David Richard-US PRESSWIRE
Based on ZIPS projections, what should the Toronto Blue Jays 2013 batting order look like?
For my debut here at Bluebird Banter (thank you, thank you, you’re far too kind), I had planned to list and discuss the 2013 ZIPS projections for the Toronto Blue Jays.
However, instead of ZIPS projections being scattered around individual Fangraph pages or at Baseball Think Factory, they are instead also available at Fangraphs via this nice tidy article. So they did that for me. Debut finished.
I kid, of course. While me simply compiling the ZIPS projections wouldn’t have been too much of a breakthrough debut, it does hint at what I hope to bring to the BB team – statistical analysis. I also write for Beyond the Boxscore here at SB Nation, but I promise I’ll keep any sabermetric analysis here to an application rather than theory/research level.
All of that is a really long way of introducing today’s topic, which is…
What should the 2013 Blue Jays batting order look like?
To determine this, we can use a couple of different methodologies. We can use the typical "baseball consensus" which has a speedster up top, a lefty in the two-hole, sluggers three-through-five, and so on. With three switch hitters in the starting nine, the Jays also have options aplenty in terms of shifting guys around to split up lefties and righties and cause opposing managers fits with their bullpen management.
First, some explanations.
ZIPS – ZIPS is a tool developed by Dan Szymbroski to predict future performance. It has shown to be among the better projection systems in terms of accuracy, although I haven’t seen how it fared in 2012 yet. In the past, it has fared well with predicting OPS, which is good since our lineup optimizer relies on OBP and SLG rates.
The Optimizer – I’m not sure when the last time it was updated was, but basically it orders your batters based on "the importance of not creating an out." That is, it looks at how often certain places in the order come up in different situations, weights them based on importance and then places players in the "proper" order.
3. Jose Reyes
4. Brett Lawrie
7. Adam Lind
8. Colby Rasmus
Obviously, this is an odd lineup configuration and probably makes you want to throw the entire optimizer out the window. And you're free to do that. However, the optimizer also shows 20 lineups that would all be worth almost the exact same in terms of runs per game (the lineup above would be predicted to score 5.291 runs per game, while the lineup below would be predicted to score 5.285 runs per game).
Since many line-ups are similar, it means that John Gibbons has some wiggle room to try and create additional runs via the basepaths, match-ups and more. This is strictly based on on-base and slugging, so there is more to be gained.
Now, the reason the optimizer likes Bautista so high is twofold: for one, he’s expected to be so good that it just makes sense to get him a ton of at bats, and, two, the lineup is pretty strong through to the bottom, so he’d still be expected to have a fair number of runners on batting high in the order.
This is probably very counter to what most people were expecting, which was probably something like this (what I had proposed initially after the trade):
5. Lind (v R), Lawrie (v L)
6. Lawrie (v L), Lind (v R)
7. Rasmus (v R), Arencibia (v L)
8. Arencivia (v L), Rasmus (v L)
I should note, too, that the optimizer would have Emilio Bonifacio in the same place as Izturis. Personally, I’m on board with Izturis being the "starter" with Bonifacio being a super-sub bouncing around the field to provide off days and keep Lind out of the line-up against lefties.
While you may quibble with the actual line-up the optimizer spits out, I think it at least confirms that the top-five in the order are pretty clear and the bottom-four pretty clear. You could argue for Lind or Rasmus higher against righties, but on the whole we have a pretty fair understanding of how the line-up is going to break down.
I’m curious to see what everyone else’s line-ups would look like, so have at it in the comments. You’re also free to plug in lefty/righty-specific line-ups in the optimizer and see how those shake out. (Or, again, throw it in the trash since it has Bautista leading off).
One really nice note about the optimizer is that it predicts that the very worst lineup configuration would still be expected to score 4.920 runs per game. This tells us that a) line-up configuration doesn’t make much difference on a day-to-day basis, b) it can actually make a difference of as much as 60 runs over the season (six wins) in a very extreme case, and, c) the Jays offense would be expected to be the third best in the American League even if Gibbons botched the lineup daily. Nails.