THE BOOK--Playing The Percentages In Baseball

Perhaps I’m biased as a guy who watches so many games in one of the least symmetrical stadiums in baseball, but I think tailoring your team to fit your specific ballpark is absolutely the right idea. 

Using the Mariners as the example, because of familiarity:

Jose Lopez is a right-handed, extreme pull power guy (61 of his 65 career home runs have been pulled).  Safeco Field is extremely pitcher friendly on balls hit to left field.  Not surprisingly, Lopez has really large home/road splits. 

Safeco is actually hitter friendly to right field, due to the short porch down the line and the way the air circulates in the stadium.  Again, not surprisingly, Ichiro shows no significant home/road split, performing just as well at Safeco as he does on the road. 

The optimal strategy for the Mariners, then, is to build a team that hits the ball to right field and pitchers who force the opponents to hit the ball to left field.  By building a roster that plays into the strengths of the home park, the Mariners will have a pretty large home field advantage simply due to the park effects and how their roster matches up with how the ball carries in Seattle.

I don’t know enough about Citi Park to know if there’s a similar optimal strategy with the Mets, but I certainly believe that there are scenarios where exploiting the park effects of your home stadium is a good roster building choice.

I’ve got to believe the Red Sox do this too.  Their home/road splits in terms of W/L are outrageous, and I thought I read a while back that they are greater than they were in the 70’s or 80’s or even the 90’s.  Since the park hasn’t changed, and to be honest the crowd isn’t *that* much better than it was back decades earlier, I feel like they are making a conscious effort to get guys like Lowell - pull-happy wall-slappers, and guys like Ortiz who have enough power to poke it over the wall when they go opposite-field.

Yes, OF COURSE that is true.  That is not the question.  The question is whether pitching, speed, and defense is properly tailoring your team to a large park (as opposed to a small park).  I have already discussed in another thread (the Dejesus one) the idea of needing or leveraging good defensive players in a large park.  I believe there is no evidence that that is true - in fact, there is contrary evidence.

It is not intuitive at all to me that better pitching (at the expense of offense of course) and more speed (at the expense of something else) is better in a large ballpark than a small one. 

That is the question I am asking…

It’s basically the same idea. 

If you have a park like Petco that turns everything under 400 feet from home runs into doubles or outs, then there’s a diminished value on all the 370-400 foot fly balls.  Therefore, a team that played in a park where those balls would clear the fences would value the type of player that hits the ball 370+ feet more than you do, because they’ll get more production from the same player. 

So, if you have a big park that is presumably anti-home run, you will value power less than other teams and therefore spend less on it.  If you’re not spending on power, what’s left to spend on? Speed, defense, and pitching.

What Dave says makes perfect sense to me when you’re talking about a park (like Safeco) that severely discriminates between left- and right-handed batters; of course left-handed power hitters will do better there than righty ones will. And it would make sense that Fenway would be the mirror image of that.

But I don’t think it extends to what the Mets are saying, namely that big park = speed and defense. If the park is going to affect hitters and pitchers of both handednesses pretty much equally, I think the best thing to do is just ignore it and build the best team you can, because (I think, but I could be wildly wrong) a park that depresses power across the board is generally going to depress Adrian Gonzalez’s and David Eckstein’s in more or less equal proportion.

So a more powerful lineup in a big park will still outslug the other guys in that park (and on the road, of course), regardless of the fact that they’ll hit fewer total homers than they would in a small park, and a great defense in a small park will still allow fewer hits.

I don’t know if that makes sense.

Bill, yes that does make sense.  I don’t think there is any reason to believe that power hitters benefit more from a small park. It might be the opposite. That is why I asked whether there is any evidence one way or the other. All the speculation in the world won’t answer the question, I don’t think…

I don’t know what Speed and Pitching has anything to do with each other, except possibly that faster players mean better defense (is that true BTW?) Offensive spped and good pitchers don’t seem to have a multiplicative effect at all.

As for Pitching and Defense going hand in hand, I always thought it was low-strikeout guys and defense that should go together.  A good infield defense seems wasted on guys like Randy Johnson in his prime...and a good outfield defense may not mean much either since he’s striking so many batters out.  But a low Walk and low K pitcher probably appreciates a better defense more since the ball will be in play more.  So it isn’t necessarily good or bad pitchers that need good defenses, but rather pitchers that have fewer walks+Ks need better defenses.

I think there’s a decent chance that Bill is “wildy wrong” (sorry) about large parks depressing power equally for all batters.  Greg R probably has the best insight there.

I think Dave has the right idea in that if power is depressed, the other facets of the game become more valuable.  Say Aaron Hill loses 10 HR after moving to a bigger park - Chone Figgins probably still puts up the same numbers. 

As for whether OF speed/defense is more valuable in a larger park, intuitively more OF flyballs staying in the park means more chances for the fielders, and so more value from good fielders.  I’d be interested to see the numbers, though.

Boston might be an interesting test, with the Monster (click my name for an old Chone link) if nobody feels like running something really detailed.

It’s intuitive to me that in a large ballpark OF defense would matter more.

I think it’s also true that in a lower scoring environment things like stealing bases and moving the runners over matter more.

I’ve also read (but forget where) that true power hitters hit a far greater % of their HR’s to center than slap hitters who, when they do hit HR’s, hit them to the corners (this is also pretty intuitive, I think). 

So, if you had a ballpark that was particularly deep to center and with a large OF I think that might favor speed guys over power guys.  A ballpark that was relatively deeper in the corners might favor power guys however.

Anyway, it probably wouldn’t be that hard to build a speed/defense team in diamond mind and simulate a numbers of season in a large park and a number of season in a small park and see if they play significantly better in the large park.  Not that this would be the definitive test.

Large parks will favour flyball pitchers over groundball pitchers, and will presumably put a premium on outfield defense.

If speed is at a premium in a lower run-environment, and outfield defense is at a premium with deeper fences, then those are correlated.  Good OFs tend to be either speed-and-defense (and often high BA) or slow sluggers.  In a larger park, I suspect that the speed-and-defense guys are at a premium over the sluggers.

IF defense, especially if you’re concentrating on FB pitchers rather than GB pitchers, is not going to be a high priority - which means that the IF is where you want your hitters.  I suggest that 3B should be a big hitting position in such a team.  Oh, look, the Mets got something right!

This reminds me of an article Brian Cartwright had on StatSpeak (click my name for the link) that found that historically, players who hit more home runs were proportionally less affected by the park they play in as far as how many home runs they hit.  Tango’s remarks on park factors for home runs and using addition instead of multiplication (Brian also mentions this in the comments on the article) and his example of the effects of AT&T on Bonds seem to imply something similar.  It could be that you can also leverage a park that suppresses home runs by building a really powerful lineup that can still hit home runs there and that intentionally going after low power guys means you are just cutting into your home run production proportionally more than if you built a normal or a powerful lineup.

I certainly wouldn’t dismiss the possibility that building against power actually hurts you in that park without having more information and research on it.  It could be that both outfield range and power at the plate are ways to leverage the park, and that you have to consider both and strike an optimal balance based on what is available, or it could be that neither one makes enough difference to consider over other factors.

This also reminds me of how so many people are convinced that speedy slap hitters are vastly improved on astroturf.

I think the correlation (or should I say, negative correlation) between outfield range and power that you frequently see at the corner outfield positions is a factor here.  In Philly, you don’t suffer as much from letting Pat Burrell play left field, compared to letting him play it in Seattle or Queens or Denver.  So there are some choices there…

Also, I’ve pondered elsewhere, is it better to reinforce or mitigate your park’s tendency?  I think of the Red Sox, who usually had great hitting, but only when their pitching improved did they win.  But it’s not a perfect example, as their hitting has usually been good along with the pitching when they won…

A great topic for discussion…

There’s no such thing as “across the board”.

I’ll trot out my two big examples:

1. Barry Bonds v other LHH at 3Com.  While in SF, Barry hit as many HR at home as he did on the road.  LHH hit one-third fewer HR in SF than away from SF.

2. Dante Bichette doubled his HR rate in Colorado.  Larry Walker and Todd Helton increased by (IIRC) 40%.  Juan Pierre was flat.

There is no surprise here.

If I had to guess, it’s the “JE” (just enough) guys that are most at risk (most to leverage).  Basically, those guys with “gap power”.

Regardless, I would not make any kind of blanket conclusion on the matter.

"In Philly, you don’t suffer as much from letting Pat Burrell play left field, compared to letting him play it in Seattle or Queens or Denver.”

That is probably not true Greg.  Why do you think that it is?

Somewhere in the archives here (I think) I have some numbers…

Maybe if you amplified the difference you could see the effect more clearly.

What if you played in a park where the fences were 500ft?

99.9% of all HR’s would be inside-the-park which would certainly favor the fast guys. And those fast guys would have many more chances to stretch doubles in triples and triples into ItP HRs.

It would also help immensely to have fast OFers since a high fly that goes 420’ would be easily caught by a fast CFer. That’s a homerun in almost any current park - so there’s a huge value in being able to catch those balls. Any time you see a fielder wait by the wall to try and rob a HR that was an easily catchable ball if the fence wasn’t there.

So is there any difference in a park that is a few percentage points bigger? I don’t know, of course, but I think it is reasonable to think so. Whether it is significant is another story.

Knowing how to leverage a park where it is virtually impossible for almost anyone to hit a home run doesn’t necessarily tell you anything about how to leverage a park where it is just more difficult to hit home runs.  Even so, I don’t think you can just assume that speed holds more value than power in such a park.  Even if a hitter can’t hit it over the fence, there is still a ton of value to being able to drive the ball into the depths of an outfield that massive.  Either the fielders play way back and you get a ton of hits in front of them, or you drive it over their heads and it keeps on going so they have to chase after it and then make a very long throw.  It could be that most of the inside the park home runs and triples come from players who can hit the ball over the outfielders heads and make them chase it to a 500 foot wall and then have to get the ball back in 500 feet to the plate rather than weaker but faster hitters.

Just trying to reason how to leverage a park without any good evidence or research is likely to lead to faulty results.  There’s just way too many factors at play to be able to just guess and properly value them all.

MGL:

Maybe I’m a bit confused about the issue, but here’s a funny story. IN a discussion a few weeks back about Ibanez’s numbers this year, someone was saying that PHI was smart to know that Ibanez’s defensive limitations would be as problematic in their smaller outfield. I responded predictably that 1) less than one season is too small a sample to say with much confidence that Ibanez is _really_ playing better/not as much of a liabililty out there, and also 2) I remembered you writing somewhere around here what you’re saying here: that there’s no evidence that you can leverage good OF defenders in a larger park, or hide bad OF defenders in smaller parks.

But then, randomly, I came across this a 2007 study at THT that concludes:

“As you can see, fast players do indeed have a much bigger advantage in large parks and slower players have an advantage in small parks, just as conventional wisdom would suggest. In a small park, a fast player basically achieves the same UZR score (1 run better) as when he plays in any other park. For a slow player, his UZR score improves by five runs, or four runs more than the fast player.”

http://www.hardballtimes.com/main/article/speed-and-defense/

Not sure I’ve ever heard of the author…

The author does go on to say it’s a small sample size and stuff, but… um, what am I missing here? Is the issues in the present post and the THT study two different things?

Ugh… grammar

Kincaid:

I think you read in a bit more to what I wrote than I intended. It’s certainly not true that all fast players are weak hitter and all power hitters are slow. In fact, with a 500’ fence, a fast power hitter would be far more valuable compared to a slow power hitter - relative to their comparable values in normal sized parks. If every hit is in play I can’t see how speed doesn’t become relatively more valuable.

So if a more normal sized but still large park size results in more BIP then speed must be more valuable. If we’re talking SIH (Speed independent hitting) then only HR, HBP and BBs would be counted - essentially the same as FIP. Any time the defense is involved then speed matters both offensively and defensively. It may be mouse nuts, but it should be there.

I’m with Cameron. I think what it all boils down to is teams need to have two different sets of “true talent” projections. The first is in a neutral setting, so they know how much a player is “worth”. The second is how much those players would produce playing half of their games in that specific team’s home park. Then put together the best team you can for the money.

I do think that in a perfectly efficient market, the speed and defense type of players would gravitate toward the bigger parks. Let’s say you have two guys worth 20 runs in a neutral setting. There will certianly be cases where a Slappy Cabrera might still be worth 20 runs in the bigger park, while Jhonny Power might only be worth 15. Naturally, that big park team, given the choice, is going to choose Slappy Cabrera.

In reality, I doubt it matters much, and when it does, it’s with the fringy/replacement type of players. There’s just not enough above average talent available for teams to be choosy when they have an opportunity to sign someone, especially if that player is coming at any kind of discount.

Devil, I’ll check out that link. I don’t recall reading it.  I got exactly opposite results.  I broke each park down into small and large LF, CF, and RF, based on the square footage of each section (LF, CF, and RF) of each park.  Then I looked at the UZR of bad fielders (and I think slow players independently, although slow players as a group are bad fielders of course) in large and small outfield sections and did the same for good fielders.  I found that the good fielders had a gap in favor of the small parks and that the bad fielders had a gap in favor of the large parks.  Basically, I found that you can leverage a fast outfielder in a small park and not a large park. I don’t see any reason why this is counter-intuitive.  I’ll say this again.  Just as you can argue that you “need” fast fielders to track down balls in large outfields, I can also argue that in large parks, much of the expanse is “wasted” in that no one can track down a ball in the far reaches and that a small park is perfectly suited to a fast player who can track down almost every ball hit in it.  I just don’t see how common sense and intuition is going to answer this question.

Matt, yes, I often amplify differences in order to get a “handle” on things in my head.  So let’s amplify the park and player differences, OK?

Super large park.  Players with no power and players with lots of power.  Players with no power can never hit it over the outfielders heads, and if they do, the ball only rolls 50 feet away, so the size of the park makes no difference as it gets larger.  On the other hand, the power hitters hit the ball over the heads of the outfielders all the time.  THEY are the ones that get all the triples and inside the park HR’s!  The non-power speedsters rarely do. We can argue both sides of the story can’t we?

Again, common sense and intuition is not gonna cut it either for defense or for offense.

And yes, if parks had a straight linear effect on HR rates, smaller parks would greatly benefit power hitters in general.  But, we don’t think that HR factors work that way.  We think that they are at least somewhat additive, so that a large park may add 5 HR to both Juan Pierre and Albert Pujols, for obvious reasons.  Most of Pujols HR are going to clear any fence and Pierre can only hit HR’s in small parks.

LOL, you got me!  I didn’t realize that was my study.  I have since redid the study using better and more data and came up with opposite results.  That is one reason why I have said that I am nowhere near certain of the true answer. But at least I am trying to get empirical evidence.  Everyone else is just blindly assuming that conventional wisdom is correct and trying to use logic and common sense to justify and support that position.  When you try and use that kind of “evidence” (supposed logic and common sense) to support a thesis, you are asking for trouble.  That is the kind of evidence that Joe Morgan uses when he spouts a thesis.  Seriously.  HIS logic and common sense of course!

Maybe the biggest problem with that study considering the sample size issues and other problems is that I did not break the outfields up into 3 sections. In my new study I did.  So, for example, in the old study, if Wrigley was a small park, but only LF is small, even the RF’er and CF’er, got classified as playing in a small park.  Over a very large sample, that is not a huge problem, but over smaller samples, like the 4 years I looked at, or whatever it was, not splitting the OF up was not a good idea. Anyway, as I have said many times, I am willing to conclude that the jury is still out on this issue.

Hey, who ever said I wasn’t a good writer? That was a well-written and clearly written article!  Oh yeah, it is those guys as BBTF that rag on my writing, presumably because they are insanely jealous…

I’m actually relieved. In the discussion I mentioned, I said “MGL somewhere said that he knows it’s ‘counterintuitive,’ but there’s no evidence that smaller parks are easier to OF in,” then a week or two later came upon that 2007 study and felt like a total ass in my pathetic appeal to authority who turned out to have shown the opposite view!

Did you ever published the results of the new study?

When I read your first response, I got really confused, because I didn’t realize you hadn’t looked at it yet… I assumed you would realize that I was referencing your own article.

hee hee…

http://www.insidethebook.com/ee/index.php/site/comments/i_could_use_some_help_with_some_data/

William, thanks!  I knew I had done something subsequent to the THT article.

One thing I didn’t mention is that the whole analysis (the first one and the second one) could be flawed if there are significant biases in the way the UZR engine computes UZR in large outfield sections relative to small outfield sections.  And that could be because of data biases. As I mentioned in the THT article, it is possible that in a large park, air ball distances are under reported and vice versa in a small park.  Or some other bias.

But, as you can see, there is absolutely ZERO evidence that CW is correct and speedy fielders are better suited to larger parks.  In fact, even if there are biases and problems with my analysis, looking at the numbers again, I would be shocked if it turned out that CW was correct.  My numbers indicate a 5 run advantage for speedy players in small parks!  It would take a huge problem with my analysis in order to turn that around in the other direction.

So whenever you hear, “We need fast/good outfielders because we play in a large park,” or “We can ‘hide’ Burrell or Dunn in the OF in our park, because it is small,” please keep this in mind.

MGL,

I’m (mostly) following you here, but how does this figure in: in a park with shorter fences, some balls that would have flown over Burrell’s head and landed on the warning track or hit the wall, balls that a rangy fielder like Carl Crawford might have fielded, will now leave the park, rendering the LF’s range irrelevant.

So, summarizing: a smaller park makes the LF’s range irrelevant on more balls than does a large park.  The converse is thus true: the LF’s range MUST be relevant on a larger percentage of balls hit to LF in a large park than in a small park.

If this is true (and I think it is), how can you say that superior range is not more valuable in a larger park?

I’m not debating whether UZR can detect any difference, mind you.  It is quite possible that UZR is not capable of detecting an effect like this, since “boundary” plays where one fielder can reach a ball and another can’t are comparitively rare.  This is a logical point…

And to follow up with some data, through July 7th of this season, I’ve identified 40 batted balls hit at Citi Field that were not home runs, but which had the distance to get out of Shea Stadium.  So, 40 balls that would have been out of reach of every fielder, slow or fast, are this year in play.

If you believe that having outfielders with excellent range is not more important in a big park like Citi Field, you would have to believe that all outfielders, slow or fast, would have an equal chance of running down those 40 balls (all of which were hit very far, either landing on the warning track at Citi or hitting the fences).  Which is of course absurd.

Sorry for the triple post, but one more point:

Take those 40 balls, and divide them up home and away.  Now you’re at 20.

Divide them up by the number of outfield zones (7).  Now you’re at 3.

Divide by the number of different outfielders who play each position for a team (especially a team ravaged by injuries like the Mets).  Now you’re at 1 or 2 plays per fielder.

Just like that, the data’s gone… UZR will never see it…

I don’t think it’s as simple as looking at the 40 balls that are now in play because they don’t clear the wall because outfielders’ positioning is informed by where the wall is.  If the outfielder were positioned in one way for a large field and could get to x number of balls on the track, and then they moved the fences in so that those balls were no longer in play, the outfielder has no reason to position himself so that he can reach those balls that were on the track before, because he can’t get to them anyway.  So he moves in slightly, and now he can’t get to those balls that went over the fence, but he can get to extra balls further in that he couldn’t reach in the larger stadium.  A fielder with more range loses more reachable balls that go over the closer fence than one who wasn’t getting to as many of those balls in the first place, but he also gains more extra balls in front of him from moving in.

The fielder will, ideally, position himself so that he can get to enough balls over his head but that the wall doesn’t cut off his range.  That means that when the fences move in, the fielder with better range can expand his range further inward than the fielder with poor range because both fielders are still ranging just as far as they were before, and the better fielder is still going to set himself so that he is not wasting range on those balls that he can no longer catch because of a wall in the way.

Do the extra balls caught by extending range inward make up for the balls lost at the track?  I don’t know, but they certainly can’t be ignored.  If 40 balls are already nearly undetectable by UZR without even considering the extra balls reached by the fielder positioning himself closer to the plate, then once those are added into the equation, I can’t imagine it makes much difference.

Maybe one possible way to look at it is to see if the number of expected outs changes in a large or a small field?  If the there are more balls in play for the fielder in a small or a large outfield, then it might stand to reason that a better fielder can leverage the extra opportunities just because his talent is coming into play more often.  I guess it could run into the bias issue for handling things differently in small and large parks, and maybe it wouldn’t say anything definitive, but it seems like it could be one piece of the puzzle.

Greg, as you can see, and as I have been trying to say, there are number of ways to look at it, and each way might favor one point of view or the other.  Which is why I keep saying that you can’t solve this one by “thinking it out”, I don’t think.

As Kinkaid pointed out, your argument only works if fielders did not reposition themselves depending upon the size of the field, which clearly isn’t the case.

So we are back at square one.

Here is an argument I have made several times:

As someone said, it often helps to “expand the differences.” Let’s say that we take a large park and make it even larger.  There comes a point at which a ball hit so far (and that point is not even that far) that no one can catch it.  So even if we use your flawed example of the outfielders who never reposition themselves, there are balls that are going to be hit in a large park, that cannot be caught by anyone, fast or slow. So at that point speed becomes irrelevant, at least as far as catching the ball.  So, even if we found out that a large park does in fact favor fast fielders, that advantage has to end at some point.  What point is that? 

Now, speed of course is important when a ball is not caught.  And I will concede and admit that in a large park, speed is an advantage because balls that are not caught will travel further and it will take a slow OF’er longer to get to them.  And UZR does not measure that aspect of fielding (unfortunately).  How much difference there is in a reasonably large and small park in MLB in terms of balls that are not caught and how far they end up, and how much difference there is, say, per season, between a slow and fast fielder tracking those down, in terms of runs, I have no idea.

But that is definitely an issue that favors the faster fielders in larger parks, incontrovertibly.  As far as catching balls, which we are mostly talking about, I still have no idea and I don’t think anyone does.  The data from my most recent study, though, seems to suggest that faster players are more suited to smaller parks, at least in terms of catching balls as opposed to running down balls that are not caught.

I’m not seeing anything here I disagree with, but I get the feeling we are not all exactly addressing the same question.

I’m NOT trying answer the question, “Is a fast outfielder more valuable in a big park or a small park?” That’s something for a player’s agent to worry about when he tries to put his client in a place where he will shine the best…

I’m trying to shed some light on the question that Omar Minaya is facing, “I have a large park.  How important is it that I have fast outfielders, given that this is my park and no other?”

I think we can all agree that no matter what park you’re in, fast outfielders are better than slow outfielders.  No matter how much ground needs to be covered, given equally intelligent positioning, the fast fielder can cover more of it, until you reach the infinity cases of infinitely small (slow can cover zero as well as fast) and infinitely large (fast can’t cover any more of infinity than slow can).  Which of course you never reach…

What I think the real issue is, though, is “Can I get more value (runs) in my large home ballpark from emphasizing rangy outfielders vs. the other possible personnel selections (e.g. slow but powerful sluggers, mid-range and mid-power players, etc.)?”

To answer that, we need to bring in measures of offensive and defensive value, establish our degree of trust in those measures, and then decide if the indicated values are statistically significant.  This is complicated stuff, made more complicated by the limitations of our defensive metrics, including outfield range.

The question is, if I have a large ballpark, am I better off with a player who is +5 on defense and -5 on offense (across all parks) or -5 on offense and +5 on offense?  That’s it.

I agree, that is the question.  And a tough one…

I think the leveraging has to be more specific to make a significant effect. Such as, FB pitchers in Citi paired with good outfielders.
+5 defense, -5 offense would be better if the park creates more defensive opportunities. But it depends on how the park effects the offense also. If the players doesn’t hit many long homeruns his value would go down. Or if left field is larger than right field.
I think I would rather have the +5 defense because of more opportunities and also chasing down balls.

I was just thinking about the issue and visualized it this way: a speedy, rangy fielder’s range is already maximized in a small ballpark, if positioned properly. His range isn’t increased in a large ballpark. Same with a slow fielder. However, the difference between a rangy and non-rangy fielder is greater percentage wise in a small ballpark. I don’t see how a large ballpark favors speed, other than chasing down balls, which isn’t part of UZR.

EK, well that’s another way to look at it…

EK, you say “the difference between a rangy and non-rangy fielder is greater percentage wise in a small ballpark.”

What difference exactly?

I meant difference in range, which would lead to a difference in UZR. I was just trying to explain MGL’s results.