But wagering volume and bankroll share are also past data, and they can change over time for various reasons. Wagering volume may be driven downwards to some extent by rising prices, while increase in casino's popularity will drive it upwards. Bankroll share can also have its own dynamic. So, even without variance it's not so easy to calculate "expected" return and it would probably require some economic modelling. But I'm not sure if it's really necessary for investors to do it, because they don't lose a single satoshi of their investment, except to potential variance, so there's no need to calculate if it's profitable or not - the math tells that it is expected to be profitable. The only reason for making  such deep analysis is if potential investor tries to compare future performance of different casinos.