Stats

The automated trading fund manager has released a signal table at 04:00 UTC. The table is based on three premises: using daily bars from a paginated provider with stocks SIP adjustment through September 18 and crypto completed through September 19, calculating 252/365 interval close return, SMA200, arithmetic ATR20, and P90 rolling60 endpoints trailing 5 years, and using an empty ledger with no queue from the supplied context. The data is sourced from https://runway.adityaperswal.workers.dev/data/stock-bars and /data/crypto-bars. The full table shows the following assets: SPY, TLT, EEM, EFA, GLD, IWM, QQQ, BTC/USD, and ETH/USD. The table includes information such as date, bars, pages, close price, annual return, SMA, ATR, P90, and rank. Based on the above-SMA ranking, EEM, QQQ, and IWM are selected. GLD and TLT fail the SMA test, and crypto ranks 6 and 7. The conditional sizes are 93.33 each, but they have not been executed. No strategy retest is required under the fixed mandate.

The final decision for R92002 is to broaden the census, which yields chemical and memory watches, but no funded entry will be made. The process started with a $120 USD supply cash and $1000 account equity, with no existing positions or orders. The goal was to reach a price that reflects an industrial bottleneck rent before guidance. Several premises were considered: Laxmi's notices on August 27, September 7, and September 15 reported allocations directly, and Celanese's Q2 results already captured a 22% sequential acetyl price increase, but August guidance forecasts moderation, and the company must deliver $543-643 million in free cash flow in the second half of the year after $157 million in the first half. Additionally, IMOS's August sales increased 33.3% year-over-year but decreased 1.3% month-over-month, with the top 5 customers accounting for 61% of sales, and non-binding forecasts and limited suppliers were noted. The net 3R hurdle of 75.185 implies an annualized Q2 EPS of 23.50x, which is not considered fair value. Frozen long-only held apart tests were conducted on CE, IMOS, BTDR, and AMTX, with results showing a 25.12% increase for CE, a 12.016% increase for IMOS, a 14.123% increase for BTDR, and a 7.886% increase for AMTX. However, AMTX fell 6.863% with a 55 basis point per side change. The analysis also considered BTDR's future customer receipts and AMTX's secured creditor claims, which constrain who owns revenue. The validity of the premises relies on the idea that a shortage plus operating leverage can create cash, but it does not necessarily entail shareholder upside at today's price. The soundness of the premises is weakened by issuer-biased forecasts, region mismatch, cash conversion, and selected historical windows. A counterexample is that sustained September allocations could defeat CE's August moderation forecast, and the memory mix could raise IMOS EPS enough to support value. The sizing of the risk fractions was estimated, but the actual size is 0, and no order will be made until the target value and current entry qualify. The binding constraint is supported by retained cash and payoff, not whole-share affordability alone. No new manager book is warranted, and industrial and power capacity leads fit the supply. The census includes all US directories with named coverage exclusions, 23 overseas samples, and 12 crypto samples, but is not a complete global market. Research was stopped because additional issuer pages repeat known claims, and paid margin and commitment gaps, as well as crypto-causality gaps, were named. Cash was chosen at $120, not the default. Existing monitors were retained, and some were resolved or armed. The process was verified, and 47 runs were left.

The premises of this analysis are as follows: first, a directory census of 13220 non-test US entries was conducted, with 12836 queried after 384 preferred-symbol exclusions, resulting in 12420 usable paired IEX snapshots and 8133 fresh entries as of September 18. This includes 3430 fresh non-ETFs with above $100k partial day dollar volume. The data was sourced from the Nasdaq trader website and a stock snapshots website. However, this is not a full SIP liquidity or OTC census. Second, 23 selected overseas stocks across Japan, Taiwan, Korea, China, India, Europe, and Australia were screened via Yahoo chart 1yr, with RSGN added after the initial 22. Only 8150.TW passed the prior 20 close/SMA 50, with different as-of dates and raw returns not FX-adjusted relative value. Third, twelve crypto pairs were refreshed as of September 19, with ETH, AVAX, AAVE, and LTC passing the price gate, while UNI/SOL was no longer above the prior 20 close. The data was sourced from a crypto bars website. No demonstrated token claim to industrial shortage rent was found, and no crypto trade was conducted. Fourth, FEAM prior funding/backtest rejection was preserved, with SGML price below trend, FPS shortlisted but with a short seasoned history, SNDK share unaffordable and spin-history short, AMTX allocation means procurement, BTDR future energized capacity requires funding, and CE/IMOS primary leads. In conclusion, indicators prioritize research, not independent confirmations. Price transformations are correlated, and issuer PR and filing are one origin. The analysis has several misses, including a non-US sample that is not all global listings, OTC/pref exclusions, IEX coverage, short IPO histories, and no fresh independent paid-fixture or crypto protocol cash study. These gaps are named, not filled by inference.

The premises for the R92002 fund are as follows: on September 18, the daily adjusted SIP was analyzed. The stock prices for CE, IMOS, BTDR, and AMTX were examined. CE's price was 45.41, which is below the prior 20-day average of 47.22 and near the Simple Moving Average (SMA) 50 of 45.40. IMOS's price was 61.30, which is higher than 58.39 and 56.99. BTDR's price was 12.97, which is higher than 12.67 and 11.01. AMTX's price was 2.16, which is higher than 2.07 and 1.75. A breakout means the prior 20-day closing high, not the intraday high. The Average True Range (ATR) for 20 days was calculated as 1.69, which gives 2 ATR stops at 42.03, 56.99, 11.11, and 1.90. The costed net 3R targets are 56.29, 75.18, 18.75, and 2.96. These targets are required prices, not supported valuations. RSGN's price was 17.40, which is lower than the prior 20-day average of 20.56 and the SMA 50 of 19.65. The daily close source is https://query1.finance.yahoo.com/v8/finance/chart/RSGN.SW?range=1y&interval=1d. In conclusion, there was no actual bracket adjustment because there were no own holdings. The price gates for CE and RSGN were false, while the other three price gates were true, but the fundamental and test gates failed. BTDR is in warm-up mode due to its operating history only, and FPS's SMA 200 is missing, not zero. There is no tape signal that can substitute for the bottleneck cash.

The premises of this analysis are based on two main points. First, the analysis uses a fixed long-only chronological holdout test from 1989 to 1992, which utilizes data from https://runway.adityaperswal.workers.dev/data/stock-bars. Second, the win rate, average win, and average loss percentages are as follows: CE 17.647%, 21.6997%, and 6.0195% with 17 trades, IMOS 29.1667%, 16.6557%, and 6.7616% with 24 trades, BTDR 33.3333%, 42.9153%, and 16.8298% with 24 trades, and AMTX 33.3333%, 41.2324%, and 15.0053% with 21 trades. The Half Kelly risk fractions, which represent the stake-at-risk fractions, not capital allocation, are calculated as 0, 0.002054, 0.035946, and 0.045360. Additionally, the 20000 IID bootstrap mean confidence intervals all cross zero: CE [-5.833, 4.665]%, IMOS [-4.275, 4.664]%, BTDR [-8.060, 15.665]%, and AMTX [-7.657, 16.539]%. It is noted that average positive trade returns can coexist with negative compounded returns, and the arithmetic mean cannot be used as a proxy for profit and loss. The conclusion is that uncertain regime-dependent distributions do not establish an actual execution edge. There are no live post-reset supply trades to estimate slippage, and costs are based on assumptions. Candidate selection uses current winners, and historical windows are not untouched prospective evidence. The actual size is zero because the cash value and current rule gates fail, not because uncertainty itself forbids a convex probe.

A diagnostic test was run using the command line interface, which is a way to interact with a computer program using text commands. The test used a node, a type of computer program, to run a backtest, which is a simulation of a trading strategy, on the stocks CE, IMOS, and BTDR. The strategy used was a breakout strategy, which means it looks for stocks that are breaking out of a certain price range, and it looked back at 20 days of data and ran the test for 1460 days. The test also used a stop loss, which is a rule to sell a stock if it loses a certain percentage of its value, in this case 8 percent, and a target profit, which is a rule to sell a stock if it gains a certain percentage of its value, in this case 24 percent. The test got its data from a website, https://runway.adityaperswal.workers.dev/data/stock-bars, and used 1002 bars of data, where a bar is a piece of data that shows the price of a stock at a certain point in time. This test also exits if the signal, or the reason to buy or sell a stock, gets weaker. This test is different from a custom test that uses a different rule to buy and sell stocks. The test did not hold out any data to test its accuracy and did not account for costs, so it is only for diagnostic purposes and its results should not be used to make actual investment decisions. The stock BTDR includes data from before it started operating, and a custom test that only looks at data from when the company was operating gives different results. The full results of the test are in a file called cli_test.json.

A diagnostic test was run using the command line interface, which is a way to interact with a computer program using text commands. The test used a node, which is a computer program, to run a backtest, which is a simulation of a trading strategy, on the stocks CE, IMOS, and BTDR. The strategy used was a breakout strategy, which means buying a stock when its price breaks out of a certain range, and it looked back at 20 days of data and ran for 1460 days. The test also used a stop loss, which is a rule to sell a stock when it loses a certain percentage of its value, in this case 8%, and a target profit, which is a rule to sell a stock when it gains a certain percentage of its value, in this case 24%. The test got its data from a website and used 1002 bars of data, which are like buckets of time. This test also exits when a signal, which is like a hint to buy or sell, gets weaker. This test is different from a custom test that only buys and holds stocks for a long time. The test did not set aside any data to check its results and did not consider any costs, it was just a diagnostic test and its results were not used to make any decisions. The stock BTDR includes data from before it started operating, but a custom test that only looks at data from when it started operating gives a more accurate result. The full results of the test are in a file called cli_test.json.

A diagnostic test was run using the command line interface, which is a way to interact with a computer program using text commands. The test used a node, a type of computer program, to run a backtest, which is a simulation of a trading strategy, on the stocks CE, IMOS, and BTDR. The strategy used was a breakout strategy, which means it looks for stocks that are breaking out of a certain price range, and it looked back at 20 days of data. The test ran for 1460 days and used a stop loss, which is a rule to sell a stock when it falls by 8%, and a target profit, which is a rule to sell a stock when it rises by 24%. The test got its data from a website, https://runway.adityaperswal.workers.dev/data/stock-bars, and used 1002 bars of data, which are like rows of data. This test also stopped when the signal, or the reason to buy or sell, got weaker. This test is different from a custom rule that only buys and holds stocks for a long time. The test did not separate the data into training and testing sets, and it did not account for costs. It was just a diagnostic test and not a real trading strategy. The stock BTDR includes data from before it started operating, and a custom test, number 3591, has more accurate results. The full results of the test are in a file called cli_test.json, located in a temporary directory.

A test called R92002 was done to see how sensitive some trading rules are. The test used daily stock prices from a website and made 20,000 fake paths for each stock over 5-day blocks. The test looked at all the stock's history and the last 252 and 60 sessions. The test used fixed price ranges for four stocks: CE with a low of 45.41, a middle of 42.027, and a high of 56.2870, IMOS with a low of 61.30, a middle of 56.99918, and a high of 75.1852, BTDR with a low of 12.97, a middle of 11.11183, and a high of 18.75245, and AMTX with a low of 2.16, a middle of 1.90454, and a high of 2.96101. The test tried to make a profit of 3% after fees and slippage. The test found that the average return for all the stocks was CE 0.682%, IMOS 1.156%, BTDR 4.464%, and AMTX 0.208%. The test also found that the probability of hitting the target was CE 18.835%, IMOS 23.395%, BTDR 32.685%, and AMTX 25.695%. The test stopped if both targets were crossed, and the stop loss was between 61.945% and 73.9%. The test found that even if there are some big losses, the trading rules cannot be calibrated to make a profit. The test used a method called Half Kelly to manage the money, but it was limited to between 0.03 and 0.5. The test found that buying one share of IMOS with a 50% gap would cost 30.834, which is more than the normal risk budget. A news scan found the note was about a 160,000 path test, not 60,000.

The premises for rejecting revenue allocation and energized capacity as immediate shareholder cash are as follows: 1. Bitdeer's August operating update reports $86m USD annualized last-day cloud contracted revenue, 92% GPU utilization, and over $800m expected A1 revenue beginning Q1 2027, which is not current cash. 2. The Q2 operating cash flow is -$158.523m, and after subtracting $266m capital expenditures, it becomes -$424.523m. Adding $195.5m digital asset sales gives an economic funding proxy of -$229.023m, while $428.9m financing supported cash. 3. Aemetis' 18m-liter procurement allocation is approximately 23.775% of its annualized 80m-gallon capacity, indicating no capacity shortage. The company's half-year free cash flow is -$27.522m, and debt sweeps mean creditors can capture improved receipts before shareholders. 4. Rands Group's backlog is 357.9m CHF, with a 17.1% increase, but order intake is down 10%, organic sales are down 9%, free cash flow is $11.5m, and net debt is $71.4m. A September 16 publication confirms earlier August 5 figures, indicating no fresh order acceleration. In conclusion, it is recommended to watch contracting and collections and reject current entries. However, a countercase could be that prepaid customers and funded commissioning could transform Bitdeer's cash, and Aemetis' successful volume and credit monetization could reduce leverage. The binding constraint is financing plus retained margin, not headline revenue. Currently, there is no calibrated target or zero-loss assumption.

A news scan found the headline 'R92002 decision: closed-session cash with two new gross-tail candidates and explicit execution gaps'. The automated trading fund manager started with 100 USD in cash and 1000 in equity, with the goal of buying qualified convex long options. The manager looked at 12836 US symbols and found 8 breakouts, including HALO Oct 16 C 130 with an ask price of 0.23 and a potential payout of 1.83266, which is 7.968 times the premium, and TH Oct 16 C 27 with an ask price of 0.18 and a potential payout of 3.58112, which is 19.895 times the premium. These two options passed the gross gate, but the quotes were indicative and the bids were lower than the ask prices, which means the actual execution risk is unresolved. The manager also looked at costed stock backtests and Monte Carlo simulations, but they are not actual option fills and the results can change depending on the sample, horizon, and regime. The manager rejected SFIX because its earnings rejection was 28 USD, which is higher than the median of 16.578%, and HZO because its cash deal invalidates its standalone history. The manager decided not to place any orders and to keep 100% in cash because the authorized options session was closed, not because all candidates failed. The mandate breach counter did not advance because this was a closed research run. On Monday at 09:50, the manager will monitor HALO and TH and refresh the capital, actual ask, bid, underlying, and eligibility, and choose the cheapest option that passes the test. The manager will also recompute the P80, hurdle, and half Kelly levels if there are any changes in the spot price. The manager rejected THP 18 because its payoffs were all 0 and HALO put MC because it prefers call watch. The manager also retained four prior collectors and tested 283 new ones, with remote verification pending. The manager saved all 1595 complete US signal rows and 127 observation batches, and found no qualifying high breaks in international and crypto markets. There were gaps in the data, including recent SIP snapshot access, low IEX liquidity, and exhaustive global, FDA, and court coverage. The manager did not propose any new funds and did not have any open trades or holdings. The sources of the data include https://runway.adityaperswal.workers.dev/data/stock-bars and /data/option-chain/HALO?feed=indicative and TH/SFIX.

A news scan found the following text: Premises for this analysis are based on a hypothetical scenario where the spot price on September 21 remains unchanged, with 20 daily steps to October 16, which is 25 calendar days, and 30,000 paths per scenario with a seed of 92002. The analysis uses a method called Bootstrap to estimate daily returns, which are scaled at 0.75, 1, and 1.5, and a constant volatility, known as BSMIV, calibrated to the ask price and a 4% interest rate. The results show that for certain stocks, including HALO and THC, the paid premiums and implied volatility percentages are as follows: HALO C130 paid 0.23 with a fit IV of 33.414%, HALO P95 paid 0.18 with a fit IV of 38.114%, and THC 27 paid 0.18 with a fit IV of 68.513%. The analysis also shows that the RV60 for HALO is 45.30 and for THC is 58.40%. The options stop at 50% and the modeled targets are filled no better than the target, with an exit friction of 0.01 per share plus a 1 USD roundtrip fee. The mean premium targets for HALO and THC are -8.90, -3.58, and 1.78 USD, and -7.89, -4.71, and 0.02, respectively. The P80 targets, which are independently derived, show mean values of -8.27, 1.42, and 16.26 for HALO, and -3.54, 7.28, and 28.75 for THC. The baseline target hits are 10.657% and 4.557%, and the stops are 89.257% and 91.873%. The analysis concludes that the whole premium should be kept at 23 or 18, which is less than 30 each, with a potential total of 41, which is less than 100, and a maximum of 4. The worst-case scenario is 41 plus costs, despite half premium stops. The terminal intrinsic Kelly is 49.7 and 235.25 USD, which is unreliable, and a hard cap of 30 binds. The prospective single contracts are minimum granularity, with a quarter of 30 and a probe of 7.50, which is unavailable. No orders were closed, and the capital and actual quote should be refreshed in the entry run. The source of the data is /data/stock-bars and /data/option-chain/HALO and THC, with a feed of indicative, from 30184 to 89.

A news scan found the following analysis: Premises are based on several indicators, including a 20-day momentum of 5.965% for HALO, a distance from the 50-day Simple Moving Average (SMA) of 17.141%, an RV of 21.407%, and a volume of 6.931 times the average, as well as TH indicators of 0.193%, 21.967%, 68.918%, and 2.420 times. The correlation between momentum and SMA over the last 252 sessions is 0.8851 for HALO and 0.8977 for TH, meaning they mostly follow one trend. A separate train using Ordinary Least Squares (OLS) regression, which is a method to predict values, was done to predict the next open-to-20 close returns using known features from the signal day. For HALO, the train had 14 sessions and the test had 3 sessions, with Mean Absolute Errors (MAE) of 21.8884 for momentum, 20.7803 for SMA, 20.6780 for RV, and 20.8975 for volume, compared to a constant value of 20.7220. This does not show a convincing improvement. For TH, the train had 2 sessions and the test had 2 sessions, with perfect correlations that are considered mathematical artifacts due to only having 2 points, and test MAE of 102.215, 39.993, 51.890, and 30.510, compared to a constant value of 28.842. The conclusion is that there is no increased confidence or size from correlated indicators or a small regression. The current breakout qualifies the universe and the mandate payoff will be tested separately, with frozen features retained for prospective observation. The source of this analysis is https://runway.adityaperswal.workers.dev/data/stock-bars, which was measured on days 30182–83.

The automated trading fund manager made certain premises for the R92002 costed breakout controls and sparse historical-option audit. These premises include a fixed 252 intraday high plus 1.5 volume rule, and a vectorbt 30 USD stock per entry with 1000 book, next open entries, 10 basis point fee plus 10 basis points slip on each side, and a 20 session exit. The training period was from 2016 to 2024, and the holdout period was from 2025 to September 18, 2026. The results for HALO train were a return of 0.49139%, a drawdown of 0.96270%, a win rate of 50%, and 14 trades. For the holdout period, the results were a return of -0.468923%, a drawdown of 1.254913%, a win rate of 33.333%, and 3 trades. The results for TH train were a return of 0.81828%, a drawdown of 0.71195%, a win rate of 50%, and 2 trades. For the holdout period, the results were a return of 1.159881%, a drawdown of 0.469995%, a win rate of 50%, and 2 trades. The results for NEO holdout were a return of -0.191897%, a drawdown of 0.43803%, a win rate of 0%, and 1 trade. A sweep of 10, 20, and 30 sessions and 50 basis points slip recorded from 30136 to 46, and the 20 session stress test results were -0.53919% for HALO and 1.10714% for TH. There was no best holdout selection. The CLI lookback used 252 days, 2500 stop, and 5 target 15, and the results were a return of 15.1221% and a drawdown of 9.515 for HALO, a return of -7.9152% and a drawdown of 22.616 for TH, and a return of -6.0274% and a drawdown of 15.971 for NEO. The CLI uses a different exit policy and is for diagnostic purposes only. Historical option bars were used to illustrate 8 OCC queries, and 2 were empty. The queries included HALO May 8, 2025, P60 open 0.80 oversized, and August 7, 2026, C100 open 3.93/P70.60 oversized. The source of the data is https://runway.adityaperswal.workers.dev/data/stock-bars and https://runway.adityaperswal.workers.dev/data/option-bars. No trade was made during this research run.

A news scan found the headline 'R92002 new breakout calls pass coarse tail gate; corrected next-open samples and signed odds'. This is a report from an automated trading fund manager. The manager uses a rule called 'fixed volume-backed 252 high rule' to make decisions. This rule looks at the past, from 2016 to September 18, 2026, and makes a 'buy' or'sell' signal based on certain conditions. The manager also looks at 'histories' of stocks, including HALO, TH, and NEO. For HALO, the manager found 17 events with a median absolute change of 10.5263% and a probability of 80% of 17.2367%. For TH, the median absolute change was 27.3899% and the probability of 80% was 44.3186%. The manager also looked at 'intrinsic' values, which are like built-in worth, and 'P80' values, which are like 80% confidence levels. For example, for HALO, the intrinsic value was 1.83266 and the P80 value was 7.9681x. The manager made decisions based on these numbers and found that some 'call' and 'put' options were good candidates for further review. A 'call' option is like a bet that a stock will go up, and a 'put' option is like a bet that a stock will go down. The manager also used a 'half Kelly' strategy, which is a way of managing risk. The manager found that the'mapped mean intrinsic' values for HALO and TH were 1.72258 and 1.86638, respectively. The manager also did a 'bootstrap' analysis, which is like a simulation, and found that the results were'sparse and unstable'. The manager concluded that HALO and TH 'call' options were good candidates for further review, but the options were closed, so the manager decided to hold cash. The manager also noted that the 'breakout' signal was not a scheduled event, but rather a signal that happened on a certain date. The manager used data from'stock bars' and 'option chains' to make these decisions.

This is a report on a trading strategy test. The test used daily stock prices from September 18, 2016, available at https://runway.adityaperswal.workers.dev/data/stock-bars. The strategy had two main rules: it would buy when the closing price was higher than the maximum price of the last 20 days and the 50-day average, and it would sell when the price reached a target or stopped losing money after 60 sessions. The test also considered costs such as fees and slippage, which are costs associated with buying and selling, and stress, which refers to the impact of unexpected events on the strategy. The test results showed a return of 59.0565% from 2016 to 2022 and 7.8856% from 2023 to September 18, 2026, with a maximum loss of 69.1088% and a win rate of 33.3333%. However, when stress was applied, the return became negative at -6.8632%. A news scan found the conclusion that the strategy has a positive baseline but is sensitive to costs, and stress can make it negative. The test was done on historical data and does not represent a current investment recommendation.

This is a report on a trading strategy called R92002 BTDR long-only breakout. The strategy is based on certain premises: it uses daily high, low, open, and close prices from September 18, 2016, available at https://runway.adityaperswal.workers.dev/data/stock-bars. The strategy has rules such as a stop loss set at 2 times the average true range over 20 days, and a target set at 6 times the average true range over 20 days, with a time exit after 60 sessions. The strategy also takes into account opening gaps and daily path ambiguity. There are costs associated with buying and selling, including a 5 basis point fee and a 15 basis point liquidity cost on each side, with an additional stress test of 5 and 50 basis points. The strategy uses a vectorbt model with long-only direction and independently compounded trade returns. The results show a sensitivity test with L15, L20, and L30, and no best-cell selection. The training data is from 2016 to 2022, but it is not available for BTDR, which only started operating on April 14, 2023. The test results from 2023 to September 18, 2026, show a return of -14.1235%, a maximum drawdown of 75.7896%, a win rate of 33.3333%, and 24 trades, with a stress test result of -27.4041%. The 2026 subset shows a return of 7.1773% with 5 trades. The conclusion is that the strategy shows a baseline long-horizon compounded loss, and any positive recent regime should be preserved, but it does not infer a durable edge. This test does not recommend adoption or ordering, and the historical 3R bracket alone is not a defended current valuation.

This is a report on a trading strategy called R92002 IMOS long-only breakout. The strategy is based on certain premises. Firstly, it uses daily high, low, open, and close prices from September 18, 2016, available at https://runway.adityaperswal.workers.dev/data/stock-bars. Secondly, it freezes the close price of the last 20 days if it is higher than the maximum price of the prior 20 days and the 50-day simple moving average, which is a calculation to smooth out price data over a period, and then it stops the trade at two times the average true range, which is a measure of price volatility, over 20 days, and targets six times the average true range over 20 days, with a time exit after 60 sessions, considering opening gaps, which are differences between the current day's price and the previous day's price. The strategy also honors stop losses, which are orders to sell when a stock falls to a certain price, first when the daily price path is ambiguous. Buying and selling costs are 5 basis points fee plus 15 basis points liquidity cost each side, with a stress test of 5 plus 50 basis points. Brackets, which are price ranges, are frozen when stressing costs. The strategy does not model partial or trailing management, which are methods to adjust the size of a trade, and it uses daily bar data as a proxy, not 15-minute execution data from a specific venue. The strategy uses vectorbt, a backtesting library, to generate orders with a long-only direction, on two marks per day, and independently compounds trade returns, which means it calculates the returns of each trade separately. The strategy was trained from 2016 to 2022 and resulted in a loss of 8.9717%, with 45 trades, and from 2023 to September 18, 2026, it resulted in a loss of 12.0162%, with a maximum drawdown of 38.5439%, a win rate of 29.1667%, and 24 trades, with a stress test result of -25.6227%, and a subset result of -2.5298% with 8 trades. The conclusion is that the strategy has a baseline long-horizon compounded loss, and it is recommended to preserve any positive recent regime, but not to infer a durable edge, which is a lasting advantage. There is no adoption or order from this test, and the historical 3R bracket alone is not a defended current valuation.

A news scan found the following text: Premises for the R92002 CE long-only breakout strategy are as follows: 1. The strategy uses daily high, low, open, and close prices from September 18, 2016, available at https://runway.adityaperswal.workers.dev/data/stock-bars. 2. The strategy freezes the close price above the prior 20-day maximum and the 50-day simple moving average, which is a calculation of the average price over a 50-day period, and then uses the next open price. It stops at 2 times the average true range, a measure of price volatility, over 20 days, targets 6 times the average true range over 20 days, and exits after 60 sessions. The strategy also honors opening gaps, which are large price movements at the start of trading, and stops first when the daily price path is unclear. Buying and selling costs include a 5 basis point fee and a 15 basis point liquidity cost, with an additional stress cost of 5 and 50 basis points. Brackets are frozen when stressing costs. The strategy does not use a 2R partial or trailing management model, uses daily bar data as a proxy, and does not account for 15-minute execution. 3. The strategy uses vectorbt, a backtesting library, to simulate trades with a long-only direction, on two marks per day, with independently compounded trade returns that reconcile exactly. The strategy discarded an initial default-direction shorting bug before records and did not select the best cell. The strategy was trained from 2016 to 2022 and resulted in a -21.9553% return, with 35 trades. From 2023 to September 18, 2026, the strategy resulted in a -25.1199% return, with a drawdown of 38.5325%, a win rate of 17.6471%, and 17 trades. The stress test resulted in a -33.5209% return, and the 2026 subset resulted in a -6.0643% return with 4 trades. The conclusion is that the strategy has a baseline long-horizon compounded loss, and any positive recent regime should be preserved, but no durable edge should be inferred. There is no adoption or order from this test, and the historical 3R bracket alone is not a defended current valuation.

The R92002 breakout is based on the fundamentals of several companies. Firstly, HALO's June 30 filing showed $420.969 million in operating cash flow for the first half of the year, compared to $253.932 million, and $231 million in cash and marketable securities. The company also reported an average mid-single-digit royalty from partner sales. On August 6, the issuer reported Q2 revenue of $481 million and royalties of $307.7 million, with an adjusted EPS guide of $8.65-$9, implying a price-to-earnings ratio of 12.49-13.0x at $112.45. While recurring paid royalties support the operating case, milestones, patents, partner concentration, and differences between adjusted and GAAP earnings constrain valuation. Secondly, TH reported Q2 revenue of $85.5 million, a 39% increase, and a net loss of $9 million, with adjusted EBITDA of $18.2 million. The company's operating cash flow for the first half of the year was $111 million, dominated by customer advances. However, the $1.4 billion multi-year awards and 9,000 beds are not annual revenue, and the 20,000 bed pipeline is not comprised of contracts. On September 10, the company issued 16.1 million shares at $18.50, with proceeds going to selling holders, not corporate cash. Thirdly, NEO reported Q2 revenue of $202 million, an 11% increase, and clinical revenue of $187 million, a 14% increase, with net income of $2 million, including an $11 million debt gain, and operating cash flow of $20 million. On August 26, the company announced additional coverage for its RaDaR test, with two more indications pending. Lastly, HZO agreed to a cash deal of $53 versus a stock price of $52.56. The validity of the breakout is supported by paid activity, but does not prove directional probability. Countercases include HALO royalty setbacks, TH advances funding future costs, NEO coverage without profitable volumes, and HZO competing bids or termination. The goal is to start with $100 cash and aim for small qualified long options, with binding constraints of current ask less than or equal to $0.30 and actual historical comparable payoff. No new dated events were invented, and HALO and TH were considered under technical breakout mandate. Sources include observations 30001-4 and NEO's press releases.

The R92002 fund manager has identified bottlenecks in four areas: chemicals, memory, power, and transformers. 1. In chemicals, a delay in imported inputs restricts Laxmi volume, which may lead to integrated CE winning supply share and pricing, affecting receivables, energy, and substitution before equity cash. Sources include Laxmi's customer information and CE's SEC filing. 2. In memory, AI demand increases assembly and test utilization, which may lead to ChipMOS price and margin recovery, but scarce equipment and materials, along with 61% customer concentration, may compete away gains, and new capital expenditures may eventually relieve constraints, as seen in ChipMOS's SEC filing. 3. In power, scarce energized land enables BTDR contracts, but GPU and build capital expenditures require prepayments, debt, or equity, and future depreciation, customer concentration, and hardware obsolescence determine owner cash, with more mining hashrate not creating BTC pricing power, as stated in Bitdeer's August 2026 production and operations update. 4. In transformers, the backlog extends to 2028, with RSGN hiring and capital expenditures ahead of deliveries, and backlog growth can coexist with lower organic revenue and margin, with installed capacity eventually easing industry pricing, as seen in RSGN's semi-annual report. These facts establish constraints and accounting, with arrows representing conditional causal hypotheses, and the manager rejects the simple'shortage = buy suppliers' shortcut, noting independent origins for Laxmi, CE, ChipMOS, BTDR, and RSGN, and that publications within one issuer are not corroboration.

A report from September 10 showed that ChipMOS' revenue was NT2785.7 million, which is a 33.3% increase from the same time last year and a 1.3% decrease from the previous month. The company cited a persistent imbalance between artificial intelligence and memory, which can be found at https://www.chipmos.com/english/news/news-detail.aspx?NID=234. On August 11, the company released its Q2 results, which included a gross margin of 18% compared to 13.8% in Q1, and an earnings per share of $0.80 USD/ADS. The company also reported a free cash flow of $23.1 million for the first half of the year, as seen at https://chipmostechnologiesinc.gcs-web.com/news-releases/news-release-details/chipmos-reports-second-quarter-2026-results-record-high. Additionally, the company's 20-F filing on April 14 showed that its top 5 customers make up 61% of its sales, with the top 3 customers making up 23%, 15%, and 9% respectively. The filing also stated that the company's suppliers have limited or no long-term contracts and can switch to other companies in 3-6 months, which can be found at https://www.sec.gov/Archives/edgar/data/1123134/000119312526153743/imos-20251231.htm. Based on this information, it appears that the company's growth and margin recovery are real, but its suppliers and large customers have the power to capture rent. The share price of $61.30 implies a price-to-earnings ratio of 19.16 times the annualized Q2 EPS of $3.20, and 24.52 times the annualized H1 EPS of $2.50. These are illustrative run-rates and not forecasts. The company's strong countercase is that continued memory pricing and utilization can accelerate earnings beyond Q2. The Taiwanese price of 94 TWD and the US ADS price cannot be arbitraged due to stale FX and asynchronous closes. The annual filing and monthly PR are from the same issuer origin, and the Taiwan and US price series have the same stock exposure.

The trading fund started with a price of CE45.41 and no position. The goal was to capture extra acetyl cash before guidance was issued. There were several premises for this decision. Firstly, Laxmi's notices on September 7, 2015, maintained allocation and immediate-payment pricing, which can be found at https://www.laxmi.com/documents/default.aspx?f=customer-info/price-announcement-1049.pdf. The prices of three products rose by 6.67%, 11.76%, and 14.05% in eight days, while n-propyl fell by 12.79%, which means the claim of uniform pricing is not true. Secondly, Celanese's Q2 disclosure report on August 4 showed a 22% increase in acetyl price and EBIT of 321 million, but the Q3 EPS of 1.35-1.75 assumes moderation, which can be found at https://investors.celanese.com/news/celanese-corporation-reports-second-quarter-earnings/4f485280-d86f-43f3-8068-de01ca04a02c. Different dates and regions can make both of these statements true, as Indian landed shortages do not necessarily mean higher Western Hemisphere realized netbacks for Celanese. Thirdly, on June 30, the debt was 12,007 million and cash was 1,364 million, resulting in a debt of 10,643 million, and the free cash flow for the first half of the year was 157 million, with accounts receivable consuming 399 million, which can be found at https://www.sec.gov/Archives/edgar/data/1306830/000130683026000117/ce-20260630.htm. The forecast for the full year is 700-800 million in free cash flow, which requires 543-643 million in the second half, or 3.46-4.10 times the first half. Fourthly, the 2025 risk factors show that Celanese is the sole source of CO, steam, and ethylene, and has a 50%-owned Fairway methanol concentration, which can be found at https://www.sec.gov/Archives/edgar/data/1306830/000130683026000031/ce-20251231.htm. In conclusion, the real bottleneck to watch is the delivered CE margin and collections, not just a shortage headline. A counterexample is Celanese's own supplier outages in 2024, which made integration a liability. Previously, there was moderate evidence of renewed rent, but a new issuer volume, netback, and cash bridge would be needed to increase confidence, as letters alone are not enough. At a price of 45.41 and a forward adjusted FY6 EPS of 7.57x, the stock is not standalone cheapness or a justified target, assuming debt and an 8% tax assumption. Laxmi's June quarter EBITDA margin of 11.8% contradicts the blanket victim classification, but precedes the July force majeure. The pending sale of Nutrinova for 152 million is financing and deleveraging, not recurring free cash flow.

A news scan found the headline 'R92002 exchange-directory screen finds eight volume-backed 52-week highs'. Premises: The Nasdaq Trader nasdaq listed and other listed directories contain 13,221 non-test rows, including one malformed symbol. After preferred-format batch failures were repaired, 12,836 alphabetic and dotted symbols were requested. IEX snapshots of 12,689 and 12,420 were taken, with usable current and prior bars. 1,643 non-ETF names with a price of $1 or more and reported dollar volume of $2 million or more received historical SIP bars. All 1,643 histories were returned, and 1,595 had 253 or more bars. A close on September 18 above the prior 252 intraday highs and volume of 1.5 times the prior 20-day mean, and prior 20-day dollar volume of $2 million or more, yielded eight results: HALO at $112.45 with a volume of 6.931 times, NEO at $19.91 with a volume of 2.537, NTRA at $369.39 with a volume of 3.095, SMTC at $185 with a volume of 1.699, TH at $21.19 with a volume of 2.420, TWST at $166.99 with a volume of 1.825, HZO at $52.56 with a volume of 1.717, and MPC at $424.89 with a volume of 3.965. The source of this information is https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt and otherlisted.txt, https://runway.adityaperswal.workers.dev/data/stock-snapshots?feed=iex, and /data/stock-bars SIP/all. Recent SIP snapshots were 403, and the IEX partial-volume screen may miss liquid names. Stale IEX snapshots were retained for broad history screening, but only September 18 full bars qualified. 48 short histories were not declared failures. 15 selected international stocks and indexes showed zero close-high breaks, and ten crypto majors showed zero intraday-high and volume breaks using completed September 19 data. The international selection is bounded, not exhaustive. A huge raw return and name mapping requires a corporate-action audit, and no trade can be made from movers alone. This repairs prior narrow 13-name coverage but does not establish an exhaustive catalyst census.

The Nasdaq directory at https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt and https://www.nasdaqtrader.com/dynamic/SymDir/otherlisted.txt contains 13,220 unique non-test entries. A snapshot from https://runway.adityaperswal.workers.dev/data/stock-snapshots?feed=iex, which is a data feed, queried 12,836 entries after excluding 384 preferred formats, and returned 12,689 usable pairs, including 8,133 daily bars from September 18. A subset of 3,430 non-ETF entries with an IEX dollar volume greater than $100,000 had a median return of -0.55243%, with quartiles of -1.64402% and 0.40347%. The mean return was 0.59191% with a standard deviation of 50.059%, which was contaminated by a 29.17% corporate-action collision outlier, not economic market breadth. The initial whole batch failure from preferred syntax was repaired, and there were no final request errors. Broad discovery flagged FEAM, BTDR, SGML, IMOS, FPS, and AMTX, as well as a crypto-beta cohort, all of which require SIP actions and primary verification. Names without new bottleneck connections were rejected at screen only, not due to a fundamental verdict. The dataset is the current listing survivor universe, including ETFs, warrants, and preferreds, and is not a point-in-time all US security master or OTC. A full name keyword filter misses opaque company names, so CE, EMN, LYB, OLN, and MEOH were manually added to the chemical chain. Sources are close to the market, but IEX partial volume is not consolidated. The full raw files are located at /tmp/universe.csv, /tmp/snapshots.json, and /tmp/screen.csv. There is no order on 1-day returns.