Interesting. I think the question should be: when, besides buying and selling straddles, does it make sense to use options viz-a-viz an outright stock position?
Say I have a model for the S&P500 with a high sharpe ratio, long-only, holding period is 1 day. Because of the high sharpe this means that volatility will go down during my trade. Therefore by selling puts I could make more money than trading the stock outright. Now from the distribution of 10'000 simulated trades I could build an implied distribution and build a combo based on this distribution. In my view it is unlikely, that a) a discretionary trader will ever have a view which is that accurate, b) even given this accuracy one will make more money by trading options than the outright stock. So I agree with the basic sentiment, that basically very few people really need options and the world wouldn't change that much if all options would disappear tomorrow. Same thing with most trading related matters. But then, one could say the same about a lot of other activities and industries.