3 prisoner puzzle

I just re-read this:
However if you thought the claim to be tested included the entire statement, the correct answer is none of the three you gave :).
While, you're right about this I think it misses a point.

If you are trying to establish the veracity of the claim but don't know if the assumptions can be trusted you don't need to turn over the 6 because a violation of the assumptions won't lead to a violation of the claim. But you need to turn over the F to prove there isn't a vowel on the other side.

Agree?
 
Same answer, probably same reasoning:

WLOG, first we overhear tom; guard says "dick" lives:
2/3 chance tom lives,
1/3 chance harry lives.

1/2 chance to overhear dick next; guard says "tom" or "harry" lives:
always know.

1/2 chance to overhear harry:
2/3 chance tom lives;
assume mental 'coin flip' : 1/2 chance guard will say "dick", (don't know), 1/2 chance guard will say "tom", (know).
1/3 chance harry lives;
guard must say "dick" (which tells us dick -- don't know).

So, 1/2 + 1/2 x 2/3 x 1/2 = 2/3.
Nice twist, Lucky. :)
Yes, that does the job. (Couldn't fathom WLOG; had to look it up!)


Without meaning to blow my own trumpet, there was never any doubt in my mind that Tom's probability of death was still 1/3 in the original problem and I couldn't really see why anyone would think otherwise. Then it was pointed out that this problem is identical to the Monty Hall problem. The latter took me a good hour to get my head around. So I was wondering why such a difference might occur? Does it say something about my ability (or lack of) at problem solving? I first learned about the Monty Hall problem about 5 years ago but I'm struggling to believe its just experience of statistics that caused the difference.
I think that's a profound and interesting question – and I don't know the answer.

In real life, we generally tackle (what ought to be) probability and other logic problems by intuition (i.e. unconscious analogy) rather than by conscious logical manipulation and calculation. But to solve a recreational puzzle we need to abstract the logic of the situation and ignore the trappings, and some (kinds of) scenarios seem to be more difficult than others for this purpose. Whilst this isn't universal (some people have difficulty with any kind of logic problem, others find them easy), there does seem to be some quality of a scenario that invites intuition and analogy rather than abstraction - and the intuition/analogy may be false.


I think the most interesting thing about these two puzzles is why people find them so non-intuitive. Personally I wasn't confused by either, but I work with probabilities (and numbers in general) every day as a profession, so I have lots of practice thinking about these things (and am quite wary of the mental traps you can fall into).
Same here. It becomes automatic to translate these problems into context-free mathematical representations, while blocking out intuitive answers. A (sometimes annoying) corollary is that you tend to do this when intuition would actually work perfectly well; so (for example in a pub conversation) you become the last (instead of the only) person to get the right answer.


It reminds me of the following famous cog psy experiment:

The experimenter says: "Here are four cards, each with a number on one side and a letter on the other" (you look down and see four cards on a table, labeled A, F, 3, 6 respectively). "The claim is, if there's a vowel on one side of a card then there's an even number on the other. Question - which of those cards must you flip over to check whether that is true?"

This is an utterly trivial problem in formal logic (in a sense it's the most simple question you can possibly ask), and yet only a tiny fraction of people answer correctly when you do the experiment.
I read about a study where this puzzle was posed in several formally identical ways, with wildly different results (unfortunately I can't remember where I read it). As you said, most people got the card scenario wrong, but with other scenarios the success rate was far higher.

One of them went something like this:
You're bar/security staff at a social event. People have to be at least 18 to be in possession of alcohol. What do you need to check: someone ordering a whisky, someone ordering an orange juice, also a young teenager and a pensioner both carrying drinks?

Plainly, it's formally the same as the card puzzle (if alcohol then >=18). Yet the large majority got this one right. Why?
 
If you are trying to establish the veracity of the claim but don't know if the assumptions can be trusted you don't need to turn over the 6 because a violation of the assumptions won't lead to a violation of the claim. But you need to turn over the F to prove there isn't a vowel on the other side.

Agree?

I'm not sure I understand what you're asking. If it's what to do to verify that the statement "each card has a number on one side and a letter on the other" is correct, then no, I don't agree - you'd have to flip them all over to be sure. If it's the other statement, then again, no - that says nothing definite about cards with even numbers on them.

One of them went something like this:
You're bar/security staff at a social event. People have to be at least 18 to be in possession of alcohol. What do you need to check: someone ordering a whisky, someone ordering an orange juice, also a young teenager and a pensioner both carrying drinks?

Plainly, it's formally the same as the card puzzle (if alcohol then >=18). Yet the large majority got this one right. Why?

Logic is formal - i.e. it depends only on form, and two problems are logically totally equivalent if they take the same form. Human thought is not like that at all, and it only occasionally coincides with logic. I think the problem stated that way is much easier because it accords with something we're likely to have to deal with in the real world - and things like that we've learned to be better at.
 
Last edited:
I'm not sure I understand what you're asking. If it's what to do to verify that the statement "each card has a number on one side and a letter on the other" is correct, then no, I don't agree - you'd have to flip them all over to be sure. If it's the other statement, then again, no - that says nothing definite about cards with even numbers on them.
No, I'm not asking about validating one or the other statement in the problem. I'm asking about validating the claim in the absence of assurance you can trust the statement about the cards being designed as claimed.

In that case you also need to turn over the F to establish it doesn't have a vowel on it's other side. But you don't need to turn over the six because there can't be a violation of the claim no matter what's on it's other side even if the problem presenter lied about the nature of the cards. No matter whether you find an even or odd number, vowel or consonant on the other side of the six, none of those situations violates the claim. Some violate the assumption, though.
 
No, I'm not asking about validating one or the other statement in the problem. I'm asking about validating the claim in the absence of assurance you can trust the statement about the cards being designed as claimed.

OK.

In that case you also need to turn over the F to establish it doesn't have a vowel on it's other side. But you don't need to turn over the six because there can't be a violation of the claim no matter what's on it's other side even if the problem presenter lied about the nature of the cards.


Sure you do. The statement was:
experimenter said:
Here are four cards, each with a number on one side and a letter on the other.
If the 6 doesn't have a letter on the other side it falsifies that statement.

The only thing I can see that might have been ambiguous is whether it's the statement above which needed to be tested, or the more specific claim concerning vowels and even numbers. A slight re-wording of the question would eliminate that, and it wasn't that confusion which caused people to get the answer to this wrong in the experiment (and evidently not in your case either).
 
Last edited:
Sure you do. The statement was:
"Here are four cards, each with a number on one side and a letter on the other.
"
If the 6 doesn't have a letter on the other side it falsifies that statement.
I guess I'm not making myself clear.

The problem presented is below with me identifying what I'm calling the assumption and the claim:

The experimenter says: "Here are four cards, each with a number on one side and a letter on the other" (you look down and see four cards on a table, labeled A, F, 3, 6 respectively).{the preceding is the assumption} "The claim is, if there's a vowel on one side of a card then there's an even number on the other. Question - which of those cards must you flip over to check whether that is true?"{that's the claim}
If you trust the presenter and are only addressing the claim, you answer A and 3.

If you don't trust the presenter (or he doesn't mention the assumption at all) then you have to also turn over the F because it could falsify the claim if it violates the assumption. Turning over the six can't refute the claim, it can only refute the assumption.

If you're testing both the assumption and the claim you then need to turn over all four cards.

So when they do this test, do they ask the participants if they trusted the presenter?
 
If you don't trust the presenter (or he doesn't mention the assumption at all) then you have to also turn over the F because it could falsify the claim if it violates the assumption. Turning over the six can't refute the claim, it can only refute the assumption.

OK, I see what you're driving at. There's an issue as to whether or not the claim can be regarded as an independent statement which might be correct even if the "assumption" is wrong. But again, the fact that we're having this conversation at all exemplifies precisely the point of this experiment, which is that people do not think according to formal logic.

To avoid this kind of confusion one simply needs to word the question a little more carefully. I don't recall precisely what their wording was, but in any case variations of the experiment have been repeated many times.
 
The new warden devised a test to determine if the 3 prisoners were psychics. The previous evening she had gathered the prisoners to tell then of the test. The warden said to them: "When we leave here, you will each be taken to separate cells where you cannot see or here the others. I will return to my office and draw the name of one of you to be executed in the morning. A guard will be sent to each of you to tell you the name of one of your comrades that will be spared." The three prisoners were immediately separated, the name was drawn and the guards were dispatched as promised. In the morning, the three prisoners were escorted to the big room and arranged in a triangle around the gallows where warden stood. The warden said: "We are here for a hanging, If any of you speak you will be executed even if your name wasn't drawn. Two of you may leave and go free now but only if the person whose name I drew last night stays in the room. The guards will now remove your blindfolds so that you may take one last look at your comrades. Remember, no speaking!"

Tom looked at Harry and saw his hands were also bound behind his back which ruled out any hope of using semaphore even if they could remember the codes. Discouraged, Tom sat down and closed his eyes.

What action should Dick and Harry take? (remember, Tom is not looking)
 
The new warden devised a test to determine if the 3 prisoners were psychics. The previous evening she had gathered the prisoners to tell then of the test. The warden said to them: "When we leave here, you will each be taken to separate cells where you cannot see or here the others. I will return to my office and draw the name of one of you to be executed in the morning. A guard will be sent to each of you to tell you the name of one of your comrades that will be spared." The three prisoners were immediately separated, the name was drawn and the guards were dispatched as promised. In the morning, the three prisoners were escorted to the big room and arranged in a triangle around the gallows where warden stood. The warden said: "We are here for a hanging, If any of you speak you will be executed even if your name wasn't drawn. Two of you may leave and go free now but only if the person whose name I drew last night stays in the room. The guards will now remove your blindfolds so that you may take one last look at your comrades. Remember, no speaking!"

Tom looked at Harry and saw his hands were also bound behind his back which ruled out any hope of using semaphore even if they could remember the codes. Discouraged, Tom sat down and closed his eyes.

What action should Dick and Harry take? (remember, Tom is not looking)


What happens if the person who was picked for hanging tries to leave? What if only one person tries to leave? What if all three try to leave? What if nobody tries to leave? I think you need to clearly specify what happens in these situations in order to determine what the best course of action for Dick and Harry is.
 
What happens if the person who was picked for hanging tries to leave? What if only one person tries to leave? What if all three try to leave? What if nobody tries to leave? I think you need to clearly specify what happens in these situations in order to determine what the best course of action for Dick and Harry is.

Those questions were all addressed in the statement "Two of you may leave and go free now but only if the person whose name I drew last night stays in the room." The condition for a prisoner to be freed is that they leave the room and that the chosen prisoner stays in the room. What happens if they stay is not specified, they will probably be subject to another experiment at a later time. There was no condition that the two must leave together. And obviously if the chosen one leaves the room the other two will not be freed and that one would simply be hauled back into the room and hanged.

Being the subjects of many logic problems, Tom Dick and Harry have all become expert logicians. If anybody can figure this problem out, they can.

The warden however is very much a believer in woo. She believes that if the prisoners are psychic, two of them will know that they can walk out of the room and go free and the third will stay behind to be executed. But if they are simply using logic, one of them will have to speak up and sacrifice them self to insure that the other can go free.

The warden is thinking at this time that maybe prison life has been too easy and the prisoners would rather stay than go free.
 
Yeah, on rereading the poster actually said that at the start. But I think it is exactly the same, as I suggested.

I dont exactly agree. Im not sure how to explain it but it differers slightly and with a different outcome to me then the Monty Hall problem.

The difference being that in the Monty Hall scenario their is an individual who has given 3 choices and must choose one. He is then shown the next choice and has the option to change based on statistical probabilities.

In this new puzzle with the 3 men and the guard the man in question is one of the options/probabilities. It is not an outsider looking to 3 possible choices but one of the choices asking about the other 2.

In this example he knows 100% that before he receives an answer to his question that one of the other 2 cell mates are going to live, because of the fact that it was originally stated that 2 would live. Any way the situation is put one of those men will live except for if he is the man to die where both of the others will live. Either way no new information will be gained by his question. No variables have changed nor did it have any statistical fluctuations.

Like i said he already knew one of them was going to live all he did was find out the name of one. Nothing changed and he still has the same percentage of being the one to hung as he did when he started.

From the start there chances of being hung were always 50% and not 33% like most will claim.
The reason being is that you can automatically eliminate one of the other prisoners from the equation once the guard informed them that 2 would go home and one would die, thus making it now down to a 50-50 chance.

To an onlooker the equation would be 33% if you had to randomly choose his pick of the 3 with no information given to him, but for those involved it was always 50%.

Sorry if my logic cant be expressed well through my typing but im almost positive i have got this right.....
 
Last edited:
This is incorrect because (You said this yourself by the way :D):



Giving him 50% chance to survive after he hears the guards answer. I agree.

Their was never a 1/3 chance of anything is this scenario. From the very start his odds were 50% with or without the guards answer to his question.

Once he was told that 2 would be set free and one would die he knew for a fact that one of his cellmates would go home and it would be between him and the remaining cellmate. Which infact makes it a 50% chance of him being hung!

So it was always a 50% chance for him when he was informed of the situation!
 
This is not the same problem. The prisoner doesn't get to choose his fate - it's already decided.

Its not because he doesnt get to choose his fate. That would be for an outsider looking at the 3 and trying to guess who will be hung.

The difference is that the man in question is one of the possible options. He is not looking at 3 options and trying to guess he is looking at 2.

It is basically from the start :

Tom vs Dick and Harry= 50% odds

When he is informed the name of one of the individuals who will survive it turns into:

Tom vs Harry= 50%
 
Last edited:
Dear god... after all this, people are STILL getting this wrong?

Im sure if your talking about me but from the start it was always 50% odds for Tom to look at and this never changed.

I am the only one who i have seen that came to this conclusion and it perfectly makes sense.
I dont see where his odds were ever 1/3 in this situation.
Actually his odds never were 1/3.
 
Last edited:
Im sure if your talking about me but from the start it was always 50% odds for Tom to look at and this never changed.

OK, maybe you're just confused about the question. Here it is again:

The warden RANDOMLY picks one of Tom, Dick, or Harry to be executed the next day. At that stage, there is a 1/3 chance that he has picked any given one.

That night, Tom asks the warden who will be set free, and the warden says "Dick". Question: knowing that Dick will be set free, what are Tom's chances of being executed? What are Harry's?

You may assume the warden tells the truth, will never answer "Tom" when Tom asks, and doesn't change his mind.
 
OK, maybe you're just confused about the question. Here it is again:

The warden RANDOMLY picks one of Tom, Dick, or Harry to be executed the next day. At that stage, there is a 1/3 chance that he has picked any given one.

That night, Tom asks the warden who will be set free, and the warden says "Dick". Question: knowing that Dick will be set free, what are Tom's chances of being executed? What are Harry's?

You may assume the warden tells the truth, will never answer "Tom" when Tom asks, and doesn't change his mind.

My friend i certainly know the scenario and i believe you are the one confused, allow me to explain.

The question is not asking me to guess as to whom the warden has chosen to be executed. If it was it would be what you are explaining to me 1/3 probability. The question is infact asking what are Toms odds when he is informed of the situation. This is very different then what you are explaining in your post.

Like i stated earlier Tom does not need to ask the warden any questions question in order to figure his odds. Tom has had 1/2 odds the entire time. From the start it has been Tom vs. Harry and Dick. Tom knows for sure one of those 2 individuals will go home but doesnt know who. So he can already assume that it will be left with him/Tom vs one individual if the rules of the scenario are followed.

So inturn from the very start Tom had 50% odds of being chosen to be executed.

Now if i asked you the spectator to guess which of the 3 would be executed it would be a 1/3 probability for you. This is what you seem to believe is going on when infact it is not.
For you are an outsider looking in and have 3 options to choose from. If i then took one individual and let him go infront of you and said "you are now left with 2 prisoners do you want to switch?" It would be statistically logical for you to do so because this would be exactly the same as the Monty Hall scenario. Tom being involved in the scenario knows for sure that one of his cellmates are going home while you the spectator have no clue who.

The big difference is that Tom is involved in the situation and you are a spectator. They are not asking what is the spectators odds of choosing who will be executed (if that was the case then your assumptions would be correct) but more so they are asking what are Toms odds of being executed.

I hope i was able to clear this situation up for you and show you that you were the one slightly off on the scenario.
 
You are correct that from the beginning it's Tom v. Harry + Dick, but it's 1/3 Tom and 2/3 Harry + Dick. When Tom receives the new information, his odds of being hanged remain at 1/3, while Harry's odds go up to 2/3, just like the Monty Hall problem.

Just like Monty Hall, it can help to understand if you increase the number of prisoners (doors). If there were 1000 prisoners, and Tom asks the guard the question, and the guard gives the names of 998 prisoners that will go free, what are Tom's chances of being hanged?
 
A) Tom Harry Dick

(Before being informed of scenario)
No Odds as of yet


B) Tom VS. Dick Harry

After being informed of scenario
= 50%


C) Tom VS. ??? Man set free

Before being told answer to his question this is what he knows for sure
Outcome = 50%


D) Tom VS. Harry DICK

After being told answer
Final Outcome = 50%
 
Last edited:
I have a game I want to play with you. I roll a six sided die, if it comes up one I pay you a dollar, if it comes up something else you pay me a dollar. Fair? 1 vs something else.
 

ISF - Join now!

Every member here is approved by hand. No bots, no spam, just people who care about evidence and honest debate.

Membership is free!

Create your free account

Back
Top Bottom