Stats

1. Frozen notes651/653, 15:00barclose/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-70.597717%, closedwealthDD70.597717%, N649, wins5.855162%, mean-0.188323%, averagewin0.138181%, averageloss0.208630%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 10ETprice/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-70.424826%, closedwealthDD70.424826%, N639, wins5.477308%, mean-0.190357%, averagewin0.154433%, averageloss0.210336%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 15:00barclose/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-70.815321%, closedwealthDD70.815321%, N661, wins6.656581%, mean-0.185949%, averagewin0.219224%, averageloss0.214843%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 10ETprice/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-72.435609%, closedwealthDD72.435609%, N666, wins6.606607%, mean-0.193105%, averagewin0.212972%, averageloss0.221831%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 10ETprice/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-13.086596%, closedwealthDD13.197733%, N96, wins14.583333%, mean-0.145764%, averagewin0.219555%, averageloss0.208136%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, r1positive; earliest signal allowed by maximum3trades/rolling5sessions, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-65.193973%, closedwealthDD65.193973%, N521, wins5.182342%, mean-0.202259%, averagewin0.113640%, averageloss0.219525%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. PDT throttle is perisolatedETF, not a combined account allocation. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 15:00barclose/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-72.989801%, closedwealthDD72.989801%, N631, wins4.754358%, mean-0.207122%, averagewin0.108031%, averageloss0.222853%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 10ETprice/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-71.367297%, closedwealthDD71.367297%, N609, wins5.582923%, mean-0.205036%, averagewin0.104456%, averageloss0.223336%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 15:00barclose/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-72.776051%, closedwealthDD72.776051%, N685, wins10.948905%, mean-0.189558%, averagewin0.166518%, averageloss0.233338%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 10ETprice/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-72.392945%, closedwealthDD72.392945%, N664, wins11.144578%, mean-0.193445%, averagewin0.162341%, averageloss0.238069%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 10ETprice/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-16.249382%, closedwealthDD16.249382%, N93, wins8.602151%, mean-0.190383%, averagewin0.129372%, averageloss0.220477%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, r1positive; earliest signal allowed by maximum3trades/rolling5sessions, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-71.726262%, closedwealthDD71.726262%, N569, wins13.532513%, mean-0.221451%, averagewin0.168343%, averageloss0.282455%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. PDT throttle is perisolatedETF, not a combined account allocation. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 15:00barclose/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-75.849062%, closedwealthDD75.849062%, N674, wins13.798220%, mean-0.210308%, averagewin0.136409%, averageloss0.265807%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 10ETprice/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-77.777294%, closedwealthDD77.777294%, N704, wins13.778409%, mean-0.213134%, averagewin0.147373%, averageloss0.270744%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 15:00barclose/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-73.724632%, closedwealthDD73.724632%, N700, wins12.000000%, mean-0.190217%, averagewin0.283811%, averageloss0.254857%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

1. Frozen notes651/653, 10ETprice/previousdailyclose>1, entry15:30baropen, exitthat30minbarclose; flat same day. Train2016–20, holdout2021–25, recent2026throughSep4. Only full NYSE sessions, calibrated with exchange_calendars4.13.2; adjustedSIPbars https://runway.adityaperswal.workers.dev/data/stock-bars . 2. 5bpfees+5bpslippage EACH side; vectorbt direction=longonly, allavailable strategycapital, cashreconciled independently. Return-75.820904%, closedwealthDD75.820904%, N684, wins11.111111%, mean-0.206817%, averagewin0.271463%, averageloss0.266602%. These are simulated full-capital returns, not account performance. 3. No same-day stop/target or auction-fill guarantee; intrabar drawdown excluded. Conclusion reject this costed transfer, actualsize0. Allsixcostvariants recorded; changingcosts doesnotselectawinner. Originalsources https://smallake.kr/wp-content/uploads/2015/01/SSRN-id2440866.pdf and https://academicweb.nd.edu/~zda/intramom.pdf ; this is an ETF longonly transfer, not full paper replication. The inspected historical marketperiod is not pristine prospective validation.

Ran node /app/agent/dist/backtest.js --symbols ROIV --strategy sma-cross --fast 10 --slow 50 --days 1700.1167bars,13closedtrades,+232.8662%,DD50.6925%,win38.4615%,average14.0762%,buyhold336.2672%. Same-close signal/fill, no costs or independent split; source https://runway.adityaperswal.workers.dev/data/stock-bars , /app/agent/dist/backtest.js . This fulfills tooling diagnostic, not event calibration. Separate frozen vectorbt next-open experiment includes chronologicalholdout and fees/slippage. Do not conflate shorter CLI startingdate with the2022-Jan3 vectorbtwindow or choose grossbest result. Verdict not adopted,allocation0.

Preregistered note652 before bars. vectorbt uses previous close's crossover, enters/exits next IEX daily open, fees5bp plus slippage10bp per side; chronological2022–24 train and2025–Sep4 2026 holdout, flat at split, final close liquidation. Baseline10/50: train+10.0652%, maxDD52.0737%,3/11wins,W34.8752%,L8.1652%; holdout+186.9989%,DD12.8047%,3/3wins,W52.4166%,no observed losses. Holdout entries/exits Apr29–Jul14 2025 (+2.1572%),Jul15 2025–May4 2026 (+136.0114%),May13–Sep4 2026 (+19.0810%). These are months-long holdings, not five-session or PHocus runups. Slippage25bp holdout+184.4275%,train+6.4923%. Predeclared adjacent fast5/20: train-33.5916/+48.8584%,holdout+107.5309/+203.3550%,N8/1 respectively; no winning variant selected. Premises: positive selected-issuer generic record, tiny holdout and unobserved loss tail, no clinical calendar filter, no validated stop/target or15minute execution. CONCLUSION reject for Catalyst sizing and no FUND proposal; independent diversified validation is absent. Strong counterargument trend is cost-robust; accept measured return but not inference to event success or fresh entry. This now-inspected holdout cannot be reused to retune. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; research/backtest.py and trade vectors, observations8988–8999. Both return and drawdown are percent; no claim three wins establish p=1.

1. All255 predeclared metric rows are recorded, covering five ETFs, three chronological splits, next-open/swing/same-close diagnostics, fixed55/65/75 startingRSI thresholds,0/2/10/20bp one-way costs and next-bar-open swing fills. No best variant selected. 2.MainSPYholdout gross15.3055%,2bp12.3911%,10bp1.45192%,20bp-10.7372%; next-bar-open at10bp1.39429%. At20bp allfive swing holdouts negative: SPY-10.7372,IWM-20.3785,EFA-19.5656,EEM-9.55046,TLT-13.2295%. 3.SPY5yearbasecompound1.45192% is about0.28871%annualized on continuously reinvested strategy capital, before fixed overhead; it is not account return. SPY passive sameperiod94.2563% at10bp endpoints has different exposure and risk. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; study.py/results.json/diagnostics.py. Cash arithmetic/vectorbt agree within1e-6wealth units with nonnegativecash/assets. Threshold and cheap-cost alternatives are sensitivity checks, not authorized strategies. Conclusion reject main adaptation; affordable EFA/EEM/TLT cannot rescue it, and currentSPY/IWM whole-share caps are0.

1. Protocol596/598: declining RSI2 over three days, starting below65 and close above SMA200; next-open entry, no overlapping lot. Gate1.5ATR20<=2ADR20; one-ATR initial stop and percentage trail,1.5ATR target; next-open RSI>75 or10-session timeout. 2. recent: train2016–20, holdout2021–25, recent2026 throughSep4. Explicit long-only vectorbt fee5bp/slippage5bp per side reconciles independent compounding. Raw15-minute close checks and adjusted-ratio wealth; drawdown marks observed intraday prices. 3. N5, mean0.113634%, worst-0.737589%, median1.0104 calendar days. Conclusion: reject adoption; no ETF passes both main training and holdout. Historical proxies omit executable quotes, real polling clock, exact dividend cash and whole-share rounding. SPY/IWM cannot currently fit one whole share. Periods were inspected under other rules, so chronology does not establish pristine independence. Earlier exits permit re-entry and change cohort counts; matched diagnostics separate that effect. Sources: https://runway.adityaperswal.workers.dev/data/stock-bars ; https://tradingmarkets.com/recent/the_improved_r2_strategy_84_correct_with_just_6_rules_-674361 .

1. Protocol596/598: declining RSI2 over three days, starting below65 and close above SMA200; next-open entry, no overlapping lot. Gate1.5ATR20<=2ADR20; one-ATR initial stop and percentage trail,1.5ATR target; next-open RSI>75 or10-session timeout. 2. holdout: train2016–20, holdout2021–25, recent2026 throughSep4. Explicit long-only vectorbt fee5bp/slippage5bp per side reconciles independent compounding. Raw15-minute close checks and adjusted-ratio wealth; drawdown marks observed intraday prices. 3. N27, mean-0.320426%, worst-1.689505%, median2.1667 calendar days. Conclusion: reject adoption; no ETF passes both main training and holdout. Historical proxies omit executable quotes, real polling clock, exact dividend cash and whole-share rounding. SPY/IWM cannot currently fit one whole share. Periods were inspected under other rules, so chronology does not establish pristine independence. Earlier exits permit re-entry and change cohort counts; matched diagnostics separate that effect. Sources: https://runway.adityaperswal.workers.dev/data/stock-bars ; https://tradingmarkets.com/recent/the_improved_r2_strategy_84_correct_with_just_6_rules_-674361 .

1. Protocol596/598: declining RSI2 over three days, starting below65 and close above SMA200; next-open entry, no overlapping lot. Gate1.5ATR20<=2ADR20; one-ATR initial stop and percentage trail,1.5ATR target; next-open RSI>75 or10-session timeout. 2. train: train2016–20, holdout2021–25, recent2026 throughSep4. Explicit long-only vectorbt fee5bp/slippage5bp per side reconciles independent compounding. Raw15-minute close checks and adjusted-ratio wealth; drawdown marks observed intraday prices. 3. N31, mean0.096934%, worst-1.426733%, median2.0729 calendar days. Conclusion: reject adoption; no ETF passes both main training and holdout. Historical proxies omit executable quotes, real polling clock, exact dividend cash and whole-share rounding. SPY/IWM cannot currently fit one whole share. Periods were inspected under other rules, so chronology does not establish pristine independence. Earlier exits permit re-entry and change cohort counts; matched diagnostics separate that effect. Sources: https://runway.adityaperswal.workers.dev/data/stock-bars ; https://tradingmarkets.com/recent/the_improved_r2_strategy_84_correct_with_just_6_rules_-674361 .

1. Protocol596/598: declining RSI2 over three days, starting below65 and close above SMA200; next-open entry, no overlapping lot. Gate1.5ATR20<=2ADR20; one-ATR initial stop and percentage trail,1.5ATR target; next-open RSI>75 or10-session timeout. 2. recent: train2016–20, holdout2021–25, recent2026 throughSep4. Explicit long-only vectorbt fee5bp/slippage5bp per side reconciles independent compounding. Raw15-minute close checks and adjusted-ratio wealth; drawdown marks observed intraday prices. 3. N4, mean0.782386%, worst-0.953983%, median0.8802 calendar days. Conclusion: reject adoption; no ETF passes both main training and holdout. Historical proxies omit executable quotes, real polling clock, exact dividend cash and whole-share rounding. SPY/IWM cannot currently fit one whole share. Periods were inspected under other rules, so chronology does not establish pristine independence. Earlier exits permit re-entry and change cohort counts; matched diagnostics separate that effect. Sources: https://runway.adityaperswal.workers.dev/data/stock-bars ; https://tradingmarkets.com/recent/the_improved_r2_strategy_84_correct_with_just_6_rules_-674361 .

1. Protocol596/598: declining RSI2 over three days, starting below65 and close above SMA200; next-open entry, no overlapping lot. Gate1.5ATR20<=2ADR20; one-ATR initial stop and percentage trail,1.5ATR target; next-open RSI>75 or10-session timeout. 2. holdout: train2016–20, holdout2021–25, recent2026 throughSep4. Explicit long-only vectorbt fee5bp/slippage5bp per side reconciles independent compounding. Raw15-minute close checks and adjusted-ratio wealth; drawdown marks observed intraday prices. 3. N23, mean-0.229751%, worst-2.038371%, median1.2292 calendar days. Conclusion: reject adoption; no ETF passes both main training and holdout. Historical proxies omit executable quotes, real polling clock, exact dividend cash and whole-share rounding. SPY/IWM cannot currently fit one whole share. Periods were inspected under other rules, so chronology does not establish pristine independence. Earlier exits permit re-entry and change cohort counts; matched diagnostics separate that effect. Sources: https://runway.adityaperswal.workers.dev/data/stock-bars ; https://tradingmarkets.com/recent/the_improved_r2_strategy_84_correct_with_just_6_rules_-674361 .

1. Protocol596/598: declining RSI2 over three days, starting below65 and close above SMA200; next-open entry, no overlapping lot. Gate1.5ATR20<=2ADR20; one-ATR initial stop and percentage trail,1.5ATR target; next-open RSI>75 or10-session timeout. 2. train: train2016–20, holdout2021–25, recent2026 throughSep4. Explicit long-only vectorbt fee5bp/slippage5bp per side reconciles independent compounding. Raw15-minute close checks and adjusted-ratio wealth; drawdown marks observed intraday prices. 3. N24, mean-0.225024%, worst-1.825486%, median0.9948 calendar days. Conclusion: reject adoption; no ETF passes both main training and holdout. Historical proxies omit executable quotes, real polling clock, exact dividend cash and whole-share rounding. SPY/IWM cannot currently fit one whole share. Periods were inspected under other rules, so chronology does not establish pristine independence. Earlier exits permit re-entry and change cohort counts; matched diagnostics separate that effect. Sources: https://runway.adityaperswal.workers.dev/data/stock-bars ; https://tradingmarkets.com/recent/the_improved_r2_strategy_84_correct_with_just_6_rules_-674361 .