Recursos
Resumen
Take one elliptic curve over Q and its quadratic twists. This class shows you how to explain why half the twists have analytic rank zero and half have rank one, and where the proof gets its leverage: a cube of twists whose corners you control.
From Selmer counts to analytic ranks
- Three numbers per curve: the analytic rank a(E), the Mordell–Weil rank r(E) and the 2^∞-Selmer corank c₂(E). The Kummer sequence gives c₂ = r + corank Sha[2^∞], so c₂ is an upper bound that Sha can inflate [0:57].
- Gross–Zagier–Kolyvagin go from analytic to algebraic. Goldfeld needs the reverse arrow [1:13].
- Theorem 1.1 (the 2-converse): if c₂(E) ∈ {0,1}, then a = r = c₂ and Sha is finite. It puts no condition on reduction, CM, torsion or isogenies. A bound on the rank alone is not enough [1:29].
- Think-first: corank 0 gives L(E,1) ≠ 0, corank 1 gives a simple zero, and for corank 2 the theorem says nothing [2:00].
- Smith's census gives corank 0 and 1 for half of the twists each. Because the converse translates every single twist, the counts carry over: densities 1/2, 1/2 and 0 for rank ≥ 2, with no BSD assumption [2:32].
The cube and interpolation
- Twist by h_x = ∏(q*)^{λ_q(x)} for x ∈ F₂^b, with q* ≡ 1 mod 4. The vertex h₀ = 1 is your curve itself [3:29].
- Arrange a unit coefficient at every nonzero vertex: a nonzero central specialization with uniformly bounded 2-adic valuation [4:00].
- A coefficient on the cube expands into monomials, with 2^|S| dividing c_S. Chevalley–Warning then finds x ≠ 0 with A(x) ≡ A(0) mod 2^M. Bounded valuation plus a deep congruence forces the base to be nonzero [4:15].
- What nonzero means at the base: at corank 0 it comes from Kato's zeta class along the real cyclotomic Z₂-extension (L(E,1) ≠ 0). At corank 1 it comes from a Heegner class over Q(√k) through Gross–Zagier, with a companion twist that has a nonzero central value [5:19].
When you read a rank statement about a twist family, look for a Selmer distribution theorem plus a pointwise converse, then find the cube that gives the converse its leverage (§11–13).