On Saturday, March 17 at the SABR Analytics Conference Cleveland Indians President Mark Shapiro held a Q&A, moderated by Ken Rosenthal of Fox Sports. When the topic of current defensive metrics was brought up Shapiro said this:
"We look at conventional statistics that are available to everybody, and we've got our own proprietary statistics. Where we feel we're at as far as objective measurement of defense today is somewhat around the equivalent of using batting average for offense. It has some value, certainly very limited value ... but we factor it in…"
The fact that a major member of the sabermetric research community said this in a public forum is fascinating. While most analysts understand that the current defensive metrics have holes, this statement has to have some resounding influences.
First, can the theory that sabermetrics only look at stats and nothing else be put to rest? The main argument I personally have with readers of my own blog and on twitter (and I know many others deal with this too) is that the sabermetric statistics used only look at numbers and not the on field performances, the whole picture, if you will.
As I constantly have to explain, sabermetrics are used to further our understanding of the game, not replace the traditional fundamentals (pitcher wins excluded, but that’s another topic for another day).
Second, without discrediting any of the defensive metrics that are currently out there, (I use them as much as anyone) I simply ask what Shapiro asked: where do we go from here? Sure, we have UZR and DRS and Fielding Percentage and Rtot (etc, etc, etc) but there are still some obvious holes that need to be considered.
On Sunday March 18th, the Tigers were playing the Nationals in a grapefruit league spring training game. Doug Fister was starting for Detroit and started the game off strong. He struck out the first two Washington batters, Ian Desmond and Danny Espinosa and looked like he had a great feel for his pitches.
Ryan Zimmerman came to the plate and grounded a ball up the middle of the infield, a playable ball for second baseman Ryan Raburn that should have gotten the Tigers out of the inning. However, Raburn bobbled the scoop and then proceeded to throw the ball behind the first baseman, Don Kelly. Zimmerman was safe at first base and with two outs, Fister was forced to throw against another batter, Jayson Werth.
This time, Fister was unaffected. His fastball had zip, his curve had a snap, and he struck Werth out to end the inning, and despite Raburn’s error, ended up striking out the side.
The top of the second inning ended for the Tigers offense like the first did; a few threats but in the end scored no runs. Fister walked out to the mound in the bottom of the inning with the score 0-0 and a positive first inning under his belt.
Chad Tracy led off the inning by promptly striking out just as the other Nationals before him. Five batters, four out, four strikeouts. The next batter, Jesus Flores, managed to fight off a pitch and single on a soft line drive near Ryan Raburn at second. Six batters, four outs, four strikeouts. Bryce Harper came to the plate and struck out on four pitches. Seven batters, five outs, five strikeouts.
Tyler Moore hit a line drive to left field that should have been caught by Andy Dirks. It should have ended the inning. It didn’t. The Tiger fielder misjudged the ball and dropped it for an error. Flores went to third base on the play.
So now the situation looks like this: Fister had struck out five of the first eight batters he faced before Roger Bernadina came to the plate. Fister had been rolling up to that point, if you like to use that term, and he looked comfortable. However, with two outs and two men on base, the latter because of the error, the point many argue, "Well it’s not Fister’s fault", isn't necessarily accurate. Instead a question should be asked, such as: "How is Fister going to respond to this situation?" The ball was still in his court. He has one out to get and can get it anywhere on the field. It should be no different than the situation in the first inning.
Yes, we know that if Dirks had made the catch, Fister wouldn’t have had to deal with the next batter, Bernadina, who in turn singled to the right fielder Brennan Boesch, scoring Moore. We know the runs are unearned on Fister’s ERA, even after Ian Desmond followed Bernadina with another single and gave the Nats a 2-0 lead.
The small rally was killed when Desmond was thrown out at second to end the inning, but the damage was done regardless of how it happened.Unearned runs are still runs and Fister still had control over allowing those runs to score even though the defense had just as much responsibility making the plays to get out of the inning.
A defensive error led to two runs, but how do we quantify the runs after the error? How do we quantify the fact that Fister lost his momentum and obviously became unsettled? He was cruising before the error. His pitches had zip on them. After the error they flattened out. The result was obvious.
Personally, I don’t think this is something that can be valued in a numerical format. Raburn made an error in the first inning and it resulted in no damage. Dirks made an error in the second inning with a man on and two runs scored before the inning ended. Where’s the difference? Where’s the reason that one inning resulted in zero runs and the other resulted in two?
How do we quantify that? How do we know that Dirk’s specific error was a direct reason for Fister’s ensuing struggles? We can't answer that with a definitive solution. How do we measure the effect error’s have on a pitcher just as the debate how we quantify "closer mentality" or "batting slumps"?
I think this is what Shapiro was referring to when he compared the defensive metrics to batting average in terms of what they tell us. It’s not the whole story. It’s why there is such a divide in how people view Fister’s teammate Miguel Cabrera’s move to third base. We just don’t know. We assume it's going to be bad. But what if Cabrera has 25 errors this year and 20 of them come in a situation where the error doesn't result in any runs such as Ryan Raburn's error in the first inning?
Frankly, unless some genius flies in on wings of the baseball gods, I doubt we ever can know fully how errors effect pitchers other than a case by case basis.