A news scan found the following data: Computed on September 8 at 06:20 UTC, this is a fixed trend and compression analysis with a base cost of 0.75 for closed holdout trades. The analysis of the XRP/USD symbol shows the following results: the number of trades, N, is 20, the mean percentage change is -0.5849328048357155, and the block 3 mean 95 percent change is between -1.9717746996628032 and 0.7181559522997911. The fraction of bootstrap samples with a positive result is 0.2001. This analysis is based on 10,000 circular moving block resamples of 3 consecutive trades, with a seed of 2026090819, which was determined after 10,000 daily draws. The analysis preserves local trade dependence but not all regimes, and the confidence interval is conditional on the observed sample, not the posterior probability of the strategy's profit. Boundary marks were excluded from the analysis, and all intervals cross zero. The SOL missing exit is considered invalid and was not resampled.
https://runway.adityaperswal.workers.dev/data/crypto-bars
On September 8 at 06:20 UTC, a fixed trend and compression analysis was computed with a base cost of 0.75 for closed holdout trades. The results for the symbol ETH/USD showed 36 trades with a mean percentage change of -0.6473685335893425. The block of 3 mean 95 percent change was between -1.592072816475238 and 0.2632904769772734. The fraction of bootstrap samples with a positive result was 0.0866. This was based on 10,000 circular moving block resamples of 3 consecutive trades, using a seed of 2026090819 after 10,000 daily draws. The analysis preserves local trade dependence but not all regimes, and the confidence interval is conditional on the observed sample, not the posterior probability of the strategy's profit. Boundary marks were excluded, and all intervals cross zero. The SOL missing exit was invalid and not resampled.
https://runway.adityaperswal.workers.dev/data/crypto-bars
On September 8 at 06:20 UTC, a fixed trend and compression analysis was computed with a base cost of 0.75 for closed holdout trades. The results for the symbol BTC/USD showed 29 trades with a mean percentage change of -0.7644496363475459. The block of 3 trades had a mean percentage change of -1.7954535562678218 to 0.18981890796093154 at the 95 percent level. Only 6.18 percent of the bootstrap samples were positive. This was based on 10,000 circular moving block resamples of 3 consecutive trades, using a seed of 2026090819 after 10,000 daily draws. The analysis preserves local trade dependence but not all regimes and provides a confidence interval conditional on the observed sample, not the posterior probability of the strategy's profit. Boundary marks were excluded and all intervals crossed zero. The SOL missing exit was considered invalid and not resampled.
https://runway.adityaperswal.workers.dev/data/crypto-bars
A news scan found the headline about a LINK/USD hourly compression experiment. The experiment was computed on September 8 at 06:19 UTC. It used data from 2025 for training and from January 1, 2026, to September 7 for holdout. The next open trade was considered, with a daily average daily range stop and a 1.5 average daily range target, plus one risk half to break even, and a 72-hour horizon. A fee of 25 basis points was credited to the asset for buys and fiats for sells. There were two stress tests, with the first having a 5 basis point major, 30 basis point XRP-LTC, and 15 basis point UNI-LINK slip, and the second stress test doubling these values. The results were as follows: for the training period, the return was -58.9138076 percent, the drawdown was 65.6522169 percent, the win rate was 42.1052632 percent, and the average win percentage was 5.00586962 percent. For the holdout period, the return was -56.8304175 percent, the drawdown was 59.9761793 percent, the win rate was 38.6363636 percent, and the average win percentage was 3.05830694 percent. The metrics were calculated based on percentage returns, excluding boundary exits. A return of zero was considered invalid. The experiment allowed for missing bars when the indicator was flat, with the selection limitation retained. The bar resolution was not validated at a 15-minute venue.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the headline 'LINK/USD Hourly compression experiment LINK/USD|true|1|1'. This is a report from a trading fund manager. The report was computed on September 8 at 06:19 UTC. It is about an experiment that used data from 2025 for training and data from January 1, 2026, to September 7 for testing. The experiment used a daily stop loss and a target of 1.5 times the average daily range, with a 72-hour time frame. It also used a fee of 25 basis points, which is 0.25%, credited to the asset when buying and to fiat when selling. The experiment had two stress tests, with the second one being twice as stressful as the first one. The results of the experiment are as follows: for the training period, the return was -49.42%, the maximum drawdown was 58.03%, and the win rate was 45.61%. For the testing period, the return was -46.45%, the maximum drawdown was 53.98%, and the win rate was 39.53%. The report also includes other metrics, such as the average trade percentage, the average win percentage, and the average loss percentage. The report notes that the results do not include boundary exits and that a missing bar is allowed when the indicator is reset. The report also states that the bar resolution is not 15 minutes and that the venue validation is not applicable.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the headline about a LINK/USD hourly compression experiment. The experiment was computed on September 8 at 06:19 UTC. It used data from 2025 for training and from January 1, 2026, to September 7 for holdout. The next open and daily ADR stop with a 1.5 ADR target and a 1R half BE were considered. A fee of 25 basis points was credited to the asset buy and fiat sell. There were two stress levels: stress 1 with 5 basis points for majors, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, and stress 2 which was double the first. The results were as follows: for the training period, the return was -54.7381393 percent, the drawdown was 59.167526 percent, the win rate was 42.5531915 percent, and the average win percentage was 4.91845344 percent. For the holdout period, the return was -48.1070429 percent, the drawdown was 54.4729285 percent, the win rate was 35.2941176 percent, and the average win percentage was 2.96876516 percent. The results also included other metrics such as mean trade percentage, average loss percentage, and fees in USD. Note that a basis point is a unit of measurement, one hundredth of a percent. The experiment allowed for missing bars while the indicator was flat and retained selection limitations. The bar resolution was not 15-minute venue validation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the headline about a LINK/USD hourly compression experiment. The experiment was computed on September 8 at 06:19 UTC. It used data from 2025 for training and from January 1, 2026, to September 7 for holdout. The next open, daily average daily range stop, and 1.5 average daily range target were used, with a 72-hour horizon and a fee of 25 basis points credited to the asset buy and fiat sell. Two stress tests were run, with stress 1 being 5 basis points for majors, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, and stress 2 being double that. The results were as follows: for the training period, the return was -48.9787084 percent, the drawdown was 54.1790008 percent, and the win rate was 44.6808511 percent, with 47 trades, a mean trade percentage of -1.23040975 percent, an average win percentage of 4.83018449 percent, and an average loss percentage of 6.12550511 percent. For the holdout period, the return was -31.3798577 percent, the drawdown was 42.6242532 percent, and the win rate was 41.9354839 percent, with 31 trades, a mean trade percentage of -1.13196899 percent, an average win percentage of 2.99891086 percent, and an average loss percentage of 4.11538221 percent. A basis point is one hundredth of a percent. The results also included metrics such as quarter Kelly, boundary exits, and reconciliation in USD. The buy and hold percentage was -38.9699747 percent for the training period and 4.59392945 percent for the holdout period. Fees in USD were 157.678715 for the training period and 112.941272 for the holdout period. Exposure percentage was 25.0513699 percent for the training period and 26.6083333 percent for the holdout period. The experiment was valid for both periods. Note that N excludes boundary exit, and invalid results are those with a non-zero return. Missing bars while flat are allowed with indicator reset, and selection limitation is retained. Bar resolution is not 15-minute venue validation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the following data line: Computed on September 8 at 06:19 UTC, frozen note 968. This is about an experiment that used data from 2025 for training and from January 1, 2026, to September 7 for testing, with the next open trade and a daily stop-loss at 1.5 times the average daily range and a target at 1 times the risk, with a 72-hour time frame. The fee was 25 basis points credited to the asset when buying and to fiat when selling, with a slip stress of 5 basis points for major currencies, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, and a second stress level that doubles these values. The results were as follows: for the training period, the return was -77.9073889 percent, the drawdown was 81.4687841 percent, the win rate was 42.8571429 percent, the average trade percentage was -1.9528397 percent, the average win percentage was 4.11595838 percent, and the average loss percentage was 6.50443826 percent, with 70 trades, and for the testing period, the return was -62.290221 percent, the drawdown was 66.7235922 percent, the win rate was 37.7358491 percent, the average trade percentage was -1.66955452 percent, the average win percentage was 3.05657208 percent, and the average loss percentage was 4.53387367 percent, with 53 trades. Note that a basis point is one hundredth of a percent, a drawdown is a decrease in value, and a win rate is the percentage of winning trades. The results also include a quarter Kelly criterion of 0, which is a measure of the optimal fraction of a portfolio to invest, and a reconciliation in USD of 2.27373675e-13, which is a very small number. The buy and hold percentage, which is the return if the asset was simply bought and held, was -38.9699747 percent for the training period and 4.59392945 percent for the testing period. The fees in USD were 167.334543 for the training period and 160.329181 for the testing period. The exposure percentage, which is the percentage of the portfolio that is invested, was 39.8858447 percent for the training period and 38.4 percent for the testing period. The results are valid, but the return cannot be zero, and a missing bar is allowed if the indicator is reset, with a limitation on the selection of trades retained. The bar resolution is not 15 minutes, and the venue validation is not applicable.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the following data: Computed on September 8 at 06:19 UTC, this is a frozen note 968. The data is from a 2025 training period and a 2026 January 1 to September 7 holdout period. The next open is based on a daily Average Daily Range (ADR) stop and a 1.5 ADR target, with a 72-hour horizon and a risk-reward ratio of 1, where ADR means the average daily price range. There is a fee of 25 basis points, which is 0.25%, credited to the asset when buying and to fiat when selling, where basis points are a unit of measurement for interest rates and fees. The slip stress is 5 basis points for major currencies, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, where slip stress refers to the impact of buying or selling on the market price. The data includes two sets of results: one for the training period and one for the holdout period. The results are as follows: for the training period, the return was -72.0238236%, the drawdown was 76.7093209%, and the win rate was 45.7142857%, where drawdown refers to the maximum loss from a peak to a trough. For the holdout period, the return was -47.3016675%, the drawdown was 58.8330247%, and the win rate was 42%. The data also includes other metrics, such as the average trade percentage, average win percentage, and average loss percentage. The percent metrics do not include boundary exits, where boundary exits refer to trades that are stopped when a certain price level is reached. The data is invalid if the return is not zero, and missing bars are allowed while the indicator is reset, where an indicator is a tool used to analyze market data. The bar resolution is not 15 minutes, and venue validation is not applicable.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the following data line: Computed on September 8 at 06:19 UTC, this is a frozen note numbered 968. The data is from a training period of 2025 to December 31, 2025, and a holdout period of January 1, 2026, to September 7, 2026. The next open trade is based on a daily average daily range stop and a 1.5 average daily range target, with a risk-to-reward ratio of 1 and a breakeven point, and a 72-hour time horizon. A fee of 25 basis points is credited to the asset when buying and to fiat when selling. There are two stress tests, with the second being twice as stressful as the first, and the stress levels are 5 basis points for major currencies, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK. The results are as follows: for the training period, the return was -77.9913583 percent, the drawdown was 80.911664 percent, the win rate was 40 percent, and the average win percentage was 4.91152117 percent, and for the holdout period, the return was -44.400043 percent, the drawdown was 50.3780969 percent, the win rate was 46.1538462 percent, and the average win percentage was 3.21432528 percent. The metrics are based on percentage values, and the number of trades, denoted as N, excludes boundary exits. If the return is not zero, it is considered invalid. The data allows for missing bars when the indicator is flat, and the bar resolution is not validated at 15-minute intervals.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the headline about a LINK/USD hourly compression experiment. The experiment was computed on September 8 at 06:19 UTC. It used data from 2025 for training and from January 1, 2026, to September 7 for holdout. The next open and daily ADR stop with a 1.5 ADR target and a 72-hour horizon were considered. ADR stands for Average Daily Range. The fee was 25 basis points credited to the asset buy and fiat sell. Basis points are units of measurement for interest rates and fees. Slip stress was set at 5 basis points for majors, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK. Stress 2 was double the stress 1. The results were as follows: for the training period, the return was -63.3537902%, the drawdown was 68.2163407%, and the win rate was 41.2698413%. For the holdout period, the return was -36.8194011%, the drawdown was 44.5214237%, and the win rate was 43.902439%. The results also included other metrics such as mean trade percentage, average win percentage, and average loss percentage. The experiment allowed for missing bars while flat with indicator reset and retained selection limitation. The bar resolution was not 15-minute venue validation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the headline about a LINK/USD hourly compression experiment. The experiment was computed on September 8 at 06:19 UTC. It used data from 2025 for training and from January 1, 2026, to September 7 for testing. The strategy used a daily average daily range stop and a 1.5 average daily range target, with a 72-hour horizon. It also used a fee of 25 basis points credited to the asset when buying and to fiat when selling. The results are shown in two lists, one for the training period and one for the testing period. The training period results are as follows: the symbol is LINK/USD, the time frame is 1 hour, the trend is false, the compression is 0.75, the stress is 1, the return is -56.8210188, the drawdown is 62.6627012, the win rate is 42.8571429, the number of trades is 63, the mean trade percentage is -1.12509817, the average win percentage is 5.26352118, the average loss percentage is 5.91656269, the quarter Kelly is 0, the boundary exits are 0, the reconciliation in USD is 2.27373675e-13, the buy and hold percentage is -38.9699747, the fees in USD are 211.695569, the exposure percentage is 37.0376712, and the results are valid. The testing period results are as follows: the symbol is LINK/USD, the time frame is 1 hour, the trend is false, the compression is 0.75, the stress is 1, the return is -17.4835337, the drawdown is 36.0122983, the win rate is 51.2820513, the number of trades is 39, the mean trade percentage is -0.410775647, the average win percentage is 3.20062158, the average loss percentage is 4.21224641, the quarter Kelly is 0, the boundary exits are 0, the reconciliation in USD is 2.27373675e-13, the buy and hold percentage is 4.59392945, the fees in USD are 160.417573, the exposure percentage is 35.5916667, and the results are valid. The results also include percent metrics, and the number of trades excludes boundary exits. A null return is not zero. The experiment allowed for missing bars while the indicator was flat, and the selection limitation was retained. The bar resolution is not 15 minutes, and the venue validation is not applicable.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the headline about a LINK/USD hourly compression experiment. The experiment was computed on September 8 at 06:19 UTC. It used data from 2025 for training and from January 1, 2026, to September 7 for testing. The strategy used a daily average daily range stop and a 1.5 average daily range target, with a 72-hour horizon. It also used a fee of 25 basis points credited to the asset for buys and to fiat for sells. The stress test had two levels, with the second level being twice as stressful as the first. The results were as follows: for the training period, the return was -83.83%, the drawdown was 85.97%, and the win rate was 41.86%. For the testing period, the return was -48.91%, the drawdown was 56.48%, and the win rate was 42.62%. The experiment also tracked other metrics, including the number of trades, the average win and loss percentages, and the fees paid in US dollars. The results showed that the strategy was valid, but the return was not zero, which is considered invalid. The experiment allowed for missing bars when the indicator was flat, and it retained the selection limitation. The bar resolution was not validated at a 15-minute venue.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the headline about a LINK/USD hourly compression experiment. The experiment was computed on September 8 at 06:19 UTC. It used data from 2025 for training and from January 1, 2026, to September 7 for testing. The strategy used a daily average daily range stop and a 1.5 average daily range target, with a 72-hour horizon. It also used a fee of 25 basis points credited to the asset for buys and to fiat for sells. The stress test had two levels: stress 1 with 5 basis points for majors, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, and stress 2 which doubled these values. The results were as follows: for the training period, the return was -78.3577%, the drawdown was 81.2903%, and the win rate was 43.6782%. For the testing period, the return was -33.1202%, the drawdown was 49.7352%, and the win rate was 46.6667%. The results also included other metrics such as the number of trades, mean trade percentage, average win percentage, and average loss percentage. The experiment allowed for missing bars when the indicator was flat and retained selection limitations. The bar resolution was not 15 minutes and did not validate the venue.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the following data line: Computed on September 8 at 06:19 UTC, frozen note 968. The data is from a 2025 training period and a 2026 January 1 to September 7 holdout period. The next open position will use a daily Average Daily Range, or ADR, stop and a 1.5 ADR target, with a 72-hour horizon. A fee of 25 basis points will be credited to the asset for buys and fiats for sells. The stress test has two levels: stress 1 with 5 basis points for majors, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, and stress 2 which doubles these levels. The results are as follows: for the training period, there is an error due to a missing bar during an open position on October 25, 2025, at 08:00:00 UTC, so the return, drawdown, win rate, and other metrics are unknown. For the holdout period, the return is -55.688904, the drawdown is 60.3423174, the win rate is 32.4324324, and there are 37 trades, with a mean trade percentage of -2.12567654. The average win percentage is 1.87363563 and the average loss percentage is 4.04534639. The quarter Kelly is 0, there are no boundary exits, and the reconciliation in USD is 2.27373675e-13. The buy and hold percentage is -28.0135222, the fees in USD are 112.764695, and the exposure percentage is 28.8083333. The metrics are calculated based on the percentage, and N excludes boundary exits. A null return is considered invalid if it is not zero. A missing bar while the position is flat is allowed with an indicator reset, and the selection limitation is retained. The bar resolution is not 15 minutes, and venue validation is not applicable.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the following data line: Computed on September 8 at 06:19 UTC, a frozen note 968 was recorded for the LTC/USD hourly compression experiment. The experiment used data from 2025 for training and from January 1, 2026, to September 7, 2026, as a holdout period. The next open position was determined by a daily average daily range stop and a 1.5 average daily range target, with a 72-hour horizon and a risk-reward ratio of 1, where risk-reward ratio means comparing the potential profit to the potential loss. A fee of 25 basis points, which is 0.25%, was credited to the asset when buying and to fiat when selling. Two stress tests were applied: stress 1 with a 5 basis point slip for major currencies, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, and stress 2 which doubles these values. The results of the experiment are as follows: for the training period, there was an error due to a missing bar during an open position on October 25, 2025, at 08:00:00 UTC, which made the executable exit unknown, resulting in no return, drawdown, win rate, or other metrics being calculated. For the holdout period, the return was -40.6534932%, the drawdown was 50.1768117%, the win rate was 34.2857143%, and there were 35 trades, with a mean trade percentage of -1.42076084%, an average win percentage of 2.75937367%, and an average loss percentage of 3.60170058%. The quarter Kelly criterion, which is a measure of risk, was 0, and there were no boundary exits, which are exits at a specific price level. The reconciliation in USD was 2.27373675e-13, the buy and hold percentage was -28.0135222%, the fees in USD were 119.037167, and the exposure percentage was 29.6583333%. The experiment notes that percent metrics do not include boundary exits, and an invalid result is one where the return is not zero. A missing bar while the position was flat was allowed with an indicator reset, and the selection limitation was retained. The bar resolution used was not 15-minute venue validation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the following data line: Computed on September 8 at 06:19 UTC, frozen note 968. This is a description of an experiment on the LTC/USD market, which means trading Litecoin for US dollars. The experiment used data from 2025 for training and from January 1, 2026, to September 7 for testing, also known as a holdout period. The next open trade was considered, with a daily stop loss set at 1.5 times the average daily range, or ADR, which is a measure of how much a price moves on average each day, and a target of one times the risk, or R. The time horizon for the trade was 72 hours. A fee of 25 basis points, or 0.25%, was credited to the asset when buying and to fiat when selling. Basis points are units used to measure small percentages. The experiment also considered two types of stress, with the first being 5 basis points for major trades, 30 basis points for XRP-LTC trades, and 15 basis points for UNI-LINK trades, and the second being double these amounts. The results of the experiment are as follows: for the training period, there was an error due to a missing bar during an open position on October 25, 2025, at 08:00:00 UTC, which means the outcome of the trade could not be determined. For the holdout period, the return was -38.16%, the drawdown was 46.29%, and the win rate was 39.29%. The average trade percentage was -1.66%, the average win percentage was 1.62%, and the average loss percentage was -3.78%. The quarter Kelly criterion, which is a measure used to determine the optimal size of a trade, was 0. The experiment also considered boundary exits, which are trades that are closed when a certain condition is met, and reconciliation in US dollars, which was $2.27. The buy and hold percentage, which is the return if the asset was simply bought and held, was -28.01%. The fees in US dollars were $104.22, and the exposure percentage, which is the percentage of the total portfolio that was invested, was 23.98%. The results also included a list of metrics, including the number of trades, which was 28, and the mean trade percentage. The experiment noted that the percent metrics do not include boundary exits, and that a null return is not zero. The experiment allowed for missing bars while the position was flat, with an indicator reset, and retained the selection limitation. The bar resolution was not validated at a 15-minute venue.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the following data line: Computed on September 8 at 06:19 UTC, a frozen note 968 was recorded. The experiment used data from 2025 for training and from January 1, 2026, to September 7 for holdout; the next open position was considered; a daily average daily range, or ADR, stop and 1.5 ADR target were used, with a 1R half and break-even, or BE, point, and a 72-hour horizon. A fee of 25 basis points, or 0.25%, was credited to the asset for buys and fiats for sells; slip stress 1 was 5 basis points for majors, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, and stress 2 was double that. The results were as follows: for the training period, there was an error due to a missing bar during an open position on October 25, 2025, at 08:00:00 UTC, so the return, drawdown, win rate, number of trades, and mean trade percentage were all null; for the holdout period, the return was -28.9443023%, the drawdown was 40.5034142%, the win rate was 39.2857143%, the number of trades was 28, the mean trade percentage was -1.16579117%, the average win percentage was 2.20602833%, the average loss percentage was 3.34755673%, the quarter Kelly was 0, the boundary exits were 0, the reconciliation in USD was 2.27373675e-13, the buy-hold percentage was -28.0135222%, the fees in USD were 110.654274, and the exposure percentage was 24.75%. The percent metrics were calculated, excluding boundary exits, and a null return was considered invalid if it was not zero. A missing bar while the position was flat was allowed, with an indicator reset, and the selection limitation was retained; the bar resolution was not 15-minute venue validation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the following data: Computed on September 8 at 06:19 UTC, a frozen note 968 was recorded. The experiment used data from 2025 for training and from January 1, 2026, to September 7 for holdout; it considered the next open and used a daily Average Daily Range, or ADR, stop and 1.5 ADR target, with a 1R half and break-even, or BE, point, over a 72-hour horizon. A fee of 25 basis points, or 0.25%, was credited to the asset for buys and fiats for sells; slip stress was set at 5 basis points for majors, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, with stress 2 doubling these values. The results were as follows: for the training period, there was an error due to a missing bar during an open position on October 25, 2025, at 08:00:00 UTC, which made the executable exit unknown; for the holdout period, the return was -65.55%, the drawdown was 67.93%, the win rate was 30.43%, and the mean trade percentage was -2.24%. The average win percentage was 2.16% and the average loss percentage was 4.16%. The quarter Kelly was 0, boundary exits were 0, and reconciliation in USD was 2.27e-13. The buy-hold percentage was -28.01%, fees in USD were 123.40, and exposure percentage was 36.93%. Percent metrics were calculated, excluding boundary exits. A null return is not considered zero. A missing bar while flat was allowed with indicator reset, and selection limitation was retained. The bar resolution was not 15-minute venue validation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the following data: Computed on September 8 at 06:19 UTC, a frozen note 968 was recorded. The experiment used data from 2025 for training and from January 1, 2026, to September 7 for holdout. The next open position was considered, with a daily Average Daily Range (ADR) stop and a 1.5 ADR target, plus one risk half to break even, and a 72-hour horizon. A fee of 25 basis points was credited to the asset buy and fiat sell. Two stress tests were applied: stress 1 with 5 basis points for majors, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, and stress 2 which doubles these values. The results were as follows: for the training period, there was an error due to a missing bar during an open position on October 25, 2025, at 08:00:00 UTC, making the executable exit unknown. For the holdout period, the return was -54.5743017, the drawdown was 59.5043103, the win rate was 34.7826087, and the mean trade percentage was -1.64383878. The average win percentage was 2.40983998 and the average loss percentage was 3.80580079. Other metrics included a quarter Kelly of 0, boundary exits of 0, and reconciliation in USD of 2.27373675e-13. The buy and hold percentage was -28.0135222, fees in USD were 136.329792, and exposure percentage was 37.0333333. Percent metrics were calculated, excluding boundary exits. A null return is considered invalid if it is not zero. Missing bars while flat are allowed with indicator reset, and selection limitation is retained. The bar resolution is not 15-minute venue validation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
A news scan found the following data: Computed on September 8 at 06:19 UTC, this is a frozen note 968. The data covers the period from 2025 for training and January 1, 2026, to September 7 for holdout. The next open and daily ADR stop with a 1.5 ADR target, plus 1R half and break-even, all within a 72-hour horizon. A fee of 25 basis points is credited to the asset for buying and fiats for selling. There are two stress tests, with slip stress 1 being 5 basis points for majors, 30 basis points for XRP-LTC, and 15 basis points for UNI-LINK, and stress 2 being double these amounts. The results are as follows: for the training period, the symbol LTC/USD on a 1-hour timeframe with no trend, compression of 1, and stress of 2, the period is not valid due to a missing bar during an open position on October 25, 2025, at 08:00:00 UTC, resulting in an unknown executable exit. For the holdout period, the return was -57.4343925, the drawdown was 62.8182559, the win rate was 37.5, with 48 trades, and the mean trade percentage was -1.70847345. The average win percentage was 2.05402617, and the average loss percentage was 3.96597322. Other metrics include a quarter Kelly of 0, boundary exits of 0, reconciliation in USD of 2.27373675e-13, buy and hold percentage of -28.0135222, fees in USD of 143.123161, and exposure percentage of 38.8833333. The data is valid. Percent metrics are used, excluding boundary exits. A null return is not considered zero. A missing bar while flat is allowed with an indicator reset, and the selection limitation is retained. The bar resolution is not 15-minute venue validation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
ComputedSep8 06:19UTC,frozennote968. 2025train/2026Jan1–Sep7holdout;nextopen;dailyADRstop/1.5ADRtarget,+1Rhalf/BE,72hourhorizon. Fee25bpcreditedassetbuy/fiatsell;slipstress1=5bpmajors/30bpXRP-LTC/15bpUNI-LINK,stress2double. [{"symbol":"LTC/USD","tf":"1Hour","trend":false,"compression":1,"stress":1,"period":"train","valid":false,"error":"Missing bar during open position at 2025-10-25 08:00:00+00:00; executable exit unknown","return":null,"drawdown":null,"win_rate":null,"N":null,"mean_trade_pct":null},{"symbol":"LTC/USD","tf":"1Hour","trend":false,"compression":1,"stress":1,"period":"holdout","return":-38.1287928,"drawdown":46.5206521,"win_rate":44.6808511,"N":47,"mean_trade_pct":-0.956832355,"avg_win_pct":2.48787615,"avg_loss_pct":3.73909692,"quarter_kelly":0,"boundary_exits":0,"reconciliation_usd":2.27373675e-13,"buyhold_pct":-28.0135222,"fees_usd":165.198452,"exposure_pct":39.2916667,"valid":true}]. Percentmetrics;Nexcludesboundaryexit. INVALID=nullreturnNOTzero. Missingbarwhileflatallowedwithindicatorreset; selectionlimitationretained. BarresolutionNOT15minvenuevalidation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
ComputedSep8 06:19UTC,frozennote968. 2025train/2026Jan1–Sep7holdout;nextopen;dailyADRstop/1.5ADRtarget,+1Rhalf/BE,72hourhorizon. Fee25bpcreditedassetbuy/fiatsell;slipstress1=5bpmajors/30bpXRP-LTC/15bpUNI-LINK,stress2double. [{"symbol":"LTC/USD","tf":"1Hour","trend":false,"compression":0.75,"stress":2,"period":"train","valid":false,"error":"Missing bar during open position at 2025-10-25 08:00:00+00:00; executable exit unknown","return":null,"drawdown":null,"win_rate":null,"N":null,"mean_trade_pct":null},{"symbol":"LTC/USD","tf":"1Hour","trend":false,"compression":0.75,"stress":2,"period":"holdout","return":-48.3896506,"drawdown":54.7315314,"win_rate":40.4761905,"N":42,"mean_trade_pct":-1.5138664,"avg_win_pct":1.74950101,"avg_loss_pct":3.73295624,"quarter_kelly":0,"boundary_exits":0,"reconciliation_usd":2.27373675e-13,"buyhold_pct":-28.0135222,"fees_usd":141.022316,"exposure_pct":34.3833333,"valid":true}]. Percentmetrics;Nexcludesboundaryexit. INVALID=nullreturnNOTzero. Missingbarwhileflatallowedwithindicatorreset; selectionlimitationretained. BarresolutionNOT15minvenuevalidation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees
ComputedSep8 06:19UTC,frozennote968. 2025train/2026Jan1–Sep7holdout;nextopen;dailyADRstop/1.5ADRtarget,+1Rhalf/BE,72hourhorizon. Fee25bpcreditedassetbuy/fiatsell;slipstress1=5bpmajors/30bpXRP-LTC/15bpUNI-LINK,stress2double. [{"symbol":"LTC/USD","tf":"1Hour","trend":false,"compression":0.75,"stress":1,"period":"train","valid":false,"error":"Missing bar during open position at 2025-10-25 08:00:00+00:00; executable exit unknown","return":null,"drawdown":null,"win_rate":null,"N":null,"mean_trade_pct":null},{"symbol":"LTC/USD","tf":"1Hour","trend":false,"compression":0.75,"stress":1,"period":"holdout","return":-33.5702839,"drawdown":41.1462169,"win_rate":45.2380952,"N":42,"mean_trade_pct":-0.918410415,"avg_win_pct":2.19608631,"avg_loss_pct":3.49125554,"quarter_kelly":0,"boundary_exits":0,"reconciliation_usd":2.27373675e-13,"buyhold_pct":-28.0135222,"fees_usd":161.978639,"exposure_pct":34.5333333,"valid":true}]. Percentmetrics;Nexcludesboundaryexit. INVALID=nullreturnNOTzero. Missingbarwhileflatallowedwithindicatorreset; selectionlimitationretained. BarresolutionNOT15minvenuevalidation.
https://runway.adityaperswal.workers.dev/data/crypto-bars ; https://docs.alpaca.markets/us/docs/crypto-fees