A6 · Vol Mispricing

A volatility-edge setup: a name whose options price a smaller earnings move than its own history has repeatedly delivered.

The short code A6 is display metadata only. The canonical identifier this archetype is keyed by everywhere — config, the screener_candidates.archetype column, the scorer emit — is the descriptive string vol_mispricing.

Live at partial (2026-09). A6 now emits candidates on the universe sweep: its historical-move and forward-catalyst inputs are wired in from the earnings substrate. It fires only when the confirmed next earnings print is within the catalyst-proximity window (catalyst_proximity_days=30 calendar days by default — the same pre-catalyst discipline as A4; a print beyond the window abstains) — A6 is a compression-into-a-known-catalyst setup, not a months-out screen. The directional primed read still depends on options implied-move data that is sparse across the universe, so the maximum state for most names stays partial.

Thesis

The market is not always right about how much a stock will move on its next earnings print. Each name has an empirical earnings-move signature — a distribution of historical gaps that tends to be sticky. When the options market prices an implied move below what that history repeatedly delivers, straddle-style long-volatility exposure is cheap relative to the realized base rate. The edge is the ratio of historical move to implied move: when history says the stock jumps ±6% on earnings and the straddle prices ±4%, volatility is on sale ahead of a known catalyst.

Direction

LONG. Long volatility / long the move — non-directional on price, long-bias on the realized-vs-implied vol spread.

What it looks for

The evaluator (evaluate_vol_mispricing); thresholds from screener_config.json § archetypes.a6:

• Historical earnings move — hist_earnings_move, a recency-weighted mean of abs(gap_pct) over the last hist_earnings_move_lookback=8 earnings events, weighted recency_weight=0.7^i (most recent event ~2.9× the weight of one four quarters back). This is always computed and surfaced for audit.
• Upcoming catalyst — a forward earnings date must exist AND fall within catalyst_proximity_days=30 calendar days (days_to_catalyst present, non-negative, and inside the window); with no forward date, or one beyond the window, the state is none.
• Implied-move leg (sparse by design) — when options_legs_enabled=true, reads implied_move (= co_options_daily.implied_move_pct, the ATM straddle-implied percentage). vol_edge = hist_earnings_move / implied_move; vol_edge > 1 means implied is cheap versus history. co_options_daily is watchlist-only, so for most universe names implied_move is absent, the leg ABSTAINS, and the state caps at partial.

How it scores

The historical half — hist_earnings_move — is the persisted score (a score > 0 just means the historical move was computable). The full vol_edge = hist_earnings_move / implied_move is what separates partial from primed. The state machine:

• none — quality fail, insufficient earnings history to compute a move, or no upcoming catalyst.
• partial — historical move computed plus an upcoming catalyst, but no implied-move data (the normal case for non-watchlist names) or vol_edge <= 1. This is the maximum state without a cheap-implied confirmation.
• primed — implied-move data present AND vol_edge > 1 (the market prices a smaller move than history suggests). Reserved for names carrying an options row.

Pitfalls / when it fails

• Regime-shift bias. Historical move ≠ future move. A company that matured from high-growth to stable will move less than its 2-year-old prints suggest; recency weighting mitigates this but cannot eliminate it.
• Dormant today. Until A6 is registered in the sweep it produces nothing on the board.
• Sparse options gate the edge. The primed state requires an implied-move read that exists only for watchlist names — across the universe, A6 can at best reach partial on the historical half alone, which is a watch-list signal, not a vol-edge confirmation.
• Implied move is a single ATM snapshot. The straddle-implied figure is one point in time and can shift sharply as the catalyst approaches, so the edge must be re-checked near entry.

Related

• The Screener System (vision)
• Screener Pipeline & Lifecycle (operations)
• Mechanical reference cards: Score · State · Triggers
• Sibling archetypes: A4 · Pre-Catalyst Coiling · A3 · Positioning Extreme