Primer #4: Measurement Not Prediction
Primer #4
As promised, moontower.ai includes a primer which is being dripped 1 post a week right here on substack.
Sign up to get early access here.
An edge means buying below and selling above fair value. In training, market makers learn that trading around fair value, not predicting the future, is the path to profits. Now this is straightforward if you trade around a "hard fair value" like cash/futures basis trades. That's just arbitrage. While arbitrage or the prospect of free money sounds tantalizing, the counterbalance is the tremendous tech, labor and capex it demands.
A more realistic goal is to trade around a "soft fair value" driven by consensus. If the fair football line means the Jets should be getting seven points and you find a counterparty who is willing to accept six, in the long run you'll make money.
This does not require a prediction of the Jets' game. It requires you to see the present clearly.
This requires mapping prices to some conception of fair value. By identifying the fair price for risk and identifying opportunities where the risk varies from consensus you become an advantage player.
The job to be done is accurate measurement.
In the wider world of gambling and investing, we see this principle manifested in benchmarking. Bettors’ records are measured against the spread, not whether they are able to just pick winners. For investors, benchmarking rigor ascends a scale.
“Did you beat the market index?” (Beta vs Alpha)
“Did you beat the market adjusted for volatility?” (Focus on Sharpe or similar ratios)
“Did you beat a market-neutral value portfolio?” (Absolute return factor investing)
“Did you beat the market net of all factor tilts?” (Pod shop focus on “idio” or pure skill)
Moneyball, Sabermetrics, and the sports analytics industry control for noise to find signal. The investing world, especially quant shops aided by large data sets and compute power, is increasingly focused on separating beta. Beta is free and abundant, whereas alpha is scarce. Measurement sophistication is an endless march towards being less “fooled by randomness”.
Measurement in Options
Making sense of what an option price means has been a subject of interest long before the words “Black-Scholes” entered the financial lexicon. Option pricing is a rich discipline that can intellectually tease, inspire, and torment its students. Options contain concrete elements such as hard arbitrage bounds (a put option cannot be worth more than the strike price) as well as murkier abstractions — what does an implied volatility of 30% actually represent?
Absent arbitrage-able prices, we cannot make definitive statements about whether options are cheap or expensive without comparisons to other options. There are few absolutes. But this is also why a devoted practitioner can hope to find opportunities that suit their goals.
Options embed unique information about assets that cannot be seen from spot prices alone.
Option term structures represent the relative demand for volatility across time.
Skew tells us something about the fair distribution of an asset over some time period.
The ratio of implied volatility to realized volatility is a measure of volatility risk premia or VRP.
Prospecting
Measures of term structure, skew, and VRP represent parameters whose values can fluctuate over time. Options and option structures allow investors to bet on these parameters.
By examining parameters from different angles you can find contradictions across assets of varying liquidity. Less liquid markets that disagree with more liquid markets are candidates for further inspection.
Less liquid markets often have idiosyncrasies that we need to normalize for. To be fair, sometimes they are so idiosyncratic that they shouldn't be compared to liquid markets. Natural gas has nothing to do with the SP500.
You will discover that prospecting for trades has little to do with prediction. Instead, the objective is to see the present more clearly. To identify implied contradictions about future states of the world, and stack the deck in our favor regardless of which future unfolds.
References
Key Measurement Concepts For Option Traders
An example of why measurement is the job to be done
The concept of time in the options world is complicated. The following example underscores why measurement requires careful implementation and interpretation.
It's Friday morning and option expiry is next Friday.
My calendar day model says there are 8 out of 365 days until expiry or 2.2% of a year remaining.
Your business day model says that there are 6 out of 251 days or 2.4% of a year remaining.
If your model thinks the IV is 30% my model sees an IV of 31.3%. We are both looking at the same option price in the marketplace but you think there's more time to expiration than I do so the implied volatility your model produces must be lower.
But prices are what we actually trade, not volatility percentages – so our rulers are causing us to disagree. In some markets, a market-maker might kick a grandma down a flight of stairs for 1.3 volatility points of edge!
[This gets more complicated when you consider that variance doesn't pass uniformly through the weekend, through weekdays, or earnings days. In fact, it doesn't even pass uniformly through all hours of the day.]
There is no such thing as a fair football line in which the Jets are only getting seven points
If I'm pricing options using calendar days(365), then I should even annualize realised volatility by multiplying 18.8(√256) instead of 16 (√256, approx trading days). In order to compare the VRP ratio on same scale, am I right?