Escape Artist or Rally Killer?

One of the first things we learn about baseball is that outs are bad. Why? Well, it limits your team’s chances of scoring more runs. By this logic, then, making two outs at once is substantially worse. Three outs? You’re a one-man wrecking crew. Triple plays are pretty rare (unless you’re Brooks Robinson) , whereas a double play is a relatively frequent occurrence. Nothing knocks the wind out of a rally better than erasing a runner from the basepaths and making an out at the same time. Some run estimators account for grounding into a double play while others don’t. Pete Palmer, for instance, excluded GIDP in his original Linear Weights equation from The Hidden Game of Baseball—stating that,

“(…) The grounded into double play is to a far greater extent a function of one’s place in the batting order than it is of poor speed or failure in the clutch, and thus it does not find a formula applicable to all batters.” (66)

Now, let’s think about this for a second. There’s some merit to this argument. A hitter in the leadoff slot is going to hit in less double play opportunities than, say, a hitter in the third slot in the order. This is true. But how the player performs in the situation is another thing entirely. Not all players perform the same when presented with a double play situation. For example, here are the players faced with the most double play opportunities in 2009:

Name	       Team	Opps   GIDP    GIDP%
Ryan Zimmerman	WAS	182	22	12.1%
Mark Teixeira	NYA	176	13	7.4%
Aaron Hill	TOR	174	17	9.8%
Ryan Braun	MIL	172	6	3.5%
Albert Pujols	STL	170	23	13.5%
Johnny Damon	NYA	160	9	5.6%
Adrian Gonzalez	SDP	154	23	14.9%
Troy Tulowitzki	COL	153	20	13.1%
Brandon Inge	DET	150	12	8.0%
Ryan Howard	PHI	149	11	7.4%

The discrepancy between the best (Ryan Braun, 3.5%) and the worst (Adrian Gonzalez, 14.9%) is an astonishing 11.4%. Mark Teixeira received only half a dozen less opportunities than Nationals third baseman Ryan Zimmerman, but he was quite better at avoiding the double play. This lends credence to the notion that some hitters do avoid the double play more than others, rather than it being a function of one’s slot in the lineup. We can change this into a plus/minus figure and convert it into runs pretty easily. Just look at a hitter’s GIDP opportunities and compare it to the league average. Subtract the player’s total GIDP from the expected double plays, and multiply by the run value*. It can be written as (Expected GIDP – GIDP) x run value.

And that’s all there is to it. If you’re familiar with Mitchel Lichtman’s Super Linear Weights or Sean Smith’s Wins Above Replacement, this should look pretty familiar to you. This tells you how many GIDP the player hit into above or below the league average, and we can identify which players were the escape artists or the escape hatch in the lineup with a runner on first.

Leaders in 2009:

Name	        Team   GIDP%    Runs
Ryan Braun	MIL	3.5%	+4.1
C. Granderson	DET	0.9%	+3.4
Chase Utley	PHI	3.7%	+3.4
Ichiro Suzuki	SEA	1.1%	+3.1
Johnny Damon	NYY	5.6%	+3.0
Stephen Drew	ARI	3.6%	+2.7
Carl Crawford	TBR	5.1%	+2.7
Michael Bourn	HOU	1.3%	+2.6
Luke Scott	BAL	3.8%	+2.5
Grady Sizemore	CLE	3.8%	+2.4

A group filled of players with respectable (if not blazing) speed, which is what I expected. I’m not so sure about Luke Scott, though. I don’t ever recall hearing about him having good speed, but I could very well be wrong. With the exception of Braun, all of the players in the top ten are left-handed.

The trailers:

Name	       Team	GIDP%  Runs
Edgar Renteria	SFG	22.1%	-2.8
Geovany Soto	CHC	21.8%	-3.1
Yunel Escobar	ATL	20.8%	-3.3
Jose Lopez	SEA	18.9%	-3.5
K. Kouzmanoff	SDP	21.0%	-3.9
Evan Longoria	TBR	19.9%	-4.0
Mike Lowell	BOS	23.1%	-4.1
Hunter Pence	HOU	23.4%	-4.3
Miguel Tejada	HOU	23.2%	-5.0
Yadier Molina	STL	27.0%	-5.2

No list that has any relationship to foot speed would be complete without a Molina brother near or at the bottom of it. In comparison, Curtis Granderson had 106 GIDP opportunities, six more than Molina. He hit into one double play—26 less than Yadier. An average hitter would be expected to hit into eleven, still less than half of his rate. That’s not good.

Unlike our leaders, our trailers are all right-handed hitters. In fact, if you look at the top 30 hitters in GIDP opportunities, 77% are lefties or switch-hitters. Conversely, 90% of the worst hitters in double play situations were right-handed hitters. Granted, this isn’t a rigorous analysis—but it seems as though lefties tend to avoid double plays better. This would validate common sense: we would expect a left-handed hitter to be more difficult to double up due to the fact that he’s closer to first base as he comes out of the batter’s box. And if you’re fast, you’re even better off.

And, of course, a look at the Major League teams in 2009:

Team	GIDP%	Runs
PHI	8.0%	+9.8
ARI	8.1%	+9.5
FLA	8.6%	+8.5
TBR	9.1%	+5.7
TEX	9.3%	+4.8
CIN	9.3%	+4.7
COL	9.6%	+3.9
MIL	10.2%	+2.0
DET	10.4%	+1.2
WAS	10.4%	+1.0
TOR	10.6%	+0.1
OAK	10.7%	+0.1
ANA	10.7%	+0.0
CHC	10.7%	-0.1
SFG	10.7%	-0.2
BOS	10.7%	-0.2
SEA	10.9%	-0.9
BAL	11.0%	-1.2
STL	11.0%	-1.3
LAD	11.0%	-1.3
NYA	11.1%	-1.8
CLE	11.1%	-1.8
PIT	11.3%	-2.3
MIN	11.5%	-3.6
SDP	11.8%	-4.1
ATL	11.7%	-4.2
CHW	11.8%	-4.3
KCR	12.0%	-4.7
NYM	12.3%	-6.3
HOU	14.4%	-12.9

The difference between the best team (PHI) and the worst (HOU) in 2009 was 23 runs, or a little over two wins. That’s not nearly as large as the gap in baserunning (about three and a half wins), but it’s certainly noticeable.

Glancing through Smith’s historical WAR, the best single season performances from 1955-2009 are at +6 runs with hitters Mike Caruso, Nellie Fox, Cristian Guzman, Vada Pinson, Harold Reynolds, and Ichiro Suzuki (two times). Jim Rice is at the very bottom with -7, followed by Paul Konerko, Rocky Colavito, and George Bell at -6. Unsurprisingly, the best seasons consist of left-handers and the worst with right-handers. All in all, we’re looking at a swing of + 5 runs for individual hitters, or about a full win at the extremes. Hitting into or avoiding the double play isn’t as important towards player value as, say, baserunning—but it’s certainly something that we should keep in mind.

Data for all players can be found here.

*The run expectancy with a man on first and no outs is .883, and with the bases empty and two outs .106 (hitting into the double play). If the batter instead made a batting out and avoided the double play altogether, you’d be left with a man on first and one out, or an RE of .533. The double play in this situation costs the difference between the two, or .533-.106 = .427 runs. Do the same thing for a man on first with one out, and you get an added value of .223 runs. Average them, and you have +.325 runs: the run value of a batting out over grounding into the double play.

The Incompleat Starting Pitcher
The end of the nine-inning start and how we got here.

Print This Post
Sort by:   newest | oldest | most voted
While I like this, I have to say I’m a little with Total. How much of this is random? These are VERY small numbers of occurrences. In addition, if there is a correlation between GIDP and HR rates in those situations, then having a higher GIDP rate might not be a bad thing. One solution would be to look at outcomes in greater detail. For those that hit ground balls but didn’t GIDP, that’s either (a) luck or (b) speed. For those that hit a fly ball or a liner, that may be a show of skill by making sure… Read more »

I seem to recall reading a similar article once, where the author found that lefties are generally better at avoiding the double play. His theory was that in a double play situation the first baseman is holding the runner, therefore leaving the batter with a bigger hole to hit into.


Good stuff JT. Two questions:

1) What were the weights that you used for the GIDP?

2) Echoing the sentiments above, it would be interesting to see these players broken down by batted ball type, as you’d expect that as players hit more FB/LD, they’d be able to avoid GIDP more on average.

Peter Jensen
Peter Jensen
JT –  It looks like you considered every time a batter had a plate appearance with a man on first and less than 2 outs as a double play opportunity.  I am not sure that doing that is better than limiting a double play opportunity to when the batter actually hit the ball. I would use the smaller number.  However, I am sure that if you use the more inclusive number, then you have to also include Ryan Braun’s 4 K+CS DP’s as part of his productivity.  The reason for this is because if he had actually hit the ball… Read more »
Matt Trokenheim
Matt Trokenheim

Echoing the above, my concerns are:

1. Sample size.
2. Accounting for FB%, K%, and BB%.
3. Focusing on the speed of the hitter would suggest that you should also account for the speed of the runner on 1st.
4. Focusing on the handedness of the hitter would suggest that you should also account for the handedness of the pitcher.

JT Jordan
JT Jordan
Total: wonderful question.  I haven’t looked too closely at multi-year data yet, although I do recall seeing certain players doing particularly well from year to year (Ichiro comes to mind) or poorly (Jim Rice).  In terms of the grand scheme of things, though, I don’t know how strong the correlation is. Michael, what weights are you referring to?  The run value?  If so, the RV in 2009 was +.325. Peter, that’s a great point.  Something more rigorous is definitely needed to really break it down.  The data source I’m using is B-Ref (it’s also on B-Pro), and sadly I don’t… Read more »
Josh Fisher
Josh Fisher

Really cool stuff, JT…do you (or does anyone) know if the speed difference between right-handers and left-handers has ever been studied? What I’m asking is how big of a 60-time advantage must a right-hander have to beat a lefty down the line?

JT Jordan
JT Jordan

That’s a great question and I wish I knew the answer!  I haven’t seen any studies done with actual 60 times, but it’d be very interesting to see just how big (or small) the difference is.


One of the things that is unaccounted for are the quality of the baserunners. If the runner on base is Carl Crawford or Ellsbury there will be fewer double plays than if one of the Molina brothers is onbase.

I have a feeling that where the ball is caught is a major factor. For instance the ss picks up a grounder throws to second and then first for the DP will be more dependent on baserunner speed then when somebody catches a linedrive and throws to first


Are the players who do well or badly consistent from year to year (ie is this more than random)?

JT Jordan
JT Jordan

You’re absolutely right- the “model” presented is quite raw and can certainly be improved upon with more information.  It’s not accounting for the speed of the batted ball or the location, nor does it account for the speed of the man on base.


I expected some discussion of FB% here although with the names on your first list (Adrian Gonzalez) it looks that coorelation may not be incredibly strong.

Nathaniel Dawson
Nathaniel Dawson

What Peter said, and I’d also add DP’s on fly balls as well. A hitter putting more balls into the air would tend to have fewer GIDP’s, but there would be a slight increase in FBDP’s to consider.

Peter Jensen
Peter Jensen

Nathaniel – I specifically didn’t include fly ball DPs because the majority of them occur when a batter is trying to advance a base after the ball is caught.  The batter shouldn’t be held accountable for this.  And in the rare times when a runner is doubled up returning to his base on a caught fly ball it is almost exclusively when he was running on the pitch.