Can platoon splits help us understand whether the Astros' offense is likely to improve in 2014? Just the question, itself, has a sabermetrics feel to it.
David recently wrote about the Astros supposedly inquiring about other teams' first basemen, including the Red Sox's lefthanded batter, Mike Carp. My immediate reaction was that Carp makes some sense because of the potential to platoon with Jesus Guzman, as well as the team's need for more effective LHBs. But this article isn't about Carp, in particular. That recent bit of news is a good jumping off point for examining whether the Astros need another lefthanded bat.
For simplicity, I will use the all-encompassing offensive stat, wRC+. As a reminder, a wRC+ of 100 is league average, and results above or below 100 represent the percentage above or below average offense. (For example, a wRC+ of 90 means that the team is 10% below average, and a wRC+ of 110 means that the team is 10% above average.)
Team Platoon Splits
The Astros had a bad offense in 2013. Based on overall wRC+, the Astros' 86 wRC+ gives the Astros a ranking of 27th out of 30 teams. But the Astros were better hitters against LHPs than RHPs. Against RHPs, the Astros ranking drops to 28th (84 wRC+). Against LHPs, the Astros ranking is almost respectable, 17th (wRC+ of 91).
If the Astros want to improve their offense against RHPs, basically two options are available. The Astros can acquire RHBs who are better overall than existing players and exhibit more balanced platoon splits. Or the Astros can acquire additional LHBs, who are expected to have better splits against RHPs. Well, there is another option: get rid of players who have bad splits against RHPs.
A sabermetric article has to have a table or chart. So here we go. The table below shows teams' wRC+ in 2013, as well as the splits vs. LHP and RHP, and the ratio of the splits.
|Team||Runs||wRC+||vs. LH||vs.RH||Ratio L/R|
(team stats from Fangraphs.com)
The Astros were among 11 teams which had a platoon split ratio favoring their offense against LHPs. The Astros and Rangers had similar platoon ratios (108% and 109%). The Rangers added substantial LHB offense over the winter (Choo and Fielder). It's worth noting that some of the best offensive teams in the majors (Red Sox, Dodgers,Cardinals, Tigers) had platoon split ratios which favor their offense against RHPs.
Let's look at some statistics we can generate from this table. Naturally, wRC+ is highly correlated with team runs (r=.856). wRC+ versus RHPs is also highly correlated with team runs (r=0.854), but wRC+ versus LHPs is substantially less correlated with runs--though a correlation still exists (r=0.52). Turning to R-squared, wRC+ against RHP "explains" 73% of a team's runs.. wRC+ versus LHP "explains" 27% of teams' runs. The wRC+ against RHP has virtually the same explanatory power for runs as overall wRC+ (R-squared of 0.730 and 0.734).
It's not unexpected that offense against RHPs is more highly correlated with scoring runs than offense against LHPs. There are more RHPs than LHPs. Last year, 57% of plate appearances came against RHPs, with 43% against LHPs. For whatever reason, the relative ability of platoon split offensive measures to predict team runs is tilted more toward offense against RHPs than even the raw plate appearances might suggest.
My takeaway is that improving the Astros offense against RHPs is likely to produce the most significant impact on run scoring by the team. This conclusion is also consistent with something else we know: baseball teams place a premium, in general, on drafting and acquiring effective lefthanded bats.
CHANGES IN ASTROS' COMPOSITION
Did the Astros' off-season lead to much improvement in the team's weak offense versus RHP? Let's compare the platoon splits for player additions and removals. For most positions on the team, the same players will hit and presumably produce similar splits to last season. The outfield and 1b provide the most significant potential for performance change due to addition/removal of players. Although the middle infield positions may be affected by roster changes, the infield back up positions are too unclear to project an impact on platoon splits. For purposes of this comparison, I have considered the incremental increase in Grossman's and Hoe's playing time as an "addition."
I have used departing players actual platoon splits in 2013 to develop a weighted average wRC+ (vs. L and R) for the hitters who are removed. (Also, I consider J.D. Martinez as a "departing" player, since he was removed from the 40 man roster.)
Dexter Fowler is the most noteworthy addition to the offense over the winter. Fowler is not a LHB, but he is a switch hitter. Fowler's wRC+ is based on his career wRC+. Barnes, Maxwell, and Crowe are the departing players who comprise the weighted wRC+.
Players Removed: 390 PA with a wRC+ of 56.
Player Added, Dexter Fowler: wRC+ of 100.
Players Removed: 212 PA with a wRC+ of 123.
Player Added, Dexter Fowler: wRC+ of 113.
The Astros' center fielders last year were horrendous against RHP, which contributed to the Astros poor hitting against righties. However, the center fielders were quite good against LHP. But it's not a good trade off. Fowler's switch hitting provides more R/L balance in CF. Fowler hits best against LHP, but his platoon split vs. lefties may produce a modest decrease in wRC+ compared to last season. However, he should significantly improve the offense against RHP. The net impact should improve the overall offense. Based on the numbers, above, Fowler would produce a 25 point increase in wRC+ out of CF.
The outgoing corner outfielders include a slice of playing time for each of the departing center fielders (Barnes, Maxwell, Crowe) as well as J.D. Martinez, Jimmy Paredes, Fernando Martinez and Rick Ankiel. Their plate appearances are assumed to be available for additional playing time for Grossman and Hoes, as well as a partial season of George Springer. The additional PAs are distributed 44% to Springer, 22% to Hoes, and 33% to Grossman. For Springer, I utilized the Steamer/ZIPS projection of a wRC+ of 110. After reviewing, Springer's career splits in the minors, I set his wRC+ at 113 vs. LHP and 99 vs. RHP.
Players Removed: 710 PA with a wRC+ of 67.
Players Added (incremental and Springer): wRC+ of 95.
Players Removed: 200 PA with a wRC+ of 56.
Players Added (incremental and Springer): wRC+ of 109.
Giving Springer around 400 or so plate appearances and utilizing Grossman and Hoes more than last year should significantly improve the corner outfield's performance against both left and right handed pitchers. Both Springer and Grossman are expected to hit better against LHPs. Hoes is quite balanced, with a wRC+ 96 / 95 platoon split. Combining the platoon improvements results in a 36 point increase in wRC+ from corner outfielders.
Carlos Pena, Brandon Laird, and Brett Wallace are the departing players at first base. We really don't know who will be the first baseman in 2014. For practical reasons, I have assumed Jesus Guzman as the full time first baseman, using his last three seasons' wRC+ (3-2-1 weighting). Though it is certainly possible that Krauss will get platoon duty at 1b, his rookie season stats were so bad that the at bats even against RHP were worse than the platoon split for Guzman. Selecting Guzman as the full time first baseman is more feasible than speculating on whether Krauss will show enough improvement to warrant platoon duty at first base. For purposes of this analysis, I assumed Carter will be the DH, even though we know that the players at first base and DH will be interchangeable over the course of the season.
Players Removed: 480 PA with wRC+ of 95.
Player Added: Guzman with wRC+ of 87.
Players Removed: 165 PA with wRC+ of 80.
Player Added: Guzman RC+ of 127.
For the first time in this exercise, we find that the removal of players has decreased the wRC+ against right hand pitchers. This is due in large part to the loss of Brett Wallace, who hit well against RHP. Against LHP, however, Guzman is much better than the departing first basemen. Guzman hits well enough versus LHP that he still manages a net gain of +4 in wRC+ relative to departing first basemen. The first base offensive picture is less than inspiring. If you are wondering why the Houston front office is still picking through first base candidates on other teams, the weak expectation against RHP may be an explanation. If Singleton can make it to the big leagues this year, his lefthanded bat could improve the team's outlook against right handed pitchers. However, it's pure speculation to project Singleton in the majors that soon.
DO THE ASTROS NEED ANOTHER LH BAT?
I will say "yes."
Changes in the Astros' lineup this season should result in marginal improvement for the offense against RHPs. The addition/subtractions set out for CF, OF, and 1b affect approximately 35% of the plate appearances against RHPs. The changes in the outfield and first base should result in approximately a 7 point increase in team wRC+ versus righty pitchers. That level of improvement would result in a wRC+ of 91, putting the Astros offense against RHPs on par with the offense against LHPs. However, the Astros' offense would still be well below average. And the most immediate way to improve the offense is by bolstering the batting against RHPs. Furthermore, Guzman's platoon splits beg for an effective lefty platoon bat at first base.
Jason Castro has been a very effective LH bat, but he is like a cannon in a pop gun lineup against RHPs. Of course, finding another effective LHB is easier said than done.
As a final comment, as I examined the Astros' platoon split data, I gained a new found appreciation for Chris Carter and L.J. Hoes. Both right hand hitters exhibited good balance against pitchers throwing from both arm sides. Carter was particularly effective, posting a 117 wRC+ against LHPs and 112 against RHPs. This type of platoon balance is uncommon for a big swinging slugger like Carter. Given some of the extreme platoon splits experienced by Astros' batters in 2013, penciling Hoes and Carter into the lineup must be comforting for the manager.