Journal entry ← All journal entries

July 12, 2026

The exam the engine can't grade itself

EVIDENCE HEALTH
Live trades
52
Live trading days
8
Live result
—
Sample status
Too small to prove edge
Record basis
Internal broker records reviewed by desk
Independent audit
No
Session summary

Paper only — 0 rehearsal trade(s), honestly scored, no capital at risk.

What happened
  • Sunday — no live trades. The desk ran its weekly portfolio review and the full nightly research loop.
  • Three research hypotheses went through the gauntlet: a position-sizing scheme, a crisis-hedge sleeve, and a rotation idea from an academic paper. All three failed honestly. Zero promoted; the graveyard grew by three.
  • The strategy-invention engine was caught memorizing: every one of its 25 current 'best' reads worked on exactly one stock out of 774. The wide-universe exam rejected them 25-for-25. Breadth guards were added to the engine the same night and the polluted hall was purged.
  • Every paper trade closed this cycle was interrogated — 93 of them. The questioning surfaced two engine bugs: a price-precision flaw that quietly disarmed stops on low-priced coins, and missing entry-context stamps. Both fixed the same night.
  • The auditor reconciled the ledger to the broker exactly, and caught one discipline violation — a page in the public tree showing more than percentages. Scrubbed same-day.
  • Discovery refresh: a handful of early-attention names queued for Monday catalyst checks; a cluster of coordinated microcap promotions was identified and excluded outright.
Closed live summary
Instrument typeU.S. listed options
Strategy familyMomentum breakout
Executed by—
Number of live trades0
Net return+0.00%
Risk multiple—
Rule followed—
Internal record matched—
Independent auditNo
What we learned
  • An engine that grades its own homework needs an outside exam — internal robustness flags passed 25 reads that the wide-universe gate rejected 25-for-25.
  • A high win rate is not enough for Kelly sizing: when the average win is smaller than the average loss, the formula pins to its floor and underperforms plain fixed sizing.
  • A crisis hedge that waits for confirmation buys protection after the storm — the inverse-ETF sleeve won 2008 but lost 2020 and 2022. The fund's crisis gap stays open, honestly.
What changed after
  • Instrument-breadth requirements added to the invention engines' halls of fame — single-stock artifacts can no longer enter.
  • Paper engine hardened: a minimum-tick stop floor and entry-context stamping on every sleeve, so future trade interrogations can answer 'did any flag predict this?'
  • The public tree shows percentages only — the standing rule, re-enforced in code.

Private ledger note. Full trade-execution detail (contracts, strikes, timestamps, prices, order IDs) remains in the private Ledger. The public journal publishes closed summaries for accountability, not signals. Every live figure above is taken from the broker-reconciled ledger, and sessions the founder traded by hand are labelled as manual rather than folded into the system's record.