Notes

A news scan found the following data: Computed on September 5 from January 1 to September 4 hourly log returns, with 10,000 simulations and 246 complete 24-hour blocks, using a seed of 20260905 after payoff bootstrap draws. This is an unconditional simulation, not conditioned on signals or weekends, and does not include 15-minute fills. The simulation results for SOL/USD are as follows: 10,000 simulations, 246 complete day blocks, 72-hour horizon, illustrative stop percentage of 6.9859, illustrative target percentage of 10.4789, terminal 5th percentile of -9.0113, terminal 95th percentile of 9.6169, path loss 95th percentile of 10.4924, stop first fraction of 0.172, target first fraction of 0.0638, neither fraction of 0.7642, nominal risk cap of $8.51, stop only notional cap of $121.83, and authorized notional of $0. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the following data: Computed on September 5 from January 1 to September 4 hourly log returns, 10,000 simulations were run on 246 complete 24-hour blocks, with a seed of 20260905 after payoff bootstrap draws. This is an unconditional simulation, not conditioned on signals or weekends, and does not include 15-minute fills. The simulation results for ETH/USD are as follows: 10,000 simulations, 246 complete day blocks, 72-hour horizon, illustrative stop percentage of 4.3367, illustrative target percentage of 6.505, terminal 5th percentile of -8.9386, terminal 95th percentile of 10.0319, path loss 95th percentile of 10.7215, stop first fraction of 0.3251, target first fraction of 0.1608, neither fraction of 0.5141, nominal risk cap of $8.51, stop only notional cap of $196.25, and authorized notional of $0. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the following data: Computed on September 5 from January 1 to September 4 hourly log returns, 10,000 simulations were run over 246 complete 24-hour blocks, with a seed of 20260905 after payoff bootstrap draws. This is an unconditional simulation, not conditioned on signals or weekends, and does not include 15-minute fills. The simulation results for the BTC/USD symbol are as follows: 10,000 simulations, 246 complete day blocks, 72-hour horizon, illustrative stop percentage of 3.3917, illustrative target percentage of 5.0875, terminal 5th percentile of -6.1842, terminal 95th percentile of 7.0062, path loss 95th percentile of 7.5084, stop first fraction of 0.3206, target first fraction of 0.1556, neither fraction of 0.5238, nominal risk cap of $8.51, stop only notional cap of $212.77, and authorized notional of $0. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the following data: Seed 20260905, 10000 independent and identically distributed resamples of closed diagnostic-holdout trades were made, but these were not independent out-of-sample tests, as serial dependence was omitted, and the bootstrap-positive fraction is not a posterior probability. The data for the symbol SOL/USD on an hourly timeframe is as follows: the number of trades is 355, the 95 percent confidence interval for the mean trade percentage is between -0.681636113661289 and -0.4652501839454183, the fraction of bootstrap samples with a positive mean is 0, the payoff ratio is 1.5413441915587014, the raw quarter Kelly fraction is 0, the authorized fraction is 0, and the median number of bars held is 1. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the following data: Seed 20260905, 10000 independent and identically distributed resamples of closed trades were used for diagnostics, but these trades are not independent out-of-sample, and serial dependence was omitted. The bootstrap-positive fraction is not a posterior probability. The symbol is SOL/USD, the timeframe is daily, and the number of trades is 12. The mean trade confidence interval at 95 percent is between -1.213095980904724 and 6.605822577529508 percent. The mean positive fraction from the bootstrap is 0.7897. The payoff ratio is 2.6975208735885015. The raw quarter Kelly fraction is 0.07866115560999794. The authorized fraction is 0. The median number of holding bars is 2. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the data line: Seed20260905, 10,000 independent and identically distributed resamples of closed diagnostic-holdout trades for ETH/USD were analyzed. The results show the following statistics for hourly trades: the mean trade confidence interval at 95 percent is between -0.6086326440213502 percent and -0.34619826394421693 percent, the bootstrap-positive fraction is 0, the payoff ratio is 1.9087119733072333, the raw quarter Kelly fraction is 0, the authorized fraction is 0, and the median holding bars is 1. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the data line for ETH/USD, which is a cryptocurrency symbol, with the following information: on September 5, 2026, a test was run with 10,000 random samples of past trades. The test looked at the results of these trades, but did not account for the fact that the trades were related to each other. The test found that the average result of these trades was between -2.67% and 0.68%. It also found that about 9.52% of the time, the trades were profitable. The payoff ratio, which is a measure of how much money was made compared to how much money was lost, was 1.66. The raw quarter Kelly fraction, which is a measure of how much of the total budget should be invested, was 0. The authorized fraction, which is the amount of money allowed to be invested, was also 0. The median number of days that a trade was held was 1.5. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the following data: Seed 20260905, 10,000 independent and identically distributed resamples of closed trades were analyzed. The results are based on diagnostic-holdout trades, not independent out-of-sample trades, and serial dependence was omitted. The bootstrap-positive fraction is not a posterior probability. The data shows the following results for the BTC/USD symbol on an hourly timeframe: the number of trades is 340, the mean trade confidence interval at 95 percent is between -0.6109 percent and -0.4430 percent, the bootstrap mean positive fraction is 0, the payoff ratio is 1.3437, the raw quarter Kelly fraction is 0, the authorized fraction is 0, and the median holding bars is 1. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the following data: Seed 20260905, 10000 independent and identically distributed resamples of closed trades were analyzed. The results are based on diagnostic-holdout trades, not independent out-of-sample trades, and serial dependence was omitted. The bootstrap-positive fraction is not a posterior probability. The data shows the symbol is BTC/USD, the timeframe is daily, and the number of trades is 18. The mean trade confidence interval at 95 percent is between -1.4245 percent and 1.3227 percent. The bootstrap mean positive fraction is 0.3275, the payoff ratio is 1.9992, and the raw quarter Kelly fraction is 0. The authorized fraction is 0, and the median holding bars is 2. https://runway.adityaperswal.workers.dev/data/crypto-bars

A computed audit was done on September 5 using the vectorbt system to check the next-open fixed signal. This system was developed from September 1, 2024, to December 31, 2025, and tested from January 1 to September 4, 2026, but this test period overlaps with history that was already checked, so it is not independent. The results are given in percent units and show the return when a position is open, but do not include the open position in the closed-trade sample. There were no stop-loss or target prices used. The audit results for SOL/USD on an hourly basis are as follows: the system looked back 20 hours, was in the holdout phase, had a fee of 50 basis points, and a slip of 10 basis points. Over 5927 bars, the return was -98.50559777346359 percent and the drawdown was 98.50559777346356 percent. Out of 355 samples, the win rate was 7.323943661971831 percent, the average win was 1.68861666312286 percent, and the average loss was 1.4042532134752974 percent. The mean trade was -1.1777331661751504 percent, there were 0 open trades, and buying and holding would have resulted in a return of -18.130533416313664 percent. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5 for the SOL/USD breakout, using a reproducible vectorbt next-open fixed signal audit. The development of this audit took place from September 1, 2024, to December 31, 2025. A diagnostic holdout was done from January 1 to September 4, 2026, which overlaps with previously inspected history and is not independent. The results are given in percent units and the return marks an open position, while the closed-trade sample excludes it. There were no stops or targets set. The audit found the following results for the SOL/USD symbol on an hourly timeframe with a lookback of 20, a fee of 25 basis points, and a slip of 5 basis points: 5927 bars were analyzed, with a return of -87.42447617299416 and a drawdown of 87.42447617299419. The sample size was 355, with a win rate of 13.521126760563378, an average win of 1.358595442097282, and an average loss of 0.8814354701161116. The mean trade was -0.5785580509999345, with 0 open trades, and a buy and hold return of -18.130533416313664. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit on September 5 found that the SOL/USD breakout strategy, which uses a reproducible vectorbt next-open fixed signal, was tested. The strategy was developed from September 1, 2024, to December 31, 2025. A diagnostic test from January 1 to September 4, 2026, was not independent because it overlaps with previously inspected history. The results are in percent units and show the return when a position is open, but the sample of closed trades does not include the open position. There were no stops or targets set. The strategy used an hourly timeframe, looked back 20 periods, and had no fees or slippage. The results were: 5927 bars, a return of 5.821752539654608, a drawdown of 12.39530201989012, 355 samples, a win rate of 31.549295774647888, an average win of 0.9816088641350075, an average loss of 0.4214337819158554, a mean trade of 0.021216292331177434, 0 open trades, and a buy and hold return of -18.130533416313664. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5 for the SOL/USD breakout, using a reproducible vectorbt next-open fixed signal audit. The development period was from September 1, 2024, to December 31, 2025. A diagnostic holdout from January 1 to September 4, 2026, overlaps with previously inspected history and is not independent. The results are in percent units and the return marks an open position, excluding the closed-trade sample. There were no stops or targets. The audit found the following results for the SOL/USD symbol on an hourly timeframe with a lookback of 30, a fee of 25 basis points, and a slip of 5 basis points: 11,682 bars were analyzed, with a return of -98.55%, a drawdown of 98.55%, 584 samples, a win rate of 19.35%, an average win of 1.19%, an average loss of 1.17%, and a mean trade of -0.72%. There were no open trades at the end of the period, and the buy and hold return was -8.00%. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5. It used a reproducible vectorbt next-open fixed signal audit. The development period was from September 1, 2024, to December 31, 2025. The diagnostic holdout period was from January 1 to September 4, 2026, which overlaps with previously inspected history and is not independent. The results are in percent units. The return is marked when a position is open, but the closed-trade sample does not include this. There were no stops or targets. The audit found the following results for the SOL/USD symbol on an hourly timeframe: it looked back 20 hours, had a fee of 25 basis points and a slip of 5 basis points, and analyzed 11,682 bars. The return was -99.32688214549539 percent and the drawdown was 99.3268821454953 percent. There were 747 samples, with a win rate of 19.544846050870145 percent. The average win was 1.26815229167138 percent and the average loss was 1.1283502305288053 percent. The mean trade was -0.6599575019595589 percent. There were 0 open trades. If the asset was bought and held, the return would be -8.000576511587521 percent. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5. It used a reproducible vectorbt next-open fixed signal audit. The development period was from September 1, 2024, to December 31, 2025. The diagnostic holdout period was from January 1 to September 4, 2026, which overlaps with previously inspected history and is not independent. The results are in percent units. The return is marked when a position is open, and the closed-trade sample does not include it. There were no stops or targets. The audit was for the symbol SOL/USD on an hourly timeframe with a lookback of 15, a fee of 25 basis points, and a slip of 5 basis points. The results showed 11,682 bars, a return of -99.65586598112374, and a drawdown of 99.6560465932145. There were 870 samples with a win rate of 19.655172413793103. The average win was 1.215670107301308 and the average loss was 1.0979141399779442. The mean trade was -0.6431751672368499. There were 0 open trades. The buy and hold return was -8.000576511587521. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5. This audit used a reproducible method called vectorbt to check the next open signal. The development of this method happened from September 1, 2024, to December 31, 2025. A diagnostic test, called a holdout, was done from January 1 to September 4, 2026, but it overlaps with history that was already checked, so it is not independent. The results are in percent units. The return is marked when a position is open, but the sample of closed trades does not include this. There were no stops or targets set. The audit found the following results for the ETH/USD symbol on an hourly timeframe: it looked back 20 periods, had a fee of 50 basis points and a slip of 10 basis points, and analyzed 5927 bars. The return was -97.48%, the drawdown was 97.48%, and there were 337 trades in the sample. The win rate was 7.72%, the average win was 1.99%, and the average loss was 1.34%. The mean trade was -1.08%. There were 0 open trades at the end, and a buy and hold strategy would have resulted in a return of -17.19%. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5. This audit used a reproducible method called vectorbt to check the next open signal. The development of this method happened from September 1, 2024, to December 31, 2025. A diagnostic test, called a holdout, was done from January 1 to September 4, 2026, but it overlaps with history that was already checked, so it is not independent. The results are in percent units. The return is marked when a position is open, but the sample of closed trades does not include this. There were no stops or targets set. The audit found the following results for the ETH/USD symbol on an hourly timeframe: it looked back 20 periods, the fee was 25 basis points, and the slip was 5 basis points. The audit analyzed 5927 bars of data and found a return of -80.99 percent and a drawdown of 81.06 percent. Out of 337 samples, the win rate was 13.95 percent, the average win was 1.56 percent, and the average loss was 0.82 percent. The mean trade was -0.49 percent. There were 0 open trades at the end, and if the asset was bought and held, it would have lost 17.19 percent. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5. This audit used a reproducible method to check the signal for the next open position. The development of this method happened from September 1, 2024, to December 31, 2025. Then, a diagnostic test was done from January 1 to September 4, 2026, but this test overlaps with history that was already checked, so it is not independent. The results are given in percent units. The return is marked when a position is open, but the sample of closed trades does not include this. There were no stops or targets set. The audit found the following results for the ETH/USD symbol on an hourly timeframe: it looked back 20 periods, was in the holdout phase, had a fee of 0 basis points, and had a slip of 0 basis points. The audit analyzed 5927 bars and found a return of 43.619932920648694 percent, a drawdown of 8.982065444455966 percent, and a sample of 337 trades. The win rate was 33.82789317507419 percent, the average win was 1.0375806726602939 percent, the average loss was 0.3568601989699852 percent, and the mean trade was 0.11484976947467894 percent. There were 0 open trades at the end, and the buy and hold return was -17.193086546273406 percent. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5. It used a reproducible vectorbt next-open fixed signal audit. The development period was from September 1, 2024, to December 31, 2025. The diagnostic holdout period was from January 1 to September 4, 2026, which overlaps with previously inspected history and is not independent. The results are in percent units. The return marks an open position, and the closed-trade sample excludes it. There were no stops or targets. The audit found the following results for the ETH/USD symbol on an hourly timeframe: it looked back 30 periods, had a fee of 25 basis points and a slip of 5 basis points, and analyzed 11687 bars. The return was -92.92567247386602 percent and the drawdown was 93.02844250804796 percent. There were 535 samples, with a win rate of 19.439252336448597 percent. The average win was 1.1711264834517694 percent and the average loss was 0.8902230274841938 percent. The mean trade was -0.48951209451720284 percent. There were 0 open trades, and the buy and hold return was 18.142164119543946 percent. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5. It used a reproducible vectorbt next-open fixed signal audit. The development period was from September 1, 2024, to December 31, 2025. A diagnostic holdout period from January 1 to September 4, 2026, overlaps with previously inspected history, so it is not independent. The results are in percent units. The return is marked when a position is open, but the closed-trade sample does not include this. There were no stops or targets used. The audit found the following results for the ETH/USD symbol on an hourly timeframe: it looked back 20 periods, had a fee of 25 basis points and a slip of 5 basis points, and analyzed 11687 bars. The return was -97.3318 percent and the drawdown was 97.341 percent. There were 696 samples, with a win rate of 18.1034 percent. The average win was 1.2279 percent and the average loss was 0.8999 percent. The mean trade was -0.5147 percent. There were 0 open trades. If the asset was bought and held, the return would have been 18.1422 percent. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5 for the ETH/USD breakout strategy, which is reproducible using vectorbt with a next-open fixed signal. The development period for this strategy is from September 1, 2024, to December 31, 2025. There is a diagnostic holdout period from January 1 to September 4, 2026, but it overlaps with previously inspected history and is not independent. The results are given in percent units, with return marking the open position, and the closed-trade sample excludes this. There are no stops or targets set for this strategy. The strategy uses an hourly timeframe, looks back 15 periods, has a fee of 25 basis points, and a slip of 5 basis points. Over 11,687 bars, the return was -98.98%, with a drawdown of 98.98%. The sample size was 836 trades, with a win rate of 17.11%. The average win was 1.16%, and the average loss was 0.89%. The mean trade was -0.54%. There are currently 0 open trades. If you had just bought and held ETH/USD, the return would have been 18.14%. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5 using a reproducible vectorbt next-open fixed signal audit. The development of this audit took place from September 1, 2024, to December 31, 2025. A diagnostic holdout was done from January 1 to September 4, 2026, which overlaps with previously inspected history and is not independent. The results are in percent units and the return marks an open position, excluding the closed-trade sample. There were no stops or targets used. The audit found the following results for the BTC/USD symbol on an hourly timeframe: a lookback of 20, a fee of 50 basis points, and a slip of 10 basis points. The results include 5927 bars, a return of -97.88%, a drawdown of 97.90%, 340 samples, a win rate of 4.71%, an average win of 1.61%, an average loss of 1.26%, a mean trade of -1.13%, and 0 open trades. The buy and hold return was -8.96%. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5 for the BTC/USD breakout strategy, which uses a reproducible vectorbt next-open fixed signal. The development of this strategy took place from September 1, 2024, to December 31, 2025. A diagnostic holdout was done from January 1 to September 4, 2026, but this overlaps with previously inspected history and is not independent. The results are given in percent units and the return is marked when a position is open, but the closed-trade sample does not include this. There are no stops or targets. The strategy uses an hourly timeframe, looks back 20 periods, has a fee of 25 basis points and a slip of 5 basis points. The results show 5927 bars, a return of -83.70%, a drawdown of 84.07%, 340 samples, a win rate of 12.06%, an average win of 0.99%, an average loss of 0.74%, and a mean trade of -0.53%. There are 0 open trades and the buy and hold return is -8.96%. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5. This audit used a reproducible method to check the next open signal. The development of this method happened from September 1, 2024, to December 31, 2025. The diagnostic test from January 1 to September 4, 2026, overlaps with previously checked history, so it is not independent. The results are in percent units. The return is marked when a position is open, but the sample of closed trades does not include this. There were no stops or targets set. The audit found the following results for the BTC/USD symbol on an hourly timeframe: it looked back 20 periods, had 0 fee basis points and 0 slip basis points, and analyzed 5927 bars. The return was 25.38 percent, the drawdown was 8.61 percent, and there were 340 samples. The win rate was 37.65 percent, the average win was 0.67 percent, the average loss was 0.30 percent, and the mean trade was 0.07 percent. There were 0 open trades at the end, and the buy and hold return was -8.96 percent. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees

A computed audit was done on September 5. It used a reproducible vectorbt next-open fixed signal audit. The development period was from September 1, 2024, to December 31, 2025. The diagnostic holdout period was from January 1 to September 4, 2026, which overlaps with previously inspected history and is not independent. The results are in percent units. The return is marked when a position is open. The sample of closed trades does not include the open position. There were no stops or targets. The audit found the following results for the BTC/USD symbol on an hourly timeframe: it looked back 30 periods, had a fee of 25 basis points and a slip of 5 basis points, and analyzed 11687 bars. The return was -96.50%, the drawdown was 96.53%, and the sample size was 550. The win rate was 12.91%, the average win was 0.79%, and the average loss was 0.81%. The mean trade was -0.61%. There were no open trades at the end. If the position had been held from the start, the return would have been 48.39%. https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees