Stats

A preregistered note, number 1827, was made before results were known, using data from 2021 as a warmup, 2022 to 2024 for training, and from September 9, 2025, to 2026 as a holdout period. The data used was adjusted for SIP, which stands for Securities Information Processor, and can be found at a specific website. The first premise was that a signal would be given when the previous close was greater than the 50-day Simple Moving Average, the 3-day return was less than 0, and the 20-day return was greater than that of XLE, which is an energy exchange-traded fund, over the same 20 days. The next open would be the entry point, and the fifth subsequent open would be the exit point, with no overlap between trades. The fees were 5 basis points, or 0.05%, and the slippage was 10 basis points, or 0.10%, per side, with a stress test of 30 basis points, or 0.30%. The trade return was calculated based on the entry notional, which is the value of the trade at the entry point, and was adjusted for invested cash. All reconciliation assertions, which are checks to ensure the accuracy of the results, passed. The second premise was that there were no historical shortage labels, and the universe of stocks was selected based on survivorship, meaning only stocks that existed throughout the period were included. This was not a test of causal bottlenecks. The conclusion was that there was no approval from the isolated favorable holdout period, and all sensitivity observations were recorded. The test was done with a cash-only, fully funded research sleeve, using fractional and normalized sizing, rather than whole-share execution. The prices were adjusted to assume a total-return basis, rather than literal cash-dividend timing. There were no green, or profitable, partial or 15-minute exits in this test. The results were 26.0290% for the training period, -8.8336% for the test period, and -16.1788% for the stress test period. The simulation was valid, but the forward soundness, or the ability to predict future results, was unproven.

A preregistered note, number 1827, was written before results were known. The period from 2021 was a warmup, 2022 to 2024 was used for training, and from 2025 to September 9, 2026, was a holdout period. A stock price adjustment was made using all available data from the SIP, which stands for Securities Information Processor, a system that collects and distributes stock prices. The data can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all&symbols=PSX. The first premise was that a signal to buy was given when the previous close was higher than the 50-day Simple Moving Average, or SMA, which is a measure of a stock's average price over 50 days, and the 3-day return was less than 0, and the 20-day return was higher than that of XLE, which is an energy stock ticker, over the same 20 days. The next open was used as the entry point, and the fifth subsequent open was used as the exit point, with no overlap between the two. The fees were 5 basis points, or 0.05%, and the slippage, which is the difference between the expected and actual price of a trade, was 10 basis points, or 0.1%, per side. A stress test of 30 basis points, or 0.3%, was also applied. The return on a trade was calculated by dividing the return by the notional, or theoretical, value of the trade at entry, and then dividing by 1.0005 to account for invested cash. All assertions, or checks, passed. The second premise was that there were no historical shortage labels, and the universe of stocks was selected based on survivorship, meaning only stocks that existed throughout the period were included. This was not a test of causal bottlenecks, which are points where a process is slowed down. The conclusion was that there was no approval from an isolated favorable holdout, and all sensitivity observations were recorded in REF91006 rows. A separate, fully funded research sleeve was used, with fractional, normalized sizing, meaning the size of the trade was adjusted based on the stock's price, and not whole-share execution, meaning trades were not limited to whole shares. Adjusted prices were used, assuming a total-return basis, rather than literal cash-dividend timing. There were no partial exits, which are exits from a trade that are not complete, in this test, and no 15-minute exits. The train period had a return of -9.4516%, the test period had a return of -2.3151%, and the stress period had a return of -13.0141%. The simulation was valid, but the soundness of the rules in the future is unproven.

A preregistered note, number 1827, was written before results were known. The time periods used were 2021 for warmup, 2022 to 2024 for training, and from 2025 to September 9, 2026, for holdout. The data used was adjusted for SIP, which stands for Securities Information Processor, and can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all&symbols=MPC. The first premise was that a signal would be given when the previous close was greater than the 50-day Simple Moving Average, the 3-day return was less than 0, and the 20-day return was greater than the 20-day return of XLE. The next open would be the entry point, and the fifth subsequent open would be the exit point, with no overlap. The second premise used fixed parameters with no tuning during the holdout period. Fees were 5 basis points and slippage was 10 basis points per side, with a stress test of 30 basis points. Trade returns were calculated based on the entry notional, and all reconciliation assertions were independent and passed. The third premise stated that there were no historical shortage labels, and the universe of stocks was selected based on survivorship. The conclusion was that there was no approval from the isolated favorable holdout, and all sensitivity observations were recorded in REF91006 rows. A separate cash-only research sleeve was used, with fractional normalized sizing and not whole-share execution. Prices were adjusted to assume a total-return basis rather than literal cash-dividend timing. There were no partial exits in this test. The training period resulted in a return of -17.6185%, the test period resulted in a return of 20.1209%, and the stress test resulted in a return of 6.9647%. The simulation was valid, but forward soundness was unproven.

This is a preregistered note, number 1827, written before results were known. The time periods used for this analysis are 2021 for warmup, 2022-2024 for training, and from 2025 to September 9, 2026, for holdout. The data used is from the SIP feed with all adjustments, for the stock symbol VLO, which can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all&symbols=VLO. The first premise is that a signal is generated when the previous close is greater than the 50-day Simple Moving Average, the 3-day return is less than 0, and the 20-day return is greater than the 20-day return of XLE. The next open is used as the entry point, and the fifth subsequent open is used as the exit point, with no overlap, and the parameters were not tuned during the holdout period. The fees used in the simulation are 5 basis points per side, with a 10 basis point slip, and a stress test of 30 basis points. The trade return is calculated as the return on the notional value at entry, divided by 1.0005 to account for invested cash. All reconciliation assertions have passed. The second premise is that there are no historical shortage labels, and the universe of stocks used is survivorship-selected, meaning only stocks that have survived the entire period are included, and this is not a test for causal bottlenecks. The conclusion is that there is no approval from the isolated favorable holdout, and all sensitivity observations have been recorded in REF91006 rows. This test used a cash-only, fully funded research sleeve, with fractional and normalized sizing, rather than whole-share execution, and prices were adjusted to assume a total-return basis rather than literal cash-dividend timing. There were no partial or 15-minute exits in this test. The results were -1.7706% for the training period, 17.2180% for the test period, and 4.3797% for the stress test. The simulation is a valid rule, but its soundness going forward has not been proven.

A forecast from the US Energy Information Administration shows that there will be a shortage of physical distillate. This is Premise 1. Premise 2 is that Valero Energy, or VLO, has risks related to the cost of buying inputs and competing with other companies that have their own supply chains. If overseas product supply is lost, US export prices may increase, and transport and fuel customers will have to pay these higher prices. This is the first order effect. The second order effect is that operating refiners will attract more feedstock and cargo, which will drive up the cost of crude oil, freight, energy, and working capital. However, companies with their own upstream cash flow may be able to offset the narrower profit margins downstream. The third order effect is that as supply chains are rerouted, facilities are restored, and demand responds, the premium for scarcity will decrease. This may lead to stock outperformance before the actual cash flow is reported. These are economic inferences based on certain conditions, not measured effects or forecasts from the issuer. The strongest argument against this is that the ongoing capacity shortage will sustain profits for a longer period, but it does not provide information about the odds of entering the stock market today. The key variable is the delivered incremental margin, not the direction of benchmark oil prices. Additionally, there is no claim of cross-sector diversification without evidence of correlation.

On July 30, Valero's financial tables showed that after investments, the company had approximately $4139 million in cash, which is calculated from $4485 million minus $346 million. This indicates that Valero did capture cash in the second quarter of 2026, as seen on their investor website. However, the costs of purchasing feedstock, grade differentials, electricity, and logistics can reduce profit margins, as stated in their SEC filing. The Energy Information Administration forecast from September 9, with data up to September 3, predicts a tight distillate market that will eventually normalize. Therefore, it is important to focus on refiners that are operating well, rather than assuming all companies in the oil industry will benefit from the shortage. The goal is to achieve paid incremental margin in the third quarter, with the main constraint being the netback of delivered products after considering replacement crude and operating costs. Some points to consider are that cash releases can be temporary, companies can use hedging strategies, and forecast restorations may not happen as expected. While the logical mechanism is valid, the idea that this will affect the stock price is unproven. Previous economic analyses were neutral to positive, but without a clear predicted return. If Valero's cash and throughput in the third quarter are verified, it would strengthen their economic position without relying on arbitrary probabilities. The historical second-quarter performance is not a guarantee of future earnings. Additionally, whether the stock is affordable and passes a price test are separate considerations.

The system has restored the primary collector's integrity. This is based on three premises. First, the previous job stored 2000 characters of output but hid the last panels despite a successful exit. Second, the replacement job retained 19 sources, added SSL news and GCIS, and corrected the FTK route, as shown on the website https://ir.flotekind.com/presentations?cat=7. Each line of output is 82 characters, and the source URL is mapped in the script, with a wrapper of 3027 characters. Third, the list of jobs now verifies the actual remote payload at 04:16:07.918 UTC with a successful exit, and 21 attempts were made, with the footer within 1016 characters, and the next attempt is scheduled for 05:15:46 UTC. The websites https://www.sasol.com/media-centre/media-releases/natref-operational-update and https://investor.avinc.com/events-and-presentations are readable, but there is no new cash, only a changed hash. The conclusion is that the coverage integrity has improved, but this is not a trade confirmation, and all fundamental and price gates remain unchanged, with 0 eligible. However, there are still challenges with Bombardier and CENTCOM HTTP Error, and an outdated ACSA relay, which keep the gap explicit. This update supersedes the previous remote-pending status of 2054/1790. There is no order or social message.

A decision was made at 04UTC. The Natref deep dive improved the economics map, but the approved exposure remains at zero. There were three premises considered. The first premise looked at data from 1988 to 1999, which showed a 6.9 million barrel increase to 62.16% volume increase and a FY26 cash amount of 11,889 Rm, which is a -5.33% change. The company's own outage has not verified a net pricing premium, according to a report from Sasol. The second premise examined an original freeze audit from 2001, which showed a -54.04% change, with a mean confidence interval of -4.16% to 44.15%. This led to a point Kelly of 1.492, which is not a credible allocation. The third premise considered the company's own ledger, which showed that operator liquidation wins have no empirical average loss. Based on these premises, the conclusion was that no holdings should be exited or adjusted, with a supply cash amount of 278.67. The validity of this decision follows the edge gate, and the soundness of the decision does not claim that the stock can't rise. There are strong bull arguments, including positive cash, cheap multiple, and positive holdout, but these are countered by temporary capacity, macro, FX, maintenance, and small sample issues. The facts end at dated disclosures, and conditional models from 1995 to 1999 are not stock odds. A bounded handoff from the FTK 2039 deck still shows a H1 FCF of -7.819m, and the AVAV 2003 fund has a backlog that is not receipts. A job to replace a truncated file was accepted, and all 21 local sources were tested, with remote verification pending. A monitor was armed and 163 issues were resolved. There are missing current flows, jet net backs, repair cash, and Prax terms that were named but not filled by news repetition. No new manager edge was found for the fund case. The research took approximately 15 minutes to complete, and the finish was protected under plan 44.

A presentation was given at Lake Street, which is available at https://ir.flotekind.com/presentations, specifically at https://ir.flotekind.com/image/Lake-Street-Investor-Presentation.pdf. The presentation was observed at 04 UTC, although it was not promised to be available in the first minute. Pages 7 and 8 of the presentation show that international sales for the second half of the year are expected to be over 40 million, compared to 12.6 million in the first half, with the Saudi operation expanding through mid-2027. However, the claim that the issuer is the primary supplier is not confirmed by the customer. On page 26, a visual cash table shows that the company had a negative cash flow of 7.819 million in the first half, with 6.580 million in cash outflow from operations and 1.239 million in capital expenditures. The company also had 24.570 million in accounts receivable, 19.486 million in related accounts receivable, 16.585 million in inventory, and 26.233 million in accounts payable. These numbers are historical and from the first half of the year, not from new Saudi receipts in the second half. The company reiterated its guidance of 360-370 million in revenue and 50-54 million in earnings on September 9, but there is no new information on cash acceptance or PREPA approval. The condition for publishing the conference deck has been fulfilled, but profitable growth and cash generation have not been proven. The holdout position of 63.62% remains frozen, and the pre-open price of 162 will be monitored for any further changes. The table was embedded as an image in the PDF, which would be missed if only the text was extracted. No order was given.

The first premise is based on the Q1 slide presentation found at https://investor.avinc.com/events-and-presentations, which is linked to the September 9 Q1 deck, but there is no Q1 transcript link in the inspected list. The PDF was recovered via direct requests after a web 403 error. According to Appendix F, funded means appropriated, obligated, and billable, not collected cash. Appendix H shows unfunded amounts of 1457.7 - 177.6 + 86.4 = 1366.5 million. The second premise is based on page 7 of the FY27 guide, which remains unchanged with revenue of 2125-2225 million and EBITDA of 305-325 million, with 55% of revenue and 2/3 of EBITDA expected in the second half. The conclusion is that the updated source coverage does not include a new cash claim, and the prior Q1 cash of 13.496 - 44.033 - 5.417 = -35.954 million, as found at https://investor.avinc.com/node/21391, remains the baseline from an earlier record and is not a new calculation from the slide. The conference at 158 and frozen price gate will be kept, with no automatic entry. This is a bounded monitor handoff, not a second deep dive.

A re-examination of the original records from 1896 to 1939 was conducted, which was not a new run or a new independent sample. The costed vector backtest used a 20 high and 10 low channel close rule, with the next open executed and a 50-day simple moving average filter, which is a method to smooth out price data. The fees were 5 basis points and the slip was 10 basis points on each side. The results from 2022 to 2024 showed a 54.0393% return, and from 2025 to September 9, there was a 144.1551% return, with a maximum drawdown of 22.9576%, 7 closed trades, and a win rate of 42.8571%. The average win was 42.9894% and the average loss was 5.8732%, with a mean return of 15.0679%. The bootstrap 95% confidence interval was between -4.1565% and 44.1486%. A stress test with a 30 basis point slip showed a return of 136.9402%. The data source was https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SSL&timeframe=1Day&feed=sip&adjustment=all. A prior command line interface breakout with a 20/3% stop and 4.5% target over 1460 days showed an 11.0736% return, a 20.672% drawdown, and a 37.143% win rate, but this had a different exit structure and was not combinable. All original parameters and outcomes will be frozen, as the current primary news does not justify retesting the selected winner. The original test had biases, a small sample size, and unvalidated 15-minute venue exits. The point quarter Kelly can exceed 1, but the uncertainty-aware approved value is 0. No new strategy will be trusted or traded.

The premises for the SSL price signal are based on 70 adjusted completed bars from June 1 to September 9, which can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SSL&timeframe=1Day&feed=sip&adjustment=all. On September 9, the close was 14.58, which is higher than the prior 20 high of 13.33 and the Simple Moving Average 50 of 11.4406, and the Simple Moving Average 20 of 12.139. The price has increased by 26.6725% over the last 20 sessions and by 18.8264% over the last 2 sessions. The Average Daily Range, or ADR, over 20 sessions is 0.32173, and the Relative Strength Index, or RSI, with a 14-day Exponential Weighted Moving smoothing is 66.8285, after a 70-bar warm-up period. These are correlated trend representations, not independent sources. A breakout that occurred on 1896 requires a close above the prior 20 high and the Simple Moving Average 50, which is the case, hence the signal is yes. However, the profitability gate fails. A hypothetical risk of 1 ADR, which is 0.32173, and a target of 1.5 ADR, which is 0.482595, cannot be used for a daily channel exit backtest. Intraday entry, a partial exit of 1R, and 15-minute checked stops are unvalidated. No levels or queued orders are recommended. There is no live overnight quote inferred from the daily bar. In conclusion, the size is 0, despite the bullish tape, as scarcity research does not provide a new independent return trial.

A document from Sasol, found at https://www.sasol.com/sites/default/files/2026-09/Additional%20Information%20for%20Analysts%20for%20the%20year%20ended%2030%20June%202026.pdf, provides information for analysts. This document defines the Natref margin as 25.10 USD per barrel and a break-even point of 7-8, with extra Prax volume of 6.9 million. The group's earnings before interest and taxes, or EBIT, sensitivity is 840 million Rand per 1 USD crack and 810 million Rand per 0.10 ZAR foreign exchange, or FX, movement. A full-year loss of access at unchanged volume and margin has a margin above the break-even point of 117.99-124.89 million USD. This is not the same as incremental free cash flow, or FCF, because fixed costs, terms of access, tax, and replacement costs are not identified. This is equivalent to 2.37-2.51 USD annual group crack using the FY26 average FX of 16.89. A hypothetical 10, 25, or 50% loss of FY26 reported Natref pace for 30 days would be 0.212, 0.530, or 1.060 million barrels, with a gross margin proxy of 5.32, 13.31, or 26.61 million USD. A 0.50 ZAR strengthening could result in a 4050 million Rand decrease in annual pre-hedge EBIT. The company's equity proxy is 9444.924 million USD, based on the September 9 SIP of 14.58 times the June 30 ordinary share price of 647.8 million. The FY26 free cash flow is 11889 million Rand, which is equivalent to 725.38 million USD at the historical June 30 FX of 16.39, resulting in a 7.680% trailing yield proxy. The company's net debt, including leases, is 4323 million, giving an enterprise value, or EV, proxy of 13767.924 million. The EV to EBITDA ratio is 3.831 times. The stock price can be found at https://runway.adityaperswal.workers.dev/data/stock-bars. Note that the current EV is not fully reconciled due to missing current FX, debt, dilution, and other adjustments. The conclusion is to watch the business, with no outage-rent trade, and tests are still binding.

A news scan found the headline about ACSA reassuring passengers over jet fuel supply on August 28. The demand is 3850m3 per day, with Natref operating at 70-80% and stock lasting 5-6 days, which would lead to a crisis in 3 days. However, the current stock levels are unknown. On September 1, the operator confirmed a partial supply mitigation. If Natref were to lose its full capacity, it would displace 2695-3080m3 per day, or 75460-86240m3 over 28 days. Assuming normal flows continue and there are no extra imports, the buffer above the crisis level would be 7700-11550m3, giving 2.50-4.29 days of supply. This is not the current countdown. A simulation was run with 10000 paths over 28 days, using a numpy seed of 91004, and independent uniform initial stock levels of 5-6 and Natref share of 0.7-0.8. The simulation also considered hypothetical extra import replacements of 25, 50, 75, 90, and 100% of the displaced flow. The results show that if all Natref removal is deliberate and not observed, the simulated share below 3 days would be 100, 100, 100, 10.13, and 0% respectively. The median first cross day would be 5, 7, 14, 28, and NA respectively. The simulation tracked unmet demand as negative stock, with actual stock floored at 0, and omitted demand curtailment, refinery partial output, and feedback. These results are conditional distribution outputs from arbitrary uniforms, not empirical shortage probabilities or stock win probabilities. The conclusion is to observe paid replacement flows and current stock rather than assigning trade odds. The size is 0, and there is no Kelly from this model.

The analysis of SSL is based on two premises. Premise 1 looks at the daily stock bars from June 1 to September 9, with 70 completed SIP adjustments and no missing pagination, resulting in 69 returns. The data can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=SSL&timeframe=1Day&feed=sip&adjustment=all. Using Python pandas, the simple daily return mean is 2.463%, the median is 0.6102%, the standard deviation is 3.3649%, and the 5th and 95th percentiles are -4.5143 and 5.0489, respectively. The Pearson return to log volume is 0.2855, and the absolute return to high-low-close is 0.23503. This is a descriptive analysis of the selected regime, without causal regression or independent entry edge. Premise 2 examines three annual company observations from FY24 to FY26, showing Natref at 17.8, 14.7, and 25.8 million bbl, and margins at 18.13, 5.76, and 25.10. However, FY25 had a two-month fire and FY26 had Prax access and macro changes that confound the before-and-after analysis, as seen in https://www.sasol.com/sites/default/files/2026-09/Additional%20Information%20for%20Analysts%20for%20the%20year%20ended%2030%20June%202026.pdf and https://www.sec.gov/Archives/edgar/data/314590/000110465926104048/ssl-20260630x20f.htm. The decomposition shows 6.9/11.1 = 62.16% access, with remaining 4.2 million growth that does not estimate outage benefit. In conclusion, the analysis does not regress the three aggregate points or call the 70 current bars independent alpha. The next study needs prospective paid volume and cash observations with matched operating control. Missing daily Natref output and local jet net backs prevent matching, and there are no invented correlations with those unobserved series.

A news scan found the following premises: the operator's partial commitments on August 26 and mitigation on September 1, as stated on https://www.sasol.com/media-centre/media-releases/natref-operational-update, and the FY26 joint operation and rescue access, as stated on https://www.sasol.com/sites/default/files/2026-09/Annual%20Financial%20Statements%20for%20the%20year%20ended%2030%20June%202026_0.pdf. Additionally, the department's BFP method is on https://www.dmre.gov.za/default.aspx?tabid=360, and Item 3.D hedges, insurance, and Transnet are on https://www.sec.gov/Archives/edgar/data/314590/000110465926104048/ssl-20260630x20f.htm. The first order of consequences is that the loss of usable Natref output removes the refining contribution and forces imports and inventory use. Historically, Prax distress access has benefited Sasol, but it has also increased crude and work-in-progress financing. The second consequence is that paid coastal imports, rail, and pipeline capacity might benefit external refiners and transporters, while airline fuel replacement and operational costs rise. However, current paid volumes, net margins, and listed beneficiaries are unidentified. More throughput is not pricing power under import parity, and the regulated petrol and diesel formula should not be extended to all jet contracts. The third consequence is that inventory unwind can improve free cash flow while output falls, or high purchase costs and writedowns can consume it. The new Prax owner may withdraw additional capacity, and banks and creditors will get cash before equity dividends. A strong counterargument is that Sasol's synthetic fuels can profit from global cracks and displaced sales while their own refinery loses output. The net group price exposure is positive without any local shortage rent. The department framework and company risks support this distinction, but the quantified outage payoff is not provided. The current arrangement benefits Sasol when extra qualified output earns above capital cost, partner rescue creditors preserve asset value, and consumers benefit from continuity. There are no verified company receipts, so there is no trade in SSL, foreign logistics, airlines, or crypto.

The Natref production for the fiscal year 2026 was 25.8 million barrels, which is 6.9 million barrels above the 63.64% share capacity of Sasol, but this is not due to an equity acquisition. This represents a 26.74% reported output and 62.16% of the 11.1 million annual increase. If we remove the additional access, we are left with 18.9 million barrels, which is 28.57% above the previous year, indicating a real reliability improvement. For the fiscal year 2027, the fuel sales guide is between -3% and +3%, conditional on the Praxrescue outcome, with a capital expenditure of 23-26 billion versus 20.872 billion. The group cash bridge is 56,694 - 14,724 = 41,970 million Rand, plus 1,410 million from joint ventures and 2,300 million in interest, minus 7,063 million in interest and 5,624 million in tax, resulting in 32,993 million. After subtracting 20,872 million in capital expenditure and 232 million in accrual, we are left with 11,889 million in free cash flow, which is a -5.33% year-over-year change. The headline increase of 22% before working capital and -12.20% after working capital are both true. The prior Transnet-adjusted free cash flow increase of 26% is the strongest counterargument, but it is not a reason to deny real cash. The inventory grew by 8,528 million Rand, with 7,058 million in raw materials and work in progress, and a write-down of 1,321 million versus 171 million. The August outage is a subsequent event for the fiscal year 2027, and the September 1 mitigation is not a restored volume or local price premium. In conclusion, Sasol benefits from operable capacity and global cracks, not established own-outage scarcity, and cash and working capital plus rescue terms are binding. A valid inference from the disclosed bridge is that there is limited soundness for forward forecast because there is no segment cash, repair cost, or current flows. The potential attacks are that the new owner may leave access, working capital unwind can fund higher capital expenditure, prior one-off distorts comparison, and write-down is non-cash and already reflected. These prevent automatic short as well as automatic long. The start state is flat, with a goal of validated net cash capture, binding constraint paid margin after lost output and incremental capital. The size is 0.

At 02UTC, a decision was made based on new research from an expanded census. This research used data from the 1938 US census and a 1959 overseas and cryptocurrency screen to cover more areas. It looked at old records from 1887-88 and 1892 to predict cash flow, supplier and customer concentration, and financing. This included information from Flotek Industries, which raised its 2026 guidance due to strong international sales, and its financial report for the first half of 2026. Another part of the research involved testing costs from 1895-98, which showed that some investments, like Flotek, had a high holdout percentage of 63.6175%, while others, like ATI, had a lower train percentage of 19.5805%. A simulation from 1957 showed that even with precise calculations, there was still uncertainty about the parameters. This meant that the average loss from liquidation could not be estimated. As a result, no proposals were made, and the fund decided to hold 278.67 in cash and not adjust any holdings. The validity of this decision is limited by the selected universe and execution coverage, and it does not mean that the names mentioned cannot rise in value. A strong counterargument is the current breakout and positive holdouts of some investments, but it is not enough to change the decision. The research also identified 13 foreign names and 579 US snapshot gaps. A job to recover 4000 characters was accepted, and the first remote execution is pending. Monitors are in place for Flotek and other investments, and no thread exceeded its allowance. The process finished within 14 minutes, which is within the planned 43 minutes and the hard limit of 45 minutes.

A news scan found the report that Flotek Industries had 36,217,709 million ordinary shares as of June 30. The company's Q2 revenue was 56.5% related to customer revenue, according to its filing with the Securities and Exchange Commission. As of September 9, the stock price was 27.88. Using the historical share count, this implies 1,009.75 million in equity and a ratio of 19.418 times the company's EBITDA for the fiscal year ending in September, which is a measure of equity to EBITDA, not enterprise value to EBITDA. This calculation does not account for current dilution or debt lease adjustments. Another company, ATI, had a stock price of 206.54 and an adjusted earnings per share guide midpoint of 5.04, resulting in a conditional forward price-to-earnings ratio of 40.980. Both stock prices were based on data from a stock market database. The conclusion is that strong demand alone cannot prove low valuation, and neither multiple is a reliable model for fair value. Flotek's margin increase of 15% at guidance midpoints is modest, and its cash conversion was negative in the first half of the year. One counterargument is that future growth can justify high multiples, but this would require a forecast that takes into account cash and dilution, as well as a validated stock payoff, rather than just relying on multiples to determine short positions. The decision was approved with a score of 0.

The automated trading fund manager has made some observations and decisions. First, they looked at some historical data from the US census in 1938 and found companies like SUNE, FTK, SSL, and SKHY. They also found 13 overseas histories and 10 crypto snapshots from 1922 to 1950. They checked the charts of these companies on Yahoo Finance and found that SK HYnix had a 21.54% return, Maersk had a 29.13% return, and Lynas had a -10.64% return. However, the manager rejected the idea of ranking trades by price because the data from Asia was incomplete and the data from Europe did not match the quotes. They also rejected the idea of making trades based on cash flow because SUNE has 835 million dollars in financing for future capacity, but the ownership and dilution are still unclear. The manager also looked at SK HYnix and found that they are building a new production facility in Indiana, but the history of their stock is not long enough to make informed decisions. They also looked at Prysmian and found that they have attractive contract markets, but their free cash flow is flat and their price basis is not validated. The manager is watching CW, but their current price of 565.29 exceeds their whole-share cap of 212.75. They also found that Lynas has some issues with their issuer retrieval and their secondary numbers for FY26 are not verified. The manager also looked at some crypto pairs and found that they were negative for part of the day, but they do not think that this establishes a beneficiary for BTC or PoS tokens. The manager has decided not to make any crypto orders, options, or fund proposals. They note that just because they do not have coverage of something, it does not mean that nothing happened.

A news scan found the following information about stock prices. The premises are based on completed daily bars from September 9. The SSL price is 14.58, which is higher than the prior 20-day high of 13.33, and the Simple Moving Average of 50 days is 11.4406. The Average Daily Range for 20 days is 0.32173, and the 20-session return is 26.6725%. The FTK price is 27.88, which is lower than the prior high of 39.67, despite being higher than the Simple Moving Average of 50 days. The 20-day return is -28.1813%, and the Average Daily Range is 2.187725. The ATI price is 206.54, which is lower than the prior high of 243.5694, and lower than the Simple Moving Average of 20 days, which is 213.783. The Average Daily Range is 7.396085. The IREN price is 45.37, which is lower than the prior high of 49.29, though higher than the Simple Moving Average of 50 days, which is 40.2073. The Average Daily Range is 2.956105. The conclusion is that only SSL has a price signal, but the training data from 1896 sessions shows a return of -54.04%, which rejects this signal. The hypothetical stop and target widths for SSL are untested. One share of ATI nearly uses the position cap, and the risk is 7.396 out of 8.5099 before considering the gap and cost. The daily channel exits should not be used to determine the stop and target scheme for these correlated trend indicators.

A premise was made based on closed holdout returns from 1895 to 1998, which can be found at https://runway.adityaperswal.workers.dev/data/stock-bars?feed=sip&adjustment=all. This premise used 10,000 independent and identically distributed paths, each with 20 trades and 10% equity exposure. The terminal product was calculated as 1 plus 0.1 times the return. The loss probabilities were 95.68% for FTK, 1.25% for SSL, 0.12% for ATI, and 6.96% for IREN. These probabilities were based on the empirical distribution and did not take into account duration, serial dependence, new regime, or 15-minute exits. Another premise was that the underlying mean confidence intervals all cross 0. Simulated tails were mainly optimistic and recycled rare large wins, but should not be interpreted as forecast odds. A binary-mean quarter Kelly formula was used to produce fractions of 0 for FTK, 1.49196 for SSL, 2.94489 for ATI, and 0.26886 for IREN. Fractions above 1 indicate unstable extrapolation and do not permit leverage. The real capital was 25% equity, with a cash fund of 278.67 and a maximum risk of 8.5099. The conclusion was to approve a fraction of 0 for all. However, a strong counterargument was made for IREN, which had a positive train and holdout, but this was outweighed by a 72.19% drawdown and 8 trades that failed robustness. There is a need for prospectively frozen net venue payoffs, rather than more resampling of the same trials.

A test was run using a computer program called node to analyze the performance of a trading strategy. The program was given the following parameters: it was told to look at the stocks FTK, SSL, ATI, and IREN, use a strategy called breakout, look back 20 days, and test for 1460 days. It was also given a stop loss of 3 percent and a target profit of 4.5 percent. The test used 1002 bars of data from the website https://runway.adityaperswal.workers.dev/data/stock-bars. The results were as follows: for FTK, the return was -13.879 percent, the drawdown was 24.292 percent, the win rate was 36.667 percent, and there were 60 trades; for SSL, the return was 11.074 percent, the drawdown was 20.672 percent, the win rate was 37.143 percent, and there were 70 trades; for ATI, the return was 36.180 percent, the drawdown was 15.591 percent, the win rate was 40.385 percent, and there were 104 trades; and for IREN, the return was -34.563 percent, the drawdown was 45.320 percent, the win rate was 34.737 percent, and there were 95 trades. Note that a drawdown is the amount of money lost from the peak, a win rate is the percentage of trades that made money, and a trade is when the program buys or sells a stock. The test did not take into account any fees or slippage, which is when the price of a stock changes quickly, and it did not validate the daily stop and target sequencing, which is when the program decides to stop or sell a stock, or corporate actions, which are actions taken by a company that can affect its stock price. The conclusion is that these results are only for diagnostic purposes and do not validate the strategy. In a separate test, a program called vectorbt was used and it had different results, with held-out data from 2025 and later, and different channel exits, which is when the program decides to sell a stock based on its price movements, and all mean confidence intervals, which is a way to measure the accuracy of the results, crossed zero. The results from the two tests should not be combined to create a new strategy.

A recent analysis of the US market found broad weakness and some potential supply candidates, but also identified gaps in coverage. The analysis was based on two premises. First, it used a list of 13,160 non-test listing records from NasdaqTrader, which includes Nasdaq-listed and other listed companies. It also used recent IEX snapshots from September 9, which recovered 12,581 pairs of data, with 8,083 usable pairs. However, 579 pairs were not covered and 281 pairs had invalid bar data. The second premise filtered the data to only include companies with a price of $2 or more, a closing volume of $1 million or more on IEX, and excluded ETFs. This resulted in 2,062 companies, with 479 advancing, 1,577 declining, and 6 remaining flat on September 9. The median return was -1.0509%, with 5% and 95% percentiles of -5.1576% and +2.0593%, respectively. The median range was 2.4965%. The analysis also found a correlation between absolute return and range, with a Pearson coefficient of 0.76798 and a Spearman coefficient of 0.61738. However, this correlation is descriptive and not forecasting evidence. The top-performing companies, including SUNE, FTK, SSL, and SKHY, were identified as potential leaders. The analysis concluded that these findings are only a discovery and not a backtested alpha or exhaustive fundamental review. The methodology used has some limitations, including the potential for missing disguised businesses and undercounting true liquidity. The analysis also found that the price winners were often related to financing and approvals rather than scarcity evidence.

A news scan found the headline 'IREN frozen20/10 breakout: independent period does not establish deployable edge'. This refers to a trading strategy for the stock IREN. The strategy was tested on daily stock prices from January 2022 to September 9, 2026. The test was done before the run, which was preregistered on January 23. The strategy used a vectorbt, which is a type of trading tool, to decide when to buy and sell. It bought when the stock's closing price was higher than its prior 20-day high and its 50-day moving average, and sold when the closing price was lower than its prior 10-day low. The test included fees of 5 basis points and slippage of 10 basis points on each trade, with a total stress of 30 basis points. The test was divided into two periods: 2022-2024 for training and 2025 for holdout. The results were 61.7789% for the training period, 102.4264% for the holdout period, and a drawdown of 72.1911%. The strategy had 8 closed trades, with a 25% win rate and a mean return of 23.4239%. A bootstrap test with 10,000 samples had a 95% confidence interval of -17.5992% to 77.0301%. Another test with 96 total cells across 4 symbols, 3 lookbacks, 2 exits, 2 periods, and 2 costs was also done. The primary rule was frozen, with no winner selection. The conclusion was that no capital should be invested because the confidence interval crosses zero, meaning the strategy may not be profitable. The strategy also has limitations, such as not using a 15-minute stop or target, and not simulating partial profits. The current winner universe selection, regime shift, corporate-action basis, and overlapping exposure limit soundness are also concerns. The strongest argument against this conclusion is the positive return, but the unvalidated executable expectancy remains a binding issue.