The Case for Defense – Examining Advances in Defensive Analytics

Making his BF&N debut, welcome the analytical genius, Russell Eassom

Baseball like any other sport you win by scoring more than your opponents and in baseball you have to score more runs to win a game. But unlike most other sports baseball has had a statistical revolution in the form of Sabermetrics. No one can deny that the landscape of baseball has changed since Bill James wrote his first Baseball Abstracts back in 1977, every major league team now spends significant amounts of money on this analysis.

Sabermetrics has traditionally looked at batting and pitching performance, how to score runs and how to stop people scoring runs. A lot of money has been invested not only improving this analysis but also invested in changing the way pitchers and batters behave. We now see more 95+ mph fastballs, home runs and strikeouts than we have ever previously seen.

The gremlin in the statistical analysis of baseball has always been about the defensive side. The main issue has always been that the data points that were historically tracked — errors,  putouts, assists, double plays, etc. — didn’t actually tell you much about whether a player is any good and therefore also the same can hold true at the team level.

These golden oldie measurements might have been fine back in the days when an average team would have more than two errors per game. Now that figure is almost at a half-error per game and it’s such a rare event that almost no one judges a fielder based on errors anymore. That’s progress.

So how do you rate a player defensively then? Unfortunately, in baseball there isn’t currently a consensus on how to rate players defensive capabilities.  It takes significantly more data points than the ones described to try and determine if a player is defensively better than another. The modern solutions of DRS (Defensive Runs Saved), UZR (Ultimate Zone Rating) and DEF (Defensive Runs Above Average), which expands on UZR and compares different positions, use play-by-play data recorded by BIS (Baseball Info Solutions). But they all interpret the data differently to give a slightly different outcome of which players are defensively better.

DRS, UZR and DEF, all have a positive correlation from season to season. In other words, a player’s score from last season tells you a bit about what you can expect next season. Sadly, that correlation isn’t that high and is significantly less than that of, say, batters power and strikeout rate. The real measure and value of any metric is in its predictability. These all have predictive value but they are not really good enough to forecast the pecking order of teams defensively with the same degree of confidence that you might with hitting statistics. These metrics are better than anything we have had before but still not enough.

Can we get to where we want to be? BIS certainly think so, at the 2018 SABR conference they introduced its enhanced DRS system which they are calling PART (Positioning, Airballs, Range, Throwing). This new metric is designed to break off each defensive skill into its own bucket, then combine them back together at the end for a new version of DRS.

This is awesome when do I get to see it? Well, I have some bad news, this new system is being marketed to teams but won’t, at first, be available publicly. Like a lot of new analytical tools the best stuff is behind the curtain so teams can try and find an edge over the others.

But fear not we fans will be getting something out of this. Thanks to this new data, ESPN’s Statcast this season should have data on infield play that is as compelling as the data that Statcast has been generating for outfield play for the past few years now, and there a few more new things coming soon, including pop time data for catchers and better wall ball catch percentages.

So who do we think the best fielders are right now? DRS, UZR and DEF usually point to different individuals but last season 2 out of the 3 pointed to Andrelton Simmons and the other had him a very close second. DEF, which is the best, in my opinion, for comparing across positions had Simmons clear by a long margin. Andrelton had a stand out season last year improving his batting performance to his best ever and getting close to his defensive performance from 2013 which brought him to everyone’s attention back when he was 23 year old Angel.  He finished 8th in the AL MVP voting last year (Personally I had him 4th behind Jose Altuve, Cory Kluber and Aaron Judge) but he is still yet to attend an all-star game.

Other names near the top are not that surprising: Mookie Betts, Byron Buxton and Tucker Barnhart  but what is surprising was that Manny Machado, who is often heralded for his defensive range and capabilities, wasn’t in the top 30 for any of those metrics last season. Maybe this shows the current hole in these metrics and/or he just didn’t have great season or the general consensus might be wrong.

Who are the worst? It came to no surprise to me that for the DEF scale 18 of the bottom 30 fielders (who had fielded 500+ innings) were first basemen, only the Giants Brandon Belt had a positive DEF score. Traditionally it has always been the place you put someone who can hit but cannot field well enough anywhere else when the DH spot has already been taken and that still seems to hold true. The honours for worst goes to the Padres Wil Myers with the Astros Yulieski Gurriel in second, the As Khris Davis (who cannot throw a ball!) was the worst OF and White Sox SS Tim Anderson was the worst non-OF/1B.

What about teams then? The Angels and the Red Sox come out around the top for all these metrics, which is not surprising when they have some of the best fielders in their teams.  Probably the biggest shock is that the second worst team defensively last year were the Houston Astros, their DEF score was heavily impacted by the aforementioned Gurriel but they also did have a single individual with a DEF score of greater than 4. Which is the definition for above average within that system.  This to me shows that defensive capabilities isn’t what is being focused on right now.

What would better defensive metrics do to baseball? A more predictable defensive metric could lead to a few changes to baseball, I think that the we might see more players with an average batting output that give you an above average defensive output. That would mean the days of the defensively poor power hitting 1B and LF could be numbered, I am thinking more of 10-15 years down the line but some teams might adopt it sooner if the new data shows its worth.

Who knows if this will lead another revolution in baseball but one final prediction. Simmons will continue like this for the rest of his career and you will able to make a strong case for him being the best defensive player of all time when he retires.

Leave a Reply