MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

Does every set of $n$ points in the Euclidean plane contain three points $A,B,C$ such that the two segments obtained by joining $A,B$, respectively $A,C$ form an angle at least equal to $(1-2/n)\pi$ at the point $A$? (Equality is of course achieved by the vertex set of a regular $n-$gone.)

Pietro Majer's example below can be generalized and shows that $(1-2/n)\pi$ has to be replaced by a somewhat smaller constant (at least for values of $n$ which are large enough). For his example we have to take $(1-2/6)\pi$ instead of $(1-2/7)\pi$. Is the best possible constant $(1-a(n))\pi$ asymptotically substantially better, ie. can $na(n)$ become for example arbitrarily large for $n$ large enough? (It is of course obvious that $a(n)$ is decreasing but how fast?)

Update: For $n=5$, one can get arbitrarily close to $(1-1/4)\pi$: Take a right-angled isocele triangle. Split the right-angled-vertex infinitesimally along a line parallel to the longest side of the initial triangle and add an additional point on the symmetry axis very high above the two infinitesimal points.

share|cite|improve this question
up vote 4 down vote accepted

The key bound is $(1 - 1/n) \pi$, due to Erdős and Szekeres:

           Erdos Furedi Quote
The above is an excerpt from this paper:
           Erdos Furedi Title
The Erdős-Szekeres result is in their 1961 paper, "On some extremum problems in elementary geometry." (PDF download link).

share|cite|improve this answer
How does the configuration of $8=2^3$ points without angles greater than $2/3\pi+\epsilon$ look like? – Roland Bacher Nov 14 '12 at 12:44
@Roland: I added a link to the Erdős-Szekeres paper. Their proof is in Section 4, p.59 ff. I cannot pursue it myself at the moment; sorry. – Joseph O'Rourke Nov 14 '12 at 13:15
Thank you. By the way, the configuration on $8$ points is easy: one can split the central point in Pietro Majer's example into two infinitesimally close points on a line parallel to a side of the initial triangle. – Roland Bacher Nov 14 '12 at 14:55
actually this way we make a new angle (1−1/6)π... – Pietro Majer Nov 14 '12 at 17:39

Consider an exagon obtained from an equilateral triangle by cutting three small equilateral triangles from its vertices. Consider the configuration of the $6$ vertices of the exagon, plus the center of the initial equilateral triangle. It seems to me that with these $7$ points one can't do angles larger than $2\pi/3+\epsilon$.

share|cite|improve this answer
Indeed. So what is the best possible constant for a given integer $n$? – Roland Bacher Nov 14 '12 at 9:14
Yes, it seems a beautiful and difficult variational problem. – Pietro Majer Nov 14 '12 at 10:01
Sorry, since I can accept only one answer and since Joseph O'Rourke mention of the Erdos-Szekeres paper kills the problem completely, I have changed the accepted answer to his reply. – Roland Bacher Nov 15 '12 at 9:50

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.