didn't watch mythbusters, did they? my question, though, is why didn't they say anything in the article about the increase in mass? even if a curveball doesn't curve in the conventional way it does at sea level (because it creates drag), doesn't the increased mass of the ball compensate in some way?

psychology seems like the cop out answer here. there's a real effect that's happening here: humid baseballs don't fly as far as dry baseballs, at least according to the experiment as run on mythbusters.

So why has the ERA at Coors Field dropped since the humidor was introduced? The drop is extremely small and the ballpark remains in the top four for home runs, notes Meyer. But the change is unexplained and merits further investigation. One intriguing possibility is that humidified balls are easier for pitchers to grip. Meyer and Bohn are now planning a follow-up study.

Baloney. The park factor has dropped from the 120-129 range, to the low teens, to now 108 in 2007

This is all due to a placebo effect? Right

Did the study take into account the effect the 'humidored' balls have on the pitcher? You can control a dampened ball easier than you can a super dry baseball.

(The authors are two of my friends)

Mythbusters is entertainment, not science. The Mythbusters "study" compared 0% relative humidity to 100% relative humidity. No controls, no nothing. The Coors field humidor changed the storage environment from 30% to 50% relative humidity.

Humidity increases the ball's mass, which means it experiences less acceleration from aerodynamic forces. The effects on the ball's travel are about a half inch less break on the curve, and about two feet more on a long drive. So whatever effect the humidor has which decreases scoring, those two factors aren't it.

Mythbusters is entertainment, not science. The Mythbusters "study" compared 0% relative humidity to 100% relative humidity. No controls, no nothing. The Coors field humidor changed the storage environment from 30% to 50% relative humidity.

so you didn't watch it either, huh? they used control balls stored in "normal" san francisco humidity. they did both a dropped ball test AND a batted ball test, measuring the distances of both and, guess what, the wetter balls didn't travel as far or bounce as high. it's fine to make the statement, as the article seems to do, that the wetter balls have more mass and less drag, but what about the effect of the ball with more mass and CONTACT with the bat? what about gravity?

and also,

(The authors are two of my friends)

doesn't exactly give you the credibility that you seem to think your answer deserves.

Who said anything about easier to hit? Also, who said anything about aerodynamics? The shape of the ball will be nearly identical no matter what the humidity is.

The real question is energy transfer. This study did nothing to answer that.

Bust. Mythbusters did much better than these clowns.

Yes, the study is baloney in that there definitely has been an effect at Coors. Whether it is a placebo effect or not (or some combo) is another story. I would think that there are so many ways that a "humidored" ball can affect run scoring/home rate rate, etc. that unless they tested (and did it well) for all of them, to conclude that the humidors have little to no effect is irresponisble.

Here are the average fly ball distances and HR per fly ball rate for both teams in Rockies' home and road games and a "fly ball distance park factor" before the humidor, after the humidor and when they "cranked it up" in 05 or 06:

Pre:
00 336/.174 317/.105 1.06
01 340/.192 320/.139 1.06

Total: 338/.183 318/.122 1.06

During:
02 334/.155 322/.117 1.04
03 334/.149 320/.121 1.04
04 336/.167 325/.143 1.06

Total: 334/.157 322/.127 1.04

Humidor "max"
05 333/.139 324/.117 1.03
06 332/.133 324/.109 1.02
07 338/.136 324/.108 1.04

Total .335/.136 324/.111 1.03

I don't think there is any doubt that the humidor has had a significant effect. Without reading the study I would think, one, large diameter, more friction, heavier, and better grip for pitchers. That creates better pitching in general and a ball which travels slower and less far, on the average. This study refutes those?

(The authors are two of my friends)

doesn't exactly give you the credibility that you seem to think your answer deserves.

My, what an appropriate handle you have. It's called a disclaimer. Look into it.

From the study

In all cases, the humidified ball actually travels slightly farther, but only by about 2
feet. Similar to the effect on lift, the acceleration due to drag is proportional to the ball’s
area, but inversely proportional to its mass. Since the fractional change in mass is greater
than that in area, the mass effect wins. The change in RH may have other effects on the
batted ball beyond aerodynamics. For example, Kagan6 found that a 20% increase in RH
would reduce the ball’s coefficient of restitution, resulting in about a 6 foot reduction on the
distance of a batted ball. However, he made a simplifying assumption that the coefficient
of restitution affected the total velocity whereas it only affects the component of velocity
along the line joining the long-axis of the bat with the center of the ball.14 Thus, 6 feet is
likely the maximum variation that can occur for a well-struck baseball.
Therefore, the effects of collision dynamics and aerodynamic effects have opposite conse-
quences for the distance of a batted ball, and the two effects may nearly cancel each other
out. This suggests that storing balls in humidity-controlled environments has little to no
effect on the range of batted baseballs. For the thin air of Denver, this conclusion depends
weakly on which drag curve we consider. However, at sea level, typical baseball velocities
are more likely to reach Reynolds numbers near the drag crisis, and humidity of the ball may
play a somewhat larger role. This general conclusion is supported by the fact that Coors
Field has experienced more HRs per game in the NL than all other stadiums save Wrigley
Field in the 2007 season4. In fact, Coors Field has finished in the top 4 of all ML ballparks
in park factor since the inception of the humidor in 2002.

The issue of the onset of the drag crisis makes it difficult to predict whether sea level experiments are even directly applicable to Coors Field. Like I said, uncontrolled experiments.

I'm not able to address the phsyics and aerodynamics. But I highlighted their claim that the drop in offense hasn't been all that much for a reason. Sometimes things are self evident. Offense is down a ton since they started using that thing. It's not even debateable, and their claim that offense is only down is just wrong. There is no other reasonable explanation for why offense would drop so precipitously. Nothing else has changed. (I don't think Denver has sunk) It must be a result of using the humidor. They simply failed to identify exactly why.

So, according to the article, one study (Kagan) showed that humidified balls traveled a bit further and the other one (these authors) showed that a humidified ball travels less far. Let's clear that up before we draw any conclusions. And I would think that actually testing the balls, in this case, would be A LOT better than putting everything into a computer model.

This is from the article:

Surprisingly, they found that drag actually dropped as balls became more humid. The reason, according to Meyer, is that the increased mass allows the ball to cut through the air more easily. This more than offsets the increased air drag from the tiny increase in surface area.

They say that the icreased weight makes the ball go further. I assume that assumes the same initial velocity off the bat. Did they test that? How about the velocity of the pitches? If we had a 100 pound ball would that go even further? No, becuase the initial velocity off the bat would be almost zero. How about the coefficient of resitution (how much the ball "bounces" off the bat and retains its kinetic energy)? Wouldn't that change with the humidity and weight of the ball? Did they test for that?

The team also dispelled the myth that humidified balls were easier to pitch as curveballs — another factor cited in the drop in home-run statistics. Again, says Meyer, the density of the humidified ball was the key factor. Because the ball was heavier, he says, it was harder to make it curve than a lighter, drier ball.

Dispelled the myth? How about if pitchers can't grip the dry baseballs as well and thus have trouble impartinmg the same spin velocity plane as with the wet ones? Did they check that (which would be hard to do unless you actually used real love pitchers)?

I'm sorry, but if this article accurately reflects the study, these guys are a bunch of research morons. Academicians (is that a word?) doing baseball research often spells disaster.

Actually, they (the author of this article, not the study, unless he was paraphrasing or quoting the study itself) lost me at, "which is done in some parks..." (emphasis mine). Are there any other parks besides Coors that uses a humidifier?

The study can be found here.

There are plenty of myths to be busted in baseball. This ain't one of them (that the humidor has significantly reduced offense in Coors Field after controlling for everything else).

I would state shoewizard's point more strongly. The study found that aerodynamics are not the cause of the humidor's effect, and that in fact they strongly counteract the effect from the changed coefficient of restitution. Similarly, the effect on breaking balls is not to increase the break experienced due to a particular spin. Neither effect is intuitively true, and in fact neither effect was expected in light of the large anecdotal effects from the humidor.

And where did you get your physics degree, mgl?

Here's a link to their group at Colorado:

http://grizzly.colorado.edu/~bohn/

Call me silly, but I'd be a bit less eager to call these guys "research morons."

How about if pitchers can't grip the dry baseballs as well and thus have trouble impartinmg the same spin velocity plane as with the wet ones? Did they check that (which would be hard to do unless you actually used real love pitchers)?

Another possible effect of the humidor is to make the horsehide surfaces of
the balls more supple, enabling the pitchers to maintain a better grip and therefore pitch
as they would with balls stored in higher RH elsewhere. This effect has been reported by pitchers, but remains unexplored quantitatively.

Nevertheless, reports from pitchers and batters alike assert that the humidified balls break
more. Our result shows that this cannot be due purely to the aerodynamics. However,
pitchers also report that the humidified balls are easier to grip. It is possible that they can
put a greater spin on the ball. Since the lift force is also proportional to spin, it would take
only an additional spin of 0.9% to overcome the aerodynamic effect arising from increased
density of the humidified ball. Any additional spin the pitcher can provide beyond this will
further overcome the aerodynamic tendency of the humid ball to break less. It may be this
effect on grip that is responsible for the ability of pitchers to perform at Coors Field more
like they do elsewhere since the introduction of the humidor.

The search for the influence of the humidor on the game at Coors Field then shifts away
from aerodynamics to other possible effects. An intriguing possibility is that the humidified
balls are easier to grip, allowing pitchers to put a greater spin on a humid ball than on a dry
one. In addition, a batted dry-ball may be able to have more spin imparted in the collision,
thereby allowing the drier ball to travel further. These will be topics for future study.

Another possible effect of the humidor is to make the horsehide surfaces of
the balls more supple

How long has it been since baseballs were made of horsehide? 20 years? 30?

#14, I have a PHD in physics from MIT.

Just kidding. I should not have called them research morons. I take that back.

Nowhere in my posts do I make any claims as to a knowldege of physics. If I did, redact that from the posts. Nice ad hominem attack. The ad hominem attack is always the province of the person who has nothing worthwhile to add to the argument. Termel, feel free to address ANYTHING of substance that I said or that anyone else has said or even, marvel of marvels, something original and worthwhile you might have to contribute to the discussion.

One of the problems of course is that it is not fair to critique a study from a synopsis of it in a newspaper.

Well, it strikes me as pretty likely that a pitcher's grip on the ball is pretty darned important. Pedro talks a lot about the importance of humidity for his ability to spin curveballs and he's a pitching genius.

Allowing pitchers a better grip on the ball is the stated purpose for using the humidor in the first place. Why wouldn't that be the first thing someone interested in its effects would test?

I agree with GMSoRP, most articles I've read about the humidor point to the ball going from a cue ball feel to a real baseball feel. Its not the aerodynamics.

Well, it strikes me as pretty likely that a pitcher's grip on the ball is pretty darned important. Pedro talks a lot about the importance of humidity for his ability to spin curveballs and he's a pitching genius.

Certainly a great point, one that I happen to agree with. However, good luck measuring grip.

Incidentally, mgl, where did you get the numbers for the "fly ball distance park factor"? What database do you use? Do you have any numbers for the error bars, or the numbers of fly balls that went into the calculations? If you could do the same calculations for every year in Denver, it might help clear up the natural fluctuation of those numbers.

One of the problems of course is that it is not fair to critique a study from a synopsis of it in a newspaper.

I have more problems with the synopsis than I do with the actual study. The study made it clear on several occasions that the conclusions are with regard to the aerodynamic effects only. To wit:

Our discussion suggests that if the humidor has altered the game at Coors Field, it does not do so by purely aerodynamic effects.

It doesn't conclude the humidor does nothing. Maybe they were quoted out of context in the synopsis, but they seem to overextend their conclusion in the article:

But these effects weren't large: the bottom line is that humidified balls don't change things much, says Meyer. "I'd say the role of the humidor is more psychology than pure physics."

Unless a physicist is suggesting that kinetics aren't in his domain - or that psychology is - this quote differs from the limited (and accurate) language in the actual study.

Nowhere in my posts do I make any claims as to a knowldege of physics. If I did, redact that from the posts. Nice ad hominem attack. The ad hominem attack is always the province of the person who has nothing worthwhile to add to the argument. Termel, feel free to address ANYTHING of substance that I said or that anyone else has said or even, marvel of marvels, something original and worthwhile you might have to contribute to the discussion.

mgl,
it's nice work to throw out some insults (research morons and Acadamecians doing baseball research) and then lecture another poster about attacking, when he really did no suh thing - he was asking about your credentials. And Termel *did* add to the discussion - he posted a link to their reseach group..

Oh, you "took it back". That's different.

Interestingly, mgl's numbers on the Coors field effect on fly balls seem consistent with the article's findings:

Pre: 338/.183 318/.122 1.06

During: 334/.157 322/.127 1.04

"Max": 335/.136 324/.111 1.03 (I'm not sure of the difference between the "during" and "max" years.

It's an odd result, because the movement upwards in the road distances is about one and a half times bigger than the movement down in the home distances, but the loss of distance on home fly balls is just about exactly what you'd get from losing six feet from coefficient of restitution and gaining back two feet from aerodynamic effects

...the increased mass allows the ball to cut through the air more easily.

For a given initial velocity, yes. But the heavier ball would not come off the bat with the same velocity - as noted by Piehole and Beano.

What the pitchers claim is that the humidor allows the covering of the ball to avoid becoming as slick as it was before the humidor. It also provides for the seams to remain raises. Both of those allow the pitcher to have a better grip. I don't know that anyone who has watched at altitude has claimed that the balls don't carry as far. What has been the consensus is the key is the pitchers get a better grip and as a result throw better pitches.

ballfan, and the others who have commented on this, are 100% right.

What the researchers appear to have done is to operate on an incorrect assumption ("Baseball believes the humidor works by changing the aerodynamics of the ball"), and then proceeded to disprove something that was, in fact, not believed in the first place.

So while they are probably correct about the aerodynamics, they have essentially disproved something that is not in contention. Not wholly without utility, mind you, but not really on point.

The idea that the role of the humidor is "psychological" is the kind of off-the-cuff comment that researchers should avoid, since it (a) makes no sense, given the demonstrated real world change in scoring; (b) has no basis in the data that they were working with; and (c) is outside their area of expertise.

It's an odd result, because the movement upwards in the road distances is about one and a half times bigger than the movement down in the home distances

Which, oddly, MGL didn't even notice.

Folks, when looking at changes in ratios, it's always a good idea to see if the changes are in the numerator or the denominator or both ... or sometimes even better, don't look at ratios to begin with.

Anyway, if distance at Coors had stayed exactly the same while road distances increased, the ratio in the "max" period would be 1.04 not 1.03 -- yep, it's clearly something to do with Coors!!

By the way it's rather curious that while fly balls outside Coors are travelling farther, fewer HRs are being hit (per flyball). Coors flyball distances have barely changed (-1%) but HR/fly ball rates have dropped by 25%. Those numbers make no sense ... or (mean) fly ball distance has no meaningful relationship to HR rates.

Anyway, for the mean distance to increase while the rate falls, beyond random noise, seems some combo of the following 3 have to be true:

1. the distribution has changed -- fewer long fly balls but also fewer short ones. Several speculative hypotheses exist regarding the first one (e.g. steroid testing) but the only one I can think of to explain the second would be some definitional change in the stats (i.e. are infield pops considered fly balls at one time but not another?). OK, an overarching explanation is that the classification of 'fly ball' vs. similar events and/or measurement of fly-ball distance are so inconsistent that the data is near useless.

2. fences have been moved back, walls raised in lots of parks.

3. some really big parks have opened in the last few years.

So the "humidor effect" is not particularly well-supported by MGL's numbers and is certainly not the interesting phenomenon in MGL's numbers. The interesting phenomena are increasing flyball distance overall couple with the overall decline in HR/fly rates. I'd be much more interested in those being explained.

However, good luck measuring grip.

True, but shouldn't PitchF/X be able to measure the movement on pitches? Couldn't you have a few pitchers known for the movement they impart on different types of pitches, like Zito (curve), Rivera (cutter), etc. throw a few dozen pitches with dry and humidored balls at Coors and measure how much their pitches move?

Basically, these guys disproved a couple of things that nobody ever thought. The curveball factor reportedly has to do with grip, not aerodynamic forces, and the coefficient of restitution, which is only covered in a citation of an article, is the explanation for why the balls fail to fly further.

Here is the cited article. Someone with a little more of a physics background will have to tell me whether or not this testing method for COR is valid.

#28: I don't know that anyone who has watched at altitude has claimed that the balls don't carry as far.

#29: What the researchers appear to have done is to operate on an incorrect assumption ("Baseball believes the humidor works by changing the aerodynamics of the ball"), and then proceeded to disprove something that was, in fact, not believed in the first place.


I'm not sure who qualifies as "baseball", but about ten seconds of googling turned up the following:

" 'It[the humidor]'s leveled the playing field for the pitchers in this ballpark,' Rockies manager Clint Hurdle said Saturday, a day before Game 3 of the National League Championship Series.

'All we wanted to do was keep the balls regulation size throughout the season. What we found was the balls were getting smaller and traveling farther.[...]' " (link)

(EDIT: misread ballfan's "I don't know that anyone" as "I don't know anyone", so I had to change my phrasing a bit.)

Computer models? According to a computer, a goose can't fly and a hummingbird can't flap it's wigs that fast. Computers models ensured these geeks on Wall Street that subprime bonds were safe investments. If they're going to rest the effectiveness of a huidor, why not just give a pitcher a dry ball vs a humidified ball and measure speed and rotation? I love wet balls with raised grippable seams to throw curveballs.

Jittery McFrong, I don't believe Hurdle's quote says that it made the ball carry less. It leveled the playing field for pitchers could also refer to the fact that a pitcher now can get a legitimat egrip on a pitch.
Keeping hte blals regulartion size is also another factor but that does not mean that the ball doesn't carry as far. There is statisitcal data that backs up the balls not in the humidor were lighter and smaller than regulations.
My point is that those regularly watching and participating in games at Coors Field remark more about the feel of the baseball and the grip than the travel. The balls still carry very well but not as often because pitchers are able to make better pitchers.

They dried all the balls then put them in 50% RH. That is not how the balls are handled at Coors. If you are trying to find out how things happen, then why do things the opposite of how it is handled in real life?

They should have measured the balls as soon as they received them, then dried some of them and kept the rest at 50% RH.

Instead they dried (shrunk) all the balls and then tried to re-expand 2/3 of the balls.

The experiment began with all balls held in a dry environment for a period exceeding
two months.

Their weights on the balls in the study did not decrease as much as was observed by Coors field staff.

The idea for the humidor bubbled up during a duck hunt six years ago. Tony Cowell, who works in the Rockies' engineering department, noticed that his leather hunting boots had constricted over the summer. He began to think about the leather covering on baseballs and how the dry air in Denver, where humidity rarely exceeds 10%, affected it.

Rockies employees conducted some basic tests. They dropped balls they'd had in storage for a while from the same height as balls they'd just received from Rawlings, Major League Baseball's supplier, and noted how much higher the older balls bounced. They started weighing balls periodically.

"You could tell over time that it was weighing less and less, to the point of being too light and not meeting Major League Baseball specifications," Kahn says.

Some balls weighed a full ounce less than the 5 to 5 1/4 ounces that MLB requires, he says.

Link

They used the mass being lowered by 1.6% or to 5.043 oz. That would not be enough.

Does anyone know where the baseballs were stored before the humidor? Could it have been a spot with lower than 32% RH?

How long has it been since baseballs were made of horsehide? 20 years? 30?

They switched from horsehide to cowhide in 1974. You would think competent researchers would have looked that up.

The baseball's outer cover is made of Number One Grade, alum-tanned full-grained cowhide, primarily from Midwest Holstein cattle. Midwest Holsteins are preferred because their hides have a better grain and are cleaner and smoother than those of cattle in other areas of the United States.

Link

My point is that those regularly watching and participating in games at Coors Field remark more about the feel of the baseball and the grip than the travel. The balls still carry very well but not as often because pitchers are able to make better pitchers.

Helton disagrees.

When they first started using balls stored in the humidor at the beginning of the 2002 season, they didn't tell their own players.

"I was hitting balls and they weren't going out of the yard," Rockies first baseman Todd Helton says. "I was wondering what the heck was going on. I thought something wrong with my swing or whatever, and then a few weeks into the season they told us and it started to make sense."

Computer models? According to a computer, a goose can't fly and a hummingbird can't flap it's wigs that fast

That's a bumblebee. For both examples, I think.

Basically, these guys disproved a couple of things that nobody ever thought. The curveball factor reportedly has to do with grip, not aerodynamic forces, and the coefficient of restitution, which is only covered in a citation of an article, is the explanation for why the balls fail to fly further.

A lot of the hypotheses in previous seasons was that breaking balls were flat due to the thin air, resulting in better contact. So the study has some merit in that area.

I am not dissing their study and I don't argue with their findings. My point is simply that I believe the help in the humidor comes from other areas, particularly dealing with gripping the baseball. I might add, as for curveballs at altitude, Burt Hooten learned the knucklecurve from the pitching coach of the Boulder Collegians, a one time summer semi pro team in Boulder, Co. Also Shawn Chacon grew up in Greeley, Co., and always has had a very good curve. A lot of times excuses are devised to cover up inabilities. Part of what has happened at Coors Field in recent years could be that the Rockies actually have better pitchers.

I don't believe Hurdle's quote says that it made the ball carry less.

Please pardon me if I'm a bit blunt here, but I think that's exactly what "the balls were getting smaller and traveling farther" means. In context it's clear that he is talking about pre-humidor balls "traveling farther" than balls kept in a humidor.

Or if you're unsatisfied with the Hurdle quote, try this one:

Boston hitting coach Dave Magadan almost makes the humidor seem like a cloak-and-dagger operation — not an atypical reaction from puzzled opponents.

"We kind had heard about it secretly," he said Friday. "There were rumors. We heard stuff in the wind. I thought it was kind of a myth, then we found out they were actually doing it.

"Yeah, you noticed a little bit of difference. Balls that a guy squares up, most of the time, it's a home run. But for the ball not to go out ... it was like, something's going on."

There is still some dispute about whether the humidor affects the distance the ball travels, or merely aids pitchers' ability to grip the horsehide.[...] (link)

That was the second hit in my google search, by the way. You don't have to look hard to find baseball people expressing the idea that the humidor makes the ball carry less. (Granted, this quote doesn't necessarily attribute a lower travel distance to aerodynamic effects, as opposed to say, a difference in the coefficient of restitution. But he was definitely not talking grip here.(And Hurdle's quote does sound like an aerodynamic claim to me.))

Keeping hte blals regulartion size is also another factor but that does not mean that the ball doesn't carry as far.

My point was that Clint Hurdle, among others, thinks/thought otherwise.

A number of posters (e.g. you, Srul) have made comments to the effect that the study was debunking a theory that no one in baseball held. But even a quick google search indicates that this is not the case.

Which, oddly, MGL didn't even notice.

Geez, some of you need to see a doctor to get your eyes or in some cases (Dial) your head examined.

That is why I posted the last number in each row which is the park factor for fly ball distance. The road numbers reflect nothing but the "strength" of the Rockies team. If the Rockies hitters in any one year happen to be "better" (in terms of longer fly balls), or the pitchers worse, then the road distance will go up, and vice versa.

We don't care about the road numbers. We only care about the ratio of the home to road numbers. Have any of you (Davis) ever done a park effect calculation before? Do we care about the raw road or home numbers? They change from year to year with the team. We only care about the ratio.

Are any park factors perfect? Of course not. There is fluctuation in the numbers and the weather affects the fly ball distances as well. So no park factors are perfect, especially in the short run.

Before the humidor, the average fly ball distance on the road was 6% higher than at home, given the same two teams (approximately - I don't know if every series at home is also played on the road, but that is how we compute park factors). From 02 to 04, since the introduction of the humidor, it was 4% higher. In 05-07 since they purpordedly left the balls in the humidor longer (it might have been in 06 and 07 only), it was 1.03. Should I repeat that more slowly for guys like Davis who do not seem to understand my numbers nor how to compute a simple park factor?

That suggests that the humidor has had an effect on the average fly ball distance at Coors relative to on the road. Are the numbers (such as the difference between the pre and post-humidor fly ball distance park factors - 1.06 and 1.35) significant, as Zach essentially asks (at least he asks rather than says something really stupid in response to my posts)? I'll let the geniuses like Dial and Davis figure that one out. Per season, there are around 1100 fly balls at home and 1100 fly balls on the road.

Over and out.

Geez, some of you need to see a doctor to get your eyes or in some cases (Dial) your head examined.

Weren't you complaining about attacks earlier? Physician, heal thyself.

"Yeah, you noticed a little bit of difference. Balls that a guy squares up, most of the time, it's a home run. But for the ball not to go out ... it was like, something's going on."

Much more likely a reference to the energy transfer than to altered aerodynamics.

Clint Hurdle might have been saying that the phenomenon is due to the aerodynamics of a moist ball (Validating this study), or he might not. The study which relates to the question most of us are interested in is the one I linked. There are a handful of small, obvious flaws in the experiment, but I'm not sure to what degree they invalidate the research, as they also do not replicate the actual conditions of Coors Field balls.

Instead they ...(shrunk) all the balls and then tried to re-expand 2/3 of the balls.

Ye gods, not another steroid thread!

Is that even legal?

We don't care about the road numbers. We only care about the ratio of the home to road numbers. Have any of you (Davis) ever done a park effect calculation before? Do we care about the raw road or home numbers? They change from year to year with the team. We only care about the ratio.

The thing I wonder about is whether the change in the measured park factors is meaningful. A fly ball is a very particular result of the bat/ball interaction. It's not a popup, it's not a grounder, it's not a line drive. It's a very limited set of initial trajectories leaving the bat. So I wonder if you shouldn't get similar answers for fly ball distance from year to year regardless of team strength.

Consider the Rockies since 2000 (the first year you give statistics for)

Year   wins HR   SLG  dist(home) dist(road)
2000   82   161  .455 336  317
2001   73   213  .483 340  320
2002   73   152  .423 334  322
2003   74   198  .445 334  320
2004   68   202  .455 336  325
2005   67   150  .411 333  324
2006   76   157  .433 332  324
2007   90   171  .437 338  324

Notice how these teams of vastly different quality hit fly balls with almost identical fly ball distances every year. It's kind of remarkable.

The best team had 33% more wins than the worst.
The best team had 42% more HR than the worst.
The best team had 18% higher SLG than the worst.
The best team had 2.4% longer fly balls than the worst at home.
The best team had 2.5% longer fly balls than the worst on the road.

It looks to me like you should expect the average fly ball distance to be the same for a particular park from year to year, with something like a 2.5% measuring error causing some random fluctuation. Thus, a change in park factor from 1.06 to 1.03 would not be significant.

This study does not address the COMPRESSION of baseballs stored at 50% humidity versus dried out balls. Pre-humidor, the balls were lighter and HARDER, making them "hot" balls with a high compression. Today's humidor balls are kept at MLB specs with normal compression. Anyone who plays golf understands the effects of higher compression balls. They fly farther.

I have been to hundreds of Coors Field games. The baseballs behaved like "Superballs" in the pre-humidor era. Now, the balls behave like normal flight baseballs. You can see the difference. Heavier, larger, softer, easier to grip. This is not rocket science, and this study only addressed their drag, which isn't the issue. It is a garbage study as far as I'm concerned.

Much more likely a reference to the energy transfer than to altered aerodynamics.

As I acknowledged. It was mentioned in response to the claim that "I don't know that anyone who has watched at altitude has claimed that the balls don't carry as far." (Although, for the record, I fail to see how "something's going on" would be more or less likely to refer energy transfer vs. aerodynamics.)

Clint Hurdle might have been saying that the phenomenon is due to the aerodynamics of a moist ball (Validating this study), or he might not.

He attributes the phenomenon to a difference in the size of the ball. ("What we found was the balls were getting smaller and traveling farther.") To test that claim you need to study the aerodynamics. Basically, I'm not seeing the "or he might not" part.

I'm not trying to say anything about your other critiques of the study, but your statement "Basically, these guys disproved a couple of things that nobody ever thought" just doesn't seem to be true.

Compression is discussed via coefficient of restitution.

I've got a dumb question: Would it make sense to store bats at a lower humidity?

He attributes the phenomenon to a difference in the size of the ball. ("What we found was the balls were getting smaller and traveling farther.") To test that claim you need to study the aerodynamics. Basically, I'm not seeing the "or he might not" part.

Well, he could simply mean that they were shrinking due to the low humidity, and simply traveling further. From what I hear, Clint Hurdle isn't a particularly bright fellow, so he might simply think that smaller spheres will travel further, or he might think the bats' energy was transferred more efficiently to low-humidity balls.

If they did this study just to invalidate Clint Hurdle's theory, well, bully for them. The majority of people have a different theory to explain the humidor's effect, and I'm not sure whether the study they cited disproves it or not.

I have been to hundreds of Coors Field games. The baseballs behaved like "Superballs" in the pre-humidor era. Now, the balls behave like normal flight baseballs.

That's the amazing thing about mgl's data. The baseballs actually aren't travelling any further.

I think the really significant effect of this research and the stuff done with coefficients of restitution is that it's starting to put a severe limit on the kinds of things that could be causing the humidor effect. It's starting to look like the biggest effect of the humidor is felt before the ball touches the bat, which I'll guarantee you is diametrically opposite to anything you would hear on the street.

[51] I thought I heard that the humididor made the balls easier for the pitcher to grip and therefore throw their pitches better...

Sparkles Peterson,

Well, he could simply mean that they were shrinking due to the low humidity, and simply traveling further.

OK, I suppose that's possible, but in context it seems unlikely to me.

If they did this study just to invalidate Clint Hurdle's theory, well, bully for them. The majority of people have a different theory to explain the humidor's effect, and I'm not sure whether the study they cited disproves it or not.

Calling it Hurdle's theory isn't really fair. When he says "what we found was...", I doubt he was using the royal we.

Calling it "smaller spheres will travel further" isn't really fair, either. It's not so obvious a priori that the phenomenon isn't aerodynamic -- the humidor affects, among other properties, the size, the mass, and shape of the ball (baseball does not equal perfect sphere). An aerodynamic explanation isn't (IMHO) so unreasonable that one would expect "nobody" to think it.

And I don't see the basis for these claims about who thinks what (e.g. How do you know what "the majority of people" think on the matter? And which people are you counting, exactly?)

My admittedly half-assed googling was aimed at that last point; it turned up a number of (non-nobody) baseball people who think that the humidified balls are different after release (not just in grip, as a number of people here think), and a subset of those baseball people who, as far as I can tell, think it can be attributed to some significant degree to the aerodynamics of humidified balls.

The latter are not physicists, so whether their claims are aerodynamic or not is a judgement call. But an analysis of the aerodynamics of humidified baseballs is of some relevence to the the former as well, and a study doesn't have to close the book on a theory to be relevent.

Of course I know how to calculate a PF MGL.

I'm a statistician you moron. You might want to listen to me.

Most of the impact in your numbers is because your denominator increased. The explanation you offered was that the numerator decreased, which it barely did.

Now, genius, what's the standard error on a PF? What's the standard error on mean flyball distance? Explain why flyballs go farther but HR/flyball goes down? Why did you include the HR/flyball numbers then ignore them? Why didn't you notice the single most interesting thing in the data you presented?

Most of the impact in your numbers is because your denominator increased. The explanation you offered was that the numerator decreased, which it barely did.

I'd say that MGL feels that the improvement of the Colorado batters from '00 to present is far and away the overriding force behind the increasing road distances (Which, as he says, would make it irrelevant and render the ratio the only number that matters), and you and others think that there is some other explanation. Do you have any theories as to what else is at play here?

Of course I know how to calculate a PF MGL.

I'm a statistician you moron.

Who wears the hat and who wears the flannel?

Edmund Meyer and John Bohn of the University of Colorado at Boulder used simulations to show that the aerodynamics of a moistened baseball are actually slightly better than those of a dry one1

This has been raised several times in this thread, but why do it with a computer model and not real MLB players? Or if MLB players were not possible, how about minor league players at least?

When cricket researchers wanted to find out the degree of straightening in the elbow during bowling, they used real elite players.

The main thing to remember is that the only thing that these people measured were the diameter and masses of the baseballs. Then they simulated the distance that these balls would travel using their model which looked to me to be a very basic one. They give no data that allows you to evaluate how good this model is and from the data that I've seen on this thread (fly ball distances), there are observed data that strongly suggest that the humidor has an effect that contradicts their predicted results. If this had been a new variant on PECOTA that jibed with the data this badly, you'd all be saying that the model was a bad one or at least be questioning it more. There's no reason not to do the same with this aerodynamic model just because these guys happen to be physicists. In my field, protein modeling, theoretical physics-based models almost always give the wrong answer for anything that isn't simple or obvious to start with.

And that last sentence about psychology was particularly arrogant and reminds me of the days when people argued that curve balls were illusions because they couldn't explain how a ball could curve.

Most of the impact in your numbers is because your denominator increased. The explanation you offered was that the numerator decreased, which it barely did.

Whether the numerator or denominator changes is of course irrelevant, as a "statistician" would know. The composition of the team will change over time, as will ballparks (for away stats). So, for example, if CO hitters have more power now than pre-humidor, the distance of home FBs may stay constant (or even increase) despite the humidor having a real impact. The point is to isolate the impact of playing in CO vs. everywhere else, which is what MGL does.

I think MGL's distance data is better presented in terms of difference in distance, rather than a ratio.
Home/Away/CO Diff
Pre: 338/318/+20
05/07: 335/324/+11
So, CO hitters are hitting the ball 6 feet further on the road now, but 3 feet shorter at home. The Coors advantage has been cut in half, from +20 to +11. That's huge.

Looking at HR/FB, we see:
Home/Away/CO Diff
Pre: .183/.122/+.061
05-07: .136/.111/+.025
A consistent story: the home advantage on HR/FB has fallen dramatically, roughly cut in half as well.

Explain why flyballs go farther but HR/flyball goes down? Why did you include the HR/flyball numbers then ignore them?

At home, distance goes down (-3 feet) as does HR/FB (-.047), so that's consistent. On the road, it's true that distance is up (+6) but HR/FB down slightly (-.011). This could be caused by any number of factors, including the introduction of Petco (where CO plays about 12% of its games) and personnel changes (Willy Taveras had a .02 HR/FB rate, Matsui not much better). So HR/FB rate is going to bounce around, especially on an annual basis. And small changes in mean distance probably have a big impact on HR rate. But again, what matters is the comparison of Coors to other parks holding the players constant. And when you do that, MGL's data shows clearly that the "Coors effect" has been dampened in a major way.

This has been raised several times in this thread, but why do it with a computer model and not real MLB players? Or if MLB players were not possible, how about minor league players at least?

Its MUCH easier and cheaper to do it with a computer model in your office on campus than to gather up some baseball players and find a place to actually do physical test.

The composition of the team will change over time, as will ballparks (for away stats). So, for example, if CO hitters have more power now than pre-humidor, the distance of home FBs may stay constant (or even increase) despite the humidor having a real impact. The point is to isolate the impact of playing in CO vs. everywhere else, which is what MGL does.

That's the really interesting point I make in #46. Fly ball distance is so stable when changing other characteristics of the team, it's hard to say that the measurement varies at all. If you accept my estimate of 2.5% measurement error, fly ball distance doesn't change measurably because of the humidor.

Whether the numerator or denominator changes is of course irrelevant, as a "statistician" would know.

It's not irrelevevant at all. The variance of the road numbers is an estimate of how well the measurement process works. If the change of your measured group is about the same as the change in your control group, you're probably getting close to the measurement limit and seeing random fluctuations.

Do you have any theories as to what else is at play here?

The set of trajectories that get classified as "fly balls" in a particular year evenly fill an area of phase space. So when you look at distance per fly ball, you get the same answer every year, unless the laws governing ball motion change, as might happen with changed aerodynamics. But there's a measurement error involved, and that's dominating any change that changed aerodynamics might have.

Fly ball distance is so stable when changing other characteristics of the team, it's hard to say that the measurement varies at all.

It's only stable at home, it changes considerably away. The apparent stability at home could be a stronger team offset by the humidor.

The variance of the road numbers is an estimate of how well the measurement process works. If the change of your measured group is about the same as the change in your control group, you're probably getting close to the measurement limit and seeing random fluctuations.

I have no idea what this means. Can you explain?

But there's a measurement error involved, and that's dominating any change that changed aerodynamics might have.

This isn't a theoretical question: CO and its opponents have actually hit fewer HRs in CO, relative to opponents' parks, since the introduction of the humidor. For example, they hit just 23% more HRs in Coors this year, compared to 79% more in 2000. So unless you have some plausible alternative explanation, the only question is how -- not whether -- the humidor suppresses HRs.

Its MUCH easier and cheaper to do it with a computer model in your office on campus than to gather up some baseball players and find a place to actually do physical test.

Well, yes, obviously.

Let me rephrase: why should people pay any attention to the study, since it is merely based on a model and since it's predicted results appear to contradict empirical results?

Year wins HR SLG dist(home) dist(road)
2000 82 161 .455 336 317
2001 73 213 .483 340 320
2002 73 152 .423 334 322
2003 74 198 .445 334 320
2004 68 202 .455 336 325
2005 67 150 .411 333 324
2006 76 157 .433 332 324
2007 90 171 .437 338 324

Numbers of homers has nothing to do with average distance of fly balls, if these numbers can be trusted. The Rockies might be hitting more fly balls overall, but I challenge you to find anything more than a random walk in the average distance per fly ball columns.

MGL gives calculated park factors year by year, shows that they've fallen since the introduction of t he humidor, and asserts that the humidor caused the fall. That's the claimed empirical result. But the whole question of empirical results is intimately tied up with your ability to measure them. Here you're taking the ratio of two average distances. But those distances can hardly be measured exactly: a spotter or a tracking system is generating them, and the world being what it is, all measurements have some attached uncertainty. So the real question you're trying to answer is "Is the effect that I think I'm seeing large enough that my imperfect measurement can tell me it's not zero?" Almost all questions that a statistician like Walt concerns himself with can be recast as variations of this question.

That's why Walt keyed on the relative changes of the home numbers compared to the road numbers. The reason we want to compare the home distances to the road distances is that we're implicitly assuming that nothing important has changed on the road, while at home we've introduced the humidor and whatever aerodynamic effects accompany it. But when we look at the numbers in the table, we should immediately be worried, because the changes in the thing that's supposedly being kept constant are actually bigger than the changes in the thing that's affected by the humidor.

In the years for which mgl has provided data, the road distance varied between a maximum of 325 feet and a minimum of 317, an 8 foot range. The home distance varied between a maximum of 340 feet and a minimum of 332 feet, also an 8 foot range. Meanwhile, according to comment #59, the Coors field advantage has fallen from +20 feet prior to 2005 to +11 feet since, for a change of 9 feet. But the Coors field addition to the distance a fly ball travels is the difference between two numbers, each of which has its own uncertainty. If we estimate the error on the fly ball average distance as 8 feet, then the error of the home distance minus the away distance is (8^2+8^2)^.5 = 11.3 feet. The error on the difference between the measured changes in fly ball distance due to Coors Field in a particular year is (11.3^2+11.3^2)^.5=16 feet. So the empirical estimate of the effect the humidor has had on the distance Coors field adds to fly ball distance (9 feet) is only half as large as a reasonable estimate of the error in that measurement (16 feet). If we wanted to be 95% confident that a 9 foot change in the Coors Field effect was real, we would need to measure home and away fly ball distances with an accuracy of 2.25 feet, which I for one am not confident has been done.

Let me rephrase: why should people pay any attention to the study, since it is merely based on a model and since it's predicted results appear to contradict empirical results?

Measuring typical flyball distances at Coors and other parks to within an error of 2.25 feet is a reasonably large technical challenge. Since this is all hobby work, you might have to wait a long time to see an experimental study that can actually resolve effects of this size.

#67

Yes, large technical challenge. So, maybe, wait, and don't jump to conclusions and make statements like "the bottom line is that humidified balls don’t change things much, says Meyer. “I’d say the role of the humidor is more psychology than pure physics.”

#66

You should be using the standard error not the max min difference to determine error estimates. It might actually turn out that the distances are significantly different if you do it properly.

#67

That's not true. A simple test would be to just eject the balls from a pitching machine wet and dry at a set velocity and see whether the aerodynamic model being used has any predictive value. Or at least show examples where the model has been used before successfully to predict distances in the past if it has been used for this purpose. It seems to me that this is a hobbyist type of thing rather than serious science, since this is the first paper (preprint?) that either of the authors have done in this field. Don't get me wrong - it's an interesting result but far from definitive or even useful and there is no excuse for overstating the results as they did.

Academicians (is that a word?)

In Soviet Russia...

That's why Walt keyed on the relative changes of the home numbers compared to the road numbers. The reason we want to compare the home distances to the road distances is that we're implicitly assuming that nothing important has changed on the road, while at home we've introduced the humidor and whatever aerodynamic effects accompany it.

OK, now I can see where you went off track. The home/road comparison does NOT assume nothing important changes on the road over the years in question. A lot could, and did, change. The point of the home/road comparison is to compare Coors to a cross-section of other parks. What is held constant (virtually) for each season are the hitters and pitchers, so any home/road difference is the park effect. However, there's no reason to think the road stats should be consistent over time, and the change in this example should not make us "worried."

As for measurement error, the changes in means tells you nothing about the accuracy of the measurements, as Scientist Guy notes. And you are looking at only single-season error estimates, when the pattern is sustained over multiple seasons. I have no idea how accurate each FB measurement is, but given the thousands of FBs in these samples it hardly matters. As long as the errors are randomly distributed (+6 feet here, - 6 feet there), the overall estimate of the mean will be quite accurate. And given the stability of these measures, the error appears to be quite small. In fact, even if the error had a systematic bias -- overestimating FB length by 2%, let's say -- that still wouldn't change our conclusion that the Coors effect has diminished, since it affects both home and road estimates. The only thing that would call these results into question is if we had some reason to think the measurement system has changed at Coors over this time period but not in other parks (or vice versa).

* *

I agree that it's hard to see a clear relationship between mean FB distance and HRs (or HR/FB). You would have to control for a lot of factors to measure it well. For example, the pre-humidor team was majority left-handed, while the current team is overwhelmingly right-handed. That may explain how the team can hit fewer HRs now with only small drop in home FB distance (left center in Coors is 390, vs. 375 for right center).

BTW, the shift to a more RH CO lineup could explain a bit of the power drop being attributed to the humidor. However, CO's opponents have also seen a big drop in the Coors advantage, so the composition of the CO offense can't be the main explanation here.

OK, now I can see where you went off track.

GuyM, the null hypothesis is that the humidor has had no effect. To disprove the null hypothesis, you need to have numbers. The error budget is 2.25 feet per fly ball. That's such a small estimate, I would personally need to see a direct test to believe that the data source was that good.

The idealizing assumptions that people make with park factors aren't good enough to measure an effect this small.

I have no idea how accurate each FB measurement is, but given the thousands of FBs in these samples it hardly matters.

The same spotter is not necessarily used for every fly ball measurement in the same park. The same spotter almost certainly doesn't judge every fly ball in every park. You can't invoke the central limit theorem, because there could be multiple spotters in the same park with different systematic errors.

And given the stability of these measures, the error appears to be quite small.

The effect is quite small.

In fact, even if the error had a systematic bias -- overestimating FB length by 2%, let's say -- that still wouldn't change our conclusion that the Coors effect has diminished, since it affects both home and road estimates.

Again, the home and road estimates are not necessarily the same theoretical distribution. As a counterexample, consider observers A and B in Coors Field, C and D in Yankee Stadium. A and C systematically measure balls going 5 feet too far, B and D systematically measure balls going 5 feet too short. Even with an infinite number of fly balls, you can't give an answer for the Coors Field or Yankee Stadium measurements without knowing the percentage of the time A spots vs B, and C spots vs D. Machine spotters might help, since then you would have one observer per park and could apply the central limit theorem on a per-park basis.

In my experience talking with experimenters, things are rarely working so well that pure statistical error is the dominant cause of the measurement error. In measuring something this small, treating everything you don't know as an independent random variable with expectation value zero is liable to make you think that you can resolve things better than you can.

#72

The central limit theorem is applicable regardless of the underlying distribution - the sample mean will converge to the true mean - you're just arguing that the true mean of the observed fly ball distances is not the true fly ball distance and that this systematic bias is different for different parks.

For the road games, the fact that you have different spotters actually makes it much more unlikely that you have systematic bias in the average of the observations. For Coors field, this is a much stronger argument - that the difference you are seeing is really a change in the systematic bias of the observation rather than any humidor effect.

I think you could control for systematic biases in distance observations by binning distances. The simplest would something like - count the proportion of balls that travel at least x feet and eliminate from your sample anything x +/- some generous error and do this for different x to see if there is a consistent result. We already know that a similar measure - comparison of hr/fb which is not subject to distance measurement error suggests that there is a difference since the humidor came in.

Zach: One last try. All you have here is a WAG that there's some bias in the distance estimates made at Coors. And even such a bias would only matter if it's changed over this time period. And since you've taken no account of sample size (probably around 3000 FBs home and road, 2005-2007), or the distribution of these measures at other parks, you really have no idea how stable they are. Also, a reduction of 9 feet in the Coors effect is HUGE, not small.

I would also note that our null hypothesis is not really no effect, given what we know about HRs. In the three pre-humidor years (1999-2001), CO pitchers'r HR allowed home/away was 145/87 (+86% at home). Over last three years it's been 86/78 (+10% at home). Given roughly 9,000 PAs home and away in each 3-year sample, that's a highly significant change (about 4.7 SDs, using back-of-envelope calculations).

There's simply no question that HRs have been reduced at Coors. If you want to argue for psychology over physics, be my guest, but the change is indisputable.

I agree that the HR/fly ball data seems to show an effect more clearly. And, Garret Atkins in game 163 aside, home runs are subject to very little measurement error. I think at this point we all understand each others' arguments, so we're reaching the point of diminishing returns. Interesting thread, though.