Image credit: © Mark J. Rebilas-USA TODAY Sports
- What benefits does RDA bring to the world of defensive evaluation?
- Very high reliability, as compared to other defensive metrics, when comparing fielder positional ranks from season to season.
- Uniquely good calibration, meaning that fielders at a position tend to be graded in a consistent range from year to year, as opposed to the more extreme differences sometimes seen with other metrics.
- The Range Out Score, which provides a simplified measure of a fielder’s range contributions at each position on a rate basis.
- The Attempt Range measurement, a count statistic which calculates a shortstop’s number of plays above or below average on which they were able to reach ground balls that the shortstop either did field or that RDA feels they should have fielded.
- Very good accuracy, even without specific fielder coordinates.
- Another strong metric to consider in the world of defensive evaluation.
- What components of fielder defense does RDA cover?
- RDA addresses fielding plays made on balls put into play that do not end up being home runs. So-called “throwing runs” (outfield assists) or deterrence of lead baserunners are not addressed by RDA and will continue to be calculated through traditional methods through DRP.
- What is the relationship between FRAA, DRP, and RDA?
- So many acronyms: we are sorry. Deserved Runs Prevented (DRP) is the new umbrella for all BP defensive metrics, including fielder range, catcher defense, and throwing defense/assists. For positions other than catcher, Range Defense Added (RDA) will be by far the largest component of DRP values. FRAA is BP’s previous defensive evaluation system, and will now be used only for leagues that do not report Statcast batted ball data. On our leaderboards, either FRAA components or DRP components will be reported for a league, but not both.
- Are these new fielder defense and team defense values available on BP’s leaderboards?
- Yes they are. Please check out player-team values here. Player-season and team-only values will arrive shortly.
- Does 2023 PECOTA project DRP values, including RDA, even though they were just released?
- Yes, for players with major league experience. Otherwise, FRAA is used.
- Are RDA and DRP values now incorporated into BP’s Wins Above Replacement Player (WARP) for all seasons where RDA is available?
- Yes. DRP and its RDA component will replace FRAA to calculate defensive contributions in applicable seasons.
- How does RDA decide to credit a fielder for a ball?
- Fielders are credited for all plays on which they are deemed the responsible fielder. For outs, the responsible fielder is the player who is deemed to have initiated the out. For non-outs, the responsible fielder is the player who is deemed to have fielded the ball, with the exception of outfield ground balls, a portion of which are charged to the shortstop at RDA’s discretion.
- How does RDA treat plays charged as errors by the official scorer?
- RDA addresses plays scored as initial fielding or initial throwing errors for infielders, but fielding errors only for outfielders. For these plays, RDA ignores the official scorer decision and recalculates the probability for success on its own. Typically, RDA agrees that these plays should easily have been made, but not always.
- How does RDA address the receiving end of fielding plays and double plays?
- At the moment, like Outs Above Average (OAA), we are focused on the initiating end of a fielding play, and we use a run value for outs that is based on the overall average value of an out in the infield or outfield, respectively. We expect these other issues (such as first-base scoops) to be addressed later, but the initiating player typically is the most important participant in these exchanges.
- Does RDA account for the dimensions of different parks?
- For outfielders, yes.
- Does RDA distinguish between singles and extra-base-hits?
- Somewhat, in that we assign a different value to an infield out versus an outfield out. At least for outfielders, we have experimented with making this more explicit, but so far have not incorporated it.
- How is Range Defense Added converted to runs?
- As noted above, we use different average run values for plays fielded by infielders versus outfielders. We expect this assignment will become more nuanced over time, such as by rewarding infielders who stop balls from going into the outfield, even if they are not converted into outs.
- Does RDA share any components or philosophy with BP’s “deserved” family of metrics, like DRC+, DRA, or catcher framing?
- Philosophy: yes, components: no. Furthermore, because fielding is complicated, there is a lot more going on.
- How does RDA deal with situations where one fielder may have “stolen” somebody else’s play?
- In general, we don’t think “vulturing” is a problem: most plays likely will only ever be fielded by a single defender, with the possible exception of some balls fielded by pitchers. However, RDA guards against this naturally because it calculates the aggregate probability that one of the considered defenders should have made the play, not just the fielder in question. Thus, busybodies who interfere with somebody else’s routine play would not be rewarded for making the play more difficult.
- How does RDA address infielders who inflate their success rates (their numerator) by getting to fewer balls (their denominator)? You know, like that one shortstop.
- This is the concept of Attempt Range. For shortstops, an additional model analyzes outfield ground balls and reassigns them to the shortstop as unconverted outs if, on average, a shortstop would have fielded the ball.
- Attempt Range sound great. Why not expand it to other infielders?
- It’s on the list for future development. Shortstop is a unique position and implementing Attempt Range for shortstops was deemed a development priority. Range Outs Made, the Range Out Score, and the rest of DRP are available for all positions now.
- How does RDA deal with plays that are missing Statcast inputs?
- RDA is accurate on a rate basis even when these plays are excluded, likely because missing measurements, especially in recent seasons, are distributed among players. However, we have developed a sophisticated imputation scheme that, by calculating plausible values for these missing measurements, actually increases the accuracy of RDA even further, and allows us to include coverage of the 2015 MLB season.
- How can RDA match or exceed the performance of Defensive Runs Saved (DRS) and OAA, when it does not have access to the additional resources used by those systems?
- Robust statistical methods are a wonderful thing. And the current information MLB provides about batted ball quality and overall fielder positioning, while limited, is sufficient to benefit from those methods.
- Which positions does RDA evaluate most successfully?
- By our preferred benchmark, RDA performs as well or better than OAA and DRS at all infield and outfield positions. Its performance advantage is greatest in the outfield. DRS is a bit more consistent at pitcher and catcher, although RDA does not see those positions as offering much range defensive value.
- How trustworthy is an RDA evaluation of a fielder?
- All RDA range rates are reported with an uncertainty estimate, which confirms that fielding ratings are, well, uncertain, even if the fielder is probably doing a good job. No more need to automatically “average 3 years of ratings” or other arbitrary rules of thumb. You may still wish to combine multiple years of estimates if desired, but now you will be able to do so using specified uncertainties to gain a more confident estimate. In our opinion, all defensive metrics necessarily have an uncertainty component, whether they report it or not.
- Are other fielding metrics besides RDA still worth consulting?
- Advanced metrics like OAA and DRS? Absolutely. Older metrics remain helpful when modern metrics are not available, or when evaluating retired players.
- Is Baseball Prospectus critical of the methods used by OAA and DRS?
- No. All methods have strengths and weaknesses. We recommend focusing less on treating any one metric as gospel. Instead, be impressed when multiple metrics agree and inquisitive about the reasons when they do not. We discuss some of the reasons why RDA, OAA, and DRS may differ in our introductory article. Obviously, we think RDA is a great choice.
- How does RDA’s coverage of seasons and leagues compare to OAA and DRS?
- RDA currently covers all positions from the 2015 MLB seasons to the present. More leagues and seasons will be covered as Statcast batted ball data becomes consistently available at other levels.
- Our understanding is that OAA currently covers all positions except pitcher and catcher from the 2016 MLB seasons to the present.
- Our understanding is that DRS currently covers all positions for the past 20 years or so of MLB seasons.
- Which “splits” will be available for RDA values?
- Player-team right now; player-season and team-only shortly.
- Can baseball organizations request that RDA be calculated for amateur or professional leagues for which there is Statcast-type batted ball data available?
- Yes. Please contact Bret Sayre for details.
- Can I approach BP with additional questions I have about RDA?
- Of course, and please do. If you have a question, somebody else is probably wondering about it too. Social media is usually the best way to reach us. Polite inquiries are both appreciated and consistently responded to. Rude inquiries have a poor track record.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.
Subscribe now