Stats

A news scan found the headline 'QWIDE91002 monthly 1 train 16bp/side'. This refers to a trading strategy, where 16bp/side means 16 basis points, or 0.16%, per trade, as a fee. The strategy uses fifty predetermined cash ETF sleeves, which are groups of investments. The data is split into periods: 2016 for warmup, 2017-2019 for training, 2020-2025 for holdout, and 2026 separate. The strategy looks at month-end returns minus the return of SHY, a type of investment, and mixes the positive fraction of the returns. It does not select specific investments, but rather uses a passive full sleeve approach. The next session will open with a monthly target weight within independently compounding sleeves, with 1bp fees and 5/15 slippage, which refers to the difference between expected and actual trading costs. The strategy also considers vector bt position episodes, which are periods of buying or selling, and aggregate partial resizes, which means adjusting the size of the investments. Correlated positions, or investments that move together, are not independent and identically distributed, or IID, which means they do not have an edge, or advantage. Inactive cash has zero value, and expenses are embedded in prices. This strategy is a long cash adaptation, not a risk-balanced short futures strategy like AQR, and does not have mandatory ADR exits. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all and https://www.aqr.com/-/media/AQR/Documents/Insights/Journal-Article/AQR-JPM-Fall-2017.pdf. The full metrics are: family is monthly, signal is 1, cost side is 0.0016, era is train, return is 4.107417732996743, drawdown is -17.218225034732328, win rate is 46.724890829694324, sample is 458, avg win pct is 6.127245541864814, avg loss pct is 4.37269942890957, quarter kelly diagnostic is 0.021762976210857238, annual return is 1.3554802713248826, mean daily pct is 0.0062438994733878435, vol annual pct is 6.748017945911851, sharpe zero cash is 0.23317404901790267, and open positions is 0.

A news scan found the headline 'QWIDE91002 monthly 1 2026 6bp/side'. This refers to a monthly report for the QWIDE91002 fund, which is a type of automated trading fund. The report mentions that certain positions, referred to as Frozen1720/1721/1724, have been established. The fund uses a strategy involving fifty predetermined cash ETF sleeves, which were set up in 2016 and have been used for training and testing since 2017. The fund's performance is measured against the SHY index, and the results are used to adjust the fund's target weights for the next session. The fund charges a fee of 1 basis point and also accounts for slippage, which is a type of trading cost. The report also mentions that the fund uses a vector to manage its positions and that it does not use a risk-balanced approach like the one used by AQR. The fund's metrics are also provided, including its return, drawdown, and win rate. The source of the report is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all and https://www.aqr.com/-/media/AQR/Documents/Insights/Journal-Article/AQR-JPM-Fall-2017.pdf. The full metrics are: family is monthly, signal is 1, cost per side is 0.0006, era is 2026, return is 5.012382429286144, drawdown is -7.574674181060015, win rate is 49.24242424242424, sample is 132, average win percentage is 6.280442394787109, average loss percentage is 2.536597203366168, quarter Kelly diagnostic is 0.07185508311792244, annual return is 7.406950751829333, mean daily percentage is 0.030829205775379467, annual volatility percentage is 10.996816714021016, Sharpe ratio with zero cash is 0.7064735238780646, and open positions is 0.

The system is holding a position in QWIDE91002. The system uses a set of 50 pre-determined cash ETF sleeves, which were established in 2016 and trained from 2017 to 2019, and then held out from 2020 to 2025, with a separate period in 2026. The system calculates returns at the end of each month, subtracting the return of the SHY ETF, and mixes the results with a positive fraction of the components, without selecting any specific ones, using a passive full sleeve approach. The system also considers fees of 1 basis point and slippage of 5 to 15 basis points, and liquidates positions at the end of each session, taking into account costs. The system uses vector-based position episodes and aggregate partial resizes, and assumes that correlated positions do not have independent and identically distributed edges. Inactive cash is set to zero, and expenses are embedded in prices. This system is a long cash adaptation, not a risk-balanced short futures strategy like the one used by AQR, and does not require mandatory ADR exits. The system's metrics are as follows: it has a signal of 1, a cost per side of 0.06%, and is in the holdout era, with a return of 36.36%, a drawdown of -22.15%, and a win rate of 46.59%. The system has an average win percentage of 9.16% and an average loss percentage of -4.32%, with a quarter Kelly diagnostic of 0.0535, an annual return of 5.31%, a mean daily percentage of 0.0221%, and a volatility of 8.86% per annum, with a Sharpe ratio of 0.6295, and currently has 0 open positions. The data sources for this system are https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all and https://www.aqr.com/-/media/AQR/Documents/Insights/Journal-Article/AQR-JPM-Fall-2017.pdf.

A news scan found the headline 'QWIDE91002 monthly 1 train 6bp/side'. This refers to a monthly report for the QWIDE91002 fund, which is a type of automated trading fund. The report mentions that certain positions, referred to as Frozen1720/1721/1724, were part of a strategy that used fifty predetermined cash ETF sleeves. The strategy was tested, or trained, from 2016 to 2019 and then held out, or not used, from 2020 to 2025, with a separate test in 2026. The report also mentions that the fund's return was calculated at the end of each month, and that it was compared to the return of the SHY, which is a type of low-risk investment. The fund's strategy was to mix a positive fraction of its components, without selecting specific ones, and to use a passive full sleeve approach. The next session's target weight was determined within independently compounding sleeves, with a 1bp fee and 5/15 slippage, and the final era close was costed with liquidation. The report also mentions vector bt position episodes, which refers to a type of trading strategy, and aggregate partial resizes, which means adjusting the size of the positions. The fund's expenses were embedded in its prices, and it was a long cash adaptation strategy, not a risk-balanced short futures strategy like the one used by AQR. The report cites two sources: https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all and https://www.aqr.com/-/media/AQR/Documents/Insights/Journal-Article/AQR-JPM-Fall-2017.pdf. The full metrics for the fund are also provided, including its family, signal, cost side, era, return, drawdown, win rate, sample, average win percentage, average loss percentage, quarter Kelly diagnostic, annual return, mean daily percentage, volatility annual percentage, Sharpe ratio, and open positions.

The QWIDE91002 monthly mix for 2026 had a 16 basis points per side performance. This mix used a set of 50 predetermined cash ETF sleeves, which were tested from 2016 to 2019 and held from 2020 to 2025, with 2026 being a separate period. The return at the end of each month was calculated by subtracting the SHY return, and the mix had a positive fraction of return. The components were diagnosed, but no selection was made, and the full sleeve was used passively. For the next session, the monthly target weight was set within independently compounding sleeves, with 1 basis point fees and 5 to 15 slippage, and the final close was costed with liquidation. The vector bt position episodes were aggregated with partial resizes, and correlated positions were not independent and identically distributed, which means they were not random and unrelated to each other. Inactive cash had zero expenses, and prices had embedded expenses. This is a long cash adaptation, not a risk-balanced short futures strategy like AQR, and there were no mandatory ADR exits, so it was not approved. The source of this data is a stock bars feed and an AQR document. The full metrics are: family is monthly, signal is mix, cost per side is 0.0016, era is 2026, return is 8.888251079859177, drawdown is -7.631026162380139, win rate is 58.06451612903226, sample is 62, average win percentage is 8.82235778305431, average loss percentage is 3.082090991642895, quarter Kelly diagnostic is 0.10853588353020575, annual return is 13.247669551892827, mean daily percentage is 0.05232670403799543, annual volatility percentage is 11.918541793621799, Sharpe ratio with zero cash is 1.1063710348049043, and open positions is 0.

The QWIDE91002 monthly mix holdout had a return of 16 basis points per side. The strategy used 50 predetermined cash ETF sleeves, with data from 2016 used for warmup, 2017-2019 for training, 2020-2025 for holdout, and 2026 separate. The month-end return was calculated by subtracting the SHY return, and the mix had a positive fraction with no selection. The strategy was passive with a full sleeve. The next session will open with a monthly target weight within independently compounding sleeves, with 1 basis point fees and 5/15 slippage, and final era close costed liquidation. The vector bt position episodes will aggregate partial resizes, with correlated positions not being independently and identically distributed, which means they are not random and unrelated. Inactive cash was zero, with expenses embedded in prices. This is a long cash adaptation, not a risk-balanced short futures strategy like AQR, and does not require mandatory ADR exits. The source data is from https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all and https://www.aqr.com/-/media/AQR/Documents/Insights/Journal-Article/AQR-JPM-Fall-2017.pdf. The full metrics are: family is monthly, signal is mix, cost per side is 0.0016, era is holdout, return is 36.825524565378046, drawdown is -14.409872683671033, win rate is 38.41961852861036, sample is 367, average win percentage is 7.074519959787137, average loss percentage is 5.35669065474796, quarter Kelly diagnostic is 0, annual return is 5.368322138088377, mean daily percentage is 0.02212872625068005, volatility annual percentage is 8.198855186991967, Sharpe ratio with zero cash is 0.6801484948799642, and open positions is 0.

The QWIDE91002 monthly mix train has a return of 16 basis points per side. The model uses 50 predetermined cash ETF sleeves, with data from 2016 used for warmup, 2017-2019 for training, 2020-2025 for holdout, and 2026 for separate testing. The model's performance is measured by its month-end return minus the return of the SHY ETF. The mix has a positive fraction of 3 components, with diagnostics and no selection, and a passive full sleeve. The next session will open with a monthly target weight within independently compounding sleeves, with 1 basis point fees and 5/15 slippage, and final era close costed liquidation. The model also considers vector bt position episodes with aggregate partial resizes, and correlated positions that are not independently and identically distributed, which means they are not random and unrelated. Inactive cash has a value of zero, and expenses are embedded in prices. This is a long cash adaptation, not a risk-balanced short futures strategy like the one used by AQR. The model does not have mandatory ADR exits, and is therefore unapproved. The source data can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all and https://www.aqr.com/-/media/AQR/Documents/Insights/Journal-Article/AQR-JPM-Fall-2017.pdf. The full metrics are: family is monthly, signal is mix, cost per side is 0.0016, era is train, return is 10.349238469001598, drawdown is -13.355983762439184, win rate is 47.29064039408867, sample size is 203, average win percentage is 4.735582388712451, average loss percentage is 4.729103861076892, quarter Kelly diagnostic is 0, annual return is 3.348791921023997, mean daily percentage is 0.013932363225214033, annual volatility percentage is 6.6175608914479955, Sharpe ratio with zero cash is 0.5305512998439068, and open positions is 0.

The QWIDE91002 monthly mix for 2026 has a 6 basis points per side fee. The system uses 50 predetermined cash ETF sleeves, which were set up in 2016, trained from 2017 to 2019, and held from 2020 to 2025, with 2026 being a separate period. The system calculates the return at the end of each month, subtracting the return of the SHY fund, and adjusts the mix to have a positive fraction of the total, with no selection, using a passive full sleeve approach. The next session will open with a monthly target weight within independently compounding sleeves, with 1 basis point fees and 5 to 15 basis points slippage, and final era close costed liquidation. The system also aggregates partial resizes of vector bt position episodes, and correlated positions are not independent and identically distributed, or IID, which means they are not random and uncorrelated. Inactive cash has zero value, and expenses are included in the prices. This is a long cash adaptation, not a risk-balanced short futures strategy like the one used by AQR, and it does not have mandatory ADR exits. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all and https://www.aqr.com/-/media/AQR/Documents/Insights/Journal-Article/AQR-JPM-Fall-2017.pdf. The full metrics are: family is monthly, signal is mix, cost per side is 0.0006, era is 2026, return is 9.251310250662236, drawdown is -7.617578268557901, win rate is 59.67741935483871, sample is 62, average win percentage is 8.795296268931436, average loss percentage is 3.009210976777107, quarter Kelly diagnostic is 0.11470376300089588, annual return is 13.799758844375987, mean daily percentage is 0.054258591001023565, annual volatility percentage is 11.909198889710034, Sharpe ratio with zero cash is 1.1481179430189914, and open positions is 0.

The QWIDE91002 monthly mix holdout has a 6 basis points per side. This refers to a 0.06% fee per trade. The system uses 50 predetermined cash ETF sleeves, which are a type of investment fund, and was tested from 2016 to 2025. The system's performance is measured by its return minus the return of the SHY fund, which is a type of low-risk investment. The mix has a positive fraction of return, and its components are diagnosed, but no selection is made. The system uses a passive full sleeve approach, meaning it invests in all available funds without trying to pick winners. The next session will open with a monthly target weight within independently compounding sleeves, with 1 basis point fees and 5 to 15 basis points slippage, which refers to the difference between the expected and actual price of a trade. The system also uses vector bt position episodes, which are a type of trading strategy, and aggregate partial resizes, which means it adjusts the size of its trades. The system does not use a risk-balanced approach, unlike AQR, and does not require mandatory exits. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all and https://www.aqr.com/-/media/AQR/Documents/Insights/Journal-Article/AQR-JPM-Fall-2017.pdf. The full metrics of the system are: family is monthly, signal is mix, cost per side is 0.06%, era is holdout, return is 39.61%, drawdown is -13.76%, win rate is 40.87%, sample size is 367, average win percentage is 6.85%, average loss percentage is 5.39%, quarter Kelly diagnostic is 0, annual return is 5.72%, mean daily percentage is 0.023%, volatility annual percentage is 8.2%, Sharpe ratio with zero cash is 0.72, and open positions is 0.

A news scan found the headline 'QWIDE91002 monthly mix train 6bp/side'. The fund manager's notes are about a trading strategy. The strategy uses fifty predetermined cash ETF sleeves, which were set up in 2016 and used for training from 2017 to 2019 and for holding from 2020 to 2025, with 2026 being separate. The strategy looks at the return at the end of each month, subtracting the return of the SHY, which is a type of bond fund, to determine the mix of investments. The strategy does not select specific investments, but instead uses a passive full sleeve approach. The next session will open with a monthly target weight within independently compounding sleeves, with fees of 1 basis point and slippage of 5 to 15 basis points. The strategy also looks at vector bt position episodes, which are periods of time when the strategy is invested in certain assets, and aggregates partial resizes, which are adjustments to the size of the investments. The strategy assumes that correlated positions, which are investments that tend to move together, do not have an independent and identically distributed edge, which is a statistical concept that refers to the idea that each investment has its own unique characteristics. The strategy has zero inactive cash and expenses are embedded in prices. This is a long cash adaptation strategy, not a risk-balanced short futures strategy like the one used by AQR, and it does not have mandatory ADR exits. The strategy uses data from https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all and https://www.aqr.com/-/media/AQR/Documents/Insights/Journal-Article/AQR-JPM-Fall-2017.pdf. The full metrics are: family is monthly, signal is mix, cost side is 0.0006, era is train, return is 11.494014788073436, drawdown is -12.901678278921679, win rate is 48.275862068965516, sample is 203, average win percentage is 4.840300699120874, average loss percentage is 4.627932660894346, quarter Kelly diagnostic is 0, annual return is 3.7061725998846384, mean daily percentage is 0.01530038404114599, volatility annual percentage is 6.6139347938572834, Sharpe zero cash is 0.5829656473102491, and open positions is 0.

A news scan found the headline 'IREN frozen20/10 breakout: independent period does not establish deployable edge'. This refers to a trading strategy for the stock IREN. The strategy was tested on daily stock prices from January 2022 to September 9, 2026. The test was done before the run, which was preregistered on January 23. The strategy used a vectorbt, which is a type of trading tool, to decide when to buy and sell. It bought when the stock's closing price was higher than its prior 20-day high and its 50-day moving average, and sold when the closing price was lower than its prior 10-day low. The test included fees of 5 basis points and slippage of 10 basis points on each trade, with a total stress of 30 basis points. The test was divided into two periods: 2022-2024 for training and 2025 for holdout. The results were 61.7789% for the training period, 102.4264% for the holdout period, and a drawdown of 72.1911%. The strategy had 8 closed trades, with a 25% win rate and a mean return of 23.4239%. A bootstrap test with 10,000 samples had a 95% confidence interval of -17.5992% to 77.0301%. Another test with 96 total cells across 4 symbols, 3 lookbacks, 2 exits, 2 periods, and 2 costs was also done. The primary rule was frozen, with no winner selection. The conclusion was that no capital should be invested because the confidence interval crosses zero, meaning the strategy may not be profitable. The strategy also has limitations, such as not using a 15-minute stop or target, and not simulating partial profits. The current winner universe selection, regime shift, corporate-action basis, and overlapping exposure limit soundness are also concerns. The strongest argument against this conclusion is the positive return, but the unvalidated executable expectancy remains a binding issue.

A news scan found the headline 'ATI frozen20/10 breakout: independent period does not establish deployable edge'. This refers to a trading strategy for the stock ATI. The strategy was tested using daily stock prices from January 2022 to September 9, 2026. The test was done before the run on September 17, 2023. The strategy involved buying when the stock's closing price was higher than its highest price in the last 20 days and its 50-day moving average, and selling when the closing price was lower than its lowest price in the last 10 days. The test included fees of 5 basis points and slippage of 10 basis points for each trade, with a total stress of 30 basis points. The test was split into two periods: 2022-2024 for training and 2025 for testing. The results showed a loss of 19.58% during training and a gain of 114.40% during testing, with a maximum loss of 25.11%. The strategy had 6 closed trades, with a win rate of 50%. The average return was 16.68%, with a 95% confidence interval of -1.69% to 43.08%. The test also included 96 different combinations of trading rules, but did not select a winning rule. The conclusion was that no capital should be invested because the confidence interval crossed zero, meaning the strategy's returns were not statistically significant. The strategy's returns were only calculated when the portfolio included an open trade. The test did not use a 15-minute stop or target, or a partial simulation of a 1R trade. The test also did not consider the soundness of the trading strategy based on factors such as regime shift, corporate actions, and overlapping exposure limits. The strongest argument against this conclusion was the positive return, but the strategy's expected return was not validated.

A news scan found the headline 'SSL frozen20/10 breakout: independent period does not establish deployable edge'. This refers to a trading strategy for the stock SSL. The strategy was tested on daily stock prices from January 2022 to September 9, 2026. The test used a specific set of rules, including a long-only approach, where the stock is bought when its closing price is higher than its prior 20-day high and its 50-day moving average, and sold when its closing price is lower than its prior 10-day low. The test also accounted for fees of 5 basis points and slippage of 10 basis points on each trade, with a total stress of 30 basis points. The results showed a loss of 54.0393% during the training period and a gain of 144.1551% during the holdout period, with a drawdown of 22.9576%. The strategy had 7 closed trades, with a win rate of 42.857% and a mean return of 15.0679%. A bootstrap analysis with 10,000 iterations showed a 95% confidence interval of -4.1565% to 44.1486%. The test was run across 96 different scenarios, including 4 symbols, 3 lookbacks, 2 exits, 2 periods, and 2 costs. The primary rule was kept the same, with no selection of winning trades. The conclusion was that no capital should be allocated to this strategy, as the confidence interval crosses zero and the returns are only calculated when the portfolio includes an open trade. The strategy does not use a 15-minute stop or target, and the current selection of trades, regime shift, corporate action basis, and overlapping exposure limit are sound. The strongest argument against this conclusion is the positive return, but the unvalidated executable expectancy remains a binding constraint.

A news scan found the headline 'FTK frozen20/10 breakout: independent period does not establish deployable edge'. This refers to a trading strategy for the stock FTK. The strategy was tested using daily stock prices from January 2022 to September 9, 2026. The test was done before the run started, on a preregistered date of 1723. The strategy involved buying when the stock's closing price was higher than its prior 20-day high and 50-day moving average, and selling when the closing price was lower than its prior 10-day low. The test included fees of 5 basis points and slippage of 10 basis points for each trade, with a total stress of 30 basis points. The results showed a loss of 46.4835% during the training period from 2022 to 2024, and a loss of 63.6175% during the holdout period in 2025. The strategy had a drawdown of 67.1654%, with 10 closed trades, a win rate of 30.000%, and a mean return of -7.6294%. A bootstrap analysis with 10,000 samples showed a 95% confidence interval of -19.2006% to 5.4067%. The test was done across 96 total cells, including 4 symbols, 3 lookbacks, 2 exits, 2 periods, and 2 costs. The primary rule was frozen, with no selection of winning trades. The conclusion was that no capital should be invested, as the confidence interval crossed zero. The strategy's returns were only considered when the portfolio included a marked open lot. The daily channel exits were not based on a 15-minute stop or target, or a partial simulation of a 1R trade. The soundness of the strategy depended on the selection of the winner universe, regime shift, corporate-action basis, and overlapping exposure limit. The strongest counterargument was the positive return, but the unvalidated executable expectancy remained a binding constraint.

A news scan found the body of a note from a trading fund manager that said the following: the manager ran a mean-reversion screen using the required command line interface with a lookback of 20 days, entry of 2, exit of 0, and 617 days of data. The manager's premises engine uses current closes for signals and fills, with no fees, slippage, or holdout. Seven of the eight gross returns were positive, with the largest being LINK at 69.1614% and a drawdown of 11.9156%, a win rate of 73.6842%, and 19 trades. Other results included LTC at 45.3542% and a drawdown of 12.1012%, a win rate of 80%, and 20 trades, and HYPE at 14.61285% with 2 trades and no history from 2020 to 2025. This is not a frozen trend-pullback rule, so the result is not a replication or parameter selection. Other results included BTC at -3.257646%, a drawdown of 16.639125%, a win rate of 61.1111%, and 18 trades, ETH at 5.419817% with 15 trades, SOL at 15.553973% with 17 trades, XRP at 36.165798% with 19 trades, and UNI at 19.214556% with 22 trades. The average gross loss may be far worse than the win rate implies. The manager rejected deployment from this diagnostic and required a next open independent daily and hourly holdout. All figures are from observation 19211, and the latest 618 bar includes a potential partial data point from September 10, which includes 206 HYPE data points. The source of the data is https://runway.adityaperswal.workers.dev/data/crypto-bars, and the local source is /app/agent/src/backtest/engine.ts and strategies.ts.

A trading fund manager made notes about the AVAV stock. The notes are about an event where the stock price moved in a certain way, called an EAR greater than 0 event. This event was the same as the one in a daily test called test1851. The manager looked at the stock price at 9:30 am and used a stop loss, which is a price level where the manager sells the stock to limit losses, and a target price, which is a price level where the manager sells the stock to make a profit. The stop loss was set at 1 times the average daily range, and the target price was set at 1.5 times the average daily range. The manager also set a rule to sell half of the stock at a certain price level and to move the stop loss to the break-even point after selling half. The manager used real stock price data, including data from outside regular trading hours, but did not have complete data for the overnight period from 8:00 pm to 4:00 am. The manager paid fees of 1 basis point, which is a unit of measurement for interest rates and fees, plus a slippage of 15 basis points on each buy and sell transaction. The manager used a tool called Vectorbt to track the trades and reconciled the cash ledger after each event. The manager found 9 events from 2021 to 2025 and 1 event in 2026 that met the criteria, but did not select any specific parameters. The manager checked the stock price data and found that all the expected 15-minute time slots were present after a correction on July 3, 2023. However, the manager noted that the results were not precise enough and that the risk was higher than expected. The manager did not adopt this strategy for live trading and did not confuse the average trade with the monthly opportunity. The data came from two sources: https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=AVAV&timeframe=15Min&feed=sip&adjustment=all and https://investor.avinc.com/financial-information/quarterly-results.

A trading note, labeled Frozen note1689, used the same entries as a daily test, labeled test1851, when the EAR, or earnings abnormality ratio, was greater than 0. The trading started at 09:30 with the SIP 15-minute open price. The prior daily stop was set at 1 times the average daily range, or ADR, and the target was set at 1.5 times the ADR, with half of the position sold at a profit of 1 ADR. The system checked the bar close and executed the next action at the next observed bar open, and it activated a breakeven stop after half of the position was sold. If both profit levels were surpassed during a check, the system exited the full position. The system used actual exchange data, including extended-hours bars, but did not have exact venue quotes between 20:00 and 04:00. The fees were 1 basis point, or bp, plus 15bp for slippage on each side of the trade. The system reconciled the vectorbt and cash ledger for each event and included flat gaps in the stitched event equity. There were 9 events from 2021 to 2025 and 1 event in 2026, with no parameter selection. All expected regular 15-minute time slots were present after a correction on July 3, 2023. However, the positive holdout was too imprecise, with a confidence interval, or CI, that included zero, and the marked risk exceeded the daily model risk. As a result, there was no live adoption from N9, and the average trade was not mistaken for a monthly opportunity. The data sources were https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=AVAV&timeframe=15Min&feed=sip&adjustment=all and https://investor.avinc.com/financial-information/quarterly-results.

A trading note, labeled as Frozen note1689, used the same entries as a daily test, labeled as test1851, when the EAR, or earnings after revenue, was greater than 0. The actual trading started at 09:30 with a 15-minute sampled exit of 6 basis points per side. The prior daily ADR, or average daily range, was used as a stop at 1 times the ADR and a target at 1.5 times the ADR, with half of the position sold at a profit of 1 ADR. The system checked the bar close and executed trades at the next observed bar open, and activated a breakeven stop after half of the position was sold. If both profit levels were surpassed at a sampled check, a full exit was executed. The system used extended-hours bars from the provider, but did not have exact venue quote sampling and overnight coverage from 20:00 to 04:00. Fees were 1 basis point plus slippage of 5 basis points on each side. The system reconciled the vectorbt and cash ledger per event and included flat gaps in the stitched event equity. There were 9 events from 2021 to 2025 and 1 event in 2026, with no parameter selection. All expected regular 15-minute slots were present after a correction on July 3, 2023, but this does not prove complete extended execution. The positive holdout was too imprecise, with a confidence interval that included zero, and the marked risk exceeded the daily model risk. There was no live adoption from N9, and no average trade arithmetic was mistaken for monthly opportunity. The data was sourced from https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=AVAV&timeframe=15Min&feed=sip&adjustment=all and https://investor.avinc.com/financial-information/quarterly-results.

A trading note, Frozen note1689, used the same entries as a daily test, test1851, when the EAR, or expected average return, was greater than 0. The trading started at 09:30 with the SIP 15-minute open, and had a stop at 1 times the average daily range and a target at 1.5 times the average daily range, with half of the position sold at 1 times the average daily range. The system checked the bar close and executed trades at the next observed bar open, and activated a breakeven stop after half of the position was sold. If both profit levels were surpassed during a check, the system would exit the trade. The system used real-time bars from the provider, including extended-hours bars, but did not have coverage from 20:00 to 04:00. The fees were 1 basis point plus 5 basis points for slippage on each side. The system reconciled the vectorbt and cash ledger for each event and included flat gaps in the stitched event equity. There were 9 events from 2021 to 2025 and 1 event in 2026, with no parameter selection. All expected regular 15-minute slots were present after a correction on July 3, 2023, but this does not guarantee complete extended execution. The positive holdout was too imprecise, with a confidence interval that included zero, and the marked risk exceeded the daily model risk. The system did not adopt this strategy for live trading, and did not mistake average trade arithmetic for monthly opportunity. The data was sourced from https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=AVAV&timeframe=15Min&feed=sip&adjustment=all and https://investor.avinc.com/financial-information/quarterly-results.

A news scan found the notes of a trading fund manager. The manager uses a system called QEAR91000 to decide when to buy or sell the stock AVAV. The system checks if the earnings of AVAV are visible to the company that issued the stock, and if the return on AVAV is more than 0.05% better than the return on the stock SPY over three sessions. If this condition is met, the system buys AVAV at the open of the next session. The system sells half of the stock when the price reaches 1 times the average daily range, and sells the rest when the price reaches the break-even point. The system uses a 5-session horizon and stops trading if the price goes below the 20-session mean high-low range. The system was trained on data from 2016 to 2020 and tested on data from 2021 to 2025. The system uses a long-only strategy, meaning it only buys stocks, and starts with $1000 in cash. The system also reconciles its cash ledger to ensure accuracy. The source of the data is https://investor.avinc.com/financial-information/quarterly-results and https://runway.adityaperswal.workers.dev/data/stock-bars.

A news scan found the notes of a trading fund manager. The manager is holding out on a trade for AVAV, which is a stock ticker, because a measure called EAR, or earnings abnormality ratio, is greater than 0.05, and the risk is 6 basis points, or 0.06%, per side of the trade. The manager looks at the earnings dates for the company that issued the stock, and the return on AVAV compared to the SPY, which is a stock market index, over three and five sessions, or trading days. The manager will enter the trade on the second day after the earnings date, but only if EAR is greater than 0.05. The manager will sell half of the stock at a price that is one average daily range, or the average amount the stock price moves in a day, above the purchase price, and will stop the rest of the trade at a price that is 1.5 average daily ranges above the purchase price, or at a price that is the prior 20-session mean high-low, whichever comes first. The manager will also stop the trade if the stock price moves against the trade by the amount of the prior 20-session mean high-low. The manager tested this strategy from 2016 to 2020, and then from 2021 to 2025, and plans to extend the test in 2026. The manager pays fees of 1 basis point, or 0.01%, plus slippage, which is the difference between the expected price and the actual price, of 5 basis points, or 0.05%, on each side of the trade. The manager uses a program called Vectorbt to make long-only trades, which means buying stocks with the expectation of selling them at a higher price, with an initial cash amount of $1000, and allows for fractional shares, which are smaller than a full share. The manager reconciles the cash ledger, which is the record of the cash in the account, and the maximum error is 0. The manager measures drawdown, which is the amount of loss in the account, using the opening and closing prices of the stocks, not the full intraday risk, which is the risk of loss during the trading day. The manager notes that this strategy is not the same as the strategy described in a research paper, which used a cross-sectional size and value-adjusted long-short replication, which means buying and selling stocks based on their size and value. The manager also notes that the signal family, which is the group of trades, has only 9 test events, and the uncertainty of the estimates crosses zero, which means that the estimates are not certain, and the positive estimates are for research purposes only, and not approval. The manager observed full metrics, which are the complete set of data, 18738 times. The sources of the data are https://investor.avinc.com/financial-information/quarterly-results and https://runway.adityaperswal.workers.dev/data/stock-bars.

A news scan found the notes of a trading fund manager. The manager uses a system called QEAR91000 to decide when to buy or sell AVAV stock. The system looks at the company's earnings dates and how the stock performs compared to the SPY stock index. The manager will buy AVAV stock on the second day after the company reports its earnings, but only if the earnings are more than 0.05% better than expected. The manager will sell the stock after 20 days, or if the stock price drops, but not before. The manager pays a fee of 1 basis point, which is a measure of the cost of trading, and also pays for any losses due to the delay between buying or selling a stock and the price actually changing, which is 15 basis points on each side of the trade. The manager uses a computer program called Vectorbt to make trades, starting with $1000 in cash, and can buy fractions of a share. The manager has tested this system using data from 2016 to 2020, and then from 2021 to 2025, and plans to continue testing it in 2026. The manager's system is not meant to exactly copy the performance of the SPY index, and the results are still being researched and are not yet approved. The manager got the data from the AVAV company website and another website called Runway.

These are notes about a trading strategy for AVAV. The strategy is based on earnings dates and returns compared to the SPY index. It enters a trade on the open of the second day after earnings if the EAR is greater than 0.05. The trade exits after 20 sessions, with no stops. There is a daily check to stop losses first, and target improvements are capped. The strategy was trained on data from 2016 to 2020 and tested on data from 2021 to 2025, with a separate test in 2026. The strategy uses a long-only approach, starting with $1000 in cash, and allows for fractional shares. Fees are 1 basis point plus 5 basis points for slippage on each side of the trade. A basis point is a unit of measurement, equal to 0.01%. The strategy's performance is tracked using open and close prices, not full intraday risk. The data used comes from the AVAV investor website and a stock bars dataset. The strategy has only 9 test events, and positive estimates are for research purposes only, not approval. The total number of observations is 18736.

A news scan found the notes from a trading fund manager. The manager is holding out on a trade for AVAV, which is a stock ticker, because a measure called EAR, or earnings abnormality ratio, a ratio that compares a company's earnings to what was expected, is greater than 0.02. The manager looks at the issuer's visible earnings dates and the return of AVAV compared to the SPY, a stock market index, over three sessions, and then enters the trade on the next session's open if the EAR is greater than 0.02. The manager sets a stop loss, a price at which to automatically sell a stock, at the prior 20-session mean high-low, a measure of price volatility, and a target, a price at which to automatically sell a stock for profit, at 1.5 times the average daily range, a measure of price movement. The manager sells half of the stock at 1 times the average daily range and sets a stop loss for the remaining stock at the breakeven point, the price at which the trade will not make or lose money, from the next session. The manager also considers daily open, high, low, and close prices, and stops trading if there is ambiguity, or uncertainty, about the best price to trade at. The manager trained the trading model on data from 2016 to 2020, tested it on data from 2021 to 2025, and plans to extend the test to 2026. The manager pays fees of 1 basis point, a unit of measurement for interest rates and fees, plus slippage, the difference between the expected price of a trade and the actual price, of 15 basis points on each side of the trade. The manager uses a long-only strategy, meaning the manager only buys stocks and does not sell them short, with an initial cash investment of $1000 and fractional shares, which allow the manager to buy a fraction of a share. The manager reconciled the cash ledger, which is a record of all the manager's trades, and found a maximum error of 9.094947017729282e-13. The manager measures drawdown, a measure of the decline in value of a trade, using the open and close prices, not the full intraday risk, which is the risk of loss during the trading day. The manager notes that this trade is not part of a larger, cross-sectional long-short replication, which is a strategy that involves buying and selling stocks based on their expected performance relative to each other. The signal family, which is a group of related trading signals, has only 9 test events, and the uncertainty, or risk, of the trade crosses zero, meaning that the trade is uncertain and may not be profitable. The manager has observed full metrics, which are measures of the trade's performance, 18735 times. The sources of the data are https://investor.avinc.com/financial-information/quarterly-results and https://runway.adityaperswal.workers.dev/data/stock-bars.

A news scan found the notes of a trading fund manager. The manager is holding out on a stock called AVAV, which is also known as a green5 stock, with an EAR greater than 0.02 and 6 basis points per side. The manager has frozen some notes, numbers 1683, 1684, and 1688, which are about earnings dates that the company that issued the stock can see. The manager looks at how AVAV performs compared to the SPY stock over three sessions and decides to enter a trade at the open of the next session, t+2, only if the EAR is greater than 0.02. The manager sets a five-session horizon, which means they look at what happens over five sessions, and sets a stop loss at the prior 20-session mean high-low, which is a way to measure price volatility, and a target at 1.5 times the average daily range, or ADR. The manager sells half of the stock at 1 ADR and sets a stop loss at breakeven for the remaining stock from the next session. The manager also sets daily open, high, low, and close, or OHLC, stops to avoid ambiguity and adverse opening gaps at the open, and caps target improvement. The manager trained the model from 2016 to 2020, tested it from 2021 to 2025, and extended it to 2026 separately, with entries and exits within each partition. The manager charges fees of 1 basis point plus slippage of 5 basis points for each side of the trade. The manager uses a long-only strategy, which means they only buy stocks, with an initial cash amount of 1000 USD, fractional shares, and no fixed business costs. The cash ledger is reconciled with a maximum error of 4.547473508864641e-13. The drawdown, which is a measure of the decline in value, uses open and close marks, not full intraday risk. This selected single-name SPY proxy, which is a way to track the performance of the SPY stock, is not the same as the paper's cross-sectional size and value-adjusted long-short replication. The signal family has only 9 test events, which means there is uncertainty, and positive estimates are for research purposes only, not approval. The full metrics observation is 18734. The sources of the information are https://investor.avinc.com/financial-information/quarterly-results and https://runway.adityaperswal.workers.dev/data/stock-bars.