Stats

Diagnostic only. Ran supplied backtest CLI momentum20, stop5%, target10%, days730 on IEX; then imported its unchanged engine and strategy for SIP train through2025 and held-out2026 throughSep8, with20 prior bars for warmup and flat warmup trading. Holdout {"fund":"catalyst","split":"holdout","symbol":"MIRM","bars":191,"returnPct":22.16509767375947,"buyAndHoldPct":40.20284547119313,"maxDrawdownPct":15.199708089039069,"winRate":50,"avgTradePct":1.5664072813636993,"N":16}. Inspection of /app/agent/dist/backtest/engine.js shows zero fees/slippage, same-close signal entry and stop fills at the exact stop even if price gaps below it. That stop treatment is contradicted by RARE's actual ledger and 15-minute gap audit17354. Positive gross results cannot be trusted for this mandate, and no parameter was selected for live use. Buy-and-hold includes warmup and must not be used as aligned holdout excess return. Source https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all ; raw and split observations17369–78/17387–91. The independent costed vectorbt calendar tests1481–82, with exposure audit, answer tonight's question; this CLI diagnostic supplies no position size.

Diagnostic only. Ran supplied backtest CLI momentum20, stop5%, target10%, days730 on IEX; then imported its unchanged engine and strategy for SIP train through2025 and held-out2026 throughSep8, with20 prior bars for warmup and flat warmup trading. Holdout {"fund":"catalyst","split":"holdout","symbol":"IONS","bars":191,"returnPct":-14.070477285394468,"buyAndHoldPct":-30.70625610948192,"maxDrawdownPct":18.321309831926,"winRate":31.25,"avgTradePct":-0.8359489928864545,"N":16}. Inspection of /app/agent/dist/backtest/engine.js shows zero fees/slippage, same-close signal entry and stop fills at the exact stop even if price gaps below it. That stop treatment is contradicted by RARE's actual ledger and 15-minute gap audit17354. Positive gross results cannot be trusted for this mandate, and no parameter was selected for live use. Buy-and-hold includes warmup and must not be used as aligned holdout excess return. Source https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all ; raw and split observations17369–78/17387–91. The independent costed vectorbt calendar tests1481–82, with exposure audit, answer tonight's question; this CLI diagnostic supplies no position size.

Diagnostic only. Ran supplied backtest CLI momentum20, stop5%, target10%, days730 on IEX; then imported its unchanged engine and strategy for SIP train through2025 and held-out2026 throughSep8, with20 prior bars for warmup and flat warmup trading. Holdout {"fund":"catalyst","split":"holdout","symbol":"RARE","bars":191,"returnPct":1.8975920520518441,"buyAndHoldPct":-59.17695473251029,"maxDrawdownPct":22.621906250000023,"winRate":33.33333333333333,"avgTradePct":0.3544089355959463,"N":15}. Inspection of /app/agent/dist/backtest/engine.js shows zero fees/slippage, same-close signal entry and stop fills at the exact stop even if price gaps below it. That stop treatment is contradicted by RARE's actual ledger and 15-minute gap audit17354. Positive gross results cannot be trusted for this mandate, and no parameter was selected for live use. Buy-and-hold includes warmup and must not be used as aligned holdout excess return. Source https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all ; raw and split observations17369–78/17387–91. The independent costed vectorbt calendar tests1481–82, with exposure audit, answer tonight's question; this CLI diagnostic supplies no position size.

Premises1 completedSep8 SIPmetrics recorded17273–81: RARE14.88 belowSMA20/50=24.4385/27.3929, RSI20.619,ATR1.949; IONS56.71 below59.478/60.7466; MIRM99.53 above99.042 butbelow108.5797; TLX11.98 above11.69075/11.2133; SRRK57.42 above55.8765/53.4903. Mixedregime, no fittedcombination. 2 TLXADR20=.218205 andATR14=.388844; 1–2ADR targetdistance .218–.436USD describes ordinarytape, not binarygap. Fund49.14 permits atmost4TLX or3RAREshares atlastclose beforecosts; otherthree0whole. 3 Fifteen-minute SIPsource https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=15Min&feed=sip&adjustment=all plusissuerHTML times17288–89/17182. IONSJun24 precompletedbarclose76.995,next18:45UTCbaropen77.15(+.2013%); RAREAug19pre26.76,next21:30bar27(+.8969%),nextregularopen28.45(+6.3154%). RARESep2pre19:45barclose26.55,next20:15baropen14.9915(-43.53484%),nextregularopen13.97(-47.3823%). Quotes/fills not observed, interveningnewsbar excluded; no causalprecision beyond timing association. LedgeractualRAREentry25.84exit13.98976 loss45.86%/6.44R. Counterexample to relyingonwhole-sharestop: afterhours eligibility cannot guarantee tradableprice24 after gap. No stop/targetproposed withoutvalidatededge.

Preregistered note1339, implementation1347 before priceoutcomes: T10_T2 means nthSPYtrading session strictlybefore issuergoal; buyopen/sellclose, goals onweekends retained. Ten selected completed decisions acrosssixissuers, not completeFDAdenominator. MIRMchenodiol2025,RAREUX1112025 andNUVL2026 resolvedbeforeentry andskipped (notpostdecisiontrades). Remaining2025N5 includes3TLX,IONS,SRRK;2026N2 IONSolezarsen/RAREDTX401. Correct advance notices verified, includingFeb23DTX401. vectorbt separateevent sleeves with equal initialcash, portfolio=sumvalues withidlecapital; overlappingevents not falselycompounded. Fees10bps+slippage10bpseachside, stressslippage30bps. Independentclosedform reconciliationpassed all28event/costoutputs. Source https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all , pairedprimary observations17167–88 withcorrections1347/17185–87, eventrows17320–47, aggregates17361–68. Train {"rule":"T10_T2","year":2025,"slippage":0.001,"N":5,"return_pct":6.537472416458501,"drawdown_pct":2.072133225198769,"win_rate_pct":80,"avg_win_pct":8.921812701239578,"avg_loss_pct":2.999888722665911,"mean_pct":6.53747241645848,"ci95":[0.7996589261197484,13.281418340512976],"inside":0,"exit_day_ambiguous":1} Holdout {"rule":"T10_T2","year":2026,"slippage":0.001,"N":2,"return_pct":3.677905521332936,"drawdown_pct":3.044835296980164,"win_rate_pct":50,"avg_win_pct":10.023035062100005,"avg_loss_pct":2.6672240194341112,"mean_pct":3.677905521332947,"ci95":[-2.6672240194341112,10.023035062100005],"inside":2,"exit_day_ambiguous":0} Stressholdout {"rule":"T10_T2","year":2026,"slippage":0.003,"N":2,"return_pct":3.264020427779446,"drawdown_pct":3.1391809797858605,"win_rate_pct":50,"avg_win_pct":9.583820034262192,"avg_loss_pct":3.0557791787033217,"mean_pct":3.2640204277794354,"ci95":[-3.0557791787033217,9.583820034262192],"inside":2,"exit_day_ambiguous":0}. Both holdoutwindows crossedactualearlyapproval: IONSJun24issuer14:37:29ETbeforeJun26exit;RAREAug19issuer17:22:01ETbeforeAug20exit. Positive3.6779% return therefore is notpurepre-newsedge;N2bootstrapCI[-2.6672,10.0230] spanszero. Gozellixtraining exitMar20sameUSdate; SECaccepted17:34:54ET, butearliestFDA/ASXnewsminuteunresolved, retainambiguous. No intraday stops/targets simulated, fees assumed not venueinvoice, dailyopen/close executionnot guaranteed. This is diagnostic calendar exposure audit, not trade-ready strategy. No retuning based onholdout; operationalallocation0.

Preregistered note1339, implementation1347 before priceoutcomes: T20_T6 means nthSPYtrading session strictlybefore issuergoal; buyopen/sellclose, goals onweekends retained. Ten selected completed decisions acrosssixissuers, not completeFDAdenominator. MIRMchenodiol2025,RAREUX1112025 andNUVL2026 resolvedbeforeentry andskipped (notpostdecisiontrades). Remaining2025N5 includes3TLX,IONS,SRRK;2026N2 IONSolezarsen/RAREDTX401. Correct advance notices verified, includingFeb23DTX401. vectorbt separateevent sleeves with equal initialcash, portfolio=sumvalues withidlecapital; overlappingevents not falselycompounded. Fees10bps+slippage10bpseachside, stressslippage30bps. Independentclosedform reconciliationpassed all28event/costoutputs. Source https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all , pairedprimary observations17167–88 withcorrections1347/17185–87, eventrows17320–47, aggregates17361–68. Train {"rule":"T20_T6","year":2025,"slippage":0.001,"N":5,"return_pct":-8.508057833350623,"drawdown_pct":9.998850991777541,"win_rate_pct":20,"avg_win_pct":1.3646854391706098,"avg_loss_pct":10.976243651480925,"mean_pct":-8.508057833350618,"ci95":[-13.845153808532212,-2.4983876640653557],"inside":0,"exit_day_ambiguous":0} Holdout {"rule":"T20_T6","year":2026,"slippage":0.001,"N":2,"return_pct":-1.8555954687694776,"drawdown_pct":4.639660087072195,"win_rate_pct":0,"avg_win_pct":null,"avg_loss_pct":1.8555954687694831,"mean_pct":-1.8555954687694831,"ci95":[-3.6015652080989558,-0.10962572944001048],"inside":0,"exit_day_ambiguous":0} Stressholdout {"rule":"T20_T6","year":2026,"slippage":0.003,"N":2,"return_pct":-2.2473906718738057,"drawdown_pct":4.929601298264319,"win_rate_pct":0,"avg_win_pct":null,"avg_loss_pct":2.247390671873811,"mean_pct":-2.247390671873811,"ci95":[-3.9863904512559833,-0.5083908924916392],"inside":0,"exit_day_ambiguous":0}. All7testedwindows finishbefore known named FDA announcements. This does NOT establish absenceofotherpipelineevents or futurehazard; negativecostedreturns rejecttrade rule. No intraday stops/targets simulated, fees assumed not venueinvoice, dailyopen/close executionnot guaranteed. This is diagnostic calendar exposure audit, not trade-ready strategy. No retuning based onholdout; operationalallocation0.

Premises from paired primaryobservations17167–17188 and FDA https://www.fda.gov/media/151712/download : performancegoals specify within periods, not waitinguntildeadline. ChainA: earlyapproval/CRL -> calendar entry may alreadybe postevent -> uncorrected calendars contaminate measured runup expectancy and position selection. Examples RAREUX11138daysearly https://www.sec.gov/Archives/edgar/data/1515673/000095017025095334/rare-20250711.htm ; GSKzidesamtinib58days https://www.gsk.com/en-gb/media/press-releases/jideytro-zidesamtinib-approved-in-the-us-for-previously-treated-ros1-positive-non-small-cell-lung-cancer/ . ChainB: sharedfillfinish deficiency -> sponsor launchcash delayed and remediation spending -> alternatecapacity demand may rise, but capacityowners capture rents only if qualified/contracted capacity and prices are documented. SRRKissuerSep23CRL andTLXZircaixSEC sources in fundamental1410-equivalent current record; no supplier stockrecommendation. ChainC: otherpipeline failure -> equityNPV reprices despite separateFDAgoal -> capitalraises/expensecuts can alter future development and counterparties' economics. RARESep2Aspire primary https://ir.ultragenyx.com/news-releases/news-release-details/ultragenyx-announces-phase-3-aspire-results-angelman-syndrome . Who benefits: patients from fasteraccess, sponsors from positive approvals; cash-rich buyers may gain bargainingpower after failures. These are mechanisms, not measured competitor revenue. Counterexample: fastpositive news can reward inadvertent holdingthrough; favorable realizedreturns do not prove safeexecution. No new book justified without quantifiedindependentedge.

Premises1 primaryAug4 RARE Q2 https://www.sec.gov/Archives/edgar/data/1515673/000119312526332802/rare-ex99_1.htm lists UX111Sep19goal and GTX102AspireSep/Oct readout concurrently; the single named FDA date cannot bound other-program first news. 2 Q2 revenue214m versus167m prior-year (+28.1437%), Crysvita156m/214=72.8972% revenue concentration, cash/investments436m andquarterOCFburn97m. Static burn436/97=4.4948quarters/13.4845months is arithmetic, not forecast: launch cost, voucher monetization, financing and planned expensecuts change it. Netloss92m is not same as cashburn97m. Existingcommercialcash does not prevent pipelinevaluation gaps. 3 SponsorSep2Aspire https://ir.ultragenyx.com/news-releases/news-release-details/ultragenyx-announces-phase-3-aspire-results-angelman-syndrome reports primary/keysecondary failure, metadata16:01:36ET. 4 SRRKSep23 2025CRL https://investors.scholarrock.com/news-releases/news-release-details/fda-issues-complete-response-letter-crl-apitegromab-treatment/ identifies thirdpartyCatalentIndiana fillfinish observations, no other approvability issue per sponsor; TLXZircaixAug28 https://www.sec.gov/Archives/edgar/data/2007191/000200719125000065/zircprcrl.htm cites comparability plus two thirdparty manufacturing/supplychain partners. These illustrate common network risk, not clinical approval odds or independent trials. Valid inference: review the whole pipeline and shared suppliers before exposure. Strongest counterexample: multiproduct commercial revenue can cushion losses; RAREactualgap shows cushioning not guaranteed. Currentmarketvaluation and incremental productNPV are not estimated; no price target or probability derives from salesgrowth.

Start state: Catalyst49.14cash, account850.99, no positions or queued orders in provided context and own queried ledger. Goal: repeatable costed run-ups exited before dated events. Binding constraint: independently credible actual-window payoff plus executable risk. Premises1 coverage audit1391 screened12647USsymbols and232JPXSeptember event rows, with explicit384syntax/124missing-symbol andUScalendartransport gaps. 2 DRI1376 fixed T5/T1 holdout+5.6622%, N6,p.5,W3.21677%,L1.31445%; meanCI[-.8285,2.9660]% still crosseszero andtrainingnoticesincomplete. Positive result meritswatch, not a forced negative conclusion. EmpiricalquarterKelly.073922 operationally0; currentwindowstartsSep17. 3 MRK regulatory p/W/L unknown;PPIH/WLTH have no remaining validated multi-sessionwindow;foreignvenue/FX/lots/edgeunverified. 4 Fundcap49.14,venuepositioncap212.7475,nominal1%equityrisk8.5099 cannotrepairfailededge; wholeDRIshare212.92unaffordable tofund. Fractional eligibilitywouldnotrepairuncertainty. Strongest counterargument: positive net DRI holdout may represent real edge; preregistered future sample, not retuning, can change allocation. NoPROPOSE/EXIT/ADJUST/FUND. Frozen earliernegativeAVOruleunchanged; no genericMRKmomentumtransport. Sources in1372/1376/1381–85/1391 and https://investor.darden.com/news/news-details/2026/Darden-Restaurants-to-Release-Fiscal-2027-First-Quarter-Results-on-September-24-2026/default.aspx . Newmonitorrejected20cap; standalonecollector rejected5jobcap. Testedcoverage-preserving replacement118 nowactive, replaces52 at240min/96runs;four other jobs unchanged. Its first sandbox result pending atcreation; local11panels passed. Static reviews explicitlyunarmed, noautomaticorders; noownrequiredtradeactionbefore04UTC. Researchendsaround02:19UTC(~18min), under43planned/45hard;12mindiscoverycutenforced andfinalrecordingreserveused. Named gapsremain, not guessed.

Premises1 NasdaqTrader Sep8 21:31 directories https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt and https://www.nasdaqtrader.com/dynamic/SymDir/otherlisted.txt yield13156 non-test rows/13155unique. Unsupportedpreferred-symbolsyntax384 excluded after initial400errors;32 repaired batches cover12771requested/12647returned,124missing. Source https://runway.adityaperswal.workers.dev/data/stock-snapshots?feed=iex . Price1–49.14USD,IEXdollars>=100k,positivepriorclose screen gives959 includingETFs,not959 eligible catalysts. 2 Nasdaq16 calendar timeouts and14Yahoo dateURLs returningidentical25Sep9rows mean broad future earnings intersection remains incomplete; do not count repeated pages ascoverage. Directissuerrecovery confirmsDRISep24,PPIHSep9BMO,WLTHSep9AMC;NUVLJuly22FDAapproval supersedesstaleSep18goal. 3 JPX https://www.jpx.co.jp/listing/event-schedules/financial-announcement/index.html July/Augquarter spreadsheets contain656datedrows,232Sep9–30events;ASKULSep15 issuerconfirmed,KGFSep22provisional,existingNXT/ITXwatches. Notanexhaustiveworldexchange census;foreignvenue/FX/lotandcostededgeunverified. 4 BTC/ETH/SOL39completeddailybars each,S ep9partialexcluded;5dayreturns-3.471/-.944/-.585%;noearnings/FDA linkage. ETHSepoliaOct6laterdeveloperdecision https://ethereum-magicians.org/t/all-core-devs-consensus-acdc-186-september-3-2026/29496 supersedesAug6overviewSep28;mainnetunconfirmed. Rejectscopecreepintoanuntestedupgradebook. Topselectedprimarysourcesandfinancialrejectionsrecorded1372–83. Allscansdescriptive,notbacktestedorqualifiedentrysignals. Discoverycutat12minasplanned;calendarcoveragegapretainedratherthanborrowingfinishreserve.

Premises1 fixed DRI holdout N6 has cash-basis p=.5, W3.216773%, L1.314447%, mean0.951163%; mean95%bootstrap CI[-.828497,2.966011]%. Bootstrap82.19% of sample means are positive is not posterior trade-success probability. 2 OLS DRI costed event return on simultaneous SPY gross return: intercept0.161299%, beta-1.272867, correlation-.730141; conventional small-sample t95% intercept[-2.35749,2.68008] and beta[-2.92652,.380785]. NetDRI minusgrossSPY mean1.571702%, but this selected six-event regression cannot establish stable alpha/hedge. 3 Standard binary-payoff quarterKelly=.25*(p-(1-p)/(W/L))=.073922; conservative uncertainty gate sets operational fraction0. 4 Queried own ledger three losses, mean-16.3806%, two operator liquidations censor planned horizons; no observed winner to calibrate average win. RARE gap loss45.86%,6.44listed-stopR demonstrates stops are not hard loss bounds. Source event observations16251–74, https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=DRI,SPY&timeframe=1Day&feed=sip&adjustment=all and runway:query_trades?fund=catalyst . Validity: positive point estimate increases research priority; soundness limited by selected issuer, N6, incomplete training notice history and untested execution. Evidence to change allocation: independent future events supporting positive costed actual-window expectancy with adequate uncertainty and executable risk plan. Monthly shortfall is not an input.

Source costed cash-return observations16251–56 from original DRI SEC releases and https://runway.adityperswal.workers.dev/data/stock-bars (URL spelling correction: https://runway.adityaperswal.workers.dev/data/stock-bars ). NumPy seed20260909 resamples six holdout whole-event returns with replacement into50000 ten-event paths, starting49.14 and hypothetically fully reinvesting at each event. Terminal p5/p50/p95=47.6642/53.8006/61.1115USD, terminal-loss frequency11.128%, p95 maximum drawdown7.4522%. Separate mean bootstrap CI[-0.8285,2.9660]% includes zero. This is conditional on the six observed outcomes being representative and independent; it cannot generate missing large gaps, regime shifts, taxes or actual execution failures. No actual capital allocated. Strongest counterexample is unobserved early news: prior RARE realized loss45.86% and6.44 listed-stop R, which this earnings bootstrap never samples. Do not interpret11.128% as calibrated live-loss probability or7.45% as worst loss. Simulation informs a research watch, not authorization; quarterKelly estimate0.073922 is overridden by evidence gates to0.

Premises1 primary date required: DRI Sep24BMO https://investor.darden.com/news/news-details/2026/Darden-Restaurants-to-Release-Fiscal-2027-First-Quarter-Results-on-September-24-2026/default.aspx ; PPIH Sep9BMO https://investors.permapipe.com/news/news-details/2026/Perma-Pipe-to-Report-Second-Quarter-2026-Results-and-Host-Investor-Conference-Call-on-Wednesday-September-9-2026/default.aspx ; WLTH Sep9AMC https://ir.wealthfront.com/node/7786/pdf ; MRK Sep21 goal https://www.merck.com/wp-content/uploads/sites/124/2026/08/MRK-06.30.2026-10Q-FINAL.pdf . 2 DRI fixed T5 entry startsSep17, so early purchase changes the tested window. Its close below20SMA/above50SMA is not a trained combined filter. 3 PPIH and WLTH pass affordability/positive prior-close screening but lack a remaining tested multi-session window; IEX price/volume are partial and closing quotes stale. 4 DRI holdout uncertainty crosses zero, training advance notices incomplete; MRK has no actual-window payoff sample. 5 Whole-share and fund/venue/risk limits apply only after edge qualification. Conclusion zero combined qualified candidates; a date or positive tape alone cannot substitute for all gates. Strongest counterexample: affordable one-session rallies occur; require a separate preregistered net rule, not inherited five-session odds.

Source https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=DRI&timeframe=1Day&feed=sip&adjustment=all . Completed Sep8 close212.92 is below SMA20 219.5185 but above SMA50 209.3456; five-session return-0.6254%. ATR14 5.4490 and ADR20 5.4510, recent low210.86/high229.7625, RV20 30.8874% annualized. Premises: daily trend is mixed, the fixed event entry is Sep17 rather than today, and one share exceeds the49.14 fund cap. A one-to-two-ADR upside reference is about5.45–10.90USD, not a proposed target; it cannot by itself validate a stop or expected return. Counterexample recovery above short average could precede earnings strength, but adding that filter now would retune the registered rule. No entry, exit or stop adjustment; book flat. IEX screen data elsewhere are a different feed from these SIP adjusted bars. Daily history cannot establish15minute execution.

Ran node /app/agent/dist/backtest.js --symbols DRI,MRK --strategy momentum --lookback5 --days1460 --stop-pct3 --target-pct5. 1001 daily bars and 133 trades. Source https://runway.adityaperswal.workers.dev/data/stock-bars . This CLI diagnostic has no chronological holdout or documented fees/slippage in output; it holds across events and uses daily stop/target assumptions. It is not evidence of an executable five-session earnings or pre-PDUFA edge. DRI actual-window vectorbt test recorded separately with costs and holdout; MRK regulatory rule has no calibrated sample. Positive MRK headline return cannot be transported to Sep21 HYPERION. Reject adoption of both generic rules. Command/output stored cli_dri_mrk.json.

Ran node /app/agent/dist/backtest.js --symbols DRI,MRK --strategy momentum --lookback5 --days1460 --stop-pct3 --target-pct5. 1001 daily bars and 127 trades. Source https://runway.adityaperswal.workers.dev/data/stock-bars . This CLI diagnostic has no chronological holdout or documented fees/slippage in output; it holds across events and uses daily stop/target assumptions. It is not evidence of an executable five-session earnings or pre-PDUFA edge. DRI actual-window vectorbt test recorded separately with costs and holdout; MRK regulatory rule has no calibrated sample. Positive MRK headline return cannot be transported to Sep21 HYPERION. Reject adoption of both generic rules. Command/output stored cli_dri_mrk.json.

Registered note1294 before price inspection: enter T-5 open and exit T-1 close; no event exposure. Primary release dates from 18 SEC exhibits, observations16251–74; six holdout advance notices verified 28–29 days before release, URLs in note1299. Training2022–24 N12; untouched2025–June2026 holdout N6. Used vectorbt with 10bp fees and 10bp slippage per side, adjusted SIP daily OHLC, 1196 bars. Source https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=DRI,SPY&timeframe=1Day&start=2021-12-01T00:00:00Z&end=2026-09-09T00:00:00Z&feed=sip&adjustment=all . Training net return1.4581%, DD6.6115%; holdout net5.6622%, DD4.9869%, p50%, cash-basis mean0.951163%, average win3.216773%, loss1.314447%. Stress25bp slippage plus10bp fee per side still +3.7773%, DD5.1290%, mean0.648762%. Six training sensitivities T3/5/10 and exitsT1/2 recorded; range returns -4.3463% to+9.3334%, not used to choose holdout parameters. Bootstrap50000 seed20260909 mean95%CI[-0.8285,2.9660]%, includes zero. Historical training notice vintages incomplete and total18 below desired20; adjusted bars do not establish actual auction or15minute stop fills. Positive evidence merits prospective watch, not adoption. Strongest counterargument is real positive costed holdout; it changes research priority but not size until uncertainty and execution gates pass. Current primary notice https://investor.darden.com/news/news-details/2026/Darden-Restaurants-to-Release-Fiscal-2027-First-Quarter-Results-on-September-24-2026/default.aspx ; paper entrySep17open/exitSep23close, not today. Local reproducible dri_test.py/dri_test.json. Trade payoff statistics use net cash invested including entry fees; vectorbt's trade-return denominator differs slightly, portfolio return unchanged.

DRI https://www.sec.gov/Archives/edgar/data/940944/000094094426000016/exhibit991-q4fy26.htm :1 traffic/mix andextraweekaffectheadlinegrowth;2 food/laborand875mplannedcapexcompeteforcash;3 cashafterinvestmentversusdividend/buybacksdeterminesfinancingneeds. Beneficiariescustomers,workers,suppliers,repurchasingshareholders;countercasepositivecompsandcashgrowthfundsustainableexpansion. MRK https://www.merck.com/wp-content/uploads/sites/124/2026/08/MRK-06.30.2026-10Q-FINAL.pdf :1 broaderHYPERIONlabelmayexpandeligiblepatient/treatmenttiming;2 physicians/payers/adverse-eventmonitoringdetermineactualuptakewhileBMSreceives22%net-salesroyalty;3 retainedcashnetofcommercial/researchcostsdeterminesMRKvalue. Countercase label adds little afterpriorZENITHapproval orpayerslimituse. No probabilityestimatedforapproval/stockorBMYbenefit;royaltyisarithmeticnottradeablearbitrage. KGF https://www.kingfisher.com/en/investors :1 housing/DIYdemandchangesvolumes;2 promotionsandfinancecostoffsetsales;3 FCF/workingcapitaldecidesbuybackcapacity;countercaseimprovedmixsustainsearnings. ASKUL https://www.askul.co.jp/news/INFO/ :1 resolvedloginissuesmayrestoreorders;2 repeatpurchasingandfulfillmentcostdecideactualrecovery;3 retainedcashandtrustdeterminevalue;countercaseresolvedbugsimmaterialtothequarter. Allconditionalhypotheses,notmeasuredpeerprofits. No separateFUND justified withoutcostedpayoffevidence.

Premises1 DRI Aug27 notice confirms Sep24BMO, not estimatedSep17: https://investor.darden.com/news/news-details/2026/Darden-Restaurants-to-Release-Fiscal-2027-First-Quarter-Results-on-September-24-2026/default.aspx . June25 SEC https://www.sec.gov/Archives/edgar/data/940944/000094094426000016/exhibit991-q4fy26.htm reports53-weekFY;Q4growth13.7% includesextraweek7.6%,OGcomps2.4vsLongHorn9.5. SimpleOCF-propertycapex-software/other1096.2m grows5.414%,but1364.7m shareholderreturns exceed268.5m;netshorttermdebt194m. FY27EPS11.10–11.35,capex875m. Counterexample cashgrowthandcompsarepositive;return/cashgapdoesnotproveinsolvency, andnext52-weekcomparisonmustbealigned. 2 MRK June30filing https://www.merck.com/wp-content/uploads/sites/124/2026/08/MRK-06.30.2026-10Q-FINAL.pdf confirmsSep21HYPERIONlabelreview. Q2Winrevair588m,3.5407%company16607m;22%BMSroyaltyimplies129.36m arithmetic,notcashreceived,remaining458.64mnotprofit. ExistingZENITHapprovaldistinct;WeliregOct4competingpipeline. Counterexample label expansioncouldaccelerateearliertreatment, butsalesgrowth75%isnotapprovalorstockwinprobability. 3 KGF Sep22explicitlyprovisional https://www.kingfisher.com/en/investors ,priorFYguidePBT565–625/FCF450–510mGBP notverifiedunchangedtoday. ASKUL Sep15dateonly https://www.askul.co.jp/corp/investor/calendar/ ;customerSep8resolvedloginnotice https://www.askul.co.jp/news/INFO/ doesnotquantifychurn. Validity: cash/royalty/datefactsinformwatchlists,nottradingedge. Soundness:primaryissuerincentivizedframing,unreadupdatedKGFQ1andforeignvenuecostgaps. Start49.14cash;bindingconstraintcostedactualwindowuncertainty. These improve candidate diversity but do not authorize allocation.

Start: Catalyst $49.14 cash, account equity $850.99, zero open/pending/closing positions. Goal: repeatable costed pre-event gains. Binding constraint: independently validated expectancy for the actual entry and exit window. Premises: (1) Ledger and coverage audit16011 confirms no live thesis to reset and no own monitor due before next02UTC run. (2) Analyses1263–1266 recover BRZE/GME primary tables, attack earnings quality and Telix pipeline economics, and reject an unverified FedEx September calendar date. Primary sources https://www.sec.gov/Archives/edgar/data/1676238/000167623826000039/a20260731-brazeincxq227ear.htm ; https://www.sec.gov/Archives/edgar/data/1326380/000132638026000050/a991-fy26q2earningsrelease.htm ; https://telixpharma.com/news-views/fda-accepts-nda-for-tlx101-px-pixclara/ ; https://www.sec.gov/Archives/edgar/data/1048911/000104891126000108/fedexrecastexhibit991.htm . (3) Saved AVO expectancy -1.212595% and TLX uncalibrated event sample imply zero quarter-Kelly or undefined positive edge. Strong issuer results cannot substitute for a costed post-event study. Conclusion: no PROPOSE, EXIT, ADJUST or FUND. Size0 before $49.14 fund cap, $212.7475 venue cap and $8.5099 risk limit. Validity: necessary edge gate fails; soundness limited to researched universe, sparse event samples, incomplete historical notices and 15-minute execution. Strongest counterargument is opportunity cost of cash; it does not supply missing p/win/loss figures. No targets or stops fabricated for nonexistent positions. Retain20 armed monitors and5 collectors:48 scheduled00:15,67 00:30,52 01:30 before next02UTC; actual delivery must be checked next run. Monitors15/17 retained for next-day review and organic/call context. Existing static AVO/LOVE/SFIX/GIS and conference overflow remain explicitly unarmed. No new dated tradeable catalyst or signal requires another job. Remaining gaps: BRZE quantified organic bridge/replay, valuation and independent post-event fills; FDX exact next results date, global search not exhaustive. Journal phase began00:04:46, within43-minute plan/45-minute hard budget. Approximate retrieval timestamp errors corrected in16019; tool clock controls elapsed time.

Premises: (1) Queried Catalyst trade history contains three losses: MIRM -0.439321%, IONS -2.842420%, RARE -45.860062%; arithmetic mean -16.380601%. Two exits were operator liquidations, so this is a descriptive selected sample, not a causal strategy estimate. No observed winner means average winning payoff is undefined. (2) RARE lost 6.440348 times the distance from entry to its listed stop, demonstrating that this stop did not bound realized loss. (3) Frozen AVO vectorbt backtest1158, with 10bp fees and 10bp slippage per side, has chronological holdout N6, p=.5, average win1.318443%, loss3.743633%, expectancy -1.212595%, net return -7.331815%, drawdown9.371961%; all16 variants lose. Python recomputation clips quarter-Kelly to zero. (4) TLX event audit500 has zero valid training events and one positive holdout; momentum503 fails training and holds through events, so its positive holdout does not estimate pre-PDUFA returns. Sources: runway:query_trades?fund=catalyst ; runway:query_research?fund=catalyst&kind=analyses&analysisKind=backtest ; https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=AVO&timeframe=1Day&feed=sip&adjustment=all . Valid inference: no calibrated positive edge supports size. Counterexample: future run-ups may win; these small selected samples do not prove universal inefficiency is absent. No new backtest, technical signal, Monte Carlo or fresh price collection is claimed. Existing failed methods remain frozen rather than repeatedly retuned. Observation16039 preserves calculations.

Noopenholdings;theseareconditionalwatchtheses,notpositions. BRZE source https://www.sec.gov/Archives/edgar/data/1676238/000167623826000039/a20260731-brazeincxq227ear.htm :1services/AIadoptionincreasesimplementationwork;2cloud/data/implementationprovidersmaycollectspendwhileBRZEgrossmixdilutes;3renewalsandnetdilutiondecideshareholdercash. Reverse:successfulautomationimprovesretentionandscaleshighmarginrecurringrevenue. GME https://www.sec.gov/Archives/edgar/data/1326380/000132638026000050/a991-fy26q2earningsrelease.htm :1collectiblesmixsupportsretailGP;2popularitycycle/suppliersandtax/receivablesaffectconversion;3eBayinvestmentconcentrationcanamplifycapitalvolatilityindependentofstoreprofits.Reverse:H1cashgrowthanddebtexchangeimproveresilience. TLX https://telixpharmaceuticals.gcs-web.com/news-releases/news-release-details/telix-2026-half-year-results-strong-commercial-execution-and :1regulatoryoutcomeaffectslaunch/CMCwork;2distributioncapacityrequiresinvestmentbeforeutilization;3pipelinecashreturnsversusfinancing/dilutiondecideretainedvalue.Reverse:commercialfranchiseabsorbsinvestmentandnewapprovalsfillcapacity. AVO priorobs15974/source https://investors.missionproduce.com/news-releases/news-release-details/mission-producer-announces-fiscal-2026-third-quarter-financial :1acquiredvolume/unitpricedown;2growermarginsandworkingcapitalpressured;3integration/seasonalitydecidedebtcapacity;countercaseforecastsynergiesrealize. CASY priorobs15975/source https://investor.caseys.com/news-releases/news-release-details/caseys-announces-first-quarter-results-3 :1fuelspreadnormalizes;2GPfundingforstoresfalls;3capexreturn/insidegrowthmustoffset;countercaseinsidegrowthandpersistentspreadsustainprofit. Sourcesissuerorigin,notindependentcausalidentification. Ratesandtotalsobs15974–75/16012–16;nopeerprofit/stockprobabilitiesestimated,nonewFUNDjustified.

Premises1 TLXAug20H1 https://telixpharmaceuticals.gcs-web.com/news-releases/news-release-details/telix-2026-half-year-results-strong-commercial-execution-and confirmsSep11PixclaraPDUFA;April10origin https://telixpharma.com/news-views/fda-accepts-nda-for-tlx101-px-pixclara/ excludesPixclarafromFYrevenueguide.2 H1EBITDA52mincludes40mupfront;subtractinggives12mARITHMETICsensitivity,notnormalizedestimate.TMSloss23mvs13m,R&D124mvs82m;cash252m/OCF23mpositive. Bearcaseupfrontandcommercialgrowthmaskinvestmentspending;anotherpipelineCMCsetback canoverwhelmPixclararunup. Countercaseestablishedimagingfranchiseandfundingcanabsorbresearch,approvalmayunlockfutureeconomics;noapprovedlabel/salesforecastinferred.3 FDX https://www.sec.gov/Archives/edgar/data/1048911/000104891126000108/fedexrecastexhibit991.htm Jun1FYchangesfromMay31toDec31/sevenmonthtransition,Freightdiscontinuedandsegmentsrecast;thirdpartySep17estimatebasedonoldcalendarfailsprimarydategate. IRrender https://investors.fedex.com/news-and-events/upcoming-events/default.aspx blankeventsnotproofabsence;nonewexactdateadopted.4 OXM https://investor.oxfordinc.com/news-releases/news-release-details/oxford-owner-tommy-bahama-lilly-pulitzer-and-johnny-was-23 andLE https://landsendinc.gcs-web.com/news-releases/news-release-details/lands-end-announces-second-quarter-fiscal-2026-results alreadySep3;runupsclosed. UKJohnLewissearchstale2025half-year,WGO/ACNnodateverified. Validitydatesandcashsourceseparatefromexpectancy;soundnessselectedsearchnotexhaustiveglobalcoverage. ExistingTLXtest500train0/holdout1 cannotestimateedge;zeroallocation. No newregulatorydecisionfoundinscopedsearch,notexhaustiveabsence;monitor22andcollector52retained.

Premises1 Sep8issuerfulltables https://www.sec.gov/Archives/edgar/data/1326380/000132638026000050/a991-fy26q2earningsrelease.htm recovered;obs16014–15/timing16019.2 OP160.2mwithinpreliminary150–170m;FYEBITDAfloor>650mraisedfrom>600m.3 Q2FCF60.7mvs113.3m,whileH1393.6mvs302.9m.4 Pretax420.2mlessOP160.2m=260mnonoperating,43.4meBaysharesworth4.9469bn;quartercashpaid4.3863bnforequity. Strongestbearcase collectiblepopularitymayfade whileinvestmentconcentrationandworkingcapital/taxcashusestresscashconversion;headlineNIoverstatesretailincome. Counterexample retailOPmorethandoubled,inventory439mvs484.9m,annualcashimproving andsubsequentdebtexchange;onequartercashdeclineisnotsecularcollapse. Validityreadretailandinvestmentsseparately; soundnessunauditedissuer,notthirdpartyconfirmation. Quarterversushalf-yearcashcontradictionresolvedbyperiodandtiming;Aug1debt4.1678bndiffersSep3postexchange~2.8bn. NoBTCsalesquantityfrom75mmarkor294.1masset/receivableaggregate. Who benefitsissueroptimisticEBITDAframing,collectiblesuppliersandconcentratedequityholders;nocrosscompanypayoffinferred. Noindependentposteventedge,price/valuationnotmeasured,cannotcallcheap. Retainmonitor15fornextdaycontext,zeroentry.

Premises1 Sep8issuerSECexhibit https://www.sec.gov/Archives/edgar/data/1676238/000167623826000039/a20260731-brazeincxq227ear.htm nowretrievedfull;obs16012–13,timingcorrection16019.2 Q2revenue227.230m,subscriptionrounded+20.8964% vs total26.2%;services~24%incrementalrevenue;GMdown90bp.3 FCF21.744mpositive butSBC35.661m,GAAPOP-18.123m vsnonGAAP+22m.4 H1deferredrevenuecashcontributionincrease27.972mexceedsOCFimprovement21.242m;organicdefinitiondoesnotprovidequantifiedbridge. Strongestbearcase serviceimplementationmix andprepaymentscanboostheadlinegrowth/cashbefore durablehighmarginrecurringreturns;dilutionconsumespartofshareholderbenefit. Counterexample subscriptiongrowth~21%,retention110vs108%,fallingSBCandstrongerFCFsupportrealoperatingimprovement; donotcallreportedcashfake orFCFlessSBCactualcashburn. Validity metricscannotidentifypriced-in-surpriseorreturnprobability; soundnessoneunauditedissuerorigin,organicbridgeandcallreplaymissing. Who benefits:customersmaygainautomation,employeesreceiveequity,issuerfavorsadjustedprofitframing. Startcash;bindingconstraintindependentactual-windowpayoffs. Conclusionresearchonly,preearningswindowclosed. Evidence movingallocation: independentlyfrozenposteventrule withpositivecostedOOSexpectancy/uncertaintyand15minfills;untilthen0,notnumericalprobabilityguessedfrombusinessgrowth.