Wind of Change

There are undoubtedly balance issues, but I have serious doubts that tying control to a specific timeframe will solve them. What parameter should be used to determine this timeframe, from which point control can be exerted? To give players a chance to catch up if they're unlucky with the game master's decisions, it should ideally be no earlier than the middle of the third week, as I understand it. But what about games where both players successfully gain control at the end of the second week? Should we artificially prolong this? What about situations where one player has accumulated a lot of resources, while the other's only chance to compete is to gain control slightly earlier? I completely agree that there is an excessive amount of frustrating randomness in Heroes, to the point where sometimes you just want to give up and only play mirror matches, where the disparity is at least limited to the quality of the heroes and very rarely extends to some overpowered direction. I would probably do that if mirror matches didn't scare off weaker players so much (and sometimes also because of the complexity of the gameplay, which doesn't allow for relaxation) – the disparity is just too obvious there.
In general, I'm saying that in your proposals, you should at least try to imagine what consequences a particular balance change might have from different perspectives. What looks good on paper at first glance is not always good in practice.