Walt Disney’s original interpretation of Alice in Wonderland is transfixingly marvelous.
Tumbling down into Lewis Carroll’s fantasy, young Alice explores a world where nothing is as it seems. On one wonderful afternoon she has unbirthday celebrations, and then plays croquet with flamingos - literally.
Once the soothing lull of a trip down memory lane faded, there were a few things I noticed about the film. The first was how many shortcuts they took with animation, and how little it actually mattered.
Relatively static backdrops were a requirement for early cartoonists. It would take years to repaint a new frame with the microscopic differences that achieve the full 360 degree camera treatment that Elsa of Arrendale gets. We barely notice, as Tweedle Dee and Tweedle Dum are just as entertaining in front of the same few shadowed oaks.
Another pragmatic pruning comes when characters are further away, their features are less detailed. The White Rabbit’s whiskers are buoyantly bushy up close, but scarcely visible as he hops to and fro. You don’t need a highly stylized image to fret alongside him.
While simple by our Pixar standards, the animation expertly tells the story of how beguiling the world is to creatures young and old. The curious mind of Alice has an insatiable optimism and toys with logic puzzles of heady proportions.
Carroll elevates the complex philosophical problems of the day over tea and crumpets. The concept of time as a fourth dimension was being contemporaneously explored, and he mocked this idea by leaving the Doormouse, the Mad Hatter, and the March Hare perpetually waiting for their fourth member Time to join them. Godot never showed up either.
The animation is just a tool, the real genius is how it’s used to tell such a brilliant story.
Not surprising to anyone following my attempts at computer generated art, this made me think of the recent developments in artificial intelligence. Doctors, lawyers, bankers, and artists all ponder how this will impact their professions. The common and oft repeated theme being that AI is a powerful scalpel which will be most successfully wielded by those with talent, expertise, and creativity.
I never had much of a career as an artist, so typing in a few prompts on Midjourney’s Discord never felt threatening. But I’ve been thinking about options since before I graduated college, and writing publicly even longer than that. A conversation earlier this week about time spreads with my friend Jay Soloff led me down a rabbit hole that made me change my tune.
I’ve been building indices to track the value of various common options spreads over time and was looking for feedback. The idea is that if the VIX is a constant 30 day volatility figure (that incorporates some measure of skewness too) what would the constant 30/60 day time spread, or 15 delta risk reversal be?
I’ve been running (flying?) with bird names here, riffing off the nomenclature of classics like the condor and seagull strategies. When I needed a bird that evoked time, Jay nailed it with the Raven. From this hommage to the excellent poem by E. A. Poe, I was inspired to see what G.P.T. Poe could do with this. I asked for a poem about a market maker in his legendary romantic style.
The results are astounding. The full text is below, but one of the better paragraphs is this:
I was intrigued by the offer, for wealth had always been a dream,
So I invited him to sit, and listen to his scheme.
He spoke of calls and puts with glee, and how to hedge with certainty,
Of volatility and strikes, and how to make them all agree.
With each word he spoke, my interest grew, and soon I was entranced.
"Teach me more," I whispered, "and take me to this options dance."
I promise this edition of the Till comes from a sentient organic being, but I’m not sure about future newsletters. You’ll find me at the options dance. Some of the style borders on plagiarism, but OpenAI is clearly getting somewhere fast with the model.
Fretting about my writing career, I decided to challenge this talented scribe with an options theory heads up. Time spreads were an ace up my sleeve, fit to trump even the Red Queen.
A time spread is a position where you’re long one month, and short a similar call or put in another month. Long time spreads involve buying the back month, and short time spreads sell the back month. It’s a bet on the timing of movement across two different dimensions. Long spreads suggest less movement in the near term, more further out, and shorts of course the opposite.
Positioning around events is a natural use case for time spreads (See The Till #99; It’s One Big Event ). When you know there’s going to be a concentrated hunk of variance right around 2pm ET on the first day of February, the near term options will hold a lot of value and not decay as much as usual. This makes certain time spreads get cheaper, because you’re short the options which are entirely a bet on the Fed move, and long options that are also mostly a bet on that same Fed move.
Given that dynamic, time spreads are unique in that long calendars are short gamma. Most of the time owning options means you’re paying time decay to benefit from something moving. A long time spread means you pay for the privilege to hold your breath and hope the night passes quietly by, setting your naked long free. The opposite short time spread is long gamma.
ChatGPT doesn’t know this! I do! A minor victory for us acolytes of Natenberg.
The description here is convincing but faulty. That’s comforting considering how many whip smart and arrogant young men I’ve seen humbled by a Texas hedge. Options have a lot of non-intuitive properties to them. Whether it's Bayesian logic, theta decay, or open interest, even the highest horsepower brains can get tripped up.
ChatGPT isn’t getting a PM job any time soon. The clever journalists at Bloomberg couldn’t even get it to build a market beating ETF. It is however already proving itself as a very practical tool for traders. My data analysis skills have leveled up thanks to the ease of asking a plain English question about Python or SQL and getting a concise, usually well contextualized answer back. Being able to quickly translate ideas into studies is a boon for any decision maker.
There are lies, damn lies, and statistics. Bad data science will bend any sample to your bias. I might have to ask the dumb questions about how to format a column, but I have enough experience and training to sniff out a miscalculated datapoint. With a better tool, I feel like a war weary surgeon who’s been blessed with the steady hands of an eighteen year old.
The most insightful stories are told not by lonely algorithms gorging on lakes of data just as timeless classics aren’t created by high octane graphics engines. The world is a complex place where short time spreads are long gamma. Holding the looking glass up to the eye, it’s Alice, not the spectacles, that makes the Cheshire Cat grin.
The Market Maker
by ChatGPT
The poetry was impressive, Mark. Your articles always impress. Interesting the AI can't do options yet.