The other thing folks need to consider: How many revolutions actually resulted in good outcome? If you did a bayesian analysis of the historic data of all revolutions looking for potential of good outcome vs bad outcome you would walk quickly away from revolution.