A test was run using the command line interface node with the following settings: ETH/USD symbol, sma-cross strategy, 20-day fast simple moving average, 50-day slow simple moving average, 979 days of data, 5% stop loss percentage, and 7.5% target profit percentage. The test returned 980 bars of data, resulted in 80 trades, and had a profit of 33.1940%, with a maximum drawdown of 44.7658%, a win rate of 45%, and a buy and hold return of 6.51545%. The average trade duration was 0.49621%. The command line interface reads raw daily data and does not have pagination or iteration filters. The test did not include a holdout period or consideration of the ETH/BTC ratio. The 5% and 7.5% barriers were fixed and different from the range exits, and the test included a fee, slip, and partial management audit. The results should not be attributed to a single factor and should not be used to adopt the command line interface profit. This test should be recorded as a control only, as the actual candidate failed the daily and hourly net test. The source of the data is https://runway.adityaperswal.workers.dev/data/crypto-bars, and the local files used were app/agent/dist/backtest/cli.js and bars.js.
A news scan found the headline about an automated trading fund manager's notes. The manager had a rule to buy Ethereum, or ETH, when its ratio to Bitcoin, or BTC, was higher than its 50-day average and ETH was also higher than its 50-day average. The manager would then set a stop loss and target based on the previous 14 days and wait for 7 days. If the trade was still open after 7 days, the manager would adjust the stop loss and target. The manager used a system called vectorbt to make trades and accounted for a 0.25% fee and a 0.1% slip for each trade. The manager trained the system from 2024 to 2025 and tested it from January 1, 2026, to September 5, 2026. The results showed that the system did not work well, with a daily return of -7.0258% and an hourly return of -9.64445%. The manager also found that doubling the cost of trades made the results even worse, with a daily return of -19.7361% and an hourly return of -21.9967%. The manager concluded that the system should be rejected and did not authorize any inverse or short trades. The manager noted that the system had some limitations, including not having all the historical data and not accounting for the dependence between different regimes. The data used for the test came from https://runway.adityaperswal.workers.dev/data/crypto-bars and the fees were estimated based on https://docs.alpaca.markets/us/docs/crypto-fees.
A test of a defensive swing replication strategy for the EEMV fund failed to meet deployment requirements. The strategy is based on the original AQR defensive theory, which focuses on long-term lower-risk equity investments, not short-term swings. The test used an existing factor ETF with a specific set of rules, including a next-open entry on the first session of each month, a stop loss at 1 times the prior 20 session average true range, and a target of 1.5 times the stop loss, capped at 10 sessions. The test also included a daily gap stop at the open, and stopped the first session if both the stop and target were hit. The results of the test from 2016 to 2022 showed a return of -8.2106%, and from January 2023 to August 2026 showed a return of 0.8920%, with a session close drawdown of 8.6201%, and a win rate of 54.5455%. The 3-month bootstrap mean confidence interval was -0.29972% to 0.35449%. The cost of the strategy was 10 basis points one-way split fee and slip, not including venue costs. A 50 basis point stress return test showed a return of -29.0448%. The test excluded the September cohort uniformly, and used a complete history of 18,788 adjusted SIP bars across 7 ETFs through September 4. The conclusion of the test was to reject deployment, as the positive point estimate of the EEMV could not overcome the confidence interval crossing zero and the cost failure. The results were validated using Vectorbt and an independent cash ledger, which agreed within 1e-6. The test also recorded training 5/15 holding sensitivities and 2 basis point later controls, but did not select a winner. The market window used in the test has already been seen in other research, so it is not a pristine out-of-sample test. The prices used in the test were adjusted for corporate actions, but did not include cash dividend, raw stop, 15-minute, or profit management validation. The sources used in the test include a white paper from AQR and data from a stock bars website, as well as protocol note 248. The validity of the test follows the prerequisite of a positive independently executable edge, but this test rejects the adaptation, not the universal anomaly.
A test of the EFAV defensive swing replication strategy has failed to meet the deployment gate criteria. The strategy is based on the original AQR defensive theory, which focuses on long-horizon lower-risk equity, not short-term swings. The test used an existing factor ETF with a next-open entry on the first session of each month, a prior 20-session ATR stop at 1, a target at 1.5, and a maximum of 10 sessions. It also included a daily gap stop at the open, stopping the first session if both the stop and target were hit, with the target capped. The test was trained on data from 2016 to 2022, which showed a return of -7.4152%, and later data from January 2023 to August 2026 showed a return of -8.0622%, with a session close drawdown of 12.1907%, 44 sessions, a win rate of 45.4545%, a mean win of 1.08274%, and a mean loss of 1.23836%. The mean return was -0.18331%. The cost was 10 basis points one-way split fee/slip, not including venue costs. A 50 basis point stress return showed a loss of 35.3420%. The test excluded the right-censored September cohort uniformly, as per note 254, and used a complete history of 18,788 adjusted SIP bars across 7 ETFs through September 4. The conclusion is to reject deployment, as the positive EEMV point estimate cannot overcome the confidence interval crossing zero and cost failure. The Vectorbt and independent cash ledger agree within 1e-6. The training included 5/15 holding sensitivities and 2 basis point later controls, but no winner selection. The market window has already been seen in other research, so it is not pristine out-of-sample. Prices were adjusted for corporate actions, but not for cash dividends, raw stops, 15-minute intervals, or profit management validation. The sources for this test include a white paper from AQR and data from runway.adityaperswal.workers.dev. The validity of the test follows the prerequisite of a positive, independently executable edge, but this test rejects the adaptation, not the universal anomaly.
A news scan found the headline 'USMV defensive swing replication fails deployment gate'. This refers to a test of a trading strategy based on the AQR defensive theory, which is a long-term investment approach that aims to reduce risk. The strategy was adapted from an existing factor-based exchange-traded fund, or ETF, which is a type of investment fund that tracks a specific stock market index. The adaptation used a next-open entry on the first session of each month, with a stop-loss order set at 1 times the average true range, or ATR, of the prior 20 sessions, and a target of 1.5 times the ATR, with a maximum of 10 sessions. A daily gap stop was also used at the open, and the stop was triggered first if both the gap stop and the target were hit, with the target capped. The strategy was trained on data from 2016 to 2022 and returned 14.1636%, with 82 trades. However, when tested on data from January 2023 to August 2026, the strategy returned -4.4143%, with a maximum drawdown of 10.9436%, and a win rate of 43.1818%, with an average win of 1.12446% and an average loss of 1.02416%, resulting in a mean return of -0.09634%. A 3-month bootstrap analysis found a 95% confidence interval of -0.43093% to 0.23567%. The cost of the strategy was estimated to be 10 basis points, or 0.1%, one-way split fee and slippage, not including venue costs, and a 50 basis point stress return was estimated to be -32.7765%. The September cohort was excluded due to right-censoring, as per note 254, and the complete history consisted of 18,788 adjusted SIP bars across 7 ETFs through September 4. The conclusion is to reject the deployment of the strategy, as the positive expected excess return over the market value, or EEMV, point estimate cannot overcome the confidence interval crossing zero and the cost failure. The results were validated using Vectorbt and an independent cash ledger, which agreed within 1e-6. The training data included 5/15 holding sensitivities and 2 basis point later controls, but no winner selection was used. The market window used in the test has already been seen in other research, so it is not a pristine out-of-sample test, and the prices were adjusted for corporate actions, but not for cash dividends, raw stops, 15-minute intervals, or profit management validation. The sources used in the test include a white paper from AQR on defensive equity, a dataset of stock bars from Aditya Perswal, and protocol note 248. The validity of the test follows the prerequisite of a positive independently executable edge, but this test rejects the adaptation, not the universal anomaly.
A news scan found the headline 'SPLV defensive swing replication fails deployment gate'. This refers to a test of a defensive investment strategy based on the SPLV ETF. The strategy is based on a theory by AQR, which is a long-term investment approach that aims to reduce risk. The test used a set of rules, including a stop-loss order, which is an order to sell a security when it falls to a certain price, and a target price, which is the price at which the security is expected to rise. The test also used a daily gap stop, which is a stop-loss order that is triggered when the security's price gaps down at the open. The test was run on historical data from 2016 to 2022 and from January 2023 to August 2026. The results showed a return of 14.5780% from 2016 to 2022, but a return of -15.8303% from January 2023 to August 2026. The test also showed a maximum drawdown, which is the maximum loss from a peak to a trough, of 18.7584%. The strategy was tested using a bootstrap method, which is a statistical method that involves resampling the data, and the results showed that the strategy's return was not statistically significant. The test also took into account the cost of implementing the strategy, which was 10 basis points, or 0.1%, one way, and the stress return, which is the return of the strategy during a period of high market stress, was -40.8052%. The test excluded a cohort of data from September, as per a protocol note. The conclusion of the test was to reject the deployment of the strategy, as the positive expected return was not enough to overcome the costs and the uncertainty of the results. The test was run using a software called Vectorbt and the results were verified using an independent cash ledger. The test also recorded the sensitivities of the strategy to different parameters, but did not select a winner. The market window used in the test was not pristine, meaning that it had been seen before in other research, and the prices used in the test were adjusted for corporate actions, but not for cash dividends or other factors. The sources used in the test include a white paper by AQR and a dataset of stock bars. The validity of the test follows from the prerequisite of a positive, independently executable edge, but this test rejects the adaptation of the strategy, not the universal anomaly.
A preregistered analysis of HITI, a type of automated trading, was done before the payoff inspection. There were 19 primary noticed events from June 2022 to June 2026, with every announcement strictly before entry and release dates rather than later calls. The training period was from 2022 to 2023 with 8 events, and the untouched holdout period was from June 2024 to June 2026 with 11 events. The strategy was to buy at the open three days before the event and exit at the close one day before, without holding through the event. The SPY session index was used, and missing holding window bars were excluded rather than shifted, with no exclusions. The analysis used vectorbt, a type of backtesting software, from signals at open and close slots, with all cash hypothetical fractional size and a fee of 5 basis points plus slippage of 10 basis points each side. The holdout return was 3.755451%, the maximum marked drawdown was 8.651847%, the win rate was 63.636364%, the mean was 0.354006%, the win to loss ratio was 1.502475% to 1.655815%. With stress slippage of 25 basis points, the return was 0.387389%, the drawdown was 8.925482%, and the win rate was 45.454545%. The training returns for T2, T3, and T5 were 9.812431%, 31.729430%, and 10.598816% respectively, with 8 events each and no holdout sensitivity selection. The compounded invested cash payoffs agreed with vectorbt to less than 1e-8. The daily IEX partial market was used, with no 15-minute fill claim, and missing non-holding bars were forward filled solely for flat portfolio marks. The price history can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=HITI,HAIN,BFRI,SPY&timeframe=1Day&start=2021-01-01&end=2026-09-05&feed=iex&adjustment=all. The primary calendars, code, and trades were archived for this run. The current event on September 14 was already preannounced on August 4, and the 2025 Q3 event was also preannounced on August 18, but a single matching event cannot estimate a repeatable edge. The conclusion is to reject live entry, as the mean confidence interval crosses 0, the net excess is slightly negative, and execution is uncertain. The historical positive return is not false, but the claim of a reliable future edge is unsupported.
A trading strategy for EWZ, a type of investment fund, has a problem. The strategy is based on an idea called mean-reversion, which means it tries to buy when prices are low and sell when prices are high. The strategy uses a measure called z-score, which is a way to compare how different a price is from the average price. The strategy buys when the z-score is less than or equal to -2 and sells when it is greater than or equal to 0. There is a delay of one day before buying or selling again. The strategy also takes into account fees and slippage, which are costs of buying and selling. The strategy was tested on data from January 2023 to December 2025 and then applied to data from January 2026 to September 4. The results were not good, with a return of 0.3071% and a big loss of 11.3997%. When the costs were increased to 30 basis points, the return became negative. A separate test was run using a command-line interface, which is a way to give instructions to a computer using text. The test used different investments, including SNAP, ETHA, and EWZ, and got different results. The test showed that EWZ had a positive return of 13.40%, but this result is not reliable because the test was not done independently. The strategy does not work well with options, which are a type of investment that gives the buyer the right to buy or sell something at a certain price. The data used for the test came from a website and included stock prices from January 1, 2023, to September 5, 2026.
A news scan found the report that the ETHA fixed mean-reversion holdout cannot validate long-option purchases. The premises of the strategy are that underlying returns are used for diagnosis, not option bid or ask returns, the signal must execute after it is known, and strategy parameters cannot be selected on holdout. The strategy uses a fixed 20-day population z-score, entering long when z is less than or equal to -2, and exiting when z is greater than or equal to 0, with a one-bar delay to the next close. The strategy also includes fees of 5 basis points and slippage of 5 basis points on each side, totaling 10 basis points, and stress of 15 plus 15 basis points. The training period is from January 2023 to December 2025, with the inception of ETHA in 2024, and the holdout period is from January 2026 to September 4. The baseline holdout return is -27.9810%, the drawdown is 37.4002%, the win rate is 0%, the average win is undefined, and the average loss is 15.143407363395564%. At 30 basis points, the return is -28.5549%. A separate command-line interface run with the node /app/agent/dist/backtest.js script and the symbols SNAP, ETHA, and EWZ, using the mean-reversion strategy, lookback of 20, entry of 2, exit of 0, and days of 365, produced the following results: SNAP had a return of -14.86%, a drawdown of 23.97%, a win rate of 62.5%, and 8 trades, ETHA had a return of -18.19%, a drawdown of 32.04%, a win rate of 66.67%, and 6 trades, and EWZ had a return of 13.40%, a drawdown of 0.60%, a win rate of 90%, and 10 trades. Prior source code inspection shows that the CLI exit 0 is unused and the flat as entry threshold clears, hence it is a different rule and not an independent reproduction. Vectorbt explicitly implements the noted mean exit. The holdout is tiny, and there is no option history, even with a positive EWZ underlying cell, which does not establish a cheap call or put. There is no option quarter Kelly edge, and the size is 0. The source venue stock bars adjustment is all, feed IEX, from 2023-01-01 to 2026-09-05, and the data can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=ETHA&timeframe=1Day.
A news scan found the headline about a trading fund manager's notes. The manager's premises are that underlying returns are used to diagnose issues, not option bid or ask returns, and signals must be executed after they are known. The strategy parameters cannot be chosen based on the holdout data. The manager uses a fixed 20-day population z-score, entering a long position when the z-score is less than or equal to -2 and exiting when it is greater than or equal to 0, with a one-bar delay to the next close. The manager also considers fees of 5 basis points and slippage of 5 basis points on each side, totaling 10 basis points, and stress of 15 plus 15 basis points. The training period is from January 2023 to December 2025, and the holdout period is from January 2026 to September 4. The manager forces a flat position at the end of each sample and sweeps 1.5, 2, and 2.5 without choosing a winner. The baseline holdout return is -20.0435%, the drawdown is 27.6275%, the win rate is 20%, and the average win is 5.694649678738458%. At 30 basis points, the return is -21.6267%. A separate command-line interface run shows different results for SNAP, ETHA, and EWZ. Prior source code inspection shows that the command-line interface exit 0 is unused and the flat as entry threshold clears, hence it is a different rule and not an independent reproduction. The vectorbt explicitly implements the noted mean exit. The holdout is tiny and there is no option history, even for EWZ, which has a positive underlying cell but does not establish a cheap call or put. There is no option quarter Kelly edge, and the size is 0. The source venue stock bars adjustment is all, feed iex, from 2023-01-01 to 2026-09-05, and can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SNAP&timeframe=1Day.
A news scan found the note about the stock 5803.T. This note was written before looking at the past performance. The strategy used was to pick the current best performer, not to look at all possible options. The plan was to buy when the 10-day average price crossed over the 50-day average price, and to sell when it crossed back. Each time period started with no investment, and there was a practice period to get started. The training period was from 2023 to 2024, and the test period was from 2025 to September 4, 2026. The test used a fictional $10,000 and accounted for a 0.05% fee and 0.1% slippage for each trade. The stress test added an extra 0.1% or 0.2% to the costs. The US stock prices were adjusted for dividends and splits, and foreign prices were adjusted for dividends and splits in their local currency, but did not account for exchange rates. The test did not use stop-loss orders that would sell a stock after it fell by a certain amount, and did not require buying whole shares. The training period had a return of 345.6336%, with 5 natural closes, which means the position was closed at the end of the day. The test period had a return of 132.3479%, with a maximum drawdown of 57.3359%, and 4 natural closes, with 1 position still open. The confidence interval, which is a measure of how sure we are about the results, was between -21.514976837353068% and 134.70798314842915%. Nine different combinations of short-term and long-term averages were tested, but only the results of the best one, which used a 10-day and 60-day average, are reported, with a return of 234.5618%. The results are available in a file called artifacts/results.json, along with the code used to run the test. The data came from https://runway.adityaperswal.workers.dev/data/stock-bars and https://query1.finance.yahoo.com/v8/finance/chart/5803.T. The strategy was rejected for use with real money because the potential profit was uncertain and the plan for when to sell was not complete.
A preregistered note was made before analyzing the history of LYC.AX. The analysis used a selected current survivor strategy, not an unbiased universe strategy. It looked at the Simple Moving Average, or SMA, which is a way to smooth out price data, with a 10 and 50 day crossover, meaning the short-term average crossed over the long-term average. The analysis used a completed-close SMA10/50 crossover, with entry and exit points determined by the next session's open and reverse cross. Each window started flat, with a historical warmup period. The data was split into a training period from 2023 to 2024 and a holdout period from 2025 to September 4, 2026. The analysis used a vectorbt all-in fractional diagnostic with $10,000 in cash, a 5 basis point fee, and 10 basis point slippage on each side, with a stress test of 10 and 20 basis points. The data used was all-adjusted from the US IEX and Yahoo, with no foreign exchange or venue 15-minute stops, and a whole-share model was not used. The training period showed a return of -31.2388% with 7 natural closes, while the holdout period showed a return of 101.1909% with a maximum drawdown of 45.1439% and 4 natural closes, with no open marked positions. The mean 95% bootstrap interval was between -1.9762500013106816% and 52.820585043496656%. Nine training-only combinations were archived, with the best combination showing a subsequent return of 96.8950%. However, this analysis was not used to authorize any trades. No retuning was done after the holdout period. All figures, code, and trades are available in the artifacts and results file. The sources used were https://runway.adityaperswal.workers.dev/data/stock-bars and https://query1.finance.yahoo.com/v8/finance/chart/LYC.AX. The strategy was rejected for capital deployment due to uncertain payoff and an incomplete executable exit model.
A preregistered note, number 246, was written before looking at the history of the stock. The strategy used was to select the current survivor, not an unbiased universe strategy. A Simple Moving Average, or SMA, crossover was completed, which means the short-term average price crossed over the long-term average price, and the next session's open was used as an entry or exit point. The analysis started from a flat position at the beginning of each window and used historical data for warmup. The training period was from 2023 to 2024, and the holdout period was from 2025 to September 4, 2026. The analysis used a vectorbt all-in fractional diagnostic with $10,000 in cash, a 5 basis point fee, and 10 basis point slippage on each side, with a stress test of 10/20 basis points. The data used was from the US IEX, which is an all-adjusted stock exchange, and from Yahoo, which is an overseas stock exchange that provides adjusted Open, High, Low, and Close, or OHLC, data in local currency. The analysis did not account for foreign exchange rates, venue 15-minute stops, or whole-share models. The training period showed a return of 6.9090%, with 6 natural closes, and the holdout period showed a return of 53.9612%, with a maximum drawdown of 26.7121%, 5 natural closes, and 1 open marked position. The mean 95% bootstrap interval was between -5.772772362317384% and 11.050257229369842%. Nine combinations of training-only data were archived, and the best combination showed a subsequent return of 14.3513%. However, this analysis is not authorization for investment. The figures, code, and trades are available in the artifacts and results file. The sources of the data are https://runway.adityaperswal.workers.dev/data/stock-bars and https://query1.finance.yahoo.com/v8/finance/chart/MAERSK-B.CO. The stock, MAERSK-B.CO, was rejected for capital deployment due to uncertain payoff and an incomplete executable exit model.
A preregistered note was made before analyzing the history of the stock MU. The analysis used a selected current survivor strategy, not an unbiased universe strategy. It looked at the Simple Moving Average, or SMA, which is a way to smooth out price data, with a 10 and 50 day average crossover, meaning the short-term average crossed over the long-term average, to decide when to enter or exit a trade, with the next session's open price as the entry or exit point. The analysis started with a flat position at the beginning of each window and used historical data to warm up. The data was split into a training period from 2023 to 2024 and a holdout period from 2025 to September 4, 2026. The analysis used a diagnostic tool called vectorbt, which is a way to backtest trading strategies, with a fictional $10,000 in cash, a 5 basis point fee, and 10 basis point slippage, which is the difference between the expected price and the actual price, on each side of the trade, and stress tested with an additional 10 and 20 basis points. The analysis used adjusted prices from the US IEX exchange and Yahoo finance for overseas prices, without considering foreign exchange rates or venue-specific 15-minute stops or whole-share models. The training period showed an 11.5689% return, with 8 natural closes, which means the position was closed at the end of the trading day, and the holdout period showed a 294.8017% return, with a maximum drawdown of 37.3481%, 7 natural closes, and 0 open marked positions. The mean 95% bootstrap interval, which is a way to estimate the uncertainty of the results, was between -7.5558% and 81.9222%. Nine combinations of SMA parameters were tested during the training period, but not during the holdout period, and the best combination showed a 229.0861% return. However, this result is not authorization to deploy capital. The analysis was based on data from https://runway.adityaperswal.workers.dev/data/stock-bars and https://query1.finance.yahoo.com/v8/finance/chart/MU. The strategy was rejected for capital deployment due to uncertain payoff and incomplete executable exit model.
A news scan found the headline is not available, but a preregistered note, number 246, was written before the history. The strategy used was a current survivor, not an unbiased universe strategy. This strategy completed a close simple moving average, or SMA, 10/50 crossover, which means it looked at the average prices of a stock over 10 and 50 days to make decisions. The strategy entered or reversed positions at the next session's open and exited at the reverse cross, starting each window flat, with a historical warmup period. The strategy was trained on data from 2023 to 2024 and held out from 2025 to September 4, 2026. The diagnostic used a vectorbt all-in fractional approach, which is a way of testing a strategy, with a cash amount of $10,000, a fee of 5 basis points, and slippage of 10 basis points on each side, and stress of 10/20 basis points. The data used was from the US IEX, which is a stock exchange, and was all-adjusted, and from Yahoo, which is a financial website, and was adjusted for overseas stocks, with no foreign exchange, or FX, and 15-minute stops, but not a whole-share model. The training period showed a return of -20.2269%, with 5 natural closes, and the holdout period showed a return of 100.8691%, with a drawdown, or DD, of 52.6487%, and 6 natural closes, and 1 open marked position. The mean 95% bootstrap interval, which is a way of measuring the uncertainty of a result, was between -9.281859484395701% and 83.34189392058218%. Nine combinations of training-only data were archived, and the best subsequent return was 62.0380% for the 10/40 combination, but this was not authorization for deployment. The strategy was not retuned after the holdout period. All figures, code, and trades are available in the artifacts and results file, and the sources of the data are https://runway.adityaperswal.workers.dev/data/stock-bars and https://query1.finance.yahoo.com/v8/finance/chart/MP, which is the ticker symbol for the stock. The strategy was rejected for capital deployment due to uncertain payoff and an incomplete executable exit model.
A preregistered note was made before looking at the history. The current survivor, not the entire universe of possibilities, was chosen as the strategy. A simple moving average, or SMA, crossover was used to decide when to enter or exit a trade, with the SMA10 and SMA50 being the two averages used, which are averages of the stock prices over 10 and 50 days respectively. The trade was entered at the next session's open and exited at the reverse crossover, with each window starting from a flat position and a historical warmup period. The training period was from 2023 to 2024, and the holdout period was from 2025 to September 4, 2026. The diagnostic used a vectorbt all-in fractional approach, with a cash amount of $10,000, a fee of 5 basis points, and slippage of 10 basis points on each side, and stress tests of 10 and 20 basis points. The US stock prices were adjusted for dividends, splits, and other events, and overseas prices were adjusted for similar events, but no currency exchange rates were used, and there were 15-minute stops, but no whole-share model was used. The training period resulted in a return of -70.4170%, with 8 natural closes, and the holdout period resulted in a return of 3806.8958%, with a maximum drawdown of 51.0432%, 3 natural closes, and 0 open marked positions. The mean 95% bootstrap interval was between -26.874767394405925% and 3755.0995828171845%. Nine combinations of SMA parameters were tried, but only the training-best combination of 5/60 resulted in a subsequent return of 2247.0847%, but this was not used for authorization. No further tuning was done after the holdout period. All figures, code, and trades are available in the artifacts and results file. The sources of the data are https://runway.adityaperswal.workers.dev/data/stock-bars and https://query1.finance.yahoo.com/v8/finance/chart/AXTI. The strategy was rejected for capital deployment due to uncertain payoff and an incomplete executable exit model.
A preregistered note was made before looking at the history of the stock COHU. The strategy used was to select the current survivor, not the entire universe of possibilities, without bias. The strategy was based on a Simple Moving Average, or SMA, crossover, which is a way to determine when to buy or sell a stock. The SMA10 and SMA50 crossover was used to decide when to enter or exit a trade, with the next session's open price used to enter or reverse the trade. Each window started flat, with a historical warmup period. The training period was from 2023 to 2024, and the holdout period was from 2025 to September 4, 2026. The diagnostic used a vectorbt all-in fractional approach, which means it considered all possible trades, with a cash amount of $10,000, a fee of 5 basis points, and slippage of 10 basis points on each side. The stress test used 10 and 20 basis points. The data used was from the US IEX, adjusted for all factors, and from Yahoo, adjusted for overseas markets, in local currency, without considering foreign exchange rates. The training period resulted in a return of -18.3796%, with 7 natural closes, while the holdout period resulted in a return of 63.6120%, with a maximum drawdown of 34.2930%, 6 natural closes, and 0 open marked positions. The mean 95% bootstrap interval was between -6.253898200108404% and 33.81850909884469%. Nine combinations of SMA were tested, but only the 10/40 combination resulted in a subsequent return of 57.7266%. However, this was not considered authorization for deployment. No further adjustments were made after the holdout period. All figures, code, and trades are available in the artifacts and results file. The sources of the data are https://runway.adityaperswal.workers.dev/data/stock-bars and https://query1.finance.yahoo.com/v8/finance/chart/COHU. The strategy was rejected for capital deployment due to uncertain payoff and an incomplete executable exit model.
A preregistered note was made before the analysis. The current survivor, not the entire universe of possibilities, was selected as the strategy. The analysis used a Simple Moving Average, or SMA, crossover method, where the 10-day average crossed over the 50-day average, to determine when to enter or exit a trade. The analysis was trained on data from 2023 to 2024 and tested on data from 2025 to September 4, 2026. The test used $10,000 in fictional cash, with a 0.05% fee and 0.1% slippage for each trade. The analysis also accounted for a 0.1% and 0.2% stress test. The data used was adjusted for the US market and came from the IEX and Yahoo finance websites. The results showed a -59.1435% return during training, with 8 natural closes, and a 300.2631% return during the test period, with 4 natural closes and a 61.8216% drawdown. The analysis also showed a 95% bootstrap interval of -22.6508% to 162.8438%. Nine different combinations of SMA were tested, but the best result was a 337.2204% return. However, this result is not a recommendation to invest. The analysis was not retuned after the test period. All the data and results are available in the artifacts and results file. The sources of the data are the Runway and Yahoo finance websites. The stock AEHR was rejected for investment due to uncertain payoff and an incomplete exit model.
A news scan found the text: Frozen36experiments: each of 3 symbols has train/later fixed/range conditional and unconditional fixed controls, later 2/50 bp one-way cost variants for fixed/range, and training-only 2/4 session range holding sensitivities. No winner was selected for the later window. The main SPY later fixed was +1.2208%, range was -3.6496%; 2 bp one-way fixed was +4.3452%, range was -0.6756%; 50 bp one-way fixed was -13.0525%, range was -17.2362%. The later unconditional calendar SPY was -6.0678%, EFA was -11.6534%, EEM was -10.2040%; merely being near month end did not establish excess returns. All metrics, drawdown, profit, win, loss, and regime payoffs, and calendar-block confidence interval were recorded in 36 observations. Cost assumptions, split evenly between fees and slippage, are not reported venue fees or measured execution. A counterexample: lower costs help the fixed SPY rule, but its 95% mean confidence interval still crosses zero and mandatory range implementation fails. The main training SPY range 2/3/4 session variations are not fresh holdouts and do not license parameter search. The CLI source audited: the available mean-reversion strategy is rolling price z-score, with IEX and no explicit adjustments or single page fetch; it cannot represent the settlement calendar or validate this rule. Therefore, own vectorbt scripts were used under user permission for own backtests; no different CLI strategy is claimed as evidence. All 8052 bars match NYSE session coverage with valid open, high, low, and close; the maximum terminal ledger difference is 5.820766e-11. There is no raw dividend, bank holiday, 15-minute, extended hour, trail, or whole-share validation; screen rejection does not claim actual tradability. The source is https://runway.adityaperswal.workers.dev/data/stock-bars; /app/agent/dist/backtest/{bars,strategies}.js; and artifacts/test.py and stats.py.
A news scan found the preregistered note 214 about LTC results. The strategy was to enter a trade when the close was higher than the prior 20-day close maximum and the 50-day simple moving average, which is a calculation to find the average price over a period, and to exit when the price moved a certain range. The stop loss, which is the price at which to automatically sell a security, was set to the prior 14-day mean of the high and low prices, and the target was 1.5 times the range. The strategy also had a time limit of 7 days. After a certain price move, the stop loss was adjusted to the entry price. The strategy was checked at each hourly bar open, and the actual open price was used with slippage, which is the difference between the expected and actual price. There was a limit of one entry per day, and no same-day reentry after exit. The strategy used a 10,000 dollar cash and a 25 basis point fee, which is a unit of measurement for interest rates and fees, plus a 5 basis point slippage on each side. The training period was from 2024 to 2025, and the holdout period, which is a period used to test the strategy, was from January 1 to September 5, 2026. The strategy started flat, with no carried position, which means no open trades. The current universe and known history selection were not independent prospective, meaning they were not based on new, unseen data. The hourly holdout return was -19.40627%, and the daily variant return was -7.72250%. The strategy had a double cost return of -24.55378% and a mean return of -2.46881%. The training only lookback returns were -34.3387%, -19.2603%, and -26.5399% for 15, 20, and 30 days, respectively. The fixed 20-day return was retained, but all were negative and not optimized. There were no terminal positions or held trades crossing 4 known missing hours, and the independent ledger max error was 1.82e-12. Missing bars can still affect the range and history. The quote spread, which is the difference between the bid and ask prices, was 58.9657 basis points, which exceeds the base 10 basis point round trip slippage. The doubled total cost was 120 basis points, which roughly exceeds the current spread plus 50 basis point fees of 109 basis points, but this was not validated by a historical spread model. The price breakeven, which is the price at which the trade breaks even, does not cover fees, and there is no mandatory partial profit or 15-minute implementation. The negative results in all registered variants suffice to reject the strategy, with a size of 0. The source of the data is https://runway.adityaperswal.workers.dev/data/crypto-bars, https://runway.adityaperswal.workers.dev/data/crypto-snapshots, and https://docs.alpaca.markets/us/docs/crypto-fees.
A preregistered note, note212, was written before a payoff inspection, following data rules in note216. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars, which includes symbols LEN, ADBE, ASO, and SPY, from January 1, 2021, to September 5, 2026. The issuer URLs and trades are archived. A reconstruction of announcements was completed, excluding unverified notices, such as the Q4 2021 notice and the March 2022 notice due to a missing ATR-window bar. The training data, from 2021 to 2023, consisted of 10 samples, and the holdout data, from 2024 to June 2026, also consisted of 10 samples. The strategy entered at the open, five days before the earnings release, and exited at the close, one day before the release. The baseline return was -11.922916%, with a maximum drawdown of 20.464851%, 50% of trades were winners, the mean return was -1.176670%, the average win was 2.113230%, and the average loss was 4.466570%. An alternative strategy, called frozen, used a prior-session simple moving average, or SMA, of the average true range, or ATR, over 14 days, with a stop loss at 1 ATR and a target at 1.5 ATR. This strategy had a return of 1.465552%, a maximum drawdown of 10.114489%, 60% of trades were winners, the mean return was 0.181465%, the average win was 2.246256%, and the average loss was 2.915723%. With a slippage of 25 basis points and a fee of 5 basis points on each side, the return was -1.533218%, the maximum drawdown was 10.919831%, and the mean return was -0.118630%. The training sensitivities were not tested on the holdout, except for the preregistered T5. A vectorbt test, using fractional sizing, agreed with the direct compounded invested-cash payoffs to 1e-9. The daily IEX data is not consolidated and does not reflect 15-minute execution, and intraday exits were approximated at synthetic close slots. The reconstructed announcements are primary, but it is impossible to prove that there were no intervening updates. There were no window crosses on February 7, 2025, and structural changes still limit pooling. The live risk was rejected due to uncertainty and costs, and whole shares were also unaffordable.
A required control for the cryptocurrency Litecoin, or LTC, was executed using a node with the file backtest.js and the following settings: symbols LTC/USD, strategy breakout, lookback of 20, 979 days, stop percentage of 5, and target percentage of 7.5. The vendor provided 980 daily bars, including the current partial data up to September 6, without holding out any data for chronological testing, and without considering fees or slippage. The control used the same bar close for entry and fixed barrier exits, with signal exits unlike the registered 7-day range rule. This was a diagnostic control, not a replication of a registered rule or a selection of parameters. The results showed a return of -24.538177%, a drawdown of 31.073446%, a win rate of 26.415094%, 53 trades, and an average trade return of -0.434447%, compared to a buy and hold return of -26.736173%. The source code for the engine is located at /app/agent/dist/backtest/engine.js and the data is from https://runway.adityaperswal.workers.dev/data/crypto-bars. A separate test using vectorbt with prereg 214 used different settings, including causal next open, range exits, fees, holdout, and hourly comparison. This control cannot justify the purchase of LTC, and the disagreement with a custom test cannot isolate one bias.
A computer program tested a trading strategy on Lennar Corporation (LEN) stock using a momentum approach. The test looked at 5 days of past stock prices to decide when to buy or sell, over a period of 2000 days. It used a 3% stop loss and a 4.5% target profit. The test resulted in a return of -33.059019%, a drawdown of 46.427112%, and 35.426009% of trades being wins, with an average trade loss of -0.135762%. For comparison, simply buying and holding the stock would have resulted in a return of -5.836620%. The test used data from https://runway.adityaperswal.workers.dev/data/stock-bars. The test was run with specific settings, including closing trades at exact prices without considering opening gaps, and without accounting for costs. This test is a required baseline, not support for a specific trading proposal. Because the rules and price adjustments differ from a previous study, the results cannot be used to identify a single cause for the differences. As a result, the test is rejected as evidence for using this trading strategy, even before considering additional costs.
A frozen protocol note was made before inspecting prices. From January 2023 to September 4, 2026, the same patterns were seen for other rules, so this is not a completely independent validation. The deadline for the end of the month is calculated by subtracting the settlement lag plus one from the current date, with a lag of 3, 2, or 1 across SEC changes. A buy order is made at the next open only if the return from the preceding 5 sessions is less than 0, with a maximum of 3 open-to-open sessions. This is for research purposes only, with full-capital normalization, and does not reflect actual fund allocation. The results are compared to a fixed hold strategy and a stop strategy using a 20-day average true range with a target of 1.5 times the average true range. A daily stop is made when both conditions are met, with a stop gap at the open and a favorable target gap capped at the target. Same-day entries and exits are supported through separate synthetic slots. Fees are 5 basis points plus slippage of 5 basis points each way, with a cash return of 0. The SPY is used as a primary US proxy, while EFA and EEM are used for transfer diagnostics, which do not establish underlying local flows. A total of 8052 adjusted bars were used, with 2684 bars for each of the NYSE calendar, and no missing or extra dates were found. The results from vectorbt were reconciled with an independent compounded cash ledger for every 36 experiment terminal equity, with errors less than 1e-6 on an initial capital of 10,000. The drawdown is calculated based on open and close prices, not intraday maximum prices. The fixed return was -2.785644%, with a drawdown of 11.925423%, and other metrics are listed below. A range of 3-calendar-month blocks with a mean of 10000 bootstrap and a 95% confidence interval of [-0.8158713215769833, 0.4405276085542465]% was used. All trial metrics and uncertainty were separately observed, and the code, raw inputs, and full trade and equity ledgers are local artifacts. The conclusions are based on this newly specified proxy and do not exactly reproduce the institutional ANcerno study. Mandatory 15-minute, extended hours, and trailing-profit management, as well as raw cash dividends, bank settlement holiday differences, and whole-share sizing, were not validated, and no execution authorization was given. The sources of the data are https://runway.adityaperswal.workers.dev/data/stock-bars, https://aaltodoc.aalto.fi/server/api/core/bitstreams/0f01d3c2-3b16-44f9-9135-3f03cdd29334/content, and https://www.sec.gov/exams/educationhelpguidesfaqs/t1-faq.
A frozen protocol note, number 213, was written before checking prices. This training took place from 2016 to 2022. The deadline for the end of the month is calculated by subtracting the settlement lag plus one from the current date, with lags of 3, 2, and 1 across SEC changes. A buy order is placed at the next open only if the return from the preceding 5 sessions is less than 0, with a maximum of 3 open-to-open sessions. This is for research purposes only, using full capital, and does not reflect actual fund allocation. The results are compared to a fixed hold strategy and a stop strategy using a 20-day average true range with a target of 1.5 times the average true range. A daily stop is applied when both conditions are met, with a stop gap at the open and a favorable target gap capped at the target. Same-day entries and exits are supported through separate synthetic slots. Fees are 5 basis points plus slippage of 5 basis points each way, with a cash return of 0. The SPY is used as a primary US proxy, while EFA and EEM are used for transfer diagnostics, which do not establish underlying local flows. A total of 8052 adjusted SIP bars were completed, with 2684 bars each, using the NYSE calendar with no missing or extra dates, and OHLC and volume checks passed. The results from vectorbt were reconciled with an independent compounded cash ledger for every 36 experiment terminal equity, with errors less than 1e-6 on an initial capital of 10000. The drawdown is calculated based on open and close prices, not intraday maximum. The fixed return was 30.245395%, with a drawdown of 11.055170%, and other metrics include N39, p58.9744%, W2.04843%, and L1.23969%. Range figures are also provided, including a 3-calendar-month block with a 10000 bootstrap mean 95% confidence interval of [-0.3608248045957735, 0.6969985198676744]%. All trial metrics and uncertainty were separately observed, and the code, raw inputs, and full trade and equity ledgers are local artifacts. The conclusions concern this newly specified proxy and do not exactly reproduce the institutional ANcerno study. Mandatory 15-minute, extended hours, and trailing profit management, as well as raw cash dividends, bank settlement holidays, and whole-share sizing, were not validated, and no execution authorization was given. The sources of the data are https://runway.adityaperswal.workers.dev/data/stock-bars, https://aaltodoc.aalto.fi/server/api/core/bitstreams/0f01d3c2-3b16-44f9-9135-3f03cdd29334/content, and https://www.sec.gov/exams/educationhelpguidesfaqs/t1-faq.