In an effort to be productive, I did a backtest of a long straddle pre-earnings. We know IV will rise up into earnings. The question is whether or not the rise in IV plus any moves in the underlying will compensate for theta decay to the point where the structure could be profitable. This is just one example. I picked it out of thin air. While it's not demonstrative of any overall long-term profitability of being long straddles pre-earnings, it's just one case study. See the image below on a pre-earnings straddle of Nike (NKE). It does not include commission costs. Here I examined buying a straddle 4, 3, 2 and 1 week before earnings. We see IV rising but still, each attempt was not profitable because the underlying barely moved and theta decay ate our lunch. There may be some other intelligent filters we can implement that would help us choose better stock candidates that would have a better chance of profit just ahead of earnings, maybe comparing ATR or historical vol as a way to project larger directional underlying moves. Again, this is just one example, but it reinforces my belief that in general options prices tend to be over-priced in the face of uncertainty, even if the date of the uncertainty event is certain. I can imagine that people that were long straddles into Brexit were profitable, but counter-consensus shocks like that don't happen very often. I can imagine there is someone out there (Nassim Taleb?) trading the extreme tails, losing small amounts over and over again (95+% loss rate) until finally something like Brexit happens and they make it all back and then some (lots!), but that's some steely patience and it would be hard to manage client money like that.
Soon you guys will be telling us that since all known information is built into option prices, and theta decay is consistently 'eating your lunch' it is impossible to make any money trading options... at least on the long side. Everybody knows that...or should. Well...I don't think it's impossible... but it isn't easy. You DO have to have knowledge that is counter to current wisdom. Sometimes it is possible to identify situations where the market is 'mispricing' for the wrong reason...you just have to be right where the market is wrong. (remember that "markets can remain irrational longer than you can remain solvent") That's why mostly I sell options. NO not selling straddles...I don't have the guts for that, but selling puts and put spreads on 'safe' stocks that you would be willing to own at the put strike IS a viable strategy...as long as you're satisfied with a modest monthly gain and don't try for a killing.
I am not going to disagree with your statement. By nature options have to be systematically overpriced and bigger the event the more the risk premium. I don't think you are about the problem properly in this case. Think of it the following way - you bought a straddle that contains an event and then contains some expectation of the ambient volatility. If you are thinking of selling it right before the earnings, you are going to enjoy the ambient realized volatility over 3 days and hope that the event vol will increase enough to help cover the decay (and then some, ideally). If you are not actively delta hedging the position, you are going to decay anyway (unless it's a monster event like the ones in biotech).
lets get real specific: Best way to buy volatility if a trader/guesser thinks its "low" and best way to sell it if trader/guesser thinks its "high" "all else being equal"
Call me old fasion, but when I traded straddles I was petrified of any flat trading. Flat trading is a buyer of straddles worst nightmare. Buying a straddle, then the underlying sitting flat for 15 days is no good. So trading low volatile stocks / underlying is no good IMO I always traded mid volatile stocks, you don't need a volatility graph for this. You profit from movement, so why trade low vol? Hopeing to catch a volatility spike? GL. Just stay away from any that have periods of 2 months or so dead flat on the 1 yr (daily bar) chart. I also only traded stocks that had 3-8 significant trends. Not ones that were bouncing all over the place. Or no apparent trends at all.