Riddle

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
By definition you are asking the victim to use answer based on disassociation. By definition disassociation would require a lack of a relationship between two numbers. Therefore, because two numbers will always share a relationship, exclusion by dissassociaation is not possible,


The problem with the curves that it has a requisite of standardized numerical shapes. If you use digitized version of the numbers, you have to modify the statement
 

CycloWizard

Lifer
Sep 10, 2001
12,348
1
81
No solution. It's logically impossible to discern between three variables with a single binary statement. Any logical test that excludes one of the choices must always allow the other two.
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
Originally posted by: CycloWizard
No solution. It's logically impossible to discern between three variables with a single binary statement. Any logical test that excludes one of the choices must always allow the other two.

I wish I was as concise as you:beer:
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
BTW...punch your TA in the face for me....


*realizes she might be 1) a woman and/or 2) handicapped and/or 3) armed*


ah nuts...
 

Pugnax

Senior member
Jan 17, 2000
517
0
0
Originally posted by: Goosemaster
Originally posted by: CycloWizard
No solution. It's logically impossible to discern between three variables with a single binary statement. Any logical test that excludes one of the choices must always allow the other two.

I wish I was as concise as you:beer:

Yeah I think so as well...This is similar to the non-linearly separable nature of the XOR problem. I'd like to think there is an answer though!
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
Originally posted by: Pugnax
Originally posted by: Goosemaster
Originally posted by: CycloWizard
No solution. It's logically impossible to discern between three variables with a single binary statement. Any logical test that excludes one of the choices must always allow the other two.

I wish I was as concise as you:beer:

Yeah I think so as well...This is similar to the non-linearly separable nature of the XOR problem. I'd like to think there is an answer though!

I also tried to think of it in terms of electrons in orbitals energy levels but got no where
 

DrPizza

Administrator Elite Member Goat Whisperer
Mar 5, 2001
49,601
167
111
www.slatebrookfarm.com
Originally posted by: yankeesfan
Originally posted by: silverpig
Originally posted by: DrPizza
I think that coming up with a math equation for which the answer is very simple to find for two of the values, and difficult to find for the 3rd value might be the way to go.

i.e. if he answers yes, the answer is 1.
if he answers no, the answer is 2
if he cannot answer the question because the problem is unsolvable in a realistic amount of time, then the answer is 3

see: http://en.wikipedia.org/wiki/Complexity_classes_P_and_NP

Give him some instructions like:

Assign the value of your number to the variable x. Compute (x-1)^(51200329123/443). Does the number which results from this computation have a 0 in the "ones" (first to the left of the decimal) place?

If x = 1, then the number is 0, and he'll answer yes.
If x = 2, then the number is 1, and he'll answer no.
If x = 3, then he'll have to do a lot of crunching and you'll know it's 3.

I think that this is the best way to go. Good job Dr. Pizza and Silver Pig.

That answer works for me :) My solution started with (x-2) , with this question:
is the product of (x-1) and the sum of numbers of any subset constructed with at least two elements from the larger set and of which your number must be one of the elements of the subset with the larger set being: {your number, a lot of really large positive or negative even numbers)} ever equal to zero?

If his number is 1, then it's trivially simple: the product is equal to zero.
If his number is 3, then it's relatively simple to deduce that the product cannot be equal to zero. Because (x-1)=2, and the second factor in the product *must* be an odd number, because it is the sum of one odd number and any number of even numbers.

If his number is 2, then determining if there is a subset of numbers such that the sum of that subset = 0 can be a problem for which no solution can be guaranteed in any reasonable amount of time.


I realized at this point, that I might have to modify the set part a little bit. Here's a similar problem (hopefully I don't screw up the numbers; I have this on a worksheet in school to use as a day killer for my advanced math students on those days when more than half the class is absent):
Given a set of 100 15 digit numbers, are there any two subsets such that the sum of the numbers in each of those subsets is the same?
It's very simple to prove that the answer is "yes" by the pigeonhole concept... Every sum must be between 15* 100000000000000 and 15*999999999999999. That's a lot of possible answers when you sum up subsets. (roughly 1.5*10^16) However, there are more subsets than there are possible sums. (100C1 + 100C2 + 100C3 +...) A whole heck of a lot more possible sums. So, imagine a giant post office with a box for every possible sum. Every time you get the sum of a subset, write it on a piece of paper and put it in the appropriate box. Not every box will necessarily have a sum in it when you're finished, but since there are more pieces of paper than boxes, it's necessary that at least 1 of the boxes has more than one slip of paper in it. Compute the number of subsets there are with 50 elements in it out of 100 elements from your original set... then figure that you can find 10^9 (a billion) sums every second. Calculate the number of years it's going to take, simply to calculate the sums of 50 number subsets.

You may need to modify my original subset restriction slightly in order for it to work out to this level of difficulty.


edit: note, the original question said "what is one yes or no question that you could ask to discern which number it is I'm thinking of" - since my solution provides a problem for which there is a yes or no answer possible for 2 of the number selections, and a complete lack of silence for the next few billion years (or an answer of "uhhhhh, I'll have to get back to you on that), then you *can* discern the original number.
 

TheChort

Diamond Member
May 20, 2003
4,203
0
76
The only thing I can think of is if the answers are Yes, No, and not possible to answer

example:
If your number was subtracted by two, would it equal 1 if the number 1 was divided by it.
I.E. Does 1/x-2 = 1 (where x is the number)?

edit: posted this before I saw DrPizza's post. I guess our solutions are essentially the same. Mine is merely humbled in his presence :p
 

HaxorNubcake

Golden Member
Jun 23, 2004
1,983
0
0
Originally posted by: TheChort
The only thing I can think of is if the answers are Yes, No, and not possible to answer

example:
If your number was subtracted by two, would it equal 1 if the number 1 was divided by it.
I.E. Does 1/x-2 = 1 (where x is the number)?

edit: posted this before I saw DrPizza's post. I guess our solutions are essentially the same. Mine is merely humbled in his presence :p

Not quite the same...
assuming that you mean 1/(x-2)

if the number is 1, function = -1 (NO)
if the number is 2, function = undefined/can't divide by 0 (NO)
if the number is 3, function = 1 (YES)

No still leaves the first 2 options
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
You turn on one light and leave it on for five minutes. Turn that light off, and then turn another one on. Run upstairs--the light that is on is the switch that is on. The light that is off but warm, belongs to the previous switch.
 

cubby1223

Lifer
May 24, 2004
13,518
42
86
Originally posted by: DrPizza
That answer works for me :) My solution started with (x-2) , with this question:
is the product of (x-1) and the sum of numbers of any subset constructed with at least two elements from the larger set and of which your number must be one of the elements of the subset with the larger set being: {your number, a lot of really large positive or negative even numbers)} ever equal to zero?

If his number is 1, then it's trivially simple: the product is equal to zero.
If his number is 3, then it's relatively simple to deduce that the product cannot be equal to zero. Because (x-1)=2, and the second factor in the product *must* be an odd number, because it is the sum of one odd number and any number of even numbers.

If his number is 2, then determining if there is a subset of numbers such that the sum of that subset = 0 can be a problem for which no solution can be guaranteed in any reasonable amount of time.


I realized at this point, that I might have to modify the set part a little bit. Here's a similar problem (hopefully I don't screw up the numbers; I have this on a worksheet in school to use as a day killer for my advanced math students on those days when more than half the class is absent):
Given a set of 100 15 digit numbers, are there any two subsets such that the sum of the numbers in each of those subsets is the same?
It's very simple to prove that the answer is "yes" by the pigeonhole concept... Every sum must be between 15* 100000000000000 and 15*999999999999999. That's a lot of possible answers when you sum up subsets. (roughly 1.5*10^16) However, there are more subsets than there are possible sums. (100C1 + 100C2 + 100C3 +...) A whole heck of a lot more possible sums. So, imagine a giant post office with a box for every possible sum. Every time you get the sum of a subset, write it on a piece of paper and put it in the appropriate box. Not every box will necessarily have a sum in it when you're finished, but since there are more pieces of paper than boxes, it's necessary that at least 1 of the boxes has more than one slip of paper in it. Compute the number of subsets there are with 50 elements in it out of 100 elements from your original set... then figure that you can find 10^9 (a billion) sums every second. Calculate the number of years it's going to take, simply to calculate the sums of 50 number subsets.

You may need to modify my original subset restriction slightly in order for it to work out to this level of difficulty.


edit: note, the original question said "what is one yes or no question that you could ask to discern which number it is I'm thinking of" - since my solution provides a problem for which there is a yes or no answer possible for 2 of the number selections, and a complete lack of silence for the next few billion years (or an answer of "uhhhhh, I'll have to get back to you on that), then you *can* discern the original number.
I have a minor in math including being near or at the top of the classes of statistics & probabilities, graph theory, and combinatorics, and I'd still go uhhhhh.... to even the "simple" solutions of the questions you pose. I do see exactly what you're getting at, but to ask someone that, it'll never be understood.

I'd propose a much simpler answer to the riddle, just ask "Are you going to tell me the number now, or do I have to beat it out of you with this baseball bat?" Quick and to the point, that's my motto.
 

cubby1223

Lifer
May 24, 2004
13,518
42
86
You could just ask the most simplistic of questions, not this crazy complex math you nerds:

"If the number is not 1, then is it 2?"

There's your:
Yes - 2
No - 3
Can't answer - 1
 

DrPizza

Administrator Elite Member Goat Whisperer
Mar 5, 2001
49,601
167
111
www.slatebrookfarm.com
Well, sorry that my solution was hard to understand... I reworded it again and again, and each time I worded it much more simply, I discovered a minor error that screwed up the logic. I put a pretty good amount of time into it, because I actually love this question and my angle of attack... I'd be interested in knowing if there was a really simple math answer (but as cyclowizard's reasoning: nope.) - so the only way I could think of a solution would be if there was that 3rd option: the yes/no question could be answered for 2 of the numbers, but the question couldn't be answered for the 3rd value.

I think I saw at a glance that someone tried the division by 0 thing to bring up this case; that was my first instinct too, but after a few seconds, I realized that with "yes/no", you still didn't have that 3rd case.
==============
Maybe this is better:

from the set:

{(thousands of positive and negative even numbers with about 20 digits in each one)}

Is the product of (x-1)(your number + a the sum of the numbers in a subset of the above set) ever equal to zero?

3 cases:
*the number is 1: trivial to show the product is 0
Answer is "yes"
*the number is 3: (3-1)(3 + even number)
=2(odd number) never equals zero
Answer is "no"
*the number is 2: (2-1)(2 + some even number)
Answer is {crickets} or "I'll have to get back to you on that one..."


Problem can be easily checked if someone had a solution; but (unless I'm mistaken), this is an NP problem - it would take an ungodly amount of time to determine that there were no sums equal to zero. It's possible that someone could stumble across a sum of zero relatively quickly; when creating the question, select even numbers carefully enough that any logical algorithm used on a super-computer wouldn't stumble across such a solution any time soon (next million years or so).


 

cubby1223

Lifer
May 24, 2004
13,518
42
86
If you want to take the approach like DrPizza of simple answers versus answers that take a long time to figure out, just ask:

Is the number either the number 1 or equal to (2Ah - 10110b) / 12o?
 

kedlav

Senior member
Aug 2, 2006
632
0
0
Originally posted by: cubby1223
You could just ask the most simplistic of questions, not this crazy complex math you nerds:

"If the number is not 1, then is it 2?"

There's your:
Yes - 2
No - 3
Can't answer - 1

False implying false is a true statement, meaning 1 is a yes.

Assuming you actually ask the ah heck the question, use the complicated function version.
 

CycloWizard

Lifer
Sep 10, 2001
12,348
1
81
Originally posted by: DrPizza
Well, sorry that my solution was hard to understand... I reworded it again and again, and each time I worded it much more simply, I discovered a minor error that screwed up the logic. I put a pretty good amount of time into it, because I actually love this question and my angle of attack... I'd be interested in knowing if there was a really simple math answer (but as cyclowizard's reasoning: nope.) - so the only way I could think of a solution would be if there was that 3rd option: the yes/no question could be answered for 2 of the numbers, but the question couldn't be answered for the 3rd value.

I think I saw at a glance that someone tried the division by 0 thing to bring up this case; that was my first instinct too, but after a few seconds, I realized that with "yes/no", you still didn't have that 3rd case.
==============
Maybe this is better:

from the set:

{(thousands of positive and negative even numbers with about 20 digits in each one)}

Is the product of (x-1)(your number + a the sum of the numbers in a subset of the above set) ever equal to zero?

3 cases:
*the number is 1: trivial to show the product is 0
Answer is "yes"
*the number is 3: (3-1)(3 + even number)
=2(odd number) never equals zero
Answer is "no"
*the number is 2: (2-1)(2 + some even number)
Answer is {crickets} or "I'll have to get back to you on that one..."


Problem can be easily checked if someone had a solution; but (unless I'm mistaken), this is an NP problem - it would take an ungodly amount of time to determine that there were no sums equal to zero. It's possible that someone could stumble across a sum of zero relatively quickly; when creating the question, select even numbers carefully enough that any logical algorithm used on a super-computer wouldn't stumble across such a solution any time soon (next million years or so).
I agree that this is the practical solution to this problem. I still don't believe that this problem has a solution in a strictly mathematical sense, but that's not necessarily the point of the riddle, either. :p
 

cubby1223

Lifer
May 24, 2004
13,518
42
86
Originally posted by: kedlav
Originally posted by: cubby1223
You could just ask the most simplistic of questions, not this crazy complex math you nerds:

"If the number is not 1, then is it 2?"

There's your:
Yes - 2
No - 3
Can't answer - 1
False implying false is a true statement, meaning 1 is a yes.

Assuming you actually ask the ah heck the question, use the complicated function version.
And true implying true is well, rather silly. And don't even get me started on the whole false implying true bit. :laugh:

Seriously though, this isn't a logic statement determining true or false. It's a question asking yes or no. There's a difference.
 

zeruty

Platinum Member
Jan 17, 2000
2,276
2
81
Originally posted by: yankeesfan
Originally posted by: RedArmy
Read the OP again, my friend just informed me it was one "yes or no" question...F'ing a.

"Could you tell me your number, but use Yes for 1, No for 2, and keep silent for 3, please?"

This assumes that they must answer.

This is the answer I came up with. However, I would have worded it a little differently.
 

Goosemaster

Lifer
Apr 10, 2001
48,775
3
81
They here is that many of you are introducing a third response in the form of an audible "wtf?!" or silence.

If the question is a yes or no question, one can just as easily assume that the respondant must be able to answer the question or is free not to answer.

Silence might, for all we know, be a display of indifferenence and and not of confusion or not being able to answer withing the boolean cosntraints.


If the flashlight is controlled by a three position, switch, is it in position 0 of 2 ( 0,1,2) because I put it there, or because I don't know how to acutalize its movment.


*head explodes *
 

thesurge

Golden Member
Dec 11, 2004
1,745
0
0
Is the reciprocal of one less than the number equal to 1?

Yes- 2
No- 3
WTF/Can't answer (still a yes or no question)- 1