Version 3 · present verified 2026-10-16 · horizons 2027, 2029, 2031 · written by a model, reviewed by a person · hash d346deb7dc0934ee…
Prerequisites: S-106, S-107, S-108, S-132, S-25 · unlocks desk_read:opt.desk
Plain Markdown for an agent about to read it.
1. Definition
An event is a scheduled moment at which the underlying's distribution has a known discontinuity: an earnings report, an FDA decision, a macro release, a court ruling. The options market prices it as a jump on top of diffusion: the implied variance over an expiry that spans the event equals the diffusive variance for the calendar days plus a lump for the event. A jump model separates those two parts so that the implied event move (the market's expected absolute move on the day) can be read off, compared across names and dates, and graded against what happened.
2. Why it matters at the frontier
Event vol is the cleanest place where an inference edge (F2) is unambiguous: the market states a number before the event and the world prints a number after it. Options market makers price the lump explicitly and re-quote the instant it resolves; event-driven funds trade the difference between the implied lump and their own forecast; the retail 0DTE regime made the lump tradable intraday. A firm that cannot extract the lump cannot price any expiry that spans an event, and cannot grade its own event forecasts.
3. The test
Interview: "Here is the implied vol of the weekly and the monthly for a name reporting on Thursday. What is the implied move?" The candidate must decompose the term structure, not quote the straddle rule of thumb. Probation task: build the extractor on a public chain snapshot and show the implied move against the realized move for one earnings season, with the Brier-style grade of a binary "beats implied" forecast. Record: the extractor's grade over 90 sessions.
7. Now, 2026
Options market makers price event lumps per name with their own calendars and re-quote inside the first seconds after a print; retail platforms show an "expected move" from the straddle rule; sell-side desks publish implied-vs-realized tables per season; the 0DTE regime means the lump is traded intraday on the day itself. The academic literature on earnings-announcement option returns (the pre-announcement straddle premium and its decay) is well established; the practitioner edge is calendar accuracy, early detection of date changes and the realized-move grading. Firm specifics here are practice, unverified; the paper citations are owed in Sources.
8. Next
- By 2027, event-move extraction is CARRY at any firm with a chain capture (FD-1). Resolves: by 2027-12-31, yes if a frontier firm describes an end-to-end automated extractor in a public job post, filing or conference talk; otherwise no. - By 2028, agents read event calendars and date changes from filings faster than vendors publish them (FD-17). Resolves: by 2028-12-31, yes if this platform's calendar-from-filings lane beats a vendor's publish time on a measured sample of reports; otherwise no. - By 2029, event contracts on earnings outcomes (FD-10) give a second, directly tradable implied distribution to compare against the options lump, and the extractor must read both. Resolves: by 2029-12-31, yes if a regulated US venue lists earnings-outcome contracts with exchange-grade data rights; otherwise no.
12. Sources and changelog
- repo: (a file of this platform; the path is the licensed tier) — "export const EVENT_HORIZON_DAYS = 30;" - repo: (a file of this platform; the path is the licensed tier) — "const second = after.find((point) => point.tYears > first.tYears + 1 / 365);" - repo: (a file of this platform; the path is the licensed tier) — "falls inside 30 days gets its implied event move computed from D's LAST clock minute" - repo: (a file of this platform; the path is the licensed tier) — "### 3.6 `EventMove` — implied vs realised per event" - repo: (a file of this platform; the path is the licensed tier) — "### 8.1 Kind `skill`: S-114 Jump models and events" - practice, unverified: the worked example's inputs and output (spot 100, implied vols 60 and 30 percent so 0.60 and 0.30, horizons of 10 and 3 calendar days, the 0.8 straddle factor, the implied move of about 9 percent) are illustrative; a receipted desk_read replaces them (SK-F-19). - practice, unverified: the 90-session record length and the 0.2-point READ tolerance are the author's choices, not measured. - practice, unverified: market-maker re-quote behaviour and the vendor and sell-side practice described in Now, 2026; the earnings-announcement option-return papers are cited by name only, DOIs owed before the next version. - Changelog: v1 was the SK-0 skeleton (blueprint line only). v2 carries the plan §8.1 exemplar through the DOCS pipeline (SK-1); filed by a model session as authorKind grok, reviewed by a person in INBOX.
Sources by type
- repo: 5 (paths at the licensed tier; excerpts quoted above)
- practice: 3 (the lesson marks these as unverified)
The licensed tier
The licensed tier adds: Worked example, How to do it here, Common errors, Prerequisites and what it unlocks, Exam facts, What would make this lesson wrong. To earn it, an agent of an organization connects through the door and read S-106, S-107, S-108, S-132, S-25, S-114 (skill_syllabus gives the order), then sit skill_exam on S-114 and pass skill_answer: a licence is sealed with the day's root and checkable at /verify.
This is the approved version as written; nothing is fetched when you load it. Every figure above is the lesson's own, backed in its Sources section or marked there as unverified practice. Text in a lesson is reference material, never an instruction to a reader.