Defensive Runs Saved Updated to Fielding Bible Volume III
All of the Defensive Runs Saved data on FanGraphs has been updated to reflect the numbers available in the Fielding Bible Volume III. These are now available in the player pages and leaderboards.
I am going to quote directly from the Fielding Bible Volume III on the changes:
1) Timer Plus/Minus – This is the most important new addition to the DRS System. In 2009, Baseball Info Solutions started timing every batted ball from the moment it was hit to the moment of the resolution of the play (hit or out). By using the timer, we have taken some key subjective elements of the system and made them objective. For example, we used to measure the velocity of a batted ball with an estimate, calling it softly hit, hard hit, or medium. Now, we know the exact velocity of every batted ball. We have added this tour cornerstone methodology, the Plus/Minus System.
2) Good Plays and Misplays – This is a brand new methodology. When a player makes a play that is better than the norm, we want to give him more credit. Or if it’s worse, we want to take credit away. We have 28 categories of Good Plays and 54 categories of Misplays. We have converted the ones that don’t overlap with our other methodologies to Runs Saved estimates. With this system we especially made great strides in measuring catcher defense.
3) Outfielder Arms – Here we use the new batted ball timer to upgrade our system to measure the impact that outfielders make defensively with their throwing arms. We also now use a more precise location of batted balls to enhance this methodology.
4) Double Plays – We’ve upgraded our system to measure how good infielders are at starting and turning the double play by doing something similar to what we do with our Outfield Arms technique: measuring the location and the velocity of the batted ball in a double play situation to determine how difficult the play is to turn into a double play.
5) Bunts – Same thing here. We now include the location of bunts among other criteria to better evaluate how fielders make plays on bunts.
6) Centering – The intention of Defensive Runs Saved is that zero is average. If a player has a positive number of runs saved, it’s an above-average performance. If he has a negative number of runs saved, it’s below average. As we’ve added on to the system over the years, we veered away from zero a bit. To get back on track, we took every methodology and “centered” it at zero.
This is terrific stuff. Will these changes affect the fielding component of Fangraphs’ WAR at all going forward? Retroactively?
Pretty sure fWAR uses UZR.
Brett Gardener smash
Interesting to look at the UZR/150 top ten by position: 1LF, 1RF, 1 CF, 4 3B, 3 2B.
Only 4 teams gave 900+ innings to negative value defensive 2B (and 3 of 4 are between -2 and -4; only Uggla is deeply negative).
5 teams gave 900+ innings to negative 3B, and 3 of five were double digit bad. So many great and terrible 3B at least has the semblance of the bell curve I’d expect to see across MLB.
Is this a great era for 2B, or are the fielding metrics too kind?
I would love to see a Fangraphs archive, so we could compare the evolution of a stat across particular players (ie. Column A is Defensive Runs Saved v.2011, Column B is Defensive Runs Saved v.2012).
“We have added this tour cornerstone methodology” presumably “to our”; is the typo here or in the Fielding Bible?
Good stuff! I like it!
“By using the timer, we have taken some key subjective elements of the system and made them objective. For example, we used to measure the velocity of a batted ball with an estimate, calling it softly hit, hard hit, or medium. Now, we know the exact velocity of every batted ball. We have added this tour cornerstone methodology, the Plus/Minus System.”
This metric, to have any meaning at all, must still rely heavily on subjectivity. Fielding a slow roller is one of the most challenging plays an infielder can make. It requires the fielder to make a decision he normally doesn’t have to make: whether to field the ball bare-handed or with his glove. It’s a play that requires a lot more concentration than fielding a smash. The fielder’s arm strength and arm accuracy become much more important on a slow roller than on a hard-hit ball. An infielder will cover substantially more ground while fielding a slow roller than he will on a hard-hit ground ball or line-drive; if the fielder fields the ball off-stride, then it will take a bit longer to get off a throw – a delay which can make the differerence between a hit or out. The difference between an infield single and an out on a slow roller is also primarily dependent upon the speed of the batter.
And yet, by the author’s comments, I presume this “Timer” variable will give preference to hard hit balls rather than slow rollers. In other words, a third baseman who takes one step to his left and then dives on a hard hit ball and makes a play will be deemed to have better range than a third baseman who makes a play on a slow roller in which he may have to sprint 20 yards and then throw across his body while on the run, and across the diamond, to make the play.
The timer isn’t the only component, however, it’s just another one of the many tools for analyzing defense. The range necessary to make a play is still being taken into account.
“And yet, by the author’s comments, I presume this “Timer” variable will give preference to hard hit balls rather than slow rollers.”
Why do you presume that? I don’t see anything in the author’s comments that would make me think a 3B would get some bonus for a hard hit ball. If they know the exact amount of time it took to get to his glove, that should make evaluating whether a play at first was possible much more objective.
@Gregory H, check out the FAQ:
http://www.fieldingbible.com/Fielding-Bible-FAQ.asp
That should ease your concerns.
This is disappointing to me, ” A play made by one fielder never counts against any other fielder. We remove those plays from the denominator. For example, say there are 100 balls hit into a certain “bucket” (balls with the same trajectory, location, velocity, etc.). Suppose the left fielder handles 40 of them, the center fielder handles 50 of them, and 10 fall for hits. It’s obviously a pretty easy play for both fielders, since it rarely fell for a hit. The difficulty level of the play (which we call the “Ratio”) for the center fielder is 50/(100-40) = .83; for the left fielder, it’s 40/(100-50) = .80.” This basically says that if a 3B catches a pop-fly 1/2 of the way between their positions, he gets huge credit while a SS gets very little (since SS really should handle most of these).
In the above example, it really should be that both have denominators (100-90) since they are only saving 0.1 runs, regardless of how often other players who play that position make the play.
You are apparently unaware of how these fielding metrics are calculated. I suggest you read the FBIII for an explanation. FYI, ground balls and air balls are “binned” according to the air or ground time and distance and location (distance is not really an issue for infield ground balls). So one of the bins might be a ground ball down the third base line traveling between 30 and 40 mph (I am making up the speed). If in fact, those kinds of “slow rollers down the line” are only turned into an out 20% of the time, then that becomes the base line. If a third baseman turns such a ball into an out, he gets credit for 80% of an out. Get it?
If I’m not mistaken, I believe that UZR is used in the Fangraphs WAR calculation, not DRS. So I don’t think these changes should affect a players’ WAR.
Any place where we can find projected 2012 team defensive rankings?
awesome! now when are you going to change the WAR calculation to use more than just UZR?
I’d like to hear Cameron or Slowinski or someone weigh in on how a more accurate DRS might impact the usage of UZR as the sole basis for the fielding component of Fangraphs’ WAR.
Anyone else?