Good teaching means more time

What would happen if we used lesson observation data to decide how many non-contacts (over and above ppa) teachers got next year? For example if a teacher averaged 3s they would only get their PPA periods, if they averaged 2s they would gain an extra 'free', anyone averaging 1s would gain 2 extra frees. Too risky? Teachers would have to agree that for each extra free they got they would commit to one good practice session and to coach a colleague.