WS Coverage: Philadelphia’s Lineup Construction
Here’s how the Philadelphia Phillies lineup card read for the opening game of the World Series last night.
1.Jimmy Rollins, SS
2.Shane Victorino, CF
3.Chase Utley, 2B
4.Ryan Howard, 1B
5.Jayson Werth, RF
6.Raul Ibanez, DH
7.Ben Francisco, LF
8.Pedro Feliz, 3B
9.Carlos Ruiz, CA
It’s possible that nothing here jumps off the page. Rollins and Victorino are typical 1 / 2 hitters; Utley is the Phillies best hitter; Howard fits the cleanup spot perfectly, and the rest of the lineup just sort of falls into place.
However, this is what we see when we look at the handedness of the batters: S-S-L-L-R-L-R-R-R Specifically, what stands out is the fact that Charlie Manuel is unnecessarily batting two left handed batters in a row: Chase Utley and Ryan Howard.
Right now, the Yankee bullpen has two left handed relievers in Damaso Marte and Phil Coke, both of whom are more than adequate against left handed batters. Joe Girardi should have no qualms about using either of these pitchers against Utley and Howard in the middle or late innings. This is the situation an opposing manager dreams of with regards to the LOOGY – you can use one of your lefty specialists to get out two batters in a row – in this case, the opponents two best hitters – and still have another one for another situation later in the game.
One of the potential arguments against this line of reasoning is that Chase Utley doesn’t show much of a platoon split and even showed a reverse platoon split this year (see graph). However, from The Book, left handed batters tend to show a platoon split of almost .027 points of wOBA. With the amount of variation present in this statistic, 1,000 PAs – roughly the amount that Utley has vs. LHPs in his career – are required to regress the observed platoon split halfway to the mean. So we should still assume that Utley will perform lower against left handed pitching.
And then consider the fact that Manuel leads his lineup off with two switch hitters. Switch hitters, intuitively, have a tiny platoon split compared to non-switch hitters. Then there’s the simple solution of merely switching Victorino and Utley in the lineup. It breaks the duo of left handed batters, and as an added bonus, batting Utley in the second spot leverages his talent slightly better. In the second spot, Utley will receive more PAs per game and will be less likely to bat with nobody on and 2 outs, as frequently happens in the first inning of games.
This decision had a minimal impact on Manuel’s Phillies in Game 1, as New York’s bullpen pitchers of either hand were ineffective. There is no reason, however, to continue to give your opponent a competitive advantage such as this, and Philly fans should hope to see a lineup change in Game 2.
Print This Post
If he were constructing an optimal lineup out of those 9 guys I’d think the first step would be to get Rollins the heck out of the leadoff spot.
Victorino
Utley
Werth
Howard
Francisco
Ibanez
Rollins
Ruiz
Feliz
looks good to me.
We all know that a lineup won’t have too much effect on a full season, but why on earth would you bat Francisco ahead of Ibanez…
SLRLRLSRR
I’m with you on this one. Are you guys completely discounting the disparity in power between Francisco and Ibanez? To me, you’re suggesting that mixing up left handed and right handed hitters is the ONLY relevant factor in creating a lineup which seems kind of silly.
Ibanez career ISO: .195
Francisco career ISO: .184
Sure, Ibanez has been playing out of his mind this year. But it’s not like Francisco sucks.
Howard and Ibanez need to be split up more than Howard and Utley do.
I do agree that the lineup looks better with Utley in the two hole, but only because I rather see Utley/Howard/Werth/Ibanez all bat before Victorino. A lefty is going to face Howard anyway, and neither Utley nor Ibanez have poor numbers against lefties.
A bit surprised this article comes out the day after Chase Utley hit two homeruns off of LHP Sabathia…
Results don’t determine analysis of decisions. Just because he did well doesn’t mean it was right. Also, he could have just easily hit them off of Sabathia in the second spot and still skipped LOOGY issues.
In addition, Werth appears to the superior hitter against lefties anyway, using career splits (didn’t weigh them, but I think it wouldn’t make a huge difference). The lineup could have easily been:
Rollins
Victorino
Utley
Werth
Howard
Ibanez
Francisco
Feliz
Ruiz
And it would not have messed much with any “old school” lineup building philosophies.
How is the point that Utley will eventually regress somewhat to the mean platoon split at all meaningful, without comparing the relationship between Utley’s perceived “real/talent-based” platoon split vs LHP and to Victorino’s splits vs LHP?
Without that comparison this analysis is meaningless. One guy is a future HOFer and one is an above average hitter. Just because you don’t think Utley is as good against LHP as he performed this year doesn’t mean he still isn’t better than Victorino
But lineup reconstruction in the playoffs will KILL DA TEEEEEEEEEEEEEEEEM
/sarcasm
i’ve wanted utley hitting second for, like, forever, but…
a) we should all know by now that lineup effects are minimal.
b) the trend in utley’s l/r splits over his career strong enough that i suspect it is statistically significant and should warn one off applying a standard regression-to-mean formula to him.
c) rollins past 3 or 4 seasons of obp and sb/cs and his babip this year say he’s not a disasterous leadoff hitter
d) given the above, for charlie to mess with his players’ minds by messing with the lineup would be an high risk for a small reward. in the ideal world it would have been rollins-utley-victorino-howard-werth-ibanez-ruiz-feliz (or something like that) all year. in the real world, it’s too late to change now.
The regression formula takes into account how meaningful his career sample is. The more significant the sample, the less the regression formula affects it. There is no reason to take a certain significance level and say all of a sudden at this point you don’t want to regress. It is a constantly shifting spectrum where the effects of regression decrease continuously as the sample grows, until eventually, the effects are small enough that it doesn’t make much difference if you regress or not. That doesn’t mean you shouldn’t regress, just that eventually, there will be a point where you gain so little that you might decide it’s not worth the effort. If throwing out the regression makes a significant difference in the projection, then the sample is not significant enough that you should consider not regressing. If the sample is big enough that the regression doesn’t change much, then you’re certainly not losing anything by regressing anyway.
This is nonsense. At this point in the season the Phils and Yankees have played over 170 games. Messing with minor changes in the lineup because of the slightest differences smacks of over managing. Joe Torre and batting Arod 8th would probably love this type of over zealous analysis.
If you were suggesting this at the beginning of the season would be fine with me, but at this point I think this is ridiculus.
Not really. How is dealing with platoon splits, a real effect, the same as moving A-Rod to the eighth spot because he struggled for two weeks?
A ‘real effect’ over a course of a season or a career but they are unlikely to make a difference in a seven game system. I think the context of this article, and your post, is nonsense, to make changes to a batting order which has had the same order for the 3 4 hitters for all but 5-10 games this years seems like over managing. I speciffically mentioned in the post above which you decided to ignore, that if you wanted to suggest a new line up to be used for a full season, then I’m fine with it, but to really suggest moving the line up around now is unneccessary.
Platoon splits, as I think you even mention in another silly post below, should be based on mulitple season’s worth of data, not two weeks. If you wanted to construct a thought experiment with the Phils playing only the yankees every game over 162 season and run the two different line ups fine, tell me how large a difference it makes. But to dicker with a lineup that was rather successul all season for a seven game season is silly.
Did Torre move Arod down because he was unsuccesful in general, or unsuccessful against the hard throwing right handers?
What is the negative effect of moving the players around in the lineup? If we move Utley up in the lineup, what’s the downside here?
Jack has already explained the upside of a move like this. As mentioned, it doesn’t even have to be drastic; it’s as simple as moving Utley up to 2nd, Werth to 3rd, and Howard to 4th. Or moving Howard down to 5th and moving Werth to 4th. Something like that. This can help avoid the platoon advantage of LOOGY management. If you can tell me that the downside to those simple moves is greater than this upside, than your point can be well taken. Otherwise, it’s just saying “it’s been done like this for a while, there’s no need to optimize it.”
The downside is unknown, just like the upside is unknown in a seven game series. The number of at bats and trials is too small. Repeat after me “Sample Size”
And this isn’t even getting at what ever other impacts this has on a player.
“oh my gosh, my manager must think I stink I’m not batting third” , “Yeah I the big piece now I’m batting third, etc, Pressperson #1: is moving to number 2 a demotion? Why do you suck now?”
You can’t quantitfy emotions/Psych/etc, and the World Series isn’t the time to experiment. You can’t quantify disctrations from the media, etc. This is not the time to mess with a line up.
Jack has not explained the upside over a seven game series and niether has you, probably becuase it is negligble, immaterial, small, etc.
I agree. the fallacious leap here is that the stats the players got this season would have been the same even with the change in lineup. It may, in fact, be that Utley would have hit lefties harder had he not batted next to Howard, or maybe he would have hit them much less. You can not know for sure, so using this seasons stats to say the lineup would be better this way or that for a game this season is really pretty meaningless. Basically, your doing a different experiment and expecting the same results.
I still can’t see the reason to bat Victorino and Rollins before Howard/Werth/Ibanez… I understand having Rollins OR Victorino lead off, but for both of them hit before those power hitters doesn’t seem reasonable to me. Again, I understand the effects are minimal throughout a whole season, but I would hate to see Victorino make the last out with Howard on deck in a World Series game.
In other words… would you walk Utley to face Victorino?
Would you walk Utley to face Howard?
Going back to the earlier suggestion:
Would you walk Howard to face Francisco?
Would you walk Howard to face Ibanez?
At least he has gotten away from 3 lefties in a row, like he had for major parts of the regular season:
Utley
Howard
Ibanez
Was a frequent 3/4/5
Doesn’t Ibanez have equal or better numbers against lefties?
For the year he has a better OPS (by about .100) against LHP, but for his career its the opposite. So, this year is probably just noise.
Max has a good point, but the discussion of what is optimal is still interesting.
The Phillies are 19-5 in the playoffs the last 2 years with Utley and Howard hitting back to back. I don’t think any Phillies fan is hoping for a change at this stage. No reason to mess with success.
What is this crap article? WTF is going on with this site? This is the world series, put more than no effort into the WS articles geez.
I think this is a case of “If it aint broke…”.
If we all lived by that slogan, we’d still be living in the stone ages. “I kill animal Ok with rock, why me need spear?”
…which is why we don’t all “live” by that phrase, we follow it in cases where it applies. This is such a case. Just like I said (“I think this is a CASE of…”).
This is piece is bizarre. The timing is odd, the statistical analysis is incomplete and the lack of context is, to be frank, baffling.
Context is king. The Phillies won the World Series last season with Chase Utley hitting 3rd and Ryan Howard hitting 4th. They are in the World Series again with Chase Utley hitting 3rd and Ryan Howard hitting fourth. They have faced a number of teams with competent left handed relievers and have somehow managed to win a hell of a lot of games with this lineup in the last two seasons. Our emphasis on left/right splits has swung so far that it’s become almost absurd. There are certain left handed pitchers that Howard can handle and there are certain left handed pitchers that he cannot. There are certain right handed pitchers that Howard can handle and certain right handed pitchers that he cannot. Obviously left handed pitchers tend to give him more trouble, but to simply argue left/right all the time is kind of reductionist, no?
I would argue that the Phillies lineup as currently constructed accentuates its strengths. Philadelphia tends to win by putting together big innings and the everyday lineup allows them to do that. It seems to me that this is exactly the kind of analysis that makes those of us who are statistically inclined seem out of touch. The Phillies have used this lineup to great success. Now they should change it before the second game of the world series? I’m sorry, that might make sense in a vacuum, but certainly not if you’re using common sense.
They also won the World Series w/ Jamie Moyer in the starting rotation. Does that mean he needs to start game 3? Moreover, the won it w/o Cliff Lee. Does that mean he shouldn’t be pitching in the series? The fact that they won a World Series (as long as we’re talking about “context”) and one thing is true doesn’t mean they won the Series BECAUSE Utley batted third and Howard fourth. You’re confusing correlation with causation. It also doesn’t mean that having Utley bat higher in the order wouldn’t optimize their lineup. Personally, I think they’d be better off w/ Utley leading off. They might even win the World Series that way.
Okay, no need to be smug. You’re not the only person here that can think. I don’t want Utley leading off and I don’t think hitting him second optimizes the lineup. The Phillies win by putting together big innings and Utley is a significant power threat. I want people to be on base ahead of him.
And I fully understand the concepts of Correlation and Causation. I’m not saying that the Phillies won the World Series BECAUSE they hit Utley 3rd and Howard 4th, I’m merely pointing out that it happened. Once again, this is statistical overthink. Are you honestly telling me that you would make this lineup change right now if you were Charlie Manual? How would you line things up? No matter how you do it, there are going to be points of contention.
The Phillies are pitted against a team that, on paper, is much better than they are. They need every advantage they can get. If everything else was equal, the Phils would gain by splitting up the lefties. (Everything else is not equal and some of the factors are intangible. How much does comfort in a certain spot in the lineup affect a hitter? I’m thinking it doesn’t affect them much at all, but I can’t prove it.) It is ignorant to argue that past success with a lineup configuration is evidence that that configuration is optimal. The Phillies won last year without Cliff Lee. Does this mean that a Phillies team without Cliff Lee is better than one with Cliff Lee? Clearly not, yet this is the exact same argument you are making.
Here’s a link that might prove helpful.
Process
To your second point, how does the current lineup help the Phillies sustain long innings? I can’t see how flipping Utley and Victorino would help or hinder this in any way. In a long inning, they would still both hit. What is your reasoning behind this?
Doh! My response was way late. Forgiveness please.
They also won the World Series with Pete Rose playing first base. Bring him back!
Once again, this is the kind of condescending stupidity that makes it hard for those of us who are reasonable to sell sabermetrics to the mainstream. Obviously, you’ve never taken a logic course.
You can think for yourself once in a while, y’know? You don’t have to drink every last drop of the saber kool-aid.
“Think for myself?” Well, it sounds crazy, but I might just try it. Sorry for making a joke that offended you so much you had to call me stupid.
Doing what they did last year = “thinking for yourself.”
A logic course seems like a great idea. Please tell me exactly which modus ponens, destructive syllogism, etc. I missed! Perhaps you prefer Aristotelian to Fregean logic? Which is more preferable for baseball?
My apologies, my nerves are a bit jangly today. I just wish it were possible to have a reasoned discussion about these things without it devolving into, “Well the numbers say this so LA-LA-LA-LA-LA-LA-LA-I-CAN’T-HEAR-YOU”
I understand the difference between correlation and causation, but clearly what you are taking my point to an absurd length. So was Chuck. If the conditions of a situation have changed, then you absolutely must adapt your approach. If Pete Rose was still an elite baseball player, I’d be happy to have him in the lineup tonight. If Jamie Moyer were still a quality starting pitcher, I’d be happy to have him pitch game three. But things change, skills erode and good teams must adapt.
I don’t think the Phillies have an optimized lineup, because that would likely entail Ryan Howard hitting fifth. I’d really hate to lose a game because he didn’t get a chance to hit being that far down in the order. With so much of their power coming from the left side, juggling the lineup is always going to be tough. As stated, I also don’t want Utley hitting second because I think it marginalizes his power.
Though I don’t know what results juggling the lineup would have, I know what results the current lineup has produced. In a seven game sample, it seems crazy to open yourself up to the potential consequences of a change when there is really no reason to do so. Make this argument at the beginning of next spring and that’s fine, but not before game 2 of the World Series.
this is weird timing for this article. after utley and howard just hit 2 homeruns/2 doubles and a walk against LHP last night.
Took the words right out of my mouth. I understand hypothetically about this kind of thing, but geez, man. Between Neyer carping about Howard’s platoon splits and generally making the case he’s only an above average player, and this website wasting time like this, has anyone noticed Howard has been on fire during the playoffs?
Look, it does the sabermetric community a disservice when you ignore the obvious in order to show how smart you are. People stop taking you seriously when you write things like this, when you should be writing articles like: “WOW! Howard is really killing the ball against lefties this postseason! Didn’t see that coming, here’s how unprecidented it is!”
Throw us a bone, man.
There’s really no ‘perfect’ lineup for the Phils.
If you bat Werth 3rd and Utley 2nd, of course Howard will bat 4th.
That means either Ibanez will have to bat 6th (to avoid the lefty-lefty problem), but there’s really no 5th hitter.
Hitting Victorino 3rd really doesn’t make that much sense given Utley’s superior stats and relatively small platoon split.
According to The Book, the third spot should belong to either your four or fifth best hitter. It should also be held by a player who doesn’t hit into a large number of DP’s. My guess (without looking at the numbers) is that Victorino would fit well there, but could also be in the fifth spot (for your fourth best hitter). This is presuming you’d reluctantly bat Ibanez sixth to avoid the platoon problems.
Jack didn’t do all the projections for the lefty splits, that’s the only issue here. He’s stating it based on the fact that, in general, players hit worse against same-handed pitching. That opened up tons of “but Chase Utley and Ibanez don’t have platoon splits!” which don’t project anything and probably only use one or two seasons worth of stats.
This isn’t to say that Utley doesn’t hit well against lefties. But it doesn mean you can do better than what Manuel did.
Beimel/Morales and Kuo/Sherrill were more formidable LOOGY combos than Coke/Marte. Why would they change now?
The guy with the pocket protector says, “Shuffle the lineup.” The guy with the World Series ring says, “Not a prayer.”
Again, just because something worked in the past does not mean it will work in the future. A good reason to change now would be the Yankees are a hell of a lot better team than Colorado or LA. The Phils need every advantage they can get. It doesn’t take a rocket scientist to see that correlation doesn’t equal causation. This is common sense. Now, you can make a strong argument in favor of keeping things as they are (See phorever’s above), but your argument, as it stands, doesn’t hold up.
These moves don’t happen in a vaccum. This is the freakin world series where every move is over analyzed. There is no reason to shuffle the line up for a fractional percentage change in an expected outcome, expesically when that expectation can only be based on playing 100s of games, not 7 (or fewer). At this point you don’t mess with the players psyche, you don’t make moves that seem desparate. These players are not robots and this isn’t a computer simulation. If you want to go run a Monte Carlo simulation (I would make a joke about doing it in your basement, but I’m in my basement so i guess it would be a joke on me too) go for it, but I can’t see how this makes any difference over a very short series.
Bill… you are missing the point. You have to consider that it may have worked in the past for the appropriate reasons, basically, assuming Manuel is smart and it taks more than reading data to put together a World Series team, which he understands and is showing that he is good at.
One thing we aren’t considering, and bloggers rarely do, is that Utley and Ibanez (or maybe just Utley in this case) not only perform well against lefties, but also feel comfortable and take good swings against lefties, two things that say a lot more about their future success than their career LOOGY splits….
What’s odd is criticizing Utley’s position in the batting order, rather than the fact that Rollins is the leadoff hitter. The optimum lineup construction might be very different than what Manuel does. Utley’s best position is probably #2. But I think the most significant single change that he could make would be to move Rollins, not Utley. Lead off with Victorino, who has a much better OBP. Then Utley, a righty (Werth?), Howard, and so on.
Hey, if you all really want to be offended — I agree with Jack that Utley should hit #2. But only against LHP.
I think he should lead off against RHP.
Wait ’til you see where else I think dudes should be hitting for the Phillies.
http://www.drivelinemechanics.com/2009/10/29/1105531/world-series-batting-order
Whiny Phillies fans make for good entertainment.
Carlos Ruiz has an .862 OPS since the AS break, over .900 from August on. Pedro Feliz is .625 since the AS. Sample size, I know I know, but given these numbers, shouldn’t Ruiz be moved up at least to 8?
Ruiz is a much better hitter for any sample size “neuter.” Feliz is one of the worst hitters in baseball. Yes, he should be batting 9th unless it is better to bat your worst hitter in an AL lineup 8th, which it may be. I don’t know. It is a little surprising that Manuel would bat Ruiz 9th as most managers automatically bat their worst hitters 9th. I can’t imagine that he thinks that Feliz is better than Ruiz.
And yes, splitting up Utley and Howard, especially in the post-season (when opposing managers are constantly looking to bring in LOOGYs and ROOGYs) is probably the correct thing to do. But, there is NO way that Manuel is going to split them since they have been together for so long and the Phillies have been so successful. No way.
By the same token, Matsui and Cano should be split up as well, even though they both have small “true” splits I think. No reason not to have Swisher hit in between them.
he probably has some sort of belief that hitting after feliz he’ll get pitched to differently. since he’s had success batting after feliz.
CAN’T WAIT TO HEAR ABOUT STAIRS STARTING OVER FRANCISCO!!!
They’ve been batting them 3 and 4 all year. If they wanted to split them up and bat Howard 5th, then they’d wind up with Howard and Ibanez back-to back–unless they wanted to move Ibanez down to the 7 spot.
The Twins did the same thing, and often had their best three (not two) lefty hitters in a row (Mauer, Morneau, Kubel). The Phillies also did this with Utley, Howard, and Ibanez in many games. I think the slight platoon advantages might be countered by the number of extra at-bats these hitters get by being one spot up in the order.
I agree that it’s not ideal, but I think that the alternatives for both the Twins and the Phillies, at least in regard to their three and four hitters, are not very compelling. It’s not as if the Yankees have an absolute shutdown lefty — Marte and Coke may be “adequate,” but they’re not much more than that.