The probability that a character value is zero for the symmetric group. (Q2509934)

The paper under review proves the striking result that if \(\chi\) is chosen uniformly at random from the irreducible characters of \(S_n\), and \(g\) is chosen uniformly at random from \(S_n\), then \(\mathbb P[\chi(g)=0]\to 1\) as \(n\to\infty\). The analogous result for alternating groups is also proved. The proof is elegant and simple enough to summarize here: a result of Erdős and Lehmer states, very roughly, that most partitions of \(n\) have about \(\sqrt n\log n\) parts. Hence the longest border-strip in a typical partition of \(n\) has length \(O(\sqrt n\log n)\). However by a theorem of Goncharov, all but \(o(n!)\) permutations in \(S_n\) have a cycle of length at least \(n/(2\log n)\). So by the Murnaghan-Nakayama rule, a typical character value \(\chi(g)\) must vanish. The final section of the paper considers the vanishing probability \(P(G)\) for a general finite simple group \(G\). The author raises the following question: `Let \(P_\varepsilon\) be the proportion of finite simple groups \(G\) of size less than \(n\) which satisfy \(P(G)>1-\varepsilon\). Then is it true that for every \(\varepsilon>0\) one has that \(P_\varepsilon\to 1\) as \(n\to\infty\)?' Another interesting question arises if rather than choose \(g\in S_n\) uniformly at random, we instead choose a conjugacy class uniformly at random.