It looks like you left out a column when you broke things down per pitch. Your pies are less than 100%. Non-foul contact will complete it, I think.

Phew!  I had been worried we might have another A-Fraud on our hands.

For some reason, Cano has been at his worst with the bases loaded, but I guess the reason(s) why are similarly difficult to tease out, and the sample size is very small.

I would still bat him 5th and hope that his splits normalize over time.

Yeah, I forgot to include in play, out and in play, no out columns.  Also looks like they have other types of fouls that I forgot to include, foul tips and foul bunts are handled separately.  I’ll update the charts in a bit.

[3] I would bat Granderson 5th against RHP and Posada 5th against LHP.

Yeah, I was thinking Granderson 5th vs. RHP as well.

I view batting Cano 5th as a show of confidence by Girardi in the hope that it would pay dividends in the mid- to long-term even if there is a price to pay in the short-term.

Guys, don’t you know?  CFers who steal 20 bases can’t bat 5th!  They must bat 1,2 or 8, 9.  It just <u>looks</u> wrong otherwise.  Second basemen can’t bat 5th either.  They must bat 1st, 7th, or 8th, even if they have poor on base skills. 

No, no matter who your players all, all lineups must be:

CF or 2b
2b or SS
1b
RF
DH
LF or 3b
3B or C
C or 2b
SS or CF

Generally, teams are allowed <U>one</U> exception, when they have a great hitter at a premium defensive position.  Teams that have more than one exception may win championships, but their lineups <u>look</u> wrong.

That’s why it’s so important to have Gardner play CF and Granderson play LF.  Not because of the defensive gain, but because when Gardner, CF bats 9th it will <u>look</u> right.

(Can you tell classes started today?)

[7] Or you put even more pressure on him and exacerbate his woes.

I’d go Jeter/NJ/Tex/ARod/Granderson/Posada/Cano/Swisher/Gardner vs. RHP
and
Jeter/NJ/Tex/A-Rod/Posada/Swisher/Granderson/Cano/4th OF vs. LHP

For whatever reason, Posada’s OPS batting 5th is .825 (1757 PA), and batting 6th it’s .919 (2240 PA). That’s approaching the 2000 PA magic number and is almost 100 points of OPS difference.

You can probably already tell this by looking at the numbers, but that type of batted ball distribution is pretty similar in both cases, and does not support a BABIP difference of .052.

Completely non-scientific, but it certainly seems that there have been numerous times (with men on base) where Cano has sent a LD into the gap and an OF made a great play on it.  Or Cano hits an at-em ball that turns into an out or a DP or something.  Of course, I like Cano so my memory is probably a bit biased, and he may very well hit more than his share of weak dribblers w/ runners on.  But I really don’t think there is anything about him to make me believe he’s unable to perform with men on base.

I know some people are big believers that batting order doesn’t matter (I’m looking at you, rilkefan), and in the abstract it seems that is correct—the difference between “optimal” and sub-optimal” order doesn’t look like all that much using just career numbers.  But Posada looks like he has a clear advantage in the 6-hole.  I wonder if our model modern managers are taking this into account, and I wonder if anyone has played around with lineup effects that examine this kind of data.  I can’t think of any intuitive reason why production would be so significantly different…but I’m sure someone else can.

I know some people are big believers that batting order doesn’t matter (I’m looking at you, rilkefan), and in the abstract it seems that is correct—the difference between “optimal” and sub-optimal” order doesn’t look like all that much using just career numbers

In the aggregate, I’d say so, but I think you need to set your lineup to try and minimize the platoon advantage as much as possible, and that can almost certainly have an effect, particularly in high-leverage situations.

In Posada’s case, he has hit .268/.378.474 vs. RHP and .299/.381/.496 vs. LHP in his career.  What I am wondering is if he faced a higher percentage of LHP when batting sixth, although intuitively I’d guess the opposite, that he was more likely to hit higher in the lineup vs. LHP because it meant that people like O’Neill and Tino were not as good vs. LHP.  But I’d check that before assuming the fifth/sixth splits are indicative of anything.

Or you put even more pressure on him and exacerbate his woes.

I would take that risk until the AS break.

I think this type of an analysis is likely to be inconclusive, and I am not sure the data exists to find anything conclusively.

Watching Cano, my general impression was (which of course could be POOMA or self-confirming), he was much more likely to swing at low-quality pitches with men-on-base. As a result, he is more often than not getting into quick 2-strike counts, and he lacks the pitch recognition skills to lay off balls when he is in a 2-strike count. Because he has such awesome contact skills, those balls that he chases more often than not ends up being foul. But still, he ends up with pitches that he cannot drive with authority. Hit/Fx data, if and when available can perhaps go a little deeper in answering this question.

I think a box diagram of Cano’s contact tendencies could be somewhat informative. What is the spatial distribution of pitches he is getting when there are men-on-base versus when there are none? Is this knowable from available data?

“For whatever reason, Posada’s OPS batting 5th is .825 (1757 PA), and batting 6th it’s .919 (2240 PA). That’s approaching the 2000 PA magic number and is almost 100 points of OPS difference.”

If every time through the lineup someone flipped a coin to decide where Posada would hit, or even before each game, I’d believe there was something interesting going on, but I bet it’s just that during long periods when he was hitting well he happened to bat 6th, or he spent a lot of time batting 5th during times when league batting was relatively weaker and his wRC+ would not show this effect.

OK, revised charts are posted including the in play columns and HBP columns.  It changes the conclusion that Cano is more likely to swing when runners are on base.

I think a box diagram of Cano’s contact tendencies could be somewhat informative. What is the spatial distribution of pitches he is getting when there are men-on-base versus when there are none? Is this knowable from available data?

Yeah, it’s knowable.  I just need to see how hard it is to plot it.

yeah, there are all sorts of ways for a player’s lineup position split to get biased. It’s a little funny that manager’s have been putting Posada higher when he has been hitting worse, but the Yankee lineup overall may have been better when Posada was doing well, and then there are righty/lefty reasons for the lineup.

I wonder if our model modern managers are taking this into account

Managers might not be, but The Magic Binder is for sure.

[16] Well, if you look at the 2 kinds of OPS+ at b-ref, Posada has hit better in the 6 spot almost every season, except 2001, 2004, and 2008, certainly for his career; OPS+ numbers should smooth out the context.  As for

“long periods when he was hitting well he happened to bat 6th”

—that’s just the inverse of “he hit well for long periods when he was 6th in the order.”  I mean maybe you can come up with some explanation of why the numbers aren’t significant, but don’t you think that given the large sample sizes we ought to give some credence to the notion that there’s some kind of batting order effect on Posada at least?

Incidentally, I had always thought Jeter hit much better leading off, but there’s not a lot to choose between his numbers for 1/2/3, though many fewer PAs at 3.

‘—that’s just the inverse of “he hit well for long periods when he was 6th in the order.”’

This is just correlation/causation though - it’s non-Abelian or something.

“I mean maybe you can come up with some explanation of why the numbers aren’t significant, but don’t you think that given the large sample sizes we ought to give some credence to the notion that there’s some kind of batting order effect on Posada at least?”

Not at all, at least until the gross selection biases are controlled, and then I’m still going to be highly skeptical of the claim that anyone radically changes effectiveness between 5th and 6th - maybe leading off and batting second.  It’s just not a safe idea to look for deviations from expected behavior, find a single example, and say “Eureka” without doing a lot of work to eliminate simpler explanations.

Zoosk?

For the stats or dating inclined, this analysis is pretty interesting.  More here.

I don’t understand the “this guy hits better when he is in this spot in the order versus this spot in the order” argument. I mean, take Posada for example, he only hits 5th or 6th in the lineup the first time through the order. After that he randomly bats lead-off through ninth in the order depending on what occurred in that first inning.

So, is the argument that mentally Posada feels more comfortable when he sees his name slotted 6th in the lineup card instead of 5th? Doesn’t that make about as much sense as oh let’s say batting ARod 8th instead of 4th in an elimination playoff game because it would take the pressure off him?

Maybe that is a simplistic way to look at it, but if Posada is the best remaining hitter (after taking Jeter, OBP Jesus, Teix, and ARod out of the equation) against LHP, then I want him in the 5th spot.

Also, I can’t see any argument (other than the brilliant tongue-in-cheek “a CF batting 5th looks wrong” argument) for anyone other than Granderson batting 5th against RHP.

“After that he randomly bats lead-off through ninth in the order depending on what occurred in that first inning.”

There is surely some difference in the distribution of men on base he sees batting 5th or 6th, as well as how many times he comes up.  Maybe he’s really tired his last AB, being a catcher and all - one could look at his splits when DHing if only there were more data, dammit.

It would be interesting to see a graph of batting 5th rate per say week vs time.

“here’s a look at the pitch locations against Cano with men on vs. with the bases empty”

Could you calculate the centroid and the 2-d rms, or even something more baseball relevant?

What about the fb etc %?  The pull %? Splits for LHP vs RHP?

“This is just correlation/causation though”

Yes, my point exactly.  I don’t object to the charge, just the way it was put.

“It’s just not a safe idea to look for deviations from expected behavior, find a single example, and say “Eureka” without doing a lot of work to eliminate simpler explanations.”

True.  I’m hoping someone smarter and with more time will do this…

“So, is the argument that mentally Posada feels more comfortable when he sees his name slotted 6th in the lineup card instead of 5th?”

Maybe.  Chicken probably didn’t make Wade Boggs a better hitter, but I venture to say if forced him to eat steak it might have thrown him off.  I guess I should have been clear, my guess is that batting order effects for individual players, if identifiable (except maybe for leadoff hitters), ARE psychological.

“Yes, my point exactly.  I don’t object to the charge, just the way it was put.”

But I think your way is biasing.

“for long periods when he was hitting well disco happened to be popular” vs “he hit well for long periods when disco was popular.”

“True.  I’m hoping someone smarter and with more time will do this…”

Assuming he didn’t oscillate, it shouldn’t be that hard from baseball reference, ignoring slight inconsistencies.

I think to really get into it, you’d want to look at a larger universe of people—20, 30, 40 hitters with say >1500 PAs at multiple spots in the order and see what’s up with them…that’s what seems time consuming.

Even more time-consuming than my posts.

“I think to really get into it”

Oh, sure, that requires a db and some easy if maybe annoying sql or the equivalent.  But it would also involve knowing all that biasing stuff I’m concerned about.  Explaining away the Posada observation would allow us to go back to the “nothing there” default.

SG is just supremely awesome. I will look forward to Cano having better luck this year. I am not convinced that the OPS difference is a mere artifact of luck, but I have been wrong before, and will be again.

Plus, Yankees won a WS last year! Who cares if Cano could not get a few runners over?

Could you calculate the centroid and the 2-d rms, or even something more baseball relevant?

<u>Bases empty</u>
Centroids
PX: -0.31876 (in terms of feet from center of plate.  A negative value indicates left when facing the pitcher)
PZ: 2.475651 (in terms of feet from the ground)
rms X: 0.941740775
rms z: 0.907928638

<u>Runners on</u>
Centroids
PX: -0.32932
PZ: 2.492107
rms X: 1.098844028
rms Z: 0.930712229

What about the fb etc %?  The pull %? Splits for LHP vs RHP?

FB?  Flyball, fastball?  As far as pull% and splits, that’s not really doable from the Pitch F/X data, it doesn’t record pitcher handedness or hit location as far as I can tell.

I am not convinced that the OPS difference is a mere artifact of luck, but I have been wrong before, and will be again.

No, I wouldn’t call it luck either.  It’s actually happened, but I guess what this tells us is that there’s a good chance that we shouldn’t expect it to continue.

All this talk about who should bat where and not one mention of A-Rod batting 3rd.  RLYW just made baby Jesus cry.

[31] Shouldn’t ARod technically lead-off to maximize the amounts of PA? Or is there a certain amount of extra runs scored by going with the highest OBP guy(s) followed by your best overall hitter?

Has anyone done any simulations to see if a team would score more runs by going with an order based on highest to lowest OPS or based on highest to lowest OBP or based on highest to lowest wOBA than compared to say a typical lineup?

Who cares if Cano could not get a few runners over?

Joe Morgan, John Kruk, <strike>Steve Phillips</strike>, Peter Gammons, and President Skroob.

[32] - Extensive research.  See David Pinto’s lineup simulator and this post for an intro.

[34] Thanks for the link. Interesting stuff.

Also, sd what is your logic for ARod batting 3rd?

[23] “tongue-in-cheek”?  Not a bit.  This is clear baseball strategy.  2b Jose Vizcaino batted leadoff for the Yankees in the world series.  Why?  Because it <u>looked</u> right.  2b Luis Sojo batted 2nd for the Yankees in the WS.  Why?  Because it <u>looked</u> right.  IN 2002, 2b Alfonso Soriano hit 39 HR and had a .332 on base percentage.  Naturally, he batted leadoff all year.  Why?  Because it <u>looked</u> right.  It was like Joe Torre was a little kid making up stories about pretend baseball teams with scrappy little second basemen who always hit at the top of the order.  He won championships that way: genius.

On a tangential note, in 1961 Bobby Richardson’s on base was .316 and his OPS+ was 67.  He batted leadoff all year in front of 240 HRs.  He also finished 24th in the MVP balloting.

look at the 2 kinds of OPS+ at b-ref

Which two kinds are those?  And what “context” are you expecting those to even out?  As I understand it, tOPS+ would just compare the Posada’s raw OPS in that lineup spot to his overall OPS.  No other adjustments are made.  sOPS+ compares the hiss raw OPS in a given lineup spot to the league average OPS for that lineup spot.  Again, I don’t think any other adjustments are made.

Sorry to have to break this to you, Froggie, but Bobby Richardson SLUGGED .316 in 1961.  His OBP was .295 (in a league that had an OBP of .332).  Thank goodness he had “only” 534 of his 704 PA batting leadoff.

MC wit’ da internet skillz!  HTML in da house!

“FB?  Flyball, fastball?”

Oops - fastball.  I wonder if he sees more or less off-speed stuff.

Thanks for the centroid data - looks like pitchers are expanding the zone a bit with men on - one would have to compare similar batters to see if this is something aimed at Cano.

Olney thinks that the MLBPA will sell out amateur players and agree to a slotting system in the draft in exchange for more money being spent on ML players, although he offers no specificity as to how that would be accomplished.

Any such move would not be good for the NYY…or the advancement of free market principles.

Does anyone else find the constant use of the word ‘sweepstakes’ or ‘derby’ annoying when it comes to a potential player acquisition? Especially if that player is a 39 year old Jim Edmonds.

[42] Yeah, at best, it’s a consolation prize. At worst, an admission of defeat in one’s pursuit of more appealing options.

[41] As soon as you have the draft you’ve abandoned free market principles.  I think either go with the slotting system, OR allow trade of picks.  If you believe that the best amateur talent should go to the “worst” teams, then a slotting system makes sense.  The worst teams are often the poorest, and having a top-pick doesn’t do them any good if they can’t afford to sign them.

Alternatively, if they can trade the pick they may get more value than drafting a “lesser” player with that spot.  E.g. KC may draft fifth but not willing to pay $8M (maybe their whole amateur budget) for the all-world SS who *should* be drafted there.  Yankees drafting 30th want the SS, and are willing to pay, but leery the Red Sox or Mets would draft same player.  So boom, Yankees get 5th pick, Royals maybe get the 30th and 60th picks, everyone happy.

[42] Yes!

[44] I think before you address problems with the draft, you have to address the problems with those 4 or 5 special teams that think it’s perfectly acceptable to field a team with a $40-$50M payroll. Then we can see what will happen in the draft and go from there, I think.

[44] You are correct that a draft is inimical to free market principles, but the absence of a slotting system mitigates it to some degree. Under the present system, the worst teams have the opportunity to draft top talent, but the contract they have to offer is subject to the pressures of the market because they understand that if they don’t sufficiently pay a player, they may not be able to sign him.  As a result, the system promotes competitive balance but not at the expense of the finances of the amateur player.

So boom, Yankees get 5th pick, Royals maybe get the 30th and 60th picks, everyone happy.

You really think so?  More like everyone bitches when that 5th pick turns into a franchise player within three years while #30 has TJS and #60 ends up in rehab.

I think that trading draft picks would be an unmitigated disaster in MLB.  There’s just way too much uncertainty in player development to expect that it would serve the useful functions that it does in other professional sports, where the value of a particular pick is much easier to quantify with at least some modicum of accuracy and precision.

Does anyone else find the constant use of the word ‘sweepstakes’ or ‘derby’ annoying when it comes to a potential player acquisition? Especially if that player is a 39 year old Jim Edmonds.

I agree. It’s not really a sweepstakes if winning it makes you feel like you have H1N1.

[45] I don’t see why you couldn’t do it in tandem.  E.g. the PA says, “okay, we’ll give you [hard-slotting|trading-picks], if you put in some mechanism to enforce a (higher) payroll floor”.

[46] I would prefer trading of picks to a slotting system, but I do have my doubts that the current system is “working”.

[47]  Well, everyone would be happy the day after the draft.  Years down the road who knows.  It’s also likely the 5th pick develops a drinking problem and never makes the majors, while the 30th is a mutliple Cy Young winner and the 60th is a stud OF. 

I don’t think MLB players are any more difficult to draft than the NBA or NFL, and I don’t have it handy but I’m pretty sure significant work has gone into researching this, and that draft position in the majors is a pretty good predictor of future success.  What it really is, is allowing teams to better manage risk.  If KC (in my example) can trade down and get two picks (would more likely be 30th and 75th b/c of supplementals) with lower risk, they may be better off.  Sure, they may miss out on on the franchise player, but they also aren’t required to shell out big $$ for a player who may bomb. 

And, why do they need to trade pick-for-pick (I’d allow pick-for-player swaps as well)?  Instead of NY’s 30th, they could take maybe Kevin Russo instead, who is VERY safe, and would likely be a starter next year for KC.  They get cheap ML talent for an expensive bet who may or may not pan out.

[49] You could.. but wouldn’t you want to see how the market would react to 1 regulation before you implement 2?