On the importance of replicability in science