Search This Blog

Friday, April 15, 2011

Defensive Metrics

Fielding metrics are easy brandy about but quite difficult to understand. They can be a very helpful tool, but, like power, they come with great responsibility. To maintain any semblance of credibility, you must know what you are talking about and how to most effectively use these advanced statistics.

Before we delve into the types of defensive statistics, I must make one thing clear – do not use fielding metrics without ample sample size. Most people like a sample size of at least three seasons, or 3500-4000 or more IP. With how radically the ratings can fluctuate, it is prudent to gather a sample size before making your judgement.

Similarly, we need to have an established baseline. If a guy comes up and posts ratings of -10, +12, and 0, you still don’t have a baseline, and metrics will be of little use. If that same guy goes on to post ratings of, say +7 and +10 in the next two seasons, at that point it is safe to say that he is a pretty good fielder.

Finally, fielding metrics do not substitute every other method of evaluation. Scouting reports, opinions from baseball minds, even your own opinions, can be taken into account, but realise: for all the inaccuracies in defensive metrics, it is, in the right circumstances, relatively accurate empirical data, and your naked eye test isn’t.

With all of that said, here we go. There are three basic types of defensive metrics I would like to address: total zone runs, ultimate zone rating, and the plus/minus system.

Rtot (Total Zone Runs) - Once described to me as basically a play-by-play-based estimate of Zone Rating for seasons prior to STATS Inc.'s exhaustive video tracking of Zone Rating data. It credits/debits fielders for their hits saved/allowed compared to an average fielder at the same position, given how many balls were hit into their zone of responsibility. These hits are then converted into run values based on the linear weights value of each type of hit (i.e., Palmer's Batting Runs). Additional run values are also applied for OF and catcher arms, and IF double plays. Finally, Rtot is the sum of the runs above/below avg. for balls hit into the zone of responsibility, OF/C throws, and IF double-play chances.

Basically, the statistic that gives us the final rating is called rtot. This is an amalgamation of a handful of other stats, namely Rtz (Total Zone Fielding runs above average) – the value of the players’ ability to get to balls and make plays on them, Rdp (Total Zone Infield Double Play runs above average) – the value of an infielder’s ability to turn double plays, Rcatch and Rof, the value of the arms of both catchers and outfielders, respectively. These are added up, according to what position the player plays, and give you rtot, a total amount of runs that player saved or cost his team. Anything above 0 is above good, below 0 is bad. If you want to know what a really good rtot rate is, Andruw Jones was annually posting rates of 25-35+ in the late ‘90s and early 2000s, but usually anything above 10 is very good. Rtot and fielding runs are featured on

There is more reading on the topic here:

UZR (Ultimate Zone Rating) – Ultimate Zone rating requires very little explanation if one already understands Total Zone, because that’s basically what it is. UZR takes into account the amount and difficulty of balls that the fielder gets to, to give a Rngr, or range run, and combines this with ErrR, or error runs, which assigns a value to the players’ propensity for making errors per balls received. These are combined into UZR. UZR is featured on under players’ defence (advanced)

UZR uses info from Baseball Info Solutions, as does a statistic featured on called rdrs, or BIS Defensive Runs Saved, right beside where they provide rtot information. My personal suggestion is that, if you want to use the combined information of rtot and UZR, you are better served saving the 5 seconds it would take to go to fangraphs and just use rdrs.

Brief explanation from fangraphs --

Now, it should be noted that rtot, UZR and rdrs are all counting stats – that is, like hits, home runs or strikeouts, the statistic value will continue to rise (or decline) as the player accumulates chances. To counter this, Baseball-reference and Fangraphs provide rtot/yr and rdrs/yr (the number of runs saved over 1200 defensive innings played), and UZR/150 (defensive runs saved per 150 defensive games played).

Plus/Minus – I think this was said best on The Fielding Bible’s own website: “My book, The Fielding Bible, goes into great length (ad nauseum to some) describing the new fielding system we developed at Baseball Info Solutions, the Plus/Minus System. Video Scouts at BIS review video of every play of every major league game and record detailed information on each play, such as the location of each batted ball, the speed, the type of hit, etc. Using this in-depth data, we’re able to figure out how each player compares to his peers at his position. How often does Derek Jeter field that softly batted ball located 20 feet to the right of the normal shortstop position, for example, compared to all other major league shortstops?

“A player gets credit (a "plus" number) if he makes a play that at least one other player at his position missed during the season, and he loses credit (a "minus" number) if he misses a play that at least one player made. The size of the credit is directly related to how often players make the play. Each play is looked at individually, and a score is given for each play. Sum up all the plays for each player at his position and you get his total plus/minus for the season. A total plus/minus score near zero means the player is average. A score above zero is above average and a negative score is below average. Adam Everett turned in the highest score we’ve had in four years of using the system with a +43 at shortstop in 2006. That means he made 43 more plays than the average MLB shortstop would make.”

The obvious detriment with the Fielding Bible is that I don’t have one, and, probably, neither do you. It is an impressively deep system that is very interesting and very useful, but unless you’re willing to pay, hard to get your hands on. On their website you can see past winners and vote totals, which, in my opinion, are worth more than Gold Glove Awards. Another detriment of this system is that it is not weighted. Adam Everett may have made 43 more plays, but how many hits did he take away. 43? Were they all singles? What is the value of a removed single? The aforementioned statistics take care of this stuff for you and simply tell you the value of the defender’s play which, as a casual fan, is all one really needs.

So there you have it. The unnecessarily complex world of defensive metrics compounded into a 1200-word nutshell. If you managed to stick with me through the entire text, congratulations, you now know more about defensively rating players than the grotesque majority of baseball fans.

No comments:

Post a Comment