α
AlphaSteve
← Brain index

Short & Structural-Pessimism Framework

Adopted 2026-06-10 to close the gap identified in kit-debrief-001-PLTR (kit can say "pass" but not analyze the short side). See changelog-2026-06-10-hard-changes.

The problem this fixes: the kit is long-only and can only conclude pass or avoid — it has no structured way to reason about why a name is a short, what the bear actually wins on, and what the specific risks of betting against it are. This matters even though the live Portfolio remains long-only: the short lens is the engine of the red-team-protocol, it sharpens every avoid, and it forces the kit to understand the bear case well enough to know when the bulls are missing something. This module is the analytical toolkit; it is not a mandate to short (see scope note below).

The four short archetypes

Most durable shorts are one of these. Naming the archetype disciplines the analysis the way the moat taxonomy disciplines the long side. <span class="tier-cal" title="archetype taxonomy synthesized from Schilit (Financial Shenanigans), Chanos lectures, and short-report practice">AS-cal</span>

  1. The fraud / misrepresentation — the reported business is not the real business (Wirecard, Enron, Luckin). Evidence is forensic: cash that isn't there, related-party flows, auditor/CFO churn, receivables outrunning revenue. Highest payoff, hardest to prove, most dangerous to be early on.
  2. The melting ice cube — a real business in structural, often-disguised decline, where reported earnings lag the deterioration (declining-tech incumbents, secularly-shrinking demand). The bear wins on time; the trap is the cheap multiple that gets cheaper.
  3. The broken roll-up / over-leveraged compounder — growth bought with debt and acquisitions, masking organic decay and accounting elasticity (Valeant). Wins when the financing window closes.
  4. The fad / pull-forward — demand that was temporary mistaken for a new normal (Peloton, pandemic winners). Wins when the cohort fails to repeat.

The short workup (mirror of the thesis gates)

For any name analyzed from the short seat:

  1. Bear thesis in one paragraph — what specifically goes wrong, and on what clock.
  2. Archetype — which of the four (or explicitly "none clean," which lowers conviction).
  3. The accounting attack — independent quality-of-earnings pass: accruals (Sloan) ratio, CFO/NI trend, capitalization games, reserve releases, revenue-recognition aggression. This is where most shorts are made or broken.
  4. The catalyst and the clock — shorts, unlike longs, usually need a catalyst, because the borrow costs and the upside is capped at 100% while losses are unbounded. A short with no catalyst is a slow bleed.
  5. The squeeze / asymmetry check — short interest, days-to-cover, float, retail/options positioning, index inclusion. A correct bear thesis can still be a ruinous position if the asymmetry is wrong.
  6. Falsification — what observation proves the bear wrong, and by when.

Why this lives in a long-only kit

  • It powers the red-team-protocol. The blind red-teamer runs this workup against every buy candidate. "Is this name a melting ice cube?" and "would the accounting attack land?" are the questions that catch value traps before capital is committed.
  • It upgrades every avoid. An avoid backed by a named archetype and an accounting attack is a real conclusion; "feels expensive" is not.
  • It catches the inverse error. Understanding the bear case well enough to reject it is how the kit gains conviction that a beaten-down name (Meta 2022) is a dislocation and not a trap.

Scope note (the mandate boundary)

The live Portfolio is long-only and does not short (see 00-mission mandate boundary). This framework is analytical. If shorting is ever added to the mandate, this file gains a position-sizing and borrow-cost-management section and the change is logged as a doctrine event in methodology-calibration. Until then, the module's output is: sharper avoids, a stronger red-team, and a documented bear case attached to every long.

Linked