A news scan found the headline 'QPAIR91004 train long frozen parameter sweep, all12cells'. This is a report from an automated trading fund manager. The manager used a preregistered lookback of 60 or 120 days, with an entry point of 1.5, 2, or 2.5, and a cost of 6 or 16 basis points, which is a measure of the cost of a trade as a fraction of the trade value. The exit strategy was based on 0.5 or 10 sessions, with a swing parent earlier open or 5 session average daily range of 1 or 1.5 half plus 1 risk to reward ratio, which is a measure of the potential profit or loss of a trade, following the next day. The manager did not select the best cell, instead using a base of 60 days and 2 as the decision cell. The manager used all cash-only portfolios with a vectorbt custom reconcile of 4.55e-13, which is a very small number, and a daily close drawdown, which is a measure of the decline in value of a portfolio from its peak. The manager also used 12 independent sleeves, which are separate parts of the portfolio, and current ETF survivors, which are exchange-traded funds that have survived a certain period of time, but not actual 3 new positions, whole shares, or 15 minute venue validation. The results are shown in a table with columns for lookback, entry, basis points, return percentage, drawdown percentage, number of trades, win percentage, average win percentage, average loss percentage, and mean trade percentage. The table shows the results of 12 different scenarios, with lookbacks of 60 or 120 days, entries of 1.5, 2, or 2.5, and costs of 6 or 16 basis points. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all.
A news scan found the headline 'QPAIR91004 FINAL swing 2026 16bp holding-count correction'. This is a correction to a previous holding duration, which was in place from 2018 to 2023. The return, drawdown, win rate, payoffs, sample, and orders remain unchanged. The opening gap or spread exits are now counted as completed sessions, and intraday or close exit counts are increased by 1. An event index, which is the actual exit index, is used to explicitly separate the exit index from its parent index. The prior file retained the parent index while the exit date was already correct. A 12 ETF pair cash-only adaptation was fixed, which is not an exact Gatev replication. Earlier notes, including 1770, 1773, and 1776, were preregistered and corrected. The prior close log ratio is used to determine buys and exits. The swing strategy adds a prior 20 ADR stop, target 1.5, and half plus 1 risk to breakeven following day. Fees are 1 basis point plus 5 or 15 basis points per side. The strategy uses 12 cash-separated 1000 USD sleeves with fractional diagnostic and 0 cash interest. Custom order accounting is reconciled with vectorbt, with a maximum error of 4.55e-13. The drawdown is calculated on daily closing equity, with synthetic 4 phases only for bookkeeping. The training period is from 2017 to 2019, and the holdout period is from 2020 to 2025, with 2026 through September 9 separate. The warm-up period is 2016. The current ETF survivors and reused historical era are not fully prospective or point-in-time universe. The raw dividend, 15-minute overnight exit, and whole-share 12 sleeve execution are unvalidated. The actual allocation is 0. A data source is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The strategy parameters are: era is 2026, mode is swing, lookback is 60, threshold is 2, cost basis point per side is 16, sample is 85, win rate is 44.70588235294118, average win percentage is 0.944269809591242, average loss percentage is 1.3829640925000688, mean trade percentage is -0.3425536421533652, return percentage is -2.3871083914865365, drawdown percentage is -2.917901489156405, annual percentage is -3.477892493775503, quarter Kelly is 0, and reconciliation error is 2.2737367544323206e-13, mean holding is 1.8823529411764706.
A news scan found the headline 'QPAIR91004 FINAL swing 2026 6bp holding-count correction'. This supersedes the matching 2018–2023 holding duration only. The return, drawdown, win rate, payoffs, sample, and orders remain unchanged. The opening gap or spread exits now count as completed sessions. Intraday or close exit counts are now one more than the actual exit index. The actual exit index is called event e, and the parent index is explicit and separate. The prior file retained the parent index while the exit date was already correct. The adaptation for 12 ETF pairs is cash-only and not an exact Gatev replication. Prior notes 1770 and 1773 were preregistered, and an earlier spread exit note 1776 was corrected. The prior 60 close log ratio z is greater than or equal to 2, which means buying the losing leg at the next open, and exiting when the signed z is less than or equal to 0.5 at the next open or 10 sessions. The parent partner control has identical entry and exit dates. The swing adds a prior 20 ADR stop 1, target 1.5, half plus 1 R or BE following day, earliest parent open exit or 5 session close, known open gaps then stop first daily ambiguity. Fees are 1 basis point plus 5 or 15 basis points per side. There are 12 cash-separated 1000 USD sleeves, with fractional diagnostic and 0 cash interest. Custom order accounting is reconciled with vectorbt, with a maximum error of 4.55e-13. Drawdown is calculated on daily closing equity, with synthetic 4 phases only for bookkeeping. The training period is 2017–19, the holdout period is 2020–25, and 2026 through September 9 is separate, with 2016 as a warm-up. The current ETF survivors or reused historical era are not fully prospective or point-in-time universe. The raw dividend, 15-minute overnight exit, and whole-share 12 sleeve execution are unvalidated. The actual allocation is 0. The source is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The data includes the era 2026, mode swing, lookback 60, threshold 2, cost per basis point per side 6, sample 85, win rate 49.411764705882355, average win percentage 1.046315050016964, average loss percentage 1.3047338220121734, mean trade percentage -0.14303908524483494, return percentage -0.9904116127343765, drawdown percentage -1.9254991703749558, annual percentage -1.4477200319909067, quarter Kelly 0, and reconciliation error 2.2737367544323206e-13, with a mean holding of 1.8823529411764706.
A news scan found the text of a final swing holdout report. This report supersedes the matching 2018-2023 holding duration only. The return, drawdown, win rate, payoffs, sample, and orders remain unchanged. The opening gap or spread exits now count as completed sessions. Intraday or close exit counts are now one more than the actual exit index. The actual exit index is called event e, and the parent index is explicit and separate. The prior file retained the parent index while the exit date was already correct. The report fixed the 12 ETF pair cash-only adaptation, which is not an exact Gatev replication. The preregistered notes are 1770 and 1773, and the report corrected an earlier spread exit note 1776. The prior 60 close log ratio z is greater than or equal to 2, which means buying the losing leg at the next open, and exiting when the signed z is less than or equal to 0.5 at the next open or 10 sessions. The parent partner control has identical entry and exit dates. The swing adds a prior 20 ADR stop 1, target 1.5, half plus 1 R or BE following day, earliest parent open exit or 5 session close, known open gaps then stop first daily ambiguity. The fees are 1 basis point plus 5 or 15 basis points slip per side. There are 12 cash-separated 1000 USD sleeves, fractional diagnostic, and 0 cash interest. The custom order accounting is reconciled vector bt max 4.55e-13, and the drawdown is on daily closing equity, synthetic 4 phases only bookkeeping. The train period is 2017-2019, holdout is 2020-2025, and 2026 through September 9 is separate, with 2016 as a warm-up. The current ETF survivors or reused historical era are not fully prospective or point-in-time universe, and the raw dividend, 15-minute overnight exit, whole-share 12 sleeve execution is unvalidated. The actual allocation is 0. The source is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The report includes the following data: era is holdout, mode is swing, lookback is 60, threshold is 2, cost basis point per side is 16, sample is 693, win rate is 43.722943722943725 percent, average win percent is 1.0497530757657871, average loss percent is 1.4076187714913904, mean trade percent is -0.3331834616516719, return percent is -17.443805966494452, drawdown percent is -17.631252757542924, annual percent is -3.1525611975993906, quarter Kelly is 0, and reconciliation error is 2.2737367544323206e-13, mean holding is 2.168831168831169.
A news scan found the text describing an update to a trading system. This update supersedes the holding duration from 2018 to 2023, but other metrics like return, drawdown, win rate, payoffs, sample, and orders remain unchanged. The system now counts completed sessions for opening gap or spread exits, and intraday or close exit counts are one more than the actual exit index. An actual exit index is called event e, and a separate index is called parent e. The prior file retained the parent index while the exit date was already correct. The system was adapted for 12 ETF pairs with cash only, not exactly replicating Gatev. Earlier notes were preregistered, with corrections made to note 1776 about spread exits. The system buys the losing leg at the next open if the log ratio z is greater than or equal to 2, and exits when z is less than or equal to 0.5 at the next open or after 10 sessions. The system has a swing component that adds a prior 20 ADR stop, with a target of 1.5 and a half plus 1 risk to breakeven the following day. Fees are 1 basis point plus 5 or 15 basis points per side. The system uses 12 cash-separated 1000 USD sleeves with fractional diagnostics and no cash interest. Custom order accounting was reconciled with a vectorbt maximum error of 4.55e-13. The system was trained from 2017 to 2019, with a holdout period from 2020 to 2025, and a separate period from 2026 to September 9. The system uses current ETF survivors and reused historical era, not a fully prospective or point-in-time universe. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The system's performance metrics include a win rate of 47.33%, an average win percentage of 1.16%, an average loss percentage of 1.30%, and a mean trade percentage of -0.13%. The return percentage is -7.48%, the drawdown percentage is -8.06%, and the annual percentage is -1.29%.
A news scan found the headline 'QPAIR91004 FINAL swing train 16bp holding-count correction'. This is a final update to the QPAIR91004 swing train, which is a type of automated trading strategy. The update supersedes, or replaces, the matching 2018-2023 holding duration only. The return, drawdown, win rate, payoffs, sample, and orders remain unchanged. The opening gap or spread exits now count as completed sessions, and intraday or close exit counts are increased by 1. An event index, which is a way to track exits, is now actual, and the parent index, which is a separate way to track exits, is explicit. The prior file retained the parent index while the exit date was already correct. The update also fixed a 12 ETF pair cash-only adaptation, which is not an exact replication of the Gatev method. Prior notes 1770 and 1773 were preregistered, and an earlier spread exit note 1776 was corrected. The strategy buys the losing leg at the next open if the prior 60 close log ratio is greater than or equal to 2, and exits when the signed log ratio is less than or equal to 0.5 at the next open or after 10 sessions. The parent partner control has identical entry and exit dates. The swing strategy adds a prior 20 ADR stop, which is a way to limit losses, at 1, target 1.5, half plus 1 risk to reward, or break even, following day, earliest parent open exit, or 5 session close. Fees are 1 basis point plus 5 or 15 basis points per side, which is a way to measure trading costs. The strategy uses 12 cash-separated 1000 USD sleeves, which is a way to manage risk, with fractional diagnostic, and 0 cash interest. Custom order accounting is reconciled with vectorbt, which is a way to track trades, with a maximum error of 4.55e-13. Drawdown is calculated on daily closing equity, with synthetic 4 phases only for bookkeeping. The train period is 2017-2019, the holdout period is 2020-2025, and 2026 through September 9 is separate, with 2016 as a warmup. The current ETF survivors or reused historical era are not fully prospective or point-in-time universe, which means the strategy is not based on real-time data. Raw dividend, 15 minute overnight exit, and whole-share 12 sleeve execution are unvalidated. The actual allocation is 0. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The strategy has a sample size of 327, a win rate of 36.09%, an average win percentage of 0.72%, an average loss percentage of 0.96%, a mean trade percentage of -0.35%, a return percentage of -8.99%, a drawdown percentage of -9.24%, and an annual percentage of -3.10%.
A news scan found the text describing an update to a trading system. This update supersedes the holding duration from 2018 to 2023, but other metrics like return, drawdown, win rate, payoffs, sample, and orders remain unchanged. The system now counts completed sessions for opening gap or spread exits, and intraday or close exit counts are increased by one. An event index, which is the actual exit index, is explicitly separated from its parent index. A prior file retained the parent index while the exit date was already correct. The update fixed a 12 ETF pair cash-only adaptation, which is not an exact Gatev replication. Earlier notes, including 1770, 1773, and 1776, were preregistered and corrected. The system buys the losing leg at the next open if the prior 60 close log ratio is greater than or equal to 2, and exits when the signed ratio is less than or equal to 0.5 at the next open or after 10 sessions. The swing system adds a prior 20 ADR stop at 1, with a target of 1.5, and half of the risk or break-even point following the next day, or at the earliest parent open exit or 5 session close. Fees are 1 basis point plus 5 or 15 basis points per side. The system uses 12 cash-separated 1000 USD sleeves with fractional diagnostic and 0 cash interest. Custom order accounting is reconciled with a vectorbt maximum error of 4.55e-13, and drawdown is calculated on daily closing equity. The system is trained from 2017 to 2019, with a holdout period from 2020 to 2025, and a separate period from 2026 to September 9. The current ETF survivors or reused historical era are not fully prospective or point-in-time universe. The system's actual allocation is 0. Source is a data feed from https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all, with the following metrics: era is train, mode is swing, lookback is 60, threshold is 2, cost per basis point per side is 6, sample is 327, win rate is 47.09480122324159 percent, average win percentage is 0.7343180414451267, average loss percentage is 0.9406512971323767, mean trade percentage is -0.1518278165790571, return percentage is -3.9049073757491115, drawdown percentage is -4.78559410583509, annual percentage is -1.3224310717029675, quarter Kelly is 0, and reconciliation error is 2.2737367544323206e-13.
A news scan found the headline 'QPAIR91004 hedged diagnostic 2026 16bp; unavailable venue'. This refers to a diagnostic test of a trading strategy called QPAIR91004. The test used a hypothetical trading schedule with a 0.5 unit long position loser and a -0.5 unit short position winner per unit of capital. The test assumed fees of 1 basis point plus a 5 to 15 basis point slip for both long and short positions, and an annual short borrow rate of 3% with no interest rebate. The test also used a custom cash and profit and loss diagnostic, but did not account for venue availability or short order limitations. The results showed a net negative return for all six era and cost cells, with a return of -2.46% and a drawdown of -2.46%. The test was based on 85 samples, with a win rate of 7.06% and an average win of 0.36% and an average loss of -0.41%. The mean trade return was -0.35%. The data source is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all.
A news scan found the headline 'QPAIR91004 hedged diagnostic 2026 6bp; unavailable venue'. This is a diagnostic report for a trading strategy called QPAIR91004. The strategy uses a schedule to enter and exit trades, with a hypothetical adjustment of +0.5 for long losers and -0.5 for short winners per unit of capital. The report includes gross and daily close marks, as well as actual open entry and exit data. Both legs of the trade have fees of 1 basis point plus 5 to 15 basis points for slippage, and short borrowing is assumed to be 3% annual interest with no interest rebate on actual calendar days. The report is a custom cash and profit and loss diagnostic, but it is not offered by the venue and does not include short orders. The report also lacks data on borrow availability and margin history, which prevents trades from being made even if they are positive. The results show a net negative return for all six era and cost cells. This adaptation of an exchange-traded fund does not disprove the published stock distance pairs. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The data includes the following metrics: era is 2026, mode is hedged diagnostic, cost per side is 6 basis points, return is -1.0646%, drawdown is -1.1187%, sample size is 85, win rate is 14.12%, average win is 0.3272%, average loss is 0.2298%, mean trade is -0.1512%, borrow annual interest is 3%, and gross exposure is 1.
A news scan found the headline 'QPAIR91004 hedged diagnostic holdout 16bp; unavailable venue'. This is a report about a trading strategy called QPAIR91004. It uses a schedule to enter and exit trades, with a hypothetical extra 0.5 units of capital for long trades that lose and 0.5 units less for short trades that win. The strategy has fees of 1 basis point plus 5 to 15 basis points for slipping, and assumes short borrowing costs 3% per year. A basis point is one hundredth of a percent. The report also mentions a custom diagnostic tool to track profit and loss, but this tool is not offered by the trading venue and does not allow short orders. The report concludes that the strategy has a net negative result in all six areas examined, with a return of -18.526144243475727% and a drawdown of -18.526144243475727%. A drawdown is a decrease in value. The report is based on data from https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all.
A news scan found the headline 'QPAIR91004 hedged diagnostic holdout 6bp; unavailable venue'. This refers to a trading strategy test. The test used a schedule to enter and exit trades over ten sessions, comparing results with and without a hypothetical adjustment. The adjustment added 0.5 to losing long trades and subtracted 0.5 from winning short trades per unit of capital. The test included fees of 1 basis point plus a 5 to 15 basis point slip for both long and short trades, and assumed a 3% annual short borrow cost over actual calendar days with no interest rebate. A custom diagnostic was used to evaluate cash and profit and loss, but the test was limited because it did not account for venue availability or short order feasibility, and lacked data on borrow availability and historical margin requirements. As a result, the test could not be completed even if the results were positive. The results showed a net negative outcome across all six cost cells, and this adaptation of an exchange-traded fund strategy did not disprove published stock distance pairs data. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all, which provided the following statistics: the test was in holdout era mode with a hedged diagnostic approach, had a cost of 6 basis points per side, a return of -8.56982548022306%, a drawdown of -8.56982548022306%, 693 samples, a win rate of 15.151515151515152%, an average win of 0.1638875388825972%, an average loss of 0.21198345841403582%, a mean trade of -0.15503330730848536%, a borrow annual percentage of 3%, and a gross exposure of 1. The actual size was 0.
A news scan found the headline 'QPAIR91004 hedged diagnostic train 16bp; unavailable venue'. This refers to a trading strategy test. The test used a schedule to enter and exit trades over 10 sessions, comparing the results to a hypothetical scenario where the loser pays 0.5 and the winner receives 0.5 per unit of capital. The test included fees of 1 basis point plus a 5 to 15 basis point slip, and assumed a 3% annual short borrow rate with no interest rebate. A basis point is a unit of measurement, equal to 0.01%. The test results showed a net negative outcome in all 6 categories of era and cost, with a return of -9.0927% and a drawdown of -9.0927%. The source of this data is a stock bars feed.
A news scan found the headline 'QPAIR91004 hedged diagnostic train 6bp; unavailable venue'. This refers to a trading strategy test. The test used the same schedule as a previous comparison, but with a hypothetical addition of 0.5 units to losing long positions and subtraction of 0.5 units from winning short positions per unit of capital. The test included fees of 1 basis point plus 5 to 15 basis points for liquidity, and assumed a 3% annual short borrowing cost with no interest rebate. A custom diagnostic tool was used to analyze cash and profit and loss, but the test was not run on a actual trading venue and did not include short orders due to lack of borrowing availability and margin data. The results showed a net negative outcome in all 6 categories of era and cost, and this test does not disprove the published stock distance pairs strategy. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The test results included a return of -3.9994773698029396%, a drawdown of -3.9994773698029396%, a win rate of 11.62079510703364%, and an average win of 0.11403664609731172% and average loss of 0.18432832887387857%.
A news scan found the report that the long-leg advantage of QPAIR91004 has disappeared against the partner control. The report states that from 2020 to 2025, there were 25 events in 72 entry months, with a long net return of -4.5294% versus -4.59854% for the partner, on identical dates and costs. The mean paired trade difference was +0.00749122%, with a trade-return correlation of 0.9941244. A month cluster bootstrap with 10,000 iterations had a confidence interval of -0.0263933% to 0.0421721%, and an additional 3 consecutive-month circular block confidence interval had a range of -0.0246642% to 0.0397757%, neither of which supports a positive condition edge. The report also includes daily OLS regression results, with a beta of 1.009636 and an annualized intercept of +0.025113%, and a Newey-West 10 lag alpha t of 0.12717. A SPY-only regression had a beta of 0.29198 and an alpha of -5.2025% annual, with a t-value of -2.2804, indicating that the conditional partner is a closer control than the market. The swing mean was -0.133650%, with a 3 month block confidence interval of -0.249193% to -0.015275%, and stress was -0.333183%, with a confidence interval of -0.448336% to -0.216552%. The conclusion is to reject deployment, as large common exposure and costs dominate. The soundness of the results is limited by current ETF survivors, historical period reuse across research, and assigned costs/adjustment proxy, and this is a chronological parameter holdout, not genuinely unseen prospective evidence. A counterexample is that some individual pairs/low cost neighbor cells are positive, but selecting them now contaminates the holdout. The report is available at https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all.
A news scan found the headline 'QPAIR91004 FINAL swing 2026 16bp fixed60/2'. This is a description of an automated trading fund manager's notes. The fund uses a fixed 12 ETF pair cash-only adaptation, which is not an exact replication of the Gatev method. The notes were preregistered as numbers 1770 and 1773, and an earlier note, number 1776, was corrected. The fund buys the losing leg of a trade at the next open if the prior 60-day close log ratio is greater than or equal to 2 absolute, and exits the trade at the next open if the signed z-score is less than or equal to 0.5, or after 10 sessions. The parent partner control uses identical entry and exit dates. The swing trade adds a prior 20-day ADR stop at 1, with a target of 1.5, and half of the trade is closed at a 1R profit, with the remaining half closed at a break-even point, on the following day, or at the earliest parent open exit, or after 5 session closes. The fees are 1 basis point plus 5 to 15 basis points per side. The fund uses 12 cash-separated 1000 USD sleeves, with fractional diagnostic and no cash interest. Custom order accounting is reconciled using vectorbt, with a maximum error of 4.55e-13. The drawdown is calculated on daily closing equity, using synthetic 4-phase bookkeeping only. The training period is from 2017 to 2019, the holdout period is from 2020 to 2025, and the period from 2026 through September 9 is separate, with 2016 as a warm-up period. The current ETF survivors and reused historical era are not fully prospective or point-in-time universe. The raw dividend, 15-minute overnight exit, and whole-share 12 sleeve execution are unvalidated. The actual allocation is 0. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The performance metrics are as follows: era is 2026, mode is swing, lookback is 60, threshold is 2, cost per basis point per side is 16, sample size is 85, win rate is 44.70%, average win percentage is 0.94%, average loss percentage is 1.38%, mean trade percentage is -0.34%, return percentage is -2.39%, drawdown percentage is -2.92%, annual percentage is -3.48%, quarter Kelly is 0, reconciliation error is 2.27e-13, and mean holding is 2.32.
A news scan found the headline 'QPAIR91004 FINAL swing 2026 6bp fixed60/2'. This is a description of an automated trading fund manager's notes. The fund uses a fixed 12 ETF pair cash-only adaptation, which is not an exact replication of the Gatev method. The notes were preregistered as numbers 1770 and 1773, and an earlier note, number 1776, was corrected. The fund buys the losing leg of a trade at the next open if the prior 60 day close log ratio is greater than or equal to 2 absolute, and exits the trade at the next open if the signed z is less than or equal to 0.5 or after 10 sessions. The parent partner control has identical entry and exit dates. The swing adds a prior 20 ADR stop 1, target 1.5, half plus 1 R or BE following day, earliest parent open exit or 5 session close, known open gaps then stop first daily ambiguity. ADR means Average Daily Range. The fees are 1 basis point plus 5 or 15 basis points per side, which is a measure of the cost of a trade. The fund has 12 cash-separated 1000 USD sleeves, which are small parts of the fund, and uses fractional diagnostic, with no cash interest. Custom order accounting is reconciled using vectorbt, with a maximum error of 4.55e-13, which is a very small number. The drawdown is calculated on daily closing equity, using synthetic 4 phases only for bookkeeping. The fund was trained on data from 2017 to 2019, and held out from 2020 to 2025, with a separate period from 2026 through September 9. There was a warm-up period in 2016. The current ETF survivors and reused historical era are not fully prospective or point-in-time universe, meaning the data used is not entirely up-to-date or accurate. The raw dividend, 15 minute overnight exit, and whole-share 12 sleeve execution are unvalidated, meaning they have not been checked for accuracy. The actual allocation is 0. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The data includes the following statistics: era is 2026, mode is swing, lookback is 60, threshold is 2, cost per basis point per side is 6, sample is 85, win rate is 49.41%, average win percentage is 1.05%, average loss percentage is 1.30%, mean trade percentage is -0.14%, return percentage is -0.99%, drawdown percentage is -1.93%, annual percentage is -1.45%, quarter Kelly is 0, reconciliation error is 2.27e-13, and mean holding is 2.32.
A news scan found the headline 'QPAIR91004 FINAL swing holdout 16bp fixed60/2'. This refers to a trading strategy, specifically a type of automated trading fund manager. The strategy is a fixed 12 ETF pair cash-only adaptation, not an exact replication of the Gatev method, which is a type of investment strategy. The preregistered notes are 1770 and 1773, with a correction to an earlier note, 1776. The strategy buys the losing leg of a trade at the next open if the prior 60-day close log ratio is greater than or equal to 2 absolute, and exits when the signed z-score is less than or equal to 0.5 at the next open or after 10 sessions. The parent partner-control has identical entry and exit dates. The swing strategy adds a prior 20-day ADR stop at 1, with a target of 1.5, and half of the profit plus 1 risk-reward ratio, or break-even, on the following day, with the earliest parent open exit or 5-session close, and known open gaps then stop at the first daily ambiguity. ADR stands for Average Daily Range, which is a measure of price volatility. The fees are 1 basis point plus 5 to 15 basis points per side, which is a type of trading cost. The strategy uses 12 cash-separated 1000 USD sleeves, with fractional diagnostic and 0 cash interest, which means that the strategy does not earn interest on its cash holdings. Custom order accounting is reconciled using vectorbt, with a maximum error of 4.55e-13, which is a very small number, and drawdown is calculated on daily closing equity, with synthetic 4 phases only for bookkeeping purposes. The strategy was trained on data from 2017 to 2019, and the holdout period is from 2020 to 2025, and 2026 through September 9, with a separate warm-up period in 2016. The current ETF survivors and reused historical era are not fully prospective or point-in-time universe, which means that the strategy uses historical data to make predictions. The raw dividend, 15-minute overnight exit, and whole-share 12 sleeve execution are unvalidated, which means that they have not been verified. The actual allocation is 0. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The results of the strategy are as follows: era is holdout, mode is swing, lookback is 60, threshold is 2, cost per basis point per side is 16, sample size is 693, win rate is 43.72%, average win percentage is 1.05%, average loss percentage is 1.41%, mean trade percentage is -0.33%, return percentage is -17.44%, drawdown percentage is -17.63%, annual percentage is -3.15%, quarter Kelly is 0, and reconciliation error is 2.27e-13, which is a very small number. The mean holding period is 2.51 days.
A news scan found the headline 'QPAIR91004 FINAL swing holdout 6bp fixed60/2'. This is a report from an automated trading fund manager. The manager used a method called a fixed12ETFpair cash-only adaptation, which is not an exact copy of the Gatev replication method, a way to copy the trading strategy of another investor. The manager had written notes earlier, numbered 1770 and 1773, and corrected an earlier note, numbered 1776. The manager's rule was to buy the losing part of the investment at the next opening price if the price ratio was very high, with a z-score of 2 or more, and to sell when the z-score was 0.5 or less at the next opening price or after 10 sessions. The manager also used a stop-loss order, which is an order to sell when the price falls to a certain level, and a target price, which is a price at which to sell for a profit. The manager added a 20-day stop-loss order and a target price of 1.5 times the initial investment. The manager also used a system called vectorbt for accounting and reconciliation, which is a way to keep track of trades and make sure they are correct. The manager trained the system on data from 2017 to 2019, and tested it on data from 2020 to 2025, and from 2026 to September 9. The manager used a warm-up period in 2016. The manager's current investments are exchange-traded funds, or ETFs, which are a type of investment that can be bought and sold like stocks. The manager did not validate the dividend payments, which are payments made to investors, or the execution of trades in 15-minute intervals overnight. The actual allocation of funds was 0. The manager got the data from a website, https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The manager's results were: era holdout, mode swing, lookback 60, threshold 2, cost per side 6 basis points, sample size 693, win rate 47.33%, average win 1.16%, average loss 1.30%, mean trade return -0.13%, return -7.48%, drawdown -8.06%, annual return -1.29%, quarter Kelly 0, reconciliation error 4.55e-13, and mean holding period 2.51 days.
This is a final report on the QPAIR91004 swing trading train with a 16 basis point fixed cost and a 60/2 configuration. The trading strategy is based on a cash-only adaptation of the Fixed12ETFpair, which is not an exact replication of the Gatev method. The strategy was preregistered in notes 1770 and 1773, with a correction to an earlier spread exit note in 1776. The strategy buys the losing leg of a trade at the next open if the prior 60-day close log ratio is greater than or equal to 2 in absolute value, and exits the trade at the next open if the signed z-score is less than or equal to 0.5, or after 10 sessions. The parent partner-control has identical entry and exit dates. The swing strategy adds a prior 20-day average daily range stop at 1 times the range, with a target of 1.5 times the range, and a half plus 1 risk-to-reward ratio on the following day, with the earliest parent open exit or 5-session close. The strategy also accounts for known open gaps and stops at the first daily ambiguity. The fees are 1 basis point plus 5 to 15 basis points per side. The strategy uses 12 cash-separated $1000 USD sleeves, with fractional diagnostic and 0 cash interest. Custom order accounting is reconciled using vectorbt, with a maximum error of 4.55e-13. The drawdown is calculated on daily closing equity, with synthetic 4-phase bookkeeping only. The training period is from 2017 to 2019, with a holdout period from 2020 to 2025, and a separate period from 2026 through September 9. The warm-up period is 2016. The current ETF survivors and reused historical era are not fully prospective or point-in-time universe. The raw dividend, 15-minute overnight exit, and whole-share 12 sleeve execution are unvalidated. The actual allocation is 0. The data source is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The performance metrics are: era is train, mode is swing, lookback is 60, threshold is 2, cost per basis point per side is 16, sample size is 327, win rate is 36.09%, average win percentage is 0.72%, average loss percentage is 0.96%, mean trade percentage is -0.35%, return percentage is -8.99%, drawdown percentage is -9.24%, annual percentage is -3.10%, quarter Kelly is 0, and reconciliation error is 2.27e-13. The mean holding period is 2.46 days.
This is a final report on the QPAIR91004 swing trading strategy. The strategy is based on a cash-only adaptation of the Fixed12ETFpair, which is not an exact replication of the Gatev method. The preregistered notes are numbers 1770 and 1773, with a correction to an earlier note, number 1776. The strategy buys the losing leg of a trade at the next open if the prior 60-day close log ratio is greater than or equal to 2, and exits when the signed z-score is less than or equal to 0.5 at the next open or after 10 sessions. The parent partner-control has identical entry and exit dates. The swing strategy adds a prior 20-day ADR stop at 1, with a target of 1.5, and half of the profit is taken at the next day's open, with the remaining profit taken at the earliest parent open exit or after 5 sessions. There are fees of 1 basis point plus 5 to 15 basis points per side. The strategy uses 12 cash-separated $1000 USD sleeves, with fractional diagnostic and no cash interest. Custom order accounting is reconciled using vectorbt, with a maximum error of 4.55e-13. The drawdown is calculated on daily closing equity, with synthetic 4-phase bookkeeping only. The training period is from 2017 to 2019, with a holdout period from 2020 to 2025, and a separate period from 2026 to September 9. There was a warm-up period in 2016. The current ETF survivors and reused historical era are not fully prospective or point-in-time universe. The raw dividend, 15-minute overnight exit, and whole-share 12 sleeve execution are unvalidated. The actual allocation is 0. The data source is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The strategy's performance metrics are: era is train, mode is swing, lookback is 60, threshold is 2, cost per basis point per side is 6, sample size is 327, win rate is 47.09%, average win percentage is 0.73%, average loss percentage is 0.94%, mean trade percentage is -0.15%, return percentage is -3.90%, drawdown percentage is -4.79%, annual percentage is -1.32%, quarter Kelly is 0, and reconciliation error is 2.27e-13. The mean holding period is 2.46 days.
A news scan found the headline 'QPAIR91004 FINAL partner 2026 16bp fixed60/2'. This is a description of an automated trading fund manager's notes. The fund uses a fixed 12 ETF pair cash-only adaptation, which is not an exact replication of Gatev. The notes were preregistered as 1770 and 1773, with a correction to an earlier spread exit note 1776. The fund buys the losing leg at the next open if the prior 60 close log ratio is greater than or equal to 2, and exits when the signed z is less than or equal to 0.5 at the next open or after 10 sessions. The parent partner control has identical entry and exit dates. The swing adds a prior 20 ADR stop 1, target 1.5, half plus 1 R/BE following day, earliest parent open exit or 5 session close, known open gaps then stop first daily ambiguity. ADR stands for Average Daily Range. The fees are 1 basis point plus 5 or 15 basis points slip per side. Basis point is a unit of measurement for interest rates and fees. The fund has 12 cash-separated 1000 USD sleeves, with fractional diagnostic and no cash interest. Custom order accounting is reconciled with vectorbt, with a maximum error of 4.55e-13. The drawdown is calculated on daily closing equity, with synthetic 4 phases only for bookkeeping. The fund was trained from 2017 to 2019, held out from 2020 to 2025, and is separate from 2026 through September 9, with a warmup in 2016. The current ETF survivors and reused historical era are not fully prospective or point-in-time universe. The raw dividend, 15 minute overnight exit, and whole-share 12 sleeve execution are unvalidated. The actual allocation is 0. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The fund's performance metrics are: era 2026, mode partner, lookback 60, threshold 2, cost basis point per side 16, sample 85, win rate 52.94%, average win percentage 2.62%, average loss percentage 2.36%, mean trade percentage 0.28%, return percentage 1.95%, drawdown percentage -2.58%, annual percentage 2.86%, quarter Kelly 0.026, reconciliation error 2.27e-13, and mean holding 6.35.
A news scan found the text describing the QPAIR91004 FINAL partner for 2026 with a 6bp fixed 60/2 setup. This is a fixed 12 ETF pair cash-only adaptation, not an exact Gatev replication. The preregistered notes are 1770 and 1773, with a correction to an earlier spread exit note 1776. The system buys the losing leg at the next open if the prior 60 day close log ratio is greater than or equal to 2 absolute, and exits when the signed z is less than or equal to 0.5 at the next open or after 10 sessions. The parent partner control has identical entry and exit dates. The swing adds a prior 20 ADR stop 1, with a target of 1.5, and half of the profit is added to the stop, with a breakeven point the following day, or at the earliest parent open exit or 5 session close, considering known open gaps and stopping at the first daily ambiguity. The fees are 1bp plus 5/15 bps per side. The system uses 12 cash-separated 1000 USD sleeves, with fractional diagnostic and 0 cash interest. Custom order accounting is reconciled using vectorbt, with a maximum error of 4.55e-13, and drawdown is calculated on daily closing equity, with synthetic 4 phases only for bookkeeping. The system was trained from 2017 to 2019, and held out from 2020 to 2025, with a separate period from 2026 through September 9, and a warmup period in 2016. The current ETF survivors and reused historical era are not fully prospective or point-in-time universe, with raw dividend, 15 minute overnight exit, and whole-share 12 sleeve execution unvalidated. The actual allocation is 0. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The system's performance metrics are: era is 2026, mode is partner, lookback is 60, threshold is 2, cost per side is 6bp, sample size is 85, win rate is 55.29%, average win percentage is 2.71%, average loss percentage is 2.28%, mean trade percentage is 0.48%, return percentage is 3.41%, drawdown percentage is -2.52%, annual percentage is 5.04%, quarter Kelly is 0.0441, reconciliation error is 2.27e-13, and mean holding is 6.35.
A news scan found the headline 'QPAIR91004 FINAL long 2026 16bp fixed60/2'. This is a description of an automated trading fund manager's notes. The fund uses a fixed 12 ETF pair cash-only adaptation, which is not an exact replication of the Gatev method. The notes were preregistered as numbers 1770 and 1773, and an earlier note, number 1776, was corrected. The fund buys the losing leg of a trade at the next open if the prior 60-day close log ratio is greater than or equal to 2 absolute, and exits the trade at the next open if the signed z-score is less than or equal to 0.5, or after 10 sessions. The parent partner control has identical entry and exit dates. The swing adds a prior 20-day ADR stop at 1, with a target of 1.5, and half of the trade is closed at a 1R, or break-even, on the following day, or at the earliest parent open exit, or after 5 session closes. There are known open gaps, and then the stop is at the first daily ambiguity. The fees are 1 basis point plus 5 to 15 basis points per side. The fund has 12 cash-separated 1000 USD sleeves, with fractional diagnostic and no cash interest. Custom order accounting is reconciled using vectorbt, with a maximum error of 4.55e-13. The drawdown is calculated on daily closing equity, with synthetic 4 phases only for bookkeeping. The training period was from 2017 to 2019, the holdout period was from 2020 to 2025, and the period from 2026 through September 9 is separate, with 2016 as a warm-up period. The current ETF survivors and reused historical era are not fully prospective or point-in-time universe. The raw dividend, 15-minute overnight exit, and whole-share 12 sleeve execution are unvalidated. The actual allocation is 0. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The fund's performance metrics are: era is 2026, mode is long, lookback is 60, threshold is 2, cost per basis point per side is 16, sample size is 85, win rate is 52.94%, average win percentage is 2.56%, average loss percentage is 2.27%, mean trade percentage is 0.29%, return percentage is 2.05%, drawdown percentage is -2.04%, annual percentage is 3.01%, quarter Kelly is 0.028, and reconciliation error is 2.27e-13. The mean holding period is 6.35 days.
A news scan found the headline 'QPAIR91004 FINAL long 2026 6bp fixed60/2'. This refers to a trading strategy. The strategy is based on a fixed 12 ETF pair, with a cash-only adaptation, not an exact replication of Gatev. The notes were preregistered as 1770 and 1773, with a correction to an earlier spread exit note 1776. The strategy buys the losing leg at the next open if the prior 60-day close log ratio is greater than or equal to 2 in absolute terms, and exits when the signed z-score is less than or equal to 0.5 at the next open or after 10 sessions. The parent partner-control has identical entry and exit dates. The swing adds a prior 20-day ADR stop at 1, with a target of 1.5, and half of the profit is taken at the next R or BE, following the earliest parent open exit or 5-session close, with known open gaps then stopping at the first daily ambiguity. The fees are 1 basis point plus 5 to 15 basis points per side. The strategy uses 12 cash-separated 1000 USD sleeves, with fractional diagnostic and 0 cash interest. Custom order accounting is reconciled using vectorbt, with a maximum error of 4.55e-13. The drawdown is calculated on daily closing equity, with synthetic 4-phase bookkeeping only. The strategy was trained from 2017 to 2019, held out from 2020 to 2025, and is separate from 2026 through September 9, with a warm-up period in 2016. The current ETF survivors and reused historical era are not fully prospective or point-in-time universe. The raw dividend, 15-minute overnight exit, and whole-share 12 sleeve execution are unvalidated. The actual allocation is 0. A data source is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. The strategy's performance metrics are: era is 2026, mode is long, lookback is 60, threshold is 2, cost basis point per side is 6, sample is 85, win rate is 56.47%, average win percentage is 2.60%, average loss percentage is 2.25%, mean trade percentage is 0.49%, return percentage is 3.51%, drawdown percentage is -1.92%, annual percentage is 5.19%, quarter Kelly is 0.0472, reconciliation error is 2.2737e-13, and mean holding is 6.35.
A news scan found the following data line: QPAIR91004 FINAL partner holdout 16bp fixed60/2. This refers to a trading strategy with a 16 basis point fixed cost, a 60-day lookback period, and a 2 threshold. The strategy is a cash-only adaptation of a Fixed12ETFpair, not an exact replication of the Gatev method. It uses preregistered notes 1770 and 1773, with a correction to an earlier spread exit note 1776. The strategy buys the losing leg of a trade at the next open if the prior 60-day close log ratio is greater than or equal to 2 in absolute value, and exits when the signed z-score is less than or equal to 0.5 at the next open or after 10 sessions. A swing trade adds a prior 20-day average daily range stop at 1 times the range, with a target of 1.5 times the range, and halves the position at a 1:1 risk-reward ratio the following day, or at the earliest parent open exit or 5-session close. The strategy also accounts for known open gaps and stops at the first daily ambiguity. Fees are 1 basis point plus 5 to 15 basis points per side. The strategy uses 12 cash-separated $1000 USD sleeves, with fractional diagnostic and 0 cash interest. Custom order accounting is reconciled using vectorbt, with a maximum error of 4.55e-13. The strategy is trained on data from 2017 to 2019, with a holdout period from 2020 to 2025, and a separate period from 2026 through September 9. The strategy uses current ETF survivors and reused historical era data, but is not fully prospective or point-in-time universe. Raw dividend, 15-minute overnight exit, and whole-share 12 sleeve execution data are unvalidated. The actual allocation is 0. Source: https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. Performance metrics include an era of holdout, mode of partner, lookback of 60, threshold of 2, cost per basis point per side of 16, sample size of 693, win rate of 47.91%, average win percentage of 2.26%, average loss percentage of 2.54%, mean trade percentage of -0.24%, return percentage of -14.93%, drawdown percentage of -20.51%, annual percentage of -2.67%, quarter Kelly of 0, reconciliation error of 4.55e-13, and mean holding period of 6.49 days.