The paper develops a simple supergame model of collusion that focuses on the role of fixed (exogenous to game played) system of quantity market shares. Conclusions implied by the model could be used to motivate data - saving markers of collusion based on market price behavior. Following conclusions of the theoretical model we propose marker of collusion based on detecting changes in seasonal parameters of prices in periods of possible collusion. An empirical application of method has been done on well known data of Lysine cartel case.