Thursday, November 22, 2018

logical deduction - Lies, damned lies, and statistics



A (hopefully) new variant on the Knights & Knaves problem.


On the island of D'Israel, the inhabitants fall into three types: Liars, Damned Liars, and Statisticians. They are visually indistinguishable except that each type of person wears a different colour of hat. If you ask them questions, their responses are as follows:



  • a Liar will always give you a false response

  • a Damned Liar will give you a false response and also stab you

  • a Statistician will give you a true or false response with 50% probability.


You are trying to escape from a prison on this island, and you are nearly out when you reach a row of three doors, each one guarded by a solid-looking sentry. The three sentries wear red, green, and blue hats, so you know that one is a Liar, one a Damned Liar, and one a Statistician, but you don't know which is which. You also know (from the prison map which was smuggled in to you and which you have now lost) that one of the doors leads to freedom, one to the prison warden's office, and one to a first-aid room; but again you don't know which is which. Your aim is to pass through the door that leads to freedom, but if you are stabbed by the Damned Liar, you will first need to visit the first-aid room. If you go to the warden's office, you will be recaptured; but going through either of the other doors allows you the chance to double back and try again if necessary.


You are being pursued, so you only have time to ask the guards one yes/no question to find out which door you should go through. What question should you ask? What is the probability you will escape?


How does the answer change if you have time for two or three questions?



This is a logic puzzle. Please, no lateral-thinking answers.



Answer




Using so called head exploding questions we can extract even more that one bit of information from the guards. In this example, we will produce a paradox if the ward is behind the second door. I.e., we ask a liar “Will you answer yes?”. If she answers “Yes” than she is telling the truth, that isn't allowed and if she answers “No” then she is telling the truth as well.


Exploiting self-referential statements can also persuade the statistician to give a useful answer. (We won't know whether we got lied to or not, but we don't care about that.) This is possible because of your wording “a Statistician will give you a true or false response with 50% probability.” instead of “a Statistician will give you a random response.”


So here is the question:



Is your statement a lie and not



door a = ward or



(door b = ward) and (is your answer no)



or is your statement true and



door a = ward or


(door b = ward) and (is your answer no)




Or with more math notation, we ask to evaluate $$ ((\lnot t) \land \lnot(a \lor (b \land p))) \lor (t \land (a \lor (b \land p)))\text{,}$$ where $t$ indicates if a guard tells the truth, $a$ and $b$ ask whether the ward is behind the respective door, $r$ the answer which will be given and $p = (r == \text{No})$ produces the paradoxes to explode heads.


With a small table we can work out that if a guard answers “No” than the ward is behind the a door, if the guard answers “Yes” than the ward is behind the c door and if her head explodes the ward is behind b.



Here is a table of all possible states of the world: τ r a b | True answer | Response 0 0 0 0 | 1 | 0 0 0 0 1 | 0 | 1 -> head explodes 0 0 1 0 | 0 | 1 0 1 0 0 | 1 | 0 0 1 0 1 | 1 | 0 -> head explodes 0 1 1 0 | 0 | 1 ……………………………………………………………………………………………………………… 1 0 0 0 | 0 | 0 1 0 0 1 | 1 | 1 -> head explodes 1 0 1 0 | 1 | 1 1 1 0 0 | 0 | 0 1 1 0 1 | 0 | 0 -> head explodes 1 1 1 0 | 1 | 1


No comments:

Post a Comment

classical mechanics - Moment of a force about a given axis (Torque) - Scalar or vectorial?

I am studying Statics and saw that: The moment of a force about a given axis (or Torque) is defined by the equation: $M_X = (\vec r \times \...