Stats

Ran required CLI node /app/agent/dist/backtest.js --symbols ETH/USD --strategy sma-cross --fast20 --slow50 --days979 --stop-pct5 --target-pct7.5 (actualflags spaced). Returned980bars,80trades,+33.1940%,DD44.7658%,win45%,buyhold6.51545%,avgtrade.49621%. CLIreadsrawdailyandhasnopaginationiteration/endfilter inbars.js; entireperiodcontrolhasnotrainholdout orETH/BTCratio. Fixed5/7.5barriersdifferentfromrangeexits andfee/slip/partialmanagementaudit. Do notattributecontrast withcustomnettest toonefactor oradoptCLIprofit. Recordascontrolonly; actualcandidatefailsdaily/hourlynettest. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; local/app/agent/dist/backtest/cli.js andbars.js.

Frozen notes247/251 beforehistoryoutput. LongETH when previouscompleteUTCday ETH/BTCratio>50daymean AND ETH>50daymean; next00open,prior14daymeanrange stop1/target1.5,7daytimeout. At completedbarclose+1R trim1/3nextopen andstoptoentry; conservativeboth-touchstopfirst,gapstopobservedopen,targetgapcappedtarget. Vectorbt explicitorders withindependentcashledgererror<1e-6;25bpfee+10bpslip eachleg,doublecoststress. Training2024-25; flatstart2026Jan1-Sep5chronologicalholdout, last7daysnoentriesfornaturalexit. Olderhistoryviewedforotherrules: diagnosticOOS, notuntouchedprospective. MA30/50/70trainallnegative totalreturn, noholdoutselection. Fixed50 hourlytrain-12.1356%,N43 includes1datagapexit; dailytrain-10.1902%,N42. Holdoutdaily-7.0258%,DD15.5417%,N21; hourly-9.64445%,DD17.3868%,N21,p47.619%,W3.60267%,L4.05042%,mean-.406093%,circular3trade95%meaninterval[-1.33974,.64794]%. Doublingcosthourly-21.9967%,daily-19.7361%. Noopenpositions/gapexits inholdout. Rawdaily sensitivity sameholdoutreturn; doesnotresolvehistoricaldefects. Conclusion REJECT; quarterKelly0. Noinverse/shortauthorization orparameterretuning. Barhigh/lowapproximationnotactual15minvenuepath; feecreditedcryptoapproximatedcashfee; missinghistoryandjointregimedependence remain. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Premises:1 Original AQR defensive theory is long-horizon lower-risk equity, not a promise of short swings.2 Frozen note248 adapts existing factorETF with next-open entry firstsession eachmonth,prior20sessionATR stop1,target1.5,max10sessions; daily gapstop at open,stopfirst ifbothhit,target capped.3 Train2016-2022 return-8.2106%,N82; later completedJan2023-Aug2026 return0.8920%,sessioncloseDD8.6201%,N44,win54.5455%,W1.13946%,L1.30363%,mean0.02897%;3monthbootstrap95%meanCI[-0.29972,0.35449]%. Cost10bponeway splitfee/slip, not measuredvenuecost;50bpstressreturn-29.0448%.4 Right-censored September cohort excluded uniformly pernote254; complete history18,788adjustedSIPbars across7ETFs throughSep4. Conclusion reject deployment; positiveEEMVpointestimate cannot overcome CIcrosszero/costfailure. Vectorbt and independent cashledger agree within1e-6. Training5/15holding sensitivities and2bp later controls recorded, no winner selection. Market window already seen in other research, so not pristineOOS; prices adjusted for corporateactions, no cashdividend/rawstop/15min/profitmanagement validation. Sources https://www.aqr.com/-/media/AQR/Documents/Insights/White-Papers/Understanding-Defensive-Equity.pdf ; https://runway.adityaperswal.workers.dev/data/stock-bars ; protocolnote248. Validity follows prerequisite positive independently executableedge; this test rejects the adaptation, not the universal anomaly.

Premises:1 Original AQR defensive theory is long-horizon lower-risk equity, not a promise of short swings.2 Frozen note248 adapts existing factorETF with next-open entry firstsession eachmonth,prior20sessionATR stop1,target1.5,max10sessions; daily gapstop at open,stopfirst ifbothhit,target capped.3 Train2016-2022 return-7.4152%,N82; later completedJan2023-Aug2026 return-8.0622%,sessioncloseDD12.1907%,N44,win45.4545%,W1.08274%,L1.23836%,mean-0.18331%;3monthbootstrap95%meanCI[-0.50147,0.14072]%. Cost10bponeway splitfee/slip, not measuredvenuecost;50bpstressreturn-35.3420%.4 Right-censored September cohort excluded uniformly pernote254; complete history18,788adjustedSIPbars across7ETFs throughSep4. Conclusion reject deployment; positiveEEMVpointestimate cannot overcome CIcrosszero/costfailure. Vectorbt and independent cashledger agree within1e-6. Training5/15holding sensitivities and2bp later controls recorded, no winner selection. Market window already seen in other research, so not pristineOOS; prices adjusted for corporateactions, no cashdividend/rawstop/15min/profitmanagement validation. Sources https://www.aqr.com/-/media/AQR/Documents/Insights/White-Papers/Understanding-Defensive-Equity.pdf ; https://runway.adityaperswal.workers.dev/data/stock-bars ; protocolnote248. Validity follows prerequisite positive independently executableedge; this test rejects the adaptation, not the universal anomaly.

Premises:1 Original AQR defensive theory is long-horizon lower-risk equity, not a promise of short swings.2 Frozen note248 adapts existing factorETF with next-open entry firstsession eachmonth,prior20sessionATR stop1,target1.5,max10sessions; daily gapstop at open,stopfirst ifbothhit,target capped.3 Train2016-2022 return14.1636%,N82; later completedJan2023-Aug2026 return-4.4143%,sessioncloseDD10.9436%,N44,win43.1818%,W1.12446%,L1.02416%,mean-0.09634%;3monthbootstrap95%meanCI[-0.43093,0.23567]%. Cost10bponeway splitfee/slip, not measuredvenuecost;50bpstressreturn-32.7765%.4 Right-censored September cohort excluded uniformly pernote254; complete history18,788adjustedSIPbars across7ETFs throughSep4. Conclusion reject deployment; positiveEEMVpointestimate cannot overcome CIcrosszero/costfailure. Vectorbt and independent cashledger agree within1e-6. Training5/15holding sensitivities and2bp later controls recorded, no winner selection. Market window already seen in other research, so not pristineOOS; prices adjusted for corporateactions, no cashdividend/rawstop/15min/profitmanagement validation. Sources https://www.aqr.com/-/media/AQR/Documents/Insights/White-Papers/Understanding-Defensive-Equity.pdf ; https://runway.adityaperswal.workers.dev/data/stock-bars ; protocolnote248. Validity follows prerequisite positive independently executableedge; this test rejects the adaptation, not the universal anomaly.

Premises:1 Original AQR defensive theory is long-horizon lower-risk equity, not a promise of short swings.2 Frozen note248 adapts existing factorETF with next-open entry firstsession eachmonth,prior20sessionATR stop1,target1.5,max10sessions; daily gapstop at open,stopfirst ifbothhit,target capped.3 Train2016-2022 return14.5780%,N82; later completedJan2023-Aug2026 return-15.8303%,sessioncloseDD18.7584%,N44,win29.5455%,W1.32691%,L1.10213%,mean-0.38446%;3monthbootstrap95%meanCI[-0.70870,-0.03333]%. Cost10bponeway splitfee/slip, not measuredvenuecost;50bpstressreturn-40.8052%.4 Right-censored September cohort excluded uniformly pernote254; complete history18,788adjustedSIPbars across7ETFs throughSep4. Conclusion reject deployment; positiveEEMVpointestimate cannot overcome CIcrosszero/costfailure. Vectorbt and independent cashledger agree within1e-6. Training5/15holding sensitivities and2bp later controls recorded, no winner selection. Market window already seen in other research, so not pristineOOS; prices adjusted for corporateactions, no cashdividend/rawstop/15min/profitmanagement validation. Sources https://www.aqr.com/-/media/AQR/Documents/Insights/White-Papers/Understanding-Defensive-Equity.pdf ; https://runway.adityaperswal.workers.dev/data/stock-bars ; protocolnote248. Validity follows prerequisite positive independently executableedge; this test rejects the adaptation, not the universal anomaly.

Preregistered249 clarified250 before payoff inspection.19 primary-noticed events2022-Jun2026; every announcement strictly before entry; release dates rather than later calls. Training2022-23 N8, untouchedholdout2024-Jun2026 N11. BuyT-3open, exitT-1close, noholdthrough. SPY sessionindex; missingholdingwindowbars excluded rather than shifted (0exclusions). vectorbt from_signals at open/close slots, allcash hypothetical fractional size, fee5bp+slippage10bp eachside. Holdout return3.755451%,maxmarkedDD8.651847%,wins63.636364%,mean.354006%,W1.502475%,L1.655815%;stressslippage25bp return.387389%,DD8.925482%,wins45.454545%,mean.053394%. Training T2/T3/T5 returns9.812431/31.729430/10.598816% (N8each); noholdoutsensitivityselection. Compounded investedcash payoffs agreevectorbt<1e-8. DailyIEXpartialmarket, no15minfillclaim; missingnonholdingbars forwardfilledsolelyforflatportfolio marks. Pricehistory https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=HITI,HAIN,BFRI,SPY&timeframe=1Day&start=2021-01-01&end=2026-09-05&feed=iex&adjustment=all . Primarycalendars/code/trades archivedthisrun. CurrentSep14 alreadyAug4preannounced;2025Q3 alsoAug18preannounced, but singlematchingeventcannotestimaterepeatableedge. Conclusionrejectliveentry:meanCIcrosses0,netexcessslightlynegative,executionuncertain. Historicalpositive returnisnotfalse;claimofreliablefutureedgeunsupported.

Premises: underlying returns are a diagnostic, not option bid/ask returns; signal must execute after it is known; strategyparameters cannot be selected on holdout. Fixed20day population zscore, enterlong whenz<=-2, exitwhenz>=0, onebar delay to nextclose; vectorbt fees5bp+slippage5bp eachside (10bp total), stress15+15bp. TrainJan2023-Dec2025 (ETHA inception2024), holdoutJan2026-Sep4; forceflat eachsampleend; sweep1.5/2/2.5 without choosingwinner. Baseline holdout return0.3071%, drawdown11.3997%, win66.66666666666666%, N3, averagewin3.4514182753635736%, averageloss6.228380384189175%. At30bp return-0.8894%. CLI separately run: node /app/agent/dist/backtest.js --symbols SNAP,ETHA,EWZ --strategy mean-reversion --lookback20 --entry2 --exit0 --days365; unseparated251bar outputs SNAP-14.86%,DD23.97%,win62.5%,N8; ETHA-18.19%,DD32.04%,win66.67%,N6; EWZ+13.40%,DD0.60%,win90%,N10. Prior sourcecode inspection showsCLIexit0unused and flatasentrythresholdclears, hence differentrule andnotindependentreproduction. Vectorbt explicitly implements notedmeanexit. Tinyholdouts andnooptionhistory; evenpositiveEWZunderlyingcelldoesnotestablishcheapcallorput. NooptionquarterKellyedge; size0. Source venue stockbars adjustmentall,feediex,2023-01-01to2026-09-05; https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=EWZ&timeframe=1Day .

Premises: underlying returns are a diagnostic, not option bid/ask returns; signal must execute after it is known; strategyparameters cannot be selected on holdout. Fixed20day population zscore, enterlong whenz<=-2, exitwhenz>=0, onebar delay to nextclose; vectorbt fees5bp+slippage5bp eachside (10bp total), stress15+15bp. TrainJan2023-Dec2025 (ETHA inception2024), holdoutJan2026-Sep4; forceflat eachsampleend; sweep1.5/2/2.5 without choosingwinner. Baseline holdout return-27.9810%, drawdown37.4002%, win0%, N2, averagewinundefined%, averageloss15.143407363395564%. At30bp return-28.5549%. CLI separately run: node /app/agent/dist/backtest.js --symbols SNAP,ETHA,EWZ --strategy mean-reversion --lookback20 --entry2 --exit0 --days365; unseparated251bar outputs SNAP-14.86%,DD23.97%,win62.5%,N8; ETHA-18.19%,DD32.04%,win66.67%,N6; EWZ+13.40%,DD0.60%,win90%,N10. Prior sourcecode inspection showsCLIexit0unused and flatasentrythresholdclears, hence differentrule andnotindependentreproduction. Vectorbt explicitly implements notedmeanexit. Tinyholdouts andnooptionhistory; evenpositiveEWZunderlyingcelldoesnotestablishcheapcallorput. NooptionquarterKellyedge; size0. Source venue stockbars adjustmentall,feediex,2023-01-01to2026-09-05; https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=ETHA&timeframe=1Day .

Premises: underlying returns are a diagnostic, not option bid/ask returns; signal must execute after it is known; strategyparameters cannot be selected on holdout. Fixed20day population zscore, enterlong whenz<=-2, exitwhenz>=0, onebar delay to nextclose; vectorbt fees5bp+slippage5bp eachside (10bp total), stress15+15bp. TrainJan2023-Dec2025 (ETHA inception2024), holdoutJan2026-Sep4; forceflat eachsampleend; sweep1.5/2/2.5 without choosingwinner. Baseline holdout return-20.0435%, drawdown27.6275%, win20%, N5, averagewin5.694649678738458%, averageloss6.583377912009972%. At30bp return-21.6267%. CLI separately run: node /app/agent/dist/backtest.js --symbols SNAP,ETHA,EWZ --strategy mean-reversion --lookback20 --entry2 --exit0 --days365; unseparated251bar outputs SNAP-14.86%,DD23.97%,win62.5%,N8; ETHA-18.19%,DD32.04%,win66.67%,N6; EWZ+13.40%,DD0.60%,win90%,N10. Prior sourcecode inspection showsCLIexit0unused and flatasentrythresholdclears, hence differentrule andnotindependentreproduction. Vectorbt explicitly implements notedmeanexit. Tinyholdouts andnooptionhistory; evenpositiveEWZunderlyingcelldoesnotestablishcheapcallorput. NooptionquarterKellyedge; size0. Source venue stockbars adjustmentall,feediex,2023-01-01to2026-09-05; https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SNAP&timeframe=1Day .

Preregistered note246 before history. Selected current survivor, not unbiased universe strategy. Completed-close SMA10/50 crossover, next-session open entry/reverse-cross exit; flat start each window, historical warmup. Train2023-24, holdout2025-Sep4 2026. Vectorbt all-in fractional diagnostic cash10000, fee5bp/slippage10bp eachside; stress10/20bp. US IEX all-adjusted, overseas Yahoo adjusted OHLC/localcurrency; no FX, venue15minute stops or whole-share model. Training 345.6336%, 5 natural closes; holdout 132.3479%, DD57.3359%, 4 natural closes, 1 open marked positions. Mean95% bootstrap interval [-21.514976837353068,134.70798314842915]%. Nine training-only combinations5/10/20 by40/50/60 archived; training-best 10/60 subsequent return234.5618%, diagnostic not authorization. No retuning after holdout. All figures/code/trades in artifacts/results.json and archive to follow. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; https://query1.finance.yahoo.com/v8/finance/chart/5803.T . Rejected for capital deployment: uncertain payoff and incomplete executable exit model.

Preregistered note246 before history. Selected current survivor, not unbiased universe strategy. Completed-close SMA10/50 crossover, next-session open entry/reverse-cross exit; flat start each window, historical warmup. Train2023-24, holdout2025-Sep4 2026. Vectorbt all-in fractional diagnostic cash10000, fee5bp/slippage10bp eachside; stress10/20bp. US IEX all-adjusted, overseas Yahoo adjusted OHLC/localcurrency; no FX, venue15minute stops or whole-share model. Training -31.2388%, 7 natural closes; holdout 101.1909%, DD45.1439%, 4 natural closes, 0 open marked positions. Mean95% bootstrap interval [-1.9762500013106816,52.820585043496656]%. Nine training-only combinations5/10/20 by40/50/60 archived; training-best 5/50 subsequent return96.8950%, diagnostic not authorization. No retuning after holdout. All figures/code/trades in artifacts/results.json and archive to follow. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; https://query1.finance.yahoo.com/v8/finance/chart/LYC.AX . Rejected for capital deployment: uncertain payoff and incomplete executable exit model.

Preregistered note246 before history. Selected current survivor, not unbiased universe strategy. Completed-close SMA10/50 crossover, next-session open entry/reverse-cross exit; flat start each window, historical warmup. Train2023-24, holdout2025-Sep4 2026. Vectorbt all-in fractional diagnostic cash10000, fee5bp/slippage10bp eachside; stress10/20bp. US IEX all-adjusted, overseas Yahoo adjusted OHLC/localcurrency; no FX, venue15minute stops or whole-share model. Training 6.9090%, 6 natural closes; holdout 53.9612%, DD26.7121%, 5 natural closes, 1 open marked positions. Mean95% bootstrap interval [-5.772772362317384,11.050257229369842]%. Nine training-only combinations5/10/20 by40/50/60 archived; training-best 20/40 subsequent return14.3513%, diagnostic not authorization. No retuning after holdout. All figures/code/trades in artifacts/results.json and archive to follow. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; https://query1.finance.yahoo.com/v8/finance/chart/MAERSK-B.CO . Rejected for capital deployment: uncertain payoff and incomplete executable exit model.

Preregistered note246 before history. Selected current survivor, not unbiased universe strategy. Completed-close SMA10/50 crossover, next-session open entry/reverse-cross exit; flat start each window, historical warmup. Train2023-24, holdout2025-Sep4 2026. Vectorbt all-in fractional diagnostic cash10000, fee5bp/slippage10bp eachside; stress10/20bp. US IEX all-adjusted, overseas Yahoo adjusted OHLC/localcurrency; no FX, venue15minute stops or whole-share model. Training 11.5689%, 8 natural closes; holdout 294.8017%, DD37.3481%, 7 natural closes, 0 open marked positions. Mean95% bootstrap interval [-7.555799047972869,81.92217897636986]%. Nine training-only combinations5/10/20 by40/50/60 archived; training-best 10/60 subsequent return229.0861%, diagnostic not authorization. No retuning after holdout. All figures/code/trades in artifacts/results.json and archive to follow. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; https://query1.finance.yahoo.com/v8/finance/chart/MU . Rejected for capital deployment: uncertain payoff and incomplete executable exit model.

Preregistered note246 before history. Selected current survivor, not unbiased universe strategy. Completed-close SMA10/50 crossover, next-session open entry/reverse-cross exit; flat start each window, historical warmup. Train2023-24, holdout2025-Sep4 2026. Vectorbt all-in fractional diagnostic cash10000, fee5bp/slippage10bp eachside; stress10/20bp. US IEX all-adjusted, overseas Yahoo adjusted OHLC/localcurrency; no FX, venue15minute stops or whole-share model. Training -20.2269%, 5 natural closes; holdout 100.8691%, DD52.6487%, 6 natural closes, 1 open marked positions. Mean95% bootstrap interval [-9.281859484395701,83.34189392058218]%. Nine training-only combinations5/10/20 by40/50/60 archived; training-best 10/40 subsequent return62.0380%, diagnostic not authorization. No retuning after holdout. All figures/code/trades in artifacts/results.json and archive to follow. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; https://query1.finance.yahoo.com/v8/finance/chart/MP . Rejected for capital deployment: uncertain payoff and incomplete executable exit model.

Preregistered note246 before history. Selected current survivor, not unbiased universe strategy. Completed-close SMA10/50 crossover, next-session open entry/reverse-cross exit; flat start each window, historical warmup. Train2023-24, holdout2025-Sep4 2026. Vectorbt all-in fractional diagnostic cash10000, fee5bp/slippage10bp eachside; stress10/20bp. US IEX all-adjusted, overseas Yahoo adjusted OHLC/localcurrency; no FX, venue15minute stops or whole-share model. Training -70.4170%, 8 natural closes; holdout 3806.8958%, DD51.0432%, 3 natural closes, 0 open marked positions. Mean95% bootstrap interval [-26.874767394405925,3755.0995828171845]%. Nine training-only combinations5/10/20 by40/50/60 archived; training-best 5/60 subsequent return2247.0847%, diagnostic not authorization. No retuning after holdout. All figures/code/trades in artifacts/results.json and archive to follow. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; https://query1.finance.yahoo.com/v8/finance/chart/AXTI . Rejected for capital deployment: uncertain payoff and incomplete executable exit model.

Preregistered note246 before history. Selected current survivor, not unbiased universe strategy. Completed-close SMA10/50 crossover, next-session open entry/reverse-cross exit; flat start each window, historical warmup. Train2023-24, holdout2025-Sep4 2026. Vectorbt all-in fractional diagnostic cash10000, fee5bp/slippage10bp eachside; stress10/20bp. US IEX all-adjusted, overseas Yahoo adjusted OHLC/localcurrency; no FX, venue15minute stops or whole-share model. Training -18.3796%, 7 natural closes; holdout 63.6120%, DD34.2930%, 6 natural closes, 0 open marked positions. Mean95% bootstrap interval [-6.253898200108404,33.81850909884469]%. Nine training-only combinations5/10/20 by40/50/60 archived; training-best 10/40 subsequent return57.7266%, diagnostic not authorization. No retuning after holdout. All figures/code/trades in artifacts/results.json and archive to follow. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; https://query1.finance.yahoo.com/v8/finance/chart/COHU . Rejected for capital deployment: uncertain payoff and incomplete executable exit model.

Preregistered note246 before history. Selected current survivor, not unbiased universe strategy. Completed-close SMA10/50 crossover, next-session open entry/reverse-cross exit; flat start each window, historical warmup. Train2023-24, holdout2025-Sep4 2026. Vectorbt all-in fractional diagnostic cash10000, fee5bp/slippage10bp eachside; stress10/20bp. US IEX all-adjusted, overseas Yahoo adjusted OHLC/localcurrency; no FX, venue15minute stops or whole-share model. Training -59.1435%, 8 natural closes; holdout 300.2631%, DD61.8216%, 4 natural closes, 0 open marked positions. Mean95% bootstrap interval [-22.650839954529136,162.84382945785072]%. Nine training-only combinations5/10/20 by40/50/60 archived; training-best 20/60 subsequent return337.2204%, diagnostic not authorization. No retuning after holdout. All figures/code/trades in artifacts/results.json and archive to follow. Sources https://runway.adityaperswal.workers.dev/data/stock-bars ; https://query1.finance.yahoo.com/v8/finance/chart/AEHR . Rejected for capital deployment: uncertain payoff and incomplete executable exit model.

Frozen36experiments: eachof3symbols hastrain/laterfixed/rangeconditional andunconditionalfixed controls, later2/50bponewaycostvariants forfixed/range, andtraining-only2/4sessionrangeholding sensitivities. No winnerselectedforlaterwindow. MainSPY laterfixed+1.2208%,range-3.6496%;2bponewayfixed+4.3452%,range-0.6756%;50bponewayfixed-13.0525%,range-17.2362%. LaterunconditionalcalendarSPY-6.0678%,EFA-11.6534%,EEM-10.2040%; merelybeingnearmonthenddidnotestablishexcessreturns. Allmetrics,DD,p,W,L,N,regimepayoffs andcalendar-blockCIrecorded in36observations. Costassumptions split evenlyfees/slippage are notreportedvenuefees ormeasuredexecution. Counterexample: lowercostshelpthefixedSPYrule, butits95%meanCIstillcrosseszero andmandatoryrangeimplementationfails. MaintrainingSPYrange2/3/4sessionvariationsarenotfreshholdoutsanddonotlicenseparametersearch. CLI sourceaudited: availablemean-reversion strategyisrollingpricezscore, withIEX/noexplicitadjustments/singlepagefetch; cannotrepresentsettlementcalendarorvalidatethisrule. Thereforeownvectorbt scriptswereusedunderuserpermissionforownbacktests; nodifferentCLIstrategyisclaimedasevidence. All8052barsmatchNYSEsessioncoveragewithvalidOHLC; maxterminalledgerdifference5.820766e-11. No rawdividendbankholiday15min/extendedhour/trail/whole-sharevalidation; screenrejectiondoesnotclaimactual tradability. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; /app/agent/dist/backtest/{bars,strategies}.js ; artifacts/test.py,stats.py.

Preregisternote214 beforeLTCresults. Fixed completedclose>prior20closemax AND>SMA50; enter nextUTCopenwhenflat; stopprior14daymean(high-low),target1.5range,timeout7days; after observed+1R moveprice stoptoentry. Check at each observedbaropen, fill actualopen withslippage, noidealbarrierfill. Oneentryperday, no same-dayreentryafterexit. vectorbt explicitorders/cash10k,25bpfee+5bpslip eachside. Chronologicaltraining2024-2025;holdoutJan1-Sep5 2026 startsflat,warmuptrainingprices, no carriedposition. Currentuniverse/knownhistory selectionnotindependentprospective. Canonicalhourly uses completedhourlyaggregateddaily signals; vendor-daily-signalhourly identicalholdoutpayoffs despitehistoricalfeeddefects. Hourlyholdoutreturn-19.40627%,DD23.92950%,N11,p27.2727%,W3.41768%,L3.86918%,E-1.88186%,tradebootstrap95mean[-3.85956,+.22622]%,quarterKelly0. Doublecostreturn-24.55378%,DD27.49468%,N11,mean-2.46881%,CI[-4.43469,-.37334]%. Dailyvariantreturn-7.72250%,DD20.19482%,N8,p25%,W9.35989%,L4.19682%,E-.80765%,Kelly0; doubled-12.04725%. Differentexitfrequency changestrades/payoffs, no independentreplications. Trainingonlylookback15/20/30 returns-34.3387/-19.2603/-26.5399%; fixed20retained,allnegative,notoptimized. No terminalpositions or heldtradescrossing4knownmissinghours; independentledgermaxerror1.82e-12. Missingbars canstillaffectrange/history. Quote spread58.9657bp exceeds base10bp roundtripslip; doubledtotalcost120bp roughly exceeds currentspread+50bpfees109bp, butnotvalidatedhistoricalspreadmodel. Pricebreakeven doesnotcoverfees; no mandatorypartialprofit or15minuteimplementation. Negativeinallregisteredvariants sufficesREJECT,size0; don'tretuneinspectedholdout. Serializationnp.int failurefixedwithoutchangingrules; partialartifactdiscarded. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://runway.adityaperswal.workers.dev/data/crypto-snapshots ; https://docs.alpaca.markets/us/docs/crypto-fees .

Preregistered note212 before payoff inspection; data rules note216. Source https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=LEN,ADBE,ASO,SPY&timeframe=1Day&start=2021-01-01&end=2026-09-05&feed=iex&adjustment=all&limit=10000 . Exact issuer URLs/calendar and trades archived this run. Reconstructed21 completed announcements; Q4 2021 unverified notice excluded; March2022 excluded for missing ATR-window bar. Train2021-2023 N10; chronological untouched holdout2024-June2026 N10. Enter T-5 open, exit T-1 close before actual earnings release, notice strictly earlier than entry. Base fees5bp/slippage10bp each side. Baseline holdout return-11.922916%, mark-to-market DD20.464851%, wins50%, mean-1.176670%, average win2.113230%, loss4.466570%. Frozen alternative uses prior-session SMA ATR14, stop1ATR,target1.5ATR; opening gaps before intraday checks, both touches stop first. Overlay holdout return1.465552%, DD10.114489%, wins60%, mean.181465%, W2.246256%, L2.915723%. With25bp slippage plus5bp fee each side, overlay return-1.533218%, DD10.919831%, mean-.118630%. Training sensitivities T3/T5/T7 were not tested on holdout except preregistered T5 and did not select a replacement. Vectorbt from_orders, all-cash fractional sizing, agrees with direct compounded invested-cash payoffs to1e-9. Daily IEX is not consolidated or15min execution; intraday exits approximated at synthetic close slots, no claim of actual fill precision. Reconstructed announcements are primary but cannot prove absence of every intervening update. No window crosses MillroseFeb7 2025 distribution; structural change still limits pooling. Reject live risk: uncertainty and costs fail, whole share also unaffordable.

Requiredcontrol executed node /app/agent/dist/backtest.js --symbols LTC/USD --strategy breakout --lookback20 --days979 --stop-pct5 --target-pct7.5 (actual flags spaced). Vendor980dailybars,currentpartialSep6included; no chronologicalholdout, fees orslippage; samebarcloseentry and fixedbarrierexits, signalexit unlike registered7dayrange rule. Thus a diagnostic control, NOT registeredrule replication or parameterselection. Return-24.538177%,DD31.073446%,win26.415094%,N53,avgtrade-.434447%,BH-26.736173%. Source engine/code /app/agent/dist/backtest/engine.js ; https://runway.adityaperswal.workers.dev/data/crypto-bars . vectorbt prereg214 separately usescausalnextopen,rangeexits,fees,holdout and hourlycomparison. ThisgrosscontrolcannotjustifyLTCpurchase, and disagreement withcustomtestcannotisolateonebias.

Executed node /app/agent/dist/backtest.js --symbols LEN --strategy momentum --lookback 5 --days 2000 --stop-pct 3 --target-pct 4.5. IEX1373bars/223trades. Return-33.059019%, drawdown46.427112%, wins35.426009%, average trade-.135762%, buy-and-hold-5.836620%. Source https://runway.adityaperswal.workers.dev/data/stock-bars . Engine inspection /app/agent/src/backtest/engine.ts: signal-close entries, exact price stop/target with no opening-gap treatment, no cost/holdout split. This is required CLI baseline, not independent support for a catalyst proposal. Rule and price adjustment differ from frozen event study, so difference in output will not isolate a single cause. Reject CLI as trading evidence even before adding costs.

Frozen protocol note213 before price inspection. Chronological later2023-Jan through2026-Sep4, already seen for other rules so NOT pristine independent validation. Endmonth deadlineD=T-(settlementlag+1), lag3/2/1 across SEC changes. Buy nextopen only if preceding5sessionreturn<0, max3open-to-open sessions; longonly full-capital research normalization, not actualfundallocation. Compare fixedhold to stop1laggedATR20/target1.5ATR20. Daily stopfirst when both hit, stopgap atopen, favorabletargetgap capped at target; same-day entries/exits supported through separate synthetic slots. Fees5bp+slippage5bp eachway, cashreturn0. SPY is primaryUSproxy; EFA/EEM transfer diagnostics do not establish underlyinglocalflows. Complete8052adjustedSIPbars,2684each,NYSEcalendar no missing/extra dates, OHLC/volume checks pass. Vectorbtfrom_orders results reconcile to independent compounded cash ledger for every36experiment terminal equity, errors<1e-6 on10000 initialcapital. Drawdown is open/close-marked, not intradaymax. Fixed return -2.785644%,DD11.925423%,N21,p38.0952%,W1.33977%,L1.02428%; range figures below. Range3-calendar-month-block10000bootstrap mean95CI [-0.8158713215769833,0.4405276085542465]%; inactive months retained. All trial metrics and uncertainty separately observed; code/raw inputs/fulltradeandequityledgers local artifacts. Conclusions concern this newly specified proxy, not exact reproduction of institutionalANcerno study. Mandatory15min/exthours/trailing-profitmanagement,rawcashdividends,banksettlementholiday differences and whole-share sizing unvalidated; no execution authorization. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; https://aaltodoc.aalto.fi/server/api/core/bitstreams/0f01d3c2-3b16-44f9-9135-3f03cdd29334/content ; https://www.sec.gov/exams/educationhelpguidesfaqs/t1-faq .

Frozen protocol note213 before price inspection. Training2016-2022. Endmonth deadlineD=T-(settlementlag+1), lag3/2/1 across SEC changes. Buy nextopen only if preceding5sessionreturn<0, max3open-to-open sessions; longonly full-capital research normalization, not actualfundallocation. Compare fixedhold to stop1laggedATR20/target1.5ATR20. Daily stopfirst when both hit, stopgap atopen, favorabletargetgap capped at target; same-day entries/exits supported through separate synthetic slots. Fees5bp+slippage5bp eachway, cashreturn0. SPY is primaryUSproxy; EFA/EEM transfer diagnostics do not establish underlyinglocalflows. Complete8052adjustedSIPbars,2684each,NYSEcalendar no missing/extra dates, OHLC/volume checks pass. Vectorbtfrom_orders results reconcile to independent compounded cash ledger for every36experiment terminal equity, errors<1e-6 on10000 initialcapital. Drawdown is open/close-marked, not intradaymax. Fixed return 30.245395%,DD11.055170%,N39,p58.9744%,W2.04843%,L1.23969%; range figures below. Range3-calendar-month-block10000bootstrap mean95CI [-0.3608248045957735,0.6969985198676744]%; inactive months retained. All trial metrics and uncertainty separately observed; code/raw inputs/fulltradeandequityledgers local artifacts. Conclusions concern this newly specified proxy, not exact reproduction of institutionalANcerno study. Mandatory15min/exthours/trailing-profitmanagement,rawcashdividends,banksettlementholiday differences and whole-share sizing unvalidated; no execution authorization. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; https://aaltodoc.aalto.fi/server/api/core/bitstreams/0f01d3c2-3b16-44f9-9135-3f03cdd29334/content ; https://www.sec.gov/exams/educationhelpguidesfaqs/t1-faq .