Quote: Original post by Zipster
The Predictor could lie, but since all his predictions are eventually proven either correct or incorrect by the final results of the game, there's no way he could falsify a large number of his predictions and gain a reputation for being "almost always" right without observably breaking the rules given to the player. And if the rules given the player don't actually correspond to how the game is played, then all bets are off in the first place.

Yeah, I dont see him lying as being very likely myself. But there were a few responses here along the lines of "Well he couldnt really be able to do that", so I figured Id present the options for that scenario as well. The initial problem presentation just says that the Predictor is "somehow presented" as having this ability, so it doesnt hurt to cover all bases.

Either they can predict your decision, or they cant predict your decision, but either way theres no reason to choose both boxes.

If the rules are correct: Choose box B.
If the rules arent correct: All bets are off, so you might as well choose box B anyway.

I didn't understand the paradox to begin with, since I thought the obvious choice would be B and that's that. If you could choose between $1000 and $1 000 000, what would you choose? With that in mind, the predictor would say:

Prediction(human choice)=B

Humans will almost always choose B, so the prediction will almost always be right.

But, now humans can reason that there will always be $1 000 000 and in all their greed they want that extra $1000, so now they will choose A.

Prediction(human choice)=A

We now realise that the predictor isn't going to put that $1 000 000 in there, but we want that $1 000 000, so forget that $1000, pick B.

Prediction(human choice)=B

and so on.


I would choose B, except that the predictor doesn't know me and isn't basing it's prediction on me- so that raises the question, is human greed almost always going to win out? The deeper you go into the thought cycle that I suggest, the more doubt and uncertainty there is in which prediction will be made.

now I read the thread
I want to apologize to pinacolada, perhaps my replies seem aggressive attacking (which I am+do do,and eye dodo :) )
but I can see u also operate on logic, thus allow me to buy the first beer (if we ever (*)meet), so sorry mate pinacolada

(*)in spirit of the thread that would be <5% chance

btw - never been to vegas (but have flown over) at night I must say it really is a city of light that just bursts out of the desert

Quote: Original post by DukeAtreides076
I'd tell the Predictor to construct the set of all sets that don't contain themselves, and then I'd steal both boxes and leave through the backdoor.

We have a winner! Not sure if you are being facetious, but this is in fact the only sensible answer.

I think the logic here is flawed

the predictor simply predicts that you change your mind, if it's flawless. so therefore you have no choice (any choice is predicted beforehand).

obviously this is an impossible scenario, so the whole paradox is absurd.

and if it's not flawless than it's random and you cannot clarify it's rate of rightfulness. you can't say 99% o 1% because it's not based on anything - so you are simply gambling with the predcitor not involved at all

This paradox isn't terribly hard to resolve.

The problem statement says it doesn't matter how the predictor operates, so we are free to make an assumption. Let's assume it does a simulation in software of you making the decision.

Now, when you are making your decision, you have no way of knowing if you are in a simulation or the real world. Therefore, the argument "what's in the box is in the box, so I might as well go with option A" doesn't hold. You are forced to choose option B if you want the big payoff.

Quote: Original post by rein
I think the logic here is flawed

the predictor simply predicts that you change your mind, if it's flawless. so therefore you have no choice (any choice is predicted beforehand).

obviously this is an impossible scenario...

I offer you to make a choice between a: losing finger, and b: walking away with million dollars (not losing finger or anything). Will you choose a or b?
I can be quite confident you choose b, confident enough as not to even bother buying axe for executing option a.
In Newcomb's paradox, I can do psychological screening and determine if your brain is organized so that it arrives at "take 1 box" or "take 2 boxes" answer.

Quote: Original post by zedz
now I read the thread
I want to apologize to pinacolada, perhaps my replies seem aggressive attacking (which I am+do do,and eye dodo :) )

Oh, no worries :) I don't think you were any more aggressive than an average internet forum dude. We all sound a little more cranky here than we would in person :)

BTW, I didn't post what answer I pick...

Firstly, the paradox is not paradox anymore once you're talking of some specific predictor. Considering 2 cases:
A:predictor uses correlations of your prior history with outcome.
B:predictor simulates your brain. (you cannot outwit predictor)
(Predictor that is almost never wrong would have to use method B.)

So, the answer is "I'd pick 1 box."
In case A, if predictor happen to includes this very post as part of my history, surely I'm better off telling that I'd pick 1 box. In case B I really should choose 1 box because I might be in that simulation, and its just prisoner's dilemma with myself - its better to cooperate.

(actually, that paradox is probably the only philosophical paradox that I like, even though it turns into non-paradox once you refer to any specific predictor)

Another way of thinking about the paradox, is that since the predictor always makes predictions with very high accuracy, we can say that no matter what choice you end up making, that's the one the predictor would have predicted. In other words, we need to reverse our inherent notion of causality, which I think is the crux of the paradox. Instead of thinking of cause being in the present and effect being in the future, we think of cause being in the future and effect being in the present. In that case, it always makes more sense to chose box B, unless box A contains enough money to make you happy enough to settle [grin]