catcher’s ability to get an extra strike based on a quadrant outside of the strike zone.  The model culminates in a “runs extra strike” category, similar to what Baseball Prospectus has with their framing runs and same with Fangraphs.

Baseball Prospectus has long been the main face of public pitch framing models.  Dan Turkenkopf, now of the Brewers, was really the first to dive into pitch framingMax Marchi, who now works for the Cleveland Indians front office continued on with pitch framing, and even looked at catcher before the PITCHf/x era and even looked at how to evaluate a catcher’s framing in the minor leagues.  This has since been updated to a called strike above average model (CSAA) that is translated to runs, with work done by Harry Pavlidis and the rest of the stats team at BPro.

Framing has long been understood as a main place where teams have placed an emphasis on, once being an inefficiency in the market place the Pittsburgh Pirates took advantage of (see Russell Martin, Chris Stewart, and Francisco Cervelli).  Framing is really that valuable.  Back in 2017, Jeff Sullivan, who is now with the Tampa Bay Rays, looked at how pitch-framing data was going insane.  As bad framing catchers were phased off the position, making the data more fuzzy, but getting an extra strike really is that valuable.

Fangraphs has now recently entered the market with their pitch framing, updating their WAR model which made Ryan Doumit the worst player in Pirates history.  Fangraphs found a correlation of 0.96 with their model and Baseball Prospectus’.  Looking at data from the 2015-2018 seasons (the first season Baseball Savant has), the Pearson correlations of the three models are:

Catcher Framing Correlation Matrix
Baseball SavantBaseball ProspectusFangraphs
Baseball Savant1.000.880.91
Baseball Prospectus0.881.000.96

The three models correlate highly, which is a good sign for Savant’s model.  Where does that leave the Pirates?  The table below shows how many runs the Pirates saved in each season from 2015-2018 by each model:

Pirates Pitch Framing
YearBaseball SavantBaseball ProspectusFangraphs

I’ve documented the Pirates struggles with defense before, especially in regards to their ability to block pitches in the dirt.  The decline in pitch framing has hit Francisco Cervelli the most, and Jake Smail at The Point of Pittsburgh has a hypothesis as to why.  The three models show Cervelli’s decline below:

Francisco Cervelli Framing
YearBaseball SavantBaseball ProspectusFangraphs

Cervelli’s framing decline came entering his age 31-year-old season (2017), about the age where Dan Brooks, Jonathan Judge, and Harry Pavlidis found, writing that,

"“Under CSAA, framing ability peaks around age 25, declines gradually until about 32, and then starts falling off a cliff.”"

Cervelli’s injuries, alongside the typical aging curve, haven’t helped his decline in saving pitchers runs by his framing.  But it’s not just the glove that hasn’t been there.  Looking at his wOBA, my own model for wOBA (est_wOBA), and Baseball Savant’s xwOBA, Cervelli’s best season was last year, and his 2015 season might have been more fluky:

Francisco Cervelli wOBA

But his 2018 season was a season of two halves, most notably the first half Cervelli hit the ball hard and got great results.  But an injury riddled second half led to the following splits in the same three metrics:

Francisco Cervelli 2018 Season
First HalfSecond Half

Cervelli has stopped hitting the ball hard, a big part into why he struggled in the second half of 2018 and a big reason for his early seasons struggles this season.  Here’s a density plot of Cervelli’s launch speed in the first half of 2018 (beginning of 2018-June), second half of 2018 (July-rest of season), and the start of 2019 to date:

While the samples are only 132, 121, and 56 with respect to the Period in the legend, you can see a clear difference in how hard Cervelli was hitting the ball.  The center of the distribution has shifted left from the first half of 2018 to 2019, and in the second half of 2018 there was a higher percentage of the balls in the left tail of the distribution.  So far in 2019, Cervelli’s launch speed is centered around 85 miles per hour.

Next. Bryan Reynolds Makes MLB History. dark

The injury history, especially a season ago, has not helped with Cervelli in regards to his pitch framing, and his decline was more drastic than the average aging curve in that regard.  After a fast start to start 2018, with a change in how he set up mechanically in the box, has since evaporated and how hard the catcher has hit the ball has declined.  In one of the most important defensive positions with how the Pirates are built, getting the extra strike is important, and with that in decline, the Pirates needed to rely on Cervelli’s bat to play.  So far, the quality of contact hasn’t been there.  This is something to keep an eye out for in the remaining five months of the season.