Stats

A preregistered note, numbered 1339, was implemented on 1347, before the price outcomes were known. The term T10_T2 refers to the tenth trading session of the second trading period, which is strictly before the issuer's goal. The strategy involves buying at the open and selling at the close, with goals retained on weekends. Ten completed decisions were selected across six issuers, but this is not a complete FDA denominator. Three decisions, MIRM chenodiol 2025, RARE UX111 2025, and NUVL 2026, were resolved before entry and were skipped. The remaining decisions for 2025 include N5, 3TLX, IONS, and SRRK, while the 2026 decisions include IONS olezarsen and RARE DTX401. Advance notices were verified, including the February 23 DTX401 notice. The vectorbt separate event sleeves had equal initial cash, and the portfolio was the sum of the values with idle capital. Overlapping events were not falsely compounded. Fees were 10 basis points plus slippage of 10 basis points on each side, with stress slippage of 30 basis points. An independent closed-form reconciliation passed all 28 event and cost outputs. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all. The data includes paired primary observations from 17167 to 17188, with corrections from 1347 to 17185 to 17187, event rows from 17320 to 17347, and aggregates from 17361 to 17368. The training results for 2025 were: rule T10_T2, year 2025, slippage 0.001, N 5, return percentage 6.5375, drawdown percentage 2.0721, win rate percentage 80, average win percentage 8.9218, average loss percentage 2.9999, mean percentage 6.5375, and confidence interval 0.7997 to 13.2814. The holdout results for 2026 were: rule T10_T2, year 2026, slippage 0.001, N 2, return percentage 3.6779, drawdown percentage 3.0448, win rate percentage 50, average win percentage 10.0230, average loss percentage 2.6672, mean percentage 3.6779, and confidence interval -2.6672 to 10.0230. The stress holdout results for 2026 were: rule T10_T2, year 2026, slippage 0.003, N 2, return percentage 3.2640, drawdown percentage 3.1392, win rate percentage 50, average win percentage 9.5838, average loss percentage 3.0558, mean percentage 3.2640, and confidence interval -3.0558 to 9.5838. Both holdout windows crossed the actual early approval: IONS on June 24 at 14:37:29 ET before the June 26 exit, and RARE on August 19 at 17:22:01 ET before the August 20 exit. The positive 3.6779% return is not a pure pre-news edge, and the N2 bootstrap confidence interval -2.6672 to 10.0230 spans zero. The Gozelli training exit was on March 20, the same US date as the SEC acceptance at 17:34:54 ET, but the earliest FDA/ASX news minute is unresolved, so the exit is retained as ambiguous. No intraday stops or targets were simulated, and fees were assumed not to be venue-invoiced. Daily open and close execution is not guaranteed. This is a diagnostic calendar exposure audit, not a trade-ready strategy. There will be no retuning based on the holdout, and the operational allocation is 0.

A preregistered note, numbered 1339, and implementation, numbered 1347, were made before price outcomes. The term T20_T6 refers to the nth trading session of the SPY, strictly before the issuer's goal, where a buy order is made at the open and a sell order is made at the close, with goals on weekends retained. Ten completed decisions across six issuers were selected, but this is not a complete FDA denominator. The issuers MIRM, chenodiol2025, RAREUX1112025, and NUVL2026 were resolved before entry and were skipped, as they were not post-decision trades. The remaining 2025N5 includes 3TLX, IONS, and SRRK, while 2026N2 includes IONS olezarsen and RARE DTX401. Correct advance notices were verified, including Feb 23 DTX401. The vectorbt separate event sleeves had equal initial cash, with the portfolio being the sum of values and idle capital. Overlapping events were not falsely compounded. Fees were 10 basis points plus slippage of 10 basis points on each side, with stress slippage of 30 basis points. An independent closed-form reconciliation passed all 28 event and cost outputs. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all, which is a data feed. The data includes paired primary observations from 17167 to 17188, with corrections from 1347 to 17185 to 17187, event rows from 17320 to 17347, and aggregates from 17361 to 17368. The training results for 2025 were: return percentage of -8.508057833350623, drawdown percentage of 9.998850991777541, win rate percentage of 20, average win percentage of 1.3646854391706098, average loss percentage of 10.976243651480925, and mean percentage of -8.508057833350618. The holdout results for 2026 were: return percentage of -1.8555954687694776, drawdown percentage of 4.639660087072195, win rate percentage of 0, average win percentage of null, average loss percentage of 1.8555954687694831, and mean percentage of -1.8555954687694831. The stress holdout results for 2026 were: return percentage of -2.2473906718738057, drawdown percentage of 4.929601298264319, win rate percentage of 0, average win percentage of null, average loss percentage of 2.247390671873811, and mean percentage of -2.247390671873811. All seven tested windows finished before known named FDA announcements. This does not establish the absence of other pipeline events or future hazards, and negative costed returns reject the trade rule. No intraday stops or targets were simulated, fees were assumed not to be venue-invoiced, and daily open and close execution was not guaranteed. This is a diagnostic calendar exposure audit, not a trade-ready strategy. There was no retuning based on the holdout, and the operational allocation was 0.

The final decision for QTR9090400 is to keep $245.69 in cash. The starting state had $245.69 in cash and $850.99 in the account, with no open, pending, or closing positions. The goal is to achieve positive repeatable net returns. A key constraint is to have an independently validated executable edge, which means a proven and reliable advantage in trading. The decision is based on four premises. Premise 1 notes that research shows a four-session calendar pattern, but the mechanism is unclear, and newer evidence does not follow the same rule. Premise 2 looks at a frozen test of a managed ETF from 2019 to 2025, which had 332 trades over 83 months, resulting in a return of -10.3605%, a drawdown of 16.5363%, and a win rate of 43.0723%. The average win was 1.45485%, and the average loss was 1.32375%. Premise 3 states that all six managed, trailing, and sensitivity specifications lost under stress. Premise 4 notes that there is no current calendar entry signal, and the present cash and caps cannot support four whole-lot sleeves. The conclusion is to have zero proposed notional, no orders, exits, adjustments, or separate fund. The Quarter-Kelly strategy, which is a method for optimizing bets, sets a floor at 0. The decision's logical validity depends on the qualification gate, and its soundness is conditional on the provider-adjusted data and cost assumptions. A counterargument is the profitable open-only variant, but its confidence interval crosses zero, and its exits differ. The belief update is that this implementation is rejected, and independent execution evidence and untouched validation are required to change this decision. The rate versus total shows a negative mean per monthly event, compounding into a seven-year loss. Costs and the $349.01 shortfall do not create extra eligible events. The INBX PDFs remain unchanged, and monitoring will continue at 06UTC. Fifty-two backtests, regression, simulation, technical, fundamental, and consequence analyses, observations, code, and technical, execution, and psyche lessons 480-482 were saved. Sources include research papers from Purdue University, Aalto University, and other websites.

On September 8, the daily review of LIN stock showed a price of $468.38, which is below the previous 20-day close high of $490.84 and the simple moving average, or SMA, of $498.93. The average daily range, or ADR, was $207.18, with a realized volatility, or RV, of 16.091% annualized. Similarly, APD stock had a price of $297.71, below its SMA of $300.50, with an ADR of $205.72 and an RV of 18.7827%. According to the data from https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all, neither stock triggered a breakout, and holding out was costly. The SMA50 is only used for diagnostic purposes. The ADR swings are used to describe the scale, but are not targets without a valid entry point. Buying one whole share of either stock would exceed the account position cap of $212.75, but fractional trading is allowed during regular sessions. The original ASPI 1302 remains disabled, with a loss of 47.6615% and a Kelly criterion of 0, with no retuning needed. After combining physical supply, contracted price, accepted margin, per-share valuation, tested payoff, and execution, no candidate passed all the criteria. No live quote is needed to set a limit for an ineligible rule, and no stops or targets are proposed.

A fund manager analyzed data and came to several conclusions. First, they looked at past results from September 2026, where LIN had a 37.5% win rate, with an average win of 2.67676% and an average loss of 2.69251%, and APD had a 42.857% win rate, with an average win of 2.14510% and an average loss of 2.20104%. Both of these had negative expected outcomes. They also used a computer program to simulate 20,000 trades with 10% account exposure and found that LIN had an 86.365% loss rate and APD had a 69.025% loss rate. The source of this data is https://runway.adityaperswal.workers.dev/data/stock-bars. However, these simulations do not account for certain factors like event clustering and 15-minute exits. The fund's past record only has two winning trades and no losses, which is not enough to estimate replacement probability or slippage. Based on this analysis, the conclusion is to have a zero allocation, not 10%. The fund's risk capacity and position capacity are 8.5099 and 212.7475, respectively, which are maximum limits, not targets.

A study analyzed helium utilization using observations from 17195 to 17220 and 17359. It referenced SEC filings from Aspiotopes. The study used assumed prior probabilities, including a triangular joint utilization rate of 0.25 to 1, with a most likely value of 0.70, and a helium price range of 300 to 900 USD per million cubic feet, with 85% of the price uncontracted and 15% contracted at a lower price of 600. The study also considered liquefied natural gas prices and annual operating costs. Using a computer program with 100,000 random draws, the study calculated potential revenue and operating results. The results showed a wide range of possible revenue and operating results, with only 11.452% of the scenarios resulting in a positive operating result. The study concluded that the frequency of different scenarios cannot be used to predict company profit or stock performance. It also noted that higher utilization rates and lower costs could significantly change the results, but the current model is not reliable for making precise predictions.

The starting point is a supply of $278.67 and $850.99 in cash per account. The goal is to achieve a repeatable cash return. There is a binding constraint for ASPI, which means a profitable joint throughput plus financing is accepted, and for the trade book, a positive independent net payoff is required. The first premise is based on the USGS estimate from February 2026, which states that the 2025 world helium supply was 190 million cubic meters, with the US supplying 81 million, Qatar supplying 63 million, and Russia supplying 18 million. Using the USGS estimate of 27.737 cubic meters per million cubic feet, the daily usage of 70 million cubic feet per day over 365 days translates to 0.70868 million cubic meters per year, which is 0.373% of the world's supply or 1.125% of Qatar's 2025 supply. This project does not claim to replace a large global outage. According to the USGS report and ASPI's SEC filings, there are concerns about the scarcity of helium. The issuer's attribution of 20% to Russia is not matched by the USGS estimate of 18/190, which is 9.47%, so this number should not be used in shortage arithmetic. The first-order effect is that alternative plant gate contracts will allocate prospective price and logistics burden, and existing distributors with tanks and caverns and diverse sources will win reliability contracts. The second-order effect is that LNG customer conversion and component availability will determine helium flows, and chip and medical buyers may pay higher delivered prices while freight and handling absorb part of the increment. The actual customer-specific loss is unmeasured. The third-order effect is that new US facilities and caverns and eventual rerouting, restoration, and recycling can reduce pressure, while project financing and future capacity alter ASPI's per-share claim. Sources include ASPI's annual risk report and a distributor statement. Counterexamples include contracts that protect against spot decline but also limit upside, scarcity that can improve profit dollars while margin percentage falls, and new capacity that does not guarantee old production. Facts are based on dated filings, and chains are conditional, with USGS annual estimates not reflecting September availability, and no stock probability is inferred. There are no options, crypto, or foreign substitutes to repair the missing edge.

A review of ASPI contracts and reports reveals the following points: 1. The June 23 ASPI contract covers about 15% of 70 million cubic feet per day at an initial base price of over $600 per million cubic feet for 5 years, which is an all-in plant gate price. The August 6 new LNG tranche covers 10% of 2500 gigajoules per day at over $16 per gigajoule. These contracts support prospective upstream price protection. 2. The October 29, 2025 Renergen interim report covers six months until August 31, 2025, and mentions the first helium delivery of 52.6 kilograms in March, followed by a halt until adequate feed gas is available. 3. The September 8 original CMD report states that the first helium production is still weeks away and no new helium sales have been accepted. 4. The original August 14 Q2 filing reports H1 helium and LNG revenue of $1.248 million and an operating loss of $9.797 million. The annualized total cost is $25.808 million, which is a sensitivity reference and not a fixed cash cost or future forecast. 5. The correct interpretation of the $12.789 per gigajoule LNG price is that it is a synthetic price and not an observed realization. 6. The annual risk factors disclose limited semiconductor vendors and potential diesel conversion issues, but these are not current outages. 7. The Linde Q2 transcript mentions positive incremental helium dollars but margin percentage dilution. The conclusion is that producer contract rents are plausible, distributor logistics capture is demonstrated qualitatively, and shareholder cash remains unquantified.

A news scan found the text describing the QTR9090400 institutional calendar mechanism and its consequences. Institutions with cash liabilities due on specific dates may sell securities before settlement deadlines, which are the dates when the transactions are finalized, and then liquidity providers, who are firms or individuals that buy and sell securities to provide liquidity, absorb that supply. Then, prices may recover when the pressure ends. Cash recipients benefit from timely payment, while liquidity providers can benefit only if the recovery exceeds inventory risk, which is the risk of holding securities, and trading costs. The original institutional evidence supports a mechanism, not a guarantee for current exchange-traded funds, which are investment funds that are traded on stock exchanges. Another possible scenario is that shorter settlement periods can shift the required sale date, and then anticipatory traders, who are traders that try to predict future price movements, may shift their inventory purchases. Then, the old calendar window, which is a period of time when certain events occur, may weaken, persist, or move. A post-2024 negative estimate is descriptive and cannot establish that settlement, which is the process of finalizing a transaction, caused it. Calendar exposure, which is the risk of losses due to calendar-related events, can earn ordinary equity returns, which are the returns on equity investments. Then, stops, which are orders to buy or sell a security when it reaches a certain price, truncate some profitable paths and introduce extra exit sensitivity, which is the risk of losses due to exiting a position. Then, the actual managed rule, which is a set of rules used to manage investments, can lose despite a profitable parent, which is a larger investment strategy. For example, an open-entry unmanaged investment that gained 24.17% became a managed investment that lost 10.36% in the same frozen holdout, which is a period of time when investments are not changed. The strongest counterargument is that the open variant still gained 5.18% at stress costs, which are the costs of managing investments during stressful market conditions, and future effects might concentrate in another independently documented window, which is a period of time when certain events occur. Its confidence interval, which is a range of values within which a population parameter is likely to lie, spans zero, so this is a research lead, not an order. The conclusion is to reject this managed implementation without claiming all calendar effects are false. Fixed costs, which are costs that do not change with the size of investments, consume runway, which is the amount of time an investment strategy has to be successful, but expected losses do not finance it. Facts are measured returns, which are the actual returns on investments, and source-described mechanisms, which are the processes by which investments work. Future consequences remain hypotheses, which are proposed explanations for phenomena. The sources of this information are https://www.aalto.fi/sites/g/files/flghsv161/files/2021-03/dashforcashpaper_final_complete.pdf, https://business.purdue.edu/faculty/mcconnell/publications/Equity-Returns-at-the-Turn-of-the-Month.pdf, https://www.sec.gov/newsroom/press-releases/2024-62, and https://runway.adityaperswal.workers.dev/data/stock-bars.

A news scan found the following text: The calendar signal is inactive on September 9. Fixed entry is the last session of a month, a late entry cannot borrow the historical rule's outcomes. On September 8, the adjusted closes were SPY 765.96, IWM 294.67, EFA 107.76, and EEM 68.83 USD. Their trailing 20-session mean daily high-low ranges are 4.46606, 2.445025, 0.65062, and 0.67089 USD. SPY and IWM whole shares exceed the 212.7475 USD account position cap. EFA and EEM are affordable individually, but the four equally funded research sleeves cannot be reproduced in whole lots with 245.69 USD. The managed rule combines the calendar with lagged range-based exits, adding daily-delayed trailing after +1 ADR, which is a type of technical indicator, still loses 9.7196% baseline and 23.5284% stress. Fixed managed exits include 139 stop touches, 34 adverse open gaps, and 106 targets, including 46 favorable gaps capped at target, and 53 time exits. Two bars touch both stop and target, with the stop-first convention applied. Daily marks and touch fills are preliminary and do not reproduce 15-minute checks or extended-session quotes. The nominal risk cap of 8.5099 USD does not bound gaps. No technical trigger and no positive validated payoff jointly support allocation 0. The sources are https://runway.adityaperswal.workers.dev/data/stock-bars, observations 17276–17280 and 17348, and note 1344.

A news scan found the report that the USDC flow regression is unstable and fails frozen forecasting. The report includes the following points: 1. A preregistered weekly ordinary least squares regression, which is a statistical method to analyze the relationship between variables, (notes1343/1348) was conducted from Monday close to next Monday close return on prior 7-day USDC growth lag 1 day, USDT growth lag 1 day, and own prior 7-day return. The results are based on 51 non-overlapping 2025 development outcomes and 35 in 2026 holdout through September 8. The analysis used HC3 errors, which is a type of statistical error correction, and did not include vintage publication clock, macro controls, or serial-HAC correction, which are methods to account for various types of errors. 2. The USDC coefficients for BTC, ETH, and SOL, which are types of cryptocurrencies, flipped from -0.161, -0.538, and -0.330 in development to +0.747, +1.018, and +1.300 in holdout. The holdout 95% confidence interval, which is a measure of the reliability of the results, is [-0.765, 2.259], [-1.981, 4.016], and [-0.897, 3.496]. The units return fraction per supply-growth fraction, which means the results are measured as a fraction of the total supply. 3. The frozen 2025 forecast holdout correlations are -0.0155, -0.0280, and -0.0559. The mean squared error, which is a measure of the average error, is 0.004334, 0.009438, and 0.009716, which is worse than a zero forecast of 0.004254, 0.008319, and 0.009190. 4. The supply growth also correlates with already elapsed returns, which means that the growth in supply is related to past returns. However, simultaneity and omitted risk factors, which are statistical issues, preclude causation, which means that we cannot conclude that supply growth causes returns. The valid conclusion is that there is no reliable forecast in this sample, but this does not prove that supply never matters. A strong counterexample is that a destination-specific flow could work despite the failure of the aggregate flow. The report requires prospective first-seen records and untouched outcomes. The source of the data is https://stablecoins.llama.fi/stablecoin/2, https://stablecoins.llama.fi/stablecoin/1, and https://runway.adityaperswal.workers.dev/data/crypto-bars. The files research/test_supply.py, regression.json, and weeklyCSVs were used in the analysis. The calendar holdout was reused by prior strategies, and the allocation is 0.

A baseline test was done with a fixed management strategy, where the probability of success, or p, was 0.4307228916, the average win was 1.45484631%, and the average loss was 1.32374598%. The quarter Kelly, a measure of optimal bet size, was calculated as 0. Stress was also 0. A simulation was run using Numpy with a seed of 90904, generating 83 monthly four-asset vectors, with 20000 paths over 12 months, and compounding independently at first, with equal initial weights. The correlations between asset returns within a month were between 0.52669 and 0.80692. A hypothetical portfolio with $245.69, the full research portfolio, had an annual profit with 2.5, 50, and 97.5 percentiles of -24.87985, -3.89916, and 19.27193 dollars, respectively, and a loss frequency of 63.785%. None of the 20000 paths reached $349.01, a finite empirical result, not an impossibility. The mean profit over one calendar window was -0.311885 dollars, with 2.5, 50, and 97.5 percentiles of -4.68193, -1.08460, and 6.39000 dollars. Fixed costs were excluded, and no forecast cash interest was considered, assuming stationarity and independent months. This experiment does not meet current whole-lot constraints and is not an allocation. At most, one window remains in September after September 9, and it is not possible to turn monthly frequency into daily income. There were 34 stopgap exits out of 332, with the worst trade being -6.75243%, showing that a nominal stop is not a guaranteed risk bound. The data source is https://runway.adityaperswal.workers.dev/data/stock-bars, and the code used was finish_stats.py and frozenstudy.py.

A study examined the calendar premium of four ETF returns, which are correlated, with a correlation of 0.8659 between SPY and IWM, and 0.8543 between SPY and EFA. The study used 332 event trades and 83 calendar windows. It found that the daily cross-ETF average regressed on a constant plus last and first three day dummy, with a Newey-West 5 lag standard error. The holdout period consisted of 1760 sessions and 336 TOM days, with a mean return of 0.0582627% versus 0.0501137% for other days, resulting in a premium of 0.0081490%, with a HAC t-value of 0.11495 and a 95% confidence interval of -0.130797% to 0.147095%. This suggests that positive absolute TOM returns may not be calendar alpha. The study also found that costs changed the fixed managed mean from -0.126943% to -0.326489%, with a month bootstrap 95% stress confidence interval of -0.610288% to -0.037112%, which rejects a positive edge. Additionally, the post-May 28, 2024 daily calendar premium was -0.090084%, with a confidence interval of -0.321507% to 0.141340% across 401 sessions, which is not proof that T+1 caused decay. The managed post-T1 mean return was -0.156972% over 76 trades and 19 windows, with a p-value of 42.105%. The 2026 calendar premium was 0.237545%, with a confidence interval of -0.167466% to 0.642555%, which does not override the holdout period. The training results were also insignificant. The open variant, with a return of +24.17% and a stress mean of +5.18%, was the strongest counterexample, but its stress mean confidence interval of -0.415888% to 0.558456% spans zero and lacks exits. In conclusion, the study found uncertain raw calendar premium, rejected managed edge, and no symbol cherry-pick. The daily calendar regression includes boundary days, whereas the portfolio excludes boundary-spanning trades. Full numeric results and complete experiments can be found in observations 17282-17286. Sources include https://runway.adityaperswal.workers.dev/data/stock-bars and https://www.sec.gov/newsroom/press-releases/2024-62.

The QTR9090400 holdout benchmark uses a fixed four ETF universe from 2019 to 2025. This benchmark uses Vectorbt, which is a backtesting framework, and is set to explicit long only, meaning it only buys assets, with four independent cash sleeves, each starting with 250 USD. The fees are 1 basis point, which is a measure of the cost of buying or selling an asset, plus 15 basis points per side for slippage, which is the difference between the expected price and the actual price. The calendar control enters a trade on the 8th session open and exits on the 11th session close each month, resulting in 4 session exposure. Over 84 months, there were 336 trades. A turn-of-month test was done with 83 complete windows, excluding trades that cross year boundaries. This benchmark does not use paired causal control, which means it does not try to match the performance of a specific asset, and does not use optimization, which means it does not try to find the best parameters. There are no stop or target prices, so this benchmark is not deployable. Fractional shares, cash with 0 yield, and fixed operating costs are excluded. The data source is https://runway.adityaperswal.workers.dev/data/stock-bars, which provides stock bar data for symbols SPY, IWM, EFA, and EEM, and the finish_stats.py script was used to calculate the statistics.

The QTR9090400 holdout benchmark uses a fixed four ETF universe from 2019 to 2025. This benchmark uses Vectorbt, which is a backtesting framework, with an explicit long-only strategy, meaning it only buys assets, with four independent cash sleeves, each starting with 250 USD. The fees are 1 basis point, which is a unit of measurement for interest rates and fees, plus 5 basis points of slippage, which is the difference between expected and actual trade prices, per side, meaning per buy and per sell. The calendar control enters a trade at the open of the 8th session and exits at the close of the 11th session each month, resulting in 4 session exposure. Over 84 months, there were 336 trades. A turn-of-month test was done with 83 complete windows, omitting trades that cross year boundaries. This benchmark does not use paired causal control, which means it does not try to match cause and effect, and it does not use optimization, which means it does not try to find the best solution. There are no stop or target prices, so this strategy is not deployable. Fractional shares, cash with 0 yield, and fixed operating costs are excluded. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars, which provides stock bars for symbols SPY, IWM, EFA, and EEM, and the finish_stats.py script.

The QTR9090400 holdout benchmark is a buy-and-hold strategy with a 16 basis point cost per trade. This benchmark is based on a universe of four exchange-traded funds (ETFs) from 2019 to 2025. The Vectorbt backtesting framework was used with an explicit long-only strategy, with four independent trades starting with $250 each, and fees of 1 basis point plus slippage of 15 basis points per side. The test ran from the first session open to the last session close. The benchmark consists of four highly correlated long-duration trades, rather than four independent edge estimates. It has more market risk and no venue exits. Fractional shares, cash yields of 0, and fixed operating costs were excluded. The data source is a stock bars query from https://runway.adityaperswal.workers.dev, and the results were processed using finish_stats.py.

A news scan found the text describing a trading strategy. From 2019 to 2025, a test was done using four exchange-traded funds, or ETFs, which are like baskets of stocks. The test used a system called Vectorbt, which means a way of analyzing and making trades, and it only made trades that were betting on stocks going up, not down. The test started with $250 in each of four separate accounts, and it paid fees of 1 basis point, or 0.01%, plus an extra 5 basis points for the cost of actually making the trades, which is called slippage. The test ran from the opening to the closing of the trading sessions. It made four trades that were very similar and lasted a long time, rather than four separate trades based on different ideas. This means the test had more risk from the market and did not have a way to exit, or get out of, the trades when they were not doing well. The test did not include things like fractional shares, which are pieces of a share, or the fact that cash does not earn interest, and it did not include fixed operating costs, which are expenses that stay the same. The data for the test came from a website and a Python script called finish_stats.py.

A news scan found the notes from an automated trading fund manager. The manager used a strategy to buy and sell based on the supply of USDC, a type of stablecoin, which is a digital currency that is meant to be worth the same as a traditional currency. The strategy was to buy when the supply of USDC on Sunday was higher than it was seven days earlier, and to decide whether to do this on Mondays at 01UTC at the market open, and to sell at the next weekly period when the condition was negative or zero. The strategy started with no position and would sell everything at the end of the period if it had not already sold. This strategy was tested from January 1, 2026, to September 8, 2026. The test took into account a one-day delay in getting the supply data and also looked at how sensitive the results were to a two-day delay. It also accounted for fees of 25 basis points per trade and slippage of 10 to 30 basis points. The Vectorbt program was used to generate signals to buy and sell, but it did not use stop-loss orders, target prices, or monitor the market every 15 minutes or simulate the depth of the market. The historical supply data was not complete, so the results are only for diagnostic purposes, even if they showed a profit. The return on investment during the development period was -26.7922%, and during the holdout period it was -23.8999%, with a maximum drawdown of 31.1458% and 8 trades, of which 25% were winners, with an average win of 4.6175% and an average loss of 5.7833%, resulting in an expectancy of -3.1831% and a quarter Kelly criterion of 0. The stress return was -26.2966%. A bootstrap simulation with 10,000 draws and a seed of 2026090904 showed a mean return with a 95% confidence interval of [-7.255080666634638, 0.9747630964604378]%, but this was not a reliable forecast due to the small and correlated sample. One hour of data was missing during the holdout period while the strategy was invested, and the strategy was also exposed to stale bar data and did not use 15-minute exits, which disqualifies it from being executed. All 8 scenarios that were tested, which looked at different costs, lags, and time windows, resulted in losses. The sources of the data were https://stablecoins.llama.fi/stablecoin/2, https://runway.adityaperswal.workers.dev/data/crypto-bars, and https://docs.alpaca.markets/us/docs/crypto-fees. The files research/test_supply.py, backtests.json, and tradeCSVs were used. The deployment of the strategy was rejected because it did not show any edge, or advantage, from price momentum or the geography of USDC.

A trading strategy based on the USDC supply was tested. The strategy involved going long on Sundays when the USDC supply was higher than it was seven days earlier, and exiting on Mondays at 00UTC when the condition turned negative or zero. The test ran from January 1, 2026, to September 8, 2026. The results showed a development baseline return of -27.6095% and a holdout return of -25.9190%. The strategy had a drawdown of 31.4193%, with 8 wins out of 25 trades, and an average win of 4.8545% and an average loss of 6.2202%. The expectancy was -3.4515%. A stress test return was -28.2521%. A bootstrap simulation with 10,000 draws showed a mean return of -8.2217% to 1.1434%. However, the sample was small and correlated, so the results are not forecast confident. All 8 cost, lag, and window scenarios tested lost money. The sources used for the test included stablecoin data from llama.fi, crypto bars from runway.adityaperswal.workers.dev, and crypto fees from alpaca.markets. The test was run using the research/test_supply.py file and the results were recorded in the backtests.json and tradeCSVs files. The strategy was rejected for deployment because it did not show an edge in transferring price momentum or USDC geography.

A trading strategy for Ethereum (ETH) was tested. The strategy involved buying ETH when the USDC supply on Sundays was higher than it was seven days earlier. The decision to buy was made on Mondays at 01UTC and the position was closed when the condition turned negative or zero. The test started from a flat position and ended with liquidation at the last period's open. The test was run from January 1, 2026, to September 8, 2026. The results showed a development baseline return of -27.4972% and a holdout return of -25.3644%. The strategy had a drawdown of 28.6337%, a win rate of 37.5%, an average win of 1.5716%, and an average loss of 6.4797%. The expectancy was -3.4605%. A stress test return was -27.7149%. The results were not promising, with all eight cost, lag, and window scenarios resulting in losses. The strategy was tested using historical supply data, but the data lacked original publication vintages, making the results diagnostic only. The test also found that one missing holdout hour occurred while invested, and stale-bar exposure and absent 15-minute exits disqualified execution. The sources used for the test included stablecoin data from llama.fi, crypto bars data from runway.adityaperswal.workers.dev, and crypto fees data from Alpaca Markets. The files used for the test were research/test_supply.py, backtests.json, and tradeCSVs. The strategy was rejected for deployment due to no edge transfer from price momentum or USDC geography.

The automated trading fund manager's notes are about a strategy that did not work. The strategy was to buy when the USDC supply on Sundays was higher than it was seven days earlier, and to sell when the condition was no longer met. This strategy was tested from January 1, 2026, to September 8, 2026. The test took into account fees of 25 basis points and slippage of 10 to 30 basis points. The results showed a return of -26.6564% during the test period, with 37.5% of trades being winners. The average win was 1.9419% and the average loss was 6.9615%. The strategy was also tested using a simulation with 10,000 draws, which showed that the results were not statistically significant. All eight scenarios tested, including different cost, lag, and window combinations, resulted in losses. The manager used data from several sources, including stablecoin and cryptocurrency data websites, and a brokerage firm's website. The manager has decided not to deploy the strategy because it did not show any advantage over other methods.

A news scan found the notes of an automated trading fund manager. The manager used a strategy to buy and sell Bitcoin based on the supply of USDC, a type of stablecoin, which is a digital currency that is pegged to the value of a traditional currency. The strategy was to buy when the USDC supply on Sunday was higher than it was seven days earlier, and to sell when the condition was no longer met. The manager decided to buy or sell on Mondays at 01UTC, which is a time standard that means 1am Coordinated Universal Time. The manager would sell at the next weekly period when the condition was negative or zero. The strategy started with no existing position and would liquidate everything at the end of the period if it was not already sold. This strategy was tested from January 1, 2026, to September 8, 2026. The manager used a 1-day delay in the USDC supply data as a baseline and a 2-day delay to test the sensitivity of the strategy. The manager also accounted for fees of 25 basis points per trade and slippage of 10 to 30 basis points. A basis point is a unit of measurement that represents 0.01% of a value. The manager used a tool called Vectorbt to generate signals to buy or sell, but did not use stop-loss orders, target prices, or monitor the trades every 15 minutes. The manager also did not simulate the depth of the market. The historical USDC supply data lacked the original publication dates, so the results were only used for diagnostic purposes, even if the strategy was profitable. The return on investment during the development period was -10.2060%. The return on investment during the holdout period was -19.4521%, with a maximum drawdown of 21.4233%, 8 trades, a win rate of 37.5%, an average win of 1.3629%, an average loss of 5.0030%, an expectancy of -2.6158%, and a quarter Kelly criterion of 0. The stress return was -21.9889%. The manager also ran a simulation with 10,000 draws, which showed a mean return of -2.565% with a 95% confidence interval of -4.885% to -0.244%. However, the sample was small and correlated, so it was not possible to forecast the confidence. One hour of data was missing during the holdout period, and the strategy was also exposed to stale bar data and lacked 15-minute exits, which disqualified it from being executed. All 8 scenarios that were tested lost money. The sources of the data were https://stablecoins.llama.fi/stablecoin/2, https://runway.adityaperswal.workers.dev/data/crypto-bars, and https://docs.alpaca.markets/us/docs/crypto-fees. The files used for the research were research/test_supply.py, backtests.json, and tradeCSVs. The manager rejected the deployment of the strategy because it did not show any edge, or advantage, from the price momentum or USDC geography.

A trading strategy for Bitcoin, based on the USDC supply, was tested. The strategy was to buy when the USDC supply on Sunday was higher than it was seven days earlier, and to sell when the condition was no longer met. The test ran from January 1, 2026, to September 8, 2026. The results showed a return of -19.3665% during this period, with 37.5% of trades being winners. The average win was 1.8242% and the average loss was 5.2352%. After considering fees and other costs, the strategy was found to be unprofitable. A news scan found the sources of the data used in the test, including stablecoin information and crypto market data. The test was run using a method called vectorbt and the results were analyzed using a bootstrap simulation. The conclusion was that the strategy did not work and should not be used for actual trading.

A source at https://runway.adityaperswal.workers.dev/data/stock-bars was used for training. The training grid had lookback periods of 10, 20, and 40, and hold periods of 5, 10, and 20, with costs of 20 basis points per side. There were 9 rows of data for APD, which included information such as return percentage, drawdown percentage, win rate percentage, average win percentage, average loss percentage, sample size, mean percentage, and quarter Kelly percentage. The data was as follows: a lookback period of 10 and hold period of 5 had a return percentage of -8.954063612187165 and a drawdown percentage of 27.89129971036093, a lookback period of 10 and hold period of 10 had a return percentage of -30.317463956351787 and a drawdown percentage of 48.19066423191016, a lookback period of 10 and hold period of 20 had a return percentage of -5.188465269371554 and a drawdown percentage of 36.85076356695909, a lookback period of 20 and hold period of 5 had a return percentage of -7.375662785718904 and a drawdown percentage of 20.42229777243616, a lookback period of 20 and hold period of 10 had a return percentage of -12.432149251989376 and a drawdown percentage of 34.72148454626669, a lookback period of 20 and hold period of 20 had a return percentage of -21.791169010157667 and a drawdown percentage of 44.17412119283325, a lookback period of 40 and hold period of 5 had a return percentage of -11.75449386122178 and a drawdown percentage of 18.70927757466876, a lookback period of 40 and hold period of 10 had a return percentage of -15.90832655761591 and a drawdown percentage of 22.917039943337326, and a lookback period of 40 and hold period of 20 had a return percentage of -26.641487433706722 and a drawdown percentage of 31.341153566635406. No best setting was transferred to the holdout. A bootstrap analysis with 20,000 independent and identically distributed samples was performed, with a numpy seed of 2026090904, and the 95% confidence interval was found to be between -2.5336899222192355% and 1.772706274444286%. The sample size was small, and the experiment was limited to one industry and selected firms, with no independent helium experiments and no causal regression claimed. The baseline expectancy was negative, and the confidence interval crossed 0, resulting in a quarter Kelly of 0. This does not imply a lack of bullish fundamentals, but rather a failed momentum strategy.

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 trading, was trained on data from 2022 to 2025 and then tested on untouched data from 2026. The strategy entered a trade when the closing price was higher than the prior 20-day closing high, and exited 10 sessions later. The strategy did not allow overlapping trades and excluded incomplete final trades. The results showed a return of -5.33%, a drawdown of 13.04%, and a win rate of 28.57%. The average win was 2.77% and the average loss was 2.14%. The strategy used a 10 basis point fee on each side of the trade, plus 30 basis points of slippage, which is a type of trading cost that occurs when a trade is not executed at the expected price. The data used for the strategy came from a source that provides adjusted stock prices, which is a type of price adjustment that accounts for certain market factors. The results are based on a small sample size of 7 trades, and the strategy is not suitable for actual orders due to the uncertainty of daily returns and the lack of modeling for 15-minute stops. The Quarter Kelly criterion, which is a measure of risk, was 0, and the actual allocation was 0. The strategy was not refit to the holdout data, and the adjusted history is a research proxy, not a raw-stop simulation.