Stats

Start allcash crypto235.86 equity850.99, goalrepeatablenetprofit, bindingconstraint entryconditioned costedpayoff. Premises:1 Liquid twoexplorertips unchanged4051232/4050335 aged184353/238173seconds08:01; operatorstillpreparingrestart. Reserve+4.000004BTCmempool not authorizedpayoutproof; confirmations and controller evidence needed. 2 Issuerstock,ETFshares andGMEaccounting don't time spotorders (fundamental1628). 3 PublishedSep9 compensation/services14UTC,Sep10PPI12:30UTC/wholesale14UTC/EIA16UTC are scheduled riskcontext, notdirectional entryedges. Firstorder: reliable recovery wouldbenefit trappedclaimants; stablecoinpayments benefitissuer/marketmakers; macro surprises change observed costs/activity. Secondorder: optional BTC/ETH/SOL flows depend onfunddestinations andallocation, with substitution/payments/hedging counterexamples. Thirdorder: priceimpact may precede publication; paying~.5%feesplusspread to chase context can erode cash. Strongcounterargument rebound underway could be missed; acknowledges opportunitycost withoutinventingedge. Conclusionsvalidgivenmandategate; soundnesslimitedbyexplorerlag,issuerbias,unknown custodyauthorization andno matchedflowpricehistory. Newverifiedpayoutor datednewspotflow updatespremises; onlyindependentdaily/hourly netedgeand15minexecution canchangesize. Sources https://blockstream.info/liquid/api/blocks ; https://liquid.network/api/blocks ; https://x.com/Blockstream ; https://www.bls.gov/schedule/2026/09_sched.htm ; https://www.census.gov/economic-indicators/calendar-listview.html ; https://www.eia.gov/petroleum/supply/weekly/schedule.php ; https://docs.alpaca.markets/us/docs/crypto-fees .

Premises:1 CircleSep7 circulation74.28bn and7day issuance8.8bn/redemption7.86bn imply+.94bn/week, average134.286m/day over reportedweek; neither buying nor destination known. Sep9 usdc.com stock74.381077962bn at08:01 vs74.436755687bn06:16 (55.678m lower), distinct live cutoff. LlamaSep9 stock74.502816709bn and chain populations differ; no contradiction after dates/definitions separated. 2 IBIT/ETHASep8 shares1.38756bn/467.68m versusSep4CSV1.38496bn/464.56m; quantities remainSep4, preventing new dated spotflow. 3 GME releaseSep8 quarterendedAug1 digitalassets/receivables294.1m, Q2loss75m, purchasecash0 andwritten-optionproceeds12.4m. Those do not identify BTCsales; latest SEC10Q remainsJun11. Cash+marketable5060.3m anddebt~2800m afterSep3 exchange don't establish forcedsale. Conclusion retain liquidity watch but allocation0. Valid inference: absent destination/timing cannot estimate entryconditioned impact. Strongcounterexample actual purchases may precede disclosure; absenceofproof notproofofabsence. New disclosure with BTCquantities/counterparty/timing would update context, still independentpriceedge required. Issuer statements are closest accounting sources but commercially interested, Llama secondary aggregation not independent flow corroboration. Sources https://www.circle.com/transparency ; https://www.usdc.com/ ; https://stablecoins.llama.fi/stablecoin/2 ; https://www.ishares.com/us/products/333011/ishares-bitcoin-trust-etf/latest-holdings.csv ; https://www.ishares.com/us/products/337614/ishares-ethereum-trust-etf/latest-holdings.csv ; https://www.sec.gov/Archives/edgar/data/1326380/000132638026000050/a991-fy26q2earningsrelease.htm ; https://data.sec.gov/submissions/CIK0001326380.json .

Startstate equity850.99,crypto235.86cash,noopen/queuedlots;goalrepeatablenetprofit,bindingconstraintvalidatedentrypayoff. Premises:1 New144costedvectorbtruns with2025train/2026holdout,nextopen,fees25bp/leg,5/15bpslip andparametercontrols do notqualifyfrozenRSItrendpullback. 2 DailybaselineBTC/ETH/SOL return+.18177/-3.75826/-3.44512%,N3/1/1;hourly+.62703/+4.48647/+2.05713%,N4/5/3. PositivebaselineCIsincludenegativepayoffs;N<30. BTCstress-.41752%daily/-.17478%hourly. ETH/SOLhourlytraining-17.0122/-23.4945%,SOLtraining6missingheldhours. 3 OLSjointtestcells2/0/0 andsimulationN1–3cannotrescuesample; no currentcompletedSep8entry. 4 Dueflow/recoveryreview1502 noverifiablenewspotbuying orauthorizedrecovery;quotes1558 cost .5385/.5309/.6303%recoveryhurdles. ConclusionnoPROPOSE/EXIT/ADJUST;approvedKelly0,235.86cash. Validity: mandate requiresvalidatededge absenthere. Soundnesslimits: reconstructedhistoricalbars,reusedcalendaracrossresearch,microstructureunmeasured,bootstrapserialdependence;inconclusivedoesnotmeanfuturepricesfall. StrongcounterargumentSOLdailyunfiltered+14.1017%N20; failsCIandhourly(-11.1073%),do notselectitaftertest. Changeansweronlywithprospectiveindependent30+samples,costedpositiveuncertaintybound,daily/hourlyand15minexecutionvalidation. No freshdatedcatalysttrade. Monitor116resolved120at08UTC;jobs123–125requested,firstsavedoutputpending06:08,manualscriptsverified;14/25retained. Lessonsrecordedtechnical/execution/psyche. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees ; https://www.circle.com/transparency ; https://www.usdc.com/ ; https://blockstream.info/liquid/api/blocks . Finish~12minunder43allowance.

06:09freshBTC/ETH/SOL bidask78903.9/78932.666,2495.48/2496.2,104.055/104.188; spreads3.645/2.885/12.774bp. Exactcreditedassetfees recoveryhurdles.5385/.5309/.6303%. Displayedasksizes78.65/82.99/98.98USD below212.7475positioncap, so fullcapfill not provenbytopquote. HYPE displayedask1.54USDdespitelargeearlierdepth; timevariation matters. StablecoinquotesUSDC/USDGstale;8/10fresh. LatestcompletedSep8dailyATR14 BTC2271.071824/ETH88.760494/SOL5.6218 (~2.8950/3.5725/5.4390%close). Measured1ATRstop/1.5ATRtarget wasonlydiagnostic; counterfactualplannednotionalriskcannotexceed8.5099USD andcash235.86. Stopchecks15min aren'tcontinuousguarantees. No currentfrozenRSIentry;positivepointKelly estimatesunsupportedbyN andCI;approvedsize0. Top10volume inventory unchangedper36pairjob14,includesUSDC/USDG andHYPE;USDGonly4/7days,sohistoryeligibilitynotimpliedbyrank. OwnonlyBTCtradeoperatorliquidated,-.081%raw,notsetupedge. No holdings soR/targetprogress/horizonreviewnotapplicable. Source https://runway.adityaperswal.workers.dev/data/crypto-snapshots ; https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

RequiredCLI node /app/agent/dist/backtest.js --symbols BTC/USD,ETH/USD,SOL/USD --strategy rsi --period 2 --low 10 --high 60 --days 616. Retrieved617bars inclcurrentpartialSep9; source https://runway.adityaperswal.workers.dev/data/crypto-bars . BTC/ETH/SOLuncostedreturn-14.5454/+53.1478/-28.7169%,maxDD33.2792/40.4938/43.7119%,N87/93/87,p59.7701/64.5161/54.0230%;avgtrade-.14368/.57454/-.29181%. Buyhold-16.4412/-25.5942/-46.3200%. Engine inspectionreveals sameclosesignalexecution,no feesandnosplit; CLIrsi logic differsfrom frozenpersistentpositionrule. ThusETHheadline53%cannotjustifytrading. Realcostednextopenvectorbt experiment1536–1541 uses2025train/2026holdout,completeddays,fees,slipandtrend/timeexit. Rejection ofCLIasvalidationdoesnotclaimnegativefutureETHreturns. StatsfiguresbelowETHillustrativeuncosted diagnostic,notfundperformance. Sources localCLIengine and https://docs.alpaca.markets/us/docs/crypto-fees .

Startstate equity850.99,cryptocash235.86,noholdings/queuedorders; goalpositive repeatablenetcashprofit; bindingconstraint evidenceprecision/execution, notavailablecapital. Premises:1 Rawliquiditysnapshots do notidentify incrementalspotbuyers.2 PullbackOOSN1–5,not30minimum;allpositivebaselinesuncertaintyincludeszero;BTCcoststressnegative.3 Sampling10000pathsfromN1doesnotreduceepistemicuncertainty. Firstorder: stablecoinissuers/paymentusers/marketmakers benefitfromcirculation/turnover; takerpaystwofees (~.5%before spread), motivatingcosthurdle. Secondorder: onlynetnewdestination-specificspotpurchases orprotocolfees maybenefitmajorcoins; paymentsorrotation canoffset. Thirdorder: chasingtinypositivebacktestestimates increases costs and shrinks runway, while cash preservescapacity fornewindependentlymeasuredflow/priceedges. Who benefitsfromincreasedunvalidatedturnover: venueandliquidityproviders, notnecessarilyfund. Strongcounterargumentcash maymisscontinuingrally; valid, but futureupsidepossibilitydoesnotestablishentry-conditionedpayoff. Resolvebyprospective independentperiods,destination/publicationdata and15minexecution,notretuningthiscalendar. Assumptions fees25bp/legandslip5/15bp;facts sources https://docs.alpaca.markets/us/docs/crypto-fees ; https://www.circle.com/transparency ; https://www.usdc.com/ ; https://runway.adityaperswal.workers.dev/data/crypto-bars .

Premises:1 Reuse actual dailybaseline entrytimestamps, gathercomplete72hour paths, freeze previousdayATR14 stop1ATRbelow entry,target1.5ATRabove; no RSIearlyexit in thisdistinct barrierdiagnostic. Hourlyclose firstcrossing, executeatobservedclosewith5bpslip andcreditedasset25bp/leg, notatidealbarrier. 2 Only3/1/1 BTC/ETH/SOL paths,none skipped:BTC0stop/0target/3time,ETH1stop,SOL1time. Netmean .224496/-3.888403/-3.406115%. 10000wholepathresamples yield0/100% empiricalfrequencies solelyfromtinyN; NOT calibrated exitodds, and hourlychecks miss15mincrossings/intrahour excursions. 3 Counterfactualbook235.86,initialallocation50=21.199%book rebalanced eachof10IIDtradecycles;10000paths baselinepayoffs. BTCdailyendingp5/median/p95=233.48/236.20/238.91,loss42.24%;BTC hourly232.13/236.67/241.26,loss38.4%. ETHdailydegenerate217.73,SOLdaily219.19;ETHhourly235.15/240.43/245.52,loss11.31%;SOLhourly230.72/240.36/247.69,loss21.93%. These are sensitivityillustrations,not22day/monthforecasts; variableholdingdurations,serialdependence,tailrisk andregimechanges omitted. Conclusion morepaths cannotreplaceindependentobservations; actualapprovedKelly/allocation0. Strongcounterargumentpositivehourlydistributions merely repeat3–5 favorablehistoricaltrades, trainingcontradictsETH/SOL. Allfiguresobservations and research/simulation.json. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

CompletedSep8daily BTC/ETH/SOL RSI2=16.5270/40.9920/36.2477; allaboveentry10whileprice>SMA100. HourlyRSI48=45.5215/50.3215/48.7828 allaboveentry40. No signal at frozencompletedcutoff. 2026dailytrendfraction26.2948/25.0996/15.5378%;oversolddays38/30/36,butjointdays6/1/1; overlappingconsecutivedays reducecompletedtrades to3/1/1. Hourlyjointbars48/24/29reduce to4/5/3trades. Controls withouttrend: BTC/ETH/SOLdailyOOSreturns-12.5166/-16.5445/+14.1017% (ETHexactinobservations), hourly-28.2809/-31.9717/-11.1073%. SOLdailycontrolN20,p60%,win4.6303/loss4.8516%,mean.8375%,CI[-1.7811,3.3112]%;15bpslipreturn9.6277%,yetfailsN/CIandhourly. No retrofit tocontrolwinner. Trendagreement isnotindependent evidence ofedge. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; observations17457–17544,17595–17600.

Premises:1 FitOLS2025 only on121 nonoverlapping3day nextopen netreturns permajor: intercept, priorclose SMA100trend,RSI2<10,interaction,20dayvol. 2 Holdout83 disjoint3dayintervals2026: MSE relativezero BTC.98188,ETH.99606,SOL.95755; predictionsmean-.7529/-1.2662/-.6976% versusactual-.5905/-.5998/-.5918%. 3 Trainjointcells6/7/5,holdout2/0/0. Positive interactioncoefficients1.3783/2.7093/2.3312pctpoints cannotvalidate a tradingcell essentiallyabsentfromtest. Conclusion regression describes mainlynegative unconditionalnetdrift, notusablejointsignal. No statisticalsignificanceclaim; different3daygrid alignment could alterresults, fullrank5 doesn'tfixsmallcellcounts. Reusedcalendar andregime dependence remain. Strongcounterargument modelmaybeatzerobyforecastingcosts/negativebase drift; thatdoesnot provide long-onlyentryedge. Allfigures andcoefficients recordedin observations; filesresearch/regression.py/json. Source https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Frozen note1384 beforefetching. Daily ewm(Wilder alpha)RSI2<10,close>SMA100; hourlydiagnostic RSI48<40,close>SMA2400 isdistinctsampling,notexactdailyreplication. Nextopenentry,exit nextopen after RSI>60daily/>50hourly or3days/72h; forcedlastopen liquidation. WarmupSep2024;train2025;OOSJan1–Sep8 2026 heldoutforthisrule, but reusedcalendarbypriorresearch isnotpristineindependentconfirmation. Fees entry.0025/.9975,exit.0025 exactlyreproducecreditedasset25bp/leg;slippage5bp/legand15bpstress assumptions,nothistoricaldepth. Isolated100cash fullyinvested portfolio. CashPnL/(entrynotional+entryfee); all144compoundedpayoffs reconcile1e-7pctpoint. Baseline {"symbol":"SOL","tf":"1Hour","period":"oos","threshold":40,"trend_filter":true,"slip_bp":5,"return_pct":2.0571309885728795,"drawdown_pct":5.428983945916399,"n":3,"win_pct":66.66666666666666,"avg_win_pct":2.928592480689437,"avg_loss_pct":3.6656875267508515,"expectancy_pct":0.7304991448760073,"ci_low_pct":-3.6656875267508515,"ci_high_pct":3.415400829960636,"quarter_kelly":0.062359234828059454,"held_missing":0}. Training return -23.49450244882479%,N10,expectancy-2.566572205239217%,heldmissing6. StressOOSreturn1.446620903042998%. Quality {"symbol":"SOL","tf":"1Hour","rows":17705,"expected":17712,"missing":7,"duplicates":0,"first":"2024-09-01 00:00:00+00:00","last":"2026-09-08 23:00:00+00:00"}. No baselineOOSinvestedgaps; forwardfillformarking/indicators may distort regime. SOLhourlytrainheld6missingintervals. Bootstrap5000IIDtrades;N1boundsdegenerate,notpopulationuncertainty. AllbaselineN<30;allpositivebaselineCIs includezero. No15minstopvalidation. Parametercontrolsare diagnostics,no winningvariantselected. Full144rows observations17457–17544. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Frozen note1384 beforefetching. Daily ewm(Wilder alpha)RSI2<10,close>SMA100; hourlydiagnostic RSI48<40,close>SMA2400 isdistinctsampling,notexactdailyreplication. Nextopenentry,exit nextopen after RSI>60daily/>50hourly or3days/72h; forcedlastopen liquidation. WarmupSep2024;train2025;OOSJan1–Sep8 2026 heldoutforthisrule, but reusedcalendarbypriorresearch isnotpristineindependentconfirmation. Fees entry.0025/.9975,exit.0025 exactlyreproducecreditedasset25bp/leg;slippage5bp/legand15bpstress assumptions,nothistoricaldepth. Isolated100cash fullyinvested portfolio. CashPnL/(entrynotional+entryfee); all144compoundedpayoffs reconcile1e-7pctpoint. Baseline {"symbol":"SOL","tf":"1Day","period":"oos","threshold":10,"trend_filter":true,"slip_bp":5,"return_pct":-3.445115910082066,"drawdown_pct":3.445115910082064,"n":1,"win_pct":0,"avg_win_pct":0,"avg_loss_pct":3.4451159100820625,"expectancy_pct":-3.4451159100820625,"ci_low_pct":-3.4451159100820625,"ci_high_pct":-3.4451159100820625,"quarter_kelly":0,"held_missing":0}. Training return -18.19790734435226%,N7,expectancy-2.5331341375826337%,heldmissing0. StressOOSreturn-3.638032905953807%. Quality {"symbol":"SOL","tf":"1Day","rows":738,"expected":738,"missing":0,"duplicates":0,"first":"2024-09-01 00:00:00+00:00","last":"2026-09-08 00:00:00+00:00"}. No baselineOOSinvestedgaps; forwardfillformarking/indicators may distort regime. SOLhourlytrainheld6missingintervals. Bootstrap5000IIDtrades;N1boundsdegenerate,notpopulationuncertainty. AllbaselineN<30;allpositivebaselineCIs includezero. No15minstopvalidation. Parametercontrolsare diagnostics,no winningvariantselected. Full144rows observations17457–17544. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Frozen note1384 beforefetching. Daily ewm(Wilder alpha)RSI2<10,close>SMA100; hourlydiagnostic RSI48<40,close>SMA2400 isdistinctsampling,notexactdailyreplication. Nextopenentry,exit nextopen after RSI>60daily/>50hourly or3days/72h; forcedlastopen liquidation. WarmupSep2024;train2025;OOSJan1–Sep8 2026 heldoutforthisrule, but reusedcalendarbypriorresearch isnotpristineindependentconfirmation. Fees entry.0025/.9975,exit.0025 exactlyreproducecreditedasset25bp/leg;slippage5bp/legand15bpstress assumptions,nothistoricaldepth. Isolated100cash fullyinvested portfolio. CashPnL/(entrynotional+entryfee); all144compoundedpayoffs reconcile1e-7pctpoint. Baseline {"symbol":"ETH","tf":"1Hour","period":"oos","threshold":40,"trend_filter":true,"slip_bp":5,"return_pct":4.486466148934809,"drawdown_pct":5.6645663087276255,"n":5,"win_pct":80,"avg_win_pct":1.895584030087242,"avg_loss_pct":3.0733323098522227,"expectancy_pct":0.9018007620993487,"ci_low_pct":-1.1393301515380472,"ci_high_pct":2.06476940466497,"quarter_kelly":0.11893442176470606,"held_missing":0}. Training return -17.012156281254462%,N9,expectancy-1.960743423358159%,heldmissing0. StressOOSreturn3.446807319115635%. Quality {"symbol":"ETH","tf":"1Hour","rows":17710,"expected":17712,"missing":2,"duplicates":0,"first":"2024-09-01 00:00:00+00:00","last":"2026-09-08 23:00:00+00:00"}. No baselineOOSinvestedgaps; forwardfillformarking/indicators may distort regime. SOLhourlytrainheld6missingintervals. Bootstrap5000IIDtrades;N1boundsdegenerate,notpopulationuncertainty. AllbaselineN<30;allpositivebaselineCIs includezero. No15minstopvalidation. Parametercontrolsare diagnostics,no winningvariantselected. Full144rows observations17457–17544. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Frozen note1384 beforefetching. Daily ewm(Wilder alpha)RSI2<10,close>SMA100; hourlydiagnostic RSI48<40,close>SMA2400 isdistinctsampling,notexactdailyreplication. Nextopenentry,exit nextopen after RSI>60daily/>50hourly or3days/72h; forcedlastopen liquidation. WarmupSep2024;train2025;OOSJan1–Sep8 2026 heldoutforthisrule, but reusedcalendarbypriorresearch isnotpristineindependentconfirmation. Fees entry.0025/.9975,exit.0025 exactlyreproducecreditedasset25bp/leg;slippage5bp/legand15bpstress assumptions,nothistoricaldepth. Isolated100cash fullyinvested portfolio. CashPnL/(entrynotional+entryfee); all144compoundedpayoffs reconcile1e-7pctpoint. Baseline {"symbol":"ETH","tf":"1Day","period":"oos","threshold":10,"trend_filter":true,"slip_bp":5,"return_pct":-3.758260817716291,"drawdown_pct":3.7582608177162857,"n":1,"win_pct":0,"avg_win_pct":0,"avg_loss_pct":3.7582608177162897,"expectancy_pct":-3.7582608177162897,"ci_low_pct":-3.7582608177162897,"ci_high_pct":-3.7582608177162897,"quarter_kelly":0,"held_missing":0}. Training return 4.145470490534578%,N7,expectancy0.7257945925852591%,heldmissing0. StressOOSreturn-3.9505521489681144%. Quality {"symbol":"ETH","tf":"1Day","rows":738,"expected":738,"missing":0,"duplicates":0,"first":"2024-09-01 00:00:00+00:00","last":"2026-09-08 00:00:00+00:00"}. No baselineOOSinvestedgaps; forwardfillformarking/indicators may distort regime. SOLhourlytrainheld6missingintervals. Bootstrap5000IIDtrades;N1boundsdegenerate,notpopulationuncertainty. AllbaselineN<30;allpositivebaselineCIs includezero. No15minstopvalidation. Parametercontrolsare diagnostics,no winningvariantselected. Full144rows observations17457–17544. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Frozen note1384 beforefetching. Daily ewm(Wilder alpha)RSI2<10,close>SMA100; hourlydiagnostic RSI48<40,close>SMA2400 isdistinctsampling,notexactdailyreplication. Nextopenentry,exit nextopen after RSI>60daily/>50hourly or3days/72h; forcedlastopen liquidation. WarmupSep2024;train2025;OOSJan1–Sep8 2026 heldoutforthisrule, but reusedcalendarbypriorresearch isnotpristineindependentconfirmation. Fees entry.0025/.9975,exit.0025 exactlyreproducecreditedasset25bp/leg;slippage5bp/legand15bpstress assumptions,nothistoricaldepth. Isolated100cash fullyinvested portfolio. CashPnL/(entrynotional+entryfee); all144compoundedpayoffs reconcile1e-7pctpoint. Baseline {"symbol":"BTC","tf":"1Hour","period":"oos","threshold":40,"trend_filter":true,"slip_bp":5,"return_pct":0.6270298370110368,"drawdown_pct":3.6590873048930606,"n":4,"win_pct":50,"avg_win_pct":1.889439687398009,"avg_loss_pct":1.5465683221991386,"expectancy_pct":0.17143568259943512,"ci_low_pct":-1.5465683221991386,"ci_high_pct":1.889439687398009,"quarter_kelly":0.02268340235240894,"held_missing":0}. Training return 2.786101604035153%,N14,expectancy0.206323737432632%,heldmissing0. StressOOSreturn-0.1747757715770888%. Quality {"symbol":"BTC","tf":"1Hour","rows":17710,"expected":17712,"missing":2,"duplicates":0,"first":"2024-09-01 00:00:00+00:00","last":"2026-09-08 23:00:00+00:00"}. No baselineOOSinvestedgaps; forwardfillformarking/indicators may distort regime. SOLhourlytrainheld6missingintervals. Bootstrap5000IIDtrades;N1boundsdegenerate,notpopulationuncertainty. AllbaselineN<30;allpositivebaselineCIs includezero. No15minstopvalidation. Parametercontrolsare diagnostics,no winningvariantselected. Full144rows observations17457–17544. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Frozen note1384 beforefetching. Daily ewm(Wilder alpha)RSI2<10,close>SMA100; hourlydiagnostic RSI48<40,close>SMA2400 isdistinctsampling,notexactdailyreplication. Nextopenentry,exit nextopen after RSI>60daily/>50hourly or3days/72h; forcedlastopen liquidation. WarmupSep2024;train2025;OOSJan1–Sep8 2026 heldoutforthisrule, but reusedcalendarbypriorresearch isnotpristineindependentconfirmation. Fees entry.0025/.9975,exit.0025 exactlyreproducecreditedasset25bp/leg;slippage5bp/legand15bpstress assumptions,nothistoricaldepth. Isolated100cash fullyinvested portfolio. CashPnL/(entrynotional+entryfee); all144compoundedpayoffs reconcile1e-7pctpoint. Baseline {"symbol":"BTC","tf":"1Day","period":"oos","threshold":10,"trend_filter":true,"slip_bp":5,"return_pct":0.18177174799714635,"drawdown_pct":1.8565435856762624,"n":3,"win_pct":66.66666666666666,"avg_win_pct":0.7451625511902433,"avg_loss_pct":1.290625790259734,"expectancy_pct":0.06656643737358425,"ci_low_pct":-1.290625790259734,"ci_high_pct":1.3954415537106202,"quarter_kelly":0.02233285786680303,"held_missing":0}. Training return 2.3869241956420666%,N7,expectancy0.36453314982556506%,heldmissing0. StressOOSreturn-0.4175198590269105%. Quality {"symbol":"BTC","tf":"1Day","rows":738,"expected":738,"missing":0,"duplicates":0,"first":"2024-09-01 00:00:00+00:00","last":"2026-09-08 00:00:00+00:00"}. No baselineOOSinvestedgaps; forwardfillformarking/indicators may distort regime. SOLhourlytrainheld6missingintervals. Bootstrap5000IIDtrades;N1boundsdegenerate,notpopulationuncertainty. AllbaselineN<30;allpositivebaselineCIs includezero. No15minstopvalidation. Parametercontrolsare diagnostics,no winningvariantselected. Full144rows observations17457–17544. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees .

Premises:1 Circle stillSep7 net7day issuance .94bn, but USDC.com newlySep9 balance74.43675568681bn at06:01 is below Sep8 74.47079165978bn and changed from same-day05:30 74.44261970278bn. Treat intraday stock snapshots as stocks, not separately timestamped net mint events. 2 ETF Sep8 shares unchanged; actual CSV quantities stillSep4, so timing of incremental spot demand unproven. 3 Liquid explorers remain897 blocks apart and stale; unchanged operator statement describes preparation, not verified authorized payouts. Conclusion no flow/recovery trade trigger. Valid reasoning from absent evidence to no authorization, not proof flows or recovery cannot occur. Strong counterargument: issuance and restart may precede publication; prospective monitoring can reveal it, reconstructed dates cannot justify anticipatory historical entries. Sources observations17414–17416 and ETF observation, https://www.circle.com/transparency ; https://www.usdc.com/ ; https://www.ishares.com/us/products/333011/ishares-bitcoin-trust-etf ; https://www.ishares.com/us/products/337614/ishares-ethereum-trust-etf ; https://blockstream.info/liquid/api/blocks ; https://liquid.network/api/blocks ; https://x.com/Blockstream . No holdings to review.

Start state: equity850.99, crypto235.86cash, no open/queued lots; owntrade10 only, operator-liquidated rawreturn-.08099%, no usableflowedge history. Goal reproducible netprofit; binding constraint positive entry-conditioned payoff with knownpublicationtime and15minexecution. Premises:1 Circle netUSDC+.94bn/week supports moreoutstandingdollars, but grossminting mostlyoffset, destinationsunidentified (1446).2 Tazapay isexistingpayments,400mstockconsideration and2027expectedclose, notimmediateUSDC/spotpurchase (1400).3 NewfrozenUSDCrule48diagnostics all lose; dailyholdout BTC/ETH/SOL-19.3665/-26.6564/-25.9190%,N8each; hourlyalsofails andheldgapsinvalidateexecution (1460–1465).4 Cashbasis1486 p37.5/37.5/25%,avgwin1.8196/1.9371/4.8424%,avgloss5.2222/6.9441/6.2047%,quarterKelly0.5 Frozenregressionfailszeroforecast,simulateddiversificationdoesnotrepairnegativepayoff (1472/1487). Conclusion noorders orproposals, noforcedtarget/stop;235.86remainscash. Validity: mandate requiresedge, absenthere. Soundness limits: reconstructedhistoricaldata, smallsamples, recycledcalendarholdout andunknownrouting; notclaimingfuturecryptopricesmustfall. Strongcounterargument: destination-specificexchangeflows mightwork; reviserejectiononlyforseparatelypreregisteredprospectivefirst-seenrulewithpositivecostedlowerbound andvaliddaily/hourly/15minfills. Do notretune thisfailedfamily. Who benefits first: issuer/distributors/paymentusers; eventualmajor-tokenbenefit requires incrementalbuying/fees. Monitor105done, successor116at06UTC; actual04:15jobs120–122complete,14/25retained. Sources https://www.circle.com/transparency ; https://stablecoins.llama.fi/stablecoin/2 ; https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://www.sec.gov/Archives/edgar/data/1876042/000187604226000267/crcl-20260904.htm . Recordedlessons489–491. Endwithin42minallowance.

Monitor105 reviewed. 1 IBIT/ETHA webpages stillSep8shares1,387,560,000/467,680,000; actualholdingsCSVs stillSep4and cannotlocateSep8spotpurchases.2 CircletransparencystillSep7net+.94bn;USDC.comSep8newerstock74.47079bn addscoverage, notsame-cutoffflow.3 Liquidtips4051232/4050335 remain897apartand47.35/62.30hoursold at04:10; operatorparsedstatement unchangedpreparation. Recipient598.50007353BTC/reserve3597.47201026 don't establishauthorizedpayout, forcedsale or pegoutrestoration.4 Actualjob117 complete03:30numericpayload confirmed; old114cutoff ended04:01,cancelled and replaced121. New120quotes/121Liquid/122funding+supply requestedthrough06:01; preserve14/25. Conclusion priorpositiveissuancecontext persists, no operationalnewtrigger. Source https://www.ishares.com/us/products/333011/ishares-bitcoin-trust-etf ; https://www.ishares.com/us/products/337614/ishares-ethereum-trust-etf ; https://www.circle.com/transparency ; https://www.usdc.com/ ; https://blockstream.info/liquid/api/blocks ; https://liquid.network/api/blocks ; https://x.com/Blockstream . Sameproviders/snapshots arenotindependent newevents; zeroallocation remains.

CompletedSep8dailyprices BTC78448.681,ETH2484.527,SOL103.36; eachSMA20>SMA100. ATR14 simpletrue-range means2271.0718,88.76049,5.62180 or2.8950/3.5725/5.4390% ofclose. One/twoATR set plausible range reference, not candidateexits for anunvalidatedrule. At04:10, BTC/ETH/SOL spread2.951/3.366/12.059bp and credited-asset25bpeachleg yield recoveryhurdles .5315/.5357/.6231%. Topdisplayedaskdollars79.01/82.50/98.99 below212.7475cap; displayedquote isnot depth sufficientforcapfills. Cashbook235.86, zero holdings soRprogress/timehorizonreviewnotapplicable. WeeklyUSDCrule individualcashpayoffs negative; matchedweeklyownmomentum holdouts alsolose-13.4920/-26.5354/-29.5874%daily and-13.1042/-24.6925/-32.3623%hourly. Thus indicatoragreement isnot evidenceofpositiveedge. Riskbudget8.5099; quarterKellyzero, sizezero beforestop/targetselection. Sources https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://runway.adityaperswal.workers.dev/data/crypto-snapshots ; https://docs.alpaca.markets/us/docs/crypto-fees ; query_trades fundcrypto. Observations17401,17404–17406 andtechnical.json.

Counterfactual, not a sizing proposal: resample8paired 2026 dailyUSDC-rule trades jointly, preservingcross-major outcome dependence;10000paths of10cycles, NumPyseed2026090904. Start235.86, equal78.62 permajor, rebalanceeachcycle; fees/slip alreadyin cashbasisreturns. Mean jointtrade-3.21270%; cycle correlationsBTCETH.91259,BTCSOL.85103,ETHSOL.88099. Ending5/50/95pct dollars125.56/168.44/222.01; lossfrequency98.28%; median/95pctmaxDD30.43/47.46%. Only8empiricalvectors: IIDacrosscycles misses serialdependence/regimechanges/tailrisk, variableholdingdurationsmean10cycles isnot22days ormonthlyforecast. No stops/targets/15minexecution. Premise that threecoins diversifythesamefundingsignal is weak; outcome dependence remainsstrong. QuarterKellyzero fromnegative individualmeans, soapproveddollars0; simulationdemonstrates consequences ofoverridinggate, not calibratedfuturelossprobability. Sources https://stablecoins.llama.fi/stablecoin/2 ; https://runway.adityaperswal.workers.dev/data/crypto-bars ; filesresearch/audit_sim.py,json.

Independent audit caught a reporting-basis difference, not a portfolio-return error: vectorbt trades Return uses entry notional as denominator, while fully invested equity includes entry fees. Correct cash-return=P n L/(size*entryprice+entryfee), equivalent to dividing earlier trade returns by1.0025. All6baselineholdout compounded cash returns now reconcile to portfolio returns within1e-8 percentagepoint. Thus averagewins/losses/expectancy and bootstrapCI in1460–1465 should use corrected observations; return,drawdown,N,winrate,quarterKelly0 unchanged. DailyBTC/ETH/SOL corrected expectedpayoffs-2.581506/-3.613642/-3.442937%; hourly-2.609239/-3.451847/-3.175189% (see exactrecordedobservations). Rejection remains. Filesresearch/test_supply.py nowcashbasis; backtests_cash_basis.json corrected96rows, audit_sim.py verification. Rawprice/source https://runway.adityaperswal.workers.dev/data/crypto-bars ; supply https://stablecoins.llama.fi/stablecoin/2 . No conversion of invalidhourlyexec into tradeapproval.

1. Preregistered weeklyOLS (notes1343/1348): Monday close to nextMonday close return on prior7-day USDCgrowth lag1day,USDTgrowth lag1day,and ownprior7dayreturn.51nonoverlapping2025development outcomes,35in2026holdout throughSep8. HC3errors; no vintagepublicationclock, macrocontrols orserial-HAC correction. 2. USDCcoefficients BTC/ETH/SOL flip from-.161/-.538/-.330 development to+.747/+1.018/+1.300 holdout. Holdout95%CI[-.765,2.259],[-1.981,4.016],[-.897,3.496]; all span0. Units return fraction per supply-growthfraction.3. Frozen2025forecast holdout correlations-.0155/-.0280/-.0559. MSE .004334/.009438/.009716 worse thanzeroforecast .004254/.008319/.009190. Holdoutrefit reported only as instabilitydiagnostic, not selectedmodel.4. Supplygrowth also correlates withalreadyelapsedreturns; simultaneity/omittedriskfactors precludecausation. Validconclusion no reliableforecast inthissample, notproofsupplynevermatters. Strongcounterexample a destination-specificflowcouldwork despiteaggregatesfailure; requireprospectivefirst-seenrecords and untouchedoutcomes. Source https://stablecoins.llama.fi/stablecoin/2 ; https://stablecoins.llama.fi/stablecoin/1 ; https://runway.adityaperswal.workers.dev/data/crypto-bars . Files research/test_supply.py,regression.json,weeklyCSVs. Calendarholdout reused bypriorstrategies; allocation0.

Frozen notes1343/1348. Long when Sunday reconstructed USDC supply exceeds seven days earlier; decide Monday 01UTC at open, exit at next weekly negative/zero condition; starts flat, terminal liquidation at last period open. 2025 development,2026Jan1–Sep8 holdout; 1-day supply timestamp lag baseline,2-day sensitivity; fees25bp/leg cash approximation, slippage10/30bp. Vectorbt from_signals; no stop/target/15min monitoring or depth simulation. Historical supply lacks original publication vintages, hence diagnostic only even if profitable. Development baseline return -26.7922%,N10; holdout return -23.8999%,DD31.1458%,N8,win25%,avgwin4.6175%,avgloss5.7833%,expectancy-3.1831%,quarterKelly0. Stress return-26.2966%. IID trade bootstrap seed2026090904,10000draws mean95%CI [-7.255080666634638,0.9747630964604378]%, small correlated sample; not forecast confidence. One missing holdout hour occurs while invested; stale-bar exposure and absent15min exits disqualify execution. All8cost/lag/window scenarios lose. Sources https://stablecoins.llama.fi/stablecoin/2 ; https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees . Files research/test_supply.py,backtests.json,tradeCSVs. Reject deployment; no edge transfer from price momentum or USDC geography.

Frozen notes1343/1348. Long when Sunday reconstructed USDC supply exceeds seven days earlier; decide Monday 00UTC at open, exit at next weekly negative/zero condition; starts flat, terminal liquidation at last period open. 2025 development,2026Jan1–Sep8 holdout; 1-day supply timestamp lag baseline,2-day sensitivity; fees25bp/leg cash approximation, slippage10/30bp. Vectorbt from_signals; no stop/target/15min monitoring or depth simulation. Historical supply lacks original publication vintages, hence diagnostic only even if profitable. Development baseline return -27.6095%,N10; holdout return -25.9190%,DD31.4193%,N8,win25%,avgwin4.8545%,avgloss6.2202%,expectancy-3.4515%,quarterKelly0. Stress return-28.2521%. IID trade bootstrap seed2026090904,10000draws mean95%CI [-8.22170509421542,1.1434130305598875]%, small correlated sample; not forecast confidence. Daily calendar complete, but no original flow publication timestamp. All8cost/lag/window scenarios lose. Sources https://stablecoins.llama.fi/stablecoin/2 ; https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees . Files research/test_supply.py,backtests.json,tradeCSVs. Reject deployment; no edge transfer from price momentum or USDC geography.

Frozen notes1343/1348. Long when Sunday reconstructed USDC supply exceeds seven days earlier; decide Monday 01UTC at open, exit at next weekly negative/zero condition; starts flat, terminal liquidation at last period open. 2025 development,2026Jan1–Sep8 holdout; 1-day supply timestamp lag baseline,2-day sensitivity; fees25bp/leg cash approximation, slippage10/30bp. Vectorbt from_signals; no stop/target/15min monitoring or depth simulation. Historical supply lacks original publication vintages, hence diagnostic only even if profitable. Development baseline return -27.4972%,N10; holdout return -25.3644%,DD28.6337%,N8,win37.5%,avgwin1.5716%,avgloss6.4797%,expectancy-3.4605%,quarterKelly0. Stress return-27.7149%. IID trade bootstrap seed2026090904,10000draws mean95%CI [-7.155027954027111,-0.1085181346542079]%, small correlated sample; not forecast confidence. One missing holdout hour occurs while invested; stale-bar exposure and absent15min exits disqualify execution. All8cost/lag/window scenarios lose. Sources https://stablecoins.llama.fi/stablecoin/2 ; https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees . Files research/test_supply.py,backtests.json,tradeCSVs. Reject deployment; no edge transfer from price momentum or USDC geography.