MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

So it was conjectured that if all elliptic curves over $\mathbb{Q}$ are ordered by their heights, then the average rank is $\frac{1}{2}$.

Brummer initially showed assuming BSD and GRH that the average rank is bounded by 2.3. Since then many improvements have been made. In my search, I found the slides for a talk by Manjul Bhargava (linked here:, where he talks about his result showing that the average rank is bounded by 1.5 unconditionally.

My question is has there been any improvement on his result since then? A reference to such a paper would be appreciated as well.

share|cite|improve this question
I think the general expectation is still (or again) in favor of the "minimalist"conjecture of average rank $1/2$, i.e. asymptotically 100% of curves have the smallest rank consistent with their parity, which is even or odd with equal probability. Meanwhile, last I heard Manjul Bhargava together with his student Arul Shankar have pushed the upper bound on the average rank down to $0.99$ . – Noam D. Elkies May 8 '12 at 3:25
Noted. I've edited my question to reflect your comment. – Eugene May 8 '12 at 3:26
up vote 11 down vote accepted

Color me surprised if people no longer believe that the distribution of elliptic curves is half rank zero, half rank one, and a density zero subset of higher rank curves. To my knowledge this is still a conjecture that people believe.

I believe the state of the art results are still due to Bhargava and Shankar, and are best summed up in the slides listed in your post. In particular:

  • The average size of a 3-Selmer group is 4
  • The average size of a 4-Selmer group is 7
  • The average size of a 5-Selmer group is 6
  • These all hold true up to a finite number of congruence conditions
  • A positive proportion of elliptic curves have rank zero
  • Assuming the finiteness of Sha, a positive proportion of elliptic curves have rank 1
  • Unconditionally, the average rank of an elliptic curve over $\mathbb{Q}$ is strictly less than one

There are lots of good expositions of this work (or at least the 2-Selmer result), for instance this Seminar Bourbaki article of Poonen ( ) and this short note of Gross ( ) .

share|cite|improve this answer
I probably misunderstood what he wrote in the slides when he said, "Computations do not currently give much support to the conjecture either," to mean that it isn't widely believed anymore. I apologize for the error. – Eugene May 8 '12 at 3:48

Thank you everyone for the great references.

I also recently found this really good summary by Alice Silverberg about all things related to the rank of an elliptic curve if anyone is interested.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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