clock menu more-arrow no yes mobile

Filed under:

Rank Strangers to Me

I figured I would mix things up and compose an entire post consisting of no reference to starting pitching, other than this reference to starting pitching in which I announce I won't reference starting pitching. Instead, I'll look at the starting lineup---and several possible permutations thereof. Although this post will feature some numbers and a table, it's not intended to be interpreted as serious analysis. Simply stated, desite the forthcoming numbers and table, I think this might be fun.

Last year, Scott Collins of Nats Farm Authority wrote a guest post for this site in which he analyzed the Nats' prospective lineup possibilities based on prior research and as applied by the Baseball Musings Lineup Analysis tool, which I'll call the "lineup analyzer" because that sounds sort of cool. I'm going to do something similar, but I won't touch on the more substantive elements, in case Scott wants to revisit the topic. What I am going to do with this post is run several possible batting orders through the lineup analyzer and see what the thingy spits out.

At this point, I should insert a disclaimer: In addition to the fact that I am running projected performance through a computer program---and heaven knows computers are evil, what with Skynet being self-aware---the numbers being inserted into this computer program are in the form of projections themselves; with one exception, I will be using the weighted mean projections prepared by the Baseball Prospectus PECOTA system. [It's PECOTA Week here at Teh Fed!] I sort of regard these types of things as "hearsay-within-hearsay" situations---not because either PECOTA or the lineup analyzer is providing hearsay testimony, but because it sounds cooler than "danger-within-danger." What I'm saying is that I'm submitting results based on a projected premise, so your mileage may vary on the utility of this activity. But PECOTA is regarded as "deadly accurate," so at least our premises (premisi?) aren't too flimsy here.

The one exception to the PECOTA projections is the ninth spot in the order, which, in light of Ryan's triumphant return to blogging, I'll dedicate to a wrestler. How about, say, Jimmy Superfly Snuka. Rad. Because I'm not going to play around with the substance and theory of lineup construction, and because my name isn't Tony La Russa, I'm going to keep it cool and make Superfly's spot essentially the pitcher's spot. Just so you know, I do this at my peril, seeing as the lineup analyzer has an incredibly persistent desire to see Nook Logan bat ninth, but I'll resist that urge. Since I need to assign OBP and SLG values for all players for the lineup analyzer, what I've done is insert the National League average for all ninth-order hitters, which was .229 OBP and .266 SLG. I will withhold the rest of the OBP/SLG numbers, since they were generated by PECOTA, which is a subscriber-based service, and the projected numbers themselves aren't that relevant to this post.

Now, on to the lineups. Nine of them will follow. The first one is what could be referred to as the "Newspaper Reader's Lineup," or for simplicity, Lineup No. 1. This lineup has the information we know about the planned real-world lineup by reading the stories and quotations in the media. It has Felipe Lopez batting first, since we know the team plans to hit Lopez leadoff. It has Nook Logan playing center, since we know Logan is at present the presumptive starter in center. It has Cristian Guzman batting second, since we know Manny Acta has openly mulled this possibility. And, perhaps less confidently, it has Kory Casto in left, since we know the team has at least professed a desire to grant him a regular spot in the lineup. Taking these suppositions---and then harmonizing them with the other regulars in a reasonable and realistic batting order---I inserted all the required information into the lineup analyzer, which yielded a projected figure for runs per game.

Lineup No. 1:
Lopez
Guzman
Johnson
Zimmerman
Kearns
Casto
Schneider
Logan
Superfly Snuka
4.485 runs per game

The first alteration I will make to this base lineup is simply flipping around Guzman's spot in the order. Instead of batting teh Guuuz in the two-hole, what if the Nats were to drop him down in the order? This would seem to create a weak bottom of the order (with Schneider, Guzman, Logan, and Superfly Snuka), but it would also place a better hitter---it's probably not revealing a trade secret to say no regular hitter other than Superfly Snuka projects to be worse than Guzman---up in the order. So I plopped Casto, by most accounts a patient hitter, into the second spot and then let the lineup analyzer do its thing.

Lineup No. 2:
Lopez
Casto
Johnson
Zimmerman
Kearns
Schneider
Guzman
Logan
Superfly Snuka
4.525 r/g

As you can see, the lineup analyzer says this little manuever boosts the Nats' runs per game by a touch. Nothing earth-shaking, but it seems to make a tiny difference.

Casto is not a sure thing to start the season in the majors, though. Although Casto is 25 years-old, he's never played above Double-A and, perhaps more importantly, he has minor league option years remaining. The recently acquired Chris Snelling, who is older than Casto by a matter of days, is out of options. He could be an asset, so there's no reason to expose him to waivers foolishly. What if Snelling were to be the starter in left? He could also bat second. So I gave that one change from the previous lineup (Snelling for Casto) a whirl in the lineup analyzer.

Lineup No. 3:
Lopez
Snelling
Johnson
Zimmerman
Kearns
Schneider
Guzman
Logan
Superfly Snuka
4.655 r/g

With Lineup No. 3, we see another slight increase in runs per game, given PECOTA projects Snelling more favorably than Casto. It's a bigger increase than the difference between Lineup Nos. 1 and 2, but it still seems rather slight.

For the next alteration, I thought about toying with the presumption Lopez would hit leadoff. Who else would do so? [And don't say "Nick Johnson." We'll get to that later.] What about Nook Logan? He's fast, right? Fast guys make good leadoff hitters, right?

Right?

Well, Logan's fast, right?

Yes. So Lineup No. 4 places Logan in the leadoff spot, and it moves Lopez to the two-hole, where he hit behind Soriano in the second half of last season.

As an added measure, we just know Snelling will get injured, so let's substitute in his injury platoon mate, Alex Escobar.

Lineup No. 4:
Logan
Lopez
Johnson
Zimmerman
Kearns
Escobar
Schneider
Guzman
Superfly Snuka
4.542 r/g

Some of that drop-off is the difference between the PECOTA projections for Snelling and Escobar, but perhaps we have learned to keep Logan out of the leadoff spot for these lineup projections.

In fact, let's go one better and just say we'll omit Logan from the lineup altogether. [Altogether: "We'll omit Logan from the lineup!"] Let's allow the universe to collapse on itself; let's place Snelling and Escobar in the lineup together.

Lineup No. 5:
Lopez
Snelling
Johnson
Zimmerman
Kearns
Escobar
Schneider
Guzman
Superfly Snuka
4.844 r/g

Hey, better. I've got a notion about that, but we'll return to it in a moment.

On to the next alteration---say, what about Ryan Church? He hasn't been traded yet, so we might as well figure him in to this activity. Let's put Church in left, re-insert Logan in center, and jumble Guzman to the two-hole (I mean to promise not to do that again).

Lineup No. 6:
Lopez
Guzman
Johnson
Zimmerman
Kearns
Church
Schneider
Logan
Superfly Snuka
4.569 r/g

The result is better than the base lineup, but not as good as some of the alterations. What if we were to trade Guzman with someone else for the two-hole? I'll be realistic about this and make the second hitter someone other than Church---who, as we certainly know by now, has had his troubles with the off-speed stuff and two-strike counts. I've seen some people suggest Brian Schneider could be okay in the second hole. I don't know about that, but let's give it a whirl with the lineup analyzer.

Lineup No. 7:
Lopez
Schneider
Johnson
Zimmerman
Kearns
Church
Guzman
Logan
Superfly Snuka
4.604 r/g

The result is a minimal increase---very minimal. Two more lineups to go. I can see two other combinations of note.

The first is placing Church and Escobar in the same lineup, with the former in left and the latter in center. Thus, Logan is omitted; this, I've found, might be a key to good run-scoring projections.

Lineup No. 8:
Lopez
Escobar
Johnson
Zimmerman
Kearns
Church
Schneider
Guzman
Superfly Snuka
4.811 r/g

That's more like it. Finally, let's put Church and Snelling together, with Church in center this time.

Lineup No. 9:
Lopez
Snelling
Johnson
Zimmerman
Kearns
Church
Schneider
Guzman
Superfly Snuka
4.889 r/g

Even better---as noted previously, PECOTA likes Snelling more than Escobar. Of course, such preferences lose a great degree of relevance when both guys are unable to play, but this is supposed to be a fun exercise, so let's not presume too many injuries.

* * * *

Nine up, nine down. That's probably a lot of information to absorb, so I'm going to pull out the aforementioned table. We'll call it Table 1.1, just to make this post seem important. Additionally, I'm going to convert the runs per game results to runs per 162 games; this puts the results in a different (though non-arbitrary) context and highlights the differences between the lineup projections in something other than decimals.

Table 1.1: Comparison of Proffered Lineups on Runs per 162 Game Basis

Rank RS/162 Lineup No.
Best 792 Nine
Second Best 785 Five
Third Best 779 Eight
Fourth Best 754 Three
Fifth Best 746 Seven
Sixth Best 740 Six
Seventh Best 736 Four
Eighth Best 733 Two
Worst 726 One

The worst projected lineup, according to the lineup analyzer, is the one that started this exercise, without the alterations. Isn't that just a kicker?

All told, the spread between the best lineup and the worst one---again, according to the lineup analyzer---is 66 runs over the course of a 162-game season. This is not an insubstantial difference. Dividing that figure by 10---which, sabermetrically speaking, is the number of additional runs over the course of a season that it normally takes to add one win---yields a difference of just over six-and-a-half wins. Note that I am not taking anything else into account: no defense, no speed, no basestealing, no "putting the pressure on the defense," no grounding into double plays, no leadership, no bad body language. There are variables that aren't analyzed in this post; those are for another post, either here or elsewhere. What I am saying is:

  • if you trust the PECOTA projections, and
  • if you apply the "weighted mean" of the projections PECOTA makes, and
  • if you trust the process by which the Baseball Musings lineup tool simulates run scoring, then
as a matter of run scoring, the manner in which the Nats make out their lineup---and the personnel making up the lineup---makes a real difference.

Of course, you don't need to subscribe to the three statements above to come to the conclusion that how you fill out the lineup---and who's in that lineup---makes a difference. That is self-evident; if we had Ichiro!, then we could tell Nook to search out Ichiro!'s old job in Japan. And yes, it would also help if the Nats didn't have to submit Cristian Guzman's wretched PECOTA projections into the mix. How would you project a guy who missed an entire season and whose prior season was beyond pathetic? PECOTA shows no mercy---and when I say no mercy, I mean absolutely no mercy. Me, I'm sort of a Law of the Cosmic Balance kind of guy; he's "due" for a bounce-back, I suppose. But I'd rather not risk cosmic kindness (and even a bounce-back wouldn't be that great). Finally, a "weighted mean" is just that---a mean. Does anyone honestly believe Soriano, while terrific last season, is as good as he was last season? I don't; I think he had a legitimately great year. On the other end of the spectrum, Brian Schneider probably had a legitimately bad year. Not everyone will perform to a reasonable mean.

Yet, that's what we have to work with here. Based on this kind of best guess, I'd say a lineup with Snelling and Church is preferable. There could be reasons to deviate from that---maybe they need to trade Church, and maybe they feel they need to play Casto---but, like I said, I did this for the fun.

* * * *

One final note: There are other ways to tweak this thing with the lineup analyzer. For every nine players you submit to the lineup tool, it spits out the "Best" and "Worst" lineup orders. Almost every conceivable combination of players above suggests hitting Nick Johnson in the leadoff spot. This is all about OBP. It ain't going to happen. Nearly as many combinations suggest batting Logan ninth. That ain't happenin', either. Whether it would be theoretically beneficial to do so is outside the scope of this post. But I thought I would note this.

* * * *

Okay, I lied---this is the final note: Nick Johnson is probably out for Opening Day, and maybe for the first month of the regular season. He'll be replaced by Robert Fick, or Larry Broadway, or Travis Lee. I could run these guys through the lineup analyzer for a runs per game estimate, but I figure I've used too many numbers in this post already.