Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

A balanced prime of order n is a prime number that is equal to the arithmetic mean of the nearest n primes above and below.

For example, 5 is a balanced prime in order 1 because it is the average of the prime before it (3) and the prime after it (7).

There is a highest order a prime could be balanced in: the number of primes that are less than it. The question of whether 5, the 3rd prime, is balanced in order 3 is undefined. So the kth prime could be said to be maximally balanced if it's balanced in order k-1; i.e. if you take all the primes less than a prime p, and that many primes greater than it, and average them, you get p again.

The first 500 primes are each less than the average of the primes surrounding them, and the difference tends to increase, but it doesn't strictly increase. So my guess would be that no maximally balanced prime exists, but I doubt there's a quick proof. Am I wrong?

share|improve this question

3 Answers 3

up vote 2 down vote accepted

If $p_k$ is $(k-1)$-balanced ($k>1$) then $$p_k=\frac{p_1+p_2+\ldots+p_{k-1}+p_{k+1}+\ldots+p_{2k-1}}{2(k-1)}. $$ Since $p_1=2$, the numerator is the sum of $2k-3$ odd numbers and $2$, hence odd. Dividing by an even denominator cannot produce an integer.

share|improve this answer
    
Very nice! That means the highest possible order the kth prime can be balanced in is k-2, and there is at least one such prime, 5. –  histocrat Apr 26 '13 at 21:57

I guess there is a simple reason why it can never happen. Prime $p_k$ is maximally balanced if $$ p_k = (\sum_{i=1}^{k-1} p_i + \sum_{i=k+1}^{2k-1}p_i)/2(k-1) $$ In this sum, you have one even number $p_1 = 2$ and $(k-2)+(k-1)=2k-3$ odd numbers. This means that $$ \sum_{i=1}^{k-1} p_i + \sum_{i=k+1}^{2k-1} \equiv 2k-3 \equiv 1 \pmod 2$$ which means that the sum will never be divisible by an even number.

share|improve this answer

Heuristically, the $(n+1)$th prime is $\sim n\log n$. The sum of the first $2n+1$ primes is $\sim S_{2n+1} = \frac{1}{2}(2n+1)^2\log(2n+1)$. We want $(S_{2n+1}-n\log n)\frac{1}{2n} \sim n\log n$. Well, this is true, which suggests (in the weakest possible sense) that there may be infinitely many maximally balanced primes. There are better estimates for the $n$th prime, or the sum of the first $n$ primes, maybe those will give a different result - I am not sure.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.