Estimating the knee/elbow point in error curves is a challenging task. However, most of the time these points represent ideal compromises or ideal parameters for several tasks, methods and algorithms. Our focus is determining the ideal number of clusters autonomously. In this paper, we formalize the notion of knee/elbow point, discuss known methods to determine it and propose our own method. Contrary to most methods, ours is resilient to long tails in the error curve. This behaviour is especially important when considering autonomous methods. The proposed method outperformed the competition on five datasets from UCI Machine Learning Repository.