by 
If you are like me, you have seen multiple defensive metrics and not known what the difference is. I usually take the first one I find and assume that others are close, at least in the same direction.
The limits of defensive stats are clear. But that is not to say they don’t provide value. Defensive value is harder to quantify, but still important. Even though they are not 100% accurate they can help us better understand the entire value of a player. Even if parts of evaluations are subjective, having some aspects that bring objectivity are important.
In this second defensive-minded piece, we will look at Ultimate Zone Rating, or UZR, and how it compares with Defensive Runs Saved. Understanding the differences and similarities, we can see if it changes how we view the Reds defense.
Ultimate Zone Rating
Ultimate Zone Rating was born from similar research and analysis as Defensive Runs Saved. John Dewan, who owned Baseball Info Solutions (now Sports Info Solutions) developed the original “Zone Rating” while working for STATS, Inc. prior to BIS. Zone Rating rated fielders on their ability to make plays within their zone based on their position. As John moved on from STATS, Inc., the metric evolved into the Plus/Minus system, now known as the Range and Positioning System, a major component of DRS.
Michael Lichtman began calculating his own “Ultimate Zone Rating” that used additional calculations and adjustments. Like DRS, this statistic is maintained on FanGraphs and uses the same SIS play-by-play data. It is important to keep this fact in mind as the differences in final outputs are coming from slightly different philosophies and interpretations of how to measure the same events.
One key difference is the way each system assigns credits and debits for fielders’ value.
DRS uses a simple equation based on the likelihood the ball is fielded. If an out is made on a similar batted ball 25% of the time, a successful play earns 0.75 runs saved, where an unsuccessful play loses -0.25. UZR uses the same approach but then multiplies that value by the difference between an average hit and an average out. For example, a typical outfield hit is worth 0.56 runs, while any batted ball out is worth -0.27 runs, making the difference 0.83. By multiplying 0.75 x 0.83, UZR would credit the fielder +0.625 runs for a successful play and -0.208 for an unsuccessful one. This adds an additional layer of specificity based on the type of batted ball and creates a smaller scale for UZR.
What is being measured?
Defense has more moving parts than one-dimensional actions of hitting and pitching. Because of this, there are multiple factors that go into measuring defensive performance. As we saw with DRS, there are ten different components that add up to one final runs saved number. For UZR, there are only four components.
The common components between both statistics include range, outfield arm, infield double play, and errors. While errors are not an explicit category in DRS, the good plays/misplayed runs saved would capture any errors. The difference is that UZR gives credit for not making errors, while DRS gives credit for making unexpected plays as well.
Other components of DRS beside those four are not components of UZR. These include: bunt runs saved, catcher stolen base runs saved, pitcher stolen base runs saved, adjusted earned runs saved and strike zone runs saved. For UZR, bunts are treated as a separate kind of batted ball for first, second and third baseman (shortstops rarely field bunts). They are therefore accounted for as any other ground ball would be. The other pitching and catching components, however, are not.
UZR does not provide values for pitchers or catchers, obviously a huge difference from DRS.
How are the components measured?
Both UZR and DRS factor in special circumstances and many are similar. Examples include when the infield is at “double play depth” or the first baseman is holding on a runner. “Wall balls” are factored in by both metrics, which is when a fly ball hits off an abnormally high wall (think Green Monster) and does not present a chance to catch the ball. “Ball hogging” is another. An example of this is if a centerfielder catches a ball that could have been caught by the corner outfielder (and similar batted balls have been caught by the corner outfielder), there is no penalty for the corner outfielder not catching the ball. In both metrics, the ratios are adjusted to better understand the likelihood each player makes the catch and the amount of credit or debit the players would receive should it fall for a hit.
There are many more adjustments that try to factor out all noise and focus on true defensive ability and value. I encourage anyone interested in learning more to read about both UZR and DRSand improve your knowledge as a fan.
Let’s look at some current Reds and see how the stats compare. To get the most accurate results, I used data from the 2016 – 2019 seasons.
Assessing Joey Votto
As the current Reds player with the most innings played since 2016, Joey Votto is our first case study. His DRS (7) and UZR (7.3) are nearly identical. For reference, below is the UZR scale. DRS does not provide a similar scale, but the numbers will naturally be larger given the difference in formulas.
From this we can see that since 2016, Joey has rated as an above average first baseman. The individual components of DRS and UZR largely agree on where Joey excels as well. Both metrics rate him average at turning double plays, which are mostly turned by middle infielders anyway. DRS’s Good Plays/Misplayed Runs Saved rates Votto -2 while UZR’s Errors Runs Saved rates him at 3.9. This could indicate that Votto is better at avoiding errors but does not make as many unexpected plays as he could. His range his where he scores big in DRS, posting a +10 plus/minus for DRS. UZR has his range at a more modest 3.2. Still, both statistics agree that the Gold Glove winner has been above average defensively the past three plus seasons.
Since UZR is a counting stat (DRS is too), a better way to compare players with different amounts of playing time is with UZR/150, which is scaled to provide a player’s value over 150 games, or roughly one season. Votto’s UZR/150 since 2016 is 2.1. How does that fare among other established first basemen?
For UZR/150, Votto comes in 7th out of 21 and is also much closer to some of these top players. Brandon Belt is the leader (5.9) with Joe Mauer (5.7) and Mitch Moreland (5.2) getting more credit from UZR. Freddie Freeman (3.5) and Anthony Rizzo (3) edge out Votto, who is slightly ahead of Paul Goldschmidt (2). According to DRS, Votto was the 9th best out of 21 players that accumulated more than 250 games worth of time at first base. Leaders include Brandon Belt (34), Rizzo (29), Freeman (22) and Goldschmidt (21).
So what does this tell us? Again, is it not advised to take any of this data completely for fact. But between these two metrics and our estimations of ability from observing each player, there are certainly some conclusions we can draw. Votto is still in the better half of the league for first basemen, even in his age 32+ seasons. Brandon Belt is probably the best defender at first, which the eye test would corroborate. The interesting question that UZR brings up is, is Joey Votto a better defender than Paul Goldschmidt? Goldschmidt is certainly more athletic, and many feels is a more “complete” player than Votto. Which number should we believe?
The eye test would probably say Goldschmidt is better. DRS gives Goldschmidt an advantage in both his double play ability and the Good Plays/Misplayed Runs Saved. UZR agrees with those two but rates Votto’s range superior. Maybe it is. Maybe the Reds defensive positioning has helped Votto’s performance. Either way, there is at least some data that suggests Votto isn’t quite as far off his new NL Central counterpart as some might think.
Beyond Votto
Besides Votto, there are not too many Reds on which DRS and UZR agree. Scooter Gennett and Jose Peraza (SS) are rated below average by both DRS and UZR with just a slight variance between the two. Three players that have rathe large variances are Eugenio Suarez, Jose Igleasis and Yasiel Puig.
Suarez has accumulated +10 DRS since 2016, putting him 10th of 19 third baseman. UZR/150, however, gives him -0.1, good for 16th. UZR really does not like his high number of errors, while DRS gives him a positive Good Play/Misplayed rating, indicating he must be making a good number of unexpected plays.
Yasiel Puig’s DRS (+32) comes in as the 3rd best out of 22 right fielders, behind Mookie Betts (+93!) and Jason Heyward (+41). UZR/150 has Puig as the 8th best at +3.9, light years behind Betts’ +19.4. Puig’s arm and range both get seriously devalued according to UZR. After watching him every day for three months, I think I lean more towards his DRS number. Maybe that is biased, but the fact that DRS shows him as an elite defender counts for something too.
Jose Iglesias gets the opposite treatment as Puig, showing more value according to UZR than DRS. DRS gives him just +14 runs saved, good for 9th out of 26 which seems low considering his incredible performance this year. For UZR/150, he posts a 10.1, good for 3rd place behind Andrelton Simmons (+18.2) and Francisco Lindor (10.9). The biggest factor in his higher UZR is his ability to make hardly any errors. Again, I may be biased, but UZR seems more accurate given the show that Iglesias has put on for the Reds.
One of the clear limitations of both DRS and UZR is the amount of data needed. Both stats should have at least three years of data to get a true representation. That leaves players like Jesse Winker, Nick Senzel, and certainly Michael Lorenzen in the outfield, with too little data to make any realistic presumptions. Directionally, UZR shows Winker about the same in both corner outfield spots while DRS shows a better fit for him in LF, which makes sense. Both metrics show that Senzel still needs time to adjust to CF, with DRS slightly more optimistic.
Like other advanced metrics, UZR and DRS will never tell the entire story and are best used in conjunction with other tools to create a well-rounded evaluation. The idea is to use the incredible amounts of data at our disposal to create the most accurate understanding. With DRS and UZR (plus many other defensive metrics), evaluating defense may not be perfect, but our evaluation of players is certainly more accurate than it would be without them.
