Earnings Volatility Plays

[maninjapan]

Quote from spindr0:

1) you said candidates needed a minimum IV gap of 15 points

5 BP's needed to evaluate a position NOT do it


2) So this would mean your selling near months and buying more far out months(higher strike), correct? this gives you a positive Vega position. I would have thought if you are betting on a close of the gap and a general decrease in the IV you would be heavy on the near options. What is the line of thinking with that?

If your belief is that the underlying isn't going to move, go heavy on the near term options.


3) Ive gone back and tried it on a few previous EA's but yet to find th eright conditions for it. Any specific EA's that worked that come to mind?

Here's a hypothetical for you.
Consider a calendar strangle.
Stock at 52
Price a near month 50/55 at .65
Price a 2nd month at .50
Overnight post EA expectation is .45 (pretend both will go there)

What would be a ratio of contracts bot/sold that would balance the maximum loss in either direction? Then, would the risk graph have an acceptable risk/reward ratio?

Do the same evaluation for diagonals.

These types of positions may not be for you but until you start modeling them, you won't know and all the word descriptions in the world will be meaningless.

Spindr0, I appreciate this exercise, however Im not sure I know how to go about calculating this manually.....
What I can gather from the above info is with a 1:1 ratio, the near month would return a profit of 20 cents (.65 -> .45) and the far mnth would return a loss of 5 cents(.50 -> .45)

I do understand the importance of modelling it. Its just Ive always used the software, never manually....

 

[spindr0]

Quote from spindr0:

1) you said candidates needed a minimum IV gap of 15 points

5 BP's needed to evaluate a position NOT do it

Typo correction. Should read...


15 BP's needed to evaluate a position NOT do it

 

[spindr0]

Quote from maninjapan:

Spindr0, I appreciate this exercise, however Im not sure I know how to go about calculating this manually.....
What I can gather from the above info is with a 1:1 ratio, the near month would return a profit of 20 cents (.65 -> .45) and the far mnth would return a loss of 5 cents(.50 -> .45)

I do understand the importance of modelling it. Its just Ive always used the software, never manually....

You don't want to do this manually.

Calculate your theoretical values. In your software, start off with a +10/-10/-10/+10 double whatever. As per previous example, you're shorting 10 each leg of the near month 50/55 strangle and buying 10 each leg of the next month 50/55 strangle. Since we were talking about selling fewer of the near month, drop it to +10/-9/-9/+10 or perhaps +10/-9/-8/+10 or perhaps +10/-8/-8/+10 etc.

When you find a risk graph that is balanced (assuming you don't have an up or down bias) and has a good risk reward ratio, you have a candidate. Then it's just a question of how many units you want to do and getting a good fill. Oh yeh, a cooperative underlying post EA sorta helps.

If you don't have software that can graph the composite position, it's an exercise in futility.

 

[bebpasco]

Quote from spindr0:

You don't want to do this manually.

Calculate your theoretical values. In your software, start off with a +10/-10/-10/+10 double whatever. As per previous example, you're shorting 10 each leg of the near month 50/55 strangle and buying 10 each leg of the next month 50/55 strangle. Since we were talking about selling fewer of the near month, drop it to +10/-9/-9/+10 or perhaps +10/-9/-8/+10 or perhaps +10/-8/-8/+10 etc.

When you find a risk graph that is balanced (assuming you don't have an up or down bias) and has a good risk reward ratio, you have a candidate. Then it's just a question of how many units you want to do and getting a good fill. Oh yeh, a cooperative underlying post EA sorta helps.

If you don't have software that can graph the composite position, it's an exercise in futility.

Probably shouldn't get in the middle of this but the poster asking the question was talking about CRBs (calendar ratioed backspreads). I think you're looking at double RC's. I think the 15 bps differential for CRBs doesn't work. In fact, I haven't found anything yet that remotely works for CRBs.

 

[spindr0]

Quote from bebpasco:

Probably shouldn't get in the middle of this but the poster asking the question was talking about CRBs (calendar ratioed backspreads). I think you're looking at double RC's. I think the 15 bps differential for CRBs doesn't work. In fact, I haven't found anything yet that remotely works for CRBs.

Hey Beb. A number of things have been discussed in this chain, so it's hard to know what I'm talking about (g). But now, I'm not talking about double RC's.

The point of this example/exercise was to get him to play with his software rather than asking tons of questions which could be more readily answered by a quick look at a risk graph. As you well know, it takes a bit of tinkering to find an acceptable combination. And whether any of it is acceptable to him can't be known until he sees the possibilities...

 

[maninjapan]

Spindr0, dont think I can set up something like that on OptionVue, I will try and find some actual scenarios that match though. I am actually spending a fair bit of time modeling these and running them on actual past EA's as well.
Im still trying to figure out how to run what-if's with different IV's for each month though. The questions try and help me understand the logic behind it all. Once I find myself a couple of examples that work, Im sure it'll all make sense though.
Once again though, the help here is much appreciated.

 

[spindr0]

Quote from maninjapan:

Spindr0, dont think I can set up something like that on OptionVue, I will try and find some actual scenarios that match though. I am actually spending a fair bit of time modeling these and running them on actual past EA's as well.

Im still trying to figure out how to run what-if's with different IV's for each month though. The questions try and help me understand the logic behind it all. Once I find myself a couple of examples that work, Im sure it'll all make sense though. Once again though, the help here is much appreciated.

Yep, examples that work become light bulb moments.
Yep, you're welcome.

And for the answer to the assignment, try a +10/-7/-8/+10 ratio. That would amuse me

 

[dd4nyc]

Quote from dd4nyc:

Let's say you're looking at BBBY. Implied vols in my software are
Sep 34.9% (7 trading days)
Oct 40.7% (27 trading days)
Nov 38.1% (52 trading days)

Using excel you can calculate that market expects 7.25% volatility jump on earnings, and non-earnings vol of 2.2%, or 34.8% annualized.

Furthermore, if you extend these numbers father out, you can forecast that october implied volatility will continue to rise until the earnings date (sep 23 , in 2 weeks ) reaching the high of 43.3%, and november vol will rise to 39.6%. After earnings both implied vols will collapse to 34.8%.

Today, Sep 23rd morning IVs are
Oct 42.8%
Nov 37.7%
Expected earnings jump about 6%