A challenge I intend to put somewhere in a story (possibly)

You have the robot sidestep its 50/50 chance of outputting a false answer, by verifying its output. (What happens to that new output, would there also be a 50/50 false chance imposed on it as well?)

This phrasing I bet is what threw many people off, it seems wonky (though I get what you're trying to mean) - maybe you can reword this?

Let's see...

"In your current state, the output for X is true".

I can't see anything wrong with it.

OK, now the player knows that the robot knows when the bug is happening; we're also assuming that the bugs are deteriministic but not predictable and that whether a bug happens or not is not affected by what question is asked, and that the player knows all this.

We're really piling up the necessary parameters to make this a workable puzzle.

FUCKING NINAJED

I'm not even considering the bugged situations for now. For the logic problem, "this statement is false," for all you Portal fans, blows out the circuit, resulting in a "wud the fuck" answer from the robot. More elegant than 0/0—if you want to think that it's a smart robot and stops if it has enough information to evaluate the expression (the first statement of an OR is true or the first statement of an AND is false).

Um, I think I'm misinterpreting your answer, then. Given that the robot goes right-to-left, the one that I'm thinking of is...

(1 OR (2 AND "This statement is false")) OR (Not 3).

Which is, in Plain English...

Are either of the following two conditions true?

• 1. Are either of the following two conditions true?
  • a. The cool door is the first one.
  • b. Are both of the following two conditions true?
    • i. The cool door is the second one.
    • ii. This statement is false.
• 2. Is 3 not the cool door?

edited 4th Oct '11 10:08:40 PM by Chubert

Hmm... Oh, I guess I made it seem like the player doesn't know about the bug. Sorry.

Doesn't "the robot knows" cover all that's necessary?

I'm going to sleep now.

edited 4th Oct '11 10:09:09 PM by Teraus

Teraus, it might work without the bug, but there is no way on God's logical Earth that it works in the bugged scenario.

Did you read post 174?

Fuck, I get it now. And, like jewelleddragon, I don't like it. It takes a hypothetical situation and asks itself if the bug would be happening in that situation if you asked the same question at the same time. That is not random, though, at least not in the strictest sense of the word. I honestly think this is a legitimate problem if you just don't include the bug. People solving it is not a bad thing! Having your answer be affirmed by the author is always nice.

I still have trouble understanding the answer to the bugged version. From the original post, I thought if the bug was occuring, the robot would give to opposite answer to what it is trying to say. So how come

?

Right; it must be a seeded pseudorandom bug, because a truly random bug might not do the same thing if you rewound time and played the scenario again the exact same way.

I have a much bigger problem with your answer than "It's math, not logic". It's this:

You're giving the robot instructions about how to answer in the case of a point of ambiguity.

When I tried that you said I couldn't.

edited 4th Oct '11 10:44:18 PM by Madrugada

Uh, right. What he said. I know what a "seeded pseudorandom bug" is, trust me.

Could somebody please explain to me what it is the bug does exactly? Anyone? I'm still confused.

Edit: Yeah, I remember Teraus saying you cannot give orders to the robot.

edited 4th Oct '11 10:50:58 PM by Nightwire

The bug was said to randomly invert the output of the robot.

The "if I said X you would say true" works like this:

If the bug isn't in effect, it works fine. If it is in effect, the logic is that you're asking it" if I said X under these circumstances when the bug is occurring, you would say true. This circumvents the bug, because a true statement will output to false and then be inverted to true and vice versa.

The way he explained it in the first post, the bug forces the robot to give the opposite answer to the one it determined was correct. The robot knows it's happening; it knows what the correct answer is, but it can't help giving the wrong one. Under those circumstances, it makes sense that the bug only kicks in after the answer has been determined but before it's announced.

edited 4th Oct '11 10:54:30 PM by Madrugada

True, but now the person talking to the robot has 2 things to consider:

1: Any answer the robot gives regarding a hypothetical situation is taking the bug into account, whether or not it can predict a bug before it happens.

2: The person asking can never know if a bug has occurred.

Abandoning the doors for a second, let's imagine that you (truthfully) say, "If I said I like ice cream, you would say 'true.'" there are a couple ways this could go. If the bug can't be predicted, it can't answer, because it can't know whether or not it will say "true." But if it can predict the bug, it will know that it will either lie or tell the truth. It will respond with either true or false, but the person asking still can't know whether it was accounting for a bug or if it knew it wouldn't bug. Not only that, but when the robot does answer, there's a chance of the bug applying to this answer.

So let's say that the robot predicts a bug. It knows it would say false, so it prepares the answer "false." But if it actually becomes subject to a bug, its answer will again be changed to "true." But there's only a 50/50 chance of a bug happening in either step of the scenario. You've basically just added complication to the situation without accomplishing anything.

What you've done is created four scenarios: 1: the robot determines it would not bug in that hypothetical situation, and does not bug when it answers. 2: the robot determines it would not bug in the hypothetical situation, but it does bug when it answers. 3: the robot determines it would bug, and doesn't bug in the answer. 4: the robot determines it would bug, and does bug in the answer.

One solution might be to use the word "when" instead, but that doesn't work either. If you say, "When I say I like ice cream, you'll say true," the robot will know this statement to be false. Why? Because it knows that if you do say you like ice cream, it will already have responded to your above statement of "When I say I like ice cream you'll say true," and it will have run out of power. It would know that it would be unable to answer your statement of "I like ice cream," so it will say false.

Even using "If" above, you would always have to start with this: "In a hypothetical situation in which you had enough energy to qualify my statements indefinitely..." because of what I just explained.

Teraus, it doesn't work.

edited 5th Oct '11 5:17:52 AM by jagillette

@Madrugada

No. I'm not. I never said, "If I say X, answer true", I said "If I said X, you WOULD answer true".

That is not an order. It's a statement.

edited 5th Oct '11 8:55:30 AM by Teraus

Answering you now.

Perhaps I should rephrase this: "If I said X, you'd answer true". It literally means "In the exact state you are at the moment you will give the answer to this statement, If I instead said the statement "X", the answer given would be true". And, yes, the robot can predict that. I said he knows everything that can be known, and that includes everything that's not a paradox or impossible to answer by definition. He is practically omniscient. Otherwise, the challenge would be impossible.

I hope that helps. I do agree that additional details were necessary, though.

edited 5th Oct '11 9:02:59 AM by Teraus

For the people who still don't like any of my answers, try this:

Portal 1: "banana"

Portal 2: "TV Tropes"

Portal 3: "an object"

"In the exact state you will be at the moment you give the answer to this statement, If I instead said the statement "the object referred by the label of the awesome portal is a fruit", the answer given would be true, or the object referred by the label of the awesome portal is a computer."

edited 5th Oct '11 9:55:49 AM by Teraus

Right, but that still assumes that the bug is deterministic—that, in identical circumstances, the bug would always occur. (That's why it works with liars and truth-tellers: because each person is always a liar or always a truth-teller.) But we don't know that that's the case. Maybe the bug is so random that, under completely identical circumstances, it might not occur. Then, as jag said, the robot might truthfully say either answer.

It doesn't even have to be identical circumstances. You're asking the robot about the specific state it will be when it gives you your answer, and he knows it. That state also includes whether the bug is happening or not, so, it must be enough.

If the very first post tells you that the robot KNOWS when the bug happens, shouldn't that be enough? I feel like you just want whatever answer I give to be invalid, no matter what.

Also, not that this is relevant, but it's a personal belief of mine that everything in existence is deterministic. There is nothing completely random. This is most definitely valid for computers: bits are never random. Generating a random number is a function of time, not a "truly" random thing.

edited 5th Oct '11 10:45:39 AM by Teraus

I would like to take you up on that.

"The randomness comes from atmospheric noise, which for many purposes is better than the pseudo-random number algorithms typically used in computer programs."

Atmospheric noise is also deterministic, as far as I can see. It simply has too many variables for humans to predict.

That still doesn't change what I've said before about my answer.

edited 5th Oct '11 10:55:09 AM by Teraus

There are quantum physicists that would like to talk to you, then.

Quantum physics is not a theory that explains everything. It works well in a certain scope, much like statistics, but it's not perfect. You still can't deny the possibility that there may be factors which are not possible to be perceived by any known means. That doesn't mean that things aren't deterministic.

And that is still irrelevant to my final arguments about my answer.

edited 5th Oct '11 11:00:44 AM by Teraus