Schizophrenia research has focused on single mechanisms mainly. Classically, a paradigm is chosen, such as a cognitive or perceptual test. The outcome is usually that schizophrenia patients perform worse than controls. Substantial research is then spent to understand what exactly causes deficient performance in the test, aimed to identify the abnormal neural mechanisms. However, schizophrenia is unlikely be caused by one single deficit. For this reason, we tested a battery of paradigms, assumed to be endophenotypes of the disease, to understand schizophrenia within a high dimensional space of potential causes. We found very low correlations between the tests. Hence, potential endophenotypes seem not to have a common factor, i.e., they are tapping in different abnormal mechanisms. This finding raises very general questions how to investigate schizophrenia. I will sketch the combinatorial problem and propose a number of research strategies how schizophrenia can be investigated and understood as disease of complexity.