Chernoff is good; it leads to some ridiculously strong concentration inequalities. If I recall correctly, the result that "a travelling salesperson tour on n un

Chernoff is good; it leads to some ridiculously strong concentration inequalities. If I recall correctly, the result that "a travelling salesperson tour on n uniformly selected points in [0,1]^2 is very, very likely to be of length very, very close to its mean" is derived using Chernoff (well, Talagrand, but Chernoff is closely related).