Research note · provenance-first

The Volatility Surface Archetype: reading the shape of the options surface, not the price, to infer direction

It reads the level and shape of the options volatility surface, centered on the term structure of implied volatility and how it is moving, and a fixed rule-based read turns that shape into a stance of long, short, or standing aside.

Published Jun 16, 2026

Method archetype

A smooth luminous three-dimensional gridded surface seen in perspective, steeper at the front and flatter at the back with one side lopsided higher, glowing blue to emerald with a coral ridge. — The evidence here is the surface's shape, not where price closed. This model reads a landscape, not a line.

Two markets at the same price can be bracing for very different things. What separates them is not the price; it is how the options market is pricing protection across expiries and strikes, and how that picture is leaning. This archetype reads that picture, the volatility surface, and treats its shape as the evidence. Price is only the thing the surface is leaning on.

What makes this its own discipline is that the read is a shape, not a single number. The model characterizes the level and shape of the surface, judges it against its own recent history, and lets a fixed rule turn that shape into a stance. It reads a landscape rather than a line, and most of the time the landscape is not leaning clearly enough to act on. The sections below walk that read from the surface to the trade.

Two panels: a greyed price line labelled what it does not read, beside a lit surface curve labelled what it reads. — The input is the options surface and how it moves, never the price tape, the closes, or the order book.

What it reads: the surface and its shape

The volatility surface has a vocabulary of shape. The load-bearing read for this archetype is the term structure: how protection is priced across nearer versus further expiries, whether that profile is steep or flat, and which way it is moving. The surface also carries skew, the balance between downside and upside protection, and curvature across strikes. Those are part of what the scan takes in as context, not separate switches that each have to fire before the model can act.

Three cards: term structure marked as the read, with skew and curvature marked as context. — The term structure carries the read here; skew and curvature are the rest of the surface's shape vocabulary, taken as context, not separate switches that each must fire.

Steep or flat, and which way it is moving

The read is a shape rather than a single reading. A steep term structure leans one way, a flat one another, and what carries the signal is how that shape compares with its own recent history and which way it is moving, not a one-off snapshot. Because the comparison is to the surface's own recent behavior, a market whose surface is always somewhat steep is read by the change in its shape, not by the raw level.

Two small curves: a steep term structure rising from near to far expiries, beside a nearly flat one. — Read the lean from the shape: a steep versus flat term structure, judged against its own recent history and which way it is moving, not from a raw level.

Why the shape is read against its own history

Surfaces differ from one market to the next, and some are structurally steeper than others. Judging the current shape against its own recent history is what lets the same method work across markets without a separate rulebook for each: it asks whether the surface is unusually steep or flat for itself lately, and which way it is moving, rather than comparing against any fixed outside level.

From shape to stance

The surface shape is the evidence; the stance is the read. A fixed rule-based regime read takes the current shape and resolves it into one of three stances: a directional-up lean becomes a long, a directional-down lean becomes a short, and a surface that is not leaning clearly keeps the model flat. This is a rule rather than a trader's judgement, and it is a single directional read, not an elaborate engine sorting the market into many states.

A flow: the surface shape feeds a fixed regime read that resolves to long, short, or flat. — The shape decides the stance: a leaning surface resolves to long or short, a neutral one to flat.

When it stands aside

A directional read does not by itself open a trade. Several filters each remove a different kind of ambiguity, and a position opens only when all of them clear. The shape must actually be leaning rather than neutral or half-formed; the move the surface implies has to be large enough to be worth the round-trip cost; the read has to have held over a prior look-back rather than just snapping into place; and the regime must not be reversing from the period before. Because every filter can veto, the model spends much of its time flat, and that standing aside is a decision rather than a gap.

Four rows naming the stand-aside filters: neutral or half-formed shape, edge too thin versus cost, unstable shape, and a reversing regime. — Standing aside is a decision, not a gap. Each filter strips out a different kind of ambiguity, which is why the model is flat much of the time.

Managing the trade

When a position opens its exits are already decided. The rulebook is fixed at entry and run mechanically, with no renegotiation mid-trade: a profit-target exit at a set favorable distance, a protective stop-loss exit at a set adverse distance, and a signal or time-based exit when the surface read no longer supports the position or a maximum holding period is reached. Whichever pathway triggers first closes the trade, and the model records which one did.

Three cards: a profit-target exit, a protective stop-loss exit, and a signal or time-based exit. — Exits are pre-committed at entry and run mechanically; whichever pathway triggers first closes the trade.

How to read it on the dashboard

A list pairing dashboard elements with how to read each: strategy label, review chip, open or flat, benchmark, walk-forward, quality gates, and one live read. — Read each element narrowly: the review chip means live but still being assessed, and the walk-forward is reproduction, not profit.

How it is validated

A three-stage track: train on an older slice, predict an unseen slice, and acceptance gates held behind a review gate. — A good walk-forward means the signal reproduced, not that money was made; the failing gates are the honest risk story.

Walk-forward verification trains the rule on an older slice of history and asks it to predict an unseen later slice, then compares its calls with what the market actually did. A close match means the surface read reproduced the later tape's direction, and stable error means even the wrong calls were small and well-behaved. That is a test of whether the signal reproduces out of sample, not a claim that it made money. The result is scored against a fixed acceptance panel, and the exemplar is held behind a human-review gate while its verification sits in a warning state.

What this archetype is not

Not a price or candle read: its evidence is the options surface and how it moves, not the price tape, the closes, or the order book.
Not a single-indicator switch: it reads the level and shape of the surface across expiries, not one number crossing a line.
Not a percentile rank: it does not rank a value into a band or ladder; that is the percentile-regime family.
Not the reversion-onset method: that standardizes one input and waits for the onset of a turn; this reads the shape of the whole surface and synthesizes a directional regime.
Not the put-skew method: that reads one downside-protection slice as a regime and confirms it with a second read; this reads the surface's broader shape, led by the term structure.
Not a volatility forecast: it does not predict how volatile the market will be; it infers a direction from the surface's shape.
Not tied to one instrument: the method is a fixed family identity independent of which market it reads.

Reusable across instruments

What makes this an archetype rather than a single trade is that it is defined by how it reads, the shape of the surface, not by which market it reads. Point the same method at a different instrument and the recipe is unchanged: characterize the surface shape against its own recent history, turn the shape into a directional read with a fixed rule, stand aside unless the filters clear, and manage the position with a pre-set exit rulebook. The documented model is one instantiation of that template, and its regime orientation and holding horizon are read from its own behavior, not inherited from the family.

Specific models built on this archetype, with their benchmarks and review state, are listed in the model catalog. The archetype itself is a method, not any one of those models.

This explainer describes a method, not a recommendation, and a model built this way is not a guarantee of future results. It is a disciplined way of reading the shape of the options surface, and acting only when that shape leans clearly enough to be worth it.

Sources

Stonewell One model archetype library: the Volatility Surface method family, which reads the level and shape of the options volatility surface rather than the price, centered on the term structure of implied volatility, to infer directional pressure.
Method basis: read the surface's shape (a steep-versus-flat term structure and how it is moving, judged against its own recent history) and collapse it through a fixed rule-based regime read into long, short, or flat, opening a position only when a set of stand-aside filters all clear.
Stonewell One model validation and governance: walk-forward out-of-sample reproduction (a signal-reproduction check, not a profit claim), an acceptance-gate scorecard, and a verification-warning, human-review, deployed-but-provisional posture.