Saturday, September 18, 2010

True Randomness

"You don't see THAT everyday..."
In my last blog post, I proved once and for all why we couldn't possibly exist in other universes short of a higher power intervening and no one could think of a single argument as to why that could be wrong.


Okay, so maybe there were a lot of comments with counter-arguments. I have a counterpoint that touches on the basis for several arguments for the very existence of other universes and what would cause them to exist at all, however, and I'm curious how it will be countered, so here we go:

Nothing is random.

Before I clarify that statement, here's why I bring it up: there is a hypothesis that every possible outcome of any given moment results in parallel universes. On the surface, this suggests an infinity of universes beyond even the most smarty-pants nerd's comprehension, with universes splitting for no more reason than a person having two different cereal options for their morning breakfast.

The problem with this is that nothing is truly random. There are tons of things that, for all human intents and purposes, are random, but that doesn't actually mean they're truly random. It just means they are beyond our ability to predict.

Three examples: the shuffling of a deck, the rolling of a die, and randomized damage dealt in an RPG video game.

Shuffling: The order of the cards is determined by what order they started in, how they were shuffled, whether they were cut, etc. While we cannot predict the final order they will wind up in, making it random for all human intents and purposes, it's not truly random at all as far as the universe is concerned. There is a clear sequence of events with every effect having a cause. A being of high enough intelligence, observational senses, and knowledge of the starting order could know the order of the cards after being shuffled without looking.

Die roll: Same as the shuffling. How the die is thrown, where it lands, etc, gives us a result that is, for all intents and purposes, random, but it landed on the number it did due to various measurable factors. There are too many factors for us to reasonably keep track of and know the ultimate outcome, but the universe is smarter than us. If it were a being capable of following all the variables, it would know the result of every die throw in advance.

Randomized computer number: One of the major issues in computer programming is generating a truly random number, and the reason for this is that it has to be based on something. In short, it's not truly random at all. Any randomization of the damage being dealt in the RPG is most likely the result of an algorithm that "randomizes" the number based on the internal system clock. If one were an android with lightning fast reflexes and could precisely follow that clock, one could time their menu selections in order to get the best possible results every time.

The end point is that the idea of multiple universes resulting from something like a coin flip doesn't make sense to me. When the coin is flipped, the outcome isn't actually random, meaning that there is only actually one possible outcome in spite of there being two sides. We don't know what this outcome is, but humanity's inability to predict something is not the same as it being random from the perspective of the universe as a whole.

This extends to human psychology and decision making. Going back to the two cereals to choose from example, there are various factors impacting the decision, more than we can even begin to comprehend if we really start to think about it, but none of those factors are themselves random. The decision will ultimately be the result of countless non-random factors that won't arbitrarily change. Regardless of how many types of cereal there are in the cabinet, as far as the universe is concerned, there is only one possible outcome.

...Wait, did I just prove the existence of fate? o_O

Okay, that was an exaggeration for my own amusement, but it's an interesting thing to consider. It's also worth noting that based on this argument, even if there is such a thing as fate, there may as well not be from our perspective. Our tiny human brains are incapable of accounting for the effectively infinite amount of x-factors, so there's no point in assuming we know what our fates are. I think we should be humble, accept the limitations of our own minds, and do our best to make our own destinies without arrogantly assuming we know what our fates are. We don't. That's what makes life both fun and tolerable.

This is the part where someone will comment about random quantum components or electrons out of sync or something like that. If one can genuinely prove quantum randomization resulting in other universes, I'm willing to consider it, but I would need some sort of real, solid proof to revise my opinion of randomness. At the moment, it sounds more to me like humans concluding that things they are incapable of measuring or predicting are truly random as opposed to just being beyond our ability to predict.


  1. now the problem with your argument is that your attempting to define an area of science (which is still heavily debated by the most brilliant people) and make it an absolute, when fiction says it's going for the infinite multiverse theory, like most of them do it should be obvious that they are using the idea put forth and accepted by most scientists that the quantum world is truly random (as what we've observed puts things on that level as going even beyond cause and effect) trying to say it's illogical and doesn't make sense is no different than saying faster than light space travel is impossible because you haven't observed enough to be convinced that it is, even though there are a number of theories about how to do it and most sci-fi clearly follows one of those theories.

    in other words for your argument about why the multiverse is unrealistic, you'd have to extend that to every facet of sci-fi no matter how hard (I hate that term)

  2. also I point I forgot to mention is that your nothing is random idea would also effectively preclude the existence of any alternate universe were anything resembling humans exist, or universe that are any different from our own, since there differences would pretty much have to stem back to the big bang itself. because if there wasn't something different about that, everything else would develop to be exactly as we know it, or if there was something different then then humans would have probably had a different evolutionary path.

  3. You are right up to a point. Most events aren't truly random. Almost all macro-scale events are not. (There is an argument about chaos theory here, but I don't really know it that well, so I won't get into it.)

    However, quantum events are truly random. That's my the Schroedinger's Cat thought experiment uses one particle decaying or not decaying to create the superposition. But it takes some set up to make the state of the cat dependent on quantum events. And it applies to one or two particles at a time. Predictions of the average behavior or a group can be accurate and deterministic. This is why we can measure the half-life of radioactive molecules and use it to tell how old something is.

    As well, we can prove that there aren't unmeasurable (or hidden) variables that encode what a particle can do. Check out the EPR paradox (
    The true randomness is what Einstein was talking about when he wrote: "I, at any rate, am convinced that He [God] does not throw dice."; however, he was proven wrong with the EPR paradox.

    Also, keep in mind that when conducting the double slit experiment, one photon can go through both slits and interact with itself. No one path is chosen and that means that no split universe, even though if it is measured, it will go through only one slip at random (which is what may (or may not) cause the universe to split). That is, only when a probability waveform is collapsed would an alternate universe be created (if they are created at all).

    Now that was about proving that there are truly random events. Now, what this means for the multi-verse.

    Well, actually, I think it would still follow pretty much what you had in your last post, except that new 'verses could be created intentionally with experiments like the Schroedinger's Cat experiment.

  4. Oh, and you might want to look up the Copenhagen interpretation[1] of QM.


  5. The uncertainty principle. Delta x times Delta p is always greater than or equal to hbar over two. At least, assuming that the Schrodinger equation holds up.

    However, the Schrodinger equation has made some remarkably accurate predictions, even about nasty pain in the arse scenarios, so I'm willing to put my money on it.

    And, if you accept the Schrodinger equation, then you've pretty much got to accept the fact that we can never be 100% certain about both the position and momentum of a particle. We can know probabilistically about both, but that involves some sort of your hated random factors.

    Now, when you get to the macroscopic levels, these random fluctuations cancel each other out, leaving you with good old Newtonian and Einsteinian mechanics.

    But, at the quantum level, we still have those lovely uncertainties.

    For more information, look up nuclear decay. Absolutely no way in hell to predict when a particular nucleus will decay. Ain't gonna happen. Just a probability of it happening in a particular time.

    </physics student rant>

  6. Dammit, Tyr, you beat me by three minutes. Oh well, let's just hope my information adds to yours. I was considering adding the EPR paradox, but I forgot the name and I couldn't find my Griffiths QM book.

  7. I'm still where Einstein was up until he dies - up until now, I can't accept the idea of a non-deterministic universe.

    That being said, as a physics student, I've heard about uncertainty and probability holding up and making accurate predictions, so I'm guessing that at some point I'll have to revise my opinion.

    So, yes Dan, I'm mostly in agreement with you, but I'm willing to consider the idea of probabilistic theories being valid models.

    On the other hand, I think it might be possible for us to at some point come up with a deterministic theory which supersedes (but does not invalidate) Quantum Theory, in a similar way to the way Relativity and Quantum superseded Newtonian Physics.

    ...Yeah, I'm very much on the fence. But my optimism and desire for determinism make me fall in Dan's camp.

    At least until I actually study Quantum Mechanics. ^^;

  8. Between these comments and the last post's comments there is a lot to read, however:

    a) I'm not arguing against ANYTHING happening in science fiction or science fantasy. Those things are fiction. Give me enough technobabble and I'll believe anything.

    b) All scientific theories are basically "fact until proven wrong". I will stand by the statement of something being illogical until given adequate information suggesting that it is logical.

    To specify, what I mean by "illogical" is, based on the knowledge currently available to me, the logic does not add up. This does not mean I cannot be convinced otherwise should the logic be proven or at least supported in some way, but until then, I will consider it illogical. I don't consider that unreasonable.

    c) Again, I do not accept "we are incapable of predicting this" or "we don't know what determines this outcome" as acceptable proof that such things are truly random. I consider those to be less random and more a sign of our own scientific shortcomings.

    An example that would support the point would be a scenario where something that can be done repeatedly under 100% identical circumstances while repeatedly getting different results. I don't know if we're at a point where we can actually do that, because when I say 100%, I MEAN 100% with no possible x-factors. If the variation can be explained in any other way, it fails to prove true randomness.

  9. Recalling the quantum mechanics system, as I recall they proved it was really, totally random. Or I might be wrong. But really, wouldn't it be a bit of a cheat for some god to put something so close to that, and make it turn out to be totally deterministic? And if it isn't deterministic, why would it be put in if it were unimportant? Clearly, it's a bit of code for the universe to account for the free will principle, accepting "they wouldn't do that" for both answers.

  10. Remarkably, I was ninja'd by the author and had the correct response to that argument. Except that there are always X-factors. Different time, place and/or spatial factors are all variables that we can never quite be sure about, until we have some overarching theory explaining that they are unimportant, that has no exceptions or oddities unaccounted for in full or in part. At least, I'm assuming that is the correct way of expressing the concept. So there are no proofs, could be the correct way of saying it.

  11. - digitrev: "Absolutely no way in hell to predict when a particular nucleus will decay. Ain't gonna happen. Just a probability of it happening in a particular time."

    ...That we know of *runs*

  12. And again, other comment-writers above me have proven the point quite succinctly, so this time I will just bring up a snippet to disprove you:

    It doesn't matter if it is random.

    Who ever said that randomness was a requirement for a universal split? All that is required is DIFFERENCE- If I have 5 boxes of cereal that I could choose from, there will be, hypothetically speaking, at least 7 splits, if not many more, from that single decision-- One for each cereal, one for not eating cereal at all, and one for mixing multiple times (which in reality would be a separate split for every single combinatorial possibility.)

    You are acting like you are right because you think we don't know the meaning of random.

    However, the truth is that you only think you are right because you don't know, or willfully ignore, the meaning of 'infinite'.

  13. Ariamaki - My argument is that if there is no randomized component, then the decision is going to based on a variety of factors that we are not even conscious of, including biological, environmental, hormonal, etc.

    If that is true and no randomization is occurring, then the outcome of person X being presented with Y cereal boxes under condition Z is going to be constant. The creation of an alternate outcome would be dependent on at least one variable being modified, hence why I consider a lack of genuine randomization a big deal.

    To put it another way, if every effect has a cause, what cause exists for them to ultimately choose one cereal over another? One does not walk up to 5 boxes of cereal and walk away with one out of the five with no reason at all behind it. Even if they're totally apathetic and just grabbed a box without looking, their behavior still follows a series of causes and effects that could be tracked and predicted if one had all the information and the means to process it.

    I also consider the idea of alternate universes being created as a result of every decision to be somewhat self-aggrandizing. I just held up my hand and chose not to snap my fingers. I find the claim that I just split the universe in twain by doing so a bit ridiculous.

    As for 'infinite'. I know what it means. A problem here is that 'infinite' has multiple definitions, and one can argue which term applies to 'infinite' multiple universes. I'm willing to accept this one:

    "Immeasurably great" - the number of alternate universes, should they exist, are beyond our ability to measure. The exponential creation of these universes is also beyond our ability to measure. The number is so vast and growing that mankind would be wiped out by the death of the sun before finishing saying what would be a long-outdated number of universes.

    Going by this definition, every POSSIBLE universe could exist. Jon Stewart could exist in many different continuities, but he would not be in all of them, nor would we all be in all universes. It would be a cosmically huge mix and match scenario.

    This one, however, I am less willing to accept:

    "Unbounded or unlimited; boundless; endless" - This implies every universe, possible or not, exists. Under this definition, a universe where the world blew up during the American civil war, was reconstructed by industrious space beavers, and ultimately resulted in an exact copy of the world as we know now, except every five seconds we all freeze, then unfreeze without realizing it, exists.

    Now, for all I know, that universe TOTALLY EXISTS, and that would be super cool. However, even if countless universes exists, I don't think that one does. I can't prove it either way, so it's just an opinion, but it's one I think logic has the back of.

    In short, I'm actually pretty open to the idea of the first definition being true if there is actually a multiverse, and short of seeing documentation proving otherwise, I'm assuming "immeasurably great" is what quantum theorists are saying.

    The second one, however, where any reality can exist for no reason, just seems silly to me. It's not something I'm going to accept has sensible from anyone without solid proof.

    Also, of course I think I'm right. I wouldn't be saying this stuff if I didn't think I was. It would be pretty cool if I was wrong, but I haven't heard anything that sounds like definitive proof of that yet. If definitive proof is found, however, I'll probably tweet something along the lines of "cool!" followed by a link to a news post about it ^^;

  14. You know, there are also different sizes of infinities.
    There are more irrational numbers than there are rational numbers, even though there are an infinite number of each. However, the infinity that is the number of integers is not greater than the infinity that is the number of fractions. Just to blow your mind some more.

    Back on topic, you do make a good point. If alternate universes exist, I do not think they would consist of every single permutation of arrangement of quarks and their positions and velocities and waveforms of both.

    As they say about for QM though: if you think you understand it, you're missing something. The behavior of subatomic particles is nothing like everyday stuff, so you *cannot* use your intuition.

    Randomness is required for superposition. Superposition is testable. Therefore, randomness is testable. Look at Bell's Inequality. A testable experiment of hidden variables vs randomness. It proves that the universe is nondeterministic (or non local, i.e. you can communicate faster than the speed of light). There is proof! GO, look it up ('s_theorem)

  15. Tyr - I'll be sure to check that out later, but it probably won't be until after the next comic is done. I will look, though--that's a promise. It sounds interesting.

  16. Ok, I don't want to delay the comic. That would be bad.

  17. Let me get away from the snark for a minute to address this topic legitimately. The problem here is that, in this case, claiming "I don't know" or even "I can't know" is probably more valid than taking one side over another.

    Let's take Heisenberg as an example, just for fun. The Uncertainty Principle says that we can't know all the information about an atom, specifically location versus movement, and that trying to measure one will actually change the other. Assuming that principle holds true, the ability to track every possible decision ever made doesn't exist, as trying to track them will likely change what the decisions were. So, for all we can fathom, randomness as a concept exists if only as an inability to understand the causes on that scale. If we don't know how something is determined, it may as well be random whether it actually is or not.

    How does this apply to multiple universe theories? Well, it doesn't, really. But multiple universes are another area where agnosticism should be the rule of the day. If tracing possibilities isn't truly possible, then neither is finding out whether or not there are truly multiple universes, or if the word universe even applies. Sure, it's an interesting discussion as far as speculative fiction goes, but trying to apply the rules of speculative fiction to reality is a fool's game, at best. Might as well try to build actual Star Trek teleporters with "Heisenberg compensators" while you're at it.

  18. This comment has been removed by the author.

  19. Matthew M. Murray: "So, for all we can fathom, randomness as a concept exists if only as an inability to understand the causes on that scale. If we don't know how something is determined, it may as well be random whether it actually is or not."

    I'm quoting this because it's pretty similar to what I was saying, but with a twist. I was saying that just because we don't know how to measure something doesn't mean it's actually random.

    A point I believe he is making, however, is that it could still be random, but we can't currently prove it one way of the other. Correct me if I'm wrong on that, Matthew.

    My gut tells me it's not random, but my gut also told me to have pie earlier, so I'm not sure if it's the most reliable source. Still, it's what I've got, and I'm going with it.

    Then again, I haven't read that article Tyr linked to yet, so we'll see what I have to say Monday-ish.

  20. While the theory that, for all respects and purposes, nothing is random is an interesting one, I can't really hold it to my world view so nyeah. :p

    Now, it's true that things are never equal, and so there is always information that can be processed about the environment, and about the current and historical state of the brain (in the way of 'preferences'), and all this information contributes to the decision (which is processed by the decision part of the brain as "for the heck of it", since it's not privy to the processes behind the decision). I would guess that, given the *exact* same variables, then the same action would be taken regardless.

    But that's boring, so let's continue. The radioactive decay of an atom is truly 'random' - it is equally likely that it will or will not decay at any given moment - rather than 'chaotic', which is the process you describe in your blog. We can only describe how much any given group of molecules will decay, and even then it's an estimate. According to the Copenhagen interpretation of QM, the atom is in both states until observed, at which point it will fall into one or the other. The many-worlds theory holds that both events happened - there was a split at the point of observation, where in one verse, the atom decayed, and in another verse it didn't. Theoretically, this could add up as you get further and further away from the 'parent' verse, to the point where the same event in two universes will have different outcomes due to the different environments involved. True, there will be a lot of 'junk' verses as well, where little will have seem to have changed, but that's to be expected.

    Of course, it could be that atomic decay is caused by some unforeseen factor. The problem is that it's not possible to prove that anything is truly random, as much as it's possible to prove that there's a not a God and he's not just playing Yahtzee on some kind of grand scale. For the purposes of modern science, the decay of an atom *is* random.

    (All of this is according to my limited understanding of QM, of course. This could be completely rong.)

    Now, if an intelligent species can create a verse-traverser, one could assume it could also build it to scan some basic environs and filter out verses that are too similar. (How the mirror from That One Stargate Episode works, one would assume, otherwise it would take far too long for our heroes to find anything different.)

  21. As it's quick and easy, I'll summarize my two posts:
    Post one:
    "What will the universe tell you about what happens when you make a choice?
    ( What mechanism effects the choice, such that it would be a certain way, would be a better question, I suppose. And I was using nondefinites in the post.)
    Post two:
    "We cannot know anything until we know everything." Something like that

  22. (Back to snark for a moment) Oh, well, if your gut is telling you it's right, Dan, then absolutely go for it. It worked for Stephen Colbert, so it can work for you too.

    Sorry. Had to get that out of the way. You're correct in translating my rambling to "reasonable," though. I'm of the school of thought that says that, if something is unknowable, there's no sense going on about it. There's plenty of knowable stuff we don't know yet, right? Why not put the energy towards that. Seems more interesting to me.

  23. okay, this is really weird.

    I was just reading a fiction novel, Infoquake, that was talking about this exact subject. In a distant future where everyone has nano-machines and nano-supercomputers all over their body, someone comes up with a program that will pre-maturely map out all possible out comes from an action and what variations the human body would need to make to accomplish the desired reaction by controlling the minute muscle control. In short, the MultiReal technology could allow anyone to, say, hit a baseball anyway and anywhere at any type of target they wanted with 100% accuracy.

    Just thought I'd let you know about this strange moment of serendipity.

  24. Matthew M. Murray: "I'm of the school of thought that says that, if something is unknowable, there's no sense going on about it."

    I agree with this in the sense that it's not worth getting into heated arguments that just go in circles due to a lack of genuine evidence on either side.

    I think it can be a worthwhile and stimulating debate and mental exercise, but as soon as tempers flare and voices are raised over something that cannot currently be proven either way, it's past the point of recreational fun and waist-deep in pointless self-defeating sludge.

    It's easy to find one's self in said sludge, but these are debates that can theoretically be had without it.

  25. Of course there are. And this blog is surprisingly tame and civilized about the debate, especially for the internet, which I like. Don't take my smart mouth as attemtping to start a flame war. I really have no intent of doing so. I wouldn't bother posting anything if I weren't interested, after all.

  26. Oh, I didn't think you were! I'm sorry if I gave that impression. I wanted to reinforce your point.

  27. On the subject of randomness, lets try a thought experiment. Take the example I described last time, the radioactive decay of an atom. Aha! you say. The problem is understanding, and a superior species could work out the variables and laws that govern it. Well, how would the species know that they are right? Obviously by predicting when it would decay and then watching it to see that they were right. Repeat the experiment a good number of times, and if they're right, it's proven.
    But here's the problem - it doesn't prove it. Not really.
    Even if they do a hundred, a thousand, a billion tests, there's no way of determining whether one test more would provide a truly random result. You would have to do an infinite number of tests to completely prove once and for all that atomic decay is not random.
    The opposite is true as well. Even if you do get an unpredicted result, there's no way of knowing whether it's completely random or that it's yet another unforeseen variable.
    In this way, we cannot really have absolute facts, only laws that are close enough that they might as well be facts. (This is why asking someone to 'prove' that God or evolution or whatever thing you're opposed to is completely utterly 100% true is a Bad Idea and ultimately serves to display one's ignorance.*) Science is at best the result of averages - lots of really precise averages that work together very well, but averages none the less.

    To sum up, it's a mess you don't want to get into. Trust me, you don't.

    * Note I'm not saying it's bad to ask for proof, just for ultimate-proof-that-accounts-for-everything. Wait, what was the original topic?

    Preview-Edit: Dammit, Jim, I swear those posts weren't there before... Silly ninjas. Well, anyway, I like friendly intellectual debate as much as the next person, even if I do hold strong convictions and get ahead of myself sometimes. May more Internet debates be this civilised, and may there be less things _wrong_ on the internet! ^^

  28. Also, to go off in a different direction than what I've seen skimming the comments so far, I note that this assumes all human actions are based in matter, energy, and the physical universe. It doesn't allow for a nonphysical decision-making component.

  29. Hey Dan, never commented here before (never read before! but follow the comic) but I feel like one big thing you're not acknowledging EXACTLY in your discussion here is that, just as we can't conceive of an infinite this or that, we also can't conceive of true randomness.

    The human ability to understand the universe is fundamentally lacking. Science is ultimately a cumulative tool we use to try to gauge what probably normally happens. So, as our cumulative knowledge currently stands, it's probably impossible to create the absolute frame of reference needed to determine for ourselves whether there is any true randomness in the universe, or whether there are infinite (to us) other universes.

    The question I found myself asking was really: What about the infinite NOT-universes, as well? For every, well, universe, is there a big emptiness?

    Thoughts like this have plagued me since I was a little girl, in all honesty, and they always lead me to this unsettled, small place where I realize how thoroughly I don't know what to think about the universe, or universes, or my miniscule affect on it (or them). I like to try not to think too hard about whether there are or aren't, and I think that's what generally leads us all back to the comfortable fictitious trope that says there are double us-es, and there are Earths in every potential universe, and our planet and our people matter. Because, if they don't, that's kind of freaky, and nothing we do really matters. There's nothing very comfortable about that!

  30. Dan, correct me if I'm wrong, but from your comments it sounds like you are saying "[the quantum stuff that shows randomness] is not logical" when you mean "[it] is counter-intuitive". There is a small, but important, difference between the two; your intuition is based on what you know of the world from other sources, while logic really only describes self-consistency.

  31. Oh, I just had a thought.
    Take a perfect random number generator (RNG) and compare it to a QM random number generator. There are a few attributescz you could compare them on, such as frequency, pairwise frequency and other things. For example a stream of truly random numbers could not be compressed into a zip file that is smaller than itself. (Well, an infinite stream if you want to be all correct about it, as it is possible that a pattern could emerge from a finite subsection of it, but that's not important right now).
    Now in all these tests, the QM RNG is indistinguishable from a perfect RNG, (but a classical software based RNG is not). And if there is no way of telling the two apart, they can be treated as the same.

    Also, I'd like to point out again that QM is counter-intuitive. It does not make sense to gut reasoning. Tiny things behave *freakishly* different compared to everyday sized things.

    And about the proof thing. We have proof. Even if it is a bit conditional. But either faster than light communication or determinism is false. That much has been shown by the Bell's inequality experiments.
    (For the rest of this comment I'm going to assume that FTL communication (of information) is not possible, so I don't have keep keep saying "or FTL communication is false", as that would get tiring.)
    Now for some clarification. What Bell's inequality disproves is that the things have a state when they are not measured. It disproves that an electron has both a position and a velocity, but we can only measure one at a time, but that it has a definite state that we just don't happen to know. That is what they call a hidden variable.
    What happens instead is each particle is described with a probability. And because they don't have a definite state, when they are measured, they _must_ pick one at random. It isn't in one state or another, but both. That is why a photon can go through both slits in the double slit experiment and interfere with itself. That is not possible if it exists in one of the two states, only we don't know which.

    (Now as the wiki link on bell's Theorem is not actually all that readable, I've tried to look up some more.

  32. The whole multiverse thing is a little above the sort of thing I like to discuss/read about. I usually like to stick to the practical.

    Which means that super-androids defeating random number generators sounds way more interesting. Most games actually 'randomize the time' once at a particular event and then use a complex algorithm to generate further random numbers from it. For many games you wouldn't actually need reflexes so much as super-human record keeping skills and an intimate knowledge of how many random numbers are generated and when.

    Some games even use this as a necessary part of the mechanics; off the top of my head I seem to recall that some Fire Emblem games seed the number generator at the start of the battle (probably seeded from total time played in seconds?) and then pull random numbers from that one seed throughout the battle. If you repeat the same moves in the same battle with the same seed it will do exactly the same thing to the last point of damage. In versions where you can save during a battle, the seed is saved as well so you can't reload after a bad damage roll and get a better one; you need to somehow make a different move or change the order of moves to get the numbers to change.

    On the other hand, even a super android who can predict this sort of thing due to exact timing and memory could probably be confounded by a programmer who expected super android exploits to become a problem. Complex conditionals and re-seeding at odd intervals could quickly make it nearly impossible to guess anything useful about the process. Even with a good understanding of how the random numbers get generated, it would probably be impractical even for an android to predict the outcomes during actual game play. Of course if the android had access to the code he might just run the game in his own memory space and he'd be effectively omniscient on all matters concerning the game.

    Ah well, maybe I can't defeat the androids after all.

  33. Oh, I know of one game (Sins of a Solar Empire),
    where, for network play, they had two separate RNGs: one for pretty stuff (random explosions and the like) and one for all the mechanical stuff. For net games, both people would have the mechanical RNG seeded the same and they would remain synchronized throughout the match. I remember there was a bug where someone used the pretty RGN for some stock market code that ended up desynchronizing gameplay very rarely.

    Also, in the fire emblems for the GBA, they used the RGN when plotting paths when it has to choose between going, say, down then left, or left then down. So it was possible to game the system be reloading then forcing it to use up the bad rolls, then continuing. It's considered cheating though.

  34. "What happens instead is each particle is described with a probability. And because they don't have a definite state, when they are measured, they _must_ pick one at random. It isn't in one state or another, but both. That is why a photon can go through both slits in the double slit experiment and interfere with itself. That is not possible if it exists in one of the two states, only we don't know which."

    Actually, Bell's theory states something even stronger than that. It states that no locally deterministic system can generate all the predictions of quantum mechanics. Which means not just that particles must choose "randomly" which is actually a *local* decision, but must choose based on *non-local* factors. If entangled particles were free to choose randomly, they wouldn't appear entangled. In fact, Bell's theorem doesn't prove the existence of randomness: if anything it proves that many things that appear random cannot be random.

    Quantum randomness actually comes from the fact that the uncertainty principle guarantees the measurement of one correlated metric guarantees erasing information about its correlated partner. So if you measure position, you are guaranteed to randomize momentum and vice versa. The best source for quantum randomness doesn't come from Bell, but rather from information theory as applied to quantum mechanics.

  35. What I was trying to get at was the Bell's Theorem disproves that reality exists separately from being measured, i.e. no hidden variables. I was using Bell, as that is the best known (to me, anyway) experimental proof of quantum mechanics. Maybe I should have gone into the double slit experiment. Can you think of any good experiments?

  36. Well, I can think of a couple of papers on the subject. This paper: outlines the thought process by which quantum indeterminancy is equivalent to "strong" randomness, i.e. incomputability.

    I should point out, though, that the many-worlds hypothesis isn't based on "decisions" or "randomness" but rather waveform collapse. Which means the correct way to address Dan's comments in the blog is to say that he's actually correct: such macroscopic "decisions" do not create alternate timelines because such macroscopic events are not quantum superimposed. Only when a quantum superposition "propagates" up to a macroscopic level do alternate timelines get created. For example, if you conduct a schrodinger's cat experiment, you will bifurcate the timeline because the cat ends up being in a superposed state: observing it creates a waveform collapse that propagates from radioactive decay to live or dead cat. But "classical" events don't do that under the many-worlds viewpoint.

  37. Yes, I agree with that. It's not the event that creates the split, but the observations.

  38. Something kinda amazing about Quantum Mechanics. There have been a number of instances where it was pretty much thought that some of the counter-intuitive stuff must just be artefacts of the math used, or approximations, and that things like this nebulous wavefunction that mathematically contains all the information about the particle depending on what equations you bracket it with can't really be how things work, can it?

    Then experiments get done that find out that yeah, that wasn't a mathematical artifact, that seems to be how things actually are.

    Heck, the idea of discrete energy levels, minimum amounts of energy difference that you can't get between, started with a guy trying to integrate blackbody radiation stuff. He basically said "well, I'll work the equations with a delta E, then send that to zero". When the difference in energy went to zero, the equations gave incorrect results. When it remained small but existent, the equations worked.

    When things get really small, our normal visualizations of things simply don't work. You can imagine a ball spinning at 60 RPM, or 61 RPM, but the idea that it is physically unable to spin at 60.5 RPM just feels wrong, but it works.

    Science: It works.