Questionable Comps
Vol #150: October 12th, 2023
Four years before the Miracle on Ice, Americans had an even greater victory over foreign adversaries. It was a wine tasting so famous they made a movie about it. The California farmers swept the Gallic vignerons in a blind tasting of the top bottles from the meilleurs chateaux.
The event was hosted by a British journalist who was looking to drum up interest in his Parisian wine shop. Somehow Steven Spurrier convinced nine of the most elite palates in French wine society to score and rank a selection of the finest Chardonnays and Cabernet Sauvignons from each shore.
There was a host of journalists nestled on the patio of the Intercontinental Hotel ready to write the obvious punchline. The Yankees win.
For both colors a California wine came out on top; Chateau Montelena for the blanc and Stag’s Leap for the rouge. Judges were confounded. Top sommeliers poo pooed Montrachets as tepid and heralded lush Californians as the return of great Burgundy.
But the results did not lie; Montrose, Meursault nor Mouton were a match for a former political philosophy professor with less than a decade in the business.
As long as I’ve known about this tasting it has grated at my understanding of wine. Clearly French wine is better. I’m not just talking my thousand gallon book, the market matches what any drinker with sophistication knows. DRC prices make a Screaming Eagle whimper, and Chateau Rayas will always be the Sina Qua Non of grenache.
Before I dissect why the results are complete plonk, I’ll stipulate that I accept them. A trades a trade, and California showed better. But the famous Judgment of Paris is flawed in two interesting ways, and the rule structure of how these wines were compared even gives us insight into the comparisons and allocations at play in the options markets.
If there is one dimension where this event was perfectly structured, it was the judge selection. The olfactory caliber here is as elevated as it is varied. The heads of Domaine de la Romanée Conti and Chateau Giscours. Two top wine journalists and three owners/sommeliers from three star Michelin restaurants. The leader of the Institut Oenologique de France and INAO’s Inspecteur Général in charge of monitoring wine quality. Nine guys that know wine, from every different angle.
However, despite the vinous luminaires overseeing this event, from the very start the deck was stacked - perhaps unintentionally - in favor of the Americans. There were six California wines against four French in each category. In any Heisenbergian measurement of quality at a particular moment in time there will be some noise. The Americans have two extra tickets in the basket to benefit from noise. Always have randomness on your side.
But the more nuanced comparable question has to do with the vintages selected. Three of the French reds were from a high yielding and thus flabby 1970 vintage. The winning California red came from 1973, a year with the perfect growing rhythm of cool spring and early summer to preserve freshness which then melted into a hot August to complete the ripening. Further, a six year old Bordeaux can be in a cranky tannic stage while a three year old California cab is rich and sumptuous.
If this sounds persnickety, it’s because the details in comps matter. Real estate agents will add or subtract pools and bedrooms with ham fisted arithmetic to make a neighboring property look like your purchase price. That’s a sales technique, not a comp. For options traders, comparables are debating whether the ATM option in the next month is a better hedge than an OTM option in the same month.
The concept of likeness is very important for derivatives traders. Every option is a slightly different bet with a different cost and payoff. Unless someone else is willing to buy or sell that same option back to you, there is some exposure to hedge (or not). Dealers might buy or sell stock immediately to hedge the most direct risk of stock movement, but that only covers one dimension.
To manage all the other variables that change as time passes, such as assumptions about volatility and rates, we hedge with other options. If I’m long the AAPL $180 Call in October, selling the $185 call in the same month is a very like option. Not only does a spread give us a nice defined risk structure, but from a vega, theta, and delta perspective our risks have been significantly neutralized.
It’s not *exactly* the same though, and that relationship can break down in a couple of different ways. In 1976 the difference between a ‘73 and a ‘70 would seem massive, not so much if those bottles were popped today. (In fact the qualities that made a Stag leap would likely punish an old hart.)
As options approach expiration their decay accelerates. The delta of that further OTM call wilts faster and the PnL of the spread will feel more dependent on the long leg - particularly if stock is between the two strikes.
Besides time, changes in implied volatility will have an effect on the likeness of comparable options. The other side of the theta coin, if you increase the implied volatility, those options will get more alike. In a world where stock could land anywhere, what's the difference between $180 and $185? Shrinking vol acts like passing time, divorcing these contracts from everything but their intrinsic value.
Yields, rates, and volatility distributions conceptualize as continuous variables. But dividends get paid on a single day, and the options expiring today have it, while the ones next week don’t. Two expirations that are nominally close will have same strike pricing that looks out of whack. Like, but for one critical point.
What’s comparable today, is not necessarily comparable tomorrow once all your pricing assumptions have been thrown out the window. This is what necessitates the constant rebalancing and pricing of a market maker’s book. If your likeness changes in ways you don’t expect or account for, it’s time for a PnL attribution.
The concept of likeness can also be self referential. If you ask how much an options price has changed today, a superficially basic question has much nuance. It’s trivial to compare the dollar price, and so the next step might be to say that the implied vol has gone up or down by so many points.
That’s for a specific fixed strike, but it can be abstracted further and say the implied vol of 32 delta calls has gone up or down. In certain respects it’s a more interesting question to compare how an options price has moved relative to like options, rather than its own change. The skew curve shows that options of different moneyness have different implied volatility levels; if the delta of a given strike changes because stock moves, implied volatility is expected to shift.
If the Bordelais were set up for an unfair fight when the teams were selected, the nuances of the scoring system further slighted them. There are many ways to referee a match or tasting and the rules for the race determine which horses will show best.
The chosen scoring system by Spurrier was simply to sum up the total points awarded by each judge. That seems simple and reasonable enough, and you end up with the scenario where Stags Leap came in at 127.5 points, just two ahead of Haut Brion. However three of the judges ranked Haut Brion as the top wine yet only one did so for Stags Leap.
To what attributes should we allocate our blank slate of twenty points? Does a point from one judge equal a point from another? Are all dollars and cents of price improvement in payment for orderflow equal?
In options market structure, the rules of the different exchanges determine which participants' strategies will be most successful, and who gets a cut of the cheese. The pro rata market model incentivizes big size. Because inbound market taking (buy side) flow will be chopped up based individual’s percent of the total size shown, you want to be the biggest part of that. And there’s safety in numbers.
The skill here is in maximizing your sizing on the good trades while avoiding the bad. Most of the time hand raising and join quoting the market will get you little slices of easy good money. But one big sweep because you’re too slow or too aggressive will trounce that PnL.
Price/time exchanges give the first person to set the market their full size before waterfalling down the line. The back of the line gets nothing. This encourages participants who are confident about pricing to set a tighter market and become the first in line. Turning a market first has nice rewards (if you’re right) but comes with the risk inherent of seeing something differently than everyone else.
For most customers this exchange nuance is a very unimportant distinction. You likely don’t even have the ability to direct to a specific exchange on an order. But likeness in options is an important factor to monitor for buy side strategies.
If you’re running a hedged equity strategy that has short upside and long downside overlays, the rates at which those out of the money options lose value will accelerate into expiration. Rolling early lets you capture some of that value. Any time a trade moves significantly away from your starting point, or significant amounts of time have passed, it’s worth reevaluating how like your spread still is.
Taste wine blind, but trade options with your eyes wide open.