Stats

Premises: 1. A forecast needs the return window that actually contains the announcement. 2. Event-gap history is a distribution estimate, not an option return backtest. 3. A two-event holdout cannot establish tail calibration. Eighteen individual issuer notices, dated March2022–May2026, confirm nine after-close releases through March20 2024 and nine before-open releases thereafter. Map after-close announcements to next-session open versus announcement-day close; before-open to same-session open versus prior close. All 18 exact dates, original notice URLs, publication dates, aligned prices and log returns saved in observations7151–7168. Primary notice establishes intended phase; historical first-news ticks and any early preannouncements were not independently reconstructed. Adjusted IEX daily bars https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=CHWY&timeframe=1Day&start=2022-01-01&end=2026-09-05&adjustment=all&feed=iex :1172 bars. Sixteen events2022–25 train; March25/June10 2026 held out before calculation, no 2026 refit. J=ln(event open/preclose). Train RMS10.3956%, mean absolute9.2054%, median absolute9.0425%, close-to-close RMS13.0700%. IID event bootstrap10000 seed2026090608 RMS95% interval[8.0483,12.4958]%; conditional on this small changing-regime sample, not a predictive interval. Deliberately wrong all-before-open mapping gives RMS6.1948%, understating variance-window exposure. Held-out gaps+10.1067% and-0.0490% log; RMS7.1466%, meanabsolute5.0779%, N2. Insufficient to validate event tails or distinguish competing forecasts; no p-value or trading edge claimed. Current before-open phase is verified by issuerAug19 https://investor.chewy.com/news-and-events/news/news-details/2026/Chewy-Announces-Fiscal-Second-Quarter-Financial-Results-Conference-Call/default.aspx . Prior was event optimism unsupported; correct timing raises measured historical variance but does not make current asks cheap. Evidence that would move the decision: prospective independent event/option bid-ask returns with uncertainty excluding zero after costs, plus execution entitlement. Counterexample to transportability: changing margins, financing or preannouncements change surprise variance; opening IEX prices can differ from consolidated auction. Valid timing logic; empirical adequacy for a paid tail bet remains unproven.

Premises: 1. Sea-Intelligence Sep2 original summary covers JULY2026, reliability56.4% down6.1pp, mean delay CONDITIONAL on late arrival6.06days up.59. 2. Derive June62.5% and5.47days from same-release differences; late share times mean delay yields diagnostic2.05125->2.64216 (+28.81%). This index is not unconditional lateness: on-time tolerance/early arrivals and vessel weights unknown. 3. Maersk Sep4 cites this same statistic in weather notice: one origin, not independent confirmation, and July cannot quantify later storms. Hapag Sep3 selected-port delay ceilings are different population/period, not contradictions. Collector index labels June62.6%, .1pp different from inferred62.5; preserve possible rounding/revision unresolved, not silently merge vintages. Raw AIS, counts, paired arrivals and spreadsheet unavailable; no defensible regression, confidence interval or causal storm coefficient from these aggregates. Prior was physical disruptions likely material; corroborated current carrier restrictions strengthen mechanism, not stock-win probability. New accepted invoices/independent throughput would strengthen cash hypothesis; normalizing delivery metrics would weaken it. Sources https://sea-intelligence.com/press-room/406-sharp-reliability-drop-from-asian-port-congestion ; https://sea-intelligence.com/press-room ; https://www.maersk.com/news/articles/2026/09/04/east-asia-typhoon-season-impacting-logistics ; https://www.hapag-lloyd.com/en/services-information/operational-updates/updates/2026/09/ops-update-middle-east-week36.html .

Premises1 reuse roundedSep5 reconstructedADRfractions fromanalysis405 BTC3.663%,ETH4.840%,SOL6.386%; these are prior measuredrangeinputs,notfresh tradetargets. 2 Documentedtier1 taker25bp eachside chargedoncreditedasset https://docs.alpaca.markets/us/docs/crypto-fees ; assume10bpslippageeachside asresearchstress,notobservedfill. k=(1-.0025)^2*(1-.001)/(1+.001)=.9930182255. L=1-(1-s)*k;W=(1+1.5s)*k-1. BTC/ETH/SOLnetloss4.3356/5.5044/7.0396%,netwin4.7580/6.5111/8.8139%,RR1.0974/1.1829/1.2521. To algebraicallyreachnet1.5RRtargetsneed7.2522/9.0177/11.3367% but NOT permitted to widen rejectedtargetswithoutnewmeasuredpayoffvalidation. Nominal1%equityriskcaps850.99*.01/L=$196.28/$154.60/$120.89 beforegap/uncertainty; actual0. IndependentquarterKelly gatefailsoverrides positiveplug-in. Snapshotbestask79.45/83.55/100.94USD andbid30/20/99.59USD underscores fillsnotguaranteed https://runway.adityaperswal.workers.dev/data/crypto-snapshots . This fee-on-receivedasset convention differs slightly from fee-on-cashentry ledger; exactaccountactivities stillrequired beforeexecution. Counterexample lower makerfees reducecost but limitfillprobability/adverseselection untested. No new strategybacktest or independentedgeclaimed.

Recomputed all3Catalyst trades with numpy from query_trades,Sep2–4 execution record. Equal-trade mean-16.3806%,wins0/3,totalPNL-31.937258 matches venue-31.94 rounding. Two operator exits censor intended strategy outcomes; one RARE stop exit-45.8601% realized-6.44035R. No winners means averagewin unavailable; do not fabricate W or infer a precise future p=0 from3selectedtrades. Live record therefore cannot calibrate positive Kelly. Saved independent CHWY event holdout analysis38 (recovered via analyses64/172) N6,p=.5,W4.523%,L4.829%,netreturn-1.719988%; clipped quarterKelly max(0,(p-(1-p)/(W/L))/4)=0. Its ten-session rule is not a validated Sep8 one-session substitute. RARE historical-loss stress on49.14cash=22.5356USD exceeds8.5099account risk cap; no claim this observed loss bounds futuregaps. Premises: failed/unknownedge+unvalidatedwindow imply no size; counterexample a future rally would not refute missing ex-ante edge. Source execution ledger query_trades fund=catalyst; historicalbars https://runway.adityaperswal.workers.dev/data/stock-bars ; dates https://investor.chewy.com/news-and-events/news/news-details/2026/Chewy-Announces-Fiscal-Second-Quarter-Financial-Results-Conference-Call/default.aspx . No new backtest or MonteCarlo claimed; resampling3censoredtrades would not create evidence.

Data: fixed six-sector panel, adjusted IEX daily bars. Historical training2020–23 with usable strategy trades startingFeb2021; chronological holdout2024–Sep4 2026. Six daily return series have mean pair correlation.570715 training and.398842 holdout; correlation-eigenvalue participation ratio2.15466 and2.87526, descriptive dimensions NOT effective independent trade counts. Baseline647training/598holdout trades cluster in132/126occupied entryweeks. Bootstrap resamples whole entry-week sums/counts jointly across symbols; robustness uses1/2/4/8-calendar-week circular blocks including empty weeks,20,000 draws. Governing four-week mean intervals:baseline[−.21028,.28597]%,trend[−.47654,−.02183]%,trendlowvol[−.57805,−.02514]%. Thus many trades do not resolve positive payoff. Regression baseline trade return on intercept,trend,lowvol,interaction uses week-cluster sandwich SE. Holdout coefficients in percentage points .86763,−1.16261,−.96635,+1.02668; t2.804,−3.358,−2.383,+2.346. Training corresponding t.662,−.408,−.921,−.281. These retrospective contrasts are unstable, not a license to invert the rule. Strong counterexample:downtrend/highvol holdout mean+.86763% over150 trades with one-week CI[.26937,1.47896]; training+.23312% over164 withCI[−.438,.942]. No untouched evidence remains for selecting this cell; record it as discovery, do not trade it. Week clusters may share multiweek shocks; block results and lack of incident labels limit inference. Source https://runway.adityaperswal.workers.dev/data/stock-bars . Every matrix,coefficient,SE,cell andbootstrap result recorded.

Premises: 1. query_trades for quant contains three closed PEAD positions, all exited by operator liquidation, and no positions awaiting a lesson. 2. Returns computed from actual entry/exit prices are SNOW -5.39236%, DELL +7.93994%, GTLB -5.13050%; their dollar P&Ls sum to approximately -$1.04605. Win rate is 1/3, average winning return 7.93994%, average losing magnitude 5.26143%. Holdings lasted 0.87414, 1.80126 and 1.96827 days, far shorter than the original 4–6 week theses. 3. No reference quotes or order-submission benchmarks are supplied, so fill slippage cannot be estimated; none is a natural stop or target exit. Conclusion: these observations describe realized book experience but do not calibrate the overnight strategy, natural exit odds, or a reliable Kelly allocation. No additional trade-linked lessons required; run-level technical/execution/psyche lessons recorded. Source: query_trades fund=quant, retrieved September 6 2026, observation5796.

Premises:1.Five predetermined ETFs have identical2684session histories2016-Jan4 to2026-Sep4, raw/alladjusted OHLCvalid, no duplicate orcrosssymbolmissingdates.2.Train2016-2020 andlater2021-Aug2026 are chronology splits, not pristine independentmarketregimes.3.SPY later grossfullnightmean3.373656bp, geometricbreakeven1.587993bp/side; grosscompound56.98655% becomes-11.04239% at2bp/side and-90.82762% at10bp.4.HAC5OLS of nextopen/close on currentintraday,laggedintraday,lagged20dayvol gives laggedbeta-.015328,t-.620745,R2.002267; price-onlypredictorsdonotidentify signedflows. Currentintraday regressor descriptive only, never usedfor samecloseconditionalfill. Strongestcounterexample:IWM lowcost+15.96% andSPYlagged+6.38%, buttheirblockmeanCIscrosszero andbasecostfails. Following-dayintradaybenchmarkcorrected so overnight*intraday exactlyreconcilespassivecloseclose. Validity:rejecttheseproxies,notallovernighteffects. Soundnesslimitedbyauction/dividendfeedsemantics,datespreviouslyinspected andbootstrapstationarity. Source https://runway.adityaperswal.workers.dev/data/stock-bars; statistics.json, diagnostics.json.

FixedT10 holdoutN6 mean-10.97667%;50000iid event bootstrap seed60606 95%meanCI[-23.21708,2.91121]%;onewinner Wilson95%winCI[3.00534,56.35028]%. Unknown serial/regime dependence and small sample mean CI is conditional,not precise forecast. Training-minus-holdout mean difference bootstrap seed60607 CI[-.72947,26.15507]points: descriptive deterioration does not prove a statistically identified break. Mean netHAIN minus grossSPY samewindow-12.39050points,minusgrossXLP-11.19342; controls cost-free intentionally favorcontrol and are not a hedgedportfolio. 420dailyreturns2025–Sep4 2026: OLS intercept-.4030566%/day,betaSPY.7719663,betaXLP.9985779,R2.05130145; correlationsHAIN/SPY.1668041,HAIN/XLP.1814433. No robustSE,causalalpha,or stationarity assumed. Single surviving issuer selected by existingdatedwatch,not representative marketcohort; repeatedfundresearch creates selection risk across names. Need prospective independent post-restructuring sample,not more bootstraps of same6. Source https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=HAIN,SPY,XLP&timeframe=1Day&feed=iex&adjustment=all ; issuer-date sources in22observations.

Premises: 1. A more flexible forecast must outperform a fixed baseline out of sample. 2. Overlapping outcomes cannot count as independent evidence. 3. Physical variance forecasts alone are not risk-neutral option values. Fixed six ETFs, adjusted IEX daily close bars Jan3 2022–Sep4 2026,1173each. Let r=log(Ct/Ct-1),v=r². Baseline f=(mean21(v)+mean63(v)+EWMA.94(v))/3. Competitor is our HAR-like OLS log(mean(v[t+1:t+5])) on [1,log(vt),log(mean5(v)),log(mean22(v))],floor1e-10; daily squared returns are not intraday realized variance. Fit976 rows beforeDec24 2025; target ends before2026, explicit purge. Arithmetic forecast exp(Xbeta)*mean(exp(trainingresidual)), coefficients and smearing recorded per symbol. Freeze coefficients for2026; select every fifth eligible date,33 nonoverlapping outcomes each. Loss QLIKE=y/f-log(y/f)-1; compare paired differences HAR minus blend. 5000 circular block bootstraps, block4 outcomes, seed2026090606; all six95% intervals include0. Deltas GLD+.185717,IWM+.005959,TLT+.004236,XLE-.016413,XLF+.025697,EEM+.031996. Thus no demonstrated general improvement; do not adopt HAR based on it making certain calls look cheap. Counterexample: XLE mean improves and GLD/EEM logMSE improves; losses rank models differently. These findings prevent claiming universal dominance of the blend. Universe chosen for economic diversity, not random market sampling; high/low regimes uneven, daily IEX noise and common market shocks limit precision; six markets are not198 independent trials. Prior qualitative expectation of fragile cheap-tail values retained, no invented posterior probability. Evidence that changes stance: prospective frozen forecast gains with uncertainty excluding0 on eligible contracts plus actual option bid/ask holdout. Source https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=XLF,XLE,GLD,EEM,IWM,TLT&timeframe=1Day&start=2022-01-01&end=2026-09-05&adjustment=all&feed=iex .

Read query_trades fund=catalyst Sep6. Three closed trades, wins0; returns MIRM-.439321%,IONS-2.842420%,RARE-45.860062%; arithmetic average loss16.380601%, dollarPnL-31.937258 before unallocated system costs. Two operator liquidations confound intended-horizon outcome; one recordedstop sale. RARE exit13.98976 versusstop24 is41.709333% below trigger, not necessarily order-routing slippage: no contemporaneous executable quote captured. All were fractional market-notional entries; no limit comparator and no reference quote to estimate entry slippage. Winnerpayoff unobserved, quarterKelly sizing0 by mandate. Do not pool these three FDA names as an earnings event sample or infer a stable0% true win probability. Source runway:query_trades?fund=catalyst&status=closed ; venue/context providesaccount850.99,fund49.14,1%risk8.5099. This is execution/outcome evidence, not a causal rule backtest.

Premise1 hourly venuebars reconstructed into UTCdaily closes; only24hour days count, nofill; training2024-2025,SOLstartsAug2024; historicalevaluationJan1-Sep5 2026. All3common trainingN484,evaluationN246. Premise2 dailyreturn correlationBTC/ETH increases.78545->.91234; BTC/SOL.76891->.88042; ETH/SOL.76842->.88022. Evaluationannualizedvolsqrt365 BTC46.925%,ETH64.509%,SOL64.059%. WeekenddailySD1.560/2.498/2.685% below weekday2.741/3.679/3.595%; observationdoesnotmeanweekendstops safe,1%tailsstilllarge. Mean dailyreturns-.01126/-.02482/-.02767%. Conclusion threecrypto positions can share one risk driver; never pool three correlated histories as independent support for Kelly. Counterargument common-date training coversshorter SOLhistory and periods differ, so correlationchange is descriptive not causal or guaranteedpersistent. Missing/incomplete days excluded, venuespecificfeedcoverage and priorinspectedholdoutbound inference. Alltrain/evaluationquantiles,counts,means,SDandcorrelations recordedobservations; source https://runway.adityaperswal.workers.dev/data/crypto-bars .

Preregistered note295:501dailytrainingreturns2023–24,250heldout2025;2factorOLSlogreturnsagainstSPY/XLB,known2026shockdatesdescribedafterfit. HoldoutpredictiveR2LIN.5952/APD.5146,RMSE.7677%/1.2194%daily. Five-sessionresiduallog%onMar2/Mar19/Jul10:LIN+.5278/-1.3739/-2.0582;APD+6.3321/-.3003/-1.1084. This is NOT an executable event backtest: includes announcement-dayclose-to-close,relatedincidents,unknownfirst-publicationminute,andearningsconfoundsin20dayJulywindow. Post-hocrobustnessusingSPYonly(samefitdates/all3events,notretuningforprofit)changesLINto-3.4158/+1.0627/-.8356andAPDto+.1830/+3.4991/+.7979;attributionsdependstronglyonbenchmark. StateStreetSep3XLBweightsLIN12.97%APD4.76%confirmbenchmarkcontainsbothandcannotbeunaffectedcontrol;weightsnotbackcast. Valid conclusion: no stable causal/equity payoff identified;positivephysicalexposuredoesnotestablishtradeedge. 2025overlappingplacebowindowsareadescriptivedistribution,notindependentNorp-values. No win probability inferred. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; https://www.ssga.com/us/en/individual/etfs/state-street-materials-select-sector-spdr-etf-xlb . Everyeventstatobs5419–5436;scripts/resultsarchived.

Premises: 1.Date-aligned adjusted closes, Jan2016–Dec2021 training and Jan2022–Aug2026 later diagnostic. Later1168 close-to-close returns; passive MRGR20.68199%, MNA15.04870%, BIL19.32620%, SPY70.66035%, before investor entry/exit friction but reflecting fund expenses. MRGR CAGR4.13930% versus BIL3.88576%; not a large current premium. 2.Circular21-session blocks,10000 bootstrap samples,seed90606: annualized arithmetic MRGR-minus-BIL mean0.31483%,95%CI[-2.96955,3.59355]; MNA mean-.66184%,CI[-4.12471,2.98204]. These are uncertainty estimates, not distributions of future deal success. 3.Numpy piecewise OLS: excess ETF return on excess SPY and hinge below-1%daily/-4%monthly; Newey-West lag5daily/3monthly. Training monthly MRGR normal beta.0682,tail.2776; MNA.0372/.6376. Later monthly MRGR normal-.0404,tail.2211,alpha t1.20; MNA.0234/.2004,alpha t.13. Only9later tailmonths,55monthlyreturns; earliermonth omitted by firstclose baseline. Daily intercepts also insignificant later. 4.Low unconditional volatility/correlation does not imply no crash exposure or alpha. MRGR sparse prints can bias estimates; ETF wrappers and eras are not independent corroborations. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; original nonlinear-model motivation https://bpb-us-w2.wpmucdn.com/voices.uchicago.edu/dist/d/2771/files/2020/09/JF_riskarb-1.pdf . Complete8regressions,8passive rows andCIs archived. Conclusion: passive exposure may diversify, but evidence cannot authorize this fund's short swing trades.

Premises: same stock onlyonecompletedFDAdeadline; arbitrary stock windows measure backgrounddrift,notdatededge. Fixed15sessionnonoverlapping windows with16sessionstep,IEX2021–Sep42026,fee5/slip10bp eachside. Train61windows mean1.2841%,median-1.5474%,win45.9016%,bootstrap95%meanCI[-6.0859,11.4398]%. Holdout25windows fullyentered/exited2025+,mean1.6538%,median.8703%,win56%,CI[-3.6993,7.1608]%; oneDec18–Jan13boundarywindowexcluded explicitly. 20000iidbootstrapresamples,seed60604; nonoverlapreducesmechanicaldependencebutregimeautocorrelationremains. Descriptive420dailyrows2025+:corrSRRK/XBI.5691,SPY.3380;twofactorOLSbetaXBI1.2164,SPY.09075,dailyintercept-.06767%,R2.3243. No robustSEorcausalclaim; XBIcontainsbiotechexposureandfactorscorrelated. Conclusion: risingpriceorpositiveplacebomeandoesnotidentifyeventpremium. Need prospective independentlyclassifiedmanufacturing-resubmissioncohort with announcements knownbeforeentry and fullpipelineoverlap. No posteriorapprovalprobability assigned; N1cannotcalibratehazard. Source https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SRRK,XBI,SPY&timeframe=1Day&feed=iex&adjustment=all . Allmeasuredrows/metrics archived.

Premises: independently forecast physical variance; hold out outcomes; do not equate physicalvariance orbetterforecastscore to risk-neutraloptionvalue. x=logadjustedclose return. Fixedforecast variance=(252var21+252var63+252EWMA(x²,lambda.94))/3. Train2023–2025; winsorclipbounds from751trainingreturns[-.04953405,.06813260] fixed before2026 comparison; 165heldout next5observedsession mean-squarelabels, trainingevaluationwouldpurge5sessions; no fittedcoefficients. CurrentRV21=.3851876,RV63=.3787668,EWMA=.3769041; blend=.3803027,winsor=.3593251. QLIKE=mean(logf+a/f); winsor-minus-blend -.038909, pairedcircular10-rowblockbootstrap10000 seed2026090604 CI[-.138576,.070907], so noestablishedsuperiority. Blend-minus-RV21 -.178529,3000bootstrapCI[-.375315,-.014350]; overlappinglabelsnotindependenttrades. Rawunadjustedforecast holdoutQLIKE-.252945 illustratesmemoryofsplitartifactevenaftercurrentRVhasrecovered. Train/testdailymeans+.15398%/-.10618%,dailySD2.11683%/2.35143%,excesskurtosis7.215/7.225; top5days18.965%/43.095%totalsquaredreturns. PosthocclassificationFeb27,Apr17,Jul17 of170testdays accounts35.751%variance; eventmean-square.0111585 versusother.00036024,ratio30.975; OLSx²=a+bI(event),b=.0107983,HC1SE.0024770. This was selectedafterextremeinspection: descriptive, notcausal/significanceevidence. Dates verifiedissuerWBDwithdrawalFeb26, earningsApr16/Jul16; notallothereventsclassified. Four-sessionnonoverlap terminalputproxy atS78.25,K72,fixedhypotheticalpremium3andfees.20: trainN187win1.604%mean-$1.1416; testN42win7.143%mean+$7.7724,avgwin150.414,avgloss3.2,10000IIDwindowbootstrapmeanCI[-3.2,28.8344]. Threewinsdominatetest; historicalconstantmoneyness/currentpremiumproxyisNOTactualoptionbacktest. Removingtailsmaydiscardgenuinejumpinsurance, while includingdatedevents mayoverpredictordinaryweek; neitheridentifiesnextweek. Sources venueadjustedstock-bars; https://www.sec.gov/Archives/edgar/data/1065280/000106528026000212/nflx-20260630.htm ; https://ir.netflix.net/investor-news-and-events/financial-releases/press-release-details/2026/Netflix-to-Announce-First-Quarter-2026-Financial-Results/default.aspx ; correspondingSecondQuarterrelease. Allocation0.

Adopt data-normalization rule, not a trade. Premises: (1) returns must compare equal share units; (2) option series must follow OCC adjustment terms and dates; (3) both sides need current executable quotes. Collected complete single-page 922 daily IEX bars Jan3 2023–Sep4 2026 independently for adjustment raw/split/all. Raw Nov17 2025 log return -2.310715 versus adjusted -.00812572; full-history annualSD125.707% versus34.338%; raw buyhold-73.465% versus adjusted+165.344%. Split and all closes identical. CurrentRV21 stays38.51876%, equalvarianceblend38.03027% exactly reproduces prior analysis283: correction to the suspicion, split artifacts do NOT explain this recent forecast. OCCNov14memo57651 effectiveNov17 divides old strikes10 and multiplies contractcount10, premiumextension100, deliverable100NFLX shares; NFLXrootunchanged. Old Sep18K1280 calls94.21/99.30 andputs217.67/225.83 retainNov14quotes; their normalizedK128series haveSep4quotes call0/.06 put48.30/50.67. Normalization cannot make oldquotes current. Sep8–25 chain748records,258quotes2025,469within120sofSep4close,21otherolder; adjacentverticalflags9all versus5fresh. Largest$827.20gap mixes pre/postsplitseries. Five remainingflags show timestampfilters are necessary but insufficient: indicativefeedprices are derivatives. Statecash850.99/fund127.65; noholdings. Sources https://infomemo.theocc.com/infomemos?number=57651 ; https://www.sec.gov/Archives/edgar/data/1065280/000106528025000450/nflx-20251114.htm ; https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=NFLX&timeframe=1Day&start=2023-01-01&end=2026-09-05&adjustment=all&feed=iex ; https://runway.adityaperswal.workers.dev/data/option-chain/NFLX?feed=indicative . Counterexamples: newrealjump remainsafteradjustment; fresh syntheticquotescanstillviolateverticals. Valid inference is artifactrejection, notefficiencyproof.

Premises1 Completed paginated2024-Jan1 to2026-Sep5 history contains23492hours each BTC/ETH,4absent versus23496expected; SOL17761hours beginsAug26 2024, notJan1. Compare976/976/737 complete24hourdays; no forward fill. 2 Max daily/open divergence6.2064/7.9068/10.2992%,allDec2 2025; volume discrepancies exceed1% on12/11/10days. Large shared clusterNov23-Dec18 2025. Median open/high/low/volume differences0. 3 Direct us endpoint Dec2 reproduces daily discrepancy; independently aggregatedminute/hour OHLCV equal to floating precision, Coinbase opens within.022% of intraday opens versus6.19/7.93/10.32%daily. 4 RecentAug1-Sep5 36days all3assets864hours each: open/high/low/volume reconcile, maxcloseBTC.002113%,ETH/SOL0. Coinbaseclosemax.02558/.04755/.06659%. Conclusion strongest evidence favors localized daily historical aggregation defect, not universal bad current tape or proxy-only failure. Soundness: both Alpaca endpoints same origin; Coinbase independent exchange has differentquotes/trades, so not executable Alpaca groundtruth. Sparse minute rows may benoactivity. SuffixscanDec2 late-day17:28/17:45/17:42matchesextremesbutnotallfields; reset/corruption plausible notproven. No exact causal explanation from primarydocs/search. This improves prior257 bylocalizing/upstreamreproducingproblem. Source https://data.alpaca.markets/v1beta3/crypto/us/bars ; https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://api.exchange.coinbase.com/products/BTC-USD/candles ; correspondingETH-USD/SOL-USD endpoints; https://docs.alpaca.markets/us/docs/historical-crypto-data-1 .

Premises:1 forecastphysicalvarianceindependentlyofoptionIV;2 evaluatenextperiodvariancebeforetrustingvaluation;3 bettervarianceforecastdoesnotidentifyrisk-neutralprices orlongpremiumprofit. For62roots with>=100adjusteddailyIEXbars throughSep4, x_t=ln(close_t/close_t-1). RV21/RV63=samplestd(x)*sqrt252; EWMAvariance v_t=.94v_t-1+.06*x_t², initializedfirstsquaredreturn; annualEWMA=sqrt252v_t. Fixedforecast sigma=sqrt((RV21²+RV63²+EWMA²)/3). No fittedcoefficientsorparameteroptimization; trainpreDec20 2025 (five-sessionlabelpurge), holdoutJan2026; predict next5observed-sessions meansquaredlogreturn*252. QLIKE=mean(logforecastvariance+realizedvariance/forecastvariance), lowerbetter; overlapping5daylabels mean165holdoutobservationsnot165independenttrades. BlendbeatsRV21alonein50/62roots, notmultiplicity-adjustedsignificance orclaimedoptionedge. SNAPforecast65.457%, stress50.445%; testQLIKE.319948 versusRV21.564146, realized/predictedvariance1.11539. MARAforecast96.768%, stress74.220%, QLIKE.843535vs.866474,ratio1.06844. ETHA53.343%/.410587stress, QLIKE.137118vs.371150,ratio.96578. EWZ23.902%, QLIKE-1.63049vs-1.57624,ratio.99352. Everyrootinputanddenominatorrecorded. Stress sigma=.8min(RV21,RV63,EWMA) isdeliberateadverseassumption, notconfidencebound. Currentlisteduniverse andIEXsamplingcausecoveragebias; crisisvariance,jumps,ETFdividendclock andriskpremiuminvalidateBSMconstantvol. SPCXonly59bars excludesits1094contractsfromvaluation;163morecontracts haveinvalidpriceconditionsorunparsedstandardformat. Decisionretainforecastasresearchbenchmarkonly, noempiricaloptionwin/loss andzeroallocation. Source https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&start=2023-01-01&end=2026-09-05&adjustment=all&feed=iex .

Premises:1 crypto price risk continueswhileUSoptionsclosed;2 equalcalendardayvariance andzerosessionclosurevariance aretestableextremes;3 UTCspotbars cannotdirectlyidentifyNYcloseETFoptionvariance. CollectedcompletepaginateddailyBTC/USD andETH/USD barsJan1 2025–Sep4 2026; logclose-to-closereturnsq_t; OLSq_t²=alpha+beta*I(Saturday/Sunday). Fixed2025train364returns/53weeks,2026heldout248returns/36weeks; weeklyblockbootstrap5000,seed20260906. BTCweekend/weekdaymeansquaredreturnratio0.5768train and0.3303test; test95%interval0.1542–0.6998. ETH0.6329train,0.4844test; testinterval0.2063–1.0165 includes1. Lowerpointestimate isnotcertaintyforETH. Testmeanreturn/day BTC-0.03688%,ETH-0.07244%; dailySD2.4534%/3.3483%, excesskurtosis7.435/5.068; heavy tailschallengeGBM. Weekendsharesoftotaltestvariance11.70%BTC,16.27%ETH, notzero. Ratesandtotalsmatter: twoquieterweekenddaysstilladdrisk. Initialhourlyrequesttruncatedat960BTChoursandnoholdout; cutthatthreadratherthanclaimproperNYclosealignment. Dailyresultsarelowerresolution andcannotcalibrateholidayFriday-toTuesdayoptionclocks orMARAoperatingleverage. Noalphastrategyreturnderived; noadjustmenttohistoricaltestparameters, nogrosscarryproposalandnospotcryptotrade. New evidence weakensliteralzero-weekendinterpretationofsessionclock, while retainingit asadversescenario ratherthanfairvaluebound. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars?symbols=BTC/USD,ETH/USD&timeframe=1Day&start=2025-01-01&end=2026-09-05 ; providerdocs https://docs.alpaca.markets/us/docs/historical-option-data .

Premises:1 wholemarket coverage requires an enumerated listinguniverse, not a handpickedwatchlist;2 missing/bad/stale data cannotbeclassifiedaszeroactivity;3 IEXliquidityrank is a triageinstrument, not an alphabacktest. Source primaryNasdaq nasdaqlisted.txt andotherlisted.txt filecreation0904202621:31, downloadedSep6:5584Nasdaq+7571other non-test listings=13155. Parserpreserves literalNA ticker, removesTestIssueY andfooters, deduplicatessymbols. Venue stock-snapshots feediex acrossallbatches initiallyhit429andunsupportedpreferredsymbols400; cachedsuccesses, pacedretries, constrainedremainingcalls toplainletters1–5.12612plainroots,543excludedformats requiringmapping(notprovednontradable);12573snapshots,39missingplainroots;12415positiveprice/volume dailybars,8056Sep4 and4359older. LatestdailyBar canbestale: neverrank oldbardollarvolume ascurrentliquidity. Top20newroots selectedfromSep4-leadingnames beyond43predeclaredglobal/sector/crypto roots andprior14; inspectedselectedtimestampsallSep4. AggregateIEXdollarvolume c*v andrange100(h-l)/c only. Initialfailurebatchesdonotenterdenominatorsaszerovolume. Coverageisalllistingmetadata and8056freshunderlyingbars, NOTalllistedoptionchains. LargestcapturedIEXdollarvolumes MU862.83m,NVDA826.96m,SPY777.49m,SNDK565.91m,TSLA537.42m. These arevenuefeedmeasurements, notconsolidatedmarketturnover. Current-universe survivor/newlistingbias bars history nothistoricalwholemarketconstituents. ForeigncontextworldCLI coveredJapan/Germany/Netherlands/HongKong andUK/Eurostoxx/FX/oil; chartpreviousClose mayreference5daywindow, sono claimofdailychangefromit. Action broaderchainselectiononly, nofunding/proposal. Sources https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt ; https://www.nasdaqtrader.com/dynamic/SymDir/otherlisted.txt ; https://runway.adityaperswal.workers.dev/data/stock-snapshots?feed=iex ; https://query1.finance.yahoo.com/v8/finance/chart/ .

Premises1 Use completedhourlyaggregateddailycloses, nofillforward acrossmissingdays; simpledailyreturns, training724pairedreturns/2026holdout246. 2 NumPyOLS ETHreturn=intercept+betaBTCreturn; BartlettNeweyWestlag7 covariance, normal95%interval. Traincorr.79072,beta1.15628(SE.05772),dailyalpha-.05332% CI[-.22539,.11874]%; holdoutcorr.91234,beta1.25422(SE.06118),dailyalpha-.01070% CI[-.13857,.11717]%. HoldoutETHdaily5/50/95percentiles[-4.77066,.00937,5.11839]%. 3 Relative-strengthrule255 stillloses-.40609%pertrade,meanblockCI[-1.33974,.64794]%; grossco-movementnotstrategyalpha. Conclusion rejectindependentETHoutperformance claim fromcorrelation/pricelevels; evidenceconsistentwithsharedmarketexposure. Valid descriptivecalculation, notcausaltestofupgrade/treasurybuying; omittedmacro,changingbeta,quoteconstruction,multipletests limitsoundness. Strongbullcase relative-strengthregimesmayhelpinunseenwindows; requirenewprospectiveevidence, notretunedhistory. Source https://runway.adityaperswal.workers.dev/data/crypto-bars . statsmodels unavailable, explicitNumPymatrixHACimplementationarchived.

Premises:1 Seven actualETF adjustedSIPdailyhistories2016-Sep4 2026 each2684bars avoid backfilling current stock constituents.2 Later2023-Sep4 SPLVvol11.2408% versusSPY14.9966%,beta0.35336,corr0.47142; passiveclose-return26.6994% versus111.5209%. USMVvol10.3056%,beta0.49859,return48.0829%; EFAVvol10.6818% vsEFA14.9815%,beta0.60455,return65.7573% vs83.0546%; EEMVvol12.2949% vsEEM19.1052%,beta0.60294,return58.8776% vs94.5624%.3 OLSdailyconstant HACNeweyWestlag5tstatistics SPLV-.0980/USMV.1594/EFAV1.2545/EEMV.6372; none establishes positive intercept. N922 dailyreturns, passiveprice ratios firstclose-lastclose exclude firstovernight; no tradingcosts or cash benchmark deducted.4 Primary44monthlytrade mean-.38446%,3monthbootstrapCI[-.70870,-.03333]%; EEMV+.02897% CI[-.29972,.35449]%. Conclusion: evidence supports lower ex-post risk exposure, not faster profits or robustalpha; negative swingadaptation does not refute longhorizon lowrisk theory. Onlymarketmodel, correlatedfunds, publicationselection and inspectedlaterwindow limit inference. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; https://www.aqr.com/-/media/AQR/Documents/Insights/White-Papers/Understanding-Defensive-Equity.pdf . All8regressionrows and27testrows saved.

Premises1 DownloadedJan1 2024-Sep5 2026 BTC/ETH hourly,each23492of23496expectedhours,976complete24hourdays and3incompletedays. Duplicateboundaryrowsdeduped,partialSep6excluded;50/14dayrollingwindowsrequirenoNaN. Initial10workerfetchhit429andfailed, replaced3workersglobal.55secondpacedrequests,retrybackoff,persist60daychunks; finalcoveragechecked. 2 Same-sourcecompleterawdailyvshourlyaggregationmaxopenabsdiffBTC6.2064%,ETH7.9068%; low3.9861/6.8874%,close.09466/.09229%,volume96.6146/99.3558%. Quote-midpointconstructiondocumentedbutdoesnotexplainaggregationdifference. 3 Hourlytrainingcandidatecrossesonegapandtagsdatagapexit; holdoutzero. Conclusion historicalbarreliability remainsbinding; matchingcashledgerdoesnotcertifyfills. ReconstructedUTCdailyis sensitivityreference, notdeclaredgroundtruth. Cause/location/pipelineprovenance unresolvedwithin7minuteinitialauditallowance; follow-ondownloadcompletedinthetestbudget. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/historical-crypto-data-1 .

Eight current-survivor names selected after broad discovery; windows chronologically split, but selection sees current winners so not independent economic-alpha confirmation. NumPyseed90602,10000 IID bootstrap resamples per ticker of natural net holdout trade returns (N3–7), 95%percentile intervals all crosszero. AEHR[-22.65,162.84]%, COHU[-6.25,33.82], AXTI[-26.87,3755.10], MP[-9.28,83.34], MU[-7.56,81.92], Maersk[-5.77,11.05], Lynas[-1.98,52.82], Fujikura[-21.51,134.71]. Five of8 fixed training returns negative; tinyN, overlapping market regimes and selection likely understate uncertainty. US420dailyOLS onSPY: betaAEHR3.447,COHU1.986,AXTI2.511,MP1.416,MU2.568; descriptive intercepts not causal supplyalpha. Overseas same-date regression is nonsynchronous and localcurrency, not hedge ratios; discard for sizing. Two live supply wins were forcedoperatorliquidations; no naturalexit or loss distribution, slippage unmeasurable without contemporaneousquote. Prior remains zeroeligibleedge; prospective synchronized evidence with positivecost-adjusted lowerbound would increase tradeeligibility, additionalheadline duplicates would not. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; https://query1.finance.yahoo.com/v8/finance/chart/ ; query_trades fund=supply IDs3,6.

Frozen11eventholdout in244: p7/11=.636364,W1.502475%,L1.655815% implies rawquarterKelly.058903788. NumPyseed2026090602,20000IIDeventbootstrap:mean.354006%,95%CI[-.732594,1.532965]%;positivebootstrapmeans72.22% is not forwardwinprobability. Same-windowSPYgrossmean.386645%;HITInetexcess-.032639%;correlation.093685;OLSintercept.284421%beta.179970,R2.008777,N11 descriptive,notcausalalpha. Stressmean.053394%,rawquarterKelly.007619052. Bindinggatepositive95%lowerboundfails=>authorizedsize0. Own3closedtrades0wins,meanloss16.380601%;2operatorliquidations+1stop cannotestimatecleanrunupedgeoraveragewin. OneknownpreannouncedmatchSep2025returned3.639981%net;N1doesnotvalidatecurrentpreannouncedregime. Premises:positivepointestimateexists; uncertainty includesloss; implementationcostandregimeshiftremain. Conclusion0isvalidunderpreregisteredgate; soundnesslimitedbyselectedissuer/smallN/dependence, bootstrapmayunderstateuncertainty. Evidence tochange:prospectivelycollectedmatchedregimeeventreturnswithpositivecoststressedCIandverifiedexecution,notmonthlyfundingpressure. Sources priceAPI https://runway.adityaperswal.workers.dev/data/stock-bars ; issuer https://hightideinc.com/high-tide-announces-preliminary-q3-2025-guidance/ ; ownrecord runway:query_trades?fund=catalyst .