Notes

A news scan found the following data line: the symbol is SPY, which is a type of stock market index fund, and the model used is an adapter, a type of computer program that helps with trading decisions. The sample used for training, which means teaching the computer program, is called train. The slippage, which is the difference between the expected and actual price of a trade, is 5 basis points, or 0.05 percent. The c_multiplier, a factor that affects trading decisions, is 1. The return, which is the profit made, is 9.570767569656734 percent. The drawdown, which is the largest loss at any point, is 14.341134246722786 percent. The CAGR, or compound annual growth rate, which is a measure of how much the investment grew per year, is 2.311317200720908 percent. The Sharpe ratio, which is a measure of risk and return, is 0.358883968134208. The annual volatility, which is a measure of how much the investment's value changed, is 7.0651066941544025 percent. There were 1008 trading sessions, which is a period of time when trades can be made, and 48 months, which is 4 years. The win rate, which is the percentage of trades that made a profit, is 47.91666666666667 percent. There were 48 trades made, and 47.91666666666667 percent of the months were positive, meaning the investment made a profit. The mean trade, which is the average profit or loss per trade, is 0.7537411734180248 percent. The average win, which is the average profit of winning trades, is 4.819735140838478 percent, and the average loss, which is the average loss of losing trades, is 2.9869732766087926 percent. The mean exposure, which is the average amount of money invested, is 0.35761137626324246. The vectorbt_max_error, which is a measure of the computer program's accuracy, is 3.637978807091713e-12. There were 102 orders, which are instructions to buy or sell, made between January 2, 2019, and December 30, 2022. A diagnostic message was found that said on February 2, the low price was 68.64, and then it changed to 678.301862, but this message is not authoritative and the correction is not fresh, and it is only for holdout, which means it is not for actual trading. The continuous controls, which are the rules that govern trading, were unaffected. The word freeze refers to a price of 2249. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/audit/change.json

A news scan found the following data line: the symbol is SPY, which is a type of stock market index fund, and the model used is a constant model, meaning it does not change over time. The sample used was a test sample. The slippage, which is the difference between expected and actual trade prices, was 15 basis points, or 0.15 percent. The c_multiplier, which is a factor used to scale trades, was 1. The return, or profit, was 89.00366944108846 percent. The drawdown, which is the largest decline in value, was 16.350101274970875 percent. The CAGR, or compound annual growth rate, was 19.09507171076641 percent. The Sharpe ratio, which is a measure of risk, was 1.4145340314520884. The annual volatility, which is a measure of price fluctuations, was 12.947869722356014 percent. There were 918 trading sessions and 44 months. The win rate, or percentage of successful trades, was 100 percent. There was 1 trade. The percentage of positive months was 70.45454545454545 percent. The mean trade, or average trade, was 97.25580118846757 percent. The average win was 97.25580118846757 percent. There was no average loss. The mean exposure, which is the average amount of capital used, was 0.8642948469117009. The vectorbt_max_error, which is a measure of the difference between expected and actual results, was 0. There were 45 orders. The start date was 2023-01-03 and the end date was 2026-08-31. A diagnostic message was also found, which mentioned a low price of 68.64 on February 2, and a price of 678.301862. It also mentioned that the correction was not authoritative and the holdout was not fresh. However, the continuous controls were unaffected. Additionally, there was a freeze at 2249. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/audit/change.json

A news scan found the headline 'SPY VM9110600 audit constant train slip15 scale1'. The audit of the SPY symbol using a constant model on the train sample with 15 basis points of slippage and a multiplier of 1 resulted in a return of 56.23209209932698, a drawdown of 29.601520994397802, and a compound annual growth rate of 11.800195274639647 percent. The Sharpe ratio, a measure of risk-adjusted return, was 0.6841544180900093. The annualized volatility was 18.940320466714926 percent. Over 1008 sessions and 48 months, the strategy had a win rate of 100 percent, with 1 trade, and 64.58333333333334 percent of months being positive. The mean trade return was 55.72972362852002 percent, and the average win was also 55.72972362852002 percent, with no average loss. The mean exposure, or the average amount of capital used, was 0.8640039985245452. There were 49 orders, and the strategy started on January 2, 2019, and ended on December 30, 2022. A diagnostic note mentioned a low of 68.64 on February 2, with a value of 678.301862, but this is not an authoritative correction and not a fresh holdout. Continuous controls were unaffected, and the freeze was at 2249. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/audit/change.json

A news scan found the headline 'SPY VM9110600 audit constant test slip5 scale1'. The audit of the SPY symbol using a constant model on a test sample with 5 basis points of slippage and a multiplier of 1 resulted in a return of 89.3532126904073, a drawdown of 16.349326779525718, and a compound annual growth rate of 19.15549316780758 percent. The Sharpe ratio, a measure of risk-adjusted return, was 1.4187412075130983, and the annualized volatility was 12.94498601405086 percent. Over 918 sessions and 44 months, the strategy had a win rate of 100 percent, with 1 trade, and 70.45454545454545 percent of months being positive. The mean trade return was 97.72489157559559 percent, and the average win return was also 97.72489157559559 percent, with no average loss return. The mean exposure was 0.864275818854325, and there were 45 orders. The test started on January 3, 2023, and ended on August 31, 2026. A diagnostic note mentioned a low of 68.64 on February 2, with a value of 678.301862, but this was not an authoritative correction and the holdout was not fresh. Continuous controls were unaffected, and there was a freeze at 2249. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/audit/change.json

A news scan found the headline 'SPY VM9110600 audit constant train slip5 scale1'. The audit results for the SPY symbol are as follows: the model used was a constant model, and the sample used was a training sample. The slippage, which is the difference between expected and actual trade prices, was 5 basis points, and the c_multiplier, a factor that scales the model's predictions, was 1. The return on investment was 56.53%, and the maximum drawdown, which is the largest decline in value, was 29.60%. The compound annual growth rate, or CAGR, was 11.85%, and the Sharpe ratio, a measure of risk-adjusted return, was 0.69. The annualized volatility, which is a measure of the investment's risk, was 18.94%. The investment had 1008 trading sessions, and 48 months, with a win rate of 100%. There was only 1 trade, and 64.58% of months were positive. The mean trade return was 56.07%, and the average win was also 56.07%. There were no losses, so the average loss is not applicable. The mean exposure, which is the average amount of capital invested, was 0.86. The vectorbt_max_error, a measure of the model's accuracy, was 0. The investment had 49 orders, and the trading period was from January 2, 2019, to December 30, 2022. A diagnostic note on February 2 mentioned a low of 68.64, which was corrected to 678.30. This correction is not authoritative and the holdout, or test, data is not fresh. However, the continuous controls were unaffected, and the freeze at 2249 is noted. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/audit/change.json

A news scan found the headline 'SPY VM9110600 audit monthly test slip15 scale1'. The provided data is a JSON output describing the performance of a trading strategy for the SPY symbol. The strategy used a monthly model on a test sample with a slippage of 15 basis points and a multiplier of 1. The results showed a return of 72.72%, a drawdown of 12.65%, and a compound annual growth rate of 16.19%. The Sharpe ratio was 1.34, and the annualized volatility was 11.72%. The strategy had 918 sessions over 44 months, with a win rate of 100% and only 1 trade. The average trade return was 13.13%, and the mean exposure was 86.55%. A diagnostic note mentioned a low of 68.64 on February 2, with a value of 678.30, but this is not an authoritative correction and the holdout is not fresh. Continuous controls were unaffected, and there was a freeze at 2249. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/audit/change.json

A news scan found the following data line: the symbol is SPY, which is a type of stock market index fund, and the model is monthly, meaning it is updated every month. The sample is train, which refers to a set of data used to train a computer model. The slippage_bp is 15, which means the expected loss in basis points, a unit of measurement for interest rates and investment returns. The c_multiplier is 1, which is a factor used to adjust the model's predictions. The return is 30.71958467425531, the drawdown is 18.121256375906757, and the CAGR_pct, or compound annual growth rate percentage, is 6.926454066989907. The Sharpe0, a measure of risk-adjusted return, is 0.6524461264049082, and the ann_vol_pct, or annual volatility percentage, is 11.23732562333039. There were 1008 sessions, 48 months, a win rate of 100 percent, and 1 trade. The positive months percentage is 62.5, the mean trade percentage is 6.127338002439735, and the average win percentage is 6.127338002439735. The average loss percentage is null, meaning there is no data. The mean exposure, which refers to the average amount of money invested, is 0.6546927280656281. The vectorbt_max_error, a measure of the model's accuracy, is 1.8189894035458565e-12. There were 36 orders, and the start date is 2019-01-02, while the end date is 2022-12-30. A diagnostic note says that on February 2, the low was 68.64, and then it changed to 678.301862, but this is not an authoritative correction and the holdout, or test data, is not fresh. The continuous controls are unaffected, and the freeze is 2249. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/audit/change.json

A news scan found the headline 'SPY VM9110600 audit monthly test slip5 scale1'. The given data is a JSON object containing information about a trading model for the symbol SPY. The model is a monthly test with a slippage of 5 basis points and a multiplier of 1. The return is 74.12482249585115, the drawdown is 12.58504847502655, and the Compound Annual Growth Rate is 16.444403333804857 percent. The Sharpe ratio, a measure of risk-adjusted return, is 1.3584936913873693. The annual volatility is 11.715084108300779 percent. There were 918 sessions, 44 months, and 1 trade, with a win rate of 100 percent. The average trade return is 13.328863054120458 percent. The mean exposure, or the average amount of capital used, is 0.8654292471826028. A diagnostic note mentions a low of 68.64 on February 2, with a value of 678.301862, but this is not an authoritative correction and the holdout is not fresh. Continuous controls are unaffected, and there is a freeze at 2249. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/audit/change.json

A news scan found the headline 'SPY VM9110600 audit monthly train slip5 scale1'. The provided data is about the stock symbol SPY, which was analyzed using a monthly model on a training sample with a slippage of 5 basis points and a multiplier of 1. The return was 31.903672590404184 percent, and the drawdown was 18.05788474595298 percent. The Compound Annual Growth Rate was 7.167776946143789 percent, and the Sharpe ratio was 0.6724341550731231. The annual volatility was 11.238884130775526 percent. There were 1008 sessions, 48 months, and 1 trade, with a win rate of 100 percent. The average trade return was 6.338650005849339 percent. The mean exposure was 0.6546411793414231. A diagnostic check on February 2 found a low of 68.64, which was adjusted to 678.301862, but this correction is not authoritative and the holdout data is not fresh. Continuous controls were unaffected, and the freeze value was 2249. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/audit/change.json

A news scan found the headline 'GLDApr19 2017 daily low 110.03 appears in 13:30Z 15min bar O 122.27/H 122.27/L 110.03/C 122.02; 57 bars for local date.' This refers to a specific data point for the GLD fund on April 19, 2017, where the low price was 110.03. The same data feed was checked and it does not prove an eligible print, meaning it does not meet certain criteria, and no correction was made to the GLD data. A screen was applied across 5 times 2687 bars of data, which is a large set of market data, and it flagged only one other instance, SPY on February 2, where the low was less than 0.8 times the minimum of the open and close prices or the high was greater than 1.2 times the maximum of the open and close prices. This threshold is a diagnostic heuristic, meaning it is a rule used to identify potential issues, but it is not proof that other bars are clean, or free of errors. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=GLD&start=2017-04-19&end=2017-04-20&timeframe=15Min&feed=sip&adjustment=raw

A news scan found the headline about a SPY price contradiction on a specific date. The daily price of SPY on February 2 was recorded as follows: the opening price was 689.58, the high was 696.93, the low was 69.005, and the closing price was 695.41. In the extended 15-minute session, the opening price was 687, the high was 697.48, and the low was 681.94, with a closing price of 697.02. The IEX adjusted daily price ranged from 685.94 to 693.21, with a low of 685.77 and a close of 691.68, compared to the SIP all price of 691.70 and a low of 68.64. There may be different definitions of daily and intraday eligibility and sessions that could cause the discrepancy. A diagnostic replacement was calculated as 681.94 multiplied by 691.70 divided by 695.41, which is lower than the IEX real-time low. The original price was retained and this is not an official correction, with no return or parameter refit. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SPY&start=2026-02-02&end=2026-02-03&timeframe=15Min&feed=sip&adjustment=raw

A news scan found the headline 'SPY VM9110600 critical OHLC outlier blocks adoption'. This refers to a problem with the SPY stock's data on February 2, 2026. The stock's daily bar data showed an open of 685.90, a high of 693.21, a low of 68.64, and a close of 691.70, which is a range of 90.29% and a close-return of 0.50%. The low on February 2 inflated the 20-day Average True Range, or ATR, which is a measure of volatility, into the March 2 entry, with a stop distance of 79.21. Although the statistical internal OHLC, or Open, High, Low, Close, ordering passed, it did not establish a valid print. The published results from the vendor, which are 2673, 2674, and 2680, remain raw and have not been validated, and the edge, or the advantage, has not been established. The system will re-fetch the raw data and 15-minute aggregation before making any corrections, and will not silently replace the low with 686.4 or optimize outlier removal. https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the following data: the empirical Kelly unit correction for the SPY fund is 2.9953113340697044 for training and 9.382307162910376 for testing. The cash-constrained Kelly is 1, and half of this value is 0.5. The expected log at half cash-constrained is 0.0006228742487276221 for training and 0.003290012749766692 for testing. The binary half risk fraction proxy is 0.031938660543686215 for training and 0.11175964318889042 for testing. The decision is to set the live fraction to 0 because the expectancy confidence interval crosses zero in both windows, venue 15-minute exits are unvalidated, and the monthly entry is not due. This decision is based on a 10,000 unit research portfolio. The binary p-q/b formula is a risk-unit fraction, not notional when wins and losses are small returns. The actual wealth Kelly maximizes the mean log of 1 plus the monthly return times a factor. Both point estimates are positive, but the live fraction is 0 due to validation uncertainty and timing, not because of a claimed negative Kelly. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/sizing.py

A news scan found the following hashes for files: the file bars.json has a hash of 917e6880861699d5205ae818148c224449dbaee28552b9a1808e18ba05915173, the file test_vm.py has a hash of b3521cfb6f39fc41907760ac216645d9d270f5cea9e11916c84447834c1c8f98, the file results.json has a hash of def99176a73ecb72fe541d3637a849dca67f5a5cc6811d95e5e0566e85935426, and the file inference.py has a hash of 3c190dbbae2043376e4be7ce5b1eda08012c67092622d32b08e6095821a002b5. This was measured on September 11, and it was noted that the primary empirical inference was frozen and not forecast. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

A news scan found the correlation data for 2022: the SPY stock has a correlation of 1 with itself, 0.8624812876997651 with EFA, 0.710881576391769 with EEM, 0.08490518149170549 with TLT, and 0.19437033055580202 with GLD. The EFA stock has a correlation of 0.8624812876997651 with SPY, 1 with itself, 0.8229418417753886 with EEM, 0.09559542486120301 with TLT, and 0.31308272726595243 with GLD. The EEM stock has a correlation of 0.710881576391769 with SPY, 0.8229418417753886 with EFA, 1 with itself, 0.05331224488779405 with TLT, and 0.29100536042057085 with GLD. The TLT stock has a correlation of 0.08490518149170549 with SPY, 0.09559542486120301 with EFA, 0.05331224488779405 with EEM, 1 with itself, and 0.34265446952136064 with GLD. The GLD stock has a correlation of 0.19437033055580202 with SPY, 0.31308272726595243 with EFA, 0.29100536042057085 with EEM, 0.34265446952136064 with TLT, and 1 with itself. This data was measured on September 11 and is based on empirical inference, not a forecast. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

A correlation test was done on September 11. The results show how closely the prices of certain stocks and funds move together. The stocks and funds are SPY, EFA, EEM, TLT, and GLD. The correlation between them is as follows: SPY and EFA have a correlation of 0.758451592357553, SPY and EEM have a correlation of 0.6722159395237813, SPY and TLT have a correlation of 0.11834935545819468, SPY and GLD have a correlation of 0.14781949029978084, EFA and EEM have a correlation of 0.7876002845720026, EFA and TLT have a correlation of 0.21903565760089785, EFA and GLD have a correlation of 0.34808110044401414, EEM and TLT have a correlation of 0.1526060670661739, EEM and GLD have a correlation of 0.3837628971166462, and TLT and GLD have a correlation of 0.17729034764039708. This is a measurement, not a forecast, based on empirical inference, which means it is based on observed data, and the primary data was frozen, meaning it was not changed or updated. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

A simulation called VM9110600 was run with the following parameters: it used 10000 paths, had 21 block sessions, and a seed of 9110600. The fund started with 187.22 USD. The results showed that over 21 sessions, the profit and loss median ranged from -7.14047330736388 to 9.14849703139084, with a 1.125154853843443 median. There was a 0.4126 probability of loss over 21 sessions and 0 gain of 349 count. Over 252 sessions, returns ranged from -7.376897558668151 to 26.104634610707905 percent, with a median of 7.643081752785874 percent. The maximum drawdown, or largest loss from a peak, ranged from 7.191690568235187 to 13.960921829242853 percent, with a median of 13.960921829242853 percent. This data was measured on September 11 and is based on empirical inference, not a forecast, and the primary data is frozen. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

A news scan found the following data: an attribution test for GLD with a seed of 9110600 and 10000 three-month blocks. The test used half Kelly as a binary risk unit proxy, not an allocation method, and log wealth sizing was separate. The results included HAC regressions, which are a type of statistical analysis that stands for Heteroskedasticity and Autocorrelation Consistent, and are descriptive. The test measured the following: adapter beta was 0.22351920148388565, annual intercept percentage was 3.136744265647853, and intercept HAC t was 0.9294568376521414. Monthly beta was 0.495164279696505, annual intercept percentage was 4.221299849098135, and intercept HAC t was 1.1668790672879124. Constant beta was 0.8934671972248269, annual intercept percentage was -0.3925229446176951, and intercept HAC t was -1.3438249468134835. The monthly minus constant mean percentage per period was -0.4589406092176363, and the monthly minus constant confidence interval percentage per period was between -1.081790913100364 and 0.3603212788200001. The correlation between lag variance and current variance was 0.5702596348777869, and the correlation between lag variance and return was -0.14183169737114673. The variance AR1 slope was 0.5735739712277674, and the variance AR1 HAC t was 5.7516902692946585. The return on lag variance HAC t was -1.003554860434899. The test also identified two regimes: one with high variance being false, lasting 5 months, with an asset mean percentage of 2.091748670133846, an asset standard deviation percentage of 5.747157445649563, a managed mean percentage of 2.019369818712524, and a constant mean percentage of 1.865802575679405; and one with high variance being true, lasting 39 months, with an asset mean percentage of 2.1561790946447785, an asset standard deviation percentage of 5.217062946774644, a managed mean percentage of 1.352509083551276, and a constant mean percentage of 1.8899763404933685. The measurement was taken on September 11. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

A news scan found the raw trade data with a probability of 0.5227272727272727, a win-loss ratio of 1.7437437571504693, and a half-Kelly value of 0.12451075149577587, which is a binary risk-unit proxy. The mean percentage change was 1.156883181190205, with a mean win percentage of 4.645716001599482 and a mean loss percentage of 2.6642194316390024. The monthly fund data showed a probability of 0.5227272727272727, a win-loss ratio of 1.8924630688304855, and a half-Kelly value of 0.13526534278698155. The mean percentage change was 0.7539657674908142, with a mean win percentage of 2.7869879747326474 and a mean loss percentage of 1.4726776023454795. The trade mean confidence interval percentage was between -0.2881342249072965 and 2.3222796686171248, while the fund month mean confidence interval percentage was between 0.01373467667745678 and 1.4555121959554778. The reasons for trade decisions included time, gap stop, target, stop, and gap target. The target hit rate was 0.25, with a target Wilson 95 of 0.14574066080685744 and 0.39440849808414946. There were 16 partials, and the R quantiles were -1.959755413307422, -1.3046508896221773, 0.15738871032945762, 2.756317595246988, and 2.7815737898311736. The same day exits were 0, and the mean calendar days were 19.704545454545453. This data is from a September 11 measurement, using a seed of 9110600 and 10,000 three-month blocks. The half-Kelly value is not an allocation, but a binary risk-unit proxy, and log wealth sizing is separate. The HAC regressions are descriptive. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

The GLD fund's attribution analysis shows the following results: the adapter's beta is 0.34407744455387385, annual intercept percentage is -0.9131940158626164, and intercept HAC t is -0.2937257327549947. The monthly analysis has a beta of 0.6993729838228316, annual intercept percentage of -0.8997592156742243, and intercept HAC t of -0.3902269627410152. The constant analysis has a beta of 0.891720594891813, annual intercept percentage of -0.11579082233651371, and intercept HAC t of -1.6881293982451433. The monthly minus constant mean percentage point is -0.2166204083689447, with a confidence interval of -0.6013459502178562 to 0.1755379277895631. The correlation between lag variance and current variance is 0.1559189870926584, and the correlation between lag variance and return is 0.16280959466108585. The variance AR1 slope is 0.15538275284595088, with a HAC t of 2.3745860181116263. The return on lag variance HAC t is 1.5317803983212177. The regimes are as follows: one regime has low variance, lasts 12 months, with an asset mean percentage of 0.4873846949298044, asset standard deviation percentage of 3.9745062947810745, managed mean percentage of 0.48588903113111664, and constant mean percentage of 0.4157041221854729. Another regime has high variance, lasts 36 months, with an asset mean percentage of 0.8860077116600936, asset standard deviation percentage of 4.306554855552968, managed mean percentage of 0.47305214263676476, and constant mean percentage of 0.7852743234439056. This analysis is based on a September 11 measurement, with a seed of 9110600, and 10000 three-month blocks. The half Kelly is a binary risk unit proxy, not an allocation, and log wealth sizing is separate. HAC regressions are descriptive. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

A news scan found the raw trade data for the GLD fund with a probability of 0.4791666666666667, a win-loss ratio of 1.4300968246654404, and a half-Kelly value of 0.057486105946237376, which is a measure of binary risk. The mean percentage change was 0.39022782352469954, with a mean win percentage of 3.394105559086327 and a mean loss percentage of 2.373339693191998. The monthly fund data showed a probability of 0.4791666666666667, a win-loss ratio of 1.3245988137716957, and a half-Kelly value of 0.042982925753962425. The mean percentage change was 0.19983464724921177, with a mean win percentage of 2.3245817233694233 and a mean loss percentage of 1.7549326627813828. The trade mean confidence interval percentage was between -0.6187108737982553 and 1.2444438822666493, and the fund month mean confidence interval percentage was between -0.5278162184117247 and 0.839501250860123. The reasons for trade exits were stop losses, which occurred 19 times, time stops, which occurred 14 times, gap stops, which occurred 7 times, target hits, which occurred 5 times, and gap target hits, which occurred 3 times. The target hit rate was 0.16666666666666666, and the target Wilson 95 percentage was between 0.08695410039453222 and 0.29578100037396254. There were 12 partial trades, and the R quantiles were -1.692227471270397, -1.331577429736301, -0.5892282000623725, 2.7157358058627517, and 3.668527393305773. One trade had a same-day exit, and the mean calendar days per trade was 16.895833333333332. This data was measured on September 11, using a seed of 9110600 and 10,000 three-month blocks. The half-Kelly value is a binary risk unit proxy, not an allocation, and log wealth sizing is separate. HAC regressions are descriptive. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

A news scan found the following data: this is a test attribution for TLT VM9110600. The attribution includes an adapter with a beta of 0.28727918837873684, an annual intercept percentage of -4.1719994547878825, and an intercept HAC t of -1.725531422978228. The monthly attribution has a beta of 0.5840040736293871, an annual intercept percentage of -2.301737510044549, and an intercept HAC t of -1.1661324024074227. The constant attribution has a beta of 0.8931008315709146, an annual intercept percentage of -0.746205473729226, and an intercept HAC t of -1.0926298801278915. The monthly minus constant mean percentage point is -0.12311707750222696, with a confidence interval of -0.678661595442801 to 0.40531402648628134. The correlation between lagged variance and current variance is 0.4509539650913504, and the correlation between lagged variance and return is 0.24158811300427543. The variance AR1 slope is 0.41628732589189393, with a HAC t of 3.8629202147792965. The return on lagged variance HAC t is 1.3312122172633625. There are two regimes: one with high variance set to false, lasting 10 months, with an asset mean percentage of -0.04718678230574147, an asset standard deviation percentage of 3.0104450948192696, a managed mean percentage of -0.06690069903038998, and a constant mean percentage of -0.0580095749627918; and one with high variance set to true, lasting 34 months, with an asset mean percentage of -0.011651905611059122, an asset standard deviation percentage of 4.318516123485614, a managed mean percentage of -0.2455806176667414, and a constant mean percentage of -0.08886767150727069. This data is from a September 11 measurement, with a seed of 9110600 and 10000 three-month blocks. The half Kelly is a binary risk unit proxy, not an allocation, and log wealth sizing is separate. HAC regressions are descriptive. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

A news scan found the raw trade data with the following statistics: the probability of winning is 0.38636363636363635, the win-loss ratio is 1.284032890053771, and the mean percentage loss is -0.27452659080791236. The monthly fund data shows a probability of winning of 0.38636363636363635, a win-loss ratio of 0.9932709403848013, and a mean percentage loss of -0.3511272856472731. The trade mean confidence interval percentage is between -1.2089827292150432 and 0.7535874981654227, and the fund month mean confidence interval percentage is between -0.8267283107660485 and 0.1767716272584237. The reasons for trade exits are: stop loss 16 times, time 15 times, gap stop 10 times, and target 3 times. The target hit rate is 0.06818181818181818. The target Wilson 95 percent confidence interval is between 0.023458970172840127 and 0.1822532133661719. There are 8 partial trades, and the R quantiles are -1.5535137227171283, -1.3055367916879892, -1.0608043307292947, 2.5681412785272917, and 2.7653524050775. The same day exits are 0, and the mean calendar days are 15.931818181818182. This data is from a September 11 measurement with a seed of 9110600 and 10000 three-month blocks. The half Kelly is a binary risk unit proxy, not an allocation, and log wealth sizing is separate. HAC regressions are descriptive. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

A news scan found the following data: TLT VM9110600 train attribution. The attribution includes adapter, monthly, and constant components. The adapter has a beta of 0.18898920319319581, an annual intercept percentage of -3.971017804322722, and an intercept HAC t of -1.5293024063451097. The monthly component has a beta of 0.5344816406319213, an annual intercept percentage of 3.4447867043977163, and an intercept HAC t of 1.6522398081945386. The constant component has a beta of 0.8990896485416479, an annual intercept percentage of -0.14177507271437362, and an intercept HAC t of -1.8918033664943312. The monthly minus constant mean percentage point is 0.37305938237054187, with a confidence interval of -0.08586043078996769 to 0.9726493356885827. The correlation between lag variance and current variance is 0.04462492816416884, and the correlation between lag variance and return is -0.05081207419134459. The variance AR1 slope is 0.04468430259685598, with a HAC t of 0.7188267460323502. The return on lag variance HAC t is -0.5522330240943266. There are two regimes: one with high variance being false, lasting 11 months, with an asset mean percentage of 1.6057403691954197, an asset standard deviation percentage of 4.847595454332281, a managed mean percentage of 1.6193661900427088, and a constant mean percentage of 1.42123429841291; and one with high variance being true, lasting 37 months, with an asset mean percentage of -0.6868055066793897, an asset standard deviation percentage of 4.1687739202319625, a managed mean percentage of -0.19713549854731716, and a constant mean percentage of -0.62220035113808. This data is from a September 11 measurement, with a seed of 9110600, using 10000 three-month blocks, and a half Kelly binary risk unit proxy, which is not an allocation, but rather a measure of risk, and log wealth sizing is separate. HAC regressions are descriptive, meaning they are used to describe the relationships between variables, with HAC standing for Heteroskedasticity and Autocorrelation Consistent. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py

A news scan found the raw trade data for TLT VM9110600, which includes a probability of 0.2708333333333333, a win-loss ratio of 1.3582093371950779, and a mean percentage change of -0.9172123763907651. The monthly fund data shows a probability of 0.2708333333333333, a win-loss ratio of 1.7623577587260084, and a mean percentage change of -0.35187396209676153. The trade mean confidence interval percentage is between -1.735295068249758 and -0.03360142317442397, while the fund month mean confidence interval percentage is between -0.7798694477865459 and 0.1740682865775698. The reasons for trade exits include 25 stop losses, 9 gap stops, 8 time stops, 5 targets, and 1 gap target. The target hit rate is 0.125, and the target Wilson 95 is between 0.05856971750400844 and 0.24700727086054822. There are 8 partials, and the R quantiles are -1.413164134902296, -1.1803605244198914, -1.0650782822396985, 2.6683140247275388, and 3.3389347150391493. The same day exits are 0, and the mean calendar days are 13.708333333333334. This data is from a September 11 measurement, using seed 9110600 and 10,000 three-month blocks. Half Kelly is a binary risk unit proxy, not an allocation, and log wealth sizing is separate. HAC regressions are descriptive. https://runway.adityaperswal.workers.dev/data/stock-bars ; research/inference.py