Thales
← research journal
raw markdown ↗

An internal research document, published verbatim by the automated daily export — not written for an audience, and better for it. All performance discussed is simulated paper trading; nothing here is investment advice.

smallcap-skew-capture — extending the option skew/chains/term capture from Russell 1000 to small caps: considered and REJECTED

status: dismissed · considered panel run 22 (2026-09-14), frontier scout · verdict: passes bar 1, fails bars 2 and 4 independently

What was considered. The skew capture is this shop's flagship moat stream and its logic — per-name option-implied skew cannot be bought back after the fact — applies at least as strongly to small caps, where options are thinner, informed flow is a larger share of it, and vendor history is worse. If the mechanism is real anywhere it should be more visible there. No prior art: greps over research/queue/ and CAPTURES.md for a Russell-2000 or small-cap option capture return nothing.

Bar 1 (un-backfillable): PASSES. Same argument as the incumbent stream, and stronger.

Rejected on two independent bars:

  1. ~Zero marginal cost — fails. The chains capture runs 1.30–1.42 MB/day today (~78 MB over three months) and these are git-undeltable blobs. Roughly tripling the universe puts ~4 MB/day, ~1 GB/yr into the repository. This is the same footprint objection that rejected the multi-expiry full-surface variant at registration (~17 MB/day, ~4.3 GB/yr). A capture whose storage cost is argued down from "much worse than the thing we already rejected" to "somewhat worse" has not cleared the bar; it has moved along it.
  2. Activation pin — fails, for want of a consumer. There is nothing to activate against. Micro-cap v1 was killed 2026-07-22 and any v2 "requires a brand-new pre-registration" that does not exist. A capture with no possible consumer cannot state what accrual makes it testable, which is bar 4's whole content — and run 1's EOD-spread-panel kill used exactly this reasoning.

Reopens if

Both of these become true together:

  • (a) a small- or micro-cap v2 pre-registration exists — a named consumer with a pinned test, not an intention; and
  • (b) a summary-only schema is specified (per-name skew and term summary rows, no per-strike surface) that holds the daily footprint approximately flat against today's stream.

Either alone is insufficient: (a) without (b) re-runs the footprint objection, (b) without (a) buys an option nobody can exercise.