O’s Defense Q&A with Ben Jedlovec, Baseball Info Solutions

To discuss the Baltimore Orioles defense, we’ve reached out to Ben Jedlovec, President of Baseball Info Solutions. Baltimore Sports and Life (BSL) thanks him for his thoughts.

(You can discuss this on the BSL Board here.)

BSL: JJ Hardy recently won his 3rd Gold Glove. Using the Baseball Info Solutions /  Fielding Bible’s Defensive Runs Saved (DRS) metric; Hardy posted the third best defensive season of his career. (By FanGraphs UZR/150, it was his best.) Any issue with him again taking home the hardware as the American League’s best? (I note that Atlanta’s Andrelton Simmons was named the Fielding Bible SS of the Year.) The 32 year-old Hardy recently signed a 3 year extension. Over those next 3 years, would you expect Hardy to remain a plus fielding SS?

Jedlovec: Hardy finished third out of all MLB shortstops in voting for the Fielding Bible Award, but he was the only AL shortstop in the top seven. He was also the most productive AL shortstop by the numbers with 10 Defensive Runs Saved. All told, he deserved the AL Gold Glove in 2014. Meanwhile, Hardy has a great defensive track record.  He has saved at least seven runs in 8 of his 10 seasons, including the most recent four.  He seems like a good bet to continue that solid defensive performance over the next three seasons.

BSL:  Baseball Prospectus tracks Defensive Efficiency (the rate at which balls put into play are converted into outs by a team’s defense). Based on the DE spectrum of the 2011 season – a .693 mark was horrendous, .710 was average, and .735 was excellent. In 2014, eleven teams were at .710 and above. That included the O’s at .720 (good for 4th overall). Any issues you see with the statistic? Do you like that as a measure of team defensive performance?

Jedlovec: Defensive Efficiency, as you say, is the rate at which balls in play are converted into outs. While this somewhat correlates with overall defensive play, it is far from a fool-proof approach. DE considers all balls in play the same, whether it was a screaming line drive in the gap for extra bases or a soft grounder that trickles between slow infielders. As a result, it can overrate defenses behind pitching staffs which induce weak contact, and it can overrate teams with poor defensive outfields allowing extra-base hits by the dozen.

We account for these variables and many more with Plus/Minus Runs Saved, estimating the number of runs a team saved with their ability to turn balls in play into outs.  Additionally, there are other important aspects of defense that we measure, such as defense on bunts and GIDP opportunities, stolen base prevention, framing, and outfield throwing.  We aggregate all nine of our individual components into Defensive Runs Saved.  By DRS, the Orioles were the best defensive team in the AL (56) this season and third overall.

BSL: Hardy was not the only Orioles Gold Glove winner – as Nick Markakis won his 2nd, and Adam Jones won his 4th. In ’12, and ’13, Markakis posted -7 DRS numbers. From ’09-’13, his UZR/150 was also negative. In ’14, Markakis’ DRS and UZR were both positive. Similarly, Jones had negative DRS numbers during ’10-’13, along with negative UZR/150s from ’09-’13. In ’14 Jones’ DRS and UZR were both positive. What do you attribute this to? Aberrations? Positioning?

Jedlovec: In Markakis’ case, the fluctuations in his total DRS numbers are influenced heavily by his tremendous throwing arm.  He’s been very consistent with his range, falling between -6 and -12 Plus/Minus Runs Saved for six consecutive seasons.  He’s also been consistent in Good Play/Misplay Runs Saved, where he’s saved between 0 and 6 runs every year of his career.  However, with his arm, Markakis has had some pretty big differences from season to season.  For example, in 2007, his best defensive season, Markakis allowed runners to take an extra base on just 44 percent of the opportunities, and he had 14 “kills” (assists without a relay man).  Since, he has allowed an extra base rate of 49 percent or higher three times and thrown out five or fewer baserunners without a relay man four times.  Though he has been above-average with his arm year in and year out, this fluctuation in results has led to a 10-run swing between his best and worst seasons with his arm.

Before 2013, Adam Jones seemed to be showing a pretty clear trend of decline.  He was a better than average defender every year from 2006-2009 and then fell well below average every season from 2010-2012.  He has since turned that around the last two seasons as a neutral defender per DRS.  For him, positioning is a factor.  He has always positioned himself shallowly in Baltimore, and he has therefore allowed a lot of balls to drop in for doubles and triples over his head.  Those balls are more damaging to the team than shallow balls that become singles, so that can quickly lead to a negative DRS total.  After allowing between 19 and 25 extra bases more than an average center fielder in three of his four seasons from 2009-2012, Jones has improved to allowing 15 and 13 extra bases over average, respectively, in the last two seasons.  The fact is, fewer extra-base hits are getting over his head.

BSL: Am I wrong in thinking there should be less fluctuation in defensive performance annually vs. offense?  I get that defense can be impacted by health, the pitching in-front of them, improvement through work. However, it seems that in the most general terms players should reach their individual limits of ability – followed by incremental regression. That there should be less variance in performance, as the players can control their defensive performance better than they can control their offensive output. If I’m seeing this incorrectly, what am I not giving proper consideration?

Jedlovec: On defense, players may have 500 opportunities over the course of a season, similar to the number of plate appearances that a regular hitter would have. However, our data has shown that on the defensive side, at least 300 of those chances are likely to be hits or outs no matter who is playing the position, so we really just have 200 plays that have some level of difficulty to them. Out of those 200, maybe 50 of them will do the most to separate the best and worst defenders. So, we’re talking about a really small sample of plays from which we try to draw broad conclusions.

I’m not convinced that defense is less prone to fluctuation than offense. Defense can slump too. The Giants were one of the league’s best defensive teams this year but almost gave away the World Series on thanks to defensive lapses. Additionally, the closer we have studied defensive metrics, the more we learn about underlying changes in ability. Often times, we have heard about injuries, coaching improvements, or weight gain or loss and the defensive numbers have followed. Every time Pablo Sandoval shows up for Spring Training overweight, his defensive ratings plummet. Mike Trout gained 40 pounds after his rookie year, and his defensive ratings took a hit. The examples are endless.

All of that being said, defensive metrics correlate better year-to-year than offensive and pitching stats like batting average and ERA over much larger samples of at-bats or innings.

BSL: If people have questions about DRS, more can be learned here. How has the increased usage of defensive shifting, impacted the measurement of DRS among individual players?

Jedlovec: We view defensive shifting as a team-initiated strategy rather than an individual decision.  Because of that, we credit teams for their making of plays when in a defensive shift in a statistic we call Shift Runs Saved rather than individuals.  So a player’s Defensive Runs Saved will be based solely on their plays made (and not made) in traditional defensive alignments.

BSL: Last year, MLB Advanced Media announced that beginning in 2015 there will be new ways for the complete and reliable measurement of every play on the field. How do you anticipate this impacting BIS, and Baseball Analytics in-general?

Jedlovec: More information is always a good thing. We’re excited for the potential of the StatCast system, though I wouldn’t hold your breath for a public release. We see Statcast as an opportunity to learn more about player and team defense, though we’re not anticipating that it will change everything we know about defense already. There is always more to learn, and thanks to the leadership of John Dewan, Bill James, and others we are confident that BIS will continue to stay ahead of the curve.

BSL: ESPN’s Orioles Sweetspot blog Camden Depot, wrote about Caleb Joseph’s pitch framing ability here. Based on observation, and all available information – how would you compare Joseph’s defensive abilities vs. Matt Wieters?

Jedlovec: Matt Wieters has traditionally been a very solid defensive catcher.  From 2010-2013, his four healthy, full seasons, Wieters averaged 4 Defensive Runs Saved per season.  He has been particularly effective at preventing stolen bases; he caught 34 percent of attempting basestealers in that time, third highest of catchers with at least 2,000 innings played (31 qualifiers). 

However, based on his limited sample in 2014, Joseph looks like the better defender.  He saved 8 runs in half of a full catcher season, which put him on a pace to match Wieter’s best defensive season.  He threw out 38 percent of attempting basestealers.  And, as you mentioned, Joseph looks to be the better pitch framer.  Baseball Info Solutions will be releasing a new component of Defensive Runs Saved called Strike Zone Runs Saved in the The Fielding Bible—Volume IV. We’ve taken the public pitch framing research to date and gone several steps further. Full details will be released in the book in March. By our measure, Joseph saved 5 runs with his framing in 2014, 1 run more than Wieters has saved with framing since 2010.