Comic 105 - Review B

19th Sep 2011, 11:54 PM in Rain
Average Rating: 5 (3 votes) Rate this comic
<<First Latest>>

Author Notes:

view 19th Sep 2011, 11:54 PM edit delete
view
Damn it, another talky page. I try to avoid these, but sometimes I can't seem to avoid it. Miwa's chatty, is the problem.

Anyway, it's time to make it clear: although she hasn't bothered to really show anybody yet, she's a real mental heavy hitter. While it may seem somewhat odd right now, by the end of the week, you'll undoubtedly know *why* she's so good at this.

BTW, the satellites aren't hers.

Comments:

jst56strong 20th Sep 2011, 1:08 AM edit delete reply
is it boink time yet??????
UmberGryphon 20th Sep 2011, 1:22 AM edit delete reply
Miwa hasn't heard of the butterfly effect, apparently. Imagine you had a network of sensors that told you everything you could possibly want to know about the atmosphere, all 1 feet apart in a giant 3-dimensional grid that covered the entire atmosphere. And you had infinite computing power to process all that data.

You still can't predict the weather a month in advance, because something happening between those 1-foot-apart sensors will turn out to be important.
view 20th Sep 2011, 2:05 AM edit delete reply
view
Your opinion, fortunately, doesn't affect how the universe works.
UmberGryphon 20th Sep 2011, 5:54 AM edit delete reply
Even my facts don't affect how the universe of this story works.
Valatar 20th Sep 2011, 4:30 AM edit delete reply
Nanites did it.
view 20th Sep 2011, 12:52 PM edit delete reply
view
I didn't hear any facts.

There's no way a car can be predicted, it has too many complicated parts. Even if you put sensors along every inch of it, things that happen between those inches will be important. You can never know if, how, or in what direction your car will run.
UmberGryphon 20th Sep 2011, 7:28 PM edit delete reply
You can design a system that does not have chaotic properties. And car designers do this.

Facts? OK. In 1961, research meteorologist Edward Lorenz was using a very primitive computer to create a very primitive model of weather. He wanted to re-run one set of calculations, so he re-entered the numbers from his printout. The new run diverged from the old run after a few weeks of virtual weather, and this happened because the printout said 0.506 where the number inside then computer said 0.506127. He then came to the realization that any inaccuracy in measurements (one part in a trillion, or even less) would cause any formula even vaguely similar to the ones he was using to change its results eventually.

He popularized this revelation by saying that whether or not a butterfly flaps its wings at a given second must eventually change the weather on a global scale, and calling this the "Butterfly Effect". See https://secure.wikimedia.org/wikipedia/en/wiki/Butterfly_effect .
view 20th Sep 2011, 8:51 PM edit delete reply
view
The butterfly effect is a myth. The idea that uncertainty compounds is not a myth, but if you get the error low enough, you can go a lot longer before the compounding gets crippling.

This is a future where they have a lot more technology to throw at the problem and a lot more information about how weather works. I was going to put in more text like "barring some catastrophe like getting hit by a huge meteor..." and "notice the uncertainty goes up as it goes on..."

But, as it turns out, that just makes an already talky chapter more talky.

I'm not going to claim she's always right. But I am going to claim she's a lot more right than you think is possible.
UmberGryphon 20th Sep 2011, 10:37 PM edit delete reply
The problem is that in nonlinear systems like the weather, uncertainly compounds exponentially.

My favorite story that demonstrates how insane exponential growth is: Imagine that a line that represents e^x is being drawn on graph paper where the grid lines are 1 meter apart, this magical graph paper is suspended vertically, and the pen is moving rightward at 1 meter a second. By the time the pen has moved right 20 meters, it is above the moon moving faster than the speed of light.
view 21st Sep 2011, 12:27 AM edit delete reply
view
Look, that's like saying that if you put $1000 in the bank at 3% compound interest then in five years, you'll have eight billion trillion dollars because exponential growth is magic!

If the exponent is low, it takes a long time for it to get out of control. She's reduced the uncertainty to that amount... or, at least, she thinks she has.
viscount_tusc 21st Sep 2011, 3:26 AM edit delete reply
I'd have to side with forview. The things we could do in the 1960s and the things we can do now is regards of simulating complex systems of many a type are remarkably different. It's entirely possible that weather predictions could be as accurate as Miwa believes they are in the near future, provided enough grant money was dumped on the initial R&D. And, hey, they've got self-aware robots (right?) so it's definitely much further than the near future.

At the very least, it's nice to see an argument on the internet that doesn't have trolls mucking it up.
JasperWB 20th Sep 2011, 2:17 AM edit delete reply
JasperWB
Well I at least still hate her lol. She reminds me of everyone in any project I've ever worked on anywhere. No matter what they know infinitely more than me. I'm not even qualified enough to move empty boxes for them.
Warmachine 20th Sep 2011, 5:11 AM edit delete reply
There is nothing in the universe that cannot be represented by a sufficiently complex math problem.
Guest 20th Sep 2011, 10:15 AM edit delete reply
Human thought there's one
view 20th Sep 2011, 12:53 PM edit delete reply
view
Bwahahahahaha... Gretel.

What is it with today's comics bringing out all the people who believe in irreducible complexity?

No such thing.
Greeb 20th Sep 2011, 7:58 PM edit delete reply
The butterfly effect implies one effect causes future effects, which is true enough in a closed system where the entire system is in balanced. Say for example, a pool of undisturbed water. A single drop into that pool creates ripples which go all the way to the furthest edges of the pool. However, as well as this works in theory, in practice, systems are rarely closed. The enviroment is effected by billions of minute abberations. The thing is that given enough data, the enviroment's reaction to these abberations can be predicted. We do this in a very general way on a humid spring day and have the gut feeling that it'll rain, because it sometimes does on humid spring days. But with more data at our disposal, we could predict the chance of rain a day, a week, or even a few weeks out.

Granted, Miwa's predictions, if accurate, are far beyond what we could possibly accomplish with our current knowledge; however, this doesn't mean that it's completely out of the realm of possibility; and in fact, given the rates of our technological growth, it's likely that similar predictions could become commonplace within the next century.

Great comic btw.
view 20th Sep 2011, 8:53 PM edit delete reply
view
This is pretty much my thinking, too. Miwa's not 100% accurate, but she's built in the uncertainties very precisely, so there are few surprises... barring a meteor hitting and taking out Moscow or something. Even that's not likely, since monitoring of potentially catastrophic things is also much improved.
WhiteAier 20th Sep 2011, 11:49 PM edit delete reply
I personally, don't see a problem with the science in this page. The whole there being a parasite/gas/bioweapon that melts people on contact works; then surely better weather prediction works in this future as well.
duLapel 21st Sep 2011, 3:42 AM edit delete reply
duLapel
This weather science accuracy strikes me a more of a plot driver than anything else... so to paraphrase the late Robert L. Forward... Ignore the science if it spoils a really good story.

As to whether weather predication can get that accurate given a good model and enough computer brute strength... well Arthur C .Clarke's first Law of Prediction says it all: When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.
elfolampo 24th Sep 2011, 9:42 AM edit delete reply
elfolampo
Two things:
1)MST3K mantra.
2)It's interesting seeing how THIS page is the one with the most heated discussion.
LazyRanger 8th Sep 2012, 12:30 AM edit delete reply
Hmm...with weather forecasts this accurate, mail delivery must be really bad! :P

Permalinks: