Attempting suicide can be a rational choice, but only if there is a high likelihood it will cause the attempter's life to significantly improve.Marcotte couldn't test the relative "life improvement" of successful suicides—since they were, of course, dead—but he could study those who had failed at suicide to determine if their lives improved after the attempt. The results are surprising. Marcotte's study found that after people attempt suicide and fail, their incomes increase by an average of 20.6 percent compared to peers who seriously contemplate suicide but never make an attempt. In fact, the more serious the attempt, the larger the boost—"hard-suicide" attempts, in which luck is the only reason the attempts fail, are associated with a 36.3 percent increase in income. (The presence of nonattempters as a control group suggests the suicide effort is the root cause of the boost.)A commenter links to this piece on a possible evolutionary link for depression. See Hagen
It seems to me possible that there's a correlation to the evolutionary explanation for such things as peacock tails and conspicuous displays, known I think as "handicapping". The idea is that animals do things that make no apparent sense except to send the signal that they are fit. The bigger the horns, the more striking the tail, the higher the jump, the more dangerous the exploit--each one is a social signal showing more evolutionary fitness.
Depression and suicide attempts might work similarly--the more you invest in showing your unhappiness, the more convincing the signal, and the greater the chance for reaction.