Wilmer Flores and the UZR/DRS discrepancy

This GIF comes courtesy of Carson Cistulli of FanGraphs, who posted it in mid-September. The play itself is from an August game (how fitting is it that Jon Niese is on the mound?) It seems safe to say that this is how we feared Flores would be at shortstop. The truth is that he wasn’t bad at all in 2014.

UZR loved his defense at SS last year, giving him a +4.0 mark in 443.1 innings, which translates to a 12.5 UZR/150. Among SS with at least 400 innings played last year, that 12.5 UZR/150 was the sixth-best mark in the majors. If that was the only piece of information we had to go on, we would be pretty excited about a full year at short from Flores in 2015.

However, we know that Flores was moved from the position earlier in his career because it was thought he couldn’t play the spot in the majors. Additionally, we know that 443.1 innings is nowhere near a large enough defensive sample to form good opinions. We also have the knowledge that another respected defensive system in DRS saw that same 443.1 sample completely differently. DRS gave him a (-3) last year.

The two systems see roughly a two-win difference over a full season, which is not terribly unusual over such a small sample. But it’s a different story over a full season of playing time. Last year there were 22 SS who played enough to qualify for the FG leaderboards and only two of those players had a one-win difference. UZR saw Andrelton Simmons as the best shortstop in baseball with a 15.5 mark. DRS saw him as otherworldly, with a 28 rating. They both saw him as great, just one took it to an even further level.

The other one that had more than a one-win difference was much more interesting. UZR liked what it saw from Erick Aybar, as it had him with a 7.5 mark. Meanwhile, DRS rated him at (-3) for the year. It’s the third time in the last four years that DRS gave Aybar a negative numbers. He finished in the red in UZR just one time in that span. But 2014 was the first time there was a significant difference between the two systems. In 2011 the difference was 2.8 runs, in 2012, the year he was positive in both systems, it was 0.5 runs and in 2013 it was 0.4 runs difference as both had him as a negative fielder.

If we go back to 2013, not one of the 21 qualified shortstops had the two systems on opposite sides of the spectrum with a player and a difference of at least 10 runs. It happened three times in 2012 (Peralta, Rollins, Desmond) out of 22, with the 15.9 difference of Jimmy Rollins being the largest gap. So, over the last three years, we have only four times in 65 seasons where there was a polar difference between the two systems of at least 10 runs and not once did that difference reach the 20-run level that we extrapolated from Flores’ partial season in 2014.

What does all of that mean? If Flores plays shortstop 100 or more games in 2015, it’s extremely unlikely there will be such a difference of opinion in the results of DRS and UZR. The difference is overwhelmingly likely to be less than half of what it appeared to be this year. But we have no idea how the two systems will move to a greater consensus. It’s possible that Flores’ DRS will move into positive numbers. It’s possible his UZR will end up in negative numbers. It’s also possible that his UZR will fall some while his DRS rises a bit. Anyone who pretends to know is just guessing.

In 2014, Flores posted a 1.3 fWAR in roughly half a season. If given the chance for Flores to play SS full-time and put up a 2.6 fWAR, I would sign on the dotted line immediately. However, fWAR uses UZR and the reason Flores had such a strong season was because of his defensive numbers. Therefore, it’s possible Flores could take a step forward offensively in 2015 and not produce as much fWAR value as he did a season ago if his UZR moves strongly to match his DRS.

The only projection available now is the one from Steamer. It shows Flores with a .697 OPS and a .306 wOBA, both numbers an improvement over what he did in 2014 but not anywhere close to the .935 OPS and .399 wOBA he posted in Las Vegas in 2014. This is how it should be. We know there is a lot of air to be removed from those Vegas numbers and research suggests to remove 19% from OBP and 34% from SLG. So, our MLE for what Flores produced in Vegas last year would be a .672 OPS. As a young player, we would expect improvement from that. Is a .697 OPS the appropriate level of improvement? That’s certainly debatable but it’s hard to look at it as some horrible prediction.

The Steamer forecast also has a UZR and fWAR projection. Overall, it has him with a positive defensive worth and a 1.9 fWAR in 512 PA. It will be curious to see what ZiPS and the other projections will forecast for Flores this year. My take is that if he can post a .700 OPS the Mets can live with his defense, even if it’s not as good as what Steamer projects here. Shortstop figures to be the weakest position for the Mets in 2015 and if they can get a 1.5 fWAR from their starter at their weakest position that wouldn’t be the worst thing in the world. However, it should be pointed out that the one year Ruben Tejada cleared 500 PA (2012, when he had 501 PA) he had a 1.8 fWAR.