a logic puzzle

extension

way

Aah wait. If you asked the knave to identify all 3, she would lie. Then when you ask her "if I was to ask you to identify all 3 of you, what would you say?", she can answer anything except that original lie, so she can still lie. Daryl's trick only works on yes no questions.

Another solution. Question 1 (to person 1): If I was to ask this person (person 2) (points to another inhabitant) who is the knave, who would he identify?

Answer 1: yes or no: person 1 is the randomiser or the knave identifies himself: person 1 is the knight and person 2 is the knave, person 3 is the randomiser identifies person 2: person 1 is the knave and person 2 is the knight, person 3 is the randomiser identifies person 3: person 1 is the knight and person 2 is the knave and person 3 is the randomiser or person 1 is the knave

Doesn't look solvable from there..

Herc

Reply to
|-|erc
Loading thread data ...

"If i were a knave, would i tell a lie?"

Reply to
Robert Baer

Rec.puzzling at its finest!

Phil

--
Marijuana is indeed a dangerous drug.  
It causes governments to wage war against their own people.
-- Dave Seaman (sci.math, 19 Mar 2009)
Reply to
Phil Carmody

I think they have to be yes/no questions.

-- Daryl McCullough Ithaca, NY

Reply to
Daryl McCullough

Then 2 questions give 4 possible outcomes. There's 6 arrangements of the 3 inhabitants, making the deduction impossible.

Herc

Reply to
|-|erc

Yes, that's what I said: 2 questions are not enough.

-- Daryl McCullough Ithaca, NY

Reply to
Daryl McCullough

from the OP "who will answer any yes/no question" I see.

If you were to ask me, "if I were to ask you 'do you think the problem is impossible?' what would you say", I would say "yes".

Herc

Reply to
|-|erc

OK, if we allow the "I don't know/cannot answer" response, and assume that they all know who they all are, then I think the following works. Call them A,B,C.

Ask A: "If I asked you and B the same question, would you both give the same answer?"

If A is unable to answer, then B must be the randomizer, and we can sort out A and C out just by asking C a question to which we know the answer, like "Is B the randomizer?".

Otherwise B is definitely not the randomizer. Also, if A says yes, then A is randomizer or knave, and if A says no then A is randomizer or knight.

In either case, ask B: "If I asked you and C the same question, would you both give the same answer?".

No answer means C is randomizer, and A's answer completes the picture.

A yes/no answer implies that C is not the randomizer, so A must be the randomizer, and B's answer tells us the rest.

Derek Holt.

Reply to
Derek Holt

Oops, mis-read the problem. Damn!

Reply to
PeterD

Mark-T wrote: > ... randomizers - > they answer any question yes or no, 50-50. ...

You get NO information from a randomizer. Worse, the presence of the randomizer pollutes the information that you could get from either a knight or knave. I.e., any question that would give you information if answered by a knight or knave is worthless, given that a randomizer might be the one answering it.

I think it's unsolvable.

Bob

Reply to
Bob Engelhardt

Yeah so they do know each other?

Reply to
a7yvm109gf5d1

As I said, I assumed that they do! If not, I have no idea!

Derek Holt.

Reply to
Derek Holt

Note that this doesn't explicitly say "you can only ask yes/no questions", it just says "they will answer yes/no questions this way".

That said, if the randomizer avoids paradoxes, then I think one of Smullyan's books suggested a question along the lines of:

"If I asked you 'Are you the knight?' twice, would you answer 'yes' exactly once?"

The knight will say no, the knave will say yes, and the randomizer will freeze up.

Reply to
Ed Murphy

BZZZZT!!

Disallowed in the problem spec.

Hope This Helps! Rich

Reply to
Rich Grise

Also vaguely Dilbertian. ;-)

Cheers! Rich

Reply to
Rich Grise

Doesn't that kind conflict with the idea of "random"? Just because I toss a head once doesn't mean the next toss is constrained to be a tail, to maintain the 50% average!

Cheers! Rich

Reply to
Rich Grise

Norman ... Coordinate ...

Cheers! Rich

Reply to
Rich Grise

I cannot see anything in the problem specs as given that disallows that. We have to try and answer the problem posed and not some other similar problem. The Smullyan-type solution in which the randomizer freezes up definitely doesn't work, because we are told that the randomizer answers every question randomly yes/no.

Maybe the OP will eventually tell us what solution he had in mind!

Derek Holt.

Reply to
Derek Holt

I understand what you are saying, but the problem specifications don't allow true randomness IMHO. If they do, then there is virtually no way to ferret out the randomizer should he either answer with the correct answer twice, or the incorrect answer twice. In both those cases, he'll appear as if he were one of the others.

OK, as an appology for not reading the original problme properly, I'll post one of my own:

In San Francisco CA, two women were sharing the same umbrella. They walked over a block, outdoors, carefully sidesteping a puddle in teh sidewalk, no less, all the while working hard to control the umbrella in the stiff breeze.

Yet, when they arrived at their destination, they were both quite dry! How come?

Reply to
PeterD

Disagree: the randomizer will answer either yes or no, based only on a random chance of 50/50. After all the problem states: "the randomizers

- they answer any question yes or no, 50-50" (and that is 'any question'...)

Reply to
PeterD

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.