The distribution of the number of chord intersections This is a follow-up to this question: Given $n$ random chords of a circle, what is the distribution of the number of intersections? Random is defined by "endpoints uniform on the circle".
Update Numerical experiments indicate that the distribution converges to a gaussian.
 A: For the expectation, linearity of expectation should help a lot. There are ${n\choose 2}$ pairs of chords, and if each chord is drawn i.i.d. then each pair has some probability of intersection of $p$, so the answer is $p{n\choose 2}$.
In the case where both endpoints of each chord are drawn uniformly at random, I believe $p=1/3$. Let's call the two chords AB and CD and imagine we first randomly place A; this divides the circle into a line segment, say, clockwise. Now the chords intersect if the ordering on this line segment is C, B, D or D, B, C, but do not intersect for any of the other four orderings. All six orderings are equally likely, so there's a $1/3$ chance that two chords will intersect. 
A: A similar question came up in 2002 in usenet:alt.math.recreational and I gave an empirical answer which translated to this question would be that for $0 \le k \le \frac{n(n-1)}{2}$ the probability of $k$ intersections among the $n$ chords is 
$$\frac{1}{(2n-1)!!} \sum_{j=0}^m (-1)^j  {(n-j)(n-j+1)/2 - 1 - k \choose n-1} \left( {2n \choose j} - {2n \choose j-1}\right) $$  
where $(2n-1)!! = \frac{(2n)!}{2^n \, n!}$ is the double factorial and $m=\Big\lfloor n+\frac12 - \sqrt{2n+2k+\frac14} \Big\rfloor$
Some values (not divided by the double factorial) are in OEIS A067310 and  OEIS A067311 with the latter entry giving many related references.
As usul has said, the expectated number of intersections is $\frac{n(n-1)}{6}$.  Empirically the variance seems to be $\frac{n(n-1)(n+3)}{45}$. I could well believe that suitably relocated and rescaled this converges in distribution towards a Gaussian distribution as $n$ increases , though for small $n$ it seems to be slightly right-skewed: the diagram shows the distributions for $n=15$, with a red Gaussian comparison illustrating this.
 
A: The number of crossings in a chord diagram depends only on the order that the ends of the chords appear on the circle.  Also, for any chord diagram every pairing of the $2n$ endpoints is equally likely.  Therefore, it suffices to take any $2n$ distinct points on the circle and pair them at random.
This is a widely studied problem. A survey and the asymptotic distribution of the number of crossings appears in this paper (note the ridiculous price Springer wants for this conference proceedings), and can be read for free here. 
