Stats

A trading strategy was tested on the stock APD from January 3, 2022, to September 8, 2026. The strategy, which is a type of momentum strategy, or a way to follow the direction of a stock's price, was trained on data from 2022 to 2025 and then tested on untouched data from 2026. The strategy entered a trade when the stock's closing price was higher than its highest closing price in the past 20 days, and exited 10 sessions later. The strategy did not allow overlapping trades and excluded any trades that were not complete. The results showed a return of -2.65%, a drawdown of 11.28%, a win rate of 42.86%, an average win of 2.15%, and an average loss of 2.20%. The strategy was tested using a type of backtesting software called vectorbt, which uses the opening price to execute trades and the closing price to mark them, and includes fees of 10 basis points, or 0.1%, on each side of the trade, plus 10 basis points of slippage, or the difference between the expected price and the actual price. The data used for the test came from a source at https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all. The results also showed a mean return of -0.34% and a quarter Kelly, a measure of the optimal fraction of wealth to bet, of 0. The strategy is not suitable for actual orders due to the small sample size and unmodeled 15-minute stops. The allocation to this strategy is 0, and there will be no refit from the holdout data.

A source at https://runway.adityaperswal.workers.dev/data/stock-bars was used to train a model called Frozen1346 with different settings. The model was trained with a lookback period of 10, 20, or 40 and a hold period of 5, 10, or 20, and costs of 20 basis points per side. The results of the training are shown in a list of 9 rows, each representing a different combination of lookback and hold periods. The list includes the return percentage, drawdown percentage, win rate percentage, average win percentage, average loss percentage, sample size, mean percentage, and quarter Kelly value for each combination. The return percentage ranged from -24.8046% to 8.3892%, and the quarter Kelly value was 0 for all combinations except one, which was 0.0081, and another which was 0.0294, and a third which was 0.0009. No best setting was transferred to the holdout set. A bootstrap analysis was performed with 20,000 independent and identically distributed samples, using a numpy seed of 2026090904, and the 95% confidence interval was found to be between -2.4286% and 1.2898%. The analysis has some limitations, including a small sample size, only one industry, and selected firms, and no causal regression is claimed. The baseline expectation is negative and the confidence interval crosses 0, which is why the quarter Kelly value is 0. This result should not be taken to mean that there are no bullish fundamentals, only that the momentum strategy did not work.

A trading strategy was tested on LIN stock from January 3, 2022, to September 8, 2026. The strategy, a type of momentum proxy, used daily price adjustments to decide when to buy and sell. It entered a trade when the closing price was higher than the prior 20-day high, and exited 10 sessions later. The test included fees of 10 basis points, or 0.1%, on each buy and sell, plus 30 basis points of slippage, which is a type of trading cost. The results showed a return of -8.55%, a drawdown of 17.34%, and a win rate of 37.5%. The average winning trade was 2.27%, and the average losing trade was -3.08%. This strategy is not suitable for actual trading orders due to the small sample size and uncertainty of daily returns. The Quarter Kelly, a measure of optimal bet size, was 0, which means no allocation was made. The adjusted history used in this test is a research proxy, not a raw simulation of stop losses.

A trading strategy was tested on LIN stock from January 3, 2022, to September 8, 2026. The strategy, which is a type of momentum trading, was trained on data from 2022 to 2025 and then tested on untouched data from 2026. The strategy entered a trade when the stock's closing price was higher than its highest closing price in the previous 20 days, and exited 10 sessions later. The strategy used a 10 basis point fee for each trade, plus a 10 basis point slippage, which is a type of trading cost. The results showed a return of -5.58%, a drawdown of 15.67%, and a win rate of 37.5%. The average winning trade was 2.68%, and the average losing trade was 2.69%. The strategy was tested 8 times, with an average return of -0.68% per trade. This strategy is not suitable for actual trading due to its high uncertainty and unmodeled 15-minute stops. The Quarter Kelly criterion, a measure of risk, was 0, which means the strategy did not meet the risk threshold. As a result, the actual allocation to this strategy was 0. The adjusted history used in this test is a research proxy, not a raw simulation of stop losses.

A command was run using a node to compare the momentum of certain cryptocurrencies over a seven-day period. The command looked at the symbols BTC/USD, ETH/USD, and SOL/USD, using a momentum strategy with a lookback of 7 days, over a period of 616 days. The command line interface, or CLI, supports rules based on price, but does not account for external factors such as supply. The data used for the comparison was approximately 617 daily bars from January 2025 to September 9, 2026, although the current bar may be incomplete. The results showed that BTC had a return of -23.5622%, a drawdown of 38.1105%, and a win rate of 32.7869% over 61 trades. ETH had a return of 31.4580%, a drawdown of 45.5262%, and a win rate of 39.6226% over 53 trades. SOL had a return of 2.82136%, a drawdown of 37.9317%, and a win rate of 45.4545% over 55 trades. In comparison, a buy and hold strategy would have resulted in returns of -16.5489%, -25.5333%, and -46.4798% for the three cryptocurrencies. An inspection of the code used for the comparison showed that it made decisions based on daily closing prices, and did not account for fees or slippage. The results of this comparison should not be used to validate the effect of USDC on trades, or to inform decisions about hourly execution. The data used for the comparison came from https://runway.adityaperswal.workers.dev/data/crypto-bars, and is stored in the file research/cli_momentum.json. No trades were made based on this comparison.

A frozen specification note1344 was made. The provider made a SIP adjustment from September 8, 2016, to September 8, 2026, for every NYSE session for each ETF, but no data from 2010 to 2015 was supplied. This was an extension test where only entries and exits within a partition were considered, and a holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, excluding boundary spanning trades. The test started with four independent cash sleeves, each with $250, and no leverage was used. The Vectorbt system was used to make explicit long-only orders with a fee of 1 basis point plus 15 basis points for slippage per side. The cash return was reconciled with the Vectorbt trade profit and loss accounting, considering the entry fee denominator. The rule sensitivity was based on the last session's open price, a lagged 20-session mean of high and low prices, a stop at 1.25 average daily range, and a target at 2 average daily range. If these conditions were not met, the target was set at the third session's close. The test managed daily high, low, open, and close prices, and stop ambiguity was handled first, with adverse gap fills at the open price and target gap capped at the target price. A bootstrap test with 10,000 complete calendar window means was run across four ETFs, not 332 independent trades. The confidence interval and per-symbol statistics were saved to match the observation title. The verdict rejected live adoption when uncertainty, cost, and execution gates failed, and a positive point estimate alone was not enough for adoption. The sources used were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note1344 was made. The provider made a SIP adjustment from September 8, 2016, to September 8, 2026, for every NYSE session for each ETF, but no data from 2010 to 2015 was provided. This was an extension test where only entries and exits within a certain partition were considered, and a holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, excluding trades that spanned boundaries. The test started with four independent cash sleeves, each with $250, and no leverage was used. The Vectorbt system was used to make explicit long-only orders with a fee of 1 basis point plus 5 basis points for slippage per side. The cash return was reconciled with the trade profit and loss accounting for the entry fee. The rule sensitivity was based on the last session's open price, a 20-session mean of high and low prices, a stop at 1.25 average daily range, and a target at 2 average daily range. If these conditions were not met, the target was set at the third session's close. The test managed daily high, low, open, and close prices, and ambiguity was resolved by stopping at the first opportunity. Adverse gap fills were executed at the open price, and target gap fills were capped at the target price. A bootstrap test with 10,000 complete calendar window means was run across four ETFs, not 332 independent trades. The results included confidence intervals and per-symbol statistics, which were saved with an observation matching the title. The verdict was to reject live adoption if uncertainty, cost, or execution gates failed, and a positive point estimate alone was not enough to adopt the strategy. The sources for this information include a study on equity returns at the turn of the month and stock bar data from a website.

A frozen specification note1344 was made. The provider made a SIP adjustment from September 8, 2016, to September 8, 2026, for every 2685 NYSE session, for each ETF, but no data from 2010 to 2015 was given. This was an extension test where only entries and exits within a certain time period were considered, and a holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, excluding trades that went across boundaries. There were four independent cash sleeves, each initially worth 250 USD, with no leverage, no periodic rebalancing between sleeves, and no idle cash. The Vectorbt system was used with explicit long-only orders, a fee of 1 basis point plus slippage of 15 basis points per side, and the raw cash return was reconciled with the Vectorbt trade profit and loss accounting, considering the entry fee. The rule sensitivity was tested with the last session's open price, a lagged 20-session mean of high and low prices, and a stop loss of 1 ADR and a target of 2 ADR, or the third session's close price if otherwise. The system was managed daily with OHLC stops to avoid ambiguity, adverse gap fills were capped at the open price, and target gaps were capped at the target price. Daily bar touch fills were preliminary, and there was no claimed 15-minute execution or whole-lot replication. A bootstrap test was done with 10,000 complete calendar window means across four ETFs, not 332 independent trades, and confidence intervals and per-symbol statistics were saved. The verdict was to reject live adoption when uncertainty, cost, or execution gates failed, and a positive point estimate alone was not enough for adoption. The sources of this information are https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note1344 was made. The provider made a SIP adjustment from September 8, 2016, to September 8, 2026, for every 2685 NYSE session, for each ETF, but no data from 2010 to 2015 was given. This was an extension test where only entries and exits within a certain time period were considered, and a holdout period from 2019 to 2025 with 83 complete calendar windows and 332 correlated ETF trades was used, excluding trades that went across boundaries. The test started with four independent cash accounts, each with $250, and no leverage was used, with no periodic rebalancing between accounts, and any idle cash was set to zero. The Vectorbt system was used to make explicit long-only orders, with a fee of 1 basis point plus a slippage of 5 basis points per side, and the cash return was reconciled with the trade profit and loss accounting, considering the entry fee. The rule sensitivity was based on the last session's open price, a 20-session mean of the high and low prices, and a stop loss of 1 ADR and a target of 2 ADR, or the third session's close price if those conditions were not met. The test was managed daily, with stops and targets set at the open, high, low, and close prices, and any ambiguity was resolved by using the first stop, and adverse gap fills were capped at the target price. The test did not claim to have 15-minute execution or whole-lot replication. A bootstrap test was run 10,000 times, using the mean of complete calendar windows across four ETFs, not 332 independent trades, and the confidence interval and per-symbol statistics were saved. The verdict was to reject live adoption when uncertainty, cost, or execution gates failed, and a positive point estimate alone was not enough to adopt the strategy. The sources used were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note1344 was made by the provider. The SIP adjustment was made for all sessions from 2016 to 2026, and it was present for each ETF, but no data was supplied for 2010 to 2015. This was an extension test where only entries and exits within a certain partition were considered, and the holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, excluding trades that spanned boundaries. The test started with four independent cash sleeves, each with $250, and no leverage was used. The Vectorbt system was used to make explicit long-only orders, with a fee of 1 basis point and slippage of 15 basis points per side. The rule sensitivity was based on the last session's open price, and a stop was set at 0.75 times the average daily range, with a target of 2 times the average daily range. If this rule was not met, the target was set at the third session's close price. The test was managed daily, with stops and targets set at the open, high, low, and close prices, and ambiguity was resolved by using the first stop. Bootstrap sampling was used to estimate the results, with 10,000 complete calendar window means across four ETFs. The results were saved, along with confidence intervals and per-symbol statistics. The verdict was to reject live adoption if uncertainty, cost, or execution gates failed, and a positive point estimate alone was not enough to adopt the strategy. The sources used were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note, number 1344, was made. The provider made a SIP adjustment from September 8, 2016, to September 8, 2026, for every NYSE session, for each exchange-traded fund, or ETF, but no data from 2010 to 2015 was given. An extension test was done, where only entries and exits within a certain partition were considered, and a holdout period from 2019 to 2025 was used, which had 83 complete calendar windows and 332 correlated ETF trades, excluding trades that span boundaries. The test started with four independent cash sleeves, each with $250, and no leverage was used, and there was no periodic rebalance between sleeves, and idle cash was set to 0. The Vectorbt system was used to make explicit long-only orders, with a fee of 1 basis point plus slippage of 5 basis points per side, and the raw cash return was reconciled with the Vectorbt trade profit and loss accounting, using the entry fee as the denominator. The rule sensitivity was based on the last session's open price, and a stop loss was set at 0.75 times the average daily range, or ADR, which is the average difference between the high and low prices, over a 20-session period, with a target of 2 times the ADR, and otherwise the third session's close price was used. The system was managed daily, with a stop loss set first, and ambiguity was resolved by using the daily open, high, low, and close prices, and adverse gap fills were capped at the target price, and daily bar touch fills were preliminary, but no 15-minute execution or whole-lot replication was claimed. A bootstrap test was done 10,000 times, using complete calendar window means across four ETFs, not 332 independent trades, and the confidence interval and per-symbol statistics were saved, matching the observation title. The verdict was to reject live adoption when uncertainty, cost, or execution gates failed, and a positive point estimate alone was not enough for adoption. The source of this information includes a publication by McConnell from Purdue University, and data from a stock bars website.

A news scan found the headline 'USDC issuance accelerates, but destination remains unidentified'. The body of the notes is as follows: Circle's net issuance of USDC on September 7 was $0.94 billion over 7 days, compared to $1.28 billion in the preceding 23 days, implying $134.29 million per day versus $55.65 million per day. This represents an increase in outstanding dollars, not $8.8 billion of net buying. Newer issuer USDC.com's circulation on September 8 was $74.47079 billion, which differs from the transparency report on September 7 of $74.28 billion and DefiLlama's report on September 8 of $74.37833 billion, due to differing timestamps and definitions. DefiLlama's same-date report from September 1 to 8 showed an increase of $0.81568 billion in USDC and $0.05974 billion in USDT, which weakens the explanation of pure USDT substitution, but does not identify the fiat source or use. The report also showed increases in ChainSOL, Hyperliquid, and ETH, but the net delta exceeds the global delta by $25.017 million, and possible bridge or coverage definitions are unresolved, so no allocation by chain can be made. The SEC's circulation includes issuer inventory, and the CCTP reshuffles global dollars, while exchange deposits can fund collateral or payments. In conclusion, potential liquidity has expanded, but the unknown share of buying majors and no justified point probability remain. A strong counterargument is that deposits may precede spot buying even without labels, and it is recommended to test future returns rather than reject the possibility. The sources of this information include Circle's transparency report, USDC.com, DefiLlama, the SEC, and the CCTP. Observations begin at 17175 and 17177, with raw files available for research.

A frozen specification note1344 was made by the provider. The provider adjusted the SIP for all sessions from September 8, 2016, to September 8, 2026, for each ETF on the NYSE, but did not supply data from 2010 to 2015. An extension test was conducted, where only entries and exits within a certain partition were considered, and a holdout period from 2019 to 2025 was used, which had 83 complete calendar windows and 332 correlated ETF trades, excluding trades that spanned boundaries. The test started with four independent cash sleeves, each with $250, and no leverage was used. The Vectorbt system was used to make explicit long-only trades, with a fee of 1 basis point and slippage of 15 basis points per side. The cash return was reconciled with the trade profit and loss accounting, taking into account the entry fee. The rule sensitivity was based on the last session's open price, with a stop loss of 0.75 times the average daily range and a target of 1.5 times the average daily range, or the third session's close price if the other conditions were not met. The test was managed daily, with stops and targets set at the open, high, low, and close prices, and adverse gap fills were capped at the target price. The results were bootstrapped 10,000 times across four ETFs, and the confidence interval and per-symbol statistics were saved. The verdict was to reject live adoption if the uncertainty, cost, or execution gates failed, and a positive point estimate alone was not enough to adopt the strategy. The sources used were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note1344 was made by the provider. The SIP adjustment was made from September 8, 2016, to September 8, 2026, and was present for every NYSE session for each ETF, but no data from 2010 to 2015 was supplied. An extension test was done, which only included entries and exits within a certain partition. The holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, and trades that spanned boundaries were excluded. The test started with four independent cash sleeves, each with $250, and no leverage was used. There was no periodic rebalancing between sleeves, and idle cash was set to 0. The Vectorbt system was used to make explicit long-only orders, with a fee of 1 basis point plus slippage of 5 basis points per side. The raw cash return was reconciled with the Vectorbt trade P&L accounting for the entry fee denominator. The rule sensitivity was based on the last session's open price, with a stop loss set at 0.75 times the average daily range and a target set at 1.5 times the average daily range, or otherwise the third session's close price. The daily OHLC stop was managed first, with adverse gap fills opening and target gap capped at the target. Daily bar touch fills were preliminary, and no 15-minute execution or whole-lot replication was claimed. A bootstrap test was done with 10,000 complete calendar window means across four ETFs, not 332 independent trades. The confidence interval and per-symbol statistics were saved, matching the title of the observation. The verdict rejected live adoption when uncertainty, cost, or execution gates failed, and a positive point estimate alone was not enough for adoption. The sources used were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note1344 was made by the provider. The provider adjusted the Systematic Investment Plan for all sessions from September 8, 2016, to 2026, for each Exchange-Traded Fund, or ETF, that was present at every New York Stock Exchange, or NYSE, session, but no data from 2010 to 2015 was supplied. This was an extension test where only entries and exits within a certain partition were considered, and a holdout period from 2019 to 2025 was used, which had 83 complete calendar windows and 332 correlated ETF trades, excluding trades that spanned boundaries. The test started with four independent cash sleeves, each with $250, and no leverage was used, with no periodic rebalancing between sleeves, and idle cash was set to $0. The Vectorbt system was used to make explicit long-only orders, with a fee of 1 basis point, or bp, plus slippage of 15bp per side, and the raw cash return was reconciled with the Vectorbt trade profit and loss accounting, taking into account the entry fee denominator. The rule for trailing stops was to use the last session's open price, and a lagged 20-session mean of the high-low range was used as a stop, with a target of 1.5 times the Average Daily Range, or ADR, and if that did not work, the third session's close was used. The system was managed daily, with open-high-low-close, or OHLC, stops, and ambiguity was resolved by using the first stop, and adverse gap fills were capped at the target, and the trail was activated after a gain of 1 ADR and updated the next day, not every 15 minutes. Daily bar touch fills were preliminary, and no claim was made that the system could be executed in 15 minutes, nor that whole lots or current caps could be replicated. A bootstrap test was run 10,000 times, using complete calendar window means across four ETFs, not 332 independent trades, and the results, including confidence intervals and per-symbol statistics, were saved and matched the title of the observation. The verdict was to reject live adoption of the system when uncertainty, cost, or execution gates failed, and a positive point estimate alone was not enough to adopt the system. The source of the information includes a study on equity returns at the turn of the month, which can be found at https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf, and stock bar data, which can be found at https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note, number 1344, was provided. The provider made a SIP adjustment from September 8, 2016, to 2026, which includes every NYSE session for each ETF, but does not include data from 2010 to 2015. This is an extension test where only entries and exits within a partition are considered, and a holdout period from 2019 to 2025 has 83 complete calendar windows and 332 correlated ETF trades, excluding boundary spanning trades. The test started with four independent cash sleeves, each with $250, and no leverage was used. The Vectorbt system was used to make explicit long-only orders, with a fee of 1 basis point plus slippage of 5 basis points per side. The rule for trailing stops is based on the last session's open price, a lagged 20-session mean of high and low prices, and a stop loss of 1 ADR and a target of 1.5 ADR. The system is managed daily, with stops and targets adjusted to avoid ambiguity and adverse gap fills. The results are based on daily bar data and do not claim 15-minute execution. A bootstrap analysis of 10,000 complete calendar window means was conducted across four ETFs. The verdict rejects live adoption if uncertainty, cost, or execution gates fail, and a positive point estimate alone is not enough for adoption. The sources for this information are https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note1344 was made. The provider made a SIP adjustment from September 8, 2016, to September 8, 2026, for every NYSE session for each ETF, but no data from 2010 to 2015 was supplied. This was an extension test where only entries and exits within a partition were considered, and a holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, excluding trades that span boundaries. The test started with four independent cash sleeves, each with $250, and no leverage was used. There was no periodic rebalance between sleeves, and idle cash was set to 0. The Vectorbt system was used to make explicit long-only orders, with a fee of 1 basis point plus 15 basis points for slippage per side. The cash return was reconciled with the Vectorbt trade P&L accounting, considering the entry fee denominator. The rule managed was to use the last session's open, with a stop at 1 ADR and a target at 1.5 ADR, based on the lagged 20-session mean of high and low prices. If this rule did not apply, the third session's close was used. The system managed daily OHLC stops to avoid ambiguity, and adverse gap fills were capped at the target. A bootstrap of 10,000 complete calendar window means was done across four ETFs, not 332 independent trades. Confidence intervals and per-symbol statistics were saved to match the observation title. The verdict was to reject live adoption when uncertainty, cost, or execution gates failed, and a positive point estimate alone was not enough for adoption. The sources used were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note1344 was made. The provider made a SIP adjustment from September 8, 2016, to September 8, 2026, for every NYSE session for each ETF, but no data from 2010 to 2015 was provided. This was an extension test where only entries and exits within a partition were considered, and a holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, excluding boundary spanning trades. The test started with four independent cash sleeves, each with $250, and no leverage was used. There was no periodic rebalance between sleeves, and idle cash was set to 0. The Vectorbt system was used with explicit long-only orders, a fee of 1 basis point, and slippage of 5 basis points per side. The raw cash return was reconciled with the Vectorbt trade P&L accounting for the entry fee denominator. The rule managed was to use the last session open, a lagged 20-session mean high-low stop of 1 ADR, and a target of 1.5 ADR. If this rule did not apply, the third session close was used. Daily OHLC stops were managed to avoid ambiguity, and adverse gap fills were capped at the target. A bootstrap of 10,000 complete calendar window means was done across four ETFs, not 332 independent trades. Confidence intervals and per-symbol statistics were saved to match the title. The verdict was to reject live adoption when uncertainty, cost, or execution gates failed, and a positive point estimate alone was not enough for adoption. The sources used were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

This is a frozen specification note1344. The provider made a SIP adjustment from September 8, 2016, to 2026, which includes every NYSE session for each ETF, but does not include data from 2010 to 2015. This is an extension test where only entries and exits within a partition are considered, and a holdout period from 2019 to 2025 has 83 complete calendar windows and 332 correlated ETF trades, excluding boundary spanning trades. The test started with four independent cash sleeves, each with $250, and no leverage was used. The vectorbt system was used to generate orders with a long-only strategy, and a fee of 1 basis point plus a slippage of 15 basis points per side was applied. The rule for opening a position is from the last session's open to the next month's third close. The system managed daily OHLC stops and ambiguity, and adverse gap fills were capped. A bootstrap of 10,000 complete calendar window means was performed across four ETFs, and confidence interval and per-symbol statistics were saved. The verdict rejects live adoption if uncertainty, cost, or execution gates fail, and a positive point estimate alone is not enough for adoption. The sources for this information are https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note, number 1344, explains how a provider adjusted its system. The adjustment, which started on September 8, 2016, and will end on September 8, 2026, happens every 2685 NYSE sessions for each exchange-traded fund, or ETF. No data from 2010 to 2015 is provided. This is an extension test where only entries and exits within a certain partition are considered. A holdout period from 2019 to 2025 has 83 complete calendar windows and 332 correlated ETF trades, excluding trades that span boundaries. The test started with four independent cash sleeves, each with $250, and no leverage was used. There was no periodic rebalance between sleeves, and idle cash was set to zero. A system called Vectorbt was used to make explicit long-only orders, with a fee of 1 basis point plus 5 basis points for slippage per side. The cash return was reconciled with the trade profit and loss accounting, considering the entry fee. The rule for opening a trade is from the last session's open to the next month's third close. Daily high, low, open, and close prices were managed to stop ambiguity, and adverse gap fills were capped. A target gap was also capped. Bootstrap sampling of 10,000 complete calendar window means was done across four ETFs, not 332 independent trades. Confidence interval and per-symbol statistics were saved to match the observation title. A verdict rejects live adoption when uncertainty, cost, or execution gates fail. A positive point estimate alone is not enough for adoption. The sources for this information are https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note1344 was made. The provider made a SIP adjustment from September 8, 2016, to September 8, 2026, for every NYSE session, for each ETF, but no data from 2010 to 2015 was provided. This was an extension test where only entries and exits within a partition were considered, and a holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, excluding boundary-spanning trades. The test started with four independent cash sleeves, each with $250, and no leverage was used. There was no periodic rebalance between sleeves, and idle cash was set to 0. The Vectorbt system was used with explicit long-only orders, a fee of 1 basis point, and slippage of 15 basis points per side. The cash return was reconciled with the trade P&L accounting for the entry fee denominator. The rule for the parent was to use the close of the second-to-last session to determine the close of the next month's third session. Daily OHLC stops were managed to avoid ambiguity, and adverse gap fills were capped at the target. A bootstrap of 10,000 complete calendar window means was run across four ETFs, not 332 independent trades. The verdict rejects live adoption when uncertainty, cost, or execution gates fail, and a positive point estimate alone is not enough for adoption. The sources used were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note1344 was made by the provider. The provider made a SIP adjustment from September 8, 2016, to 2026, which includes every NYSE session for each ETF, but does not include data from 2010 to 2015. This is an extension test where only entries and exits within a partition are considered. The holdout period from 2019 to 2025 has 83 complete calendar windows and 332 correlated ETF trades, excluding boundary spanning trades. The test started with four independent cash sleeves, each with $250, and no leverage was used. There was no periodic rebalance between sleeves, and idle cash was set to 0. The vectorbt system was used with explicit long-only orders, a fee of 1 basis point, and slippage of 5 basis points per side. The raw cash return was reconciled with the VBT trade P&L accounting for the entry fee denominator. The rule for the parent is to use the close of the penultimate session to the close of the next month's third session. The test was managed daily with OHLC stops, and ambiguity was handled by filling adverse gaps and capping target gaps. Daily bar touches were used for preliminary fills, but no 15-minute execution or whole-lot replication was claimed. A bootstrap of 10,000 complete calendar window means was used across four ETFs, not 332 independent trades. Confidence intervals and per-symbol statistics were saved to match the title of the observation. The verdict rejects live adoption when uncertainty, cost, or execution gates fail, and a positive point estimate alone is not enough for adoption. The sources used were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

A frozen specification note, number 1344, was made. The provider made an adjustment to the SIP, which is a system that provides stock prices, from September 8, 2016, to September 8, 2026, for every NYSE session for each exchange-traded fund, or ETF, which is a type of investment fund. No data from 2010 to 2015 was provided. A holdout test was done, which means that only the entries and exits within a certain partition, or section, were considered. The holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, and trades that spanned boundaries were excluded. Four independent cash sleeves, or pools of money, were initially equal, with $250 in each, and there was no leverage, which means borrowing money to invest, and no periodic rebalancing, or adjusting, of the sleeves. The vectorbt system, which is a backtesting framework, was used to make explicit long-only trades, which means buying stocks with the expectation of selling them at a higher price, with a fee of 1 basis point, or 0.01%, plus slippage, or the difference between the expected and actual price, of 15 basis points, or 0.15%, per side. The cash return was reconciled with the vectorbt trade profit and loss accounting, which means that the cash return was compared to the profit and loss of the trades, and the entry fee was considered. The rule sensitivity was tested, which means that the rules for making trades were adjusted to see how they affected the results. The rules were based on the last session open, the lagged 20 session mean, which is the average of the high and low prices over the last 20 sessions, and the stop, which is the price at which a trade is automatically sold, was set at 1.25 times the average daily range, or ADR, and the target was set at 2 times the ADR. If these conditions were not met, the trade was made based on the third session close. The daily open, high, low, and close, or OHLC, prices were used to manage the trades, and the stop was set first to avoid ambiguity, or confusion. Adverse gap fills, which are trades made at a worse price than expected, were filled at the open, and target gap fills were capped at the target price. Daily bar touch fills, which are trades made when the price touches a certain level, were preliminary, and it was not claimed that the trades were executed at a specific 15-minute interval, nor was it claimed that the trades were replicated exactly. A bootstrap test, which is a statistical test, was done 10,000 times, and the results were saved. The verdict was that the strategy should not be adopted if the uncertainty, cost, or execution gates, which are the conditions that must be met for a trade to be made, failed. A positive point estimate, which is a single estimate of the expected return, alone was not enough to adopt the strategy. The sources of the information were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

This is a frozen specification note1344 from a provider. The provider made an adjustment to the Systematic Investment Plan, which includes all NYSE sessions from September 8, 2016, to September 8, 2026, for each Exchange-Traded Fund, but does not include data from 2010 to 2015. A holdout test was conducted, which only included entries and exits within a specific partition. The holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, excluding trades that span boundaries. The test started with four independent cash sleeves, each with $250, and no leverage was used. There was no periodic rebalance between sleeves, and idle cash was set to 0. The Vectorbt system was used to generate orders, with a long-only strategy, a fee of 1 basis point, and slippage of 5 basis points per side. The cash return was reconciled with the Vectorbt trade profit and loss accounting, considering the entry fee denominator. The rule sensitivity was based on the last session's open, a lagged 20-session mean of high-low, with a stop at 1.25 times the average daily range and a target at 2 times the average daily range. If this condition was not met, the target was set to the third session's close. The system managed daily high-low-close stops, with ambiguity resolved by using the first stop, and adverse gap fills were capped at the target. A bootstrap analysis was conducted with 10,000 complete calendar window means across four ETFs, not 332 independent trades. The confidence interval and per-symbol statistics were saved, matching the title of the observation. The verdict rejected live adoption when uncertainty, cost, and execution gates failed, and a positive point estimate alone was not enough for adoption. The sources of this information are https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.

This is a frozen specification note1344. The provider made a SIP adjustment from September 8, 2016, to September 8, 2026, with data present for each ETF in every 2685 NYSE session, but no data was supplied for 2010-2015. A holdout test was conducted, where only entries and exits within the partition were considered, and the holdout period from 2019 to 2025 had 83 complete calendar windows and 332 correlated ETF trades, excluding boundary-spanning trades. The test started with four independent cash sleeves, each with $250, and no leverage was used. The Vectorbt system was used to generate orders with a long-only strategy, with a fee of 1 basis point plus 15 basis points for slippage per side. The cash return was reconciled with the trade profit and loss accounting, considering the entry fee. The rule sensitivity was based on the last session's open, with a stop at 1 ADR and a target at 2 ADR, using a 20-session mean of high-low values. Other rules were based on the third session's close. Daily OHLC stops were managed to avoid ambiguity, with adverse gap fills and target gap capped targets. A bootstrap analysis was conducted with 10,000 complete calendar window means across four ETFs, not 332 independent trades. The results included confidence intervals and per-symbol statistics, saved to match the observation title. The verdict rejected live adoption when uncertainty, cost, and execution gates failed, as a positive point estimate alone was not enough for adoption. The sources used were https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf and https://runway.adityaperswal.workers.dev/data/stock-bars.