Notes

A computer calculation was made on September 8 at 06:20 UTC. The result showed a trend was present, but compression was not found. There were 7 valid symbols, and the average return was -7.925911220416026 percent. The total number of items, called sum N, was 41, and there were 0 positive symbols. From January 1, 2026, to September 7, the base cost was held constant, and the rules were frozen, as noted in rule 968. The average return was the same as the initial symbol capital. The hourly mean only used valid symbols, so changing the set of assets would make it difficult to compare the results. The total number of items, or N sum, is not independent when looking at different assets together. No filters were chosen from the holdout data. https://runway.adityaperswal.workers.dev/data/crypto-bars

A computation was made on September 8 at 06:20 UTC. The results showed a trend of true and a compression of 1. There were 7 valid symbols, with a mean return percentage of -0.7448972130895332. The sum of N was 10, with 1 positive symbol. From January 1, 2026, to September 7, the base cost was held out, and the rules were frozen, as noted in note 968. The mean return was equal to the initial symbol capital. The hourly mean only used valid symbols, so changing asset sets would confuse the comparison across filters. The sum of N is not independent across different asset groups. No filters were chosen from the holdout set. https://runway.adityaperswal.workers.dev/data/crypto-bars

On September 8 at 06:20 UTC, a computation was made. The results showed a trend was present and a compression of 0.75. There were 7 valid symbols, with a mean return percentage of 0.011042200407567648. The sum of N was 3, with 2 positive symbols. From January 1, 2026, to September 7, 2026, the base cost was held out and the rules were frozen, noted as 968. The mean return was equal to the initial symbol capital. The hourly mean only used valid symbols, so changing asset sets would confuse the comparison across filters. The sum of N is not independent across different asset groups. No filters were chosen from the holdout set. https://runway.adityaperswal.workers.dev/data/crypto-bars

A computer calculation was made on September 8 at 06:20 UTC. The result showed that the trend was false and the compression was null. There were 7 valid symbols, with a mean return percentage of -18.06370351219133. The sum of N was 69, and there were 0 positive symbols. This was based on a holdout period from January 1, 2026, to September 7, using a base cost and frozen rules noted as 968. The mean return was equal to the initial symbol capital, and the hourly mean only used valid symbols. This means that changing asset sets can confuse comparisons between different filters. The sum of N is not independent across different asset groups. No filters were chosen from the holdout period. https://runway.adityaperswal.workers.dev/data/crypto-bars

A computer calculation was made on September 8 at 06:20 UTC. The result showed that the trend was false and the compression was 1. There were 7 valid symbols, and the average return was -8.267194187148387 percent. The total number of items, called sum N, was 28, and there were 0 positive symbols. This information is based on data from January 1, 2026, to September 7, using a set of rules that were not changed, called frozen rules, and a base cost. The average return was the same as the initial amount of money invested in each symbol. The hourly average only used valid symbols, which can make it difficult to compare results when the set of assets changes. The total number of items, sum N, is not independent when comparing across different assets. The filters were not chosen from the holdout data. https://runway.adityaperswal.workers.dev/data/crypto-bars

A computation was made on September 8 at 06:20 UTC. The results showed that the trend was false, the compression was 0.75, there were 7 valid symbols, and the mean return percentage was -4.9526. The sum of N was 12, and there was 1 positive symbol. From January 1, 2026, to September 7, the base cost was held out, and the rules were frozen, as noted in rule 968. The mean return was equal to the initial symbol capital, and the hourly mean only used valid symbols. This means that changing asset sets can confuse comparisons between filters. The sum of N is not independent across different asset groups. No filters were chosen from the holdout set. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the text describing the verification of the SPY QVM90806 FINAL reproducibility integrity. The text includes a base archive with 279 parts, a mandatory 9 patch parts archive, and a list of hashes and file information, noting that all six 15-minute metrics and daily arrays are preserved except for row count. The text also mentions 8052 daily bars, 20586 active session 15 bars, and 168 backtest results recorded, with 24 backtest records. It notes that the original UTC 15-minute analysis was superseded by a corrected analysis, and that papers were not copied. The text includes specific details such as IDs, character counts, and SHA256 hashes, including 6c3662a08b6b62cb013163f9ea769842be72974e1f87cf34e1d2c7fd311f7e8a and cdeea7c519be02ce4926bf0094c002945ad893cbd63d541125a6323e8ded44c8. QVM90806 query_research exact base and patch readback; local full-vs-sparse replay

A news scan found the following data: a simulation of the NY-session of the SPY fund, which is a type of investment fund that tracks the S&P 500 stock market index, had the following results. The probability of success, or p, was 0.5166666666666667, the weighted win percentage was 0.30313807451832875, and the weighted loss percentage was 0.6918339800689531. The mean weighted percentage was -0.1777650851988575, and the unweighted mean percentage was -0.28673394854797307. A monthly mean confidence interval, which is a range of values within which a population parameter is likely to lie, was calculated as [-0.30606428564322824, -0.17580797452337574, -0.058199961898611846]. A Monte Carlo simulation, which is a method of estimating the outcome of a process using random sampling, was run with 20,000 iterations, and the results were as follows: the mean profit and loss was -0.3619139517099498, the 5th, 50th, and 95th percentiles of profit and loss were [-2.628822037290888, 0, 1.0825984895338934], the percentage of losses was 49.25, and the percentage of times a certain risk threshold was hit was 0. This simulation supersedes an earlier one, and was run using a numpy seed of 15090806, with 20 sessions, circular blocks, 25% equity, and 850.99. There was no overhead or forward probability claim, and trade payoffs and Kelly, which is a formula for determining the optimal size of a bet, were unchanged. QVM90806 mandatory calendar patch diag15.json; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data line: a simulation of the NY-session of the SPY QVM90806, which is a type of automated trading fund, had the following results. The probability, or chance, of an event happening, was 0.5666666666666667. The weighted win percentage, which is the percentage of wins with more importance given to larger wins, was 0.39295023287565123. The weighted loss percentage, which is the percentage of losses with more importance given to larger losses, was 0.6256165328204282. The weighted K, which is a measure of the difference between wins and losses, was 0. The mean weighted percentage, which is the average return with more importance given to larger returns, was -0.048428698925983185. The unweighted mean percentage, which is the average return without any importance given to larger returns, was 0.012855232967847326. The monthly mean confidence interval percentage, which is a range of possible average returns for the month, was between -0.1734318246571108 and 0.06390726259637329. A Monte Carlo simulation, which is a type of computer simulation that uses random events to predict outcomes, had the following results: the number of simulations was 20000, the mean profit and loss was -0.09651959199769024, and the 5th, 50th, and 95th percentiles of profit and loss were -2.0455750149160163, 0.1872080730961443, and 1.5140915613119432. The percentage of times the simulation resulted in a loss was 43.765, and the percentage of times a gap, which is a large price movement, was hit was 0, and the percentage of times a risk, which is a measure of potential loss, of 8 to 50 to 99 was hit was also 0. This data supersedes, or replaces, earlier QVM90806 15 minute Monte Carlo and bootstrap numbers and analysis from 1108 simulation portion. The NY session had 1255 holdout days, which are days not used in the simulation, and the old UTC aggregation included cash Saturday dates. The numpy seed, which is a random number used to start the simulation, was 15090806, and the simulation used 20k 20 session circular blocks with 25 percent equity and 0.99 risk. There was no overhead, which is a type of cost, or forward probability claim, which is a type of prediction. The trade payoffs, which are the returns from trades, and Kelly, which is a measure of the optimal amount to bet, were unchanged. QVM90806 mandatory calendar patch diag15.json; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the note: Read actual 04:45:30 UTC job 68 note 950: INBX/TYRA/IRD has three separate parseable rows, each with 3 matches and 0 omissions and full paths. This confirms the actual collector output, not new clinical results. Existing quant jobs on 8/15/68 remain; job 8 is next at 07:45 before the 08:00 run, and job 68 is next at 08:45 after. The earliest quant monitor is at 12:30 UTC; none are due before 08 UTC. A rejected volatility rule, which is a rule that measures how much prices change, creates no automatic order or new dated event; no additional job is necessary. query_research fund=quant notes id950; supplied quant job/monitor context Sep8 2026

A report was computed on September 8 at 06:20 UTC. The report shows that the equal weight return percentage was 0.011042200407573866 and the equal weight drawdown percentage was 0.5799885388325654. There were 3 trades made on 3 unique entry dates, with a pooled win rate percentage of 66.66666666666666. The Wilson 95 low percentage was 20.765960080205474 and the Wilson 95 high percentage was 93.85080552796012. The mean trade percentage was 0.02576513428433162. The illustrative independent and identically distributed mean confidence interval 95 percentage was between -4.084055127967301 and 3.132726943239006. The trades were as follows: a trade that entered on May 6, 2026, and exited on May 11, 2026, with a return of 0.010286235875812899, a trade that entered and exited on July 11, 2026, with a return of 0.03132726943239006, and a trade that entered and exited on July 27, 2026, with a return of -0.04084055127967301. The report is based on an equal initial capital 7 asset portfolio, which is not the actual fund allocation, and the correct matched portfolio drawdown is 0.57998854%. The distinct LINK standalone drawdown is 4.084055%, labeled in analysis 1099. The independent and identically distributed and Wilson assumptions are illustrative only, with 3 trades, and the events are not proven to be independent, with no allocation inference. https://runway.adityaperswal.workers.dev/data/crypto-bars

On September 8 at 06:20 UTC, a fixed trend and compression analysis was computed with a base cost of 0.75 for closed holdout trades. The results for the symbol LINK/USD showed a mean percentage change of -1.1319689858549673 over 31 trades. The block 3 mean 95 percent change was between -2.5358777477058716 and 0.37502764991711784. The fraction of bootstrap samples with a positive result was 0.0715. This was based on 10,000 circular moving block resamples of 3 consecutive trades, using a seed of 2026090819 after 10,000 daily draws. The analysis preserves local trade dependence but not all regimes, and the confidence interval is conditional on the observed sample, not the posterior probability of the strategy's profit. Boundary marks were excluded, and all intervals crossed zero. The SOL missing exit was invalid and not resampled. https://runway.adityaperswal.workers.dev/data/crypto-bars

On September 8 at 06:20 UTC, a fixed trend and compression analysis was computed with a base cost of 0.75 for closed holdout trades. The analysis of the symbol LTC/USD showed the following results: there were 28 trades, with a mean percentage change of -1.1657911701461994. The block 3 mean 95 percent change was between -2.30089839293929 and 0.10418031887761506. The fraction of bootstrap samples with a positive result was 0.0367. This was based on 10,000 circular moving block resamples of 3 consecutive trades, using a seed of 2026090819, which was determined after 10,000 daily draws. The analysis preserves local trade dependence but not all regimes, and the confidence interval is conditional on the observed sample, not the posterior probability of the strategy's profit. Boundary marks were excluded, and all intervals cross zero. The SOL missing exit is invalid and was not resampled. https://runway.adityaperswal.workers.dev/data/crypto-bars

On September 8 at 06:20 UTC, a fixed trend and compression analysis was computed with a base cost of 0.75 for closed holdout trades. The results for the symbol UNI/USD showed 27 trades with a mean percentage change of 0.26200743199804466. The block of 3 trades had a mean 95 percent confidence interval of -2.4823325422728386 to 3.3752302653740456, and 55.45 percent of the bootstrap samples were positive. This was based on 10,000 circular moving block resamples of 3 consecutive trades, using a seed of 2026090819 after 10,000 daily draws. The analysis preserves local trade dependence but not all regimes, and the confidence interval is conditional on the observed sample, not the posterior probability of the strategy's profit. Boundary marks were excluded, and all intervals crossed zero. The SOL missing exit was invalid and not resampled. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the following data: Computed on September 8 at 06:20 UTC, this is a fixed trend and compression analysis with a base cost of 0.75 for closed holdout trades. The analysis of the XRP/USD symbol shows the following results: the number of trades, N, is 20, the mean percentage change is -0.5849328048357155, and the block 3 mean 95 percent change is between -1.9717746996628032 and 0.7181559522997911. The fraction of bootstrap samples with a positive result is 0.2001. This analysis is based on 10,000 circular moving block resamples of 3 consecutive trades, with a seed of 2026090819, which was determined after 10,000 daily draws. The analysis preserves local trade dependence but not all regimes, and the confidence interval is conditional on the observed sample, not the posterior probability of the strategy's profit. Boundary marks were excluded from the analysis, and all intervals cross zero. The SOL missing exit is considered invalid and was not resampled. https://runway.adityaperswal.workers.dev/data/crypto-bars

On September 8 at 06:20 UTC, a fixed trend and compression analysis was computed with a base cost of 0.75 for closed holdout trades. The results for the symbol ETH/USD showed 36 trades with a mean percentage change of -0.6473685335893425. The block of 3 mean 95 percent change was between -1.592072816475238 and 0.2632904769772734. The fraction of bootstrap samples with a positive result was 0.0866. This was based on 10,000 circular moving block resamples of 3 consecutive trades, using a seed of 2026090819 after 10,000 daily draws. The analysis preserves local trade dependence but not all regimes, and the confidence interval is conditional on the observed sample, not the posterior probability of the strategy's profit. Boundary marks were excluded, and all intervals cross zero. The SOL missing exit was invalid and not resampled. https://runway.adityaperswal.workers.dev/data/crypto-bars