The Astros and Grounding Into Double Plays

By now, almost everyone who regularly follows the Astros is frustrated by the announcer saying, "and he grounds into a double play."  At times over the season, we have talked about the Astros' GIDP proclivity as one of the reasons that the offense can't seem to score as many runs as it should, based on battting average and  on base percentage.  I set out analyze the team's GIDP tendency.

The Astros continue to lead the NL in GIDP, with 133, which outdistances the Mets (129) and Dodgers (128).  In fact, the Astros lead the majors in GIDP; the Twins (128) and Yankees (122) lead the AL.  If we are looking for a cause, maybe all we need to know is the groundball to flyball rate: the top three NL teams in GB/Fly are Houston, New York Mets, and LA Dodgers.  That ranking is exactly the same as the GIDP ranking.

Miguel Tejada currently leads the NL in GIDP (27), but he is in a tight "contest" with Yadier Molina (26) and Kevin Kousmanoff (25).  Evan Longoria leads the AL with 26 GIDP.  Hunter Pence is No. 6 on the leaderboard with 20 GIDP.  Carlos Lee is No. 11 with 18 GIDP.  However, it also worth noting that Tejada leads all Astros in plate appearances (589) and he has the most plate appearances among his "competitors" in the GIDP rankings.  Albert Pujols and Adrian Gonzales are the only hitters in the top 5 for GIDP who have as many plate appearances as Tejada.  In addition to raw plate appearances, "opportunities" for a GIDP are based on  runners on first base with less than 2 outs when the hitter bats.

Bill James' web site calculates a GIDP rate based on opportunities for a DP, and I separated the GIDP for 2009 to show how Astros' hitters rank :

2009 GIDP Rate Per Opportunity

Quintero 47%

Tejada 24%

Pence 22%

Pudge 19%

Keppinger 16%

Lee 15%

Berkman 12%

Blum 9%

Matsui 4%

Bourn 1%

As with the team rankings, GB/Fly ratio has an impact on GIDP tendency.  Pence, Molina, and Tejada are 11, 12, and 13 in GB/Fly ratio in the NL.  Most of the hitters ranked higher are speedy lead off type batters (like Bourn and the Mets' Luis Castillo). Also, keep in mind that the GIDP rate/opportunity doesn't take into account the speed of the batter hitting in front of each player---which I would argue may affect how poorly a player ranks on this measure.

The current Astros' team is on almost exactly the same GIDP pace as the 2000 Astros team, which holds the team record with 154 GIDPs.   If it's any consolation, the 2009 Astros are turning DPs on the defensive side at a higher rate than they are grounding into them.  The 2009 Astros have turned 141 DPs, which puts the team on pace to be the 6th best Astros team at turning the DP. The Astros' are 2d in the NL in defensive DPs.

The 2000 team had 154 GIDPs, with the following players leading this dubious stat: Alou (21), Bagwell (19),Bogar (15), Hidalgo (13), and Biggio (10). The 1998 team is second in GIDP (146) with the following team leaders: Gutierrez (21), Ausmus (18), Bell (14), Alou (14), Bagwell (14), C. Everett (14), and Biggio (10).

My next step is to examine the relationships between standard team batting stats and the GIDPs among all of the Astros' teams.  This turned out to be a more time consuming task than I expected, because the GIDPs were not easily imported from the B-Ref team page (only DPs, as in "DPs turned" are shown), which meant I had to go to each Astros' team season on B-Ref and manually enter the data.  The results of my correllation investigation is after the jump.

Based on the population of Astros' (and Colt 45s) teams since inception, I investigated possible correllations between DP/G and offensive stats.  The only reasonably strong correllation was between HR/Game and DP/Game, with a correllation coefficient of .52 (R-sq. .27).  I'm not sure this reflects causation though.  Teams in the 1995 or later era seem to have a higher GIDP tendency.  That era also coincides with an increase in HRs, overall, in baseball.  The 12 lowest GIDP rates by Astros' teams were all pre-1993 teams. 

Beyond the HR correllation, the moderate, weak, and non-existent correllations are shown below.

Moderate Correlation

Slugging (.44)

R/G (.38)

K/G (.36)

OBP (.38)

Batting Avg. (.34)

SB/G (-.33)

Singles/Hits (-.33)

Weak or Non existent Correlation

Winning % (-.03)

Hits to BB ratio (-.06)

I think many of the stats which are moderately correlated with GIDP, like OBP, batting average, and R/G, simply reflect the fact that teams which have a lot of traffic on the base paths will have more opportunities for the GIDP.  We expect team speed to counter the GIDP tendency, and using stolen bases as an indicator of team speed, SB/G is at least correlated in the right direction (inverse), although it's only a moderate relationship.  K/G and Slugging/G are correlated with GIDP, but I think this falls within the discussion of HRs, above, since both stats are normally correlated with home runs.  Surprisingly, singles as a ratio of hits are inversely correlated with GIDP.  I expected the opposite.  I tried to use hits/BB as an indicator of plate discipline, but no correlation was shown.  GIDP apparently had no effect on Astros' teams' winning percentage.

OK, this may not show much, other than the need to perform more sophisticated analyses, which I'm not inclined to do at the moment.  The most significant point is that the Astros teams which scored the most runs, slugged the most, had the highest OBP, and the highest batting averages tended to bring the most GIDP baggage along the way.  Although there is some scoring era impact which distorts the simple correlations, the corelation between time periods and GIDP is relatively weak.

  Despite the results, above, the three leading NL teams in GIDP in 2009 (Astros, Dodgers, and Mets) are among the bottom teams for rates of extra base hits, home runs, and extra base hits as a percentage of hits.  Those three teams are also among the league leaders in attempting sacrifice bunts.  This may indicate that the manager makes a sub-optimal move in order to avoid DPs.  The Astros lead the NL in percentage of DP opportunities which turn into GIDP (14%).  The Astros are followed by the Padres, Pirates, and Mets---all bad teams.   The Dodgers are league average in GIDP rate per opportunity, which means that their high ranking in number of GIDP is partly an artifact of lots of baserunners (the Dodgers lead the league in baserunners, and have an above average team OBP).

I also think--but can't prove--that lineup structure can be used to reduce GIDPs.   In another thread, Timmy and I had a discussion as to whether hitting Tejada behind Lee increases Tejada's chances of a GIDP.  I think so, but the data isn't really available to test that question.  In particular, the "GIDP rate per opportunity" would be more useful if we knew that rate based on lineup position and the hitter in front of the batter for each opportunity.  However, the problem with filling out the Astros' lineup card is that so many Astros' hitters have high GIDP rates.  It's hard to space them out.