Because Someone Has to Bat Second
Mon, Feb 27, 2006
by Geoff Young
Last week in the discussion we were trying to figure out who would and/or should bat second for the Padres in 2006. Reader Steve pointed us to a projected batting order at Padres.com, which shows Mike Cameron in the #2 hole.
A debate then ensued over whether Cameron strikes out too much to bat so high in the order. Lance Richardson eschewed the notion, claiming that “the worst thing a #2 hitter can do is ground into a double play.” Readers Anthony and Pat suggested some criteria for a #2 hitter, most notably the ability to get on base. Seems reasonable.
My own personal opinion is that everyone not named Brian Giles is best suited to hitting #6. Unfortunately this isn’t a viable option, so I looked around to see what others who had studied such matters might have to say.
First stop was Dan Agonistes, who notes that lineup construction is “one of those things that your intuition tells you must be one way when hard data shows something else.” Beyond stating the conundrum quite eloquently, he also points to a thorough evaluation of lineups (potential number overload warning) done by Tom Ruane at Retrosheet.
Ruane presents a great deal of data showing various scenarios and then concludes that, “If anything, my approach shows that batting orders matter even less than people have believed.” From this follows the assertion that, “Since all but the most pathologically weird lineups produce just about the same number of runs, I might be inclined to select the lineup that makes the most intuitive sense to the players and fans. Simply put, it’s not worth all the fuss you’d cause trying to be clever with lineups.”
Heady stuff. Potentially useful, but not terribly satisfying. Sure seems like lineup construction should matter, doesn’t it?
Well, let’s assume it does. Now what?
A funny thing happened over the weekend. I discovered this cool lineup analysis tool over at Baseball Musings.
I couldn’t resist. I grabbed my copy of Ron Shandler’s Baseball Forecaster and plugged his projections for the Padres starting eight (assuming Mark Bellhorn at second base) into the tool. I used actual 2005 figures (.180 OBP/.190 SLG) for the pitcher’s spot.
According to the numbers I fed into the tool, the Padres’ optimal lineup is as follows (please be seated for this part):
- Brian Giles, rf
- Mike Piazza, c
- Mark Bellhorn, 2b
- Mike Cameron, cf
- Ryan Klesko, 1b
- Khalil Greene, ss
- Vinny Castilla, 3b
- pitcher
- Dave Roberts, lf
Aside from the ghastly notion of batting Bellhorn third and the unconventional employment of the pitcher’s spot at #8, I’m struck by the idea of Piazza in the #2 hole. Although he gets on base at a decent clip (or did prior to 2005), Piazza also is one of the slowest players in all of baseball and a perpetual threat to ground into a double play.
Here is what the Padres.com article referenced earlier projects:
- Dave Roberts, lf
- Mike Cameron, cf
- Brian Giles, rf
- Mike Piazza, c
- Ryan Klesko, 1b
- Khalil Greene, ss
- Vinny Castilla, 3b
- Mark Bellhorn, 2b
- pitcher
The Baseball Musings Lineup Analysis tool shows the above lineup scoring 4.777 runs per game, which comes out to 774 runs for the season. Using the optimal configuration, the Padres score 4.997 runs per game, or 810 for the season. In other words, tweaking the lineup in a way that will make sense to no-one should net the Pads an additional 36 runs over the course of 162 games. In a tightly contested division, that could be the difference between making the playoffs and going home early.
It’s worth noting here that we’re still talking theory. Even if it made sense, nobody would use the “optimized” lineup because it looks so damn weird and is the kind of thing that could get a guy fired, or worse. That said, it’s fascinating stuff, and Dan Scotto has given some guidelines for characteristics of each spot in the lineup over at Beyond the Boxscore. Here is what Scotto says about the #2 hole:
The 2-hitter should be the lineup’s most balanced hitter, a good combination of OBP and SLG. David Wright fits the bill here, as does the player I chose, Chase Utley. The first guy I thought of was Mike Lowell in his prime, when I looked at the results and coefficients.
On the Padres, is that really Piazza? I’m not so sure. A better choice might be Klesko. If there is one skill of his that hasn’t deteriorated due to age and injury, it’s the ability to reach base.
Klesko also doesn’t ground into many double plays — just 14 in 1000 plate appearances over the past two seasons. The only guys in the projected lineup to get doubled up at a lower rate during that time are Bellhorn and, you guessed it, Cameron.
Again, I don’t see any real solid candidates for the #2 hole on this team. Klesko is probably my favorite option. But if you’ve already got lefties leading off and batting third, then it makes some sense to break that up with a right-handed hitter. And on a team devoid of traditional #2 type hitters, Cameron seems as good a choice as any to me.
Then again, maybe it doesn’t matter.
Tunes: Pinback, Jason Falkner, The Chills, Loretta Lynn, Rush, Keali’i Reichel, Thingy, Modest Mouse, XTC, This Mortal Coil, Elliott Smith, Dead Can Dance, Live.
Picking up where last year's version left off, the Ducksnorts 2008 Baseball Annual provides in-depth analysis of and commentary on the San Diego Padres. Get your copy today.
February 27, 2006 at 8:36 am
Does Shandler project a comeback for Piazza and Bellhorn? I could see a lineup based on descending OBPs, but if that’s how it works I’d have thought Klesko would have been #2.
The 6-7-8 spots, now known as “Get this over with fast.” Like Don Coryell’s theory of defense applied to baseball.
February 27, 2006 at 10:59 am
That “optimized” lineup is a perfect example of why traditional baseball people have so much contempt for the sabermetric crowd. It’s possible to be so smart that you circle all the way back around to stupid. The #3 is the guy that doesn’t fit anywhere else? That makes no sense because if nothing else, the #3 hitter is going to get more at bats than the #4 through 9 hitters. Why do we want Bellhorn to get more at bats than Khalil or Dave Roberts?
I think the lineup does have some merit though, it would maximize Giles’ ABs and just sticking the pitcher in the 8 hole and Roberts 9th would probably raise all kinds of interesting situations.
It probably doesn’t make much difference in probability models but I think in the real world you would see some big differences because the opposing team would change their strategies based on the unusual lineup. For example, Giles would possibly see more fastballs as a leadoff hitter because he’d start every game with no one on base ahead of him.
February 27, 2006 at 11:21 am
You might want Bellhorn to get more at-bats than Greene if you think Greene is going to have problems keeping his OBP decent. It was 296 last year. Bellhorn stank in 2005 and his OBP was still better.
But Cameron, yeah. Same OBP, more power.
The pitcher hitting 8th does seem based on the low OBPs of Greene and Castilla. Dave Roberts getting to 1b with 2 outs doesn’t help.
February 27, 2006 at 11:41 am
TW: Shandler has Piazza at .345/.465 and Bellhorn at .338/.382, so yes, slight bouncebacks for both. Numbers are not adjusted for their move to Petco.
Also, I should note that pretty much all the “optimal” lineups have the pitcher batting eighth regardless of the OBP of surrounding hitters.
Anthony: Good point about traditional thought vs sabermetric thought. What would be really cool is if both sides could look at studies like these as starting points of ways to “think outside the box” rather than as some kind of pissing contest to prove who knows more about whatever.
That said, I agree that some of these conclusions are downright baffling. The thought of Bellhorn batting third terrifies me.
February 27, 2006 at 12:12 pm
What this study points out to me is that Giles better not get hurt. I bet we’re as dependent on Giles as SF is on Bonds.
February 27, 2006 at 12:19 pm
If 06 is anything like 05 (which its looking that way) the lineup will change on a nightly bases. Personaly I think Giles would be best fit in the #2 hole, (good at moving runners, getting on base and not hitting in to DP’s) but then there would be no one fit for the #3 hole. Maybe green will have a break out year at the plate and solve that problem.
February 27, 2006 at 12:36 pm
Traditional lineup says…
#1 Get on base…create Havoc (Think Ricky)
#2 Handles the bat well, does not K, can move runners, maybe can hit doubles
#3 Best hitter. Gets on base, has power, speed, bat control.
#4 Power…more power. Drive that ball over the wall. Scares other teams
#5 RBI guy. Picks-up an inning with a timely hit
#6 Think #2, but a bit more power
#7 Where no one else fits
#8 good on base…controls the bat well
Don’t try to get fancy…just get men on base with Giles/Piazza/Cameron coming up and we will score more runs than last year.
Oh yea, Barfield should end up lead-off by June with Roberts jettisoned and a LF named Johnson.
February 27, 2006 at 12:43 pm
Don’t see Barfield as a leadoff. Same goes for Greene. Heavily reliant on batting average. Juan Pierre has made that work some years, but they don’t have Pierre’s speed.
One good thing about having two possible 20 HR hitters at second and short, we can afford to carry a less-powerful, higher-OBP hitter at more traditional power positions. Too bad those Sean Burroughs - Wade Boggs comparisons never panned out.
February 27, 2006 at 2:08 pm
My theory about lineups is similar to my theory about closers. Straying too far from mainstream would result in mental mutany within the players’ ranks. Keep the hitting speed guy at the top of the order (but only if he can also get on base a decent clip) bunch the power guys 3-5 and move downhill from there…
For what it’s worth, I thought Loretta was the IDEAL 2nd hitter…
February 27, 2006 at 2:19 pm
I’ll go off-color here. Traditional lineup for the last forty years makes the leadoff hitter “fast dark-skinned guy.” Ranges from Ricky to Tom Goodwin.
Peter may be right, the distraction from such a weird lineup might actually give back some of those runs. It might work better in a sleepy burg like San Diego, but even here the media would be buzzing all the time.
So Dave, what does it feel like to hit BEHIND Chris Young? Do you take a pitch to give him a chance to steal?
Rhino, does it bother you to see Mark Bellhorn’s name in the third spot when you have, oh, 200 more career HR?
February 27, 2006 at 3:22 pm
Here’s a fun toy too:
http://home.comcast.net/~stein.....ineup.html
from BTF, I think.
Remember when LaRussa bat his pitchers eighth in the McGwire era? Even he succumbed back to the pressure of traditional lineup.
February 27, 2006 at 7:06 pm
That tool on Baseball Musings is worthless. It relies on some very questionable findings and applies them in ways that I really don’t think you can do. I was actually pretty surprised that Pinto put it up on his site. We discussed this on Primer:
http://www.baseballthinkfactor.....ion/36584/
“That “optimized” lineup is a perfect example of why traditional baseball people have so much contempt for the sabermetric crowd. It’s possible to be so smart that you circle all the way back around to stupid.”
Agreed. That’s what’s so damaging about bad analysis. People who are knowledgeable on statistical analysis can look at the methodology and discredit it. But outsiders don’t really see the distinction; they just see all these stats guys that want to bat Piazza #2 and the pitcher #8, and then paint broadly with that brush.
February 27, 2006 at 8:56 pm
Sometimes I think a lot of the sabermetric types are so hellbent on overturning conventional wisdom that they contruct their studies to generate a desired outcome, which is just as bad as blindly following conventional wisdom. Rather than starting by asking if a commonly held belief is true, they assume it’s wrong and set out to prove it.
Don’t get me wrong, I think the flood of analysis we’ve seen in the last few years is great and has led to a lot of new insights but sometimes it seems the people running these studies don’t actually watch any baseball games.
February 27, 2006 at 9:17 pm
Cy’s goal was to see if there seemed to be any difference in the valuation of obp relative to slg between the different lineup positions. He wasn’t seeking to determine “ideal” lineups. He clearly stated that there were issues with drawing concrete conclusions from the numbers. I don’t see the need to be a jackass in response to his trying to answer a question.
February 27, 2006 at 9:44 pm
TW: Agreed about Giles. Without him, we’re hosed.
Didi: Yes, I do remember LaRussa batting the pitcher in the #8 spot. Didn’t McKeon do that with Dontrelle Willis in Florida last year? In fact, IIRC, Willis actually hit seventh once or twice.
Vinay: This is a danger of trying to be transparent with these studies. The upside is it allows folks to build on each other’s work. The downside is it allows other folks to see research that may or may not be complete.
Anthony: There probably is some of that out there, but I think a lot of people genuinely are seeking answers to difficult questions.
February 27, 2006 at 9:57 pm
Richard: you’re exactly correct. Cy even commented in a different thread about the lineup optimizer that it was extending his work in a way that didn’t really work. It’s a textbook example of conusing correlation with causation.
Re Dontrelle: depending on your team, he’s good enough to bat 7th. Come on, who would you up in a big situation, Dontrelle or Dave Ross?
February 27, 2006 at 10:00 pm
And I should say, I didn’t mean to come off as much of a jackass as I did in my earlier post today. Guess I should read my comments before hitting “Post”. My apologies.
February 28, 2006 at 9:06 am
I remember hearing a story about some light-hitting middle-infielder who was traded and went to his new manager and said, “Whatever you do is ok, just don’t bat me 8th.”
So his manager batted him ninth; behind the pitcher. After the game, the player went to his manager and said, “Ok, I’ll bat 8th…”
February 28, 2006 at 9:19 am
Vinay: Wasn’t calling you a jackass. I had just finished reading some of the btf comments and was referring to that. I’m sure you know which ones I’m talking about.
February 28, 2006 at 11:04 am
Sometimes when I read BTF I wonder if some of the commenters there even like baseball. There’s a tremendous amount of valuable research presented there and a few people who seem to hate everything.
I just wanted to add that I think this type of computer analysis of lineup construction is something that could make a real difference. Unfortunately if the above lineup really is the optimal one it has zero chance of being adopted over the course of a full season. The first time the #3 hitter strikes out with two on in the first inning is the last time any manager would use that lineup.
February 28, 2006 at 11:37 am
Oh,yeah. Dontrelle did hit 7th a few times. Shoot, he’s a stud with the bat. I’m suprised Bobby Cox never hit Hampton 8th in Atlanta.
Well, if the #3 hitter bunts the runners over everytime, would that be a bad thing?
Here’s praying that Giles stays healthy for the season and that he stays away from Donovan McNabb’s mom and her Chunky Campbell Soups. That’s what happened to McNabb. My story, I’m sticking to it.
February 28, 2006 at 1:16 pm
Steve - Bobby Hill is ideal for the 2-hole, as he fits your description perfectly! He bunts, handles the bat well, puts the ball in play, has decent speed and he switch hits. Bellhorn is not the answer and Barfield needs at least a half season at the bottom of the order. Hill could split time between 2nd and 3rd base, easing Barfield into the starting role and giving Vinny a much needed day off every week.
February 28, 2006 at 4:14 pm
Didi - re: “if the #3 hitter bunts the runners over everytime, would that be a bad thing?” … I suspect so … I don’t know where to find “expected outcomes”, but it’s my understanding that 1st & 2nd with 0 outs has a higher “expected outcome” than 2nd & 3rd with 1 out … so I think actively wanting to make that transition is “a bad thing”.
February 28, 2006 at 4:16 pm
Hey NC … in my book, *YOU* were a *STUD* … http://www.retrosheet.org/boxe.....TL1972.htm
:-) 
February 28, 2006 at 4:38 pm
LM, and Bobby Hill can bunt.
All joking aside, where is Bobby Hill in the Padres’ plan? 2B and 3B or UTIL?
I’m playing him as 2B on my Padres game from MVP 2004 after realigning the team with the 2006 roster. Not bad so far.
I just hope Blum won’t be taking much time at 2B.
February 28, 2006 at 7:18 pm
Ah, here it is … http://www.nssl.noaa.gov/users.....druns.html … found via googling for “expected runs” (not “expected outcomes”) …
This tells me that 1st & 2nd with 0 outs is worth 1.50 runs, and is worth 1+ runs 63.6% of time … whereas 2nd & 3rd with 1 out is worth only 1.40 runs but is worth 1+ runs 68.6% of the time … so I guess that says that if you can guarantee a successful bunt and you need just 1 run, then you improve your chances by 5% … doesn’t seem likely that the *BEST* bunter is good enough to not eat the whole 5% bump …
February 28, 2006 at 11:38 pm
Ah but here we see the fallacy of just pulling out the average numbers. What is the expected runs in such a situation when the proposed #3 hitter, the guy who doesn’t fit anywhere else, is at bat? Those numbers are an average of all games, correct?
I’m not really being serious, I think we all know those numbers aren’t accurate for every situation. It kind of makes me sad though to see that “small ball” has pretty much been discredited in the modern game, it was a lot of fun to watch.
March 1, 2006 at 9:02 am
Good information, LM.
I think the reality is depending on who’s bunting and how good the baserunner at 3B is and how good the defenders are. Baseball wouldn’t be interesting if everybody plays using the same stategy.
Anthony, I agree with you. The excitement of a successful squeeze play is amazing. Those two squeeze plays I watched that David Eckstein did last season was exciting. I supposed those wouldn’t have worked had the defenders were expecting them and Eckstein was a terrible bunter.
Speaking of bunting, I just read that Finley won the bunting competition. That old fox still got it.
March 1, 2006 at 9:21 am
One of my absolute favorite moments of Padres baseball in recent years was when Eric Owens did a straight steal of home off Brett Tomko. It may not always be the most effective strategy, but I have to admit that small ball can be very exciting.
March 2, 2006 at 10:25 am
http://www.hardballtimes.com/m.....g-lineups/