Stats

Python NumPy seed20260906; 20000 IID event bootstrap samples for each fixed holdout. Source price API https://runway.adityaperswal.workers.dev/data/stock-bars ; exact issuer URLs in calendar archive; own trades runway:query_trades?fund=catalyst. LEN baseline N10 mean-1.176670%,95%bootstrap mean interval[-3.803275,1.275506]%; overlay N10 mean.181465%,interval[-1.496710,1.790803]%. Fraction of resampled means positive=.17765/.5879, not forecast win probability. Overlay p=.6,W2.246256%,L2.915723% implies raw quarterKelly .0201963; preregistered uncertainty gate and negative stress mean reduce authorized size to0. Baseline raw quarterKelly-.139203 clipped0; stress overlay-.015288 clipped0. Paired gross SPY window mean+.098397%; baseline net excess-1.275067%, correlation.455788; OLS intercept-1.298713% beta1.240315,R2.207742,N10, descriptive not causal alpha or robust inference. Own three trades won0%, average loss16.380601%; two forced liquidations and one stop make this a selected mixed-catalyst sample, not setup-specific p=0 forecast. No average winning payoff can be estimated from it. RARE45.860062% loss applied to49.14 implies22.535634 loss versus8.5099 equity-risk budget; stress, not a promised worst case. Counterargument: positive overlay payoff could reflect real edge. Rebuttal:10 outcomes, bootstrap spans0, costs reverse it, and dependence may make IID interval optimistic. Need new independent prospective evidence, not parameter revision.

Numbered premises:1 Freeze dates, rule, costs and controls before reading price returns (note213).2 In SPY the82training monthly events have prior5day/rebound correlation-0.09716 and OLSslope-0.06601, versus44later events correlation+0.06206/slope+0.04214. EFA latercorr-0.04610 and EEM-0.05156 are similarly small; these are descriptive, not causal regressions or significant estimates.3 SPYrange later19trades mean-0.186537%,p42.1053%,W1.28528%,L1.25695%,95%CI[-0.96363,0.57048] from10000 circular3calendar-month blocks including inactive months,seed906.4 T+1 range12trades return-8.791%,mean-0.75713%; priorT+2 seven+5.63693%,mean+0.79162%. No causal attribution to settlement change because marketregime also changed. EFA range+0.62029%/19trades and EEM-4.60343%/21 do not have positive lowerconfidencebounds. Conclusion no reliable positiveedge; primaryrangequarterKelly clips0. FullholdpositiveSPY1.22081% annual0.3313% is notrobust, and frictionlessadjustedbuyhold109.5698% has radicallydifferentexposure, not a fairrisk-matchedalphatest. All3-marketobservations correlated; not57independentSPYtrades. Counterexample: larger signed-flow shocks may exhibitstrongerreversal thanournegative-returnproxy; lackofproxyedge doesnotrejectthepaper'smechanism. Initialevent-block bootstrap replaced beforefinalrecordingwithregisteredcalendar-monthblocks; savedstats.json containsfinalnumbers. Sources https://runway.adityaperswal.workers.dev/data/stock-bars and primarypaper https://aaltodoc.aalto.fi/server/api/core/bitstreams/0f01d3c2-3b16-44f9-9135-3f03cdd29334/content .

Premises1 Refetch directproxy UTCdaily/hourly Dec1-3 forBTC/ETH/SOL; all72hour rows each complete. DailyDec2opens exceed hourlyaggregate by6.2064/7.9068/10.2992%; dailyvolume underreports sums82.7396/89.8276/80.2971%. Dailyclose matches hourlylast across3days; this does not validate open/range/volume. 2 CoinbaseBTCDec2open86282.35 agrees approximately Alpaca hourly86270.82, not daily91625.132. Crossvenue comparison is corroboration of morninglevel, not samevenue executiontruth. 3 LTC979calendar days/23492hour rows: four missing hours on3days;976complete. Maxcompleteopen discrepancy6.55235%,low6.20774%,volume98.4400%; dailyvsaggregatedclose tiny differences558days,max.240743%. Fixed20breakout+SMA50signals identical77days; stopranges can stilldiffer. 4 Alpacaofficialdocs include quote midprices; no-trade prices possible. This alone cannot explain omission of morning range and most volume. No causal vendor explanation obtained within8minuteaudit allowance; fullmajorhistoricalexpansioncut, namedgap. Initial30pagecap failed because responsepages~200rows despite requested10000, not repeatedtokens; rejectedtruncatedinputs. LTCfetch repartitioned90daywindows, paginationplusendcoverage, final979days completeendSep5 23UTC. Valid conclusion: rawdaily figures cannot support exact stopsize/volume claims; use prereg same-source hourlyaggregation only as sensitivity, not certified repair. Require 15minute/executablequotes and prospective evidence before adoption. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://api.exchange.coinbase.com/products/BTC-USD/candles ; https://docs.alpaca.markets/us/docs/historical-crypto-data-1 .

Premises: fixed diagnostic has5closed holdouttrades andzero closed trainingtrades. NumPyseed906,10000IIDresamples of5 natural net payoffs:mean23.11798%,95%percentileinterval[-6.49303,58.89909]%. IID tiny-sample assumption ignores clustering/unseen crashes; interval is exploratory and likely understates structural uncertainty. 2025+daily OLS onSPY420alignedreturns givesbeta1.78518,annual arithmeticintercept45.17725%,corr.554865; descriptive selectedhistory, not causal supplyalpha. The huge averagewin66.1459% reflects long trend holds and cannot set a short swing target. Two live fundwins are forcedliquidations with no observed loss sample; they cannot replace this uncertainty. Prior cautious; genuine prospective net naturalexit sample with actualvenue risk control and positive robust lowerbound would change eligibility. Archive2958–2997. Source https://runway.adityaperswal.workers.dev/data/stock-bars . Deductionzero authorizedsize valid under edgegate; no forecast that GEV falls.

Premises: (1) future underlying variance affects long-option value; (2) a historical variance forecast may inform value but physical and risk-neutral distributions differ; (3) executable quote and model uncertainty gates are necessary. Method: IEX adjusted daily bars Jan2023-Sep4 2026, IBIT Jan2024 onward. Let r_t=log(C_t/C_t-1), v_t=r_t^2. OLS predicts mean(v_t+1...v_t+10) from intercept and means of prior5/21/63 daily v; train features before2025-12-15, purging10 future label sessions before2026. Variance floor1e-8; annualized sigma=sqrt252*prediction. Frozen holdout Jan2026 onward,16 nonoverlapping ten-session labels per symbol. Coefficients/each MSE and inputs recorded in per-symbol observations2915-2933. QQQ MSE1.1259e-8 vs RV21 baseline1.6753e-8 (32.79% lower); SPY2.1827e-9 vs3.7162e-9; FXI2.5531e-8 vs5.8587e-9, EEM8.1043e-8 vs4.1328e-8 and TLT6.4004e-10 vs6.3178e-10 worse. Forecast estimatesQQQ16.3489%, SPY13.3867%, IWM21.2358%, PATH54.5551%. Soundness: small holdout, unconstrained negative coefficients, changing regimes, corporate actions, incomplete IEX coverage and physical/risk-neutral wedge limit inference. No volatility MSE translates into a known option win rate. Counterexample: steady low volatility after a large historical jump leaves elevated forecasts and overpriced tails. Source template https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=QQQ&timeframe=1Day&start=2023-01-01&end=2026-09-05&adjustment=all&feed=iex . No causal macro claim or economic profit established. Conclusion: keep as research diagnostic; size zero.

1. RepairedhourlyseriesendSep4 23UTC; BTC/ETH24428rows eachfromNov22 2023,SOL17737fromAug26 2024 16UTC. Missing4/4/7hours; holdout7823/7822/7823vs7824expected. No missingholdoutmidnight; SOLmissingOct11midnightisinpurge. 2. Comparingcompleteholdoutdays325/324/325: largestabs(dailyopen/hourlyopen-1) BTC6.2064%,ETH7.9068%,SOL10.2992%,allDec2 2025. Daily/hourlyopens91625.132/86270.82,3020.1375/2798.84,139.716/126.67 respectively. Causes unresolved; do notassume identicalinputprices or attribute everydaily/hourlyreturn differencetomanagementfrequency. 3. Actualholdoutentry comparatorsBTC19,ETH19,SOL30: maximumopendifference.467143/0/0%; oneBTCentrydiffers>10bp. 4. BTCMar10-Mar13 2026timeoutwin+2.335254%crossesmissingMar12 17UTC; ETH/SOLnaturalholdouttradescrossnogap. ThatBTCoutcomehasunobservedbarrierpath; cannotclaimexactstoptest. 5. Bothdaily/hourlyimplementationslossesandno currententry preserve rejection, but reportedreturnmagnitudesarediagnosticswithdatauncertainty, not executable validated performance. Independentcashledgerreconciliationprovesaccountinggiveninputs,notinputtruth. Primarydata https://runway.adityaperswal.workers.dev/data/crypto-bars ; fields/ranges,missinglistsandcomparisonsarchivedaudit.py/audit2.py. Need vendoraggregation reconciliation or independentquote/tradehistory before futurepositive adoption. RemainingmissingglobalOI/borrow/liquidationparticipantdata limits causalclaims. No optimistic repair/holdoutretuning performed.

Premise1:549latertrades share184rotationdecisions and44entrymonths; treating549asindependent exaggerates information. NumPy10000circular3monthblock resamples(seed2026090504) retain concurrenttradeclusters; mean95%intervalfixed[-.1050,.3364]%,range[-.2139,.0512]%. Theseareconditionalhistorical intervals, notposteriorprobabilities. Premise2:equalnineETFbuyholdlater+72.671%,DD16.389%,vsfixed+19.897%,DD18.425%; dailyOLSbeta1.0168,annualarithmeticintercept-9.903%. Rangebeta.4178/intercept-10.540%; descriptive,no causalalpha/significanceclaim. Trainingbasket+85.453% versusfixed-21.436%,range-16.849%. Premise3:evenrangezero-costmeanCI[-.0178,.2502]% and2bponeway[-.0561,.2102]% span zero; fixedlowcostmeanspositivebutexecutionandbenchmarkremainconstraints. Conclusion no robustdeployableedge forrange; fixedpositiveheadlineisnotselectionalpha. Strongcounterargument smalleractualcosts/turnovernettingcouldimprovefixedreturns; repeated177/546slots meansforcedexit/rebuyoverstatesoptimizedturnover. Nettingisuntested, donotclaimallreversalfails. Current-survivingETFselection,sharedresearchdates,stationarityandcorporateactionassumptionslimitsoundness. Threeownforcedliquidationsprovide0naturalexits,notaconditionaledge. Source https://runway.adityaperswal.workers.dev/data/stock-bars ;query_trades fundquant. Fullstats/symbol/yearrowsandseedincode.

Fixedtest132holdout13naturalexits,p4/13=30.7692%,W7.54887%,L7.74074%,mean-3.03624%. NumPyseed2026090504,20000IIDresamples yieldsmean95%[-7.58618,+1.62648]%. Wideconditionalintervalpermitspositivepopulationmeanbutdoesnotestablishit; selectedname,regimeclustering andgenericpriceeventsmakerealuncertaintylarger. NoquarterKellyriskfromnegativeplug-inmean. Daily2026OLS170alignedGRAL/SPYreturns:beta2.34322,intercept.0226788%/day,correlation.309332; descriptiveexposure, notcausalalphaorpanelreactionforecast. Returnmayreflectstock-specificbinariesunrelatedtoSeptemberpanel. Catalystlive0wins/3closedtradesincludes2forcedliquidationsand1gapstop; nonrandomtinyrecorddoesnotestimatepanelwinchance orfillslippagewithouttimestampedquotes. Needcomparablepoint-in-timepanelannouncementsandFIRSTbriefingdates, frozenentry/exit/prospectiveholdout beforeanyeventprobability. StrongcounterexampleGRALcanrallydespitepoorpastprice-rulepayoffs; truebutnotnumericedge. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; fixednote158,tests.json/stats.json, directquery_tradesfundcatalyst. Nohypothesistuningoninspectedholdout.

1. October2025PATHFINDER2initialmatrix TP133,FP83,FN196,TN22749,N23161 givesprevalence1.4205%,PPV61.5741%,episode sensitivity40.4255%,specificity99.6365%. WilsonPPV95%54.9386–67.8050. May2026full173/287=60.2787%,Wilson54.5176–65.7683. Cohortsoverlap; doNOTtreatasindependentdifference-test. 25578initialenrolledvs23161analyzable9.4495%differenceiseligibility/followupgap, notautomaticallyloss-tofollowup. Source https://investors.grail.com/static-files/599b1d07-9463-4a3e-b49f-fa074ab63083 ; https://grail.com/press-releases/grail-presents-pathfinder-2-results-of-more-than-35000-participants-showing-the-galleri-test-substantially-increased-cancer-detection-with-robust-performance-and-favorable-safety-at-2026-a/ .2. Holdinginitialsensitivity/specificityconstant, hypotheticalprevalence.5/1/2% givesPPV35.85/52.90/69.41%. Bayesillustration, notfutureclinicalforecast: patientmix,testversionandfollowupmaychangebothinputs. Specificity99.6%isnotprobabilityapositivepersoniscancerfree. 3. NHS197146pooledroundobservationsTP937/FP864/FN2114/TN193231, NOTuniquepeople; repeatedpeoplemeannaiveindependentbinomialCIwouldunderstateuncertainty. SponsorPPV52%andepisode sensitivity30.7%addressdifferentconditionals. First-roundFP303with54laterdiagnoses(44correctCSO)doesnotretroactivelychange12monthresult. 4. All-screenedPF2invasive213/35335=.6028% differsfromslide17conditional91%cancer/50%no-cancer. Denominatorsnotfullyreconciled, nofalseprecisecount. NHSAEsrecordedcollectiontoreferral, PF2postlockseriousworkupeventshowsascertainmentwindowmatters. Source https://investors.grail.com/static-files/995005e4-2c8e-4a02-bef0-400403856cc7 .5. Primaryp.6234slidesversus.6324releaseunresolvedbutbothnon-significant,IRR1.03same. Currentregistryfixedsequenceconfirmedbutpost-resultversionnotoriginalSAP. Conclusionuncertainutilityandconditionalburdencannotbecompressedintoasingleapprovalscore. Validityisconditionalarithmetic; soundnessboundedbycohortselection,roundeddisclosuresandversiondifferences. Factsarecounts; hypotheticalprevalencesexplicitassumptions.

Premises: net naturalpayoffs from registeredtest have smallsamples and selectedcurrentnames; the two liveSupplywins were operatorliquidations and contain nolossestimate/referencequotes. NumPyseed90504 circular3tradeblocks,10000 draws; TSN95%meanCI[-1.4654,.2683]%,PPC[-1.5547,.7230]%,JBS[-.6432,1.7526]%. Allincludezero; no positive lowerbound forrisk. Blockresampling retainslocaldependence but cannotremove selection/regimes or unknowntails. Descriptive2025+stockdailyOLSvsSPY uses420TSN/PPCdays and308JBSdays: betas.16053/-.04664/.28185; annualarithmeticintercepts-2.817/-7.212/-.659%, not causalalpha or policyeffect. Priorcautious; replicatedprospectivefixedrule with actualvenueexits and positive netconfidence wouldchange it, repeatingheadlinewon't. Validdeductionzeroauthorizedrisk under fundcriteria; soundnessconditional onhistoricaldata. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; observations1723–1737, savedrawdata/code.

1. Oct14 2025-Sep1 2026 next3day open/open overlapping observations N323 perasset. Extreme negative previousday funding per fixed rule yields grossmean BTC-.6221%(28windows),ETH-.0827%(29),SOL-3.5426%(44), versus otherdays-.2193,-.3447,+.0184%. 2. Seven-calendar-day circular blockbootstrap10k draws gives extreme-minus-other95%CI BTC[-2.863,2.351]pp,ETH[-3.003,3.021],SOL[-6.204,-.904]. Returns overlap; event counts are NOT independent trades. 3. OLS forward3 return on previousday funding inbp and prior3day return, intercept; HAC7 SE funding coefficients BTC-.003560(SE.002443),ETH-.004857(.003435),SOL+.003588(.001685) fractionalreturn perbp. SOL positive coefficient means more negative funding associates with weaker subsequentreturns; association, not causal identification or validatedshort. 4. ExploratorySMA50 strata show BTC/ETH uptrend extreme means+1.52/+1.51% on6/9windows; SOL-2.12% on10. These inspectedholdout strata cannot justify adding a trendfilter and calling it OOS. 5. Net naturaltrade circular3trade intervals BTC[-2.133,1.622]%,ETH[-3.370,1.723],SOL[-4.589,-.656] reinforce zero sizing. Currentuniverse, multiple comparisons, selectedregimes and datarevisions limit inference. Conclusion reject fixedrebound rule; Solana evidence favors continuedweakness in this period, not universal prediction. Prior unknown edge revised against this rule; independent preregistered prospective positive net mean with positive uncertainty bound and exact exits would change sizing, a freshnegative quote would not. Sources https://fapi.binance.com/fapi/v1/fundingRate ; https://runway.adityaperswal.workers.dev/data/crypto-bars ; rawmodels androws saved.

Fixed test113 yields training mean+.297502% (N7), bootstrap95%[-.740855,1.462600]%; holdout mean-.853688% (N5), interval[-2.351641,.378149]%. Python NumPy seed20260905,20000 IID resamples per partition, train then holdout. Very small selected event histories; intervals conditional, not robust population guarantees. Pooled12-event OLS of net Copart window return against matched SPY open return: intercept-.222577% per event, beta.613775, Pearson.258024; descriptive and not causal alpha. Existing selected fund0/3live record includes forced liquidations and one gap stop, so cannot replace event calibration or estimate order slippage without reference quotes. Holdout win40%,W.633108%,L1.844885% implies negative plug-in expectancy and clipped quarter-Kelly0; even training positive Kelly is unstable and above capital cap. Prior cautious unchanged: independent positive executable prospective mean evidence would change it; more repetitions of current calendar dates would not. Counterexample a broader complete event sample could differ, so reject this implementation, not the universal anomaly. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; issuer notices in observations1208–1220; payoff observations1221–1226. Actual capital-return denominator independently reconciled.

Benchmark falsification, not a new optimized strategy. Buy all19 preregistered country ETFs equally at the partition's first open, charge5bp fee+5bp slippage, drift holdings thereafter, mark daily adjusted closes with no terminal sale. Training2017-2022 +31.9001%,DD39.1696%; holdout2023-Sep4 +101.8003%,DD14.0588%. The monthly rotation +113.2853% uses higher risk: DD19.2653%, daily OLS beta1.23544 to this country buyhold and annual arithmetic intercept-2.07476%. SPY-based +3.00% intercept in analysis98 therefore does not establish selection alpha; benchmark choice materially changes interpretation. Premise1: global equity exposure itself earned substantial returns. Premise2: turnover and concentration add risk. Conclusion: profitable monthly evidence remains WATCH but is not proven superior to passive exposure. Validity is descriptive; no significance or causal claim without residual/dependence inference. The current country sample survives to today and is not a point-in-time comprehensive index universe. Formula q_s=(10000/19)/(first_open_s*1.0005^2), daily value=sum(q_s*close_s); OLS with intercept on matched daily returns. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; artifacts/benchmark.json.

1. Registered holdout 2023-Sep4 2026 produces 130 completed monthly trades and 133 range trades, but simultaneous countries are dependent. Group by entry month: 44 and 45 clusters. NumPy seed90502, 10,000 independent month-cluster resamples gives mean-return 95% intervals monthly [0.21096,3.50982]% and range [-0.56978,0.72463]%. 2. Stronger serial-dependence sensitivity, seed90503, stationary blocks averaging three entry months, 10,000 samples: monthly [0.35802,3.46067]%, range [-0.51449,0.60859]%. Positive fractions .9938/.6264 are bootstrap frequencies, not current win probabilities. 3. Daily OLS against SPY: monthly beta .93818, annual arithmetic intercept 3.00129%, R2 .38703; range .06893, -.53141%, R2 .01884. Descriptive estimates do not establish causal alpha. EWY/EWT correlation .76154; EWY/EWP .47974, EWT/EWP .52041. 4. Monthly annual returns 2023 +14.83%,2024 -8.26%,2025 +40.23%,2026-to-date +44.37%; range +3.25,-2.08,+6.00,-4.38%. Conclusion: distinguish positive monthly historical evidence from failed range implementation; do not reject all momentum or use monthly payoff to size range exits. Soundness bounded by current surviving wrappers, one unusually strong recent chip cycle, historical rather than prospective holdout, and monthly dependence beyond resampling assumptions. Lower mean bound would have to remain positive on independent implementation-compatible outcomes to reconsider range; presently size0. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; scripts artifacts/stats.py,monthblock.py; backtests70-77.

Premise1:419complete common daily return rows2025-Sep4 from all-adjustedIEX; OLS versusSPY beta AXTI2.511,AEHR3.447,PSIX2.892,VSH2.387,POWL1.899,ETN1.539. POWL/ETNcorr.6593,AEHR/VSH.5628; distinct products do not eliminate shared market exposure. Intercepts descriptive, not causalalpha; no annualized alpha claim. Premise2: 10000 circular3-trade block mean bootstrapseed905 on naturalexit samples40/39/29/38/34/32; five means have95%interval crossingzero. PSIXmean3.7507%,CI[.4832,7.1587]%, but only4trainingbars; this is selected postlisting history. Additional20000draws conservative six-test tail sensitivity[.025/6,1-.025/6] gives[-.6271,8.3548]%; cannot correct all3012-name screening/unknown prior research and is not a formal population guarantee. RemovebestPSIXtrade mean2.5700%; not entirely one outcome, still smallsampleandregime risk. Premise3: contemporary survivorship, selected highmovers, assumed costs, conditionalbarrierfills and missing natural ledger losses undermine transportability. Conclusion PSIX merits prospective collection, not size; others rejected/watch perbacktests83-88. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; stats.json,tests.json,derived.json/trades archived. Prior on PSIX shifts fromunknown tointerestingdiagnostic, never adopted; genuinely independent netexecution outcomes would change verdict.

Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; Jan1-Sep4 2026 net natural-exit samples XRP6,LINK14,UNI13,PAXG8. Numpy20,000 IID resamples seed2026090502: mean trade95% intervals XRP[-1.194,18.765]%,LINK[-1.726,2.184]%,UNI[-2.933,2.634]%,PAXG[-1.402,.419]%. IID diagnostic assumes independent trades; current-universe selection and regimes limit interpretation. Positive-resample frequencies94.375/51.585/38.65/13.8% are not posterior probabilities. Daily OLS onBTC with intercept: betas1.237/1.176/1.181/.299, correlations.871/.874/.677/.446; sample247 each exceptPAXG204. Similar XRP/LINKbeta weakens diversification narrative. Rawbinarymean-payoff quarterKelly notional formula .25*(p/loss-(1-p)/win) gives3.889/.281/-.468/-7.550 capitalfractions, clearly unstable and constrained beforecaps. Lowermeanboundsnegative => authorizedallocation0. PremisespositiveXRPreturn andrate66.67% wins depend on6trades; strongestbullcase largepayoff, but no precise populationedge established. Need prospective fixedrule evidence and testedvenueexits; do not pickbestlookback frominspectedholdout. UNI terminalopenmark excludedfrompayoffs rather than countedwinner.

Data https://runway.adityaperswal.workers.dev/data/crypto-bars ; Jan1-Sep4 2026 daily returns. BTC/ETH correlation0.9124, BTC/SOL0.8805, ETH/SOL0.8801; OLS beta via covariance/variance ETH1.254,SOL1.202. Daily SD BTC2.452%,ETH3.370%,SOL3.347%; worst close-close -13.986/-14.890/-14.931%. Premise1: three major longs are strongly correlated risk, not three independent bets. Premise2: registered net daily closed-trade bootstrap (10000 IID resamples, seed20260905) mean95% intervals BTC[-1.424,1.323]%,ETH[-2.670,.678]%,SOL[-1.213,6.606]%. SOL N12, win50%,mean win6.123%,mean loss2.270%; naïve quarter-Kelly .07866, but authorized size0 because interval crosses0, history selected, prospective validation absent and venue exits untested. Bootstrap-positive fraction .7897 is resampling stability, NOT probability strategy profitable. IID assumption ignores serial dependence, so intervals may understate uncertainty. All hourly20 net mean intervals entirely negative: BTC[-.611,-.443],ETH[-.609,-.346],SOL[-.682,-.465]%. Strong counterexample: SOL daily may have real positive momentum edge despite inconclusive sampling; insufficient confidence does not prove negative return. Goal repeatable net gains, binding constraint independent payoff validation; do not convert raw Kelly into dollars. Methods/results artifacts return-statistics.json,bootstrap.json and risk_analysis.py; no predictive regression claimed.

Premise1: all-adjusted daily returns Jan2025-Sep42026,420 paired rows, give FRO/STNG Pearson0.7506; OLS stock return on SPY plus intercept slopes0.6291/0.4025. Low market beta does not mean low shared tanker exposure. Premise2: corrected natural-exit samples41/38 trades, FRO mean0.53025%, STNG0.37899%, differ from two operator liquidations which cannot calibrate average loss. Premise3:10000 circular moving-block resamples of3 consecutive trades, seed501, yield95% percentile mean intervals FRO[-0.87436,1.95854]%, STNG[-0.75304,1.45383]%. Resample means positive76.17%/74.69%; these are resampling frequencies, not Bayesian probabilities that current trades win. Conclusion: evidence cannot robustly separate positive mean from zero; original two-ticker diversification is overstated. Valid inference only for selected histories; current two-name survivor selection, earlier historical inspection, small N, overlapping market exposures and possible regime shifts limit soundness. Bootstrap cannot invent missing tail events or audit dividend/15-minute fills. Prior neutral-to-cautious unchanged. Independent prospective net outcomes with a positive lower mean interval and actual execution/corporate-action handling would change eligibility; a lower bound<=0 keeps policy size0. No multiple-test corrected significance claimed. Source https://runway.adityaperswal.workers.dev/data/stock-bars retrieved Sep5; code archived notes97-99; observations110/111 superseded for FRO payoff estimates by198; backtests36/37 contain final figures.

Descriptive OLS on18 matched9-session event returns: netCHWY=alpha+beta*rawSPY; alpha-4.33%,beta3.02,correlation.848,R2.720. Holdout meanCHWY minusSPY-.611%. No regression filter fitted or used for trading; full-sample descriptive relationship could be driven by market regimes/outliers,not a causal catalyst premium.1. Event study selected one surviving issuer and announced dates rather than full market.2. Six OOS events have mean-.153%,95%IID bootstrap[-4.53,3.69]%; uncertainty not resolved.3. Own3tradeledger0wins,meanloss16.38%,total-31.94,oneRAREstop and2operator exits; meanwin undefined,forced exits censor strategy horizons. Conclusion no numerical positive edge supports size. Validity: no qualifying evidence; soundness restricted by small and selected samples. Strong countercase future events may rally and exogenous liquidation could hide wins; neither is observed payoffs. Source https://runway.adityaperswal.workers.dev/data/stock-bars and query_trades fund=catalyst readSep5; full per-event issuerURLs and returns saved. Rate/total distinction: mean-.153% perholdoutwindow produces-1.72% compounded across6,not six times mean because volatility drag. Priorcashunchanged; independently preregistered prospective positive net sample with implementable exits would change eligibility,not monthlyshortfall. Gap loss was-6.44R versus intendedstop,so normalizedstop-risk alone is misleading.

SIP-derived2023-2026-09-04 fixed10month returns, source https://runway.adityaperswal.workers.dev/data/stock-bars.18closed trades average+2.3818%,IID bootstrap20000 seed20260905 mean95%CI[-1.1377,6.3696]%;10.375% of draws nonpositive. Trades overlap across correlated markets, so IID interval is only diagnostic. Remove best+24.5168%trade and mean falls to+1.0797%. Joint stationary block bootstrap expected20sessions,3000 full-horizon paths gives annual arithmetic strategy meanCI[1.8463,13.6657]%,active versus benchmark[-9.7113,-0.7710]%. This is conditional resampling of one history, not a population confidence guarantee. Daily OLS versus equal-initial-sleeve buy-hold beta0.6631, annualized intercept-0.7287%,R2.7887; descriptive no causal alpha inference. SPY/EFA daily correlation.7598; SPY/IEF.0528 in test versus.1931 in2022. Counterexample: observed lower drawdown can follow reduced exposure alone; a different crisis may break diversification. Validity: positive historical net returns support further study but not calibrated venue edge. Soundness limited by18closed trades,4open omitted from payoff statistics, one selected asset-class wrapper set and missing2008. Preserve10month rule; do not select12months after looking. Policy allocation0.

On the seven qualified events with 20-day outcomes, OLS of SPY-adjusted 20-day return on opening gap and log(1+EPS surprise) gives gap coefficient -1.45, surprise coefficient +0.0083, R-squared 49.7%; simple gap/return correlation is -0.703 and log-surprise/return correlation -0.155. These estimates are descriptive only: N=7, heteroskedastic, a single HPE outcome is influential, universe chosen from current news/candidates, and adjusted-versus-GAAP EPS definitions conflict (e.g. GTLB). Counterexample: DELL Feb continued +28.0% over 20d after a 13.1% gap; HPE reversed -28.1% after 33.7%. Conclusion: neither larger gap nor larger headline surprise supports stronger expected drift here; if anything the gap may flag exhaustion. No sizing edge.