It has been proposed that it is possible to decompose changes in variability of human motor behavior into 3 independent components: covariation, task tolerance, and stochastic noise. The authors simulate learning to throw accurately and show that for this task the proposed analysis does not give an unambiguous answer to the question of what the 3 components contribute to the simulated skill improvement. It is argued that this is caused by the fact that the component covariation depends on the choice of control variables. The authors conclude that it is not possible to distinguish between the 3 components of noise reduction without knowing the controlled variables.