The Thing Adam Dunn Was Surprisingly Great At
This morning, David Appelman announced some minor upgrades to the way we calculate WAR, including the addition of a factor for doubles plays grounded into. While the results aren’t dramatic, some players do tend to hit into more double plays than others, and those rally-killing ground balls do have a tangible harm on a team’s offense, so they should be reflected in a metric attempting to assign a single value to a player’s performance. The swing isn’t huge, but the best player at double play avoidance in 2014 — Mike Trout, because of course he was — created an additional three runs above the league average, while the worst player — Casey McGehee — took five runs off the board by hitting into 31 double plays.
These names at the extremes make plenty of logical sense, as Trout as one of the fastest players in baseball, while McGehee is a slow ground ball machine who makes a ton of contact. And if you look at the leaderboard for valued added by double play avoidance since 2002, you’ll find names that make a lot of sense. At the top, there’s Ichiro Suzuki, adding 23 runs above the league average by almost never hitting into double plays. Right behind him is Johnny Damon, another speedy left-handed batter who made his living off his legs.
Now, if I worked for BuzzFeed, I’d have some cliffhanger sentence here, something like this.
“You’ll NEVER BELIEVE who is ALSO at the top of this list!”
But since I already put his name in the headline, you’ve probably already figured out that I’m talking about Adam Dunn. That’s right; when it comes to the best double play avoiders during the years in which we have batted ball data, Adam Dunn has created as much value as any other hitter in baseball.
On the one hand, this seems kind of shocking, because Adam Dunn is not exactly anyone’s idea of a speedster, and we don’t exactly have a lot of embedded memories of him beating out a bang-bang play at first base. On the other hand, though, Dunn is a pretty great reminder that there’s more to double play avoidance than just being fast. For reference, here are the wGDP leaders from 2002 to 2014, plotted against their Speed Score.
There is a correlation between speed score and double play avoidance, but it’s clearly not the only variable impacting things here. Dunn stands out as a bit of an outlier, but you’ll notice there are five other dots on that chart with similar or even lower speed scores than Dunn; Carlos Pena (+16 wGDP), Jim Thome (+11), Jason Giambi (+11), Russell Branyan (+11), and Dan Uggla (+10). The leaderboard is mostly comprised of guys who are too fast to double up, but there are this handful of slow sluggers who also managed to avoid GIDPs. What do they have in common?
Well, besides Uggla, they’re all left-handed, which can’t hurt. If you’re going to be slow, better to start closer to first base so you don’t have to cover as much ground when trying to get down the line than having to start from the opposite batters box. But there are a lot of left-handed batters in baseball, and most of them are faster than guys like Dunn, Giambi, Pena, and Thome, so what else do these guys have in common that would help them add value by avoiding double plays?
You’ve probably guessed the answer by now; strikeouts. Among players with at least 2,500 plate appearances since 2002, no player struck out at a higher rate than Branyan’s 32%. Dunn ranks 7th on that list, Pena 11th, and Thome 17th. These guys are some of the biggest whiffers of the past decade, making a contact-for-power trade-off that allowed them to hit a bunch of home runs and avoid hitting into a bunch of double plays. After all, it’s hard to hit into a double play if you don’t hit the ball in the first place.
This is one of the reasons why statistical types have long argued that strikeouts aren’t demonstrably worse than other types of outs. While they have no value in runner advancement, they also rarely create multiple outs on one play, offsetting some of the negatives associated with swinging and missing. There are absolutely some scenarios where you’d take a flyout over a strikeout, but there are also other scenarios where you’d prefer the whiff rather than a groundball right at an infielder. While there’s a lot of emphasis on the failures of players to put the ball in play when they just need to move a runner over, there isn’t as much of a stigma about adopting an approach that leads to more rally-killing double plays; both skillsets have their pluses and minuses, however.
This is one of the positives of having a high strikeout guy in your line-up. Of the hitters with the 20 highest strikeout rates since 2002, every single one has a positive wGDP. This is basically a collection of lumbering sluggers, as these guys are only in the majors because they can hit a baseball a long way, but their uppercut swings that produce so many Ks also tend to help them not hit many balls on the ground in double play situations. It’s a little thing, but over a long career, you’re looking at as much as a couple of extra wins from this little thing.
Hitting into double plays isn’t a large enough factor that it should change your opinion about a player’s value, but it does matter. While Ichiro and Adam Dunn got there in very different ways, both of them added a little over two wins of value by not hitting into double plays. And I’d bet that this is probably the only category in baseball that Dunn and Ichiro would ever find themselves co-leading.
Dave is the Managing Editor of FanGraphs.

What this bring to mind: the Cespedes Family Barbecue guys regularly do drafts of baseball players for non-baseball things, and eventually one of them will always pick Mike Trout “because he’s good at everything”.
Never change, Mike Trout.
#BASEBALLGODMIKETROUT or something like it needs to happen.
The trouble with the phrase “Baseball God Mike Trout” is that it’s redundant.
Now I’d like to see Mike Trout attempt killer facial hair, to see if he’d also be league-best in that.
And maybe become an impeccably-dressed Dwyane-Wade-style fashion icon.
Given the point of this article, one has to assume that besides his speed, Trout’s 180+ Ks last year are another big reason he led the league in avoiding GDP. Remember, he’s a RHB.
I saw this and honestly thought it was a hiccup in the algorithm that would get fixed later this week…Adam Dunn is weird.
Nah, it makes perfect sense. Hard to GIDP when you strike out.
In other news, DH’s rarely accrue defensive errors or negative UZR. And, if I played professional baseball, I’d never get thrown out on the basebaths (because I’d never get on the basepaths).
not to mention that when he made contact it was usually in the air. so, if you combine whiffing a lot with walking a lot and hitting fly balls a lot you end up with a low GIDP. makes sense to me
Three True Outcomes: strikeout, walk, HR. Dunn was the poster child for TTO. This is not at all surprising.
I thought this was gonna be HBP (which iirc, Dunn was also pretty good at). Alternatively, I thought it might be some random factoid, like he’s really good at cunnilingus.
On a more relevant note: I was surprised to learn a while ago that Albert Pujols has led baseball in GDP multiple times, and that he is also the active career leader in that category. I think I thought of him as such a great hitter that it seemed hard to believe that he could be so high on such a negative category. However, once I thought about it, it makes perfect sense. He has all the features that lead to GDP:
-Plays on a good offensive team and bats in the middle of the order (IOW, has lots of runners on ahead of him)
-Rarely strikes out (IOW puts a lot of balls in play)
-Has an average batted ball profile, not an extreme fly-ball one
-Hits the ball hard
-Slow
By baseball reference, he’s lost 26 runs compared to the average player from this, so he’s already close to -3 WAR from double plays, assuming a replacement players isn’t particularly worse than an average one in this regard. But, then again, over that same period he was worth 683 runs above average on offense, so I’d say the skillset is worth the slight penalty in the end.
The all-time leader in GIDP wasn’t such a bad hitter either.
Unfortunately wGDP numbers aren’t available that far back. Presumably, most of what caused Ripken to top the list was his extremely high number of career plate appearances. But it’s still my favorite surprising career-leader stat.
It also helps that Trout, who is so often the runner on first when Albert bats, didn’t steal many bases last year. Given Pujols’ GDB, you’d think Trout would run more. That he didn’t last year only fed the rumors that Pujols didn’t want him to go because of the distraction.
I take it that the value of a walk not being a double play is already part of the value of a walk and so is not part of wGDP, i.e., that wGDP looks at outs only, and marks down the relatively better and worse ones.
I don’t think it even has to do that. If you’re looking at a player’s offensive value based on linear weights, all you’re really doing is adding up the expected run value of each thing the player does. So you add in double plays just like you add in strikeouts, singles, walks, etc. At the end of it you get a total number of runs added/lost and compare that to average or replacement level.
Bat avoids ball. Ball fails to ground to SS or 2B.
Pretty open-and-shut case.
Also, not looking forward to what this does to Paul Konerko, AKA “GIDPK.
“
“While they have no value in runner advancement, they also rarely create multiple outs on one play, offsetting some of the negatives associated with swinging and missing. ”
Is the “value in runner advancement” of batted ball outs factored into WAR like the value of “rarely creat[ing] multiple outs” is now? If not, what was the thinking behind that choice?
Yeah, it seems really weird to include GIDP adjustments if runner advancement isn’t also included. Honestly, if you’re going to include GIDP adjustments, it’s not that crazy to include RBI adjustments above-and-beyond mere runner advancement, since both RBI and GIDP permanently take a runner off the basepath.
Great point.
Agreed. I can appreciate that a high gb% hitter who is slow-footed should be uniquely penalized in WAR, because they are more inclined to GIDP than other profiles. But retrospective GIDP itself is awfully team-dependent. I guess we don’t yet have all the details of wGDP, so perhaps it will be cleared up then.
Nice Buzzfeed slam!
Yeah, but only after the purely BuzzFeedian headline.
Not judging, but I did think the headline was particularly teasy for FanGraphs (instead of teasing the ‘who’ it teases the ‘what’). I don’t have a problem with it, but it’s funny that this is the article with the Buzzfeed jab.
There were two potential headline choices for this article, which followed the ‘buzzfeed clickbait’ pattern:
1. The surprising player who was great at avoiding double plays
2. The thing Adam Dunn was surprisingly great at
Both were plays on the fact people don’t expect Adam Dunn to be good at avoiding DP’s, and created the tease by giving you one piece of information and withholding the other. This article basically used one as it’s actual headline and made fun of the other. It also revealed the surprising link in the first sentence of the article.
I don’t care about this style of headline. It bothers me when an article is trying to trick me into clicking on an article I wouldn’t want to read otherwise, but fangraphs isn’t likely to have that problem for me. I just thought it was funny that it was kind of a “suspense fail.” So now it’s become a comforting assurance that fangraphs is not poaching editors from Buzzfeed.
What Bip said. What’s most obnoxious is the “you won’t believe what happens next!!!” headlines and then it turns out from reading the article that I can totally believe what happened next.
In the case of this article, I was legitimately surprised that Dunn was good at GIDP avoidance. Even better, I assumed it was from an extreme fly-ball profile, but it turned out that it was also his large number of strikeouts.
This seems an uncharacteristically team-dependent stat for FanGraphs WAR. Will RBI be added in soon?
RBI’s are already factored in, in a sense. The run value of a single, double, or HR used to calculate a player’s wOBA/RC are based on averages for those events, which include cases where runners are on base. A HR is worth more than 1 run, for example, because the actual value is based on the expected number of runners on base.
So, one the one hand, you’re right that factoring actual double-plays is probably too context-specific. However, DP’s happen, and they affect player value, so it should be factored it.
Ideally, what we would do is calculate a player’s expected number of DP’s in a context-neutral way, which considers things like GB/PA and speed, and use that. However, I’m not sure how reliably we can do that. I don’t know if we have reliable speed data for players, or if someone has done the work to make a good DP rate predictor.
Agreed.
Perhaps the correct way to calculate the negative run value of outs is to apply linear weights to out events just like the positive outcomes in wOBA. In this way ground ball outs would be weighted more costly than other types of outs relative to their league-wide rate of producing extra outs on the bases.
In any case, writing a post about context neutral player value that includes double plays is a bit out of place for FanGraphs. Also, naming a bunch of players who ground into the fewest double plays, and then failing to mention that many of them are leadoff hitters that have zero chance of a double play in 1/4 of their at bats, and a low chance in the others, since they are following the part of their lineup that is least likely to get on base is just odd.
I think the issue is that is only part of the equation. A grounder hit by Billy Butler is more likely to be a double play than one hit be Dee Gordon. So it also wouldn’t make sense to give them the same run value for every ground ball they hit. At the very least, we would want to have some way of accounting for speed, and the likely direction of the ball.
Ultimately, we would prefer to not even use singles and doubles to calculate wOBA, we’d just use batted ball trajectory information, along with speed and baserunning to measure how good a hitter someone is without involving defense, but I think we’re not there (plus we don’t have all the data.)
“Ultimately, we would prefer to not even use singles and doubles to calculate wOBA, we’d just use batted ball trajectory information, along with speed and baserunning to measure how good a hitter someone is without involving defense, but I think we’re not there (plus we don’t have all the data.)”
BIP, glad you posted that. I’ve been thinking along those lines for a while.
Yes, as Cameron points out, the reason the negative value of a K is not much worse than that of a BIP out is because statistically, the runs created by advancing or driving in men on base are almost cancelled out by the runs lost by DP. If you’re going to give players more or less runs according to how many DP they avoided or hit into vs. average, you should also give them more or less runs according to how often they advanced/drove in runners on base by GB or FB. It seems to me that the two go hand in hand.
It’s true that some players avoid DP by their speed, and so you could argue that they deserve extra runs that otherwise wouldn’t be taken into account, but that’s not the case for players like Dunn. If he’s avoiding DP by not putting the ball into play, he’s also not advancing/driving in an average number of runners by putting the BIP.
He plays a pretty mean center field too.
I love Fangraphs’ inside jokes. #notsarcasm
Well, besides Uggla, they’re all left-handed, which can’t hurt. If you’re going to be slow, better to start closer to first base so you don’t have to cover as much ground when trying to get down the line than having to start from the opposite batters box.
The main value of being a left handed batter in avoiding double plays is not a faster time to first but that the first baseman has to hold the runner on first, letting more ground balls reach the outfield.
Wait, first basemen don’t hold runners when a righty is batting?
I think he meant that lefties are more likely to hit grounders to the right side, thus having a 1B holding a runner on will affect lefties more than it affects righties.
I assume you mean left-handed pull hitters have an advantage due to the hole caused by the first baseman holding the runner on first and the second baseman is playing at double play depth.
Of course, the problem with your hypothesis is that ground balls hit towards the hole, even if they are fielded, are more difficult to turn into a double play. And if the first baseman is too far off the base and fields the front end of a double play, then the pitcher has to cover first. The 3-6-1 double play is just about the most difficult play in baseball to complete.
3-2-3?
Stephen Drew #10 on the list and J.D. Drew #18. Neither fits the TTO nor blazing speed type.
Stephen Drew has a pretty far below average ground ball rate. That might not explain all of it, but it’s a factor.
Dunn also appears to be fairly bad at hitting sacrifice flies, only 40 in 14 years. (worse than Larry Bowa) Maybe all that striking out hurts his team more than not hitting into DP’s?
A lot of would-be SF’s probably became HR’s.
Surprised you didn’t mention flyball rates too. Dunn and company also put the ball on the ground less often when they do put the ball in play.
Implied here: “This is basically a collection of lumbering sluggers, as these guys are only in the majors because they can hit a baseball a long way, but their uppercut swings that produce so many Ks also tend to help them not hit many balls on the ground in double play situations.”
I was hoping for Adam Dunn – the pitcher
Great! McGehee will look exactly like Pablo Sandoval 2011 at the plate. Or AJ Pierszinski the year he was with the G.
Of course Ichiro never hit into double plays. He spent the majority of his career batting leadoff on terrible hitting teams. Hitting into double plays when you rarely ever bat with runners on base is difficult even for the magical Ichiro.
I’m sure that is factored in, and the lack of opportunities would suppress the positive value of his ability to avoid double plays.
Do you know the guy doesn’t really like baseball that much? Do you know the guy doesn’t have a passion to play the game that much?
“The swing isn’t huge, but the best player at double play avoidance in 2014 — Mike Trout, because of course he was”
This made me laugh with incredulity. One paragraph in, and I’ve already gotten my money’s worth (even if money were actually involved).
31 double plays is only 5 runs?
Seems low.
The wGDP stat is based upon double plays above or below average. I’m assuming that “average” means the number of DPs that the batter being measured would be expected to hit into given the game situations when he bats (i.e. runner on 1st with 0 or 1 outs would have a higher expected GDP rate than other situations).
Baseball Reference has an analogous GDP adjustment in their WAR calculation and McGehee rated 4 runs below average in 2014.
I think this is all interesting, but I don’t think there’s any correlation to GOOD baseball player here. This doesn’t change the fact that absolutely no manager on earth would ever want Adam Dunn batting with the game on the line and a runner on first. I know that’s not your point and you are just passing along an interesting statistic. But does it have value? 29 other GMs for the last 4 years think otherwise.
I bet the 2011-2014 White Sox Front Office wishes they had this stat while on the phone with other GMs: “Hey he may make 200 worthless outs but at least he makes a handful less double outs than most of the league! Just give us 2 of your top prospects for him.”
What are you talking about? Adam Dunn, career: runner on first, less than 2 outs:
1316 ABs, 103 HRs, .271 BA, .382 OBP, .561 SLG (.943 OPS)
And check his clutch stats too, if you care about such things.
http://www.baseball-reference.com/players/split.cgi?id=dunnad01&year=Career&t=b
I think fangraphs should have an annual Adam Dunn week around this time.
What about defenses putting on a shift? Obviously teams have much less extreme shifts with a runner on first than with no runners on, but is it possible that Dunn also hit into fewer double plays because the middle infielders were positioned far enough from second base with Dunn and the others batting to prevent them from turning two?
Dunn has only one season of more than 10 GIDP. He grounded into 100 over his 14 year career. Billy Butler’s four worst GIDP seasons add up to 104 GIDP.
According to Baseball Reference, Adam Dunn has gained +1 run from GDP avoiding in 2002-2014.
21.5 less than Fangraphs’ number. Why?
B-R calculates GDP rate as follows. GDP divided by GDP opportunities, which are “any infield ground ball with a runner on first, less than two outs and at least one out is recorded on the play”.
That is, B-R excludes hits, walks, flyouts and strikeouts from denominator. I guess it is the main difference between B-R and FG.
B-R considers GDP avoiding as not batting, but running skill. Usually baseball fans don’t think, that strikeouts/popups with runner on base are GDP “avoiding”. If a batter hits a grounder and he runs to first base as fast as avoids a GDP, then he deserves praise.
I think B-R’s method is more reasonable than Fangraphs’.