Why We Go To WAR: The Use of Modern Stats For Evaluating Defense

This is the final part of a three part series detailing the use of advanced metrics on BurningRiverBaseball.com by explaining their value in evaluating players on a deeper level. During the second part (found here) I covered the pitchers side of the game, focusing on FIP and more advanced analysis of thrown baseballs. Today, it’s time to head into the field as we take a comprehensive look at fielding percent, UZR and the even more advanced Inside Edge Fielding.

UZR vs FLD%

It has been known for years that fielding percent is an inaccurate portrayal of any defenders fielding prowess, but there has never been a significantly better replacement. While there have been stats that could help qualify fielding percent, like range factor, these didn’t actually consider individual plays, but plays made per game assuming all fielders at each position should be making the same amount. While still nowhere near perfect, Ultimate Zone Rating (provided by fangraphs.com) is the best we have right now. The biggest problem with fielding percent is that it completely ignores range. A slow fielder who doesn’t even come close to some balls can easily have a better fielding percent than a high flyer who is constantly diving around the field. UZR includes four parts, all judged in the total runs saved above the average fielder. These parts are arm strength, double plays, range and actual errors.

Not to pick on Kipnis again, like I did in the offensive edition, but he provides the perfect example of how fielding percent doesn’t tell the whole story. The chart above shows Kipnis’ numbers for fielding percent, range factor per nine innings (and the same for the leagues second basemen as a whole), his UZR and his UZR averaged for 150 games. An average second baseman in the Major Leagues has held a fielding percent of between .983 and .985 over the past three years, so considering fielding percentage alone, Kipnis would appear to be among the league’s best. Unfortunately, this is not the case. Both range factor and UZR confirm that Kipnis’ limited range has actually made him a below average defender. Breaking down UZR further, he is given credit for his sure handedness as he saved 2.5 more runs than the average fielder by not committing errors. However, he allowed 8.5 more runs than the average fielder, leading to most of his -8.0 total UZR.

Accuracy must be still brought into question with UZR as defense is much more subjectively judged than offense. A hit is a hit and a foul a foul, but who is to say exactly how far a positions range should extend. In this case, UZR succeeds by splitting up the field into zones and judging players by their ability to field balls in certain zones. This is broken down even further by fangraph’s Inside Edge Fielding, where we can see that Kipnis made 0% of 12 plays that would be considered impossible for the average second baseman, 7.7% of those considered remote (would be made between 1% and 10% of the time by a normal fielder), 10.5% of those considered unlikely (10-40% conversion rate), 42.9% of those considered even (40-60%), 75% of those considered likely (60-90%) and 98.3% of those considered routine (90-100%). Thanks to these advanced metrics, we can go far beyond the simple statements of “Jason Kipnis sucks on defense” or “he should be a Gold Glover because he doesn’t commit any errors.” Taking all the numbers into account, Kipnis is a great fielder on balls hit to him (in that 60-100% range), but is simply not capable of going out of his comfort zone.

Unlike BABIP, by considering actual events, UZR is a valuable now and in the past. The problem here is that no one was recording it for games taking place prior to 2002 and there are no Inside Edge numbers from any year before 2012. This makes it very difficult to compare modern players to the greats of the past, an age where Gold Gloves were decided by fielding percentage, flashy glove work and who was being snubbed from another award instead of who was the best at converting batted balls into outs. It is possible that even the hardest critics of advanced baseball metrics may be coming around on UZR as well as this year, five of the top ten players in UZR in the AL took home Gold Gloves with another, Jackie Bradley, Jr., likely being excluded because of limited innings. In all, the only real problem was Eric Hosmer robbing Albert Pujols and Mark Teixeira at first base and considering the past coast bias, this could just be seen as retribution.