Just One Extra Coin...

Your Math Professor says,

"I want to play a game with you. I have a 100 coins and you have 101 coins, just one more than me.

Both of us will throw all of our coins simultaneously and observe the number that come up heads.

Assuming all the coins are fair, what is the probability that you obtain more heads than me ?"


The answer is 0.5.

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

Satyen Nabar
Oct 23, 2014

Either you throw more heads than your prof, or you throw more tails than your prof, since you have only one extra coin, but not both!

These two mutually exclusive possibilities occur with equal probability.

The probability that you obtains more heads than your professor is 1/2.

The probability is independent of the number of coins held by the players.

This is a fun problem, Satyen, and a clever solution. I did a few binomial calculations to get the same result, but logic is always preferable to brute force. :)

Brian Charlesworth - 6 years, 7 months ago

Tx Brian :)

Satyen Nabar - 6 years, 7 months ago

What about throwing equal number of heads?

Guiseppi Butel - 6 years, 7 months ago

Log in to reply

My take on Satyen's answer is this: if they throw an equal number of heads then the student must have thrown one more tail, and if they throw and equal number of tails then the student must have thrown one more head, so the initial "either/or" statement holds true. That these events occur with equal probability is clear by symmetry. I took a less elegant approach in my solution, but I think that it lends itself better to a generalization of the problem.

Brian Charlesworth - 6 years, 7 months ago
Srinjoy Srimani
Nov 28, 2014

Take m as the number of coins you have. And n as the number of coins the professor has. For a better generalization, n>m.

1/2^n) / (1/2^m)

=(1/2^101)/(1/2^100)

=(1/2^101)*(2^100)

=1/2 paragraph 4

For sake of variety, here's another approach....

Let A A be the probability that both you and your professor get the same number of heads after each tossing 100 100 coins. By symmetry, the probability that you have tossed more heads at this point in the proceedings is then 1 A 2 \dfrac{1 - A}{2} , and this status won't change regardless what the result is of your 101 101 st toss.

If you and your professor are tied after 100 100 tosses, however, then there is a probability of 1 2 \frac{1}{2} that your 101 101 st toss will leave you with more heads. Thus the probability that you end up with more heads than your professor after tossing all your coins is

1 A 2 + A ( 1 2 ) = 1 2 = 0.5 \dfrac{1 - A}{2} + A*(\dfrac{1}{2}) = \dfrac{1}{2} = \boxed{0.5} .

This gets more complicated if you have 100 + n 100 + n coins for n > 1 n \gt 1 . For example, with n = 2 n = 2 , if you are tied after than 100 100 tosses then there is a 3 4 \frac{3}{4} probability that at least one of your extra two tosses will yield a head. That's easy enough to deal with, but we would also have to deal with the possibility that you have one less head after 100 100 tosses, in which case there is a 1 4 \frac{1}{4} probability that your extra two tosses yield heads and hence leave you with the greater number of heads. As noted in my comment to Satyen's solution, I found that that the probability P ( n ) P(n) of having more heads when given n n extra coins to toss is 0.52806 0.52806 in the case of n = 2 n = 2 and 0.54885 0.54885 in the case of n = 3 n = 3 . A general formula for P ( n ) P(n) is in the works; it would probably be worthwhile to introduce the variable k k as well for the number of coins the professor starts with as well, (giving you k + n k + n coins). So now I'm looking for an equation for P ( k ; n ) P(k;n) ..... Or maybe I'll stop being so obsessive about this..... sigh.

Yes you are correct Brian... i worked on it with smaller number of coins...

Difference of 2 coins...

1 coin versus 3 coins -- prob ability of 3 coins getting more heads is 11/16 (0.6875)

2 v/s 4 --- 42/64 (0.65625)

3 v/s 5 --- 163/256 (0.63671875)

Difference of 3 coins----

1 v/s 4 --- 26/32 (0.8125)

2 v/s 5 --- 99/128 (0.7734375)

3 v/s 6 --- 382/512 (0.74609375)

Don't quite see a pattern apart from the fact given the difference is the same, as the quantum of coins increases, the probability of more heads steadily reduces...

Satyen Nabar - 6 years, 7 months ago

Please check my generalization - If the professor has m m coins and the student has n n coins ( m < n ) (m<n) , then the required probability that I get is

P = 1 2 m + n i = 0 m [ ( m i ) j = i + 1 n ( n j ) ] P=\dfrac{1}{2^{m+n}}\sum_{i=0}^{m}\left[ \binom{m}{i} \sum_{j=i+1}^{n} \binom{n}{j} \right]

Pratik Shastri - 6 years, 7 months ago

Log in to reply

Yes, I think that must be right. I made a mistake with my initial calculation in the case where the student had 3 more coins; my new calculation is exactly the same as the value your formula gives. It also agrees with my calculation for the case of 4 more coins. Your formula is much more concise than mine so I'll just go ahead and consider yours as the official generalized formula. :)

So if the professor has 100 coins and the student 111, the student will have end up with more heads with a probability of 0.754, and the probability goes up to 0.900 if the student has 120 coins.

Brian Charlesworth - 6 years, 7 months ago

Log in to reply

@brian charlesworth How do you compute these results? I plugged the formula I got into wolfram alpha, but it wasn't able to calculate the probability even for m = 100 m=100 and n = 101 n=101 ..

Pratik Shastri - 6 years, 7 months ago

Log in to reply

@Pratik Shastri Here is the calculation for m = 100, n = 111. It took a while to get the bracketing right, but from here you can just alter the m and n values to get whatever result you want. This calculation does seem to push WA to its limit, however.

Brian Charlesworth - 6 years, 7 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...