How are the apa_p of an elliptic curve distributed?
Sagebrush
November 2026
Sagebrush
The Sato–Tate law
November 2026 1 / 10
An elliptic curve
E: y2+y=x3−x E:\ y^2 + y = x^3 - x
  • •
    Conductor N=37N = {37}, discriminant Δ=37\Delta = {37}.
  • •
    jj-invariant j=11059237j = {\frac{110592}{37}}, not an integer, so EE has no complex multiplication.
  • •
    Rank 1{1}: E(Q)E(\mathbb {Q}) is infinite.
Sagebrush
The Sato–Tate law
November 2026 2 / 10
Counting points modulo pp
For a prime p∤Np \nmid N, reduce EE modulo pp and count:
#E(Fp)=p+1−ap. \#E(\mathbb {F}_p) = p + 1 - a_p.
pp22335577111113131717191923232929
#E(Fp)\#E(\mathbb{F}_p)55778899171716161818202022222424
apa_p−2-2−3-3−2-2−1-1−5-5−2-200002266
Sagebrush
The Sato–Tate law
November 2026 3 / 10
Hasse’s bound
Hasse (1936):  ∣ap∣≤2p|a_p| \le 2\sqrt {p},  so  xp=ap2p∈[−1,1]x_p = \dfrac {a_p}{2\sqrt {p}} \in [-1, 1].
a plot
The points look random. Is there a law?
Sagebrush
The Sato–Tate law
November 2026 4 / 10
The Sato–Tate law
Theorem (Sato–Tate, proved 2008–2011)
If E/QE/\mathbb {Q} has no complex multiplication, the xpx_p are equidistributed in [−1,1][-1, 1] with density
μ(x)=2π1−x2. \mu (x) = \frac {2}{\pi }\sqrt {1 - x^2}.
Writing ap=2pcos⁡θpa_p = 2\sqrt {p}\cos \theta _p, the angles θp∈[0,π]\theta _p \in [0, \pi ] follow 2πsin⁡2θ dθ\frac {2}{\pi }\sin ^2\theta \,d\theta .
Proved by Clozel, Harris, Shepherd-Barron and Taylor when jj is not an integer, and by Barnet-Lamb, Geraghty, Harris and Taylor for every curve without complex multiplication.
Sagebrush
The Sato–Tate law
November 2026 5 / 10
Evidence: the moments
The even moments of μ\mu are ∫−11x2k μ(x) dx=Ck4k\displaystyle \int _{-1}^{1} x^{2k}\,\mu (x)\,dx = \frac {C_k}{4^k}, with CkC_k the Catalan numbers.
For the 1227{1227} primes p<10000p < {10000} with p∤2Np \nmid 2N:
2k2kmean of xp2kx_p^{2k}Ck/4kC_k/4^k
220.2470.24714\frac{1}{4}
440.1200.12018\frac{1}{8}
660.07440.0744564\frac{5}{64}
Sagebrush
The Sato–Tate law
November 2026 6 / 10
Evidence: the distribution function
The fraction of the xpx_p at most tt (blue, joining the sorted points), against the Sato–Tate distribution function (red)
F(t)=12+t1−t2+arcsin⁡tπ. F(t) = \frac 12 + \frac {t\sqrt {1 - t^2} + \arcsin t}{\pi }.
a plot
Sagebrush
The Sato–Tate law
November 2026 7 / 10
Complex multiplication is different
G: y2+y=x3−7G:\ y^2 + y = x^3 - 7 has j=0j = {0} and complex multiplication by Q(−3)\mathbb {Q}(\sqrt {-3}).
For p≡2(mod3)p \equiv 2 \pmod 3, GG is supersingular and ap=0a_p = 0: that is 616{616} of the 1227{1227} primes 5≤p<100005 \le p < {10000}, about half.
So the xpx_p of GG can’t follow μ\mu : half of them sit at 00.
Sagebrush
The Sato–Tate law
November 2026 8 / 10
Try it
Pick a curve and a bound. The curves 27a1 and 32a1 have complex multiplication.
Sagebrush
The Sato–Tate law
November 2026 9 / 10
References
  • •
    H. Hasse, Zur Theorie der abstrakten elliptischen Funktionenkörper I–III, J. reine angew. Math. 175 (1936).
  • •
    L. Clozel, M. Harris, R. Taylor, Automorphy for some ℓ\ell -adic lifts of automorphic mod ℓ\ell Galois representations, Publ. Math. IHÉS 108 (2008).
  • •
    R. Taylor, Automorphy for some ℓ\ell -adic lifts of automorphic mod ℓ\ell Galois representations II, Publ. Math. IHÉS 108 (2008).
  • •
    M. Harris, N. Shepherd-Barron, R. Taylor, A family of Calabi–Yau varieties and potential automorphy, Ann. of Math. 171 (2010).
  • •
    T. Barnet-Lamb, D. Geraghty, M. Harris, R. Taylor, A family of Calabi–Yau varieties and potential automorphy II, Publ. RIMS 47 (2011).
  • •
    A. V. Sutherland, Sato–Tate distributions, Contemp. Math. 740 (2019).
Sagebrush
The Sato–Tate law
November 2026 10 / 10