Struggling to Grasp Vomma/Volga of OTM Options

[HBWTech]

Hello, everyone. Disclaimer: I'm a rookie at the quant things

I've been recently working on some strategy development and have come up against a problem that I can't seem to find a satisfactory answer to:

The structure I'm brainstorming about involves creating a very deep OTM iron condor (or emulates one in terms of non-vol parameters) but with positive Volga. Not sure I'd actually trade it - it has become more a learning exercise and exercise of my stubbornness at this point. However, my understanding of Volga is at odds with what I'm finding in my research (using the analyze tab)

For example: 62 DTE in \GC. Short strikes around 12 Delta and long strikes around the 5 Delta. I've poked around in some AAPL option chains as well so I'm thinking the futures aspect doesn't factor in (though I could be wrong on that certainly)

Now, this is where I'm guessing I've got a mistake in my thinking:

everything that I've poured over would tell me that this trade would have a negative Volga - I.E. the overall magnitude of vega for the position should increase (become more negative as the case is) as volatility increases.
I claim that for this reason: If the bimodal peaks of the Volga curve occur around 1 stdv away from the money, the long options would have a lower Volga than the short ones, and thus, the short options vega would increase (in magnitude) more quickly than the long options' vega would increase (in magnitude) - resulting in an overall vega that is becoming more negative as vol expands.

My confusion comes from using Tos' analyze tab to simulate spikes in volatility for such a position, and what I find is that the vega moves towards zero as volatility gets pushed up. As a matter of fact, I keep finding that the vega of further OTM options is more sensitive to volatility than closer to ATM across all OTM strikes.

So would this be an issue with the model being used to estimate these values under various conditions? I know they use BS, but I've come across a good number of references to Vanna-Volga pricing models that do better at accounting for vol behavior.

Any suggestions would be super appreciated! I've spent quite a number of hours searching, so I'm placing my hope with you all.

 

[tommcginnis]

Not knowing what you're seeing in ToS, it's hard to say, but your intuition is right -- the inside strikes control the situation, and the vega of the position will increase because the net impact on the inside/shorts will out-do the lesser impact of the outside/longs.

FWIW, time does much the same thing. (Although, dayyyyy-um, it cannot reverse. )

I have always thought it useful to think in terms of a circus tent that, over time (or, decreasing volatility), collapses from the outside-in, as the center-pole is lowered, lowered, lowered. To the point where, when the far-OTM longs have hit the ground, the only thing that flops with market movement or time or vol, are the (damn) shorts.

Anywho, your intuition is right-on with BSM, etc.: this Volga flows uphill.

 

[HBWTech]

Thanks so much! It is somewhat relieving to know that my head is at least somewhat effective in synthesizing this information, though I'm still not sure I grasp the entirety of the situation. I'll include a picture of the ToS result which is causing me grief - the middle strike is the current spot price.

So what you're saying is that the Vega of those shorts Does, in actual practice, increase more quickly than the long, further out options, such that the increasingly negative vega of the shorts should overcome the increasingly positive vega of the longs (equivalently, a net negative Volga for the position)? But because ToS is using BSM, wherein Volga is strictly increasing as a function of volatility, this doesn't appear to happen?

Best,
Hollis

 

[Flynrider]

Hello, everyone. Disclaimer: I'm a rookie at the quant things

I've been recently working on some strategy development and have come up against a problem that I can't seem to find a satisfactory answer to:

The structure I'm brainstorming about involves creating a very deep OTM iron condor (or emulates one in terms of non-vol parameters) but with positive Volga. Not sure I'd actually trade it - it has become more a learning exercise and exercise of my stubbornness at this point. However, my understanding of Volga is at odds with what I'm finding in my research (using the analyze tab)

For example: 62 DTE in \GC. Short strikes around 12 Delta and long strikes around the 5 Delta. I've poked around in some AAPL option chains as well so I'm thinking the futures aspect doesn't factor in (though I could be wrong on that certainly)

Now, this is where I'm guessing I've got a mistake in my thinking:

everything that I've poured over would tell me that this trade would have a negative Volga - I.E. the overall magnitude of vega for the position should increase (become more negative as the case is) as volatility increases.
I claim that for this reason: If the bimodal peaks of the Volga curve occur around 1 stdv away from the money, the long options would have a lower Volga than the short ones, and thus, the short options vega would increase (in magnitude) more quickly than the long options' vega would increase (in magnitude) - resulting in an overall vega that is becoming more negative as vol expands.

My confusion comes from using Tos' analyze tab to simulate spikes in volatility for such a position, and what I find is that the vega moves towards zero as volatility gets pushed up. As a matter of fact, I keep finding that the vega of further OTM options is more sensitive to volatility than closer to ATM across all OTM strikes.

So would this be an issue with the model being used to estimate these values under various conditions? I know they use BS, but I've come across a good number of references to Vanna-Volga pricing models that do better at accounting for vol behavior.

Any suggestions would be super appreciated! I've spent quite a number of hours searching, so I'm placing my hope with you all.

May not matter, but You can change some of the modeling options in ToS (BS, binomial, berklystensly,, Volatility smile vs, individual IV vs uniform vol across strikes).

 

[cvds16]

as an ex-option vol market maker and arber I think you should stick to analysing things with simple instruments (not t(hat I would trust the model known as Bull&Shit ... read Nassim Taleb on Risk. If you don't trade exotics, skip the second part. It's expensive but money well spent.

 

[tommcginnis]

I don't really think you have an volatility question -- I think you have a ToS question.

Your post posited a iron condor -- the thing at the bottom of those two screen shots. But what's above them -- and apparently giving you grief -- are three Stk Prices -- which might be Stock Prices (which don't make sense in an IC analytics framework), or Strike Prices (which don't make sense quoted in pennies). Some sort of butterfly to start with? Ugh.

Again, I see your intuition as spot-on. (Google "Option volatility over time" and scan the images: nothing will be new to you.) But if you're getting strange numbers, I'd question the ToS, not BSM basics.

 

[cvds16]

I agree though BS is a rather robust framework if you really know it flaws. But I programmed better myself in Visual Basics while i was an options trader. I was really good how to model skew on DIA.

 

[HBWTech]

I don't really think you have an volatility question -- I think you have a ToS question.

Your post posited a iron condor -- the thing at the bottom of those two screen shots. But what's above them -- and apparently giving you grief -- are three Stk Prices -- which might be Stock Prices (which don't make sense in an IC analytics framework), or Strike Prices (which don't make sense quoted in pennies). Some sort of butterfly to start with? Ugh.

Again, I see your intuition as spot-on. (Google "Option volatility over time" and scan the images: nothing will be new to you.) But if you're getting strange numbers, I'd question the ToS, not BSM basics.

Those numbers simulate what the total value of the position for each parameter would be if the underlying were at that price - the middle one is the current price. Basically just summing the 4 options' delta, vega, theta, etc. at that spot price. So I held everything constant (including the DTE) except for boosting vol by 10% - which resulted in a more positive vega for the overall position. At this point, I believe you're probably right though. The question has more to do with the model that is being used to calculate these variables than it does with my conceptualization of them..

Again, I thank you for taking the time to share your input.

Best,
Hollis

 

[HBWTech]

I agree though BS is a rather robust framework if you really know it flaws. But I programmed better myself in Visual Basics while i was an options trader. I was really good how to model skew on DIA.

Thank you! You mentioned "simple instruments" in your previous post. What is our operational definition of an instrument here? Like an analytics software, or option modeling framework?

I've started looking into learning about some GARCH techniques, but my knowledge of code is pretty sparse. Being that I only trade vanilla options at the moment, I'm not sure if the time spent on that would be worthwhile. However, I will certainly begin reading Taleb as I hear that referenced very frequently.

 

[sle]

as an ex-option vol market maker and arber I think you should stick to analysing things with simple instruments (not t(hat I would trust the model known as Bull&Shit ... read Nassim Taleb on Risk. If you don't trade exotics, skip the second part. It's expensive but money well spent.

Personally, I always prefer a model I can understand than a model I have to trust. BS is a reasonable model, you just have to know how to apply it well and know when it breaks down.