heh, in this day and age the alien would be reported to homeland security and his money seized. Can aliens from another plantet own money?

"In the open box there is a thousand-dollar bill. "

Take the closed box only. Theres not such thing as a "thousand=dollar bill"

Destroy both boxes. Remind him that people are always more complex than "superior beings" can easily predict.

Kill the alien, take the $1000, sell the corpse to FoxTV for an Alien Autopsy TV special, and use its spaceship to conquer Botswana for its rich diamond industry.

If you don't take the closed box you will always wonder what would have happened if you did, so take it. It's worth $1000 to put your mind at ease forever, and you may get $1 million!

The being claims that he is able to predict what any human being will decide to do.

(emphasis added)

All that's required it to remove our decision process from the game and you'll have a 1-in-2 chance of winning extra money.

How's that? By introducing chance into the equation, we remove the alien's ability to predetermine our chosen action.

For example, with a coin and a sharpie, we could write "both boxes" on one side of the coin and "closed box" on the other. Flip the coin and follow the directions.

You have to take only one box.
Why ? Because you are doing the experiment twice. First you do it as you are being simulated by the alien, then you do it again in the real world. The best would be to take only one box the first time and both box the second time. But since you don't know if you are in the real world or just in the alien mind, you can't take the risk and so you have to take only one box.

To my thinking, there is absolutely no paradox here.

Case 1)

If the alien (even though he has claimed to have predicted your actions) has already placed a certain amount in both boxes before you make your decision, and nothing can change that state once you make your decision, than the only logical answer is to take both boxes. There is absolutely no case whatsoever for taking one box (other than maybe to be proud enough to say that you did something other than what the alien predicted). The reason being that you will always get more or the same amount of money by taking both boxes. This is true regardless of how he made his prediction or what it was.

The above case seems to be the one described and to me there is only one correct answer. The dominance principle is the only one that makes sense in that case.

Case 2)

If there is a chance of the amount changing after you make your decision, than there's an element of chance. You could attempt to take the one he didn’t predict so that he will put the million dollars in the box. The correct decision in this case is up to the odds of correctly guessing the box that he didn't predict. If the odds are worse than 1:1000 than it's only worth taking the open box.

But this isn’t the case that was described.



The point is that his prediction means nothing in the case described (case 1). His prediction can be disregarded because it has no bearing on the outcome.

The only case in which you must bear in mind his prediction is in case 2. I didn't see it presented as an option

Am I missing something here? It seems pretty simple to me.

Kiekeben is wrong about how scientific predictions work. He describes a process of induction, wherein one moves from singular statements about observation to general statements which provide predictions - a process which can only function if some principle of induction is assumed (which would be distasteful :-P). Another common version of the inductivist view is that each time the same observation is made it becomes more likely that it will be made again. This is also wrong; how could each sunrise be more probably (i.e. higher percentage chance of occuring) for each sunrise which has already happened?
A better description of scientific predictions is that based on observation we postulate universal statements which we assume to be true. We then deduce predictions from our new universal statements and compare those predictions to new observations. If the predictions are incorrect, our universal statement is shown to be wrong. If the predictions are born out, then we can use our universal statement to make more predictions (and to create technological applications of the theory) until the universal statement is falsified by experiment. If the stament/theory were true, that would never happen, but we have no way of knowing for sure that it wont happen eventually.

I don't think my nit-picking affects Kiekeben's eventual decision.
Effectively, the alien is testing how trusting each of his subjects is. The subjects are rewarded with a million dollars if they trust the alien. I say take the closed box, but don't trust him. He's an alien; his real motivations are probably totally inscrutable.

Ryan Smith
notes from $500 to $100000 were used,last printed around 1934
look them up

or try here
http://cgi.ebay.com/1934A-1000-RICHMOND-FE...3417QQrdZ1QQcmdZVie Item

Ryan Smith: $1000 notes were last issued in 1945 and stopped being circulated (officially) in 1969 -- but some still exist, and they are still legal tender.

Now, regarding the problem....me, I will take both boxes, thanks. I get $1000, whether or not said alien is telling the truth, and that's good. In fact, I'm not bucking for getting the highest amount of money, because of external factors -- I'm a dual US-UK citizen living in the UK. As it stands, I only pay UK taxes. However, if I ever make more than $80,000 in a year, I invoke s***loads of US taxes, on TOP of my UK taxes. Frankly, I just can't be bothered dealing with the IRS.

If I'm wrong, and the closed box actually DOES contain $1 million, then I can afford an accountant to help me offshore it quickly.

Either way, I don't lose out.

And thus do I thumb my nose at predicted motivations. ;-p

The question is if the 999 humans that were tested before all had the same information about the problem as you have now. if they didn't and were not informed about "the catch" i.e. the previous outcomes, their decision is a simple one: $1000 for sure and a 50:50 chance for 1 million; they of course take both boxes. so the alien 999 times predicted they would take both boxes. so the other box was always empty. so you have to take both boxes because $1000 is as much as you can get.
in fact we have no information about wether the other 999 ppl had the same amount of knowledge that we have. this aspect of the problem is however an essential one as i have demonstrated.

As Robert points out, this is not a logic problem. The fact that the predictor of your choice is so good at it is a *red herring*. All the hokum and hand-wringing about omniscience, backward causation, free will and the predictability of human behavior are *red herrings*! Take a close look at the payoff matrix, as a function of the relative dollar amounts and the predictive accuracy of the predictor. How good does he have to be in order for the paradox to arise? Not very good at all! Any psych grad student with a 20-question survey about your attitudes towards games and risk, that will take you under 5 minutes to fill out, will do just fine. Any reasonably streetwise con artist can also probably figure you out in under 30 seconds.

All it really boils down to is which arguments about maximizing expected utility versus maximizing minimum winnings sway you the most. It makes a difference whether you're risk-averse or risk-tolerant. It also makes a difference how many times you get to play the game; if you get to play many times, then maximizing expected gain may make a lot more sense than maximizing minimum payoff from a one-shot trial.

What's very interesting about this problem is not that there are two contradictory approaches to deriving a "rational" coursre of action, but that there are two camps of people that strongly adhere to the one camp or the other, and "rational" arguments don't work very well at causing people to switch camps. For some people, the negative impact of "losing" an otherwise certain $1000 has a huge negative utility that can swamp the dollar amounts involved. Those people are risk-averse; uncertain value is very steeply discounted against certain value. For others, the uncertain chance of winning a large amount of money greatly outweighs the uncertain chance of not winning anything at all, and the expected gain calculation is not seriously distorted by non-monetary considerations. This just goes to show that the expected utility payoff matrix is too simplistic an analysis of what human rationality really is.

Closed box, for sure. The paradox doesn't matter. It's how one values money. What is 1000 dollars? In LA, it's a month's rent. I have a month's rent. Another month's rent isn't going to change my life. Now, a million dollars? That's another matter, entirely. That's the opportunity to quit my job and "Live The Dream" as a full-time BoingBoing and Slashdot reader.

 Since I am a theologian by training with an interest in economic
theory, it appears to me that Newcomb's Paradox is just the window
dressing on the theological and philosophical problem of
predeterminism and foreknowledge, muddled together with an economic problem of rational-choice.

To answer Newcomb’s Paradox, it is helpful to describe the problem in
economic terms. The problem of rational choice was described to me by
economist Peter Boettke. In the Economist's Paradox, the researcher
solicits participants by offering them one cent out of a dollar.
Since something is logically better than nothing, rational-choice theory
dictates that the participant should choose the penny. Invariably,
however, participants try to bargain the researcher up to a 50/50
split. The outcome has occurred wherever the experiment is tried,
regardless of culture or other controlling factors.

Since the alien claims to have a prediction capacity of 100%, his
interest in the experiment is high. However, any solicited
participant has a high participation threshold. Intuitively, I believe most people realize that the “opportunity cost” (the intangible valuation of a person’s time, motivation, or experience) of participating in the
Economist’s Paradox implies that the researcher has overvalued the
experiment: the supply of “research units” ($1.00) is higher than the
demand; the money is simply a way of keeping score.

The participant was indifferent to the researcher’s problem before it
was presented, so what benefit is gained by simply accepting it as a
given? The researcher must lower the cost of participation by
increasing the incentive. The problem doesn’t change no matter what
amount of money is affixed to it. To put it another way, if you
consider an abstract unit of absolute supply (1), and a unit of
absolute demand (0), then the point at which supply and demand meet is
always (0.5).

Thus, one way to answer Newcomb’s Paradox would be to bargain the
alien up to $500,500 in the open box in order for you to fulfill his test.
If you choose both boxes, you have a 100% chance of gaining at least
half of the money. However, if you choose the closed box, you have at
most a 50% chance of gaining the same amount.

It doesn’t matter if the alien claims he has a 100% chance of
predicting your choice. From your point of view, you know with
certainty that he has at least half of the money to offer. Suppose he
had lied, and possessed only a maximum of $500,500? Then you have
already gained a 100% chance at all of the money. But if you take the
closed box, you have a 50/50 chance he is lying.

Of course, if the alien refuses to bargain at all, then you must
figure your opportunity costs into the process of convincing him to change
his mind. This will be individually determined by each person’s
experience and self-valuation. After a certain point (depending upon your
patience), it will be simply wiser to choose both boxes and move on
with your life.

The problem doesn't say that the alien CLAIMS that he guessed right 999 of 999 times. The problem states that the alien DID guess right 999 of 999 times. This being the case, you don't need to make any kind of assumptions of whether the alien is being truthful or not. Many folks are making some assumptions that just aren't in the problem.