Friday, December 10, 2010

Re-envisioning the 2011 Red Sox Lineup

Remember that post I wrote a month ago about the 2011 lineup? If you haven't read it already, you can probably ignore most of what was written after two absolutely massive moves in the last two weeks. If you didn't understand what Theo meant when he said 2010 was a "bridge year", now it has to be clear. After several years of many players moving in and out of roster spots, the core of the Red Sox team is locked up for a very long period of time. Jacoby Ellsbury and Kevin Youkilis are under contract through 2013, Pedroia through 2015, Crawford through 2017, and if the Adrian Gonzalez extension goes through he'll be under contract until after the 2018 season. On the pitching side, Buchholz, Beckett, and Lackey are under team control through 2014, as is Lester, although the Red Sox have an affordable option for the 2015 season. If you're a Red Sox fan who is a hot stove junkie, be ready to get disappointed over the next few years. The team is going to be very similar for a very long period of time.

Left field was a disaster for the Red Sox last year. While Daniel Nava, Darnell McDonald, and Ryan Kalish all did a lot more than the team expected, the offensive bar is set very high for left field. The Red Sox left fielders ranked 28th in batting average, 28th in OBP, and 18th in slugging. Carl Crawford is an obvious improvement offensively, and according to all the advanced metrics his defense is off the charts. Also, catcher was a weak point for the team last year. Current reports indicate that Russell Martin is close to signing with the team. While he is not the All Star he was in 2008 (thanks Joe Torre! Who knew that 449 games in 3 years would harm a young catcher?), Martin still posts an excellent OBP for a catcher (.347 last season, .360 for his career). For a lineup as deep as the 2011 Red Sox, a catcher who can work the count would be a nice addition.

But what does that mean for the 2011 season? With the additions of Adrian Gonzalez and Carl Crawford, just what does the Red Sox lineup look like now? Quite simply, it looks pretty damn good.

Average OBP Slugging OPS+ UZR/150
C Russell Martin 0.257 0.358 0.347 92 N/A
1B Adrian Gonzalez 0.286 0.390 0.524 152 1.4
2B Dustin Pedroia 0.300 0.371 0.478 120 9.6
3B Kevin Youkilis 0.308 0.406 0.560 150 6.9*
SS Marco Scutaro 0.275 0.350 0.387 96 -3.5*
LF Carl Crawford 0.298 0.349 0.457 117 20.6
CF Jacoby Ellsbury** 0.292 0.347 0.406 93 4.8
RF JD Drew 0.269 0.375 0.492 124 6.4
DH David Ortiz 0.258 0.357 0.501 122 N/A

*Kevin Youkilis' UZR/150 is his career at 3B, not a weighted mean. Marco Scutaro's UZR is his career at SS, as a very strong 2008 in a small sample size skewed his 2008-2010 weighted mean.
** Jacoby Ellsbury's statistics are a weighted mean from 2008 and 2009 (5, 4), and his defense is for all outfield positions.

Looking at the weighted averages (5, 4, 3) using 2008-2010 stats for the projected Red Sox lineup you can see how there are no weak spots in the lineup. This is a simple model, and I haven't adjusted numbers for park factors, aging, or similar corrections. While Ellsbury, Scutaro, and Martin have low slugging percentages, no expected starter has an OBP lower than .347. Just how good is this lineup? Baseball Musing's lineup analysis tool predicts that the ideal lineup would score 5.893 runs per game, or 954 for the entire season. While that doesn't quite challenge the 1931 Yankees' record of 1,067, it would have topped the 2010 Yankees' league leading total by almost 100 runs. The lineup tool only looks at OBP and slugging, so the speed of Ellsbury and Crawford doesn't contribute anything. It also typically ends up with some odd lineup results - the "ideal" lineup has Youkilis and Adrian Gonzalez hitting 1/2, followed by Marco Scutaro. Pedrioa and Crawford, who are likely to hit 1/2 according to Buster Olney.

Update December 14, 2010:
Looks like the reports of signing Russel Martin were premature. I suppose it serves me right for actually believing a baseball rumor before the ink is dry on a contract. Re-running the lineup analysis tool with Jason Varitek's weighted means (.259 OBP, .305 slugging) drops the runs scored per game from 5.893 to 5.718, or from 954 to 926. This is a substantial drop, about 3 wins using the rule of thumb of 10 runs are worth a win over the course of a season, although there is reason to believe that Saltalamacchia should beat those numbers. Using Saltalamacchia's career numbers (.315 OBP, .386 slugging) softens the offensive blow substantially, with the Red Sox projected to score 5.851 runs per game, a difference of only 6 runs over an entire season and would still have easily topped the AL in runs scored last year.

Not only would this lineup score a lot of runs, there are no major weaknesses defensively as well. Using a weighted mean of UZR/150 (available from Fangraphs, it is the number of runs the fielder would save (positive) or cost (negative) compared to an average fielder over 150 plays, which is about a season's worth of playing time) Marco Scutaro is the only fielder below average.

Where does Jed Lowrie fit in?
After a strong finish to the season, many folks are pushing for Jed Lowrie to be given the starting shortstop job, pushing Marco Scutaro to a super utility role. I agree with this, as Scutaro is a free agent after the 2011 season and in an admittedly small sample size, Lowrie was the best hitting shortstop in the AL. I've put Scutaro in my lineup for a couple of reasons, one of which is based on practicality. The front office has made it clear that they're going to start the season with Lowrie as the super utility player. Also, putting Lowrie in the lineup would make the runs per game calculation a bit tricky; due to his age and injuries Lowrie hasn't accumulated enough MLB at bats to make a good prediction using my admittedly basic model.

What about Ryan Kalish and Josh Reddick?
For 2011, both Kalish and Reddick appear to be blocked. As I wrote in my original post about the 2011 lineup, the Red Sox feel that Kalish could use more time in AAA, despite his strong showing last September. If a major injury occurs to an outfielder, it wouldn't surprise me to see the Red Sox turn to Kalish again. In 2011, the path to the big leagues is clear for at least one of them. JD Drew's contract expires after the 2011 season, and Kalish's great athleticism and arm strength seems like a great fit with the spacious right field in Fenway. The future isn't as clear for Reddick. He is organizational depth for the time being, but may end up as part of a trade package, as the younger Kalish has passed him in front office's mind. His line at AAA (.266/.301/.466) isn't particularly impressive, but the potential is definitely there, as shown by his post All Star Break numbers (.363/.385/.643). Reddick didn't impress in the Dominican Winter League either, although he showed a good approach at the plate (10 BB in 76 PA, much higher than his AAA walk rate). Next year is a make or break year for Reddick with the Red Sox; another mediocre showing in AAA could lead to him getting traded for some in season help.


  1. Good post. I think the main reason some of the projected best lineups start with Youkilis and Gonzalez is that they are the two best hitters on the team, and putting them there gives them a few more at bats over the course of the season.

  2. It does, although it still doesn't explain why Marco Scutaro, who has the lowest slugging and one of the lowest OBPs hits third, over someone like Pedroia, Crawford, Drew, or Ortiz.

  3. Has there been any significant research as to whether teams really should put their best hitters in the leadoff spot to maximize at bats?

  4. The lineup tool at Baseball Musings takes the number of at bats into account. Over the course of a season the difference between hitting first and 3rd or even 5th is relatively small, though. I can't find the study with some Googling, unfortunately. There is a trade off between OBP and slugging - where you want your high OBP hitters hitting ahead of your high slugging hitters. Often times, they are they same players though.

    In the end, differences in lineup order end up being pretty trivial. Fangraphs got into the differences in optimal lineups a bit. They used Bill James' projections, which are typically more optimistic, so their optimal lineup is different and scores more runs than the one I put together using weighted means.

  5. For those that are curious, I found this for the AL, 1998-2002:
    1. 4.83
    2. 4.72
    3. 4.61
    4. 4.49
    5. 4.39
    6. 4.26
    7. 4.14
    8. 4.02
    9. 3.90

    Taken from this article: