4
$\begingroup$

I need a little help understanding Hasse's theorem for elliptic curves over finite fields, as well as the proof of this theorem. (Sorry about my editing)

Hasse’s Theorem: Let $E$ be an elliptic curve defined over $F_q$. Then |#$E(K)-q-1|≤2\sqrt{q}$

So Hasse's theorem is estimating the number of points of the E over $F_q$. It says that there are approximately q points with an error of no more than $2\sqrt{q}$ right?

I also read somewhere that "Hasse’s theorem on elliptic curves, provides a bound for the number of points on an elliptic curve when it is reduced modulo a prime p. It’s also referred to as the Hasse bound, because as a result the value is bounded both above and below." but I don't completely understand this result.

Proof: Consider the Frobenius endomorphism on $E$ in $F_q$ where $p$ is prime. This is the map $ϕ:(x,y)↦(x^p,y^p)$.

I guess I also am struggling with the concept of the Froebnius endomorphism. i can see that (x,y) gets mapped to a power p of itself $x^p,y^p)$, but I don't really see the important of it, although I see Frobenius endomorphism constantly being mentioned in elliptic curve literature.

From Fermat’s little theorem we have $x^p≡x modp$.

I understand this.

So the map fixes $E$ pointwise, that is, $ϕ(P)=P$.

This just means every x maps to an $x^p$ which is congruent to x (modp) right? so actually every x maps to itself...?

Then $ϕ(P)-P=0$, so $(ϕ-1)(P)=0$.

Just alegbra I think.

$P∈ker(ϕ-1)$. Thus $E$ is isomorphic to the kernel of the map $(ϕ-1)$.

I don't get where this result comes from, the concept of kernel has always confused me.

The isomorphism yields $N_p (E)=ker(ϕ-1)=(ϕ-1)$.

is $N_p$ just a different notation for the # of points?

So $|(ϕ-1)-deg⁡(ϕ)-deg91)|≤√(deg⁡(ϕ) deg⁡(1) )$.

Totally confused.

But $deg⁡(ϕ-1)=N_p > (E)$,$deg⁡(ϕ)=p$ and $deg1=1$ so $|N_p (E)-p-1|≤2√p$.

Where did these values for the degrees come from??

Any help would be greatly appreciated. Thanks

Here is the proof I used: enter image description here

Alternatively, here is Silverman's proof, which I understand even less: enter image description here

$\endgroup$
  • $\begingroup$ What proof are you following? Silverman's? $\endgroup$ – Álvaro Lozano-Robledo Apr 29 '15 at 18:48
  • $\begingroup$ @ÁlvaroLozano-Robledo I actually can't remember where I found this proof, but I know it's not the one Silverman gives in his "Arithmetic of Elliptic Curves". I totally didn't understand his, this one seemed a bit easier to me $\endgroup$ – Math Major Apr 29 '15 at 18:51
  • $\begingroup$ @ÁlvaroLozano-Robledo I used this proof: math.washington.edu/~morrow/336_09/papers/Igor.pdf $\endgroup$ – Math Major Apr 29 '15 at 18:58
4
$\begingroup$

The idea is, $E$ has lots of points in $\overline{\mathbb F_p}$, not all of which are in $\mathbb F_p$. However, if I have an $\overline{\mathbb F_p}$-point, then it is actually an $\mathbb F_p$-point iff it is fixed by the $p^{th}$ power map. This follows because the solutions to $x^p=x$ in $\overline{\mathbb{F_p}}$ (of which there are $p$) are canonically identified with $\mathbb F_p$.

So computing the number of points of $E$ mod $p$ is the same as counting solutions to $\phi(P)=P$ which is the same as counting solutions to $(\phi-1)P=0$, where $(\phi-1)P = \phi(P)-P$.

Saying that $(\phi-1)P=0$ is exactly saying that $P$ is in the kernel of $(\phi-1)$.

The value of the degrees of these maps is worked out elsewhere in Silverman. Does this help?

$\endgroup$
  • $\begingroup$ Sorry...I know what $F_p$ is, but what is $\bar{F_p}$? $\endgroup$ – Math Major Apr 29 '15 at 19:25
  • $\begingroup$ The algebraic closure of $\mathbb{F}_p$. $\endgroup$ – Álvaro Lozano-Robledo Apr 29 '15 at 20:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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