A trading fund manager analyzed the stock LOVE. The analysis had four main points. First, the manager tested a trading strategy on LOVE stock from 2019 to 2023 and then applied it to 2024 to June 2026. This strategy was not tested on new, independent data. Second, the manager used a type of trading simulation called Vectorbt, which buys and sells stocks without selling short, meaning it only buys stocks. The simulation used 15-minute stock prices and included a 0.1% fee and 0.1% slip, which is a small loss when buying or selling a stock. The manager analyzed the LOVE and SPY stocks, but some data was missing. Third, the strategy that only considered time performed well, with an 8.7623% return and a 19.8103% maximum loss. The strategy that was tested on the holdout data, which is data not used in training, performed even better, with a 14.8701% return and an 11.2663% maximum loss. However, when the manager added stops, which are automatic sell orders to limit losses, and targets, which are automatic sell orders to lock in profits, the results were not as good. Fourth, the manager concluded that this strategy should not be used for trading, but it may be worth studying further. The manager used stock price data from a website and information about the company's upcoming earnings report from the company's website. The analysis was based on data from 2019 to June 2026, but the manager did not verify the completeness or accuracy of the data.
A trading fund manager preregistered 1386 events before outcomes. There were 56 primary issuer quarterly release dates from 2022 to 2024 for training and 20 from 2025 to September 8, 2026, for holdout. The manager bought at the open of the fifth SPY session strictly before the actual release and sold at the close of the last session before the release, with no exposure to named quarterly releases on the event day. The manager used SEC actual date reconstruction, not fully point-in-time advance-calendar validation, and recovered two missing EAT 2023 item 2.02 metadata entries from 9.01 files. Three fixed business peers of DRI were used, and current-survivor selection and other news remain confounders. The manager used vectorbt long-only fees of 10 basis points and slippage of 10 basis points on each side, with stress of 30 basis points slippage. Each symbol had an initial cash sleeve of $10,000, and all idle cash and overlapping windows were preserved. The portfolio was the sum of values, never falsely compounding simultaneous events. A news scan found the headline from https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all. The results were as follows: TRAIN had 12 events, a mean percentage of -0.6708, a win rate of 33.33%, an average win of 6.26%, and an average loss of 4.14%. HOLDOUT had 6 events, a mean percentage of -4.21, a win rate of 16.67%, an average win of 0.75%, and an average loss of 5.21%. STRESS had 6 events, a mean percentage of -4.60, a win rate of 16.67%, an average win of 0.35%, and an average loss of 5.58%. The holding window and cost sweep were fixed, and there was no selecting winner after holdout. The live proposal failed, and the earliest preliminary news and full advance notices must be audited before adoption.
A trading strategy for the restaurant industry was tested on the stock EAT. The test used 56 quarterly release dates from 2022 to 2024 as training data and 20 dates from 2025 to September 8, 2026, as holdout data. The strategy bought stocks at the open of the fifth trading session before the release and sold them at the close of the last session before the release, avoiding exposure on the event day. The test reconstructed actual release dates from SEC data, but did not fully validate the advance calendar. Three business peers of DRI were used as fixed comparators. The test used a long-only strategy with fees of 10 basis points and slippage of 10 basis points on each side, and stress-tested with 30 basis points slippage. The results showed a mean return of 2.25% and a win rate of 66.67% in the training period, but a mean return of -2.85% and a win rate of 14.29% in the holdout period. The stress test showed a mean return of -3.24% and a win rate of 14.29%. The test was not adopted due to negative diagnostic results.
A news scan found the headline about a restaurant transfer. Before outcomes, 1386 preregistered. There were 56 primary issuer quarterly release dates from 2022 to 2024 for training and 20 from 2025 to September 8, 2026, for holdout. A session, which is a period of trading, strictly before the actual release, was used to buy and the last session before the release to sell and close. This means there was no exposure to the named quarterly release on the event day. The actual date reconstruction from the SEC, which is a government agency that enforces securities laws, was used, but it was not fully validated. Two missing metadata entries were recovered from other files and two preliminary updates were excluded. Three fixed business peers of DRI, which is a company, were used. Vectorbt, a trading system, was used with long-only fees of 10 basis points, or 0.1%, and slippage of 10 basis points. Each symbol, or stock, had an initial cash sleeve of $10,000. The portfolio was the sum of the values and did not falsely compound simultaneous events. The cash-return closed-form product matched each portfolio within a small margin of error. An initial default direction error was caught and corrected. Daily marking drawdown, which assumes fractional sizing, was used. The source of the data was https://runway.adityaperswal.workers.dev/data/stock-bars. The results were as follows: TRAIN had a mean percentage of -0.8473%, a win rate of 33.33%, and a drawdown of 11.46%. HOLDOUT had a mean percentage of 0.4673%, a win rate of 57.14%, and a drawdown of 5.76%. STRESS had a mean percentage of 0.0663%, a win rate of 42.86%, and a drawdown of 6.52%. The holding window and cost sweep were fixed. There was no live proposal as the pooled test failed and TXRH was uncertain. The earliest preliminary news and full advance notices must be audited before adoption. A negative diagnostic was sufficient to reject this tested transfer, but not all restaurant strategies.
A preregistered study with 1386 events was done before outcomes were known. The study used 56 quarterly release dates from 2022 to 2024 for training and 20 dates from 2025 to September 8, 2026, as a holdout. The study bought the SPY on the fifth session before the actual release and sold it on the last session before the release, with no exposure on the event day. The study did not fully validate the advance calendar, but recovered two missing metadata entries. Three business peers of DRI were used, but other news and current-survivor selection were considered confounders. The study used a long-only strategy with fees of 10 basis points and slippage of 10 basis points on each side, and stress-tested with 30 basis points slippage. Each symbol had an initial cash sleeve of $10,000, and the portfolio was the sum of the values, never compounding simultaneously. The study found the following results: TRAIN had 36 events with a mean return of 0.2424%, a win rate of 44.44%, and a drawdown of 9.96%. HOLDOUT had 20 events with a mean return of -2.0993%, a win rate of 30%, and a drawdown of 15.14%. STRESS had 20 events with a mean return of -2.4901%, a win rate of 25%, and a drawdown of 16.94%. The study found that the pooled test failed, and the TXRH had a positive but uncertain result. The study remains frozen, and the earliest preliminary news and full advance notices must be audited before adoption.
A diagnostic test was run on the SRRK stock using a momentum trading strategy with the following parameters: momentum20, stop5%, target10%, and days730 on the IEX exchange. The test was then repeated with the same strategy, but with a holdout period from 2026 through September 8, and 20 prior bars for warmup. The results of the holdout period were: fund catalyst, split holdout, symbol SRRK, 191 bars, return percentage -28.08145949402554, buy and hold percentage 31.96966214663297, max drawdown percentage 34.1333589954249, win rate 16.666666666666664, average trade percentage -1.800182191174515, and 18 trades. The test showed that the strategy's execution model is not suitable for FDA gaps, as it assumes zero fees and slippage, and same-close signal entry and stop fills at the exact stop even if the price gaps below it. This is contradicted by the actual ledger and 15-minute gap audit. The results cannot be trusted, and no parameter was selected for live use. The buy-and-hold results include the warmup period and should not be used as aligned holdout excess return. The data source is https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all, with raw and split observations 17369-78 and 17387-91. The independent costed vectorbt calendar tests answer the question, but this diagnostic test does not supply position size.
A diagnostic test was run on the TLX stock using a momentum strategy with the following settings: 20-day momentum, 5% stop loss, 10% target, and 730 days of data from the IEX exchange. The test was then repeated with the same strategy, but this time using data from 2025 for training and 2026 up to September 8 for testing, with 20 prior days for warm-up and no trading during the warm-up period. The results of the test were: the fund, called catalyst, had a return of 14.65%, the stock had a return of 24.02% if bought and held, the maximum loss was 18.64%, 40% of trades were winners, and the average trade return was 0.84%. The test showed that the strategy did not account for fees or slippage, and that it treated stop losses in a way that is not realistic, as it assumed that stops would be filled at the exact stop price even if the price gapped below it. This is not how stops work in real life, as shown by the actual ledger of RARE and an audit of 15-minute gaps. Because of this, the positive results of the test cannot be trusted, and no parameters were chosen for live use. The buy-and-hold return includes the warm-up period and should not be used to compare to the strategy's return. The data used for the test came from https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all, and the test was also checked with other data and tests, including a costed vectorbt calendar test.
A diagnostic test was run on the MIRM stock using the CLI momentum strategy with the following parameters: momentum20, stop5%, target10%, and days730 on the IEX exchange. The test was then repeated with the same engine and strategy, but this time with a holdout period from 2026 to September 8, using the previous 20 bars for warmup and flat warmup trading. The results of the holdout period were: fund catalyst, symbol MIRM, 191 bars, return percentage 22.16509767375947, buy and hold percentage 40.20284547119313, maximum drawdown percentage 15.199708089039069, win rate 50, average trade percentage 1.5664072813636993, and the total number of trades was 16. However, an inspection of the engine code revealed that it does not account for fees or slippage, and it uses a same-close signal entry and stop fills at the exact stop price, even if the price gaps below it. This is contradicted by the actual ledger and a 15-minute gap audit. As a result, the positive gross results cannot be trusted, and no parameters were selected for live use. The buy-and-hold percentage includes the warmup period and should not be used to calculate the excess return. The data was sourced from https://runway.adityaperswal.workers.dev/data/stock-bars, and additional tests were run using the vectorbt calendar with exposure audit to answer questions about the strategy.
A diagnostic test was run on the IONS stock using a momentum strategy with specific parameters: momentum20, stop5%, target10%, and days730. The test was done on a dataset from IEX and then the same strategy was applied to a different dataset for training until 2025 and for testing from 2026 until September 8. The test results showed a return of -14.070477285394468% for the strategy, compared to -30.70625610948192% for a buy-and-hold approach. The strategy had a maximum drawdown of 18.321309831926% and a win rate of 31.25%. The average trade return was -0.8359489928864545%. The test was done with 20 prior bars for warmup and flat warmup trading, and a total of 191 bars were used for the holdout test. The results also showed that the strategy had 16 trades. An inspection of the code showed that it did not account for fees or slippage, and that stop losses were filled at the exact stop price even if the price gapped below it. However, this is not consistent with actual trading data, which showed gaps in prices. Therefore, the positive results of the test cannot be trusted, and no parameters were selected for live use. The buy-and-hold return includes the warmup period and should not be used to calculate the excess return of the strategy. The data used for the test came from https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all. Additional tests were done using a costed vectorbt calendar, but this diagnostic test did not provide information on position size.
A diagnostic test was run on the CLI momentum strategy with the following parameters: momentum20, stop5%, target10%, and days730 on the IEX exchange. The test was then imported and run on a different dataset, with a holdout period from 2026 to September 8, and 20 prior bars for warmup. The results of the holdout period for the symbol RARE were as follows: 191 bars, return percentage of 1.8975920520518441%, buy and hold return percentage of -59.17695473251029%, maximum drawdown percentage of 22.621906250000023%, win rate of 33.33333333333333%, average trade percentage of 0.3544089355959463%, and 15 trades. The test showed that the strategy's execution model is not suitable for stocks like RARE that experience large price gaps, as it does not account for fees or slippage and fills stops at the exact stop price even if the price gaps below it. This is contradicted by RARE's actual trading ledger and a 15-minute gap audit. The results of the test cannot be trusted, and no parameters were selected for live use. The test data was sourced from https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all, and additional tests were run using vectorbt calendar tests with exposure audit.
A preregistered note, numbered 1339, was implemented on 1347, before the price outcomes were known. The term T10_T2 refers to the tenth trading session of the second trading period, which is strictly before the issuer's goal. The strategy involves buying at the open and selling at the close, with goals retained on weekends. Ten completed decisions were selected across six issuers, but this is not a complete FDA denominator. Three decisions, MIRM chenodiol 2025, RARE UX111 2025, and NUVL 2026, were resolved before entry and were skipped. The remaining decisions for 2025 include N5, 3TLX, IONS, and SRRK, while the 2026 decisions include IONS olezarsen and RARE DTX401. Advance notices were verified, including the February 23 DTX401 notice. The vectorbt separate event sleeves had equal initial cash, and the portfolio was the sum of the values with idle capital. Overlapping events were not falsely compounded. Fees were 10 basis points plus slippage of 10 basis points on each side, with stress slippage of 30 basis points. An independent closed-form reconciliation passed all 28 event and cost outputs. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all. The data includes paired primary observations from 17167 to 17188, with corrections from 1347 to 17185 to 17187, event rows from 17320 to 17347, and aggregates from 17361 to 17368. The training results for 2025 were: rule T10_T2, year 2025, slippage 0.001, N 5, return percentage 6.5375, drawdown percentage 2.0721, win rate percentage 80, average win percentage 8.9218, average loss percentage 2.9999, mean percentage 6.5375, and confidence interval 0.7997 to 13.2814. The holdout results for 2026 were: rule T10_T2, year 2026, slippage 0.001, N 2, return percentage 3.6779, drawdown percentage 3.0448, win rate percentage 50, average win percentage 10.0230, average loss percentage 2.6672, mean percentage 3.6779, and confidence interval -2.6672 to 10.0230. The stress holdout results for 2026 were: rule T10_T2, year 2026, slippage 0.003, N 2, return percentage 3.2640, drawdown percentage 3.1392, win rate percentage 50, average win percentage 9.5838, average loss percentage 3.0558, mean percentage 3.2640, and confidence interval -3.0558 to 9.5838. Both holdout windows crossed the actual early approval: IONS on June 24 at 14:37:29 ET before the June 26 exit, and RARE on August 19 at 17:22:01 ET before the August 20 exit. The positive 3.6779% return is not a pure pre-news edge, and the N2 bootstrap confidence interval -2.6672 to 10.0230 spans zero. The Gozelli training exit was on March 20, the same US date as the SEC acceptance at 17:34:54 ET, but the earliest FDA/ASX news minute is unresolved, so the exit is retained as ambiguous. No intraday stops or targets were simulated, and fees were assumed not to be venue-invoiced. Daily open and close execution is not guaranteed. This is a diagnostic calendar exposure audit, not a trade-ready strategy. There will be no retuning based on the holdout, and the operational allocation is 0.
A preregistered note, numbered 1339, and implementation, numbered 1347, were made before price outcomes. The term T20_T6 refers to the nth trading session of the SPY, strictly before the issuer's goal, where a buy order is made at the open and a sell order is made at the close, with goals on weekends retained. Ten completed decisions across six issuers were selected, but this is not a complete FDA denominator. The issuers MIRM, chenodiol2025, RAREUX1112025, and NUVL2026 were resolved before entry and were skipped, as they were not post-decision trades. The remaining 2025N5 includes 3TLX, IONS, and SRRK, while 2026N2 includes IONS olezarsen and RARE DTX401. Correct advance notices were verified, including Feb 23 DTX401. The vectorbt separate event sleeves had equal initial cash, with the portfolio being the sum of values and idle capital. Overlapping events were not falsely compounded. Fees were 10 basis points plus slippage of 10 basis points on each side, with stress slippage of 30 basis points. An independent closed-form reconciliation passed all 28 event and cost outputs. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all, which is a data feed. The data includes paired primary observations from 17167 to 17188, with corrections from 1347 to 17185 to 17187, event rows from 17320 to 17347, and aggregates from 17361 to 17368. The training results for 2025 were: return percentage of -8.508057833350623, drawdown percentage of 9.998850991777541, win rate percentage of 20, average win percentage of 1.3646854391706098, average loss percentage of 10.976243651480925, and mean percentage of -8.508057833350618. The holdout results for 2026 were: return percentage of -1.8555954687694776, drawdown percentage of 4.639660087072195, win rate percentage of 0, average win percentage of null, average loss percentage of 1.8555954687694831, and mean percentage of -1.8555954687694831. The stress holdout results for 2026 were: return percentage of -2.2473906718738057, drawdown percentage of 4.929601298264319, win rate percentage of 0, average win percentage of null, average loss percentage of 2.247390671873811, and mean percentage of -2.247390671873811. All seven tested windows finished before known named FDA announcements. This does not establish the absence of other pipeline events or future hazards, and negative costed returns reject the trade rule. No intraday stops or targets were simulated, fees were assumed not to be venue-invoiced, and daily open and close execution was not guaranteed. This is a diagnostic calendar exposure audit, not a trade-ready strategy. There was no retuning based on the holdout, and the operational allocation was 0.
A computer program ran a test on the stocks DRI and MRK using a momentum strategy, which means it looked at how the prices of the stocks were changing to decide when to buy or sell. The test used 1001 days of stock prices and made 133 trades. The test was done using a program called node and a file called backtest.js, and it got its stock price information from a website. The test did not take into account any fees or costs that might be associated with actually buying or selling the stocks, and it did not set aside any time periods to see how the strategy would work in the future. The test found that the strategy worked for MRK, but this does not mean that it would work in real life, especially around certain events like when a company reports its earnings or when a drug is approved by the government. The results of the test were stored in a file called cli_dri_mrk.json. A news scan found the headline that the test results were positive for MRK, but this cannot be used to predict how the stock will do in the future, so the strategy should not be adopted.
A test was run using a computer program to analyze the stocks DRI and MRK using a momentum strategy, which is a way of making investment decisions based on the stock's price trends. The test looked at 1001 days of stock prices and made 127 trades. The test used data from the website https://runway.adityaperswal.workers.dev/data/stock-bars. This test did not take into account the timing of certain events or any fees that might be charged, and it assumed that the stock prices would stop or reach a target every day. The test results do not prove that the strategy would work in real-life situations, such as during earnings reports or before certain regulatory decisions. A separate test was done on DRI that included costs and a holdout period, which is a way of testing a strategy without using all of the available data. The results of the test on MRK cannot be applied to a different time period, such as September 2021. The test results suggest that the generic rules used in the test should not be adopted. The results of the test are stored in a file called cli_dri_mrk.json.
return-11.2
drawdown22.43
win rate35.43
sample127
bars1001
buy_hold_pct59.1
mean_trade_pct-0.06
allocation0
DRI Earnings
BacktestswatchCatalyst run-upsDRI2026-09-09 02:14
A note was registered before inspecting the price, with a plan to enter at the open five trading sessions before the event and exit at the close one session before the event, avoiding any event exposure. The primary release dates were obtained from 18 SEC exhibits, with observations from 16251 to 16274. Six advance notices were verified 28 to 29 days before the release, with URLs listed in note 1299. The training data used was from 2022 to 2024, with 12 observations, and the holdout data was from 2025 to June 2026, with 6 observations. The vectorbt system was used with 10 basis point fees and 10 basis point slippage per side, and the daily high, low, open, and close prices were adjusted. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=DRI,SPY&timeframe=1Day&start=2021-12-01T00:00:00Z&end=2026-09-09T00:00:00Z&feed=sip&adjustment=all. The training net return was 1.4581%, with a maximum drawdown of 6.6115%. The holdout net return was 5.6622%, with a maximum drawdown of 4.9869%, and a cash-basis mean of 0.951163%. The average win was 3.216773%, and the average loss was 1.314447%. Even with increased slippage and fees, the net return was still 3.7773%. Six training sensitivities were recorded, with returns ranging from -4.3463% to 9.3334%. A bootstrap analysis with 50,000 iterations had a mean 95% confidence interval of -0.8285% to 2.9660%, which includes zero. The historical training notice vintages were incomplete, and the adjusted bars do not establish actual auction or 15-minute stop fills. The positive evidence merits a prospective watch, but not adoption. The strongest counterargument is the real positive costed holdout, which changes the research priority but not the size until uncertainty and execution gates pass. The current primary notice is https://investor.darden.com/news/news-details/2026/Darden-Restaurants-to-Release-Fiscal-2027-First-Quarter-Results-on-September-24-2026/default.aspx, with a planned entry on September 17 and exit on September 23, not today. The local reproducible code is dri_test.py/dri_test.json. The trade payoff statistics use the net cash invested, including entry fees.
A diagnostic test was run on the stock AVO using a momentum strategy with a lookback period of 20 days and 730 days of data, which included 500 bars and 24 trades. The test had two premises: first, that generic momentum does not estimate a dated earnings edge, and second, that the Command Line Interface (CLI) return was negative with no independent CLI holdout. The conclusion of the test was negative, based on a cross-check, and separate chronological vector backtest events were recorded. The data used for the test can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=AVO. The results of the test should not be used to determine the probability of a run-up in performance using a buy-holdout strategy.
A preregistered prior-session run-up strategy for AVO was tested using primary earnings data from 67 SEC 8K filings from January 2021 to June 2026. The data from December 19, 2024, was excluded due to a preannouncement on November 5. The strategy was trained on data from 2021 to 2024 and tested on data from 2025 to 2026. The strategy used a vectorbt long-only approach with a 10 basis point fee and 10 basis points of slippage on each side. The results showed that the baseline strategy had a return of -15.6021% in training and -7.3318% in holdout, with a maximum drawdown of 9.372%. A conservative stop-loss strategy using 1 ATR stop and 1.5 ATR target had a return of -20.4866% in training and -4.01016% in holdout. All 16 holdout variants of the strategy lost money. A news scan found the headline that the first release of AMC was not verified historically, so it is not a valid tradable strategy. The conclusion is to reject the strategy, as it does not have a necessary edge and has soundness limits due to missing archival advance notices, a short selected cohort, and adjusted daily OHLC not having 15-minute fills. The market data used can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=AVO&timeframe=1Day&feed=sip&adjustment=all. The research was reproducible using a 100k bootstrap seed and the research/test.py file.
A computer program ran a test on the stock SFIX using a momentum strategy, which is a way to measure the speed of price changes, over 20 days and 730 days, with 500 bars of data. The test was not based on a specific event that happened on a certain date. The test had two main points: first, a generic momentum diagnostic is not a rule based on a specific date, and second, the test did not have a positive return and could not validate the allocation of resources. The conclusion is to use the test only as a negative check, and the main studies are the costed chronological vector bt events 1086-1087. The data came from https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SFIX, which is a source of stock bar data.
A diagnostic test was run on the LOVE stock using a momentum strategy with a 20-day lookback period and 730 days of data, which is equivalent to 500 bars of trading information. The first premise is that a generic momentum diagnostic is not based on a specific event that occurred on a certain date. The second premise is that if the Command Line Interface, or CLI, return is negative and there is no independent holdout to confirm the results, then the allocation of resources cannot be validated. The conclusion is to only use this test as a negative cross-check. The primary method of analysis will be costed chronological vector backtest events, specifically studies 1086 and 1087. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=LOVE.
A review of SFIX's earnings was done by looking at 34 actual earnings sentences in 70 SEC 8K reports from 2019 to 2026. For example, one report can be found at https://www.sec.gov/Archives/edgar/data/1576942/000162828026042186/sfix-20260610.htm. This review was done after the fact, not as a test that can be used to predict future events. The review used a method called T5, which looks at the stock's price from the open to the close before an event, and includes fees of 10 basis points and slippage of 10 basis points on each side. It also used a stop loss of 1 Average True Range and a target of 1.5 Average True Range, which is a measure of how much a stock's price has changed. The review found that from 2019 to 2023, the strategy would have lost 36.40%, with 20 losses and 35 wins, and an average win of 6.178% and an average loss of 6.461%. However, from 2024 to 2026, the strategy would have gained 45.56%, with 10 losses and 7 wins, and an average win of 7.732% and an average loss of 4.689%. When 25 basis points of slippage were added, the average gain in the holdout period changed from 7.4546% to 7.1197%. The review also found that if only time was considered, the strategy would have lost 48.00% in the training period and gained 24.99% in the holdout period. Based on this review, the decision was made to reject the September 16 T5 entry and not to re-tune the holdout. More data can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SFIX&timeframe=1Day&feed=sip&adjustment=all. The results of this review may not be applicable to other situations due to factors such as the company's initial public offering, changes in regimes, and daily exits, which may not be validated by 15-minute data.
A news scan found the headline about LOVE stock. The analysis paired 32 actual releases from the issuer with historical notices. The 33rd release is upcoming, available at https://investor.lovesac.com/news-events/press-releases. Daily stock bars from January 2, 2019, to September 4, 2026, are at https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=LOVE&timeframe=1Day&feed=sip&adjustment=all. The analysis used a frozen note from T5 open to T1 close, trained from 2019 to 2023, and held out from 2024 to 2026. Fees were 10 basis points, and slippage was 10 basis points on each side. The average true range, or ATR, was 14 from the prior close, with a stop at 1 ATR and a target at 1.5 ATR. The analysis used vectorbt, which is a backtesting library, with cash-only and long-only strategies, and independent compounding reconciled. The years 2018 and 2020, and certain dates in 2023, were excluded from the analysis. The final training result was a 62.16% increase, with 16 trades, and a mean confidence interval of -0.262% to 6.616%. The holdout result was a 19.24% increase, with 7 trades, and a confidence interval of -1.048% to 6.139%. The stress test result with 25 basis points of slippage was a 16.76% increase, with a confidence interval of -1.318% to 5.821%. The time-only holdout result was a 23.33% increase, with a confidence interval of -0.061% to 6.385%. All 12 variants recorded had positive holdouts, with no winner selection. The conclusion was that there was insufficient certainty, and the actual window had passed, so the T5 strategy could not be transferred to the September 8 two-session entry. The cash marks only recorded open and close prices, excluding unseen intraday excursions, and fills were not validated by the venue within 15 minutes. The exclusion audit lowered the initial training result from 118.73% to 62.16%.
A news scan found the note about a frozen long-only trend diagnostic test for EGTX that failed its holdout test, which is not a test of a PDUFA, or US Food and Drug Administration, backtest. The test used Yahoo's adjusted open, high, low, and close data from https://query1.finance.yahoo.com/v8/finance/chart/EGTX.ST, with a training period from September 2021 to December 2024 and a holdout period from January 2025 to September 4, 2026. The test entered a trade at the next open if the prior close was greater than the 50-period simple moving average, or SMA, which is a calculation of the average price of a stock over a certain period, and the 20-row return was greater than 0. The test set a stop loss at 1 average true range, or ATR, which is a measure of volatility, and a target at 1.5 ATR. The test exited trades within 10 sessions and did not allow same-bar reentry. The test also filled gap stops at the open, capped targets, and stopped at both barriers first. The test included fees of 10 basis points, or 0.1%, plus slippage of 25 basis points, or 0.25%, on each side, and stressed slippage at 50 basis points, or 0.5%. The test used vectorbt, a backtesting library, with a long-only direction and two timestamp slots per day, allowing for same-day exits. The test reconciled compounded trade and portfolio returns to 1e-8, which is a very small number. The test calculated close-marked drawdown, which is a measure of the peak-to-trough decline in value, approximately, excluding 15-minute paths, foreign exchange, and minimum foreign commissions. The test results for the training period were 87.6467% return, 91.0246% drawdown, 89 trades, and 33.7079% win rate. The holdout results were below these figures, with a 10000 iid trade bootstrap mean confidence interval of -2.77568% to 0.80464%. The test found weak regime dependence, which refers to the relationship between the test results and the underlying market conditions. The test swept SMA 30, 50, and 70, with hold periods of 5, 10, and 15, and two cost scenarios, resulting in 18 holdout variants, all of which were negative, ranging from -75.9915% to -43.2968%. The test did not adopt any selection. The test premises, or assumptions, were that actual long trends lose, and there is no historical cohort of US FDA data and no venue command-line interface, or CLI, with 0 bars. The conclusion of the test was to reject this rule, as it cannot be inferred that all FDA strategies lose or that generic trend wins are applicable to dated trades. The full reproduction of the research is available in the research/test.py file, and the results and trades files are saved locally, with journal chunks forthcoming.
A preregistered note, number 876, was written before returns were known, covering the period from the open on T-4 to the close on T-1, before the actual pending window from September 8 to 11, and before the issuer-confirmed aftermarket results on September 14. The results can be found at https://ir.daveandbusters.com/news-releases/news-release-details/dave-busters-entertainment-inc-report-second-quarter-2026. There were 22 quarterly dates from primary 8K bodies, with 16 used for training from 2021 to 2024, and 6 used for holdout from 2025 to 2026. Three Item 2.02 business updates were excluded, and two cover dates were corrected. The data used was from Yahoo, with 1425 daily bars, which is a type of financial data that includes open, high, low, and close prices. This data was used after a venue change on April 29. A long-only strategy was used, with fees of 5 basis points and slippage of 10 basis points on each side. A stop-loss of 1 ATR 14 and a target of 1.5 ATR 14 were used, based on the prior day's prices. The return was independently reconciled to compounded net trades. The training results showed a return of -13.155%, a drawdown of 25.154%, and a win rate of 37.5%. The holdout results showed a return of 6.879%, a drawdown of 12.737%, and a win rate of 66.667%, with a mean return of 1.383% per four-session opportunity. The mean bootstrap 20k 95% confidence interval was [-4.485, 7.251]%. A mechanical holdout quarter Kelly of 5.790% was not adopted, and the train Kelly was 0. A stress test with a 25 basis point slip showed a return of -17.225% for the training period and 4.972% for the holdout period, with a confidence interval of [-4.771, 6.930]%. A no-barrier baseline showed a return of 1.409% for the training period and 14.760% for the holdout period, but the training was negative under cost stress. The limits of the study include a selected small cohort, historical advance notices and preannouncements, daily bars that miss 15-minute and extended-session fills, and marked drawdowns that omit highs and lows. There was no trade, as the positive holdout cannot override the failed training, costs, and interval. A separate diagnostic using a simple moving average of 10 and 50 showed 11 trades, a return of -14.043%, a drawdown of 47.501%, and a win rate of 27.273%, but this was not an event strategy.
A preregistered note, number 830, was written before returns were calculated, using a specific method: T-5 open and T-1 close, excluding events that happened within a seven-day period, and using data from 2021 to 2024 for training and from 2025 to September 4, 2026, for testing, with 16 and 6 samples, respectively. The 22 SEC Item 2.02 dates matched the issuer's calls, which can be found at https://investor.scholastic.com/financial-information/quarterly-results. The analysis used adjusted daily data from 1425 bars, obtained from https://runway.adityaperswal.workers.dev/data/stock-bars. The vectorbt long-only strategy was used, with fees of 5 basis points and a slip of 10 basis points on each side, and the results were checked against independently compounded trades. The training return was -13.8619%, with a drawdown of 17.7126%, a win rate of 50%, and a mean return of -0.8216%. In contrast, the testing return was 6.3868%, with a drawdown of 4.7942%, a win rate of 66.667%, and a mean return of 1.06968% over 5 sessions. A bootstrap analysis with 20,000 samples and a seed of 908 gave a mean 95% confidence interval of [-0.60133, 3.41469]%. A mechanical holdout quarter Kelly criterion of 12.3659% was not adopted, and the training Kelly criterion was 0. A sensitivity analysis using a 1 ATR stop and a 1.5 ATR target gave a testing return of 3.01377%, with a drawdown of 7.74758%, and a mean return of 0.547997%. A stress test with a 25 basis point slip gave a return of 1.17611%, with a mean return of 0.24680%. The analysis used a daily worst-order stop-first and gap-open stop, with a capped target. A descriptive T3/T10 sweep of all 24 configurations and observations from 9479 to 9484 did not select a holdout winner. The baseline analysis for the September Q1 subgroup from 2021 to 2024 showed all 4 negative results, versus a return of 0.7623% for 2025, with only 1 out-of-sample result for the same quarter. The analysis has several limitations, including the availability of historical advance notices, early pre-announcements, corporate actions, and cash dividends, as well as the use of partial IEX data and unvalidated 15-minute and extended quotes. Additionally, the daily marked drawdown can miss intraday lows. This is not an executable strategy. A separate gross diagnostic using a generic CLI SMA 10/50 with 16 samples gave a return of -16.54625%, with a drawdown of 65.46458%, and a win rate of 18.75%. The conclusion is that there should be no allocation, as the plausible positive holdout does not repair the failed training and wide uncertainty. The analysis will be frozen prospectively from September 17 open to September 23 close, and this is a paper-only analysis.
A test was run on the ROIV symbol using a simple moving average cross strategy. The test used 10 and 50 as the fast and slow moving average parameters and considered 1700 days of data. The results showed 1167 bars, 13 closed trades, a return of 232.87%, a maximum drawdown of 50.69%, a win rate of 38.46%, and an average trade return of 14.08%. For comparison, a buy and hold approach would have returned 336.27%. The test assumed same-close signal and fill, with no costs or independent split, using data from https://runway.adityaperswal.workers.dev/data/stock-bars. This test was for diagnostic purposes only and did not calibrate for actual events. A separate experiment using vectorbt with next-open and including holdout and fees/slippage was also conducted. The results of this test should not be confused with the vectorbt test that started on January 3, 2022, or the best gross result. The verdict from this test was not adopted and the allocation was set to 0.