With over a quarter of the games in the 2011 season having been played, the time has come for us to cast off the shackles of small sample size and start making some adjustments to the lineup based on results rather than the assumption that "he'll turn it around."
The Orioles as a team have a slash line of .250/.315/.380 with a wOBA of .312. That's just below league average in every one of those categories, although it's no longer among the bottom feeders, either.
There are two basic assumptions you can make about lineup construction. The first is that you can maximize your expected runs by following a certain formula in making the lineup. The second is that even if you optimize your lineup according to this formula, the contribution over an entire season is a few more runs at most.
On top of that, there are the psychological factors that can't really be quantified that go into making lineups, the massaging of the egos and the building up of players' confidence. Or as Yogi Berra would say, "Ninety percent of baseball is half-mental." But it's not much fun to try to do a little analysis if you go around thinking that nothing matters, even if nothing does actually matter. So let's just ignore the fact that we rubes can never possibly know all; much, to us, is unknowable. Instead let us take a step or three away from reality and pursue that flighty temptress, adventure.
According to The Book (or at least my understanding of it), the ideal lineup construction will meet these criteria:
-
The best three hitters should bat 1st, 2nd and 4th. You want better OBP from the first two hitters and you want better SLG from the cleanup hitter.
-
The next two best hitters should bat 3rd and 5th.
-
Then in descending order, you want the hitters going 6th, 7th, 9th, 8th.
This is challenging the standard "old school" baseball lineup, where you might hear some manager or writer who's stuck 30 or more years in the past talking about the need to have a speedy leadoff hitter, a second batter who's good at sacrificing to move the runner up, and so on. This sort of thinking belongs in the past, with the sacrifice bunt. A takedown of sac bunting is another one of the core messages of The Book.
I am working with the assumption that these criteria are the best to use just because it gives us something to talk about. I'm not going to pretend to have a greater understanding of the principles than Tom Tango and company. With that aside, let's take a look at the Orioles in this context.
OPENING DAY
On April 1, 2011, the first game of the season, the lineup for the Orioles was as follows:
Brian Roberts (2B)
Nick Markakis (RF)
Derrek Lee (1B)
Vladimir Guerrero (DH)
Adam Jones (CF)
Luke Scott (LF)
Mark Reynolds (3B)
Matt Wieters (catcher)
J.J. Hardy (SS)
Let's also assume that given how everything looked on the Opening Day roster, these nine players should have been starting in some order. I think this is a safe assumption. So, it's Opening Day. How would we gauge the best possible lineup with these Orioles? For this exercise, we'll use 2010 stats. Even this is not perfect, but it's the easiest one to make, and probably the safest: for a given player, the most likely outcome would seem to be a small increase or a small decrease compared to the previous year's statistics. Columnists the world over have to tell us all about their predictions for breakout players, but I see this stuff as throwing things at the wall to see what sticks.
For the leadoff spot, the best OBP from this bunch in 2010 is Nick Markakis, who put up a .370 on the season. The next best is Luke Scott, with a .368 OBP, but he's also our best slugging percentage (.535) so we'll bat him cleanup and put our 3rd-best 2010 OBP player, Brian Roberts (.354) in the 2nd position.
This leaves us with six spots to fill. How are we going to judge the "best" hitter for these spots? To make my podcast co-host Andrew G happy, I'm going to use wOBA (weighted on base average), which gives different values to each outcome, with more bases being worth more, just not quite as much as SLG.
By wOBA, our next two best hitters from 2010 are Vlad (.360) and Lee (.340), so let's plug in Vlad as our #3 hitter and Lee 5th. Now we're left with Wieters (.303), Jones (.333), Hardy (.313) and Reynolds (.328). We would want to line them up descending, except with the lowest as 8th, so 6-9 would go Jones, Reynolds, Wieters, Hardy, and that would have given us an Opening Day lineup that would look like this:
Markakis
Roberts
Guerrero
Scott
Lee
Jones
Reynolds
Wieters
Hardy
Nobody is more than two places from where they "should" be, and 7-9 are locked in. Not bad. Note that in our optimal lineup we haven't considered things like platoon splits and whether to stagger our batters. I'm not going to sweat the platoon splits right now, but as far as staggering, our lefties are Markakis and Scott, and we have switch-hitters in Roberts and Wieters. So 5-7 are all righties, but the righty-righty platoon split is not as bad for hitters as the lefty-lefty one is. We are probably OK without making any further adjustments to account for this.
TODAY
The date is now May 23rd. The Orioles have played 45 games, which is over one-quarter of the season. Seven of the nine batters from the Opening Day lineup have 150+ plate appearances (Scott has 145, and Hardy, who spent time on the DL, has only 73). As fans, we are still hopeful that some players will improve on their current season numbers, with underperformers including Markakis, Scott, Reynolds and Hardy. This will lift up the entire lineup and perhaps give us a different "ideal" construction after a month. For now, we'll go with today's season-to-date stats.
To play with our current lineup, I am making an assumption that Roberts will be missing further time with his concussion issues and that we will see Robert Andino in his place. Though Lee is currently on the DL, signs point to him being ready when his 15 days have passed.
We still want our highest OBP batter in the leadoff spot. At the moment, this is Andino (.366 over 103 PA). I find this fascinating. This is still less than 1/6th of a full season's plate appearances for an every day player, but we've never seen this level of performance from him over any stretch. Whether he will sustain it, who knows? At this time, though, this is where he should be. Our next best OBP is Wieters (.342) so we will plug him in second. Until recently, the best slugger was Scott, but he's mired in quite a slump and so Guerrero has the best SLG on the team (.444) and he stays in the cleanup spot he's occupied all season long.
As we did for the Opening Day lineup, let's turn to wOBA to fill out the rest of the spots. The best wOBA on the team belongs to Adam Jones (.344). Jones is actually only a few points shy (.440 SLG) of being the cleanup hitter, and Guerrero (.341 wOBA) is close to Jones there. But for the sake of consistency in our experiment, because this is just meant to be a snapshot of a moment in time, we'll plug in Jones third. Next highest who isn't slotted yet is Scott (.327 wOBA) and he becomes our 5th place hitter.
Descending by wOBA we get Reynolds (.308) in 6th, and Markakis (.307) in 7th. Our worst hitter by this metric is Hardy (.286) so we'll drop him in 8th and put Lee (.295) in 9th.
That would mean the lineup on most nights would look like this:
Andino
Wieters
Jones
Guerrero
Scott
Reynolds
Markakis
Hardy
Lee
Here's where we have gotten fairly radical in our lineup placement. Wieters hitting second just looks weird, and Lee all the way down in 9th when he's been batting 3rd all season long is also something.
Would we need to make any adjustments to avoid having hitters from the same side of the plate bunched together in the lineup? Our lefties are separated by a batter and each lefty is two or three places away from our switch hitter (Wieters), so as for staggering, this is probably not too bad. With Lee on the DL, we might as well just plug the Pie/Reimold LF platoon into the bottom of the order and we'll get from it what we get.
Our optimized lineup still hasn't considered platoon splits. Perhaps we might be better off having a separate lineup for a right-handed starting pitcher or a left-handed one. For the purposes of this game, let's suppose that the gain from tweaking the regulars around every day is offset by the loss of mental edge of not having a consistent position in the lineup. This may not be true, but I can only juggle so many balls at once.
The bigger question is what to do with Roberts when he returns from the DL. He has 178 plate appearances and with a .277 wOBA, he is the worst regular hitter on the team. His .273 OBP is also the lowest among everyone who was in the Opening Day lineup. What is his talent level in 2011? Roberts hasn't had a sub-.340 OBP since 2003 (and even then it was .337), but when I've watched him this year I wonder if the bottom's fallen out. Can we simply assume he'll climb his way up even as far as a .320 OBP when he's back? Even if he gets to that point, according to our little game here he has no business being the leadoff hitter any more.
The Book concedes that the proper lineup construction strategy will only net you a few runs over the course of the season. But hey, if you squeeze those runs out in one-run games, that's no small thing. Of course, it's even easier to score more runs if you have better players in the lineup, but we can only fight one battle at a time.