Notes

A news scan found the following data: on 2023-03-23, a calendar mode with a train split and 0.001 slippage, which is a small loss due to buying or selling, had a return percentage of -0.8144287601460243. Another calendar mode with the same train split but 0.0025 slippage had a return percentage of -1.1115405489618202. A barrier mode, which is a type of trading strategy, with the same train split and 0.001 slippage had a return percentage of -1.7879981225065178. The barrier mode with 0.0025 slippage had a return percentage of -2.0821935704556815. The fees were 5 basis points, or 0.05%, for slippage of 10 or 25 basis points on one side. The trading was frozen from T-5 to T-2, which refers to a specific time period before an event. The barrier was set at 1 ATR, or average true range, and 1.5 ATR. If the ATR was undefined, there was no barrier. The trading window was scheduled, but it has not been proven to be free of news that could affect the market. https://investors.generalmills.com/press-releases/press-release-details/2023/General-Mills-to-Webcast-Fiscal-2023-Third-Quarter-Earnings-Results-on-March-23-2023/default.aspx ; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data: on 2022-12-20, a trading fund manager recorded the payoff vector for GIS, which is a type of investment. The results were as follows: on 2022-12-20, in calendar mode, with a train split, and a slippage of 0.001, the return percentage was -0.40694295688784576, and with a slippage of 0.0025, the return percentage was -0.7052753752207153. In barrier mode, with the same split and dates, and a slippage of 0.001, the return percentage was -1.8950952351204586, and with a slippage of 0.0025, the return percentage was -2.1889698721461293. The fees were 5 basis points, or 0.05 percent, per slippage of 10 or 25 basis points. The trading was frozen from T-5 to T-2, which refers to a period of time before an event. The barrier was set at 1 ATR, or average true range, and 1.5 ATR. If the ATR was undefined, there was no barrier. The trading occurred during a scheduled window, but it is not proven to be free of news events. https://investors.generalmills.com/press-releases/press-release-details/2022/General-Mills-to-Webcast-Fiscal-2023-Second-Quarter-Earnings-Results-on-Dec.-20-2022/default.aspx ; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data: on 2022-09-21, a trading system tested different modes, including calendar and barrier modes, with different slippage rates, which is the difference between expected and actual trade prices, of 0.001 and 0.0025. The system had return percentages of 1.9209494638818159 and 1.6156438107058246. The system also considered fees of 5 basis points, or 0.05 percent, per slippage of 10 or 25 basis points, and used a barrier of 1 or 1.5 average true ranges, which is a measure of price volatility. The system did not use barrier rows when the average true range was undefined. The trading window was scheduled, but it is not proven to be free of news that could affect trading decisions. https://investors.generalmills.com/press-releases/press-release-details/2022/General-Mills-to-Webcast-Fiscal-2023-First-Quarter-Earnings-Results-on-September-21-2022/default.aspx ; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data: on 2022-06-29, a trading system tested different modes and splits, which are ways to divide data for training. The results were as follows: for the calendar mode with a train split, a slippage of 0.001, which is the difference between expected and actual trade prices, resulted in a return percentage of 4.251309446377194, and a slippage of 0.0025 resulted in a return percentage of 3.9390231667418174. For the barrier mode, which uses a specific price level to trigger trades, with a train split, a slippage of 0.001 resulted in a return percentage of 2.900623174251016, and a slippage of 0.0025 resulted in a return percentage of 2.592382894547085. The fees were 5 basis points, or 0.05 percent, per slippage of 10 or 25 basis points. The system used data from T-5 to T-2, where T is a specific time, and a barrier of 1 or 1.5 ATR, which stands for Average True Range, a measure of price volatility. If the ATR was undefined, no barrier was used. The system operated during a scheduled window and was not proven to be free of news influence. https://investors.generalmills.com/press-releases/press-release-details/2022/General-Mills-to-Webcast-Fourth-Quarter-and-Full-Year-Fiscal-2022-Earnings-Results-on-June-29-2022/default.aspx ; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data line: on 2022-03-23, there were two events with the following details: the first event had a slippage of 0.001 and a return percentage of -0.8287661804221025, and the second event had a slippage of 0.0025 and a return percentage of -1.1258350212916213. Fees were 5 basis points per slippage of 10 or 25 basis points per side. The system was frozen from T-5 to T-2, with a barrier of 1 ATR or 1.5 ATR. There was no barrier when the Average True Range, or ATR, which is a measure of price volatility, was undefined. This occurred during a scheduled window, which may not be free of news events. https://investors.generalmills.com/press-releases/press-release-details/2022/General-Mills-to-Webcast-Fiscal-2022-Third-Quarter-Earnings-Results-on-March-23-2022/default.aspx ; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data line: a list of costed event payoff vectors for 2021-12-21. The list includes four items with the following details: date, mode, split, slippage, and return percentage. The modes are calendar and barrier, which refers to a type of trading strategy. The slippage, or the difference between expected and actual trading costs, ranges from 0.001 to 0.0025. The return percentages range from 1.965478029678036 to 2.4066566378274867. The fees are 5 basis points, or 0.05%, per slippage of 10 or 25 basis points. The trading was frozen from 5 days before to 2 days before the event, and the barrier was set at 1 or 1.5 times the average true range, or ATR, which is a measure of price volatility. If the ATR was undefined, no barrier was set. The trading window was scheduled, but it is not proven to be free of news that could affect trading. https://investors.generalmills.com/press-releases/press-release-details/2021/General-Mills-to-Webcast-Fiscal-2022-Second-Quarter-Earnings-Results-on-December-21-2021/ ; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data: on 2021-09-22, a trading system tested different modes, which are like settings, and splits, which are like divisions of data. The results were: for the calendar mode with a train split, a slippage, which is a small loss due to buying or selling at a bad time, of 0.001, the return percentage was -0.38204986202644564, and with a slippage of 0.0025, the return percentage was -0.6804568479789053. The same results were found for the barrier mode, which uses a barrier, like a limit, to make decisions. The system also considered fees of 5 basis points, which is a measure of cost, and slippage of 10 or 25 basis points. The system used data from the 5 days before and 2 days before, and a barrier of 1 or 1.5 average true ranges, which is a measure of price movement. If the average true range was undefined, no barrier was used. The system was scheduled to run during a certain time window, but it was not proven to be free of news that could affect the results. https://investors.generalmills.com/press-releases/press-release-details/2021/General-Mills-to-Webcast-Fiscal-2022-First-Quarter-Earnings-Results-on-September-22-2021/default.aspx ; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data line: a list of results for a costed event payoff vector on 2021-06-30. The results are as follows: on 2021-06-30, in calendar mode, with a train split, and a slippage of 0.001, the return percentage was 0.43279311586394975, and with a slippage of 0.0025, the return percentage was 0.13194525618624997. On the same date, in barrier mode, with a train split, and a slippage of 0.001, the return percentage was -1.7162867333556653, and with a slippage of 0.0025, the return percentage was -2.010696993790595. Fees were 5 basis points per slippage of 10 or 25 basis points per side. The terms frozen T-5 and T-2 refer to periods when trading was not allowed, specifically five and two days before an event. Barrier 1 ATR and 1.5 ATR refer to a type of stop-loss order, where ATR stands for average true range, a measure of price volatility. No barrier was used when the average true range was undefined. The results were based on a scheduled window, which has not been proven to be free of news events. https://investors.generalmills.com/press-releases/press-release-details/2021/General-Mills-to-Webcast-Fourth-quarter-and-Full-year-Fiscal-2021-Earnings-Results-on-June-30-2021/default.aspx ; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data: on 2021-03-24, a trading system, which is a computer program that automatically makes investment decisions, tested different ways to buy and sell stocks. The system tried two modes: calendar, which refers to a regular schedule, and barrier, which means setting a specific price level to buy or sell. The system also tested two slippage levels, which is the difference between the expected price and the actual price of a trade, set at 0.001 and 0.0025. The return percentage, which is the profit made, was 3.14340238027242% and 2.8344348520051854% for the calendar mode, and 2.604885434978854% and 2.2975310418775985% for the barrier mode. The system also considered fees of 5 basis points, which is a unit of measurement for interest rates, and slippage of 10 or 25 basis points. The system used a scheduled window, which means it only traded at specific times, and did not account for news that may have affected the market. The system also used a barrier of 1 ATR, which stands for Average True Range, a measure of price volatility, and 1.5 ATR. If the ATR was undefined, the system did not use a barrier. The system froze trading from T-5 to T-2, which refers to a specific time period before a scheduled event. https://investors.generalmills.com/press-releases/press-release-details/2021/General-Mills-to-Webcast-Fiscal-2021-Third-Quarter-Earnings-Results-on-March-24-2021/default.aspx ; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data: on 2020-12-17, a trading system tested different modes, which are like settings, and splits, which are like divisions of data. The system tested a calendar mode, which is a schedule-based setting, and a barrier mode, which is a setting that uses a barrier, or a limit, to make decisions. The system also tested different slippage, which is the difference between expected and actual trading costs, values: 0.001 and 0.0025. The return percentage, which is the profit or loss, was -0.9484512033678172 and -1.2451615260494875 for the different tests. The fees were 5 basis points, which is a unit of measurement for interest rates, plus a slippage of 10 or 25 basis points. The system used a frozen time period from T-5 to T-2, which are like milestones in time. The barrier was set at 1 or 1.5 ATR, which is the average true range, a measure of price volatility. If the ATR was undefined, there was no barrier. The system used a scheduled window, but it is not proven to be free of news, which can affect trading decisions. https://investors.generalmills.com/press-releases/press-release-details/2020/General-Mills-to-Webcast-Fiscal-2021-Second-Quarter-Earnings-Results-on-December-17-2020/ ; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline 'TLT QR290706 sameclose/65/base/20bp/recent'. Measured on September 7, 2026, the protocol used was 596/598 with a full fixed grid and no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to September 4. The market period had been previously inspected. The system used an explicit long-only vectorbt, which was reconciled with cash arithmetic. Daily marks were used as the baseline, and 15-minute marks were used for swing trades. Raw barriers and an adjusted-ratio wealth proxy were used. Fees and slip were equally split as a one-way cost, but exact quotes and 15-minute venue polling were unavailable. The results included a sample size of 3, a return percentage of -1.5408550006766997, a drawdown of 4.129387319827893, a win rate of 66.66666666666666, an average win of 0.5689809693076087, an average loss of 2.6516816672246923, a mean of -0.5045732428698251, a Kelly criterion of 0, a median days of 5, a worst trade of -2.6516816672246923, a vectorbt error of -2.842170943040401e-14, the symbol TLT, the mode sameclose, a threshold of 65, no delay, a split of recent, and a cost of 20 basis points. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline 'TLT QR290706 sameclose/65/base/2bp/recent'. This refers to a trading report for the TLT symbol. The report is based on data measured up to September 7, 2026, using a protocol called 596/598 with a full fixed grid and no optimization. The data was split into training from 2016 to 2020, holdout from 2021 to 2025, and recent data from 2026 to September 4. The market period had been previously inspected. The report uses a long-only strategy, which means it only considers buying, not selling short, and it uses a vectorbt system, which is a type of backtesting framework, to reconcile cash and calculate returns arithmetically. Daily marks are used as a baseline, and 15-minute marks are used for swing trades. Raw barriers and an adjusted-ratio wealth proxy were used. Fees and slippage, which is the difference between expected and actual trade prices, are split equally as a one-way cost. Exact quotes and 15-minute venue polling were not available. The report includes the following metrics: a sample size of 3, a return percentage of -0.47173297200373554, a drawdown of 4.025691019612343, a win rate of 66.66666666666666, an average win of 0.9316819049635883, an average loss of 2.30059601632594, a mean of -0.14574406879958773, a Kelly criterion of 0, a median days of 5, a worst trade of -2.30059601632594, a vectorbt error of -7.105427357601002e-14, a symbol of TLT, a mode of sameclose, a threshold of 65, a delay of false, a split of recent, and a cost in basis points of 2. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline 'Measured Sep 7 all-window 1460 days IEX 1002 bars/symbol'. This refers to a measurement of aluminum diagnostic figures on September 7, covering a period of 1460 days, using 1002 bars or units of data per symbol on the IEX exchange. The results for two symbols are given: AA and CENX. For AA, there were 71 trades with a return of -29.873974%, a drawdown of 43.124386%, a win rate of 33.802817%, a mean of -0.419490%, and a buy and hold return of 2.240638%. For CENX, there were 84 trades with a return of 14.870239%, a drawdown of 27.479756%, a win rate of 35.714286%, a mean of 0.274645%, and a buy and hold return of 511.357702%. The measurement used a different exit and feed sample than a previous study that was frozen, and it did not find a trade edge, which refers to a strategy or method that gives a trader an advantage. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline about TLT QR290706 sameclose/65/base/0bp/recent. Measured on September 7, 2026, the protocol used was 596/598 with a full fixed grid and no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to September 4. The market period was previously inspected. The system used an explicit long-only vectorbt, which is a type of backtesting framework, reconciled cash arithmetic. Daily marks were used as the baseline, and 15-minute marks were used for swing trades. Raw barriers and an adjusted-ratio wealth proxy were used. Fees and slip, which refers to the difference between expected and actual trade prices, were equally split as a one-way cost. Exact quotes and 15-minute venue polling were unavailable. The results included a sample size of 3, a return percentage of -0.3522, a drawdown of 4.0257, a win rate of 66.67%, an average win of 0.9721, an average loss of 2.2615, a mean of -0.1058, a Kelly criterion of 0, a median days of 5, a worst trade of -2.2615, and a vbt error of 0. The symbol was TLT, the mode was sameclose, the threshold was 65, the delay was false, the split was recent, and the cost in basis points was 0. https://runway.adityaperswal.workers.dev/data/stock-bars

On September 7, 2026, it was observed that Alcoa's home page listed the actual Q2 release date as July 16, 2026. The scheduling page announced the release on July 16 after the market closed. The source webpage had a label of 4:10 EST, which is not a guaranteed timestamp. The local release was inaccessible due to an HTTP 403 error, even with a User Agent. The indexed official release provided operational figures. A Securities and Exchange Commission document had a 60,000 character cap, which cut off the Management's Discussion and Analysis section, and did not provide a full review. No new catalyst for September was inferred. https://www.alcoa.com/global/en/home/?offset=675 ; https://news.alcoa.com/press-releases/press-release-details/2026/Alcoa-Corporation-Reports-Second-Quarter-2026-Results/default.aspx

A news scan found the headline is not available, but a data line was found. Measured on September 7, 2026, the protocol used was 596 and 598, with a full fixed grid and no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to September 4, which is a market period that was previously inspected. The long-only vector, which is a type of investment strategy that only buys assets, was reconciled with cash using arithmetic, meaning simple mathematical calculations were used. Daily marks, which are the values of the assets at the end of each day, were used as the baseline, and 15-minute marks, which are the values of the assets at 15-minute intervals, were used for swing trades, which are trades that are held for a short period of time. Raw barriers, which are unadjusted levels that an asset's price should not cross, and adjusted-ratio wealth proxy, which is a way to measure wealth that takes into account the ratio of assets, were used. Fees and slip, which is the difference between the expected price of a trade and the actual price, were equally split as a one-way cost, meaning they were only subtracted from the profit, and exact quotes and 15-minute venue polling, which is the process of checking prices at specific intervals, were unavailable. The results were as follows: the sample size was 20, the return percentage was -13.508523831086228, the drawdown, which is the percentage of loss from the peak, was 17.88516188597944, the win rate, which is the percentage of winning trades, was 45, the average win was 0.8552797511111956, the average loss was 1.9768258428699115, the mean, which is the average return, was -0.7023783255784133, the Kelly criterion, which is a way to determine the optimal size of a trade, was 0, the median days, which is the middle value of the number of days that trades were held, was 7.5, the worst trade, which is the trade with the biggest loss, was -5.909877207586622, the vectorbt error, which is an error in the vectorbt system, was -1.8474111129762605e-13, the symbol, which is the code for the asset being traded, was TLT, the mode was sameclose, meaning the trade was closed at the same price it was opened, the threshold, which is the level that the asset's price must cross before a trade is made, was 65, the delay, which is the time between trades, was false, meaning there was no delay, the split, which is the way that the data was divided, was holdout, and the cost in basis points, which is a unit of measurement for the cost of a trade, was 20. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the report for TLT QR290706. The report was measured on September 7, 2026, using Protocol 596 and 598 with a full fixed grid and no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to September 4. The market period had been previously inspected. The report used an explicit long-only vectorbt reconciled cash arithmetic method. Daily marks were used as the baseline, and 15-minute marks were used for swing trades. Raw barriers and an adjusted-ratio wealth proxy were used. Fees and slip were equally split as a one-way cost. Exact quotes and 15-minute venue polling were unavailable. The results included a sample size of 20, a return percentage of -7.051470442958974, a drawdown of 15.471963673072475, a win rate of 55.00000000000001, an average win of 1.0288485172863615, an average loss of 2.0225093730984676, a mean of -0.34426253338681156, a Kelly criterion of 0, a median days of 7.5, a worst trade of -5.570542203454387, a vectorbt error of -9.947598300641403e-14, a symbol of TLT, a mode of sameclose, a threshold of 65, a delay of false, a split of holdout, and a cost in basis points of 2. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the report for TLT as of September 7, 2026. The report uses Protocol 596 and 598 with a full fixed grid and no optimization. The training period is from 2016 to 2020, the holdout period is from 2021 to 2025, and the recent period is from 2026 to September 4. The market period was previously inspected. The report uses an explicit long-only vectorbt, which is a type of backtesting framework, reconciled cash arithmetic. Daily marks are used as the baseline, and 15-minute marks are used for swing trades. Raw barriers and adjusted-ratio wealth proxy are used. Fees and slip, which refers to the difference between expected and actual trade prices, are equally split as a one-way cost. The report notes that exact quotes and 15-minute venue polling are unavailable. The results show a sample size of 20, a return percentage of -6.3049, a drawdown of 15.201, a win rate of 55, an average win of 1.069, an average loss of 1.983, a mean of -0.304, a Kelly criterion of 0, a median days of 7.5, a worst trade of -5.532, and a vbt error of -8.526e-14. The symbol is TLT, the mode is sameclose, the threshold is 65, the delay is false, the split is holdout, and the cost basis point is 0. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline is not available, but a data line was found for TLT, measured on September 7, 2026. The protocol used was 596/598 with a full fixed grid and no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to September 4. The market period had been previously inspected. A long-only vector, which is a way to measure investment returns, was reconciled with cash arithmetic, meaning it was calculated using simple addition and subtraction. Daily marks, or valuations, were used as the baseline, and 15-minute marks were used for swing trades, which are short-term trades. Raw barriers, or price levels, were used, and an adjusted-ratio wealth proxy, a way to measure wealth, was applied. Fees and slip, or the cost of buying and selling, were split equally as a one-way cost. Exact quotes and 15-minute venue polling, which is a way to get price data, were not available. The results included a sample size of 31, a return percentage of -0.6943051464882943, a drawdown of 7.977167325169554, a win rate of 51.61290322580645, an average win of 1.069736577103575, an average loss of 1.170246240255317, a mean of -0.014126076457179266, a Kelly criterion, which is a way to measure investment performance, of 0, a median days of 7, a worst trade of -2.4965016574796217, a vectorbt error, which is an error in the calculation, of -1.5631940186722204e-13, a symbol of TLT, a mode of sameclose, a threshold of 65, a delay of false, a split of train, and a cost in basis points of 20. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline is not available, but a data line was found for TLT. Measured on September 7, 2026, the protocol used was 596/598 with a full fixed grid and no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to September 4, which was a market period previously inspected. The explicit long-only vectorbt was reconciled with cash arithmetic, which is a method of calculating returns. Daily marks were used as the baseline, and 15-minute marks were used for swing trades. Raw barriers and an adjusted-ratio wealth proxy were used. Fees and slip were equally split as a one-way cost. Exact quotes and 15-minute venue polling were unavailable. The results included a sample size of 31, a return percentage of 11.030280419669115, a drawdown of 5.82661561966118, a win rate of 61.29032258064516, an average win of 1.224713903008728, an average loss of 1.044077963975539, a mean of 0.34647188998256023, a Kelly criterion of 0.07072506671382399, a median days of 7, a worst trade of -2.144856351564106, a vbt error of -2.1316282072803006e-13, the symbol TLT, the mode sameclose, a threshold of 65, no delay, a split of train, and a cost of 2 basis points. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data line: Measured Sep7 2026. Protocol596/598 was used with a full fixed grid and no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to Sep4. This was a market period that was previously inspected. The long-only vectorbt was reconciled with cash using arithmetic, which is a method of calculating returns. Daily marks were used as the baseline, and 15-minute marks were used for swing trades. Raw barriers and an adjusted-ratio wealth proxy were used. Fees and slip were equally split as a one-way cost. Exact quotes and 15-minute venue polling were not available. The results were: sample size was 31, return percentage was 12.415627301290533, drawdown was 5.751246768384865, win rate was 61.29032258064516, average win was 1.2652118877618908, average loss was 1.004487677499747, mean was 0.3866185076606116, Kelly criterion was 0.07639402368099076, median days was 7, worst trade was -2.1057064645188417, vectorbt error was 8.526512829121202e-14, symbol was TLT, mode was sameclose, threshold was 65, delay was false, split was train, and cost in basis points was 0. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data line: Measured Sep7 2026. This is a report about a trading protocol called Protocol596/598, which uses a full fixed grid with no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to Sep4. The market period was previously inspected. The report uses a long-only vectorbt, which is a type of trading strategy, and reconciles cash using arithmetic. Daily marks are used as a baseline, and 15-minute marks are used for swing trades. The report also mentions raw barriers and an adjusted-ratio wealth proxy, which is a way to measure wealth. Fees and slip, which refers to the difference between the expected and actual price of a trade, are equally split as a one-way cost. Exact quotes and 15-minute venue polling are not available. The report includes the following statistics: sample size is 5, return percentage is 0.3195, drawdown is 1.0393, win rate is 40, average win is 0.9073, average loss is 0.4941, mean is 0.0664, Kelly criterion, which is a way to determine the optimal bet size, is 0.0183, median days is 2.2708, worst trade is -0.7376, vbt error, which is an error in the vectorbt strategy, is -4.2633e-14, symbol is TLT, mode is swing, threshold is 75, delay is false, split is recent, and cost in basis points is 10. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline is not available, but a data line was found for TLT as of September 7, 2026. The protocol used was 596/598 with a full fixed grid and no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to September 4. The market period was previously inspected. The long-only vectorbt was reconciled with cash arithmetic. Daily marks were used as the baseline, and 15-minute marks were used for swing trades. Raw barriers and adjusted-ratio wealth proxy were used. Fees and slip were equally split as a one-way cost. Exact quotes and 15-minute venue polling were unavailable. The results showed a sample size of 29, a return percentage of -8.557262504777952, a drawdown of 9.940304003311473, a win rate of 24.137931034482758, an average win of 1.0934746834692024, an average loss of 0.7483151291671956, a mean of -0.30374517439289256, a Kelly criterion of 0, a median days of 2.1666666666666665, a worst trade of -1.6895048805490087, and a vectorbt error of 4.263256414560601e-14. The symbol was TLT, the mode was swing, the threshold was 75, the delay was false, the split was holdout, and the cost was 10 basis points. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline is not available, but a data line was found for TLT. Measured on September 7, 2026, the protocol used was 596/598 with a full fixed grid and no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to September 4, which is a market period that was previously inspected. The system used an explicit long-only vectorbt, which is a type of backtesting framework, and reconciled cash arithmetic, meaning it accounted for cash transactions. Daily marks were used as the baseline, and 15-minute marks were used for swing trades. Raw barriers and an adjusted-ratio wealth proxy were used. Fees and slip, which refers to the difference between expected and actual trade prices, were equally split as a one-way cost. Exact quotes and 15-minute venue polling were unavailable. The results included a sample size of 35, a return percentage of 4.9606, a drawdown of 3.9357, a win rate of 45.7143, an average win of 1.1280, an average loss of 0.6843, a mean of 0.1442, a Kelly criterion of 0.0319, a median days of 2.1979, a worst trade of -1.4267, a vbt error of -1.2789e-13, a symbol of TLT, a mode of swing, a threshold of 75, a delay of false, a split of train, and a cost in basis points of 10. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline about a trading report for TLT. Measured on September 7, 2026, the report used Protocol 596 and 598 with a full fixed grid and no optimization. The training period was from 2016 to 2020, the holdout period was from 2021 to 2025, and the recent period was from 2026 to September 4. The report used an explicit long-only vectorbt, which is a backtesting framework, reconciled cash arithmetic. Daily marks were used as the baseline, and 15-minute marks were used for swing trades. Raw barriers and an adjusted-ratio wealth proxy were used. Fees and slip, which refers to the difference between expected and actual trade prices, were equally split as a one-way cost. Exact quotes and 15-minute venue polling were unavailable. The report included the following statistics: a sample size of 5, a return percentage of 0.5407562606016381, a drawdown of 1.0447521625477019, a win rate of 40, an average win of 0.9072714348841604, an average loss of 0.4211136502594785, a mean of 0.11024038379797707, a Kelly criterion, which is a measure of risk, of 0.03037690253414968, a median days of 1.0104166666666667, a worst trade of -0.7431062996850257, a vectorbt error of -1.4210854715202004e-14, a symbol of TLT, a mode of swing, a threshold of 65, a delay of true, a split of recent, and a cost in basis points of 10. https://runway.adityaperswal.workers.dev/data/stock-bars