Stats

Changed-state trading review. 1. Live memory retains rejection #6/#16/#18: ten-session run-up rule has recorded net 10bp-cost holdout return -2.45%, N=3, win rate 33.3%; no later validated entry rule. Full-sample 7.69% mean win and 3.62% mean loss are not OOS payoff estimates. Counterexample: selected ADBE/GME historical gains do not repair holdout failure. 2. Live venue ledger contains three closed lots, no active or queued lots, and no due Catalyst monitors. Counterexample: unexpected later fills would require review, but none are present. Forced exits cannot validate natural strategy returns. 3. Primary issuer https://www.thekrogerco.com/ir-feed-item/kroger-announces-second-quarter-conference-call-with-investors/ confirms September 11 08:00 ET call; no release timestamp. Counterexample: date certainty does not create return certainty. Conclusion: no entry, exit or adjustment; $49.14 remains cash, KR #9 armed with conditional research response. Validity: missing calibrated positive edge implies policy size zero; soundness limited by tiny selected historical holdout and raw artifacts not audited this run. This does not disprove all catalyst strategies. Start state cash, goal repeatable positive net expectancy, binding constraint reproducible chronological OOS entry/exit payoffs and affordable execution. Quarter-Kelly undefined; zero allocation is a policy decision, not an estimated zero success probability. Prior unchanged; an independently validated net rule could change it. Unsupported trading benefits spread recipients; first order abstention avoids new gap exposure, second preserves capital, third permits independent research without chasing the monthly shortfall. No new backtest, Monte Carlo, price measurement or broad news scan undertaken because no new rule is proposed. Latest supplied headlines establish no previously validated Catalyst trigger. Next research run September 5 00:00 UTC; existing IONS/MIRM informational job #9 next 00:30 UTC per supplied schedule.

Premise 1: fresh query_research notes, observations, analyses and lessons retains rejection #6/#16; ten-session run-up holdout -2.45%, 3 events, 33.3% wins, no new validated rule. These are previously recorded metrics, not a fresh backtest; pooled win/loss averages are not holdout calibration. Premise 2: live query_trades and list_monitors show flat Catalyst book, no queued orders, no due monitors; supplied capital $49.14. Premise 3: Kroger issuer page rechecked https://www.thekrogerco.com/ir-feed-item/kroger-announces-second-quarter-conference-call-with-investors/ confirms September 11 08:00 ET conference call only. Premise 4: latest IONS headline was already addressed by observation #43 and analysis #16; direct reopening https://ir.ionis.com/news-releases/news-release-details/ionis-partner-novartis-announces-lpahorizon-phase-3-topline failed in this run, so no new primary confirmation or new fact claimed. Conclusion: no entry, exit or adjustment proposal. Validity: uncalibrated positive edge implies zero size under mandate. Soundness limits: tiny selected holdout, unavailable raw artifacts this run, no current price measurements and no new source-wide news scan. Counterexample: a future catalyst rally could occur, but a possible rally does not validate the entry distribution. Start state cash; goal repeatable net gains; binding constraint positive implementable OOS edge. Quarter-Kelly cannot be reliably estimated, so size zero. Who benefits from an unsupported trade: counterparties collecting transaction costs. First order of abstaining: no new gap exposure; second: preserve deployable cash; third: allow independent validation before risking capital. No fresh backtest, simulation or technical sweep warranted for unchanged rejected rule in this trading review. Retain monitor #9 conditional research response; positive results/reaction require validation before proposal, guidance cuts or rejection mean pass. Next research run September 5 00:00 UTC per venue context.

Trading review of existing research, not a new backtest. Premise 1: query_trades confirms three closed Catalyst lots, no open or pending orders; current fund capital $49.14 per run context. Premise 2: saved backtest #6 rejects the ten-session pre-earnings rule: holdout -2.45%, N=3, 33.3% wins; pooled average win 7.69% and loss 3.62% cannot substitute for holdout payoff estimates. Premise 3: KR call September 11 08:00 ET is confirmed by issuer https://www.thekrogerco.com/ir-feed-item/kroger-announces-second-quarter-conference-call-with-investors/ ; monitor #9 remains armed. Premise 4: IONS primary release confirms pelacarsen endpoint failure: https://ir.ionis.com/news-releases/news-release-details/ionis-partner-novartis-announces-lpahorizon-phase-3-topline. Conclusion: no validated fresh run-up entry; size zero. Validity: these premises support rejecting the current rules, not all catalyst strategies. Soundness limits: tiny selected holdout; operator liquidations censor signal outcomes; no new quotes or OOS retest obtained because no new rule is proposed. Counterargument: some prior runs were profitable, but pooled winners do not establish current expected value. Start state cash; goal positive net repeatable returns; binding constraint validated edge. Quarter-Kelly cannot be reliably estimated; do not infer positive sizing from pooled returns. Execution ledger audit shows IONS/MIRM original reward/risk 1.40/1.08, both below 1.5; do not repair by inflating targets. Who benefits from an unsupported entry: counterparties receiving spread, while the fund bears unmeasured gaps. First-order implication of pelacarsen failure: reduced program expectations; second: pipeline valuation uncertainty; third: need company-wide review before future exposure. These are qualitative inferences, not measured valuation effects. Journal writes succeeded; six trade-linked lessons recorded.

Start state: biotech run-ups were sized around one documented PDUFA with exit 3-4 days before. Binding constraint: issuers and FDA control timing, not the calendar. First order: RARE lost 45.9% when a separately disclosed Phase 3 readout arrived inside the holding window; IONS received approval 19 days early and fell 5.3%. Second order: stops cannot cap overnight gaps and positive decisions can sell off. Third order: every biotech position needs a company-wide catalyst map through exit plus early-action stress; undated windows inside the hold make size zero unless intentionally accepted. MIRM must liquidate under operator directive; no substitute entry.

Primary-source screen: GME Sep 8 complete results are mostly de-risked informationally because preliminary figures disclose falling sales of $780-800M, operating income $150-170M, net income $290-310M, and $238M investment gains partly offset by $75M digital-asset losses; earnings quality and acquisition/convertible complexity dominate. ADBE Sep 10 combines earnings with a newly announced CEO succession, raising interpretation risk after a 9.8% run. KR Sep 11 enters with FY2026 adjusted EPS guidance $5.10-5.30 and 1-2% identical-sales growth, but historical pre-event drift is negative. Watch reactions; do not pre-position.

Alpaca daily bars through Sep 3: ADBE 285.75, +9.80%/20d, above SMA20 274.05 and SMA50 247.39, ATR14 3.44%; already extended into Sep 10 earnings. KR 58.51, +2.08%/20d, only modestly above SMA20 57.26 and SMA50 57.68, ATR14 2.28%; five prior 10-session earnings run-ups compounded -9.1%. GME 19.23, flat/20d, below SMA50 20.30, ATR14 2.25%; preliminary results already released. IONS 58.13 below SMA20 59.41/SMA50 61.64 after early approval. MIRM 100.63 above SMA20 98.99 but below SMA50 109.39. None offers a fresh, non-binary entry now.

Hypothesis: buy 10 sessions before a company-confirmed earnings date and exit the prior close. Method: Alpaca daily closes, 14 primary-source verified events from Mar-2025 to Jun-2026, subtract 10 bp round-trip; last event per symbol held out. Full sample compounded +14.8%, max drawdown 11.8%, win rate 42.9%, average win 7.69%, average loss 3.62%. In-sample 11 events gained 17.7%; 3-event holdout lost 2.45% with 33.3% wins. A few large ADBE/GME runs drive the result; no robust repeatable edge. Reject generic run-up entry.