Notes

On September 11, around 00:02-04 UTC, a direct issuer response with a 403 status, which means forbidden access, was received, and the web page was readable. The latest acquisition, 25-00060-UT, was updated on August 18. According to the original PDF document, page 3, an operative clause allows staff and parties to respond by September 9, 2026, but this is not an approval, and there is a typo in footnote 2027 that is separate. The regulator sent 20042301 bytes of data, but the JavaScript shell did not provide readable case documents, and the web page had 0 lines. No actual response or accepted protection has been verified, and the absence of readable evidence does not mean that there was no filing. The full PDF document, with 2002589549 bytes and a SHA hash of bfccac357506321b2fb05f8a0394e1fbfcb246456e516b277f8e982aa64c8178, remains unchanged from the previous version, 22965. The access scope is bounded, and a timestamp clarification observation was made at 00:03 on paper, which was completed and not scheduled for the future at 00:05. https://www.txnmenergy.com/investors/rates-and-filings/pnm-nmprc-filings.aspx ; https://e360.prc.nm.gov/portal/public/

A news scan found a release on September 9 about the fiscal first quarter that ended on August 1. The financial report shows that the company's cash flow from operations was $13.496 million, and the cash spent on property was $44.033 million, while the cash spent on software was $5.417 million, resulting in a total of -$35.954 million. If we exclude the software expenses, the free cash flow, which is the amount of cash a company has after paying for its operating expenses and capital expenditures, was -$30.537 million. The company's sales from its Aerospace Systems segment were $345.969 million, and its earnings before interest, taxes, depreciation, and amortization, or EBITDA, a measure of a company's overall financial performance, were $62.285 million. The sales from its Space and Cyber Defense Electronics segment were $134.521 million, with a loss of $8.896 million. The company's cash on hand was $278.390 million, compared to $377.325 million on April 30. The inventory was $410.773 million, compared to $312.856 million, and the unbilled revenue, which is the amount of money the company has earned but not yet received, was $637.832 million, compared to $570.408 million. The company's backlog, which is the amount of orders it has received but not yet fulfilled, is approximately $1.5 billion, and it expects to recognize 78% of its fiscal year 2027 revenue in the first quarter, according to its management's timing expectations, but this is not the same as cash. The company is the primary issuer of this information, and there is no new announcement, as the file remains unchanged. https://investor.avinc.com/node/21391 ; https://investor.avinc.com/node/21396/html

A news scan found the research paper by Faber, updated in February 2013, on the author's website on September 11 at 00:03-05 UTC, pages 21-22. The paper says to buy when the last day of the month closes above the 10-month Simple Moving Average, and to sell when it closes below. The paper uses total return series and cash equivalent to a 90-day T-bill, and does not include taxes, commissions, or slippage, which is a difference in price between expected and actual trade execution. This research is associated with the author and is based on historical data from 2006, 2009, and 2013, but it is not independently confirmed for live trading in 2026. Our own method, which was planned in advance, uses a 200-day daily cross with a stop loss of 2 Average True Ranges and a target of 6 Average True Ranges, with execution at the next open and including costs, is different and needs to be tested separately. We cannot use the return claims from the paper. https://mebfaber.com/wp-content/uploads/2016/05/SSRN-id962461.pdf

On September 11 at 00:01 UTC, a new ledger showed 3 closed lots that were liquidated by the operator and no open, pending, or closing lots. These exits affected the planned strategy and outcomes for stop and target prices. The quant cash was $187.22 and the account balance was $850.99 per run context. Job 195 completed 3 panels and ended at 21:15:23, 22:30:21, and 23:45:20 on September 10, which were approximately 75-minute intervals, despite being configured for 60 minutes. There were no remaining jobs to be done. Job 85, which ran at 22:30:27, found 74 news items, 2 pages, 5 matches, and 1 omitted item, with secondary discovery only, and the next run is scheduled for 02:30. Job 15 is scheduled to run next at 02:45, with 54 jobs left. The latest country, number 19, had 0 eligible items as of September 9. There were no holdings to sell or exits to move, and any watch-thesis attacks were recorded separately, a term which refers to a strategy of monitoring and reacting to market changes. query_trades(fund=quant); list_jobs(fund=quant); query_research(fund=quant)

On September 10, at 21:03:01 local time, a script returned a successful result, indicated by the code HTTP 200, and ended with a true status. The script checked link hashes for AVAV, FTK, and LOVE, which were found to be unchanged since 20 UTC, although it did not verify the full page. The successor, number 195, was accepted at 21:04:12 and is set to run every 60 minutes for 3 runs with a 60-minute timeout. The initial next run is scheduled for 21:04:12, but the actual remote execution has not been verified yet. The predecessor script, number 189, completed at 20 UTC, retaining 85 percent of its data and discarding 15 percent. This was for research purposes only and did not result in any orders being placed. list_jobs(fund=quant); local execution of job189 script; schedule_job(fund=quant)

A news scan found the following data line: 21:01–03UTC ledger3closed operatorliquidations,0open/pending/closing; intendedexits censored. This means that at 21:01 to 03 UTC, the ledger collector closed and there were no open, pending, or closing operator liquidations, but the intended exits were not shown. Job189 was done with 0 remaining, and it actually ended at 20:00:19, which is 29 minutes and 59 seconds after 19:30. The prior 45-minute interval was not erased. There are 85 next at 22:15 and 15 left. On September 11 at 02:45, there were 54 left, and all 19 eligible were at the top, including EWY, EWT, and EWP, which are types of exchange-traded funds. The results were: a return of -26.212876%, a drawdown of 27.351241%, a win of 43.979544398%, a win percentage of 61.7581664%, and a loss percentage of 68.2768517%. The Python mean was -0.1108804318176%, and the halfKelly, which is a betting strategy, was 0. The fund had 187.22 in the account and 850.99 in position, with a cap of 212.7475, which is a normal 3% risk of 25.5297. No new backtest was done. query_trades(fund=quant); list_jobs(fund=quant); query_research(fund=quant)

A news scan found the headline 'TXNM QTR9102100 response evidence remains unknown'. On September 10 at 21:02 UTC, a website became readable after a direct 403 error, which is a type of error that means access to a website is forbidden. This was related to an acquisition with the code 25-00060-UT, which was last updated on August 18 and had a declaratory case separated on August 27. The original order from August 18, on page 3A, verified that staff and party responses were due by September 9, but this did not mean approval was given. A website at https://e360.prc.nm.gov/portal/public/ had zero readable lines. It is unknown what the actual responses were, and it is not because of non-filing or acceptance. A typo in the original footnote from 2027 remains different from the operative text from 2026. The issuer is a mirror of the primary procedure, but the regulator's actual responses were not retrieved. https://www.txnmenergy.com/investors/rates-and-filings/pnm-nmprc-filings.aspx

A news scan found the headline 'FTK QTR9102100 full issuer documents unchanged'. On September 10, at 21:01:38 UTC, a file cache was checked. The file's unique code, or SHA256, was bfccac357506321b2fb05f8a0394e1fbfcb246456e516b277f8e982aa64c8178, which matched a known code, 22898. This means there was no new information about cash in this file, and it did not confirm that cash was completely absent. The company AVAV, which can be found at https://investor.avinc.com/node/21391, also had a file of 299723 bytes with a unique code, or SHA, of 87a4f3355ebdf8d90bad4fe1ea75f3aee74e07575c8107e685ed7f75a52d7647, which matched a known code, 22885. This file was released on September 9 and its information about the company's identity was unchanged. The issuer of the information was the primary source, but there was no independent confirmation of the cash information. https://ir.flotekind.com/image/Lake-Street-Investor-Presentation.pdf

A news scan found the headline 'AVAV QTR9102100 frozen closing EAR positive'. Retrieved at 21:02 UTC on September 10, the data showed a SIP adjustment for all 28 daily bars with no next page. On September 4, the prices were AVAV 144.65 and SPY 770.19, and on September 10, the prices were AVAV 147.065 and SPY 757.83. The returns were +1.66954718% for AVAV and -1.60479882% for SPY, with a difference of +3.274346 percentage points. September 7 was a holiday, and the dates September 4, 8, 9, and 10 were used to verify the t-2 and t+1 for the September 9 release. The prior 20-day mean of the difference between the high and low prices for AVAV through September 10 was 8.759085 USD, based on dates from August 13 to September 10, while for SPY it was 4.36456. The vendor's daily aggregates were used at the cutoff and are subject to corrections. The first request at the end of September 11 returned 403 recent SIP data, but after fixing the issue at 20:30, it returned 200 with the same feed and no IEX substitution. A frozen note, number 1683, had a positive branch, which means a hypothetical open on September 11 was considered only after the actual print, with a planned five-session exit at the close on September 17, but no actual order was placed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=AVAV,SPY&timeframe=1Day&start=2026-08-01&end=2026-09-10T20:30:00Z&feed=sip&adjustment=all

A review of the TXNM QTR9101940 response remains limited to those with access. On August 18, the acquisition 25-00060-UT issuer mirror was updated, and on August 27, a declaratory order was issued in a separate case. The original order from August 18, page 3A, states that staff and all parties may respond by September 9, which is not a deadline for approval. A website, https://e360.prc.nm.gov/portal/public/, was checked but did not provide any readable information. Although responses were submitted on September 9, they have not been verified, and it is not clear if they were accepted or not. The main procedure is focused on the interested issuer mirror, but there is no direct evidence of a response from the regulator. A footnote on page 2 mentions August 17, 2027, which conflicts with the operative dates of 2026, so the signed order from August 18, 2026, and the response deadline of September 9, 2026, will be used, even though there appears to be a typographical error. https://www.txnmenergy.com/investors/rates-and-filings/pnm-nmprc-filings.aspx

The actual collector completion for QTR9101940 was finished at 19:30:19, which is 44 minutes and 58 seconds later than the scheduled 30 minutes. The selected link hashes for AVAV, FTK, and LOVE remained unchanged. There is one remaining scheduled task at 20:00:15, but the receipt has not been received yet. A previous task was completed at 16:45. The system processed 117 pages, found 3 matches, and did not omit any pages, with the next update scheduled for 22:15. The daily update is set for September 11 at 02:45, with 54 items left to process and all 19 items are eligible. The top 3 items are EWY, EWT, and EWP, with no monthly catch-up needed. list_jobs(fund=quant); query_research(fund=quant,kind=notes)

A news scan found the headline 'DirectGET19:41:13UTC200'. This is a report about a file check. The file's identity was verified using a SHA256 code, which is a type of digital fingerprint, and it matches the expected code, which is 'bfccac357506321b2fb05f8a0394e1fbfcb246456e516b277f8e982aa64c8178'. This code has 22874 characters. The check confirmed that the entire file is unchanged, but it did not compare the file to a selected link. The file does not contain any new information about cash, but that does not mean that no other information exists. The issuer of the report is interested in the primary forecast, which is a prediction of future events, but it is not an independent verification of paid invoices, which is a confirmation of payments made. https://ir.flotekind.com/image/Lake-Street-Investor-Presentation.pdf

At 19:41 UTC, a new ledger was created with 3 closed trades that were liquidated by the operator, and there were no open, pending, or closing trades. The saved holdout, which was not rerun, showed the following results: N2151 had a return of -26.212876%, a drawdown of 27.351241%, and a win rate of 43.979544398%. The monthly cluster mean confidence interval was between -0.165928% and -0.052452%. The stressed return was -56.878654%. Using Python, the probabilities were recalculated as p = 0.43979544398, W = 0.617581664%, and L = 0.682768517%, resulting in a mean of -0.11088043181763163% and a half Kelly of 0. The current account balance is $187.22, with a total account value of $850.99. The position cap is $212.7475, with a normal 3% risk of $25.5297. Note that operator liquidations affected the intended exits, making it impossible to determine the stop odds or slippage from the three trades. query_trades(fund=quant); analyses2269/2208/2209; https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the original earnings release of AVAV for the fiscal quarter ended August 1. The release was issued on September 9 at 16:10 EDT. Revenue was 480.5 million dollars, which is a 6% increase from 454.7 million dollars. Bookings were 0.7 billion dollars, with a book-to-bill ratio of 1.4 and funded backlog of 1.5 billion dollars, which is a 37% increase. Gross margin was 124.6 million dollars, compared to 95.1 million dollars, with an intangible purchase accounting expense of 18.5 million dollars, compared to 37.4 million dollars. The issuer's primary interest is noted, rather than independent cash verification, which is a method of verifying financial information. The full HTML release was 299723 bytes, with a SHA code of 87a4f3355ebdf8d90bad4fe1ea75f3aee74e07575c8107e685ed7f75a52d7647, but no prior comparison was made. The current close on September 10 was unobservable at 18:30, and no earnings per share or price was inferred from headlines. https://investor.avinc.com/node/21391

A news scan found the following data line: 189 actual 18:00:57 all 3 panels/END true; preceding 17:15:52/16:45:55 => 30 then 45 min interval versus 30 scheduled. This means that the system had 3 panels end at 18:00:57, with the previous panels ending at 17:15:52 and 16:45:55, resulting in a 30-minute interval, then a 45-minute interval, instead of the scheduled 30 minutes. Hashes for AVAV, FTK, and LOVE, which are stock tickers, remained unchanged, with the values being AVAV 701 acbd7ad08, FTK bbd0666457a, and LOVE ab7020afb5ec. The selected links also remained the same. There are 3 runs left, with the next one scheduled for 18:30:53. The system is expected to be completed by 19:30, but will be reviewed by 19:40, although the actual cadence is not guaranteed. A previous task was completed at 16:45:44, with 0 tasks left, and an old issue with a missing footer was resolved. At 18:15:38, the system had seen 117 pages, with 3 more pages to process, and 4 matches found, but 0 pages were omitted. The system also processed syndicated transcripts for CULP, DBI, and LOVE, but the new event signals for YB were not verified. The next update is scheduled for 22:15. There are 54 daily tasks left, with the next one scheduled for September 11 at 02:45. The top 3 eligible stocks, EWY, EWT, and EWP, had 0 monthly catch-up tasks. list_jobs(fund=quant); query_research(fund=quant,kind=notes)

A news scan found the headline 'TXNM QTR9101828 response deadline original reverified; actual filings unavailable'. At 18:29 UTC, the issuer mirror acquisition, numbered 25-00060-UT, was updated on August 18. The original order from August 18, page 3A, states that staff and all parties may respond by September 9, but there is no approval deadline. On August 27, a declaratory order was issued for a distinct case. The website https://e360.prc.nm.gov/portal/public/ has no readable lines, and the actual responses from September 9 remain unavailable, and are not assumed to be absent or accepting protections. The original regulator order on the issuer mirror is the primary evidence of the procedure. https://www.txnmenergy.com/~/media/Files/P/PNM-Resources/rates-and-filings/2025%20Blackstone%20Infrastructure/PNM/Order%20Granting%20Joint%20Movants%20Motion%20for%20Extension.pdf

A news scan found the headline of a full investor PDF identity for FTK QTR9101828. The direct issuer is GET at 18:29:51 UTC on 200. The SHA256, a type of digital fingerprint, is bfccac357506321b2fb05f8a0394e1fbfcb246456e516b277f8e982aa64c8178, which matches 22846. This means the entire file, not just a list of its contents, has been verified. The file does not include any new information about cash payments. The issuer's forecast is considered important evidence, but it has not been officially checked and is not a record of actual cash paid. https://ir.flotekind.com/image/Lake-Street-Investor-Presentation.pdf

At 18:29 UTC, a fresh ledger was created, showing 3 closed trades that were liquidated by the operator and 0 trades that are open, pending, or closing. Using Python, the saved values of p = 0.43979544398, W = 0.617581664%, and L = 0.682768517% were used to calculate a mean of -0.1108804318176%. The half-Kelly formula, which is a method to determine the optimal bet size, was also calculated as max(0, 0.5 * (p - (1-p) / (W/L))) = 0. No new backtest was performed. The user's context includes a fund value of 187.22 and an account value of 850.99. The current mandate is to take a normal risk of 3%, with a convex risk of 5% and a position risk of 25%. The specific values for these risks are 25.5297, 42.5495, and 212.7475, respectively. A prior mandate of 1% with a value of 8.5099 is no longer the current venue mandate. There has been no change in sizing since the last update, which is referred to as edge 0. query_trades(fund=quant); saved analyses2208/2209/2255; https://runway.adityaperswal.workers.dev/data/stock-bars

The Freshledger3 is closed with 0 open or pending items. The Python exact returns are as follows: SNOW had a return of -5.39236448%, DELL had a return of 7.93994131%, and GTLB had a return of -5.13049905%, with a mean return of -0.86097407%. The operator liquidations do not identify planned stops, targets, or slippage, which refers to the difference between expected and actual trade prices. The saved gap is 0.43979544398, with a win percentage of 61.7581664% and a loss percentage of 68.2778517%, resulting in a mean of -0.1108804318%. The quarter Kelly, a measure of investment performance, is 0. The carried N, or number of trades, is 2151, with a return of -26.212876% and a drawdown, or maximum loss, of 27.351241%. The stress, or volatility, is -56.878654%, and this is not a new backtest. The current cash balance is 187.22, the account balance is 850.99, the position cap is 212.7475, and the risk cap is 8.5099. query_trades(fund=quant); https://runway.adityaperswal.workers.dev/data/stock-bars

A news scan found the headline 'Direct20016:32UTC SHA256bfccac357506321b2fb05f8a0394e1fbfcb246456e516b277f8e982aa64c8178 matches22834 exactly'. This is a data line indicating that a file, identified by a unique code called a SHA256, has been compared to a previous version and found to be identical. The file, named ftk.pdf, does not contain any new information. A previous forecast, made over 40 minutes ago, for the second half of the period, still stands and has not been affected by any new payments. The file has been stored for review. https://ir.flotekind.com/image/Lake-Street-Investor-Presentation.pdf

A review from 16:31 to 33 found that the latest order from August 18, known as acquisition 25-00060-UT, is still in effect. This order is separate from a declaratory order made on August 27 in a different case. The original order from August 18, page 3A, allowed staff and parties to respond by September 9, but this was not a deadline for approval. The regulator's website, https://www.prc.nm.gov/pnm_acquisition_case_information/, still has a summary that is not available, and the other website, https://e360.prc.nm.gov/portal/public/, has no readable information. An official search did not find the actual responses, and this is not because they were not filed or accepted. There will be no new attempt to reverse engineer the API after a prior audit failed. https://www.txnmenergy.com/investors/rates-and-filings/pnm-nmprc-filings.aspx

A news scan found the headline about the August Producer Price Index. At 16:31, the web reported that the final demand for August increased by 0.4% from the previous month and 5.4% from the same time last year, excluding food and energy, the increase was 0.3% from the previous month and 4.7% from the same time last year. This is less than the 0.1% and 4.8% reported at 13:52, which remained unchanged. The prices of energy increased by 4.2%, goods by 1.1%, services by 0.1%, and transportation and warehousing by 2.3%. This data comes from an agency producer survey, which does not measure surprise or direct deflator for July wholesale. https://www.bls.gov/news.release/ppi.nr0.htm

A news scan found the headline 'SPY QTR9101630 July wholesale exact table and vintage bridge'. The table shows that in July, sales were $801,278 million and inventory was $958,854 million. This revises the June numbers to $794,907 million in sales and $946,687 million in inventory, which replaces the previous rounded numbers of $794.1 billion in sales and $944.7 billion in inventory. It is noted that July growth should not be calculated using the old numbers. For electrical sales, there was a 0.1% month-over-month increase to $111,485 million and a 2.3% increase in inventory to $100,128 million, resulting in a ratio of 0.90. For petroleum sales, there was a 3.1% increase to $100,152 million and a 6.5% increase in inventory to $35,998 million, resulting in a ratio of 0.36. These numbers do not include issuer cash or delivered scarcity. Using Python, the ratio of inventory to sales was calculated to be 1.1966804, and the difference in growth between 1.3 and 0.8 is 0.5 percentage points, but no test of significance was performed on this difference because the covariance is unknown. https://www.census.gov/wholesale/pdf/mwts/currentwhl.pdf

A news scan found the headline with the following information: On September 10, at 16:31-33 UTC, the July wholesale data was released. The nominal sales for July were 801.3 billion dollars, which is an increase of 0.8 percent from the previous month, with a 90 percent margin of error of plus or minus 0.4 percent, and an increase of 13.0 percent from the same month last year, with a margin of error of plus or minus 0.5 percent. The inventories for July were 958.9 billion dollars, which is an increase of 1.3 percent from the previous month, with a margin of error of plus or minus 0.2 percent, and an increase of 5.7 percent from the same month last year, with a margin of error of plus or minus 1.2 percent. The ratio of inventories to sales was 1.20, compared to 1.28 in July 2025. The change in sales for June was revised from -3.0 to -2.9 percent. This survey does not include sales from manufacturers' sales offices, branches, and agents or brokers, and the numbers are not adjusted for price changes. Both reports come from the same original survey, and do not provide independent confirmation of the data. https://www.census.gov/wholesale/current/index.html