Fair Game

Algebra Level 3

Alice and Charlie are playing a game where they take turns flipping a coin. If Alice flips a heads, she wins, and if Charlie flips a tails, he wins. Alice goes first, and they continue to take turns until one of them wins.

Obviously, if the coin is fair, Alice should win more often because she goes first. But if the coin is weighted, the odds are changed. How should the coin be weighted so that the probability that either wins is 50%?

Give your answer as the percentage probability of flipping heads.


The answer is 38.1966.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

3 solutions

Jacopo Piccione
Sep 2, 2018

If p A p_A is Alice's probability and p p is the probability of flipping head,

p A = p + ( 1 p ) p 2 + ( 1 p ) 2 p 3 + . . . = x = 0 ( 1 p ) x p x + 1 p_A=p+(1-p)p^2+(1-p)^2p^3+...=\displaystyle\sum_{x=0}^\infty(1-p)^xp^{x+1}

because Alice wins if occurs H or THH or THTHH and so on, where H means Head and T means Tail.

So let's solve the series:

x = 0 ( 1 p ) x p x + 1 = p x = 0 ( 1 p ) x p x = p x = 0 ( ( 1 p ) p ) x = p 1 1 ( p p 2 ) 1 2 \displaystyle\sum_{x=0}^\infty(1-p)^xp^{x+1}=p\displaystyle\sum_{x=0}^\infty(1-p)^xp^x=p\displaystyle\sum_{x=0}^\infty((1-p)p)^x=p\frac{1}{1-(p-p^2)} \equiv\frac{1}{2}

And now let's solve for p p :

p 2 3 p + 1 = 0 p^2-3p+1=0

p 1 = 3 5 2 p 2 = 3 + 5 2 p_1=\frac{3-\sqrt5}{2}\;\vee \; p_2=\frac{3+\sqrt5}{2}

The only valid solution is p 1 p_1 because p 2 > 1 p_2>1 . Therefore p = 3 5 2 38.19 % p=\frac{3-\sqrt5}{2}\simeq\boxed{38.19\%}

Mark Hennings
Sep 1, 2018

Condition the problem on the outcome of the first one or two tosses. If Alice throws H, she wins. If Alice throws H and Charlie T, he wins. If Alice throws H and Charlie throws H as well, we are back at the original state, with Alice about to roll again. Thus, if the probability of throwing H is p p and if the probability that Alice wins is q q , then q = p × 1 + ( 1 p ) ( 1 p ) × 0 + ( 1 p ) p × q q = p 1 p ( 1 p ) \begin{aligned} q & = \; p \times 1 + (1-p)(1-p)\times0 + (1-p)p \times q \\ q & = \; \frac{p}{1 - p(1-p)} \end{aligned} Since we want q = 1 2 q=\tfrac12 we obtain p 2 3 p + 1 = 0 p^2 - 3p + 1 = 0 , and so p = 1 2 ( 3 ± 5 ) p = \tfrac12(3 \pm\sqrt{5}) . Only p = 1 2 ( 3 5 ) p = \tfrac12(3-\sqrt{5}) is a valid probability, making the answer \boxed{38.19660113\} %.

Jeremy Galvagni
Sep 1, 2018

If p p is the probability fo heads,

Alice's probability is p + p 2 ( 1 p ) + p 3 ( 1 p ) 2 + . . . = 1 2 p+p^{2}(1-p)+p^{3}(1-p)^{2}+...=\frac{1}{2}

Charlie's probability is ( 1 p ) 2 + p ( 1 p ) 3 + p 2 ( 1 p ) 4 + . . . = 1 2 (1-p)^{2}+p(1-p)^{3}+p^{2}(1-p)^{4}+...=\frac{1}{2}

Multiply all of Charlie's probability by p ( 1 p ) 2 \frac{p}{(1-p)^{2}} to get p + p 2 ( 1 p ) + p 3 ( 1 p ) 2 + . . . = p 2 ( 1 p ) 2 p+p^{2}(1-p)+p^{3}(1-p)^{2}+...=\frac{p}{2(1-p)^{2}}

This is the same as Alice, so 1 2 = p 2 ( 1 p ) 2 \frac{1}{2}=\frac{p}{2(1-p)^{2}}

p 2 3 p + 1 = 0 p^{2}-3p+1=0

p = 3 ± 5 2 p=\frac{3 \pm \sqrt{5}}{2}

The minus solution is about p = 0.381966 = 38.1966 % p=0.381966=\boxed{38.1966}\%


Incidentally p = 1 1 ϕ p=1-\frac{1}{\phi}

If you multiply Charlie's probability by p ( 1 p ) 2 \tfrac{p}{(1-p)^2} , you get Alice's probability. Thus you want p ( 1 p ) 2 = 1 \tfrac{p}{(1-p)^2} = 1 , so that two probabilities are the same.

Otherwise, why don't you just add up the GP to determine Alice's probability as p 1 p ( 1 p ) \tfrac{p}{1-p(1-p)} ,and then solve p 1 p ( 1 p ) = 1 2 \tfrac{p}{1-p(1-p)}=\tfrac12 ?

Mark Hennings - 2 years, 9 months ago

Log in to reply

There are many related ways of thinking about this problem. I just shared the way that came to me first.

Jacopo shows the 'proper' way of dealing with GPs.

Considering only the first two flips as you did is most elegant.

I like the variety of approaches.

Jeremy Galvagni - 2 years, 9 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...