Stickiness wrt Vega, straddles, etc

[ducatista]

Let’s say you expect some binary event and IV just happens to be relatively low. You don’t know how market will react, but you know itll be a strong move, so you open a long Straddle

How is your Straddle affected by the smile dynamics?

For example if equity markets are usually sticky strike / sticky local vol and there’s negative skew, this would mean vol deflates as spot increases, and increases when spot falls.

So if spot crashes after your binary event, it makes sense that your Straddle would be profitable. Vol increases in a crash, and your long Straddle makes money

But in an upside move vol decreases, so wouldn’t that theoretically mean your Straddle wouldn’t be profitable? What am I missing here?

 

[ETJ]

Devils in the detail - the propensity for volatility to be inverse may not hold in an event - otherwise, the reversal would have a great return. In 87 call volatility went up - again lots of theories here, but generally, an event will pop all volatility. A gradual rally is a different story.
If the put volatility was "too high" relative to the call you would short the stock(if available to short)buy the "cheap" call and sell the "expensive" - except there are 10.000s of machine that are going to identify that opportunity and bang it back in line. Devil is still in the details and overall liquidity and the availability of short stock is gonna be an issue.
Parity is Stock+Put-Call=T-bill and conversely -stock-put+Call=-T bill. Natenberg's book does a great job on parity and volatility - worth the read if you haven't read it.
Think of volatility as an imbedded cost of pricing the hedge and they are generally inverse, but all rules are off for an event. In 87 covered writers underperformed on both assets.

 

[ducatista]

Hm, interesting and makes sense. Sinclair (I think it was Sinclair?) says that in a low vol env vol can be sticky delta, so that would make sense.

Sticky Delta would also imply that vol falls when spot falls too, right? I dont think I've ever seen that personally. Maybe I'm adhering to the definitions a little too strictly.

Is it fair to say it is also product/market specific? I feel like this is a cop-out followup since everything is always product specific lol

 

[Matt_ORATS]

Let’s say you expect some binary event and IV just happens to be relatively low. You don’t know how market will react, but you know itll be a strong move, so you open a long Straddle

How is your Straddle affected by the smile dynamics?

For example if equity markets are usually sticky strike / sticky local vol and there’s negative skew, this would mean vol deflates as spot increases, and increases when spot falls.

So if spot crashes after your binary event, it makes sense that your Straddle would be profitable. Vol increases in a crash, and your long Straddle makes money

But in an upside move vol decreases, so wouldn’t that theoretically mean your Straddle wouldn’t be profitable? What am I missing here?

You are missing delta.
If the stock moves more than you paid for your ATM straddle you are profitable. That's the delta component.
Vega will decrease as the stock moves away from strike, so the vega drop on the upside will be mitigated. The longer duration of the straddle, the more vega and the more pronounced vega will be in relation to delta. You said you bought the straddle in a relatively low IV environment so that will mitigate the vol crush effect as well.

 

[ducatista]

^ so the takeaway being that the increase in both Delta and Gamma could offset the losses due to Vega? (again provided that the underlying has sufficient momentum and moves far enough go the upside)

In either case, the straddle would be more profitable to the downside vs a commensurate move to the upside, correct? For example assuming a 50tick move in either direction.

 

[jamesbp]

The longer duration of the straddle, the more vega and the more pronounced vega will be in relation to delta.

Probably better to compare both Raw Vega and some form of Time Weighted Vega if comparing strategies across different expiry dates ...

 

[ducatista]

Hm alternatively could I just use a synthetic straddle to eschew the issue altogether?

Eg long 2 puts + long 100 shares (under the assumptions of sticky strike I'd be protected from skew on the upside no?)