Riddles and Puzzles: Extend the following to a valid equation

  • Context: Challenge 
  • Thread starter Thread starter fresh_42
  • Start date Start date
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
592 replies · 55K views
fresh_42 said:
81. Bart and Lisa are sitting in front of a huge heap of skittles. Since both of them want to eat as many as possible, they decide to play a game. Bart has to write two (different) numbers (positive integers) on two pieces of paper. Then Lisa turns around one of the papers and guesses, whether the other number is higher or lower. If she is right, then she will get 10 skittles, otherwise Bart will get them.

Is there a strategy for Lisa which improve her chances in comparison to a 50:50 guess? And can Bart counter this strategy?

D95
Suppose that Bart has chosen numbers i and j. One is chosen at random. 0.5 probability for i, 0.5 probability for j. Without loss of generality, assume i > j.

Lisa decides on a strictly monotone increasing function f from the positive integers to the reals between 0 and 1. After seeing the chosen number, she uses this function to determine her probability of staying with the number that has been revealed.

Lisa's probability of seeing i and then winning is ##0.5 \times f(i)##
Lisa's probability of seeing j and then winning is ##0.5 \times (1 - f(j))##
Total ##0.5 + 0.5\times (f(i) - f(j))##

Since f is monotone increasing and i > j, f(i) > f(j) and Lisa's winning probability exceeds 0.5

Bart can counter this strategy by choosing very large numbers. Lisa's function is bounded and monotone increasing. It must therefore have a least upper bound that is approached arbitrarily closely for large n. (i.e. ##\lim_{n \to \infty}f(n) = U##. It follow that for any epsilon, Bart can choose a delta large enough that that f(i)-f(j) < epsilon for all i, j > delta.

Edit: You'd only asked about the existence of a counter-strategy for Bart rather than for an actual implementation. The following mixed strategy should perform adequately.

Suppose that Bart wishes to obtain a winning probability within epsilon of 0.5. Bart arbitrarily picks a positive integer n that is larger than 1/epsilon. He then picks uniformly and at random an integer i between 1 and n. He writes i on one piece of paper and writes i+1 on the other.

Lisa's chance of winning will depend on her choice of f, of course. But averaged over all the n possible picks by Bart, her winning probability exceeds 0.5 by a total of:
$$\frac{1}{n}\sum_{i=1}^n 0.5 \times(f(i+1)-f(i))$$
Since f(1) and f(n+1) are bounded by 0 below and 1 above, this sum is no more than 0.5/n. This is less than epsilon.
 
Last edited:
on Phys.org
I think you got the basic method, but I don't follow your reasoning, since ##f(i)=0## for any ##i##, considering the amount of numbers available. I don't think it can easily be put into a "proof", since the probabilities involved are very hard to determine appropriately without additional assumptions.

So I was looking for an answer to: What must Lisa do to improve her chances?

Very large numbers is a practical counter strategy based on assumptions of human behaviour, but there is a counter strategy which always work.
 
fresh_42 said:
I think you got the basic method, but I don't follow your reasoning, since ##f(i)=0## for any ##i##, considering the amount of numbers available.
Huh? There is no guarantee that f(i) = 0 for any i. ##f(i)=0.75 - \frac{1}{2i}## works just fine and is bounded below by 0.25.
I don't think it can easily be put into a "proof", since the probabilities involved are very hard to determine appropriately without additional assumptions.
Again, what? Nothing hard about any of it.
Very large numbers is a practical counter strategy based on assumptions of human behaviour, but there is a counter strategy which always work.
Yup. Found one and had already edited it in.
 
Maybe I didn't understand your solution in detail, esp. the definition of ##f##.
Anyway, I said I think you have it. Here's my wording:

Lisa chooses a third random number ##s## secretly and adds ##0.5##. If the displayed number is less than ##s##, then she answers "greater" and vice versa. This does not affect her chances if ##s## is smaller than the smaller or greater than the bigger of the two numbers. But if ##s## is in the interval of the two, she will always win. So all these instances increase her chances.

The counter strategy for Bart would be to choose intervals of length one.
 
fresh_42 said:
Maybe I didn't understand your solution in detail, esp. the definition of ##f##.
Anyway, I said I think you have it. Here's my wording:

Lisa chooses a third random number ##s## secretly and adds ##0.5##. If the displayed number is less than ##s##, then she answers "greater" and vice versa. This does not affect her chances if ##s## is smaller than the smaller or greater than the bigger of the two numbers. But if ##s## is in the interval of the two, she will always win. So all these instances increase her chances.

The counter strategy for Bart would be to choose intervals of length one.
So, roughly speaking, my monotone increasing f is the inverse of the cumulative distribution of your random variable s. Sure, that works.

Intervals of length one is a good start. But it is inadequate, by itself, to assure a particular expected value against all instances of the Lisa strategy we have discussed.
 
82. Fill in all numbers ##\{\,1,2,\ldots,15\,\}## such that the sum of all numbers in every ring is the same.
246804


D96
 
83. A dealer shuffles a standard deck of ##52## cards. The player has ##\$\,100## at the beginning. He can bet any non negative amount on the color (red and black) of each card the dealer turns around. E.g. he could wait for ##51## cards to be drawn and put the entire ##\$\,100## on the missing color. In case he is right, he will receive twice the amount he has set, and nothing if he loses.

What is the optimal strategy?

[optimal: guarantees ##\$\,y## at the end of the game and no other strategy can guarantee ##\$\,z > y##. No statistical assumptions must be made and must work even if the dealer recognizes the strategy and cheats.]

D98
 
84. A red ant and seven black ants are placed on a five-meter-long, straight branch. Each ant chooses (randomly) to go to the left or to the right, and then runs in that direction at a speed of 1 meter per 10 seconds. Whenever two ants hit each other, both turn and then continue in the opposite direction. When an ant reaches the end of the branch, then it falls off the branch and out of the game.

How long can the red ant stay on the branch at most?

D99
 
fresh_42 said:
83. A dealer shuffles a standard deck of ##52## cards. The player has ##\$\,100## at the beginning. He can bet any non negative amount on the color (red and black) of each card the dealer turns around. E.g. he could wait for ##51## cards to be drawn and put the entire ##\$\,100## on the missing color. In case he is right, he will receive twice the amount he has set, and nothing if he loses.

What is the optimal strategy?
If we want to maximize the minimal money we walk away with then we must be indifferent about the outcome of every bet: If we would favor winning then we should bet less money to make losing less bad and vice versa.
Let f(r,b) be the factor we can multiply our money with if there are r red and b black cards left. Clearly f(0,0)=1 as we can't bet any more and f(r,0)=2r as we bet all on red for r rounds.
Let's assume r>=b, we bet on red. Our bet fraction p must be chosen such that winning and losing will lead to the same result: (1+p)f(r-1,b) = (1-p)f(r,b-1). Solve: p = (f(r,b-1)-f(r-1,b))/(f(r-1,b)+f(r,b-1)). While negative bets were not allowed bets with p between -1 and 0 are effectively a -p bet on red, so we can use this formula everywhere.
Fill out the table: f(26,26)=9.08132954942779. Multiply by the initial $100 and we get $908.13.

I also looked at a different question: What if you bet everything when there is an imbalance in the remaining cards? You get $100*226 with probability 26/51 * 25/49 * ... = 26!/51! and zero otherwise.. A one-in-7 million chance to win 6.7 billion. The expectation value? 9.08132954942779.

This is not an accident, but I don't have a simple argument why the highest minimal payout for n red and black cards is (2n)!/(2n-1)!.
 
  • Like
Likes   Reactions: fresh_42
Not sure whether this one will be too easy or too hard. (revised)

86. I am a three digit even natural number which can be written as a sum of two primes in six different ways. None of them is in my factor decomposition, but if I sum up all products of those six pairs, then it is a five digit number and again a product of two primes.

D100
 
Last edited:
87. There are four six sided dice with an unusual numbering:
Die 1: 0-0-4-4-4-4
Die 2: 1-1-1-5-5-5
Die 3: 2-2-2-2-6-6
Die 4: 3-3-3-3-3-3
John chooses first which one he will play with, Jane second. Then they roll their dice once and the higher number will win. Who has the better chances to win, and with which strategy?

D101
 
Hey, 87 looks familiar...
fresh_42 said:
86. I am a three digit even natural number which can be written as a sum of two primes in six different ways. None of them is in my factor decomposition, but if I sum up all products of those six pairs, then it is a five digit number and again a product of two primes.
I had tried the first three examples when it said eight different ways but then didn't check more.
With six different ways the first numbers are 60, 72, 100, 106, 110. The first three lead to a sum of products that is too small. 106 is twice a prime. 110 leads to a sum of products of 10334=2*5167. I didn't check if there are other numbers, but that is a solution.
fresh_42 said:
84. A red ant and seven black ants are placed on a five-meter-long, straight branch. Each ant chooses (randomly) to go to the left or to the right, and then runs in that direction at a speed of 1 meter per 10 seconds. Whenever two ants hit each other, both turn and then continue in the opposite direction. When an ant reaches the end of the branch, then it falls off the branch and out of the game.

How long can the red ant stay on the branch at most?
I have seen this before, but as no one answered:
It doesn't matter which ant is red, for any arrangement of ants we can simply label the one red that will stay on the branch the longest. This also means we can replace the collision process with the ants passing each other without changing the survival times. Now the problem became trivial: 10 seconds per meter at 5 m gives 50 seconds. This maximum will always be reached if one ant starts at the end, facing towards the rest of the stick. Which ant will reach it depends on the details of the arrangement.
 
  • Like
Likes   Reactions: fresh_42
#86
I did the solution with brute force with Mathematica.

110 = {{3, 107}, {7, 103}, {13, 97}, {31, 79}, {37, 73}, {43, 67}}
Sum of products = 10334 = 2 * 5167

166 = {{3, 163}, {17, 149}, {29, 137}, {53, 113}, {59, 107}, {83, 83}}
Sum of products = 26186 = 2 * 13093

182 = {{3, 179}, {19, 163}, {31, 151}, {43, 139}, {73, 109}, {79, 103}}
Sum of products = 30386 = 2 * 15193
 
  • Like
Likes   Reactions: fresh_42
#87
I first calculate the chance of each die winning against each other die. For the first die in columns and the second die in rows, I find for the probability of the first die winning to be
$$\begin{pmatrix} 0 & -\frac13 & -\frac19 & \frac13 \\ \frac13 & 0 & -\frac13 & 0 \\ \frac19 & \frac13 & 0 & -\frac13 \\ -\frac13 & 0 & \frac13 & 0 \end{pmatrix}$$
A negative number means that it is for the second die.

For the first player, the worst loss is 1/3 in all cases. That means that the first player has no reason to choose one of the dice over the others.

For the second player, which die to choose depends on the one that the first player chose. Here is a table of which one to choose:
  • 1: 2
  • 2: 3
  • 3: 4
  • 4: 1
That will give the second player a 1/3 more chance of winning than losing in all cases. This is equivalent to a 2/3 chance of winning and 1/3 chance of losing, also in all cases.
 
Let Shepherd 1 = x and Shepherd 2 = y.
We get the system of equations x+8 = 2(y-8) and x-8 = y+8. Solving, we get x = 56 and y = 40
Thus shepherd 1 has 56 sheep and shepherd 2 has 40 sheep.
 
  • Like
Likes   Reactions: fresh_42
Mr 42, would you care to revise #86? There's more than 1 solution (the 110 already mentioned)
182: (3,179),(19,163),(31,151),(43,139),(73,109),(79,103): 30386=2x15193
 
bluej said:
Mr 42, would you care to revise #86? There's more than 1 solution (the 110 already mentioned)
182: (3,179),(19,163),(31,151),(43,139),(73,109),(79,103): 30386=2x15193
Well, ##110## was what I was looking for. It doesn't really matter whether there is more than one solution. This thread is more for daily fun than mathematical rigor. E.g. I don't demand proofs, as long as the solution is correct.
 
fresh_42 said:
89. What is the secret behind the following sequence?
##\quad \;4-8-12-2-1-7-6-5-3-11-10-9##
Shouldn't this be 4-8-12-2-1-7-6-3-5-11-10-9?
 
  • Like
Likes   Reactions: fresh_42
fresh_42 said:
90. We have an elliptic camembert. Is it possible to cut it into two pieces with one straight cut such that the pieces form a heart?
Yes.
fresh_42 said:
I don't demand proofs, as long as the solution is correct.
Cut the ellipse on a nice diagonal, flip one piece top for bottom then rejoin the sliced faces.