Friday, February 21, 2014

What is the Tigers Best Line-up?

 Where should Miguel Cabrera bat in the Tigers line-up?
(Photo credit: Hardball Talk)

Every fan has his own idea of the ideal line-up.  Traditionalists tend to like to have a speedster lead off, a bat-control guy hit second, the best hitter third and the best slugger (who is not also the best hitter) bat fourth.  Some just want the numbers one and two hitters to get on base a lot and don't care as much about speed. Others follow The Book by Tom Tango, Mitchel Lichtman and Andrew Dolphins which claims that the best hitter should not bat third, bat rather first, second or fourth.  Still others toy with the idea of having the best hitter on the team lead off, the second best hitter bat second, etc. with the reasoning that the best hitters should get the most at bats.   

One thing I like to do before every season is check out the line-up tool at Baseball Musings.  Developed by analysts Cyril Morong, Ken Arneson and Ryan Armbrust, it estimates the number of runs a line-up would score based on every batter's on-base percentage (OBP) and slugging average (SLG).  Since getting on base (OBP) and advancing runners with hits (SLG) are the two most important elements of run scoring, their method makes some sense.

However, the line-up algorithm also has limitations.  Perhaps most importantly, it does not consider the speed of base runners.  It also does not address psychological factors such as batters feeling comfortable in certain spots.  What it does do is try to determine the best line-ups based purely on hitting which is a good place to start.

Using the Steamer projections, I plugged OBP and SLG for the nine Tigers starters into the line-up analyzer.  One possible line-up is shown in Table 1 below.  The line-up tool says that line-up would score 5.050 runs per game or 818 runs in 162 games.  If that sounds like a lot of runs, it's because we are assuming that all nine players are going to play 162 games which, of course, won't happen.  That's OK though.  The goal is just to compare different line-ups.

Table 1: Tigers Possible Line-up

The line-up tool considers every possible permutation of those nine batters and estimates that the best line-up would score 5.117 RPG or 829 runs, while the worst would score 4.856 RPG or 787 runs.  That is a difference of 42 runs which is not huge, but not insignificant either - an estimated four wins.

Table 2 shows that the top five line-ups have the same players in the top five spots of the order - Miguel Cabrera, Torii Hunter, Austin Jackson, Ian Kinsler and Victor Martinez.  That is not surprising and manager Brad Ausmus will likely also have the same batters atop the order.  What is interesting is that Cabrera is second in each optimal line-up.  In fact, the thirty most productive line-ups have Cabrera batting second.   In theory, this makers sense because Cabrera would maximize his plate appearances, but still have many opportunuities to knock in runs batting second. 

Table 2: The Five Top Run-Producing Line-ups 

Runs per Game123456789
5.117 Jackson Cabrera Hunter Martinez Kinsler Dirks Avila Castellanos Iglesias
5.117 Jackson Cabrera Hunter Martinez Kinsler Castellanos Avila Dirks Iglesias
5.117 Jackson Cabrera Kinsler Martinez Hunter Dirks Avila Castellanos Iglesias
5.116 Jackson Cabrera Kinsler Martinez Hunter Castellanos Avila Dirks Iglesias
5.116 Martinez Cabrera Hunter Jackson Kinsler Dirks Avila Castellanos Iglesia

Table 3 looks at the worst line-ups.  Right away, you see a big problem - that Cabrera is batting ninth.  I think we can all agree that this would not be a good idea. As bad as those line-ups are, they would still produce only 5% fewer runs than the best line-ups.  We want those five percent though, so those line-ups are out.

Table 3: The Five Lowest Run-Producing Line-ups

Runs per Game123456789
4.856 Castellanos Iglesias Kinsler Dirks Avila Martinez Hunter Jackson Cabrera
4.856 Castellanos Iglesias Avila Dirks Kinsler Martinez Hunter Jackson Cabrera
4.856 Castellanos Iglesias Jackson Dirks Kinsler Martinez Hunter Avila Cabrera
4.857 Castellanos Iglesias Avila Dirks Hunter Martinez Kinsler Jackson Cabrera
4.857 Castellanos Iglesias Kinsler Dirks Jackson Martinez Hunter Avila Cabrera

Getting back to the best line-ups, The Book would agree that Cabrera, as best hitter, should not bat third, but rather first, second or fourth. The second best hitter - Martinez - should also bat in one of those spots. The line-up optimizer cooperates by putting Martinez clean up in the top four line-ups.  It wouldn't make much sense to have either of those lead-footed runners batting lead off, so second and fourth seems to be the way to go.  The next best hitter is either Jackson or Kinsler, so either one of them could leadoff.  So, I think any of the top four line-ups seems reasonable.


  1. Well, the Tigers got a bit of an offensive bump when Cabrera moved up to second in the ALCS.The effect might have been more pronounced if Miggy hadn't been a crippled shell of himself at the time. I always thought that batting your best hitter second made more sense in the American League, since the top of the lineup sees more baserunners without the pitcher batting ninth and giving away two or more outs a game.

  2. Great breakdown! And there's other things that are also important to consider. Like I don't think Dirks is a good long-term keeper so my preference would be to see him be below Castellanos in the batting order because I want to challenge Castellanos and find out more information about how he can perform in the batter's box, whereas with Andy Dirks I don't really care as much.

    This is good stuff to think about, and that's the problem with the Tigers, is they don't know how to play the numbers or even understand how the numbers work. Further, if they could figure that stuff out then they should look at how you can get exponential benefits by creating a string of better performing players, so the tool to calculate the best batting order or to project runs will get slightly better than the sum of all the parts as you escalate into higher calibers of talent efficiency.

    I still think we have a severe shortage of value/efficient offensive talent, and that's my #1 gripe as of right now.

    1. I agree about the shortage of offense. I can imagine a very long slog through the back half of the batting order--Dirks/Davis, Castellanos, Avila, Iggy (and age could well catch up to Hunter and maybe VMart). It is not at all inconceivable that as many as four or even five of these guys have down years. We may find ourselves trying to score more runs in three or four innings, when our big guns are aligned, as the other guys do in nine. I sure hope the rotation stays healthy.

    2. Yeah it's a lot to be concerned about, I'm much more worried though about 2015-2025 than I am about 2014. To me we only have a shortage of offense now because the last 10 years we didn't do a good job of building out for medium-term and long-term plans. So now is the time that we can have an impact on the next 5 years and soon we need to start worrying about the 5 years after that. But we seem to take things one day at a time and not consider what the consequences are as early as tomorrow.

      We just don't know how to create efficient leverage in transactions or strategies in general, as well as in evaluating what players are worth, and thus we overpay and underachieve.

    3. So to clarify, I would have Castellanos locked in to anywhere between 2-6 all season no matter what, and any spot could be eligible for him. I want to see him show a sign of failure or triumphant, one or the other it will be good for him either way. Let's see what he's got.

  3. Berdj Joseph RassamFebruary 28, 2014

    I like the option with Kinsler in the 3 spot and Dirks in the 6 spot.



Blog Archive


My Sabermetrics Book

My Sabermetrics Book
One of Baseball America's top ten books of 2010

Other Sabermetrics Books

Stat Counter