Introducing FanGraphs Stats: Offense and Defense

FanGraphs has a lot of metrics, ranging from the standard counting numbers that have filled baseball cards for decades to more complicated calculations and formulas. Odds are pretty good that if there’s something you want to know about a particular Major League player, we probably have a metric for it. However, up until today, we’ve lacked one number that answered two of the most basic questions that people ask: Who has been the best offensive player in the game, and who has been the best defensive player in the game?

You could get at those questions with the numbers we had, using a combination of metrics found on the site, but there wasn’t just a single number that simply answered those two questions. And, unfortunately, the lack of an available sum of offense or defensive value led to people occasionally substituting in a number that didn’t actually answer the question they were trying to answer. This was our fault, not theirs, but as of an hour ago, this has been rectified with the introduction of two new numbers here on FanGraphs.

On the dashboard and in the value section, you will find two columns simply labeled Off and Def, which stand for Offense and Defense. These numbers are the total runs above or below average based on offensive contributions (both batting and baserunning) or defensive contributions (fielding and position). While both numbers are simply the addition of two metrics previously available, this gives everyone the ability to now simply quote, in runs, how far a player was from the league average in either creating or saving runs.

For instance, here are the 2013 leaders in offensive runs, combining their batting and baserunning totals:

# Name Off
1 Mike Trout 69.6
2 Miguel Cabrera 64.0
3 Chris Davis 53.6
4 Andrew McCutchen 47.7
5 Paul Goldschmidt 44.9
6 Joey Votto 44.1
7 Matt Carpenter 40.6
8 Shin-Soo Choo 40.0
9 Josh Donaldson 35.8
10 Jayson Werth 33.7

I’m going to skip over the top two names on the list because we all know how this will go if I open that can of worms. But look down at #4 and #5; by combining batting and baserunning, we can plainly see that Andrew McCutchen has been the best offensive player in the National League this year. Without even considering the run prevention side of things, McCutchen has a slight lead over Paul Goldschmidt. This helps make the case for McCutchen as MVP even for people who are skeptical of defensive metrics and prefer to base their decision almost entirely on what a player did in the parts of the game that are easier to measure.

Now, for the same leaderboard, only focusing on the defensive side.

# Name Def
1 Manny Machado 33.4
2 Andrelton Simmons 31.1
3 Gerardo Parra 23.6
4 Nolan Arenado 22.8
5 Carlos Gomez 22.7
6 Russell Martin 22.1
7 Shane Victorino 19.6
8 Alcides Escobar 18.1
9 Yunel Escobar 18.0
10 Yadier Molina 17.1

This is where the upgrade was needed the most. Prior to this rollout, we listed “Fld” as our primary defensive rating on both the dashboard and the fielding section, which solely rated players relative to average at their specific positions, but that treats all positions as if they are equal in value, which is clearly not true. Pretty much every shortstop is a better defender than every first baseman, even the ones with negative ratings at SS compared to the guys with positive ratings at 1B. But, with the Fld metric, this wasn’t always obvious, and it led to people using a metric that rates players at their position as if that was their standing in the league overall.

By adding the positional adjustment into the fielding rating to create Defense, we’re now presenting a much better view of who the best defenders actually are. Catchers get a big boost here as well, as they have the largest positional adjustment of any spot on the field, and this better recognizes their contributions, even if they aren’t necessarily rated dramatically higher than other catchers. Just being a catcher has a lot of value, and this helps display that value more correctly.

There are some other ancillary benefits to displaying things in this matter as well. Since these are counting stats, playing time is explicitly included, so a player who has posted a .400 wOBA over 150 games will grade out better than a player who had the same .400 wOBA in only 100 games. Rate stats can be very useful, but when counting total value, you also want to give players credit for staying in the line-up every day. Now, instead of having to just cite wOBA or wRC+ and then adjust for both playing time and baserunning value, those are already baked into the Offense column.

This also helps level the playing field for guys who shift between positions. Utility players who spend parts of their season at multiple positions, or even regulars who bounce around between spots like Ben Zobrist, do not have their defensive performance easily described by a single UZR rating. If a guy plays five positions and has a +5 UZR, you don’t really know what that +5 is relative to. Now, with the fielding component added in, you know that the baseline is an average defender across all positions.

And, finally, if you don’t want to try to spend time talking about what a replacement level player is, this gives you an easy way to simply compare players relative to league average, a baseline that is easily understood without any explanation. Replacement level is an important concept, but there are times it is not necessary to add in those extra runs, and a comparison to average will do just fine. For those times, these metrics now make those citations easier.

You will find Off and Def — the shorthand labels for these numbers — in the Dashboard and Value sections of the player pages and the leaderboards, starting immediately. These will be added to the custom dashboard options in the near future. As always, a big thanks to site founder/owner/overlord David Appelman for his hard work, and we hope you find these two new numbers as useful as we do.




Print This Post



Dave is a co-founder of USSMariner.com and contributes to the Wall Street Journal.


91 Responses to “Introducing FanGraphs Stats: Offense and Defense”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. IZZY2112 says:

    In the Value Section, I see a League Column that wasn’t there before? Is that a new modification to WAR that gives AL guys some extra credit b/c it’s the tougher league?

    Vote -1 Vote +1

  2. Well-Beered Englishman says:

    Truly awesome!

    *raises glass*

    Vote -1 Vote +1

  3. AJT says:

    Can of worms? What can of worms?

    +14 Vote -1 Vote +1

  4. Sylvan says:

    Very nice!

    Tiny quibble: Juan Lagares ought to be among your top 10 defenders with his 21.7 runs. I assume he got left out because he doesn’t qualify for the batting title.

    +5 Vote -1 Vote +1

    • IZZY2112 says:

      Yeah, that’s why. Same reason why Wright doesn’t show up. Perhaps there’s a way to program it so that the qualified thing only applies when you’re sorting by a rate stat.

      Vote -1 Vote +1

      • ralph says:

        Yeah, definitely. If a player’s good enough to be ranked well in a counting stat without being qualified, that accomplishment should definitely be recognized by default.

        Vote -1 Vote +1

  5. ralph says:

    But FanGraphs still does have a stat for how many times Prince Fielder has failed to catch a foul ball, then grabbed a nacho from a fan immediately thereafter said failure.

    Slackers.

    +5 Vote -1 Vote +1

  6. IZZY2112 says:

    Also, the ZIPS/Steamer Projections now kind of display the projected positional value RoS, right? Subtracting Fld from Def. Is that number from the projection systems or is it an extrapolation of what they did in 2013.

    Vote -1 Vote +1

  7. John says:

    Terrible, bring back the baserunning adjustment to the leaderboard. I feel as though this new change is catering to the fans who immediately focus on the RBI column.

    Vote -1 Vote +1

  8. Fabian says:

    I like this new number, but I also miss the baserunning number. The nice thing about this number is that it’s now quite easy to answer the question “Would Miguel Cabrera be more valuable as a DH?” Since David Ortiz has almost the same Defensive Value as Cabrera while hardly ever playing defense I guess the answer is a resounding “It’s too close to tell”.

    Vote -1 Vote +1

  9. Eminor3rd says:

    Any reason why you’re basing these numbers against average instead of replacement level? Either is fine, but why not strive for consistency?

    Vote -1 Vote +1

    • Dave Cameron says:

      There’s no such thing as a “replacement level hitter” or “replacement level fielder”, really. There are replacement level players, and they come in all shapes and sizes, but having separate replacement levels for offense and defense has never been something we’ve done. The average baseline is consistent with how the value section has always been.

      Vote -1 Vote +1

      • Eminor3rd says:

        Makes sense. Why not just go with raw runs created then, instead of having an “above average” component at all?

        Vote -1 Vote +1

      • pft says:

        If you can separate them for average players you can do the same for replacement level players.

        There is actually no such thing as replacement level fielder since most replacements are at least as good defensively as MLb fielders. What separates a replacement level player from a MLB player is the hitting.

        Vote -1 Vote +1

  10. ralph says:

    Okay, serious question, based on this quoted portion:

    These numbers are the total runs above or below average based on offensive contributions (both batting and baserunning) or defensive contributions (fielding and position).

    I bolded “average” because that got me wondering what baselines are being used here. Is hitting now being rated against “average” instead of replacement? (I know baserunning and defense are rated against average, but it would seem weird to make batting now be rated with respect to average.)

    Vote -1 Vote +1

  11. Iron says:

    Would there be any value in a Off + Def column? Because I would not mind being able to sort by such a column.

    Vote -1 Vote +1

  12. tylersnotes says:

    Zach Greinke is ranked 9th among Dodgers in Offense (min 1 PA) this season.

    Vote -1 Vote +1

  13. August Fagerstrom says:

    I would like to add the Def metric to the custom stats that I use that show up when I go to a players page. Found under the “Customize” link. I don’t see Defense in there, though, just “Fld”. Help?

    Vote -1 Vote +1

  14. Hi mom says:

    Used this to find out that Machado’s is not just the best this year, but currently tied for the best defensive year 2000-2013.

    Vote -1 Vote +1

  15. Bryrob58 says:

    Bitter-sweet, but mostly sweet. Thanks guys!

    Vote -1 Vote +1

  16. adam Pollio says:

    The UZR and other defensive stats still give a better value when comparing players who field the same position correct? Or would the Defensive value metric give more information?

    Vote -1 Vote +1

    • Bryrob58 says:

      If I’m understanding correctly, the ol’ Fld stat (and UZR) is best when looking relative to those that play the same position. The Def stat just adds in positional value, but is extremely useful, as mentioned in the article, for those who play multiple positions and looking for a players defensive value as a whole. Like, is a “good defensive” shortstop more valuable than a “great defensive” centerfielder. It just makes comparing apples to oranges easier.

      I hope I’m on the right track here.

      Vote -1 Vote +1

    • adam Pollio says:

      Just wanted to expand on my concern. Looking at Logan Morrison and Prince fielder Morrison has a worse UZR rating, but his defensive value is better than fielder’s. I understand fielder has played more games but I assumed defensive metrics like UZR already take that into account. So how does Morrison end up being a supperior defender even though he isn’t based on UZR.

      Vote -1 Vote +1

      • Bryrob58 says:

        Fielder has roughly double the negative positional value because he has played roughly double the games there, with pretty equal defense as you mentioned. I think it is assuming that in the games where Morrison was not playing, they had an average defender at that position.

        It is accumulative, which is why a player that has a very small sample size would not be near the top of the list for Def value (or Off).

        Vote -1 Vote +1

  17. Harry says:

    This Defense number seems to be a little off… Simmons has the highest single season DRS ever and the highest dWAR, 5.4, (per b-r) since like 1906. He has obviously been the most valuable defensive player in baseball this season (if not this century), yet he is over 2 runs behind Machado, can you explain that?

    Vote -1 Vote +1

    • Dave Cameron says:

      DRS and dWAR are the same thing, since DRS is the fielding component that Baseball-Reference uses in their WAR. We have DRS listed here on the site, but it isn’t part of our defensive calculation.

      Vote -1 Vote +1

      • bstar says:

        No, B-Ref dWAR is not just DRS; its includes the positional adjustment also.

        Simmons DRS runs = 41
        Simmons pos. runs = 7
        Simmmons dWAR = 5.4

        Vote -1 Vote +1

      • bstar says:

        So if you came up with a leaderboard for this new metric using B-Ref numbers, the top 5 would be:

        1. Andrelton Simmons – 48
        2. Carlos Gomez – 37
        3. Manny Machado – 36
        4. Gerardo Parra – 34
        5. Nolen Arenado – 32

        Vote -1 Vote +1

    • I heart Andrelton says:

      Wondered how long it would take before someone complained that Simmons wasn’t in first by some defensive metric…

      Vote -1 Vote +1

    • Billy says:

      They seem a bit off to me too. Look at Brett Gardner and Shane Victorino. One went from a corner outfield spot to center, the other from center to a corner. They both rate average or slightly above average in center, and elite on the corners. This would imply their outfielding abilities are of similar overall caliber. Yet their scores are pretty different. I guess I’m questioning how we’re calculating the positional adjustment. I’m sure someone will call me a moron for this. Please don’t. I accept very well reasoned, sensible answers.

      Vote -1 Vote +1

  18. leeroy says:

    Timing is questionable. I only say that because I searched Billy Hamilton today and couldn’t see the base running right away. Made me a little salty

    Vote -1 Vote +1

  19. Grand scheme of things, I can just made one more click to see BsR, but I’ll miss it on the main dashboard.

    Vote -1 Vote +1

  20. Joe says:

    Can someone please explain the “League” column under the Value section?

    Vote -1 Vote +1

  21. Brandon says:

    BsR is back now?

    Vote -1 Vote +1

  22. Bryrob58 says:

    This brings up a related question I have had for a while. Hopefully I’ve just missed a stat that answers this, but…

    Is there an offensive stat that is positionally adjusted to show how good a hitter is relative to others at the same position? For example, a .330 wOBA from a catcher is obviously better than a .330 wOBA from a first baseman, because the average wOBA for a first baseman is greater than that of the shortstop. wRC+ is the same way, right?

    God I hope I’m just being blind to something obvious.

    Vote -1 Vote +1

    • Bryrob58 says:

      Like, for example, if we take away Baserunning from the new Offense stat, it is the Batting metric in the Values leaderboard. Isn’t Trout’s Batting of 61.6 more impressive than if a first baseman had the same number? Is this incorporated into Positional value? If so, why is Positional value only measured under the new Def stat?

      Vote -1 Vote +1

  23. Røark says:

    This looks great. Could we also get Off/150 and Def/150?

    Vote -1 Vote +1

    • Røark says:

      Or is Off/150 essentially wRC+?

      Vote -1 Vote +1

      • tangotiger says:

        It would have to be that. And if it’s not, I’d like to know the reason.

        That is, to convert from wRC+ to Off, you would do something like:
        (wRC+/100 – 1) * .12 * PA

        The .12 is because there’s around .12 runs per PA.

        Vote -1 Vote +1

  24. tangotiger says:

    For those asking why is the positional adjustment added to the fielding part and not the offensive part:

    Because when you discuss players, say a prospect, on a 20-80 scale, you don’t say “at SS, he’s a 60, but in LF, he’s a 70″. You would simply say “he’s a 65 fielder” or something like that.

    And you know that to play SS, you really need to be at least a 50 fielder.

    Similarly, you wouldn’t say that he’s got “30 power for a 1B, but 50 power for a SS”. You’d simply say “he’s got 40 power”.

    To the extent that you want to compare players across positions, there are better reasons to include the positional adjustment as part of his overall defense than as part of his overall offense.

    Vote -1 Vote +1

  25. JD Rocks says:

    I always wondered why I had the feeling Keith Hernandez was overrated, especially for his defence. The pundits would go on and on about him. Now I know – over the course of his entire career, with all the good and bad years averaged out, Keith Hernandez saved less than one run (0.3 career def) with his defensive skills. For a first baseman he was great; for his defence being the difference between winning and losing, not so much. The next time someone tells me how important it is to have a great glove at first base, I’ll be able to point them to Keith Hernandez and this fantastic new stat.

    Vote -1 Vote +1

  26. Dave M. says:

    Ortiz gets a hefty -13.6 Def rating this year, but he’s only played 26 innings in the field this season. Is his rating because he was unimpressive, and the positional adjustment gives him the big negative? Is the positional adjustment scaled by playing time at all? Or is he racking up some of that due to being a DH?

    Vote -1 Vote +1

    • tangotiger says:

      DH.

      Essentially, a DH is treated as a poor-fielding 1B, or at least, about as much defensive value as a poor-fielding 1B.

      Another good reason to rollup the positional value on the defensive side.

      Vote -1 Vote +1

      • adam Pollio says:

        Why wouldn’t the DH position be treated as a Zero, and actually give value to what he did in the field?

        Vote -1 Vote +1

        • tangotiger says:

          That would be suggesting that Ortiz and Frank Thomas would INCREASE their defensive value by playing LESS on the field (i.e, moving from 1B to DH).

          Vote -1 Vote +1

        • pft says:

          Completely agree, DH should be 0. MLB created the position of DH, a player should not be penalized because his manager uses him there.

          Vote -1 Vote +1

  27. GBSimons says:

    It’s amazing to see that the best player in the league – on either Off. or Def. – scores basically twice as well as the 10th-best player.

    Best of the best, indeed.

    Vote -1 Vote +1

  28. pft says:

    Have to say I like this, it addresses a number of problems. Thanks.

    Vote -1 Vote +1

  29. Cory says:

    David Ortiz played 26 innings in the field (1B at that) and has a -13.6 defense. He’s bad, but he could rack up triple digits in negative fielding.

    Vote -1 Vote +1

  30. coninefan says:

    While I like the stat, I still want to see the old fielding number on the player’s page, too, because I don’t know the positional adjustment numbers off the top of my head and certainly can’t scale them to playing time or whatever I’d need.

    When I go to see someone, I want to see how he ranks v. guys of the same position, not just an overall “defense” stat.

    Vote -1 Vote +1

  31. Jay says:

    So Goldschmidt goes from good defender to bad defender just like that?

    Vote -1 Vote +1

  32. chuckb says:

    Wow! This is awesome!

    Carlos Beltran has been worth MINUS 22.7 runs on defense. YIKES!

    Vote -1 Vote +1

  33. Bob M says:

    “Pretty much every shortstop is a better defender than every first baseman, even the ones with negative ratings at SS compared to the guys with positive ratings at 1B.”
    ———————-

    I think this year’s Brewers have proven this assumption to be false. They took several poor fielding SS (Alex Gonzalez, Yuniesky Betancourt), played him them 1B, and ended up with poor fielding 1B.

    Really, a poor fielder is a poor fielder. Just because you happen to stand in a particular spot on the field of play doesn’t automatically make you better than someone who happens to stand at a different spot.

    Vote -1 Vote +1

  34. pft says:

    So I am assuming the offense number has park adjustment, and no other adjustment (ie-average hitter is average hitter regardless of position played, same value for SS and 1Bman.)

    Vote -1 Vote +1

  35. Matt says:

    Da faq? It seems like I’m seeing non stop articles about Andrelton Simmons’ defense but Machado has been more valuable, according to this? Didn’t realize Machado was so good.

    Vote -1 Vote +1

  36. gabriel syme says:

    Late in the game, but I’d really urge that UZR be put back onto the dashboard (or put DRS as well). The def. number is incredibly opaque (yes, I know what it is), and it doesn’t answer the question the vast majority of readers want answered: has the player been a good or bad defender at his position? Most people will know Andrelton Simmons is a SS or Joey Votto is a 1B, and they are not usually interested to know what the positional adjustments are, but instead want to see how they’ve performed relative to their fellows.

    If you want all the components of WAR to be available on the dashboard, keep BsR and add wRAA, positional value and UZR. And leave off and def down in the Value box.

    Vote -1 Vote +1

  37. emilio says:

    In the final section (“Value”) Bambino leaf columns appear whose final sum (row) leads to the RAR, ie:
    Offense (1347.3), Defense (-18.6), League (47.8) and Replacement (325.2). This result (1701.8) agrees with the figure reported in col. RAR (penultimate right). Now, based on the theory, 1701.8 / 10 should give us the final figure of Babe Ruth WAR, but in the last column, we published Fangraphs WAR of 168.4.

    This discrepancy is even more noticeable in other tabs I’ve reviewed. For Cap Anson, the RAR reaches 1086.2, while 91.2 WAR is obviously far from 1086.2 / 10.

    Vote -1 Vote +1

  38. emilio says:

    clear to me that this site has the offensive in its two components (Batting + – Base Running) and Defensive already set (Fielding + – Positional). Although not shown separately, also includes a correction factor for the park, but do not understand the contents of the “League” column. Please let me illustrate on this adjustment.

    Another issue of concern and that has nothing to do with sabermetric calculations. When we refer to the list of Fangraphs WAR highest achieved (position players) we see that Barry Bonds appears in a second historic place with 164.1. But what really surprised me is a rather elementary fact: Baseball and Baseball-reference-prospectus give an official number of games (G) for life in 2986 while both Fangraphs publishes 2976. Hello I think I have found even greater differences in other cases: Fangraphs “games” Rickey Henderson 3068,
    Reference and Prospectus 3081.

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>