Stats

Vectorbt: long when close exceeds prior20calendar-day CLOSE maximum, exit when false, following exact OPEN; hourly480bars. Train2024–25 (SOLstartsAug26), cashresetJan1, known2026holdout throughSep12; no claim of untouched new OOS. Fees25bp eachside creditedasset equivalentbuyfee .0025/.9975, slippage10bp/30bp stress. DailyOHLC from24completehours; missingcalendarwindows remainunavailable, no filling. Baseline train return-69.973193%,DD70.504557%,N197,win17.766%; holdout return-20.782205%,DD25.913218%,N45,win13.333%,open0;stress-33.831787%. HIGHthreshold control holdout-11.419497%N21. All15/20/30day xCLOSE/HIGH xcostcells recorded26572–83,knownholdoutrange-38.410to-9.158%;none selected. Premise: same-close grossCLI gains are not evidence for costed next-open rule. Counterexample daily SOL remainspositive; hourly disagreement and missing required3R/partial/15minpolicy prevent adoption. This short-hold toggle is NOT prior10LOWexit Donchian or managedETHpullback. Conclusion reject as live replacement. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Vectorbt: long when close exceeds prior20calendar-day CLOSE maximum, exit when false, following exact OPEN; hourly480bars. Train2024–25 (SOLstartsAug26), cashresetJan1, known2026holdout throughSep12; no claim of untouched new OOS. Fees25bp eachside creditedasset equivalentbuyfee .0025/.9975, slippage10bp/30bp stress. DailyOHLC from24completehours; missingcalendarwindows remainunavailable, no filling. Baseline train return5.185949%,DD31.646448%,N41,win29.268%; holdout return-21.109976%,DD29.758847%,N13,win15.385%,open1;stress-25.256777%. HIGHthreshold control holdout-16.281467%N8. All15/20/30day xCLOSE/HIGH xcostcells recorded26572–83,knownholdoutrange-25.257to-7.238%;none selected. Premise: same-close grossCLI gains are not evidence for costed next-open rule. Counterexample daily SOL remainspositive; hourly disagreement and missing required3R/partial/15minpolicy prevent adoption. This short-hold toggle is NOT prior10LOWexit Donchian or managedETHpullback. Conclusion reject as live replacement. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Vectorbt: long when close exceeds prior20calendar-day CLOSE maximum, exit when false, following exact OPEN; hourly480bars. Train2024–25 (SOLstartsAug26), cashresetJan1, known2026holdout throughSep12; no claim of untouched new OOS. Fees25bp eachside creditedasset equivalentbuyfee .0025/.9975, slippage10bp/30bp stress. DailyOHLC from24completehours; missingcalendarwindows remainunavailable, no filling. Baseline train return-78.978048%,DD78.978048%,N208,win14.423%; holdout return-25.349461%,DD26.444395%,N57,win12.281%,open0;stress-40.568953%. HIGHthreshold control holdout-12.189031%N28. All15/20/30day xCLOSE/HIGH xcostcells recorded26572–83,knownholdoutrange-40.569to-8.841%;none selected. Premise: same-close grossCLI gains are not evidence for costed next-open rule. Counterexample daily SOL remainspositive; hourly disagreement and missing required3R/partial/15minpolicy prevent adoption. This short-hold toggle is NOT prior10LOWexit Donchian or managedETHpullback. Conclusion reject as live replacement. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Vectorbt: long when close exceeds prior20calendar-day CLOSE maximum, exit when false, following exact OPEN; hourly480bars. Train2024–25 (SOLstartsAug26), cashresetJan1, known2026holdout throughSep12; no claim of untouched new OOS. Fees25bp eachside creditedasset equivalentbuyfee .0025/.9975, slippage10bp/30bp stress. DailyOHLC from24completehours; missingcalendarwindows remainunavailable, no filling. Baseline train return-2.113871%,DD26.189310%,N45,win28.889%; holdout return-10.194852%,DD17.371016%,N18,win22.222%,open0;stress-16.433561%. HIGHthreshold control holdout5.123294%N8. All15/20/30day xCLOSE/HIGH xcostcells recorded26572–83,knownholdoutrange-17.855to5.689%;none selected. Premise: same-close grossCLI gains are not evidence for costed next-open rule. Counterexample daily SOL remainspositive; hourly disagreement and missing required3R/partial/15minpolicy prevent adoption. This short-hold toggle is NOT prior10LOWexit Donchian or managedETHpullback. Conclusion reject as live replacement. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

RequiredCLI node /app/agent/dist/backtest.js --symbols CNQ --strategy breakout --lookback20 --days2100 --stop-pct6 --target-pct18 yielded1439IEXdailybars,106trades,return23.1053%,DD21.2717%,win39.6226%,buyhold104.1599%. CLI uses signal exits and no explicit fee/slippage arguments; not samefrozen3R strategy or heldouttest, therefore not sufficient for trade. Independent vectorbt audit uses SIPall and explicitcosts. Sensitivity10/20/40: trainingbase.5375/15.6616/4.6417%; holdout20.9743/15.3639/-7.8131%. Stress15bps train-6.8202/8.9260/-.8591%; holdout17.7518/12.7409/-9.7288%. L40negativeheldout and L10weaktraining indicate parameterfragility. KeepL20frozen, do not choosebetterheldout10. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; scripts and cachedbars in research/ availablethisworkspace, persistence via recordedrecipe/manifest.

Preregistered2804 before results: prior20close breakout AND SMA50; nextopen entry, baseline5bpentry-2ADR20stop/+6ADRtarget,max40sessions. 2025-01-01 to 2026-09-11. SIPadjustmentall with warmup2020; dividend-adjusted proxy, not raw cash entitlement. Vectorbt longonly cash portfolio with1bpfees/5bpslippageeachside; open/closemarks, entrydayexits, conservative stopfirst dailyambiguity, gapopenexit. 15bp stress fixes baseline barriers and sameentry/exitdates; earlier path-changingdiagnostic2808 superseded. Reportreturn includes markedopenposition wherepresent; winrate closedtradesonly. Manual netcash/trade-denominatorreconciliation error<4e-16. No +2Rquarter or15minute/off-hours validation, no corporate-action execution claim. Main L20 positive but smallsample; L10/L40 sensitivity not selected. Source https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=CNQ&timeframe=1Day&feed=sip&adjustment=all . No order; current50.07<51.33prior20.

Preregistered2804 before results: prior20close breakout AND SMA50; nextopen entry, baseline5bpentry-2ADR20stop/+6ADRtarget,max40sessions. 2021-01-01 to 2024-12-31. SIPadjustmentall with warmup2020; dividend-adjusted proxy, not raw cash entitlement. Vectorbt longonly cash portfolio with1bpfees/5bpslippageeachside; open/closemarks, entrydayexits, conservative stopfirst dailyambiguity, gapopenexit. 15bp stress fixes baseline barriers and sameentry/exitdates; earlier path-changingdiagnostic2808 superseded. Reportreturn includes markedopenposition wherepresent; winrate closedtradesonly. Manual netcash/trade-denominatorreconciliation error<4e-16. No +2Rquarter or15minute/off-hours validation, no corporate-action execution claim. Main L20 positive but smallsample; L10/L40 sensitivity not selected. Source https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=CNQ&timeframe=1Day&feed=sip&adjustment=all . No order; current50.07<51.33prior20.

Required CLI run: breakout/lookback20/days981 BTC return46.460751%,DD14.692428%,win34.328358%,N67;ETH41.521926%,24.943256%,31.481481%,N54;SOL72.151727%,18.795394%,52.777778%,N36. Same-close prior20CLOSE policy, gross/nofees/noholdout, partialSep13day included: diagnostic, not an executable HIGH/pullback policy. Frozen managedETH studies2883/2884/2897/2899 retained unchanged: train42.312262%,DD37.521903%,7/13wins;known2026holdout19.206747%,DD14.542745%,3/3wins,1open,stress17.549943%;dailycontrol-1.925730%N2. Prior20HIGH thenhalfATRpullbackwithin3days,next exactopen,2ATRstop,20daytimeout;25bpfees/10bpslip,15minpointchecks,quarterat2R thenstopentry,residualtargetnet3R. Daily disagreement andtrainingmeanCI[-3.77,12.80]% limitconfidence. Missing15minbounds retained from prioraudit, notnewquoteproof. ETHSep11 22UTCentry passed/noarm;size0. This record retries two500failedwrites; neitherfailure counted asrecord. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Edge test: corrected10 consecutive issuer-dated AMC earnings Mar2024–Jun2026, daily SIP release-close to next-session close, median absolute16.578014%. Sep11S2.95, Oct16P2.5 .12Sep11/C2.5 .50Sep10 plus assumed .05/share costs require -21.016949%/+3.389831%, covered2/10(20%)/4/10(40%). Callpassescoveragebut50USDpremium exceeds42.5495 convictionrisk; even45USDintrinsicexceedscap. Putfails40%; no documentedexpectedactualoptionpayoff>=3xpremium. IssuerSep2 quote 'after market close on Wednesday, September 23, 2026' https://investors.stitchfix.com/news-events/press-releases/news-details/2026/Stitch-Fix-Announces-Date-for-Fourth-Quarter-and-Full-Fiscal-Year-2026-Financial-Results-and-Conference-Call/default.aspx . Oct16expiry23dayslater; defaultSep22RTHexit. Sources option_contracts and https://runway.adityaperswal.workers.dev/data/stock-bars /data/option-bars; historicalsources recorded26537–46. OriginalN8preserved. 1. RequiredCLI momentum20/days1000:685bars,32trades,+41.383224%,DD55.802626%,win28.125%; unsplitstockdiagnosticonly. 2. Vectorbt fixedlag2/3/5openentry,releaseT-1closeexit;fees10bp/side,slippage10/30bp/side. Restoredmissingquartersintraining; preserve2026Mar/Junlast2holdout. Trainingbestlag5+14.174754%,DD14.475874%,win37.5%,N8; holdout+9.472964%,DD6.811989%,win100%,N2; stress+8.600670%. All12gridrows26549. 3. AlreadyseenholdoutcannotbeclaimedfreshOOS; positivecontrolcannotvalidateoptionpremiumdecay,spread,3Rtargetor15minmanagement. NooptionPnLbacktestavailable. Nochoosewinninglag2testafterseeingit. Wouldneedprospectiveoptionsseriesbeforetrustingoptionrunupconversion.

Reran required CLI: node /app/agent/dist/backtest.js --symbols FRO --strategy breakout --lookback20 --days2080 --stop-pct10 --target-pct30 (actual flags separated by spaces). Returned1426 IEX daily bars,117 trades. CLI includes signal exits, percentage barriers, no independent holdout or explicit cost model. This differs from L20/SMA50/2ADR/40-session SIP policy; its negative result cannot be substituted into that policy or hidden because another result is positive. No order based on CLI. Source https://runway.adityaperswal.workers.dev/data/stock-bars .

Frozen specification note2629, clarification2631. Raw SIP daily close exceeds prior20 closes and SMA50; enter next scheduled NYSE09:30 15-minute OPEN, stop2 prior ADR20, target3R, max40 sessions. Barriers use actual slipped entry. Sample subsequent RTH15-minute OPEN prices; stop before target then quarter at2R with breakeven effective for subsequent samples. Exit at observed sampled price plus slippage, never intrabar extreme; final scheduled close-minus15-minute sample on session40. Full-equity fractional long-only vectorbt orders reconcile manual cash within1e-9. Fee1bp/side, slip5bp and15bp stress. Calendar handles early closes. Zero dividend credit and static raw barriers: PRICE-only account sensitivity, not total shareholder return. 16 closed trades plus1 open mark; win rate excludes open trade. Training contains one trade with missing entry-day samples (Mar8,2022); day low9.39 stayed above stop8.96647 and high9.76 below2R11.30167, so envelope rules out missed exit/partial triggers; drawdown sampling remains incomplete. No holdout missing scheduled bars. Historical dates already viewed for other rules; this is chronological holdout for the frozen implementation, not globally unseen evidence. Costs alter barriers and can alter dates, explaining non-monotone variant behavior. Source https://runway.adityaperswal.workers.dev/data/stock-bars . Positive results advance execution watch, not a forecast for the forthcoming6.93% distribution.

Frozen specification note2629, clarification2631. Raw SIP daily close exceeds prior20 closes and SMA50; enter next scheduled NYSE09:30 15-minute OPEN, stop2 prior ADR20, target3R, max40 sessions. Barriers use actual slipped entry. Sample subsequent RTH15-minute OPEN prices; stop before target then quarter at2R with breakeven effective for subsequent samples. Exit at observed sampled price plus slippage, never intrabar extreme; final scheduled close-minus15-minute sample on session40. Full-equity fractional long-only vectorbt orders reconcile manual cash within1e-9. Fee1bp/side, slip5bp and15bp stress. Calendar handles early closes. Zero dividend credit and static raw barriers: PRICE-only account sensitivity, not total shareholder return. 31 closed trades plus0 open mark; win rate excludes open trade. Training contains one trade with missing entry-day samples (Mar8,2022); day low9.39 stayed above stop8.96647 and high9.76 below2R11.30167, so envelope rules out missed exit/partial triggers; drawdown sampling remains incomplete. No holdout missing scheduled bars. Historical dates already viewed for other rules; this is chronological holdout for the frozen implementation, not globally unseen evidence. Costs alter barriers and can alter dates, explaining non-monotone variant behavior. Source https://runway.adityaperswal.workers.dev/data/stock-bars . Positive results advance execution watch, not a forecast for the forthcoming6.93% distribution.

Frozen specification note2629, clarification2631. Raw SIP daily close exceeds prior20 closes and SMA50; enter next scheduled NYSE09:30 15-minute OPEN, stop2 prior ADR20, target3R, max40 sessions. Barriers use actual slipped entry. Sample subsequent RTH15-minute OPEN prices; stop before target then no partial management. Exit at observed sampled price plus slippage, never intrabar extreme; final scheduled close-minus15-minute sample on session40. Full-equity fractional long-only vectorbt orders reconcile manual cash within1e-9. Fee1bp/side, slip5bp and15bp stress. Calendar handles early closes. Zero dividend credit and static raw barriers: PRICE-only account sensitivity, not total shareholder return. 16 closed trades plus1 open mark; win rate excludes open trade. Training contains one trade with missing entry-day samples (Mar8,2022); day low9.39 stayed above stop8.96647 and high9.76 below2R11.30167, so envelope rules out missed exit/partial triggers; drawdown sampling remains incomplete. No holdout missing scheduled bars. Historical dates already viewed for other rules; this is chronological holdout for the frozen implementation, not globally unseen evidence. Costs alter barriers and can alter dates, explaining non-monotone variant behavior. Source https://runway.adityaperswal.workers.dev/data/stock-bars . Positive results advance execution watch, not a forecast for the forthcoming6.93% distribution.

Frozen specification note2629, clarification2631. Raw SIP daily close exceeds prior20 closes and SMA50; enter next scheduled NYSE09:30 15-minute OPEN, stop2 prior ADR20, target3R, max40 sessions. Barriers use actual slipped entry. Sample subsequent RTH15-minute OPEN prices; stop before target then no partial management. Exit at observed sampled price plus slippage, never intrabar extreme; final scheduled close-minus15-minute sample on session40. Full-equity fractional long-only vectorbt orders reconcile manual cash within1e-9. Fee1bp/side, slip5bp and15bp stress. Calendar handles early closes. Zero dividend credit and static raw barriers: PRICE-only account sensitivity, not total shareholder return. 31 closed trades plus0 open mark; win rate excludes open trade. Training contains one trade with missing entry-day samples (Mar8,2022); day low9.39 stayed above stop8.96647 and high9.76 below2R11.30167, so envelope rules out missed exit/partial triggers; drawdown sampling remains incomplete. No holdout missing scheduled bars. Historical dates already viewed for other rules; this is chronological holdout for the frozen implementation, not globally unseen evidence. Costs alter barriers and can alter dates, explaining non-monotone variant behavior. Source https://runway.adityaperswal.workers.dev/data/stock-bars . Positive results advance execution watch, not a forecast for the forthcoming6.93% distribution.

Edge test: nine issuer-dated AMC earnings Feb2024–Jun2026, SIP daily release-close to next-session close; median absolute5.543964%, mean-0.908716%. Sep11 S39.02, Oct16 C45 .30/P35 .40 catalog closes plus assumed.05/share roundtrip reserve require+16.222450%/-11.455664%, covered0/9 and1/9 (11.11%), below40%. Without reserve put2/9. No documented expected actual-option payout>=3xpremium. Oct16expiry31daysafterissuerSep2 schedule: 'Tuesday, September 15, 2026, U.S. Time, after the market closes.' https://investors.trip.com/node/15681/pdf ; defaultSep14RTHexit. Hurdles exceedmedian10.678486/5.911700pp, noholdexception. Sources option_contracts and https://runway.adityaperswal.workers.dev/data/stock-bars . CorrectedN9 includesMay19 2025; August26 2024hourunverified sensitivityN10median6.067764%,0/10C1/10P. OldN8/S38.7/.32C/.48Pforecast retained separately. 1. Vectorbt24prespecifiedcells:lags2/3/5sessions,directionslong/short,first7training,last2chronologicalknownholdout,10bpfeesandslippage10or30bp eachside. Entryopenatdayindex-lag,exitprior-sessionclose; markdailyclose,fullcashstockproxy, shortsresearchonly. Theseareunderlyingcontrols, notallowedshortordersoroptionPnL. Actualoptionhistory,IV/spreads,15minmanagementnotreproduced. 2. BaseT3shorttrain+2.471362%,DD4.871260%,win57.142857%,N7;holdout-.282352%,DD2.955683%,win50%,N2. T2shorttrain+2.337936%,holdout-1.199408%. T5shorttrain+1.717228%,holdout+6.670505%N2, butknownholdoutalreadyseenin24034; cannotselectT5asfreshOOSwinner. Longsalltrainnegative;T3longtrain-8.144169%/holdout-1.318695%. All24rowsrecordedobservations26450–26455approximately; exactIDsavailableviaquery. 3. RequiredCLIdiagnostic node /app/agent/dist/backtest.js --symbols TCOM --strategy momentum --lookback20 --days365 succeeded:251bars,7trades,return-14.664032%,DD21.044477%,win42.857143%,buyhold-47.183861%. Fullwindowdiagnosticnotheldoutstrategyvalidation. Conclusion noentry; preservefrozenlast2holdout,notretuneonmonthlyshortfall.

Reuse audited C91204 experiments2883/2884/2897/2899, not a rerun or new independent sample. Policy: prior20dayHIGH breakout then halfATR pullback within3days, next exact open,2ATR20 stop,20day timeout;25bp credited-coin entry fee and USD exit fee,10bp slip eachside/30bp stress. Point-check15minute exit proxy:25% sale at gross+2R, remainingstopentry, residual target solved for net3R. Hourly train2024–25 return42.312262%,DD37.521903%,win7/13,oneopen; known2026holdout19.206747%,DD14.542745%,3/3wins,oneopen,stress17.549943%. Daily coherent control training-27.126920%;holdout-1.925730%N2/stress-2.707198%. Eighteen missing15min checks inside hourly holdout trades are bounded away from thresholds by larger bars; retain modeled gains, not literal executablequote validation. Required CLI prior2942 is prior20CLOSE/sameclose/noexplicitcost/partialdaily diagnostic, not this policy. Premise1 favorable hourly outcome is a lead; premise2 daily disagreement, known holdout and training trade-meanCI[-3.77,12.80]% limit confidence; premise3 latest new entry absent. Conclusion no new order; preserve candidate without selecting new parameters or inflating N. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Command: node /app/agent/dist/backtest.js --symbols FRO,STNG --strategy breakout --lookback 20 --days 2080 --stop-pct 10 --target-pct 30. CLI returned1426IEXbars and uses signal-driven exits as well as percentage stops/targets. No explicit execution costs or independent holdout. It is a diagnostic, not a replication of frozen SMA50/2ADR3R/40session SIP policy, and cannot adjudicate its validation. Source https://runway.adityaperswal.workers.dev/data/stock-bars . Both tools' results retained rather than selecting the positive implementation; method differences explain disagreement.

Ran required CLI --strategy breakout --lookback20 --days980 onBTC/USD,ETH/USD,SOL/USD. BTCreturn46.460751%,DD14.692428%,win34.328358%,N67;ETH40.761782%,DD24.943256%,win30.909091%,N55;SOL72.151727%,DD18.795394%,win52.777778%,N36. Inspected strategies.ts/engine.ts:prior20CLOSE threshold,samecloseentry,noexplicitfees/noheldoutwindow,currentpartialdailybarincluded;SOL748barsversus981majors. Premise gross performance is diagnostic only;counterexample ETHcostedHIGH/nextopen2026daily-4.8672%,hourly-.4895% and these also lackmanaged3R. Conclusion no cross-policy borrowing of CLIreturnsfororders. Source https://runway.adityaperswal.workers.dev/data/crypto-bars . CLIandrawstatssavedinobservations;notadoption.

Command: node /app/agent/dist/backtest.js --symbols FRO,STNG --strategy breakout --lookback 20 --days 2080 --stop-pct 10 --target-pct 30. CLI returned1426IEXbars and uses signal-driven exits as well as percentage stops/targets. No explicit execution costs or independent holdout. It is a diagnostic, not a replication of frozen SMA50/2ADR3R/40session SIP policy, and cannot adjudicate its validation. Source https://runway.adityaperswal.workers.dev/data/stock-bars . Both tools' results retained rather than selecting the positive implementation; method differences explain disagreement.

Premise1 preregistered prior20calendar-dayHIGH breakout,prior10dayLOWexit,fillfollowingOPEN,longcashonly,fees25bp/slip10bp eachside,stress30bp. Hourly uses480HIGH/240LOW calendarhours. No stops/targets/partial management: diagnostic comparator,not user's executable3Rpolicy or previousETHpullbackpolicy. Train2024-2025 (SOL starts2024-08-26),resetcashJan1,known2026holdoutthroughSep1206UTC;2026alreadyresearched,not untouchednewOOS. Train return9.334791%,DD36.795757%,N8,win50%,meanCI[-4.938225,8.002702]%. Holdoutmarked return1.322187%,DD24.525746%,closedN4,win0%,mean-5.897838%,CI[-7.941673,-3.062724]%,openN1;stressreturn-0.484906%. Premise2 all9lookback15/20/30 x SMA0/100/200cells retained;holdoutbasecost range-2.935467to19.188094%. MeanCI5000IIDtradebootstrap,smallNandmultiplecomparisonsunadjusted. Counterexample: markedopenprofitcanmaskclosedlosses;strictrollingSMAbecomesunavailableaftergaps,N0isnotvalidation. Conclusion no winningparameterchosenfromholdout and no currentlateentryauthorized. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ;fees https://docs.alpaca.markets/us/docs/crypto-trading . All216cellfiguresobservations26302–26325.

Premise1 preregistered prior20calendar-dayHIGH breakout,prior10dayLOWexit,fillfollowingOPEN,longcashonly,fees25bp/slip10bp eachside,stress30bp. Daily OHLC reconstructed only from24completehours. No stops/targets/partial management: diagnostic comparator,not user's executable3Rpolicy or previousETHpullbackpolicy. Train2024-2025 (SOL starts2024-08-26),resetcashJan1,known2026holdoutthroughSep1200UTC;2026alreadyresearched,not untouchednewOOS. Train return-10.120708%,DD42.459529%,N7,win42.857143%,meanCI[-8.221019,5.529944]%. Holdoutmarked return4.581108%,DD17.406778%,closedN3,win0%,mean-4.373566%,CI[-5.151662,-3.875026]%,openN1;stressreturn3.12758%. Premise2 all9lookback15/20/30 x SMA0/100/200cells retained;holdoutbasecost range0to19.561908%. MeanCI5000IIDtradebootstrap,smallNandmultiplecomparisonsunadjusted. Counterexample: markedopenprofitcanmaskclosedlosses;strictrollingSMAbecomesunavailableaftergaps,N0isnotvalidation. Conclusion no winningparameterchosenfromholdout and no currentlateentryauthorized. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ;fees https://docs.alpaca.markets/us/docs/crypto-trading . All216cellfiguresobservations26302–26325.

Premise1 preregistered prior20calendar-dayHIGH breakout,prior10dayLOWexit,fillfollowingOPEN,longcashonly,fees25bp/slip10bp eachside,stress30bp. Hourly uses480HIGH/240LOW calendarhours. No stops/targets/partial management: diagnostic comparator,not user's executable3Rpolicy or previousETHpullbackpolicy. Train2024-2025 (SOL starts2024-08-26),resetcashJan1,known2026holdoutthroughSep1206UTC;2026alreadyresearched,not untouchednewOOS. Train return16.493028%,DD41.889093%,N14,win42.857143%,meanCI[-6.826896,12.704494]%. Holdoutmarked return-0.489511%,DD22.644888%,closedN5,win40%,mean1.326191%,CI[-5.247297,10.986248]%,openN1;stressreturn-2.654456%. Premise2 all9lookback15/20/30 x SMA0/100/200cells retained;holdoutbasecost range-12.836788to13.949464%. MeanCI5000IIDtradebootstrap,smallNandmultiplecomparisonsunadjusted. Counterexample: markedopenprofitcanmaskclosedlosses;strictrollingSMAbecomesunavailableaftergaps,N0isnotvalidation. Conclusion no winningparameterchosenfromholdout and no currentlateentryauthorized. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ;fees https://docs.alpaca.markets/us/docs/crypto-trading . All216cellfiguresobservations26302–26325.

Premise1 preregistered prior20calendar-dayHIGH breakout,prior10dayLOWexit,fillfollowingOPEN,longcashonly,fees25bp/slip10bp eachside,stress30bp. Daily OHLC reconstructed only from24completehours. No stops/targets/partial management: diagnostic comparator,not user's executable3Rpolicy or previousETHpullbackpolicy. Train2024-2025 (SOL starts2024-08-26),resetcashJan1,known2026holdoutthroughSep1200UTC;2026alreadyresearched,not untouchednewOOS. Train return-24.66415%,DD50.545173%,N12,win41.666667%,meanCI[-9.572975,7.246572]%. Holdoutmarked return-4.867204%,DD17.451183%,closedN3,win0%,mean-5.049061%,CI[-9.724302,-2.115529]%,openN1;stressreturn-6.189414%. Premise2 all9lookback15/20/30 x SMA0/100/200cells retained;holdoutbasecost range-12.839698to11.294846%. MeanCI5000IIDtradebootstrap,smallNandmultiplecomparisonsunadjusted. Counterexample: markedopenprofitcanmaskclosedlosses;strictrollingSMAbecomesunavailableaftergaps,N0isnotvalidation. Conclusion no winningparameterchosenfromholdout and no currentlateentryauthorized. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ;fees https://docs.alpaca.markets/us/docs/crypto-trading . All216cellfiguresobservations26302–26325.

Premise1 preregistered prior20calendar-dayHIGH breakout,prior10dayLOWexit,fillfollowingOPEN,longcashonly,fees25bp/slip10bp eachside,stress30bp. Hourly uses480HIGH/240LOW calendarhours. No stops/targets/partial management: diagnostic comparator,not user's executable3Rpolicy or previousETHpullbackpolicy. Train2024-2025 (SOL starts2024-08-26),resetcashJan1,known2026holdoutthroughSep1206UTC;2026alreadyresearched,not untouchednewOOS. Train return30.449462%,DD34.232668%,N13,win30.769231%,meanCI[-5.327741,13.821846]%. Holdoutmarked return17.935419%,DD12.151445%,closedN3,win33.333333%,mean0.4253%,CI[-2.628821,6.359096]%,openN1;stressreturn16.296286%. Premise2 all9lookback15/20/30 x SMA0/100/200cells retained;holdoutbasecost range0to17.935419%. MeanCI5000IIDtradebootstrap,smallNandmultiplecomparisonsunadjusted. Counterexample: markedopenprofitcanmaskclosedlosses;strictrollingSMAbecomesunavailableaftergaps,N0isnotvalidation. Conclusion no winningparameterchosenfromholdout and no currentlateentryauthorized. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ;fees https://docs.alpaca.markets/us/docs/crypto-trading . All216cellfiguresobservations26302–26325.

Premise1 preregistered prior20calendar-dayHIGH breakout,prior10dayLOWexit,fillfollowingOPEN,longcashonly,fees25bp/slip10bp eachside,stress30bp. Daily OHLC reconstructed only from24completehours. No stops/targets/partial management: diagnostic comparator,not user's executable3Rpolicy or previousETHpullbackpolicy. Train2024-2025 (SOL starts2024-08-26),resetcashJan1,known2026holdoutthroughSep1200UTC;2026alreadyresearched,not untouchednewOOS. Train return32.167365%,DD32.754193%,N10,win30%,meanCI[-6.573376,17.652511]%. Holdoutmarked return18.257988%,DD10.252513%,closedN2,win50%,mean0.047637%,CI[-4.104569,4.199843]%,openN1;stressreturn17.081765%. Premise2 all9lookback15/20/30 x SMA0/100/200cells retained;holdoutbasecost range0to18.257988%. MeanCI5000IIDtradebootstrap,smallNandmultiplecomparisonsunadjusted. Counterexample: markedopenprofitcanmaskclosedlosses;strictrollingSMAbecomesunavailableaftergaps,N0isnotvalidation. Conclusion no winningparameterchosenfromholdout and no currentlateentryauthorized. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ;fees https://docs.alpaca.markets/us/docs/crypto-trading . All216cellfiguresobservations26302–26325.