I've read some conflicting views on this regarding the best time to buy strangle/straddle (as a pre-earning trade). In one training video, the instructor said the best time to buy a strangle or straddle is 30 days before the earning release because implied volatility would still be low then. In another forum, I read that it's best to buy it the day before the earning release because we don't want to keep a lot of unnecessary risk on regarding time, so entering the position the day before will leave us set up for the announcement and nothing more. Just curious, what are your thought?
I had a really good run on straddles a couple of years ago. It seems to have changed over time. Now it seems like you can get some time decay coming into play, now personally, I put them on the day before earnings. Oops, I had to edit. You do want to shop around for the actual expiration to strip off as much premium as possible, for as long as possible. Generally, ATM or slightly ITM for a straddle, two or three months out.
Tough to answer. You certainly want to buy before the vol pop. Day before is also well into the vol pop and a poor day to buy. Great debate but many will sell day before and ideal time to sell. Just before earnings can be an ideal time to a spread. Very dependent on the underlying and the street's expectation. Will certainly prompt a much argued discussion. Simply put buy the back month (after earnings) and sell the front month. Most will sell the front straddle and buy the back. You are looking to benefit from the vol crush. This example is from a 2013 Group One training manual. This is often referred to as a trade more suitable for MMs as opposed to a retail investor. 2013 AMZN prices. The manual sits on our desk and pretty much get's used a few times a day.
@ajacobson mostly said what I was going say, but here are some of the considerations (a) stock might move away from your strike if you buy too early (b) the straddle might gain/lose some value if the move is under/overpriced (c) you might not have sufficient liquidity if you wait to the last day (d) the call date might get postponed if you buy too early etc. So, your mileage will vary. PS. Beware of the "vol ramp up" stories, a lot of the increase in IV is simply an event pricing artifact.
@Secret Santa is spot on. There is no monopoly on ideas and you question is probably one of the most asked.
As secret sanata mentioned. The vol ramp is an illusion. The expirations that expire after earnings have 3 "different" vols priced into them. The ambient volatility (volatility leading into the earnings), the event vol or jump and the diffuse vol(the volatility that the stock will experience after earnings) which can be quite large if the stock has a bad earnings. It might actually be a good idea to sell a straddle leading into earnings if you believe these events/volatilities are over priced. As @ajacobson mentioned this is a very popular question (I am curios about it as well), however I have never seen a thread/paper about selling the "vol ramp".
If the earnings is not confirmed the vol is usually not all there. So the "best" time to buy the vol would be before the earnings is confirmed. That is a whole other discussion all together with different risks
Do you guys model the option prices with a more advanced jump model to account for earnings? Or do you trade using experience?
The pre earnings vol pop is stale as a standalone idea itself...routinely faded, and money being made on the spreads rather than price movement. It's the "all else equal" perfect economics class example that never has (and never will) manifest itself in reality beyond a contrivance of statistics. Every single market participant is the "all else", and it's decidedly not equal. Where it works is in tandem with other trade ideas that will exploit an expected volatility increase to offset adverse price movement (for example buy a put subject to earnings and a call a week earlier to catch an upwards pre earnings price move with little theta risk). I've only been successful on long straddles when a stock went vertical, and I knew it was going straight up then straight down...just didn't know where the top would be in between; and also on catching unexpected jumps in volatility unrelated to earnings.
Earnings is 1 jump, so you dont need to use a jump model. That being said you have to have a model for this. I can't imagine doing it on experience. To many numbers involved