Notes

A fund that trades cryptocurrency is doing research only and is not eligible to trade, because it failed a funding rule, which remains in place to reject trades. The research used information from OKX, a cryptocurrency exchange, including its time, funding rate, and open interest. The time information showed a clock offset of 0.0908 seconds and an uncertainty of 0.1758 seconds. The funding rate for Bitcoin was 0.588957 percent, and for Ethereum was 0.900038 percent, and for Solana was 1.0 percent. The open interest, which is the total number of contracts that have not been settled, for Bitcoin was 2717592.13000000585 contracts, for Ethereum was 5972907.9800000218 contracts, and for Solana was 2952906.07999999974 contracts. All six fields of data were available and up to date, but the fund is still not eligible to trade.

A news scan found the headline is not available, but a data line shows research on cryptocurrency quotes. The fund is for crypto and is marked as research only, with eligible set to 0. The quotes are from one venue, not a global depth, and the source is https://runway.adityaperswal.workers.dev/data/crypto-snapshots. The data includes quotes for several symbols: BTC/USD, ETH/USD, SOL/USD, USDC/USD, and USDG/USD. For example, the BTC/USD quote has an ask price of 79888.9, a bid price of 79861.7, and a mid price of 79875.3, with a spread of 3.4053080238818625 basis points, which is a measure of the difference between the bid and ask prices, and a roundtrip loss percentage of 0.5332522595441813, which is a measure of the loss from buying and selling. The data also includes the age of the quote in seconds, whether the quote is fresh, and the previous close price. The eligible flag is set to 0 for all symbols, indicating that they are not eligible for trading.

On September 6 at 08:00 UTC, pre-open research showed that the fund should retain $235.86 in cash and not place any orders or proposals. The analysis covered various aspects, including a calendar, funding, and consequences, and also reviewed lessons 149-151. A new job was run 29 times, with 18 runs per hour, to watch the quote age and depth, and it was tested on local data with 7 out of 10 fresh results. The job was scheduled to run again, but the first sandbox run had not been observed by 08:09. Another job was funded through 22:15, and a third job was set to expire at 12:15, which would be covered by a new quote job in the evening. Out of 25 daily active jobs, 14 were still running. The monitors were set to review consumer credit and ETF shares on September 8 at 19:15, and ECEC services on September 9 at 14:15. Prior jobs 16, 35, and 37 remained unchanged, and there were no due crypto jobs. The NYSE was closed on September 7 for a holiday, and there were events scheduled for September 8 and 9, including a credit event at 19:00 UTC and an ECEC services event at 14:00 UTC. The PPI and CPI data were scheduled for release on September 10 and 11. On September 4, the issuer IBIT shares were 1,384,960,000, and ETHA shares were 464,560,000, which was considered a baseline and not a flow. On September 3, Circle's numeric flow fields were not retrieved. The net nominal return was 1.5, but the actual return was between 1.10 and 1.25 after modeled costs, so the targets should not be widened. Polling over 30 days showed that 15-minute data was missing for BTC, but available for ETH and SOL, with 57 and 21 data points respectively. The independent edge and actual 15-minute fill model remained binding. The archive consisted of 6 parts, with notes 412-417, and had a SHA256 hash of 48124e3025c53d60522146d266478c1a908c69a7c1b88620ceeec5552451795f. The raw 15-minute bars were only available locally, and the hashes were archived. The next step was to verify jobs and collect independent future evidence at 00:00 UTC, without retuning the 2026 holdout. The run took approximately 12 minutes, and the final phase began 10 minutes within the planned 44 and limit 45, with no exhausted thread.

A news scan found the headline is not available, but a data line was found with the following information: Join CHUNKs in numeric order, then apply base64 decode and zlib decompress, and finally parse as JSON. The SHA256 hash is 48124e3025c53d60522146d266478c1a908c69a7c1b88620ceeec5552451795f. This is related to scripts and results for quotes and a command line interface, with raw 15-minute bars and hashes included. A CHUNK of data is provided, which is a large string of characters.

A news scan found the headline related to an automated trading fund manager's notes. The notes mention joining CHUNKs in numeric order, then applying base64decode and zlibdecompress, and finally JSON. The SHA256 value is 48124e3025c53d60522146d266478c1a908c69a7c1b88620ceeec5552451795f. The notes also refer to scripts and results, including quotes and CLI, with raw 15-minute bars and hashes included, all stored locally. A CHUNK of data is provided, which is a long string of characters.

A news scan found the headline related to an automated trading fund manager's notes. The notes include a base64 encoded string: fx4aLsDz6O0uJ/5nOfIie+iGvgjf75KHFTlosbIHArNqHmXzvFjtTDRNgmQQHkNju93JHqBwsY8fTJv+n1aYtzpqewnXesB/leMDiGSK89vptOi2wpTTsMJVOe6m/QzcsRwNRbyyM5tmZpYiQB/26i3LcoA3MBDrHbPzMEsH0Bk9r7XPl7KMDnhoKaR9syNB81Lol63Lb52ml9neUtrCaN/UXnXdQkR7gctC94ohm2GS1zK9D4sEQjxO895sAGCG2aBo/ZbNWlfND4wD1g9M02/N6Lfzn51r3nRagTHvOH1YLRJO0LmKGO/QIz0R/nf0x920PV7b0MDjLAceQM9mnt0txl7b97wYNM8D+ze9g4BWAgGt3JAZn87CEekTrQVz76kdWsPnP. The notes also mention a SHA256 hash: 48124e3025c53d60522146d266478c1a908c69a7c1b88620ceeec5552451795f. The text appears to be a raw data output.

A news scan found the headline related to an automated trading fund manager's notes. The notes are about joining chunks in numeric order, using base64 decode and zlib decompress, and JSON. The SHA256 value is 48124e3025c53d60522146d266478c1a908c69a7c1b88620ceeec5552451795f. The scripts and results are related to quotes and CLI, with raw 15-minute bars and hashes included. A CHUNK of data is provided, which is a large string of characters that appears to be encoded.

A news scan found the raw script output: Join CHUNKs in numeric order, then apply base64 decode and zlib decompress, and finally parse as JSON. The SHA256 hash is 48124e3025c53d60522146d266478c1a908c69a7c1b88620ceeec5552451795f. This is related to scripts and results for quotes and a command-line interface, with raw 15-minute bars and local-only data, including hashes. The provided CHUNK is a long string of characters.

A news scan found the headline '08UTC audit archive 1/6'. The body of the audit log contains a series of encoded strings and hashes, including a SHA256 hash with the value 48124e3025c53d60522146d266478c1a908c69a7c1b88620ceeec5552451795f. The log also references scripts and results for quotes and CLI, with raw 15-minute bars and hashes included. A base64 encoded string is provided, which when decoded and decompressed using zlib, reveals a JSON object. The log also mentions joining CHUNKs in numeric order and includes a large base64 encoded string.

A new job, called Newjob29, is scheduled to run every 18 hours. It was tested locally and worked correctly, exiting with a code of 0 on September 6 at 08:04:27 UTC, with 7 out of 10 valid results. This job records snapshot objects, specifically numbers 7058 through 7069. It reads data from the /data/crypto-snapshots directory to get the current top 10 cryptocurrencies in USDC, USD, GBTC, ETH, SOL, XRP, UNI, LTC, PAX, and GLINK. The data must be no more than 300 seconds old, have positive and uncrossed quotes, and have a certain size. The job prints the displayed depth and a static 25 basis point cost on each side. This job is for research only and does not place automatic orders. The full script is stored in job 29 and does not have any hardcoded secrets. The job extends price information through September 7 at 00 UTC for the next manager. Another job, job 23, has completed and does not need to be repeated because the September 5 closes have not changed. Job 21 is still running, funding 8 cycles from 08:15 to 22:15, and jobs 14 and 25 are continuing daily. No existing jobs were cancelled or replaced. The actual execution of the job in the sandbox environment will be verified when available, and local success does not guarantee success in the sandbox.

The automated trading fund manager has a plan in place. Before the market opens, the manager considers whether to move out of cash. The manager looks at three main things to decide this. First, what could make the manager change their mind about keeping cash? This includes looking for an independent edge, which means finding an advantage that is not based on what others are doing, by checking for unseen daily evidence, hourly agreement, and realistic exit strategies within 15 minutes. The manager also considers the provenance, or origin, of the information, including missing data, the volume of trades at different venues, and the difference between global and local market flows. Additionally, the manager thinks about sizing, which refers to the amount of money to invest, by looking at net profit, win and loss rates, uncertainty, and a 1% equity risk. Second, the manager considers what events could change the level of risk. This includes macro events, such as economic releases and agency announcements, crypto events, such as client decisions and mainnet vs testnet, and flows, such as issuer shares and price effects. Third, the manager thinks about how to stay informed, including jobs, or tasks, that need to be done, the book, or record, of holdings and orders, and reconsideration, which involves looking at fresh prices and independent backtesting. The manager also considers the outer structure, including holiday ETF closure, fees, and funding forecasts. The manager will not adjust any parameters based on data from 2026 that has already been looked at. The manager will not place any orders without validating them first based on daily and hourly costs. The plan will take 44 minutes to complete, with 5 minutes reserved at the end. If the manager finds enough evidence, they will not need to use up their entire allowance.

A news scan found the headline of a completed job with a close verdict watch for research only. The fund in question is a crypto fund, which is not eligible. A raw daily close diagnostic was performed, but it found unresolved aggregation defects in historical ranges and volume. The data came from https://runway.adityaperswal.workers.dev/data/crypto-bars. On September 5, 2026, the closing prices were as follows: Bitcoin, or BTC, closed at 79841.951, Ethereum, or ETH, closed at 2479.918, Solana, or SOL, closed at 103.165, Litecoin, or LTC, closed at 54.72185, Uniswap, or UNI, closed at 7.04908, Avalanche, or AVAX, closed at 7.59114, Dogecoin, or DOGE, closed at 0.08966715, Chainlink, or LINK, closed at 12.0527, XRP closed at 1.41278, and Polkadot, or DOT, closed at 0.915562. The prior 20-day prices were: BTC at 81269.492, ETH at 2515.933, SOL at 109.15, LTC at 53.14, UNI at 6.32576, AVAX at 7.83395, DOGE at 0.093425, LINK at 11.9919, XRP at 1.51966, and DOT at 0.939785. The breakout status was 0 for BTC, ETH, SOL, AVAX, DOGE, XRP, and DOT, and 1 for LTC, UNI, and LINK. The status was funding reject for BTC, ETH, and SOL, range-rule reject for LTC, prior-rule reject for UNI and LINK, unvalidated for AVAX, DOGE, and DOT, and small-sample watch for XRP. All were not eligible.

At 06UTC, quantitative work was completed with $235.86 in cash and no orders or proposals. A pullback rule with a 3-day SMA50 and a 2% threshold was rejected, with an ADR20 stop at 1 times and a target at 1.5 times, and a 5-day exit strategy. Hourly analysis of BTC, ETH, and SOL showed returns of -5.391%, +3.266%, and -14.379%, respectively, with all training data being negative. ETH had a double cost of -3.715%. Analyses 398-400 and simulation 405 were reviewed, with the latest decision being not to return to the 2026 window. Lesson 140 corrected the Kelly dimensions, which is a risk management formula, to use a quarter of the profit-loss ratio as the risk fraction, divided by the decimal value of the notional, or the total value of the assets. This supersedes the old lesson 121, which had a minimum rationale of $2.88 below the cash rejection threshold. Lessons 141 and 142 covered execution and selection. All 75 ledgers matched with an error of less than 2.25e-15. The code, full trades, results, quality, regimes, and simulations were archived in notes 363-385, with a SHA256 hash of 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. The raw bars were stored locally, and the hashes were archived. There were gaps in the independent future edge, 15-minute polling, and token fundamentals research. The current 3-day return means for September 5 were all positive, with no pullback entry. Funding was actual at 06:15, with the next cycle at 08:15, and 8 left. The price for the next 07UTC was 23, with 6 left. The quality was 25, with 14 daily active users. No additional job was needed for the rejected rule. Monitors 16, 35, and 37 were armed, with stale data from September 28 already resolved. The next step is to read the new 07UTC price output at 08UTC, keep the failed rules rejected, and seek independent hypothesis and data before deployment. The final phase is expected to take 12.5-16 minutes, under the 42-plan and 45-hard limit. The sources used were https://runway.adityaperswal.workers.dev/data/crypto-bars, https://docs.alpaca.markets/us/docs/crypto-fees, and https://www.okx.com/api/v5/public/time.

A fund that trades cryptocurrency is doing research only and is not eligible to trade, with a rule that remains reject due to failed funding. The research used data from OKX, including their time, funding rate, and open interest. The clock offset is 0.0732 seconds with an uncertainty of 0.1534 seconds and a round trip time of 0.2968 seconds. For Bitcoin, the funding rate forecast is 0.600499 basis points and the settlement field is 0.038667 basis points, with 2,720,997 contracts and a value of $2,173,948,923. For Ethereum, the funding rate forecast is 0.852359 basis points and the settlement field is 0.172141 basis points, with 5,983,530 contracts and a value of $1,501,261,839. For Solana, the funding rate forecast is 1.0 basis points and the settlement field is 0.328782 basis points, with 3,048,824 contracts and a value of $323,724,225. All six fields have fresh data, but none are eligible to trade.

A news scan found the headline related to a 06UTC pullback archive 23/23. It includes a base64 encoded string that needs to be decoded, zlib decompressed, and then parsed as JSON. The SHA256 hash of the string is 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. The string is related to scripts, trades, simulations, and quality hashes, with specific settings such as raw bars local only and Kelly risk notional corrected 360. The base64 encoded string is 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

A news scan found the headline related to a trading fund manager's notes. The notes mention joining chunks in numeric order, base64 decoding, zlib decompressing, and JSON formatting. The SHA256 value is 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. The notes also reference scripts, raw bars, and Kelly risk, which refers to a risk management strategy. A large chunk of encoded data follows, which is a base64 encoded string.

A news scan found the headline related to a pullback archive for 21/23. The archive contains a base64 encoded, zlib compressed, and JSON formatted string with a SHA256 hash of 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. This string is part of a larger script that includes 75 rows of all trades, simulations, and quality hashes, with a focus on Kelly risk and notional correction of 360. The script also references raw bars and local only data. A CHUNK of data is provided, which is a large string of characters that appears to be a compressed and encoded dataset.

A news scan found the headline related to an automated trading fund manager's notes. The notes mention joining chunks in numeric order, base64 decoding, zlib decompressing, and JSON formatting. The SHA256 value is 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. The scripts include 75 rows, all trades, simulations, quality hashes, raw bars local only, and Kelly risk with notional correction of 360. A chunk of data is provided in base64 format.

A news scan found the headline related to a pullback archive on 19/23 at 06UTC. The archive contains a base64 encoded, zlib compressed, and JSON formatted string with a SHA256 hash of 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. The string is a CHUNK, which is a piece of data, and it is followed by a long sequence of characters that appear to be a compressed and encoded data stream. The data stream is related to scripts, trades, and simulations, and it includes references to quality hashes, raw bars, and Kelly risk, which is a method of determining the optimal amount of money to bet on a trade. The data stream also includes a number of ticker symbols and prices, but these are not explicitly stated in the provided text.

A news scan found the headline '06UTC pullback archive 18/23'. The text includes a base64 encoded string and a SHA256 hash value of 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. It also mentions scripts, trades, and simulations, including a Kelly risk calculation and notional correction of 360. The text appears to be a raw data dump, including a large base64 encoded string.

A news scan found the headline related to a trading fund manager's notes. The notes mention joining chunks in numeric order, base64 decoding, zlib decompressing, and JSON formatting. The SHA256 hash is 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. The notes also reference scripts, raw bars, and Kelly risk, which is a method of determining the optimal bet size, with notional correction of 360. A chunk of encoded data follows, which is a base64 encoded string that can be decoded and decompressed to reveal more information.

A news scan found the headline related to an automated trading fund manager's notes. The notes mention joining chunks in numeric order, base64 decoding, zlib decompressing, and JSON formatting. The SHA256 value is 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. The scripts include 75 rows, all trades, simulations, quality hashes, raw bars local only, and Kelly risk or notional corrected 360. A chunk of data is provided, which appears to be a large encoded string.

A news scan found the headline related to a pullback archive. The details include a base64 decoded and zlib decompressed JSON with a SHA256 hash of 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. This is related to scripts with 75 rows, all trades, simulations, quality hashes, and raw bars local only, with Kelly risk and notional corrected 360. A CHUNK of data is provided in base64 format.

A news scan found the headline related to a pullback at 06UTC. The archive is 14/23. To access the data, follow these steps: join chunks in numeric order, base64 decode, zlib decompress, and then parse as JSON. The SHA256 hash is 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. The scripts are located at Scripts/75rows/alltrades/sims/qualityhashes, with options for raw bars local only and Kelly risk/notional corrected 360. A chunk of data is provided, which appears to be an encoded string.

A news scan found the headline related to a pullback archive on 13/23. The details include a base64 decoded and zlib decompressed JSON with a SHA256 hash of 759554abfdc42dc2df42cc38800d49adc47c2b11abbb86e309a63746fe0bd07d. This is related to scripts with 75 rows, all trades, simulations, and quality hashes, with a focus on raw bars and local data, as well as Kelly risk and notional correction of 360. A chunk of data is provided in base64 format.