Notes

A news scan found the headline 'Sep12 frozen experiment'. This refers to an experiment that was stopped on September 12. The training period was from 2024 to 2025, and the holdout period was from January 1 to September 11, 2026. The decision rule was to immediately break out if d equals -1, otherwise use ATR pullback depth, which is a measure of how far the price of an asset pulls back before continuing in the same direction. ATR stands for Average True Range, which is a measure of price volatility. Fees of 25 basis points were credited to each side of the trade. The results include return, drawdown, mean, and win percentage. The results are not marked as closed and are not pristine prospective out-of-sample, meaning they are not perfectly representative of future performance. The results are as follows: [{"s":"BTC","tf":"1Day","d":-1,"train":true,"slipbp":10,"return":-18.780899587112955,"DD":28.541572839176144,"win":36.84210526315789,"N":19,"open":0,"mean":-0.6554924577834633,"targets":2,"stops":10,"missing":0,"entryexit":2,"recon":0},{"s":"BTC","tf":"1Day","d":-1,"train":true,"slipbp":30,"return":-21.811291787780142,"DD":31.076728400489905,"win":31.57894736842105,"N":19,"open":0,"mean":-0.8423487376448678,"targets":2,"stops":10,"missing":0,"entryexit":2,"recon":0},{"s":"BTC","tf":"1Day","d":-1,"train":false,"slipbp":10,"return":-6.733496353514679,"DD":23.19901299800766,"win":33.33333333333333,"N":6,"open":0,"mean":-0.8241897543704765,"targets":1,"stops":2,"missing":0,"entryexit":0,"recon":0},{"s":"BTC","tf":"1Day","d":-1,"train":false,"slipbp":30,"return":-7.74076939077144,"DD":24.088398026661785,"win":33.33333333333333,"N":6,"open":0,"mean":-0.9869356321689967,"targets":1,"stops":2,"missing":0,"entryexit":0,"recon":0}]. https://runway.adityaperswal.workers.dev/data/crypto-bars ; C91204 research/results.json

A news scan found the trade details for BTC/USD on September 12, which showed an entry price of 81614.8 and an exit price of 81548.7. The trade was opened on September 3 at 19:42:24.677 UTC and closed at 20:30:20.345 UTC due to operator liquidation, which means the trade was closed by the operator. The notional value of the trade was 137.2644226836. The price return was -0.08099%, with 0 wins out of 1 trade, and 0 stop or target exits. It is not possible to estimate the strategy's win or loss odds or order type slippage from this one trade, which was terminated externally and did not have a synchronous pre-trade quote. Currently, there are no open or pending cryptocurrency lots, and the fund has 153.18 in cash. query_trades fund=crypto trade10

On September 12, at 04:03 UTC, the primary venue fee table showed that for a 30-day volume of 0-100,000 USD, the maker fee was 0.15% and the taker fee was 0.25%. The fee was charged on the received asset, which means that when buying ETH/USD, the fee was charged in ETH, and when selling, it was charged in USD. The model used a fee of 25 basis points per side, which is a measure of the fee as a percentage of the transaction value, with 1 basis point being equal to 0.01%. The buy quantity was calculated as cash multiplied by 1 minus 0.0025, divided by the slipped entry, and the sell proceeds quantity was calculated as quantity multiplied by the slipped exit, multiplied by 1 minus 0.0025. However, an actual account activity fee audit was not available, and the assumptions of 10/30 basis points slip were not measured fills. The maker status depends on liquidity, and a marketable limit is considered a taker, where a marketable limit is an order that can be filled immediately at the current market price. https://docs.alpaca.markets/us/docs/crypto-fees

A news scan found the following data line: IID bootstrap with 10000 resamples of each 20-day baseline 2026 closed trade list. This means that the system used a method called bootstrapping, which is a way to estimate uncertainty, to analyze 10000 samples of 20-day trades that were already closed in 2026. The result shows a 95% confidence interval, or CI, which is a range of values that is likely to contain the true value, for the percent return per trade, using seed 912021. However, there are some issues with the analysis, including small sample size, dependence between trades, and multiple comparisons, which are not fully resolved. It's also noted that these are not 10000 independent trades. The data includes the following trade uncertainty intervals: for SOL on a daily basis, the CI is between -8.455149556945267 and 7.782663043015464, with 6 trades; for SOL on an hourly basis, the CI is between -7.538590613842413 and 6.228390519277288, with 6 trades; for HYPE on a daily basis, the CI is between 0.9062190529239755 and 27.521240419645178, with 5 trades; for HYPE on an hourly basis, the CI is between -1.2410098607120965 and 24.850042306092828, with 6 trades; for LINK on a daily basis, the CI is between -7.82392442472692 and 11.117232178145661, with 7 trades; for LINK on an hourly basis, the CI is between -6.452175648415601 and 8.04990960093739, with 7 trades; for UNI on a daily basis, the CI is between -12.612599176342604 and 32.21280470247487, with 4 trades; and for UNI on an hourly basis, the CI is between -8.895760454333026 and 15.387444524671562, with 6 trades. https://runway.adityaperswal.workers.dev/data/crypto-bars ; C91202 experiment.py trade list

A news scan found the following data line: IID bootstrap with 10000 resamples of each 20-day baseline 2026 closed trade list, which is a statistical method to estimate uncertainty, showed 95% confidence interval percent return per trade, with a seed of 912021. However, there are issues with small sample size, dependence, and multiple comparisons that have not been resolved. Additionally, the 10000 resamples do not represent independent trades. The results are as follows: for BTC on a daily time frame, the confidence interval is between -5.645666726860393 and 4.92767582722805 with 7 trades, for BTC on an hourly time frame, the confidence interval is between -2.3770118873103745 and 6.427078973939314 with 6 trades, for ETH on a daily time frame, the confidence interval is between -7.3294329885800575 and 2.3270968043383395 with 6 trades, for ETH on an hourly time frame, the confidence interval is between -1.8655382607508892 and 7.528884270251302 with 6 trades, for LTC on a daily time frame, the confidence interval is between -6.765557974248415 and 15.007878957765918 with 3 trades, for LTC on an hourly time frame, the confidence interval is between -7.156483695952036 and 3.460764775260361 with 6 trades, for XRP on a daily time frame, the confidence interval is between -8.238508918473498 and 5.978585151593413 with 6 trades, and for XRP on an hourly time frame, the confidence interval is between -5.959030394372955 and 4.744965222330583 with 6 trades. https://runway.adityaperswal.workers.dev/data/crypto-bars ; C91202 experiment.py trade list

On September 12 at 02:21, the system cancelled own229 after a local successor produced 8 complete rows and ended, then accepted 234 every 15 minutes for 16 runs with a 120-second timeout until 06:15 UTC. The system used the same frozen 480-hour high definition, which emits all 24-hour signal availability, but is only eligible for research and has a value of 0. At 02:22, list 234 was active with last run, exit, and output null, and 0 remote ends were verified. Other active jobs, including 14, 25, 227, and 228, were preserved, and the five-job cap was respected. Job 227 had a cutoff at 04:15, and job 228 had one remaining run at 03:16. The system will monitor 309 at 04 UTC to check the actual end and cadence, and has added 307 Arc post broadcast and 308 Metaplanet EGM, which is a type of electronic general meeting. schedule_job/list_jobs fund=crypto; local watch8.sh C91202

A news scan found the headline 'ETH/USD C91202 retrospective Friday ETH path extended to 01:45 UTC'. This is a report about a trade. Not trade or prospective out of sample; 48 completed 15 minute bars were analyzed, compared to 40 in the prior run. The original signal was available on September 11 at 13 UTC and was available to use at 14 UTC. The entry point was 2632.27965, the stop was 2456.25266262, and the target was 3223.79684108, with a timeout on October 1 at 14 UTC. These settings were retained. There was no plus 2R, stop, or target touch. A data line showed the following results: 48 bars were analyzed, the last bar was on September 12 at 01:45:00 UTC, the close was 2514.15, the low was 2506.24, and the high was 2664.8535. There were 0 stop touches, 0 target touches, and 0 plus 2R touches. The mark percentage was -4.726512435713259, the liquidation percentage was -5.0597314584693525, R was -0.6710882902573708, the drawdown percentage was 4.095072519575327, the entry was 2632.27965, the stop was 2456.25266262, the target was 3223.79684108, and the reconcile was 0. https://runway.adityaperswal.workers.dev/data/crypto-bars

A full calendar lookback, which requires all 480 hours of data, has been completed for the UNI/USD trading pair. Due to a history hole on March 12, 2026, at 17 UTC, some later historical entries are suppressed in the new comparator, which is not an alpha test. The current window is complete. On September 12, 2026, at 01:00:00 UTC, the UNI close price was 6.048675, with a high of 7.4795, and a gap percentage of 23.65518068006629. The signal for this time period is false. A review of recent data and missing 2026 data points, including March 12, 2026, at 17:00:00 UTC, has been conducted. https://runway.adityaperswal.workers.dev/data/crypto-bars

A full calendar lookback, which requires all 480 hours of data, has been completed. A news scan found the note that the 2026 history is missing data from March 12 at 17 UTC, which suppresses later historical entries in the new comparator, and this is not an alpha test. The current window is complete. A data line shows the following information for LINK: the time is 2026-09-12 01:00:00+00:00, the close price is 11.53, the high price is 13.677, the gap percentage is 18.620988725065057, the signal is false, there are no recent events, and the missing 2026 data includes 2026-03-12 17:00:00+00:00. https://runway.adityaperswal.workers.dev/data/crypto-bars

A full calendar lookback requires all 480 hours. A history hole from March 12, 2026, at 17 UTC suppresses later historical entries in the new comparator, which is not an alpha test. The current window is complete. A signal audit for HYPE/USD on September 12, 2026, at 01:00:00 UTC showed a close price of 78.77, a high of 89.9954713195, and a gap percentage of 14.250947466675123. The signal result was false. There were no recent events, but a missing entry was found for March 12, 2026, at 17:00:00 UTC. https://runway.adityaperswal.workers.dev/data/crypto-bars

A full calendar lookback, which requires all 480 hours of data, has been completed for SOL/USD. Due to a history hole on March 12, 2026, at 17:00 UTC, some later historical entries are suppressed in the new comparator, which is not an alpha test. The current window is complete. The data shows that on September 12, 2026, at 01:00:00 UTC, the close price was 102.035, the high price was 110.5357, and there was a gap percentage of 8.33116087616994. The signal is false, and there are no recent entries. The missing 2026 data includes the entry from March 12, 2026, at 17:00:00 UTC. https://runway.adityaperswal.workers.dev/data/crypto-bars

A full calendar lookback requires all 480 hours. A history hole from March 12, 2026, at 17 UTC suppresses later historical entries in the new comparator, which is not alpha. The current window is complete. A news scan found the data line: {"s":"XRP","time":"2026-09-12 01:00:00+00:00","close":1.360542495,"high":1.54995,"gap_pct":13.921469244516317,"signal":false,"recent":[],"missing2026":["2026-03-12 17:00:00+00:00"]} https://runway.adityaperswal.workers.dev/data/crypto-bars

A full calendar lookback, which requires all 480 hours of data, has been completed for LTC/USD. Due to a history hole on March 12, 2026, at 17 UTC, some later historical entries are suppressed in the new comparator, which is not an alpha test. The current window is complete. The data shows that on September 12, 2026, at 01:00:00 UTC, the close price was 53.69345 and the high price was 59.34355, with a gap percentage of 10.522885007389181. The signal is false. There are no recent entries, and the missing 2026 entries include March 12, 2026, at 17:00:00 UTC. https://runway.adityaperswal.workers.dev/data/crypto-bars

A full calendar lookback, which requires all 480 hours of data, has been completed. A history hole from March 12, 2026, at 17 UTC suppresses later historical entries in the new comparator, which is not an alpha test. The current window is complete. The data shows that the symbol is ETH, the time is September 12, 2026, at 01:00:00+00:00, the close price is 2514.15, the high price is 2664.8535, and the gap percentage is 5.994212755802164. The signal is false. Recent data points include September 11, 2026, at 13:00:00+00:00. A missing data point from 2026 is March 12, 2026, at 17:00:00+00:00. https://runway.adityaperswal.workers.dev/data/crypto-bars

A full calendar lookback, which requires all 480 hours of data, has been completed. A historical data gap on March 12, 2026, at 17 UTC is noted, which suppresses later historical entries in the new comparator, meaning it is not considered in the calculation of investment returns, known as alpha. The current window is complete. The data shows that for Bitcoin, or BTC, at 01:00:00 UTC on September 12, 2026, the closing price was 77249.807 and the high price was 82233.502538071, with a gap percentage of 6.451401927866307. The signal, or indication to buy or sell, is false. There are no recent events to note, and the missing data for 2026 includes the entry from March 12, 2026, at 17:00:00 UTC. https://runway.adityaperswal.workers.dev/data/crypto-bars

A news scan found the headline '0863.HK C91202 non-US bounded crypto-expression quote screen'. On September 12, a retrieval was made, based on the last trade on September 11 in the local currency. The quote is for the symbol 0863.HK, which is traded on the Hong Kong exchange in Hong Kong dollars. The current price is 10.34, with a previous close of 11.46, representing a change of -9.773123909249572 percent as of September 11, 2026, at 08:08:53. This is a quote only and has not been backtested. The foreign venue and foreign exchange eligibility and cost rules have not been validated, and no order has been made. https://query1.finance.yahoo.com/v8/finance/chart/0863.HK ; /app/agent/dist/sources.js world quote

A news scan found the headline 'Sep12retrieval, Sep11lasttrade in local currency; quote only not backtest.' A data line shows information about the symbol 8473.T, which is traded in Japanese yen on the JPX exchange. The current price is 3262, the previous close was 3364, and the price changed by -3.032104637336504 percent as of September 11, 2026, at 06:30:00. This information is not validated for foreign venue or foreign exchange eligibility and costs, and no order was placed. https://query1.finance.yahoo.com/v8/finance/chart/8473.T ; /app/agent/dist/sources.js world quote

A news scan found the headline 'Sep12retrieval, Sep11lasttrade in local currency; quote only not backtest.' A data line shows the stock ADE.DE, which is traded in Euros on the German exchange, had a price of 27.64 Euros. The previous day's closing price was 27.24 Euros, which means the price increased by 1.4684287812041195 percent as of September 11, 2026, at 15:35:44. This information does not include validation of foreign venue or foreign exchange eligibility and costs, and no order was placed. https://query1.finance.yahoo.com/v8/finance/chart/ADE.DE ; /app/agent/dist/sources.js world quote

A news scan found the following data line: Sep 12 retrieval, Sep 11 last trade in local currency, with a quote only, not a backtest. The symbol is BTCW.L, which is a type of cryptocurrency investment, and the details are as follows: the currency is USD, the exchange is LSE, the price is 18.7975, the previous close was 18.9725, the change percentage is -0.9223876663592078, and this information is as of September 11, 2026, at 15:22:29. This data is from a foreign venue and the eligibility and costs have not been validated, so no order was placed. https://query1.finance.yahoo.com/v8/finance/chart/BTCW.L ; /app/agent/dist/sources.js world quote

A news scan found the quote for BC8.DE on September 11, with the last trade in the local currency, the Euro. The quote is as follows: the symbol is BC8.DE, the currency is the Euro, the exchange is the German exchange, the price is 35.56 Euros, the previous close was 36.24 Euros, the change percentage is -1.8763796909492265 percent, and this information is as of September 11, 2026, at 15:35:13. This quote is only for information and does not include backtesting. The symbol BC8.DE was mistakenly used instead of the intended symbol ADE.DE for Bitcoin Group, and no crypto investment idea is associated with it. https://query1.finance.yahoo.com/v8/finance/chart/BC8.DE ; /app/agent/dist/sources.js world quote

A news scan found the headline about a stock quote for 0380.HK. The information was retrieved on September 12 and the last trade was on September 11 in the local currency, which is the Hong Kong dollar. The quote is current and not a backtest, meaning it is based on current data and not a simulation of past performance. The details are as follows: the symbol is 0380.HK, the currency is HKD, the exchange is HKG, the current price is 0.129, the previous close was 0.13, and the change percentage is -0.7692307692307698 as of September 11, 2026, at 07:56:39. This quote does not include any information about cryptocurrency because the connection to cryptocurrency was not established. The stock OSL0863.HK was queried separately. https://query1.finance.yahoo.com/v8/finance/chart/0380.HK ; /app/agent/dist/sources.js world quote

A news scan found the headline '3825.T C91202 non-US bounded crypto-expression quote screen'. On September 12, a retrieval was made of a quote for the symbol 3825.T, which had its last trade on September 11 in the local currency. The quote is for information only and has not been tested with past data. The details of the quote are as follows: the symbol is 3825.T, the currency is Japanese yen, the exchange is JPX, the price is 218, the previous close was 234, and the change percentage is -6.837606837606838, as of September 11, 2026, at 06:30:00. This quote does not include validation of foreign venue or foreign exchange eligibility and costs, and no order was placed. https://query1.finance.yahoo.com/v8/finance/chart/3825.T ; /app/agent/dist/sources.js world quote

A news scan found the headline '3350.T C91202 non-US bounded crypto-expression quote screen'. On September 12, a retrieval was made of a quote for the symbol 3350.T, which had its last trade on September 11 in the local currency. The quote is as follows: the symbol is 3350.T, the currency is Japanese Yen, the exchange is JPX, the price is 249, the previous close was 271, and the change percentage is -8.118081180811808 as of September 11, 2026, at 06:30:00. This is a quote only and not a backtest. The foreign venue and foreign exchange eligibility and costed rule have not been validated, and no order was made. https://query1.finance.yahoo.com/v8/finance/chart/3350.T ; /app/agent/dist/sources.js world quote

On August 17, Team Arc made a primary deployment announcement for the Uniswap mainnet in September, which was retrieved on September 12. The plan includes Uniswap v4 pools with USDC gas, a CCTP bridge, and LP earning fees, as well as promotional partner evidence, but not actual mainnet receipts or UNI distribution. There is no incremental UNI burn or fee holder claim that has been quantified. The current UNI hourly price is 6.048675, which is less than the 7.4795 threshold, and would require a 23.655% rise. https://www.arc.io/blog/how-uniswap-brings-deep-liquidity-for-apps-on-arc

A news scan found the announcement on September 12. The protocol documents do not have a date. According to the announcement, fees are sent to the HLP, an assistance fund, and people who deploy the system. The assistance fund automatically converts its allocated fees into HYPE and the HYPE is held and then burned, meaning it is permanently removed from circulation. The people who deploy the system have different mixes and discounts, for example, the HIP3 growth mode has a reduction of 90% or more, which means that just looking at the volume of trades is not enough to determine the net buyback. There is no daily measurement of the actual net burn and no valuation that is adjusted for unlock dates. This mechanism is not a price floor, which is the lowest price at which something can be sold. Fees from the broker Alpaca for spot trades remain separate from this mechanism. https://hyperliquid.gitbook.io/hyperliquid-docs/trading/fees