If I am not mistaken, the equality of the $p$-Selmer rank and the free rank of an elliptic curve are conjectured to be equal. This is one of the many implications of the Birch and Swinnerton-Dyer conjecture.

I want to ask, and excuse me if this is "stupid": Is it enough to show the equality of the ranks for a certain set of primes $p$ to prove the validity of the Tate-Shafarevich conjecture for a given elliptic curve or does one have to prove it for all primes $p$?

  • 2
    $\begingroup$ They are equal if and only if the $p$-primary part of Sha is finite. This is part, sort of an algebraic part of BSD. The more important point is that has all something to do with $L$-functions. $\endgroup$ Jul 23 '16 at 21:45

You first statement is correct, both ranks are expected to be equal. In particular we have:


So if either $Ш$ or its $p$-primary part are finite, then the equality holds. And conversely, the equality for any prime implies the Tate-Shafarevich conjecture.

But none of this implies the Birch and Swinnerton-Dyer conjecture, although it is all closely related (see for example the partity (and p-parity) conjectures).

The $p$-Selmer rank is the more accessible tool which allows us to bound the rank, as for example in the recent breakthrough of Bhargava and Shankar. But notice that they don't prove new cases of BSD, they quantify the ones known from Gross-Zagier-Kolyvagin-Breuil-Conrad-Taylor-Wiles ($\leq 1$).

  • 2
    $\begingroup$ "Wiles" being in turn shorthand for Breuil-Conrad-Taylor-Wiles. $\endgroup$ Aug 2 '16 at 17:09
  • 2
    $\begingroup$ @NoamD.Elkies Of course. Now, that's a mouthful... $\endgroup$
    – Myshkin
    Aug 2 '16 at 18:21
  • 2
    $\begingroup$ Just to be clear: for the displayed equality to be true, Sel$_p$ should mean the $p$-infinity Selmer group (the $p$-Selmer group also sees the $p$-torsion of the elliptic curve) , which is, in general, not of finite rank over $\mathbb{Z}_p$, but of finite co-rank (i.e. the Pontryagin dual has finite rank) so that "rank" in front of Selmer and in front of Sha should be "co-rank". $\endgroup$
    – Alex B.
    Aug 2 '16 at 20:13
  • $\begingroup$ It is of finite rank over $\mathbb{Q}_{p}/\mathbb{Z}_{p},$ sure. $\endgroup$ Aug 2 '16 at 20:32
  • $\begingroup$ Does this mean that for a fixed elliptic curve and prime $p$, one may computationally verify finiteness of the $p$-primary part of Sha by finding independent $r$ independent rationally points and simultaneously showing that the mod $p$ Selmer group has rank at most $r$? $\endgroup$ Feb 14 at 0:30

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.