You are correct. Surfaces, when properly normalized, are fairly stable over time. But, if you can identify a distortion, that presents a possible opportunity. The distortions, should they occur, would be difficult for a retail trader to exploit. There is no "edge" for a retail trader in in the typical surface. In an homage to TastyTrade, reversion of volatility is one of the few things that retail can take advantage of.
I mean isn't it such that for a retail trader most pnl will be attributed to direction of vols more the then anythings else by a large factor... The frictional cost of extracting some abnormal skew or kink mostly outweighs the slight anomalies... I'm still not clear on what he is trying to get across... But I'm not the smartest guy in the room in any room... I've always though in my surface constructions that I didn't account with the decreasing liquidity amoung strikes etc... I just normalized it by taking a mid or on eod just a close... Whose to say that unknown wouldn't cause so much friction going in and out of the market that the kink is irrelevant... Sometimes I've felt these are lower return higher risk play... As for one it relies on alot of assumptions and the other more influencing factors negate the trades entirely .. For example a hedged risk reversal to trade skew.... Something a market maker might do as a result of large order flow for calls... Erect a proverbial fence and trade hedge with Delta until it normalizes .. Most likely a guy like me doesn't have a automated hedge algo and takes a hit on Delta trading and gets killed with commissions.. Times the fact that profit can be very much in what the hedge frequency is.....