In stepping through an oil put trade, I wrote:
“As a junior trader I remember selling calls because “it’ll never get there”. I promise you there are many people who think like that. They don’t understand vol trading.”
A reader asked:
What’s wrong with selling OOM call options if vol is too rich? If it’s priced as a 1/100 event and it’s closer to 1/1000…that’s a good sell.
Here’s my response:
Given your assumption then I’d agree. But can you think of a scenario even with your assumption where it can still be wrong?
See this post How Much Extra Return Should You Demand For Illiquidity?
It looks superficially unrelated but at its heart is deeply related to the question. It’s like adding another dimension (axis) to your reasoning. The effect of path and many permutations of cross contamination is hard to model but you’d be falling for a streetlight effect to think it doesn’t matter.
Here’s a phrase to inhale:
“options on options”
What is the option to sell another option at some point in its life worth? Where do those “option on options” exist, and how are their value distributed across cross-asset states of the world?
I understand that’s a bit abstract but I’d maintain it’s the best avenue of reflection. But I can also address the question more concretely, even if I don’t think it’s the best answer to the question.
The concrete response is:
If someone is paying as if something is 1/100 when you think its 1/1000 why do you think you’re right? How can you parse the difference between 1/100 and 1/1000 in a non-physical system? Did the odds of GME going to $50 change when someone started betting that it would?
[2 years ago I wrote about CVNA in A Socratic Dissection Of An Option Trade. The stock is up 50x in 18 months.]
There is a good reason why one of the first things option market makers learn is the only way to price a far OTM option is via another option, ideally of similar moneyness. Because as soon as someone says “I bet I can drink 20 beers in an hour” your belief about how likely that is requires a massive update.
[In nerd language — the Bayesian prior on what a tail option is worth, based on some underpowered frequentist sample, is so low confidence that any real bid renders it stale and worthless.]
Something I’ve always found amusing — market-maker firms will interview kids out of college and pose a proposition like “I’ll give you 2-1 odds that I will get more than 5 heads in 10 flips” and the kids will say they’ll take the bet. The interviewer will up it to 3-1 and instead getting suspicious, some of the kids get even more excited. This is the most un-street-smart instinct imaginable. (I’ve heard that SIG has or used to have an employee who could reliably flip heads better than chance who was born to give these interviews. Maybe a reader can verify this.)
In the physical world, buying homeowners insurance doesn’t make a fire more likely. Well, stock and option prices are not the world of physics and chemistry. If someone says something is 100-1 when you insist it’s 1000-1 then I think you just enrolled in a epistemology bug bounty exercise.
(The wickedness of trading is that you’re unlikely to ever get enough trials to find out if you are right or wrong, but since the odds are 100-1 even when you’re wrong you’ll just go about your life as if you were right when in fact you learned nothing.
The corollary is you should be restrained in what you think track records can tell you if a strategy doesn’t do a huge sample of trades. A truth so inconvenient the entire asset management industry, GPs and allocators alike, ignore it.)