In suicidology, the common view is that ‘rational’ suicides occur only rarely, because the competence of people who want to end their lives is compromised by mental illness. In the Netherlands and Flanders, however, patients’ requests for euthanasia or assistance in suicide are granted in 5300 and 1400 cases a year respectively, and in all these cases at least two doctors have confirmed the patient's competence. The combination of these two findings is puzzling. In other countries one would expect at least some of these people to end their own lives. The article argues that we can distinguish between two types of suicide with clustering characteristics. In cases of the first type, the agent doesn't carefully plan his action, doesn't communicate his plans to relatives or others, and uses violent means. In such cases it is reasonable to presume lack of competence. The other type has the opposite characteristics. The most plausible explanation of our problem is that suicides of the second kind are invisible to suicidology, because they tend not to be registered as suicides at all.