Saturday, May 30, 2009

Magic > Magick

Well, I'm not completely a believer in the title. In fantasy games, I'm usually one to take up spellcasters. I guess I'm a sucker for scholarly people making explosions and creative sorts of battlefield manipulations. This, however, is a post about the real world.

There's a real talent involved in stage magic. It takes dexterity, planning, coordination with Lovely Assistants, showmanship, and an understanding of human perception. I'm especially fond of "street magic," where the magician has to perform under conditions he can't control as easily, and often alone. We know we're going to be fooled somehow, and yet a good magician still manages to fool us.

It's fun in a humbling manner. A guy gets to show off our perceptual flaws with good humor. When we don't know the trick, we can make educated guesses as to how he pulled it off. Maybe he palmed the watch when he was asking for a handkerchief and we were distracted. Maybe instead of sticking our chosen card into your back pocket, he put one in there at the start and manipulated our choice in a separate deck. That sort of "magic" is solvable. We can look at the results and figure out answers. When we do, it's often surprisingly simple and clever.

The other kind of magic woos are fond of usually isn't like that. It isn't something to ponder, just to blindly accept. Questions about conditions and details are discouraged. Don't try to measure it. Don't try to understand it, except for the vague jargon they throw out. And don't you dare suggest there's something going on other than what your eyes see. Your eyes can't deceive you, so trust them.

I like the magic of fantasy worlds because there's usually an implication that there's something to be studied. Arcanists have towers full of books for a reason. Magical power is associated with knowledge, and it's dangerous to let an uneducated apprentice play with your pointy hat for the same reason you don't lend a child your power tools. Fantasy magic requires you know what you're doing. Bring a fantasy wizard's library into the information age, and I wouldn't be surprised if some careful, studious individuals could learn their own spells and give us some breakdown of the principles.

Yet with woos who claim to wield magic or psychic powers, they tend not to be terribly forthcoming. They're just not interested in basic tests that exist just to show that they're doing something science doesn't know about, yet.

55 comments:

James K said...

I like the magic of fantasy worlds because there's usually an implication that there's something to be studied. Of course, there are intuitive wizards in fiction, like the D&D sorcerer. But even then, there's often a linkage of magic to art, like a powerful sorcerer is like a virtuoso musician.

That and untrained intuitves are often described in terms usually reserved for unstable explosives.

Dunc said...

I used to work with a guy who did card tricks... He was pretty good - even when you know exactly how the trick is done, it's usually very difficult to spot. And then there's the sheer technical skill of card manipulation involved... Even the basic skills like doing a perfect interleave shuffle (or whatever the technical term is) require a huge amount of practice. Very impressive.

King of Ferrets said...

I like the magic of fantasy worlds because there's usually an implication that there's something to be studied.Hmmm... I can't decide if the usually applies to Discworld magic or not.

MWchase said...

Well... Mostly the witches do magic, right? Just on a grander scale. The wizards do magick, and it at least tends to have consistent consequences. In terms of too much magick -> the dungeon dimensions. Except, I haven't read recent books as much, so I'm not sure how much that's been phased out.

(Reasonably, it's hard to derive conflict from an ongoing storyline where any failure leads to, at best and at minimum, the horrible death of the main characters.)

King of Ferrets said...

Failure=horrible death by dungeon dimensions? They've killed dungeon dimension creatures multiple times, I think.

Also, when Rincewind went through the portal made by having a Sourceror around, he outran every single denizen of the Dungeon Dimensions.

MWchase said...

Well, Rincewind's thing is getting into horrible horrible danger, then getting back out.

Also, it depends. If it's just wizards who aren't thinking straight, wouldn't they try to use magic on it? That would be... 7A.

King of Ferrets said...

7A? Is that code for "really, really bad?"

I'm not sure even drunkest wizard in the history of the Disc would be quite stupid enough to use magic on something from the Dungeon Dimensions.

Actually, wait, a wizard might do that just to see what would happen. Nevermind.

MWchase said...

7A is code for 8, unless I botched remembering trivia. In the beginning, wizards avoided the number 8, because it had such uncontrolled magical resonance.

King of Ferrets said...

Don't remember reading any books that mentioned that.

MWchase said...

Maybe it was only in The Colour of Magic...

King of Ferrets said...

I don't remember that in The Color of Magic... I mean, they mentioned the Octavo and octarine and stuff in there, didn't they? Those are at least related to the number 8...

MWchase said...

They didn't say eight when Rincewind ended up in that temple place.

King of Ferrets said...

So they only avoided saying eight in a very dangerous highly magical temple to an ancient evil god-thing?

MWchase said...

The way it was introduced, it was a general wizarding thing. There's no room eight at UU. Just 7A.

King of Ferrets said...

Well, they never mentioned it again.

There are a lot of rooms that don't exist in the UU, though. For example, the room where every single lecture happens does not exist.

James K said...

It was an aspect of magic that Pratchett quickly dropped. It was mentioned druid the Temple of Bel-Shamboroth section in The Colour of Magic, but was never mentioned again.

debra said...

BD- What do you know about the "game"...
Sentinent World Simulation?
And please, no arguments ... just knowledge.
Thank you.

MWchase said...

Clarification, please... Where are the sentences supposed to end, there? Is that supposed to be "What do you know about the game, [called] 'Sentient World Simulation'?" or "What do you know about the 'game'? [Is it a] 'sentient world simulation'?"?

Also, how did this come up?

(Never mind that, strictly speaking, "no arguments" is an odd request. Or possibly even loosely speaking.)

Anonymous said...

Oh Debra's back, awesome.

I have a post up about people like you on my blog Debra, should bring back some old memories:

On conspiracy theories

MWchase:

Debra was here a while ago talking about her conspiracy theories and she didn't take kindly to being questioned about them - hence her request for no arguments probably.

debra said...

Is this a site to learn sentence structure?
I didn't think so.
So... do y'all know about Sentinent World Simulation or not?
If you do... I would like to hear your thoughts on it.

MWchase said...

Here is my not-very-researched opinion on Sentient World Simulation: experience has taught us that even merely aspects of humanity interact chaotically. What this means is that a small error in a simulation can magnify itself over and over.

This, by the way, is the true meaning of the butterfly effect. Chaotic systems have 'attractors', states that 'pull' iterative recalculations of state towards themselves. The Lorentz attractor, one of the simplest attractors, is a kind of double attractor that will send the system between two nodes. From a certain angle it looks like a butterfly.

The point is, any simulation that big is bound to have, on average, an attractor for every degree of freedom, I think. Reality is one of those attractors. The slightest error could send the state into the wrong attractor entirely.

Even constant recalibration can't take care of everything. If bad data takes hold, it could be very hard to remove, or it might not.

There are further problems: they can't model everything in real-time perfectly, not for quite some time.

That would take a computer the size of the earth, or bigger.

(And sentence structure is important if it obscures meaning. Ever noticed, sometimes, that giving Google a phrase will have it split across sentences in a few results? Such results are typically useless, for the purposes of that search.)

Valhar2000 said...

The Sentient World Simulation project sounds interesting. Like MWChase, I don't think it wil be the solutin to the world's problems, but it does sound like it's a good idea that should be looked into

Jimmy Blue said...

The Sentient World Simulation sounds like exactly the sort of training tool that state departments, law enforcement and the military should have access to.

The ability to run an infinite variation of scenarios with real world parameters and almost real world reactions is a potent capability.

Now, what do you think of it Debra?

Dark Jaguar said...

It's the S3 program, Simulation for Societal Sanity.

James K said...

Dark Jaguar, if you start going no about turning the console off and needing scissors, I'm going to get very cross with you ;)

As a professional economist and policy analyst I am very sceptical that a system like Sentient World Simulation could have any value. Human interaction is too complicated to be modelled, you'd basically need a real-time feed from the inside of the brains of every person on earth to make sure you had enough data. And since a lot of human activity is chaotic (and I mean that in the technical sense) even a small hole in your data will FUBAR the whole model, and you generally won't find out until its too late. And the worst part is even if you got around this problem, you're still screwed because as soon as you start using it people will change their behaviour in response and your model is invalidated.

Technical system for making easy money on the stock market are uniformly failure and the success of macroeconomic forecasting is poor. These things are both true because predicting the future in the social sciences is essentially impossible unless you're focusing on a very narrow area on a short time horizon.

MWchase said...

And just for the sake of trying to invoke a Doggerel in the name of skepticism, I'll point out that it's unlikely that most of those using it would have access to comprehensive and accurate information on love.

Or food preferences. Or susceptibility to the common cold. Or clothes style. Or even, in the case of the truly paranoid, internet habits. (Some things just can't be subpoenaed.)

debra said...

MW- Face Book has dozens and dozens of personality quizzes. Everybody takes them. They love doing it! People voluntarily give away enormous amounts of information about themselves so that some cool revelation about them can be known to other members. This would be the place to find intel on love, food choices, clothing etc.

James K said...

debra:
That data isn't rich enough. Its not enough to ask them whether they prefer chocolate over hokey pokey ice cream, you need to get a picture of how much of one good a person is willing to trade off for another, for every good that exists or could exist if the prices were right.

And this data would quickly lose its value over time, requiring frequent resurveying.

And since even small gaps in knowledge could prove fatal, you would need to get this information from everyone.

And people won't always answer surveys honestly, often they answer to signal positive characteristics rather than honestly reporting preferences.

So really, no. The data requirements are too large. This debate was had by economists in the 50s between the pro-planners like Samuelson and Galbraith and the free marketers like Friedman and Hayek. Since every attempt at central planning has resulted in Epic Fail, I'd say the balance of evidence is with Friedman and Hayek at this time.

MWchase said...

I was going to say...

In that case, everybody who doesn't use Facebook, or doesn't take all those quizzes, is a confounding variable. In any case, I fail to see how it will help anybody who wants to model me to know that I scored "Water" on the Which Element Are You quiz.

Dark Jaguar said...

Water's not an element, I call shenanigans on that test! I'd try for Uranium!

Dark Jaguar said...

Okay in addendum, I don't bother with facebook. I don't have a mysapce page, and I don't share information with anyone. Further, those tests don't record anything as a general rule. It's a simple script. Even the embed codes aren't linking to any saved data, it just "fills out" the answers you gave all over again and puts that in your sig, though it's beyond me why anyone would put one of those on their page as a mark of pride or whatnot.

Besides those things are bunk! The Final Fantasy quiz always says I'm Kefka! What was I SUPPOSED to do? NOT burn down that guy's house and poison his family?

Debra: I hear its amazing when the famous purple stuffed worm in flap-jaw space with the tuning fork does a raw blink on Hari Kiri Rock. I need scissors! 61!"

Tom Foss said...

Debra, have you ever taken a personality quiz on Facebook? They aren't actual personality quizzes. Actual personality quizzes that psychologists use to find out actual things about your personality are developed using what we actually know about psychology and neuroscience, are tested and verified extensively, and are phrased in meticulously careful ways so that the questions actually measure what they're purported to measure.

The quizzes on Facebook are developed by teens with universally poor spelling, using character traits defined in the broadest possible swaths and awkward questions drawn from quotations and actions to determine which Naruto character best matches your choices from eight nonrepresentative answer possibilities.

If the government is using online quizzes to find out information about the populace, then the government is made up entirely of fourteen-year-olds of average intelligence with too much time on their hands.

debra said...

No Tom, i have never taken a quiz on YourTwitFace. I was on Face for a month and was embarrassed by the posts that my old High School friends were posting, like "Philip is at the Dentist". So I closed my page.

James K said...

Dark Jaguar:
I hear its amazing when the famous purple stuffed worm in flap-jaw space with the tuning fork does a raw blink on Hari Kiri Rock. I need scissors! 61!

That's it buddy! My flying monkeys are on their way right now!

Tom Foss said...

Debra: The quizzes on Facebook are really no different from the ones that have been circulating the Internets since people realized you could combine HTML forms and Buffy quotes. The only thing sinister is in how much time they waste, but Wikipedia and TV Tropes are far worse offenders at that.

Valhar2000 said...

Don't knock TvTropes man!

Bronze Dog said...

Finally getting back online from my new apartment with cable modem.

I didn't know anything about that program or whatever before reading the responses, and I'm along the same general lines.

Human beings are very chaotic systems (specifically meaning that human behavior is very sensitive to a large number of inputs).

As for online personality quizzes, yes, that's incredibly incomplete data, on top of its shallowness, it's especially bad because the subjects know they're being tested. Some may try to bias results. Kind of like I did recently, indulging in a bit of guilty pleasure: Picked up Pokemon: Mystery Dungeon and biased the starting test to end up a Mudkip, since I hear people like mudkips.

MWchase said...

The trick to dealing with embarrassment is to ignore most of your Friends™. Some people will actively try to gather as many Friends™ as possible, but the actual point is...

No, scratch that. I'm just on FaceBook so I can grab my friends' personal information at a moment's notice. (That, and it gives me access to a sometimes-insanely-buggy chat setup. Which I access from Adium, so I don't have to deal with FaceBook's... everything that isn't an actual message in chat.)

Back to magic, I have to echo the sentiment about the skill that sleight-of-hand takes. I've never tried, but the whole idea of something not being impressive because it's based on deception and misdirection... The act creates a conflict between what seemed to happen, and how the world works. The more skilled a magician is, the bigger the conflict they can create and sustain.

And people are all about conflict.

Tom Foss said...

Who's knocking anything? There's a trope about it. And an XKCD comic. I'm pretty sure those are among the rules that come after #34.

Lifewish said...

I agree re fantasy worlds. I think that's a big part of why science geeks tend to like fantasy: because it gives them new material to sink their pattern-recognition neurons into.

Facebook has definitely gone horribly wrong in a law-of-unintended-consequences kinda way. It should have been fairly obvious from the get-go that people would start to see Friends[tm] as a way of scoring social points. It's very similar to the spam problem: people engage in both direct marketing and "friend-farming" in real life, too, but it's massively easier online. The result is that the signal/noise ratio drops like a guillotine.

Dunc said...

The fundamental problem with something like the Sentient World Simulation program is that people are weird. I can't emphasise that enough - really weird. If there are universal laws of human behaviour (and I'm not convinced there are), we don't know what they are yet.

MWchase said...

And like I said... Every variable you come up with is going to matter. This is a system that, in its ultimate form, will easily be dealing with trillions, if not more, degrees of freedom.

Anyway, what it actually sounds to me is more like a crowd simulation, or something, with a high degree of verisimilitude.

Alternatively, this is the problem of strong AI plus the problem of simulating the body, plus stuff like ballistics and other physics stuff, all times several billion.

debra said...

Why would the D.O.D want this? If the data is so extremely limited and in flux, what purpose does the outcome serve.
It seems like working on this project would be the ultimate job for a gamer, no?

MWchase said...

It looks to me like the taxpayers here are displaying considerably more realistic attitudes than the people paying for things, with tax money. Near as I can tell, the people who write the checks figure this will work just fine. ... Somehow.

Bronze Dog said...

Gaming would be one good application for the thing. It'd probably produce more realistic NPCs.

Of course, MW's got the usual good point: There's no shortage of government officials willing to throw money at stupid things that'll be of no use to the government or the people.

Just because the air force spent millions on teleportation doesn't mean they'll accomplish anything. The real world isn't like the Civilization series, where research never ends in failure, or all options have a result.

Dunc said...

Why would the D.O.D want this? If the data is so extremely limited and in flux, what purpose does the outcome serve.

Well, the military have a policy of looking at all possibilities, not just the likely ones... Then, of course, there's the usual bureaucratic turf-wars stuff - departmental funding, who's got the most staff, whose got the shiniest toys...

I've worked in both government and the private sector, and in both cases a lot of stuff goes on that's got nothing to do with the ostensible purpose of the enterprise, and everything to do with managerial empire-building. For example, I've once saw a company bring in contractors on 4-figure day rates (plus substantial expenses) because some manager didn't want to admit that his staff had nothing to do, because he was worried they'd get re-assigned to someone else's team, thus shrinking his empire. So I ended up getting charged out at something like $2000 a day on a four-week contract, whilst that organisation's permanent salaried staff sat around reading books. Funnily enough, that company (a major UK financial services company) is no longer in business. Since the military can more-or-less never go broke, they can indulge their fantasies as long as they like, with the taxpayer picking up the tab. Have you seen some of the crazy shit DARPA gets up to?

Dilbert is a much more realistic portrayal of organisational reality than any conspiracy theory. As always, Hanlon's Razor applies: "Never attribute to malice that which can be adequately explained by stupidity." Although I do subscribe to the later modification which notes that sufficiently advanced stupidity is indistinguishable from malice...

debra said...

Are any of you familiar with the comic book series "Harsh Realm"? I net- flixed the TV series and watching it tonight.

MWchase said...

Checked Wikipedia... The scenario that Harsh Realm is about sounds ridiculous. How would a virtual person influence the real world like that? I mean, I get that you're worried about this, but the computational requirements are... They're insane.

Not only would you need to simulate over six billion people to a high level of detail, you'd also need a real-time or better weather simulation (weather affects stuff like driving, where people walk, and therefore congestion, etc.) Researchers can still get publicity by trying to simulate part of a single brain. If you look at the budget-per-amount simulated on this, there's no way the DoD can afford this thing. If it could, AI would be a garage industry. I mean, I'm not accounting for economies of scale, but I'm also using their entire budget to guess the cost, which is clearly wrong. My point is, 3D printing hobbyism has gone further than AI research, and people are buying stuff that's nearly five times the cost I'm projecting.

debra said...

I'm curious, not worried,
When this country socializes health care, won't our records become government property? And wouldn't that add enormous amounts of info to the DOD World Sentient Simulation project.
Does anyone know what this project looks like?
Is it just data being submitted or its it 3D?
Also, what is a node?

MWchase said...

Well, aside from the fact that there'd be less reason for government insurance to try to get access to health records than traditional insurers do (since government coverage is supposed to be relatively unconditional), all the data in the world won't help without the power to process it, and I just don't think the current DoD budget (which is pretty big) can buy that much processing power.

Bronze Dog said...

The big problem with making accurate simulations of very chaotic systems, like humans, is that for any long-term accuracy, essentially you'd need a computer bigger than what you're simulating.

And meanwhile, they have yet to create a realistic simulation of a single human.

Even getting past all the laws-of-physics barriers to such a wide scale simulation, you can bet if info about such a project got leaked, there'd be riots.

Berlzebub said...

@ Debra:
I'm curious, not worried,
When this country socializes health care, won't our records become government property?

Not without violating HIPAA. Even then, that's just your medical history.

And wouldn't that add enormous amounts of info to the DOD World Sentient Simulation project.
How would that information add anything other than your medical history?

Does anyone know what this project looks like? Is it just data being submitted or its it 3D?
From what I read of it, it seems to give some sort of readout of scenarios (graphs, colored maps, etc.), but I can't find anything about 3D. There is some suggestion that they hope to be able to allow "play" in a simulation, but it's too vague to figure out if it would be virtual reality or at a keyboard.

Actually, everything I've read about it is vague. There's talk of 'predicting', but nothing that talks about the accuracy. How far in the future has it predicted an occurence, how much information do they put into the program, how do they make sure the information is reliable...?

The whole thing sounds like a Sims experiment. Considering the Butterfly Effect, which MWchase alluded to earlier, at best it would be able to perform calculations on likely outcomes for a large populace, and even then the results would be limited to a finite period of time. Each time it started to depart from the current happenings, it would have to be reset and fresh data entered.

It's a cool toy, but that's about it.

Also, what is a node?
I believe this is what MWchase was referring to.

MWchase said...

Actually, I was just being slightly vague, forgetting which word was correct, and thinking of one that kind of works. What I meant was, a chaotic system will tend towards specific states: these are attractors. However, sometimes an attractor is actually several states that the system oscillates between.

I don't think node is the word here, but I was thinking of, like... acoustics and such. Wave theory.

Berlzebub said...

Oh, so you meant this one. ;-)

James K said...

I don't know but I've been told
Diedre's got a network node.
Likes to press the on-off switch
dig that crazy Gaian witch.