Original post on Mathoverflow here.
Given a list of identical and independently distributed Levy Stable random variables, $(X_0, X_1, \dots, X_{n-1})$, what is the is the probability that the maximum exceeds the sum of the rest? i.e.:
$$ M = \text{Max}(X_0, X_1, \dots, X_{n-1}) $$ $$ \text{Pr}( M > \sum_{j=0}^{n-1} X_j - M ) $$
Where, in Nolan's notation, $X_j \in S(\alpha, \beta=1, \gamma, \delta=0 ; 0)$, where $\alpha$ is the critical exponent, $\beta$ is the skew, $\gamma$ is the scale parameter and $\delta$ is the shift. For simplicity, I have taken the skew parameter, $\beta$, to be 1 (maximally skewed to the right) and $\delta=0$ so everything has its mode centered in an interval near 0.
From numerical simulations, it appears that for the region of $0 < \alpha < 1$, the probability converges to a constant, irregardless of $n$ or $\gamma$. Below is a plot of this region for $n=500$, $0< \alpha < 1$, where each point represents the result of 10,000 random draws. The graph looks exactly the same for $n=100, 200, 300$ and $400$.
For $1 < \alpha < 2$ it appears to go as $O(1/n^{\alpha - 1})$ (maybe?) irregardless of $n$ or $\gamma$. Below is a plot of the probability for $\alpha \in (1.125, 1.3125)$ as a function of $n$. Note that it is a log-log plot and I have provided the graphs $1/x^{.125}$ and $1/x^{.3125}$ for reference. It's hard to tell from the graph unless you line them up, but the fit for each is a bit off, and it appears as if the (log-log) slope of the actual data is steeper than my guess for each. Each point represents 10,000 iterations.
For $\alpha=1$ it's not clear (to me) what's going on, but it appears to be a decreasing function dependent on $n$ and $\gamma$.
I have tried making a heuristic argument to the in the form of:
$$\text{Pr}( M > \sum_{j=0}^{n-1} X_j - M) \le n \text{Pr}( X_0 - \sum_{j=1}^{n-1} X_j > 0 )$$
Then using formula's provided by Nolan (pg. 27) for the parameters of the implicit r.v. $ U = X_0 - \sum_{j=1}^{n-1} X_j$ combined with the tail approximation:
$$ \text{Pr}( X > x ) \sim \gamma^{\alpha} c_{\alpha} ( 1 + \beta ) x^{-\alpha} $$ $$ c_{\alpha} = \sin( \pi \alpha / 2) \Gamma(\alpha) / \pi $$
but this leaves me nervous and a bit unsatisfied.
Just for comparison, if $X_j$ were taken to be uniform r.v.'s on the unit interval, this function would decrease exponentially quickly. I imagine similar results hold were the $X_j$'s Gaussian, though any clarification on that point would be appreciated.
Getting closed form solutions for this is probably out of the question, as there isn't even a closed form solution for the pdf of Levy-Stable random variables, but getting bounds on what the probability is would be helpful. I would appreciate any help with regards to how to analyze these types of questions in general such as general methods or references to other work in this area.
If this problem is elementary, I would greatly appreciate any reference to a textbook, tutorial or paper that would help me solve problems of this sort.
UPDATE: George Lowther and Shai Covo have answered this question below. I just wanted to give a few more pictures that compare their answers to some of the numerical experiments that I did.
Below is the probability of the maximum element being larger than the rest for a list size of $n=100$ as a function of $\alpha$, $\alpha \in (0,1)$. Each point represents 10,000 simulations.
Below are two graphs for two values of $\alpha \in \{1.53125, 1.875\}$. Both have the function $ (2/\pi) \sin(\pi \alpha / 2) \Gamma(\alpha) n (( \tan(\pi \alpha/2) (n^{1/\alpha} - n))^{-\alpha} $ with different prescalars in front of them to get them to line up ( $1/4$ and $1/37$, respectively) superimposed for reference.
As George Lowther correctly pointed out, for the relatively small $n$ being considered here, the effect of the extra $n^{1/\alpha}$ term (when $1 < \alpha < 2$) is non-negligible and this is why my original reference plots did not line up with the results of the simulations. Once the full approximation is put in, the fit is much better.
When I get around to it, I will try and post some more pictures for the case when $\alpha=1$ as a function of $n$ and $\gamma$.