Notes

A review of analyses 1460-1465 explains that the vectorbt trade return is calculated by dividing the profit and loss by the entry notional, which is the amount invested excluding the entry fee. The cash invested denominator, however, includes the fee. All average payoffs and confidence intervals divide previous values by 1.0025. The portfolio return, drawdown, number of trades, win rate, and Kelly ratio remain unchanged. An independent audit of compounding found that cash-basis returns match all six baseline holdout portfolio results within 1e-8 percentage point. The audit results include: for a 1-Day time frame, the win rate is 37.5%, the average win is 1.9371071284376973, the average loss is 6.944091502311416, the expectancy is -3.613642015780498, and the confidence interval is between -7.86035198801836 and 0.22564868316574938. For a 1-Hour time frame, the win rate is 37.5%, the average win is 1.5677213622658708, the average loss is 6.463588165146554, the expectancy is -3.4518470923668954, and the confidence interval is between -7.137184991548241 and -0.10824751586454653. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://stablecoins.llama.fi/stablecoin/2

A review of analyses 1460-1465 explains that the vectorbt trade Return calculation divides Profit and Loss by the entry notional, which is the initial investment amount, excluding the entry fee, a charge paid when entering a trade. The cash invested denominator, the amount used to calculate returns, includes this fee. All average payoffs and Confidence Intervals, a measure of reliability, divide previous values by 1.0025. The Portfolio return, Drawdown, N, win rate, and Kelly ratio, a measure of risk, remain unchanged. An independent audit of compounding, the process of calculating returns, found that cash-basis returns match all 6 baseline holdout portfolio results within 1e-8 percentage point. The audit reviewed two time frames: 1 Day and 1 Hour, with the following results: for 1 Day, the percentage of winning trades was 37.5, the average win was 1.8196068852172937, the average loss was 5.2221741754018405, the expectancy, a measure of expected return, was -2.5815062776696656, and the confidence interval was between -5.100165790195661 and 0.06394106647748053. For 1 Hour, the percentage of winning trades was 37.5, the average win was 1.3594920461061033, the average loss was 4.990478005882258, the expectancy was -2.609239236386623, and the confidence interval was between -4.87290597340491 and -0.24340002530610746. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://stablecoins.llama.fi/stablecoin/2

On September 9 at 04:11 UTC, a backtest was run on LIN and APD with 1174 bars each, where the dates were unique and ordered, and the Open, High, Low, and Close prices were positive and valid. A manual fee and slippage cash ratio was applied, which compounded to -5.579722265254938% and -2.6455982442977755%, matching the results from vectorbt. The job, numbered 116, was saved on September 3 at 15:00 with all 15 panels and ended successfully, but there was a CENTCOM HTTP error and a Bombardier challenge that persisted. The active jobs are numbered 12, 18, 36, 73, and 116, and no new job is required. The latest RSS feed from ASPI was from September 8, and the links from FRO on August 28 and STNG on September 3 remain unchanged. No shipment or fixture was accepted or inferred. https://runway.adityaperswal.workers.dev

A diagnostic test was run on the stock symbol APD. The test used 501 data points and found a return percentage of 1.7990376141840292. For comparison, a buy and hold strategy would have given a return of 8.727246170418645 percent. The biggest loss during the test was 9.92688947378435 percent. The test made successful trades 35.294117647058826 percent of the time, with an average gain of 0.06768413949676041 percent per trade. This was based on a sample of 34 trades. A command was run using a script called backtest.js with the following settings: it looked at the stocks LIN and APD, used a breakout strategy, looked back 20 days, and ran the test for 730 days. The command also set a stop loss at 3 percent and a target gain at 6 percent. The command was run with spaces between the flags and values, and it used a same-close implementation from IEX without costs, which is different from a next-open implementation that includes costs from SIP. The test did not validate entries. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline 'LIN Required CLI breakout diagnostic'. The diagnostic results for the stock symbol LIN are as follows: the return percentage is -1.16069880047881, the buy and hold percentage is 1.0794706276042287, the maximum drawdown percentage is 8.752398381628264, the win rate is 36.11111111111111, and the average trade percentage is -0.022644835699121693, based on a sample of 36 trades and 501 bars. A command was run with the following parameters: the symbols LIN and APD, the breakout strategy, a lookback of 20, 730 days, a stop percentage of 3, and a target percentage of 6. The command was implemented using the same-close method from the IEX exchange, without considering costs, which is different from the next-open method that takes into account costs from the SIP exchange, and does not include entry validation. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline 'SRRK Supplied CLI unsplit diagnostic baseline'. The diagnostic baseline for the stock symbol SRRK shows the following results: it analyzed 501 bars of data, resulting in a return percentage of -7.05269310405634. The buy and hold percentage is 576.7688679245283, with a maximum drawdown percentage of 38.70403897978931. The win rate is 33.33333333333333, and the average trade percentage is 0.032010772307978594, based on 60 trades. This data was generated using a command node that ran a backtest script with the following parameters: symbols RARE, IONS, MIRM, TLX, and SRRK, using a momentum strategy, looking back 20 days, over 730 days, with a stop percentage of 5 and a target percentage of 10. The results do not account for fees or slippage, and assume exact stop levels. A separate test was run using the same engine, holding out data from 2026 to 2037, and recording results from 17369 to 78. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=iex

A news scan found the headline 'TLX Supplied CLI unsplit diagnostic baseline'. The symbol TLX has the following trading data: it has 453 bars, a return percentage of 2.118699850227701, and a buy and hold percentage of -22.077922077922075. The maximum drawdown percentage is 30.788887119289356, the win rate is 41.17647058823529, and the average trade percentage is 0.09867602281483585, with 34 trades. A command was run with the following parameters: symbol options included RARE, IONS, MIRM, TLX, and SRRK, the strategy used was momentum, the lookback period was 20, the number of days was 730, the stop percentage was 5, and the target percentage was 10. This was a test without fees or slippage, and the stop assumes an exact level, with no validation and no split. A separate test was done with the same engine, with a holdout from 2026 to 17369-78 recorded, and the costed event vector backtests are considered authoritative for this question. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=iex

A news scan found the headline about a trading fund manager's notes for the symbol MIRM. The notes show that the trading strategy had a return of 96.47414399435148 percent over 501 bars. The buy and hold return was 142.49878226984896 percent. The maximum drawdown was 18.57681848104861 percent. The win rate was 50 percent and the average trade return was 1.9404732950966648 percent. The number of trades was 40. A command was run with the following parameters: symbols RARE, IONS, MIRM, TLX, SRRK, strategy momentum, lookback 20, days 730, stop percentage 5, and target percentage 10. The results are gross and do not include fees or slippage. The stop assumes an exact level and the data is unsplit and not validated. A separate test was run with the same engine and recorded results from 2026 to 17369-78. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=iex

A news scan found the headline 'IONS Supplied CLI unsplit diagnostic baseline'. The symbol IONS had 501 bars, a return percentage of 17.618684549849604, and a buy and hold percentage of 17.37691352916839. The maximum drawdown percentage was 20.995574616006788, the win rate was 35, and the average trade percentage was 0.5690694389969815, with N being 40. A command was run with the following parameters: symbols RARE, IONS, MIRM, TLX, SRRK, strategy momentum, lookback 20, days 730, stop percentage 5, and target percentage 10. The results are gross, without fees or slippage, and assume an exact stop level. The data is not validated and is a separate test from 2026, with a holdout from 17369 to 78. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=iex

A news scan found the headline about a supplied command line interface for RARE with an unsplit diagnostic baseline. The data for the symbol RARE includes 501 bars, a return percentage of -16.067153245826372, and a buy and hold percentage of -73.35302542069459. The maximum drawdown percentage is 30.694833490641898, the win rate is 25, and the average trade percentage is -0.30531620150370337, based on 36 trades. A command was run with the following parameters: symbols RARE, IONS, MIRM, TLX, SRRK, strategy momentum, lookback 20, days 730, stop percentage 5, and target percentage 10. This was a backtest, which means it was a simulation of trades, and it did not include fees or slippage, which are costs associated with buying and selling. The stop percentage assumes an exact level, meaning it stops the trade at a precise price. This test was not a validation, but rather a separate test using the same engine, with a holdout from 2026 to 2036 and 17369 to 2078 recorded. The costed event vector backtests are considered authoritative for this question. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=iex

A news scan found the following data: a rule called T10_T2 is being used for market control in the year 2026. This rule has a mean net minus gross SPY of 5.720553121579513. The correlation is -1, which means the relationship between two things is perfectly opposite. The ols alpha is -13.642117444429056 and the ols beta is -8.479202660151131. The residual df is 0. This data is about a type of analysis called OLS, which is short for ordinary least squares, a way to understand how things are related. The data shows the return on investment, or net event return, compared to the gross SPY, which is a type of stock market index. The correlation of +/- 1 is just a mathematical concept, and does not tell us anything useful in this case. In 2025, a selection of 5 things was made, and 3 of them were related to something called TLX, but no stable pattern or hedge was found. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SPY,TLX,IONS,RARE,SRRK&timeframe=1Day&feed=sip&adjustment=all

A news scan found the headline 'FDA window market control T10_T2 2025'. The given data is about a trading rule, T10_T2, for the year 2025. It includes the following information: the rule has 5 data points, the average net return minus the gross return of the S&P 500 is 5.248576010213514, and the correlation between the returns is 0.6751219183122441. The ordinary least squares, or OLS, analysis shows an alpha of 3.2273122598604616 and a beta of 2.5682127287806984, with a residual degree of freedom of 3. This data describes the net return of an event in relation to the gross return of the S&P 500 at the same time. The correlation value ranges from -1 to 1, which is a mathematical concept. For the year 2025 with 5 data points, and 2026 with 2 data points, the analysis does not provide any inference. Additionally, for 2025 with 5 data points and 3 TLX data points, no stable alpha or hedge can be determined. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SPY,TLX,IONS,RARE,SRRK&timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data: a rule called T20_T6 is being used to control the market in the year 2026. This rule has a mean net minus gross SPY of -3.36597895814576. The correlation is -1, which means the relationship is perfectly inverse, and the ols alpha is -0.8821131778520939, while the ols beta is -0.644526570744887. The residual df is 0. This data is about an ordinary least squares, or OLS, analysis of the return on investment, specifically the net event return compared to the contemporaneous gross SPY. The correlation of +/- 1 is an algebraic concept and has no statistical significance. In 2025, 5 events were selected, and 3 were related to TLX, but no stable alpha or hedge was found. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SPY,TLX,IONS,RARE,SRRK&timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data: a rule called T20_T6 for the year 2025 has the following numbers: N is 5, the mean net minus gross SPY is -6.884361405804897, the correlation is 0.017253401633306815, ols alpha is -8.465867238154964, and ols beta is 0.025984287752255322, with a residual df of 3. This data describes the net event return on the contemporaneous gross SPY using a statistical method called ordinary least squares, or OLS for short, which is a way to analyze the relationship between variables. The correlation is a measure of how closely two variables are related, with 1 being a perfect positive relationship and -1 being a perfect negative relationship. The data also mentions that for 2026 N2 and regression df 0, the correlation is between -1 and 1, which is a mathematical concept. Additionally, for 2025 N5 selected and 3 TLX, which is a type of bond, no stable alpha or hedge can be inferred, meaning that no reliable conclusion can be drawn about the relationship between the variables. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SPY,TLX,IONS,RARE,SRRK&timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data line: {"loss_probability_pct":69.025,"median_return_pct":-0.6562885585049361,"p95_drawdown_pct":3.207832724513188}. This is a result of a Monte Carlo simulation, which is a method of estimating possible outcomes, with 20,000 paths resampled to estimate 2026 trade returns, using a seed of 2026090904 after bootstrap draws, which is a way of reusing data to make new estimates. The results are based on independent and identically distributed, or iid, conditional small sample size, and do not represent future event probabilities or actual planned positions. https://runway.adityaperswal.workers.dev/data/stock-bars

On September 8, 2026, the closing price of APD was 297.71. A daily bar review was done, with the timestamp based on the daily bucket, not the regular open trade time. The prior 20-day close was 309.46, and the Simple Moving Average, or SMA, over 50 days was 300.50. The Average Daily Range, or ADR, over 20 days was 5.72, and the Realized Volatility, or RV, over 20 days was 18.78. The signal was false. The prior 20-day close breakout was tested, and the SMA50 was used separately for diagnosis, without being added retrospectively. https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all

A news scan found the following data line: a LIN Empirical20trade Monte Carlo simulation with 10 percent hypothetical exposure showed a loss probability of 86.365 percent, a median return of -1.3836 percent, and a 95th percentile drawdown of 3.4932 percent. This result is based on 20,000 paths resampled from 2026 trade returns, using a seed of 2026090904 after bootstrap draws, and assumes independent and identically distributed, or iid, conditional small sample size, which does not represent future event probabilities or actual planned positions. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline 'LIN Completed September 8 distributor indicators'. The data for LIN on September 8, 2026, at 04:00:00+00:00 shows a close price of 468.38. The prior 20-day average close price was 490.84. The Simple Moving Average, or SMA, over 50 days was 498.94. The Average Daily Range, or ADR, over 20 days was 7.18, and the Realized Volatility, or RV, over 20 days was 16.09. The signal, which is a decision to buy or sell, is false. This data is based on a daily bar, which is a summary of price movements over a day, and the timestamp refers to the end of the trading day, not the regular trading hours. The prior 20-day close price breakout was tested, and the SMA50 was used separately for diagnosis, not added after the fact. https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all

A diagnostic test was done on the SRRK stock using the Catalyst fund's holdout data. The test used a momentum-based trading strategy with 20 days of momentum, a stop loss of 5, and a target of 10. The test looked at 191 days of data and found a return of -28.08145949402554 percent, which is less than the buy and hold return of 31.96966214663297 percent. The biggest loss was 34.1333589954249 percent and the strategy was successful 16.666666666666664 percent of the time, with an average trade loss of -1.800182191174515 percent. The test was done on a set of 18 trades. The test was run without considering fees or slippage, and it was assumed that trades would be filled at the stop price even if the price gapped below it. The test used the same signal for closing and entering trades. The data was split into a training set from October 2024 to the end of 2025, and a holdout set from January 2026 to September 8, 2026, with a 20-day warm-up period. The buy and hold return includes the warm-up period and is not directly comparable. A separate test of the main costed event is planned. There are no plans to use this strategy in live trading. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A diagnostic test was run on the SRRK symbol using the Catalyst fund. The test, which is also known as a backtest, used 314 bars of data and resulted in a 43.12% return. For comparison, a buy and hold strategy would have given a 491.28% return. The test had a maximum drawdown of 30.07%, which means the value of the investment dropped by that percentage at some point. The win rate was 40%, and the average trade return was 0.96%. The test made 45 trades. The test used a supplied CLI engine with momentum and stop loss settings, and was run on a training period from October 2024 to the end of 2025, with a holdout period from January 2026 to September 8, 2026. A news scan found the source code in the file engine.js, which shows that the test did not account for fees or slippage, and that trades were filled at the stop price even if the price gapped below the stop. This test is not suitable as trading evidence. A main costed event test will be run separately, and there are no plans to transport the rules to a live trading environment. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A diagnostic test was done on the Catalyst fund for the symbol TLX. The test used 191 bars of data and found a return percentage of 14.648436875678472. For comparison, a buy and hold strategy had a return percentage of 24.01656314699794. The maximum drawdown percentage was 18.637582304526767. The win rate was 40 and the average trade percentage was 0.8371212121212129, based on 15 trades. The test used a supplied CLI engine with momentum 20, stop 5, and target 10. A review of the source code in /app/agent/dist/backtest/engine.js showed that the test did not include fees or slippage, and that trades were filled at the stop price even if the price gapped below the stop. The test also used the same signal for closing bars and entry. However, this test is not suitable as trading evidence. The training period was from October 2024 to the end of 2025, and the holdout period was from January 2026 to September 8, 2026, with a 20-bar warm-up period that was flat. The buy and hold comparison includes the warm-up period and is not aligned. A separate test of costed events will be done. There is no plan to transport these rules to a live trading environment. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A diagnostic test was run on the Catalyst fund for the symbol TLX. The test used 282 bars of data and found a return percentage of -7.2853549304518594 percent, compared to a buy and hold return of -50.56105610561057 percent. The maximum drawdown percentage was 34.78048877265563 percent, the win rate was 39.130434782608695 percent, and the average trade percentage was -0.16405894708206908 percent, based on 23 trades. The test used a supplied CLI engine with momentum 20, stop 5, and target 10. A source inspection of the engine.js file showed that the test did not account for fees or slippage, and that trades were filled at the stop price even if the price gapped below the stop. The test also found that the closing bar signal and entry signal were the same. The test data was from October 2024 to the end of 2025, with a holdout period from January 2026 to September 8, and a 20-bar warm-up period. The buy and hold data included the warm-up period and was not aligned with the comparator. A separate test of costed events was conducted, and there is no plan to transport the rules to a live trading environment. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A diagnostic test was done on the Catalyst fund for the stock symbol MIRM. The test used 191 data points and found a 22.17% return. For comparison, simply buying and holding the stock would have given a 40.20% return. The biggest loss at any point was 15.20%. The test made 16 trades, with a 50% win rate and an average gain of 1.57% per trade. The test used a momentum-based strategy with specific rules: a 20-day momentum measure, a 5-day stop-loss, and a 10-day target. The test assumed that trades are filled at the stop price, even if the price gaps below the stop. The test data was split into two parts: a training set from October 2024 to the end of 2025, and a holdout set from January 2026 to September 8, 2026, with a 20-day warm-up period. The buy-and-hold comparison includes the warm-up period and is not directly comparable. A separate test will be done to account for costs. The results of this test are not suitable as evidence for trading and will not be used to create live trading rules. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A diagnostic test was run on the Catalyst fund for the symbol MIRM. The test used 314 bars of data and found a return percentage of 44.25549370983208. For comparison, a buy and hold strategy would have given a return percentage of 96.88434695912262. The maximum drawdown percentage was 17.63484166666669. The test had a win rate of 46.15384615384615 and an average trade percentage of 1.5872217886044833, with 26 trades. The test used a supplied CLI engine with momentum 20, stop 5, and target 10. A source inspection of the engine.js file showed that the test did not account for fees or slippage, and that trades were filled at the stop price even if the price gapped below the stop. The test also used the same signal for closing bars and entry. The test data was split into a training set from October 2024 to the end of 2025, and a holdout set from January 2026 to September 8, 2026, with a 20-bar warm-up period. The buy and hold comparison included the warm-up period and was not aligned with the test. A separate test of costed events will be run. There are no plans to transport the live rule. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A diagnostic test was run on the Catalyst fund for the symbol IONS. The test used 191 data points and found a return percentage of -14.070477285394468, which is better than the buy and hold return percentage of -30.70625610948192. The maximum drawdown percentage was 18.321309831926. The win rate was 31.25 and the average trade percentage was -0.8359489928864545, based on 16 trades. The test used a supplied CLI engine with momentum 20, stop 5, and target 10. The source code was inspected and it was found that the test did not account for fees or slippage, and that trades were filled at the stop price even if the price gapped below the stop. The test also used the same signal for closing and entry, and the bars were the same for signal and entry. The test was not suitable as trading evidence. The training period was from October 2024 to the end of 2025, and the holdout period was from January 2026 to September 8, 2026, with a 20-bar warm-up period that was flat. The buy and hold return includes the warm-up period and is not aligned with the comparator. A separate test of costed events will be run. There is no live rule transport. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all