0
$\begingroup$

Is it possible to tell if a point on an elliptic curve is less than half of the curve's order?

If I have a point $𝐴 = [a]𝐺$ on a curve with prime order q, is there an efficient way to know that $a < q/2$?

I understand that range proofs would work for this, but is there a quicker way? Specifically, I am working with secp256k1, but any advice is greatly appreciated.

$\endgroup$
1

1 Answer 1

2
$\begingroup$

If there were a polynomial time solution then this would provide a polynomial time solution to the elliptic curve discrete logarithm problem. We strongly believe this not to be the case.

To see the reduction to the elliptic curve discrete logarithm, suppose that I have a point $P_0=[x_0]G$ where the order of $G$ is $q$ and I wish to know $x_0$. I calculate $2^{-1}\pmod q$, calculate $[2^{-1}]P_0$ and run my magic algorithm. If the algorithm says that there exists an $a$ with $0<a<q/2$, then I know that $x_0$ is even other wise I know that is is odd. Writing $b_0$ for the low bit of $x$, I write $x_1=(x_0-b_0)/2$ and compute $P_1=[2^{-1}](P_0-[b]G)=[x_1]G$. I can now repeat the process to recover the low bit $b_1$ of $x_1$ and so on, terminating when $P_n=G$. This will take at most $\log_2 q$ steps.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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