Notes

A news scan found the following data: two sets of results for a trading strategy, one with a 20 rsi and bp_side, and one with a 50 rsi and bp_side. The first set had a return of 52.76717799836802, a drawdown of 32.3563551154957, and a win rate of 43.75 over 16 trades, with 15 closed. The second set had a return of 38.78335414496807, a drawdown of 33.96049159676051, and a win rate of 43.75 over 16 trades, with 15 closed. This data was observed on September 20, and was trained on data from 2016 to 2022, with a holdout period from 2023 to 2025, and recent data from 2026 to September 18. The strategy uses a next open, 2 true average true range, and 6 average true range, with a stop at the first ambiguous bar, and a gap stop at the open, with a window end mark. The strategy is a long only strategy, using vectorbt, with 5 basis point fees, and 15 and 45 basis point slippage on each side, reconciled to 1e-9. The strategy starts on November 1, 2016, and the holdout period is historical, not prospective. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data line: two sets of numbers for a trading strategy, one with a 10 rsi, or relative strength index, a measure of stock price movement, and a 20 bp_side, or basis point side, a measure of position size, and another with a 10 rsi and a 50 bp_side. The first set had a return of -45.15%, a drawdown, or largest loss, of 65.97%, a win rate of 29.73%, and an average win of 22.03%. The second set had a return of -56.07%, a drawdown of 70.97%, a win rate of 29.73%, and an average win of 21.30%. The data was observed on September 20, and was based on training from 2016 to 2022, holding out from 2023 to 2025, and recent data from 2026 to September 18. The strategy used a next open, 2 true average true range, or measure of price volatility, and 6 average true range, with 60 sessions, a stop at the first ambiguous bar, and a gap stop at open, with window end marks. The strategy was a vectorbt long only strategy, with 5 basis point fees and 15/45 basis point slippage on each side, and was reconciled to 1e-9, or a very small number. The strategy started on November 1, 2016, with a separation, and the holdout was historical, not pristine prospective, meaning it was based on past data, not future predictions. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data: the results of a costed sweep of CENX, which is a stock ticker, using a strategy called rsi_only, which refers to a trading strategy based on the Relative Strength Index, or RSI, a measure of stock price movement. The results are as follows: a list of different scenarios with varying RSI values, which measure the speed and change of stock price movements, and bp_side values, which refer to the percentage of the stock's price used to determine buy and sell points. The results include return, which is the profit or loss, DD, which stands for Drawdown, a measure of the biggest decline in value, win, which is the percentage of winning trades, N, which is the number of trades, closed, which is the number of closed trades, mean, which is the average return, halfKelly_risk, which is a measure of the optimal bet size, avgwin, which is the average winning trade, avgloss, which is the average losing trade, targets, which are the number of trades that reached a certain goal, and buyhold, which is the return if the stock was bought and held. The specific results are: for RSI 5 and bp_side 20, the return was -4.314049411338385, for RSI 5 and bp_side 50, the return was -5.455429966893405, for RSI 10 and bp_side 20, the return was 2.709222530187026, for RSI 10 and bp_side 50, the return was 0.8769783288332861, for RSI 15 and bp_side 20, the return was -17.498933447683974, and for RSI 15 and bp_side 50, the return was -19.455406532861524. The data was observed on September 20, and was trained on data from 2016 to 2022, with a holdout period from 2023 to 2025, and recent data from 2026 to September 18. The strategy used a next open, 2 true ATR, or Average True Range, which measures the volatility of a stock, and 6 ATR, with 60 sessions, and a stop at the first ambiguous bar, and a gap stop at open, with a window end mark. The strategy also used vectorbt, which is a backtesting library, with long only, which means only buying stocks, and 5 bp fees, which is a 5 basis point fee, and 15/45 bps lip each side, which refers to the slippage, or the difference between the expected and actual price of a trade. The results were reconciled to 1e-9, which is a very small number. The AA, or Autocorrelation Analysis, starts on November 1, 2016, and the separation is based on historical data, not prospective data, which means the data is based on past events, not future predictions. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the headline 'CENX R92006 costed sweep CENX/rsi_only/holdout'. The results of a trading strategy for CENX are as follows: the strategy was tested with different parameters, including rsi values of 5, 10, and 15, and bp_side values of 20 and 50. The results show returns ranging from 61.51% to 192.22%, with drawdowns ranging from 38.54% to 53.38%. The strategy had a win rate of 40.91% to 45%, with an average win of 30.48 to 33.38 and an average loss of -12.71 to -13.50. The strategy was trained on data from 2016 to 2022, and tested on holdout data from 2023 to 2025, with recent data from 2026 to September 18. The next open trade will use a 2 true ATR and 6 ATR, with a 60 session stop and a gap stop at open. The strategy uses a vectorbt long only approach, with 5bp fees and 15/45bps slip on each side, and has been reconciled to 1e-9. The strategy started on November 1, 2016, and the holdout data is historical, not prospective. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data line: a list of results for the stock CENX, including return, drawdown, win rate, and other metrics, for different combinations of parameters, such as rsi and bp_side. The results are as follows: for rsi 5 and bp_side 20, the return was 19.99892569001976, the drawdown was 81.5236197479961, the win rate was 34.883720930232556, and the average win was 37.074500048980575. For rsi 5 and bp_side 50, the return was -7.289635370082817, the drawdown was 83.21488910492046, the win rate was 34.883720930232556, and the average win was 36.25450749140964. For rsi 10 and bp_side 20, the return was 111.5847561601767, the drawdown was 81.43445488254177, the win rate was 32.6530612244898, and the average win was 41.63963513496148. For rsi 10 and bp_side 50, the return was 57.68869269297593, the drawdown was 83.23478078932163, the win rate was 32.6530612244898, and the average win was 40.79233350974142. For rsi 15 and bp_side 20, the return was 6.170191617439491, the drawdown was 87.51490194807316, the win rate was 29.310344827586203, and the average win was 41.11284434341498. For rsi 15 and bp_side 50, the return was -25.033601376078025, the drawdown was 88.9267243392313, the win rate was 29.310344827586203, and the average win was 40.26869403027929. The data was observed on September 20, and the training period was from 2016 to 2022, with a holdout period from 2023 to 2025, and a recent period from 2026 to September 18. The next open, 2 true ATR, 6 ATR, 60 sessions, stop first ambiguous bar, gap stop at open, and window end marks were also noted. The vectorbt long only strategy was used, with 5 bp fees and 15/45 bps lip each side, and the results were reconciled to 1e-9. The AA starts on November 1, 2016, and the holdout is historical, not pristine prospective. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data: a list of trading results for CENX, which is a stock ticker symbol, with different settings. The results include return, which is the profit or loss, and drawdown, or DD, which is the largest decrease in value. The results also include win percentage, number of trades, or N, and average win and loss. The settings used were different relative strength index, or RSI, values and different buy point side, or bp_side, values. The data was observed on September 20 and was trained on data from 2016 to 2022, with a holdout period from 2023 to 2025, and recent data from 2026 to September 18. The next open trade will use a true average true range, or ATR, which is a measure of price volatility, and a 60 session window, with a stop at the first ambiguous bar, which is a bar with conflicting signals, and a gap stop at the open. The trading strategy used was a long only strategy, which means only buying, not selling, with fees of 5 basis points, or bps, and a slip of 15 to 45 bps on each side. The results were reconciled to an accuracy of 1e-9, which is a very small number. The strategy started on November 1, 2016, and the holdout period is historical, not prospective, meaning it is based on past data, not future predictions. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data: the results of a trading strategy for CENX, which is a stock ticker, from September 20, with training data from 2016 to 2022, holdout data from 2023 to 2025, and recent data from 2026 to September 18. The strategy used a relative strength index, or RSI, which measures stock performance, and a buy/sell side, or bp_side, which refers to the percentage of the portfolio invested. The results showed returns ranging from -60.60135491455735 to 91.39408786425125, with a maximum drawdown, or DD, of 76.92589874606259, which is the largest decline in value. The strategy had a win rate, or win, of 23.52941176470588 to 46.15384615384615, with an average win, or avgwin, of 28.10101296031206 to 34.91093787568268, and an average loss, or avgloss, of -11.76103089350332 to -14.36447719670618. The strategy used a half Kelly risk, which is a measure of risk, and a mean, which is an average, of 3.1753728437343054 to 7.765541967873656. The strategy also used targets, which are price levels, and a buy and hold, or buyhold, return of 369.7841726618705. The next open trade will use a true average true range, or ATR, which measures price volatility, and a stop loss, which is a price level at which to sell, at the open, with a gap stop, which is a stop loss at the opening price, and a window end mark, which is a price level at the end of a trading period. The strategy used vectorbt, which is a trading library, and long only, which means only buying, with fees, or bpfees, of 5 basis points, or bps, and a slip, which is a small price movement, of 15/45 bps each side. The results were reconciled to 1e-9, which is a small number. The strategy started on November 1, 2016, and the holdout data is historical, not prospective, meaning it is based on past data, not future predictions. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data: the results of a trading strategy test on CENX, which is a stock ticker, are as follows: the strategy was tested with different parameters, including rsi, which stands for relative strength index, a measure of stock price movement, and bp_side, which refers to the side of a trade. The results include return, which is the profit or loss, DD, which stands for drawdown, the largest decline in value, win, which is the percentage of winning trades, N, which is the total number of trades, closed, which is the number of trades that were closed, mean, which is the average return, halfKelly_risk, which is a measure of risk, avgwin, which is the average winning trade, avgloss, which is the average losing trade, targets, which are the profit targets, and buyhold, which is the return if the stock was bought and held. The results are: a return of -24.99349694514956 percent with an rsi of 5 and bp_side of 20, a return of -32.672163843361965 percent with an rsi of 5 and bp_side of 50, a return of 85.60622255549765 percent with an rsi of 10 and bp_side of 20, a return of 61.681120103146846 percent with an rsi of 10 and bp_side of 50, a return of 26.96135930623734 percent with an rsi of 15 and bp_side of 20, and a return of 9.27649483364778 percent with an rsi of 15 and bp_side of 50. The data was observed on September 20, and the training period was from 2016 to 2022, with a holdout period from 2023 to 2025, and recent data from 2026 to September 18. The next open trade will use a 2 true ATR, which stands for average true range, a measure of price movement, and a 6 ATR stop, with a 60 session limit, and a stop at the first ambiguous bar, with a gap stop at the open, and a window end mark. The strategy uses vectorbt, which is a trading library, and long only trades, with 5 basis point fees, and 15 and 45 basis point slippage on each side, and was reconciled to 1e-9, which is a small value. The AA, which stands for autonomous agent, starts on November 1, 2016, and the holdout is historical, not prospective, meaning it looks back at past data, not forward at future data. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data line: two sets of trading results for CENX, which is a stock ticker, with the following details. The first set has a relative strength index, or RSI, of 10, a buy or sell side, called bp_side, of 20, a return of -29.41%, a drawdown, or DD, of 46.32%, 20% win rate, 5 total trades, 5 closed trades, a mean of -5.76%, a half Kelly risk of 0, an average win of 22.84%, an average loss of -12.91%, 1 target, and a buy and hold return of -2.38%. The second set has an RSI of 10, a bp_side of 50, a return of -31.49%, a DD of 47.43%, 20% win rate, 5 total trades, 5 closed trades, a mean of -6.32%, a half Kelly risk of 0, an average win of 22.10%, an average loss of -13.43%, 1 target, and a buy and hold return of -2.38%. These results were observed on September 20, and were trained on data from 2016 to 2022, with a holdout period from 2023 to 2025, and recent data from 2026 to September 18. The next open trade will use a true average trading range, or ATR, of 2, and a 6 ATR stop, over 60 sessions, with a stop at the first ambiguous bar, and a gap stop at the open, with window end marks. The vectorbt long only strategy was used, with 5 basis point fees, and 15/45 basis point slippage on each side, and was reconciled to 1e-9. The AA, or automated trading algorithm, starts on November 1, 2016, and is a separation of the data. The holdout is historical, not prospective, meaning it looks back at past data, rather than looking forward to future data. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data: the trading results for CENX are given in two scenarios. The first scenario has an RSI of 10, a bp_side of 20, a return of 84.00226144125746, a drawdown of 52.22978959542116, a win rate of 40.909090909090914, 22 trades, 21 closed trades, a mean of 4.406014173212882, a half Kelly risk of 0.08311243779530678, an average win of 26.506346643713073, an average loss of -10.89421599867187, 8 targets, and a buy and hold return of 369.7841726618705. The second scenario has an RSI of 10, a bp_side of 50, a return of 61.24851735870408, a drawdown of 54.46861975655322, a win rate of 40.909090909090914, 22 trades, 21 closed trades, a mean of 3.7814475721522425, a half Kelly risk of 0.07342738185311007, an average win of 25.74957377424235, an average loss of -11.427255183140911, 8 targets, and a buy and hold return of 369.7841726618705. The data was observed on September 20, trained from 2016 to 2022, and held out from 2023 to 2025, with recent data from 2026 to September 18. The next open is based on 2 true average true ranges and 6 average true ranges, with 60 sessions, and a stop at the first ambiguous bar or a gap stop at the open, with window end marks. The strategy uses vectorbt long only, with 5 basis point fees and 15/45 basis point slippage on each side, and was reconciled to 1e-9. The AA starts on November 1, 2016, as a separation point. The holdout is historical, not pristine prospective. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the following data line: a list of two sets of trading results for CENX, which is a ticker symbol, or a stock identifier. The first set has a return of 25.274143299162503 percent, a drawdown, which is a measure of loss, of 71.07998198918727 percent, and a win rate of 32.432432432432435 percent. The second set has a return of 0.3337689467547079 percent, a drawdown of 73.41005329147362 percent, and the same win rate. These results were observed on September 20, and are based on training data from 2016 to 2022, with a holdout period from 2023 to 2025, and recent data from 2026 to September 18. The trading strategy uses a next open, 2 true average true range, or a measure of price volatility, and 6 average true range, with 60 sessions, and a stop at the first ambiguous bar, or a bar with conflicting signals, and a gap stop at the open. The strategy also uses a window end mark, and a vectorbt long only, or a strategy that only buys, approach, with 5 basis points fees, or a 0.05 percent fee, and 15 and 45 basis points slippage, or the difference between the expected and actual price, on each side. The results were reconciled to 1e-9, or a very small number. The AA, or autonomous agent, starts on November 1, 2016, and the separation, or the point at which the training data ends and the holdout period begins, is also on this date. The holdout is historical, or based on past data, rather than pristine prospective, or based on future data. https://runway.adityaperswal.workers.dev/data/stock-bars?timeframe=1Day&feed=sip&adjustment=all

A news scan found the headline about an issue with Alcoa's separation boundary. On November 1, 2016, there was a big gap in the stock price, with a loss of 74.35646%. This was because Alcoa split into two companies, with the new company being called Arconic, and the parent company keeping the Arconic shares. This change means that the original Alcoa stock is not the same as it was before, so it's not just an ordinary loss. To fix this issue, the rows of data before November 1, 2016, need to be removed and the indicators need to be recalculated. The original data should be kept in a file called AA_unrepaired.csv. After fixing the issue, the worst gap in the stock price was found to be -13.2465% on March 16, 2020. There was an incomplete audit of corporate actions outside of the identified boundary, and there is no actual size or stop based on the incorrect 74% gap. https://news.alcoa.com/press-releases/press-release-details/2016/Alcoa-Corporation-Launches-as-an-Independent-Industry-Leader-in-Bauxite-Alumina-and-Aluminum-Products/default.aspx

The information is dated August 12, 2026, despite later dates mentioned on the homepage. In the first half of the year, sales were 939 thousand tons versus 1373 thousand tons; cast production was 1006 versus 1420. The operating cash flow was 404 million USD versus 937 million USD, and the metal premium was 2405 USD per ton versus 857 USD per ton. The company has established alternative routes outside of Hormuz, but shipments are contingent on the Strait reopening due to previous incidents. The reported EBITDA was 1204 million USD, and the adjusted EBITDA was 1227 million USD, with the adjustment excluding the impact of an attack. The plants' restart and export recovery face different constraints. This report covers the first half of the year, not the current flow as of September. https://media.ega.ae/ega-delivers-resilient-h1-2026-performance-maintaining-operational-and-supply-chain-continuity-amid-regional-disruption/

On August 26, the primary operator of the Al Taweelah aluminum smelter reported that 315 out of 1262 cells were restarted, which is approximately 24.9604% of the total cells. The operator expects about one week of stabilization, and some hot metal will be used for the next restarts, so not all of the 25% export recovery will be achieved. All three potlines have been energized, and full production is expected by the first quarter of 2027. The refinery is currently operating at about 50% capacity, but this is not dependent on the smelter's recovery. The forecast is not a guarantee of completion, and the direct issuer has a bias towards an optimistic recovery. https://media.ega.ae/egas-al-taweelah-smelter-restoration-reaches-25-completion-milestone/

A news scan found the headline that the primary operator for the June quarter reported the Oswego hot mill was restarted early in June and is ramping up to normalize shipments. The Bay Minette commissioning is underway, and there is a positive free cash flow forecast for fiscal Q4, which is not current. Constellium may have benefited from a shortage from April to June, and Novelis may have also restarted in the same quarter, but restarting a mill does not mean instant full qualified output. This is a case against a permanent sheet shortage, according to an independent operator. https://investors.novelis.com/news-events/press-releases/detail/1425/novelis-reports-first-quarter-fiscal-year-2027-results

A news scan found the headline about CSTM R92006 H1 cash versus noncash metal lag. On July 29, the issuer reported its June quarter and first half results. The second quarter EBITDA, which is earnings before interest, taxes, depreciation, and amortization, a measure of a company's overall financial performance, was 439, including 129 of metal lag, which refers to the delay between paying for metals and receiving them. The first half EBITDA was 798, including 226 of metal lag, leaving the segment at 572. The issuer's operating cash flow was 234, capital expenditures were 149, and property, plant, and equipment inflows were 10, resulting in a free cash flow of 95. The networking capital uses were 457, calculated as -495 - 418 + 480 - 24. The issuer provided a full-year EBITDA guidance of 980 to 1020, excluding metal lag, and a free cash flow of more than 300, which requires a second half of more than 205. The automotive rolled volume was 68 kilotons in the second quarter, compared to 59 in the previous period, and the issuer attributes part of this increase to a North American supply shortage. The net debt was 1760 million, and the number of shares on June 30 was 135,527,728. The issuer is primary, but the durability of its rent is uncertain. https://www.sec.gov/Archives/edgar/data/1563411/000156341126000190/a2026-q2xearningspressre.htm

A news scan found the headline about CENX R92006 Q2 operating scale and Q3 EBITDA guidance. On August 6, the issuer reported the following information for Q2: shipments were 130,632 tons versus 1,128,065 in Q1; adjusted EBITDA was 326.9 million. For Q3, the guidance is 325 to 345, with a midpoint of 335. The Mt Holly last 90 pots restart is complete and Grundartangi Line 2 is near full. In July, a 45X refund of 94.3 million was made and cash was above debt, but it was asserted that there will be no September cash balance. The guidance is not based on cash and is not confirmed future income. https://www.sec.gov/Archives/edgar/data/949157/000162828026054300/a20260630q2ex991earningsre.htm

A report filed on August 6, for the period ending June 30, showed that the company had an operating cash flow of $236.0 million and capital expenditures of $134.4 million, resulting in a simple free cash flow of $101.6 million. The company also had a separate cash sale of $200 million from Hawesville, which resulted in an income gain of $287.9 million, but this is not part of the operating rent. The company recognized insurance gains of $73.1 million, which should not be automatically subtracted as it is not the same as cash. The manufacturing credit receivable was $225.2 million, which is the sum of $176.8 million in current credits and $48.4 million in non-current credits, with $52.6 million in credits recognized in the first half of the year. Glencore accounted for 45.5% of the company's sales in the first half of the year, compared to 59.1% in the prior year. The company had $343.4 million in unrestricted cash, and the primary issuer released $94.3 million in cash on July 45, which is later than the June cash flow. https://www.sec.gov/Archives/edgar/data/949157/000162828026054308/cenx-20260630.htm

A news scan found the filing from March 3, 2026, about the company's fiscal year 2025. The company, CENX, has two large alumina supply contracts. One contract is with Glencore for 500,000 tonnes per year through 2028, with prices linked to the London Metal Exchange, or LME. The other contract is with Concord for approximately 540,000 tonnes per year through 2029, with prices fixed and also linked to the LME and other components. Glencore accounts for 36.4 percent of the company's stock and 54 percent of its fiscal year sales. The company's power costs are also linked to the LME, with about 70 percent of the power for Grundartangi and a cost-of-service agreement with Santee Cooper for Mt Holly through 2031. These contracts do not provide for upside if spot prices rise, and are not a pure play on spot prices. The company is the primary issuer of these contracts, and the quantities listed are contracted annual rates, not verified delivered tonnes. https://www.sec.gov/Archives/edgar/data/949157/000162828026013788/cenx-20251231.htm

A report filed on February 25, 2026, for the fiscal year 2025, shows that the top 10 metal suppliers accounted for 49% of the purchase volume, and the top 10 customers accounted for 56% of the revenue. The two largest customers generated revenues of 943 million USD and 852 million USD. Most of the slab and billet used was produced internally. The company's price model aims to achieve metal neutrality through passing costs on to customers and hedging, but the risk clauses state that complete pass-through is not guaranteed. The incentives for the primary issuer are favorable. The full filing was reviewed, but extracting information from the CLI section was unreliable. https://www.sec.gov/Archives/edgar/data/1563411/000156341126000057/cstm-20251231.htm

A report was retrieved at 06:01 UTC on September 20. A job, numbered 282, was completed at 05:46 UTC, with all 16 parts finished. The latest information from Laxmi, with numbers 1049, 1048, and 1047, and the CE board from September 16, with an 8K filing, was unchanged since 04 UTC. Notes, numbered 3866 and 3858, were recorded from an issuer notice index, which is a list of announcements from companies, and the collector is one step removed, meaning it is not directly from the source. This report does not indicate a new shortage or a paid-spread observation, which refers to the difference between two prices. Some corrections were made to H1 FCF 143 and H2 need 557-657, but no holdings were affected and no automatic order was placed. https://www.laxmi.com/customer-info/

A list of possible prices for a stock was created based on different assumptions. The list includes the following prices: for a free cash flow of 286, the prices are 26.059480527399423 at a 10% yield, 21.71623377283285 at a 12% yield, and 17.372987018266283 at a 15% yield; for a free cash flow of 500, the prices are 45.558532390558426 at a 10% yield, 37.96544365879869 at a 12% yield, and 30.372354927038955 at a 15% yield; for a free cash flow of 700, the prices are 63.781945346781804 at a 10% yield, 53.151621122318176 at a 12% yield, and 42.52129689785453 at a 15% yield; for a free cash flow of 750, the prices are 68.33779858583765 at a 10% yield, 56.948165488198036 at a 12% yield, and 45.558532390558426 at a 15% yield; for a free cash flow of 800, the prices are 72.89365182489348 at a 10% yield, 60.74470985407791 at a 12% yield, and 48.59576788326233 at a 15% yield. The free cash flow, or FCF, is the amount of money a company has after paying expenses, and the yield is the return on investment, with 10%, 12%, and 15% being the assumed yields. The prices are in US dollars per share. The yields of 10%, 12%, and 15% are assumptions and not actual market quotes or fair value estimates. The free cash flow of 286 is an annualized figure based on seasonal data from the first half of the year and is not the base forecast. The free cash flow range of 700 to 800 is the issuer's guide for 2026, but the durability is unknown. No single cell in the grid was selected to force a 3R decision. https://www.sec.gov/Archives/edgar/data/1306830/000130683026000114/q220268-kex991b.htm

A news scan found the following data: the company has 109.748926 million shares, a market capitalization of 4983.698729659999 million USD, net debt of 10643 million USD, and an enterprise value of 15626.698729659998 million USD. The net debt to equity ratio is 2.135562476250665. The price to earnings guide is 7.5683333333333325. The free cash flow yield guide, which is a measure of a company's ability to generate cash, is between 14.045792853288221 and 16.052334689472254 percent. The free cash flow yield for the first half of the year, annualized, is 5.73870965148633 percent. The company needs 557 to 657 million USD in the second half of the year, which is 3.895104895104895 to 4.594405594405594 times the amount needed in the first half. The value of Nutrinova per share is 1.3849793846729763 USD, and the multiple, which is a ratio used to evaluate the company's performance, is 38. The company can pay down 1.4281687494127595 percent of its net debt. The earnings per share lost is 0.036446825912446745 USD. The Acetyl segment has declined by 26.791277258566982 percent, with a difference of 86 million USD. The hurdle free cash flow at yield, which is the minimum amount of cash flow required to meet the yield, is between 617.743955726382 and 926.6159335895729 million USD. The incremental earnings per share for one quarter is between 0.7209182165481967 and 0.5485247299823235 USD. The current stock price is 45.41 USD. The free cash flow issuer definition has been corrected to 30351. The tax rate is forecasted to be 8 percent, versus a normalized assumption of 20-30 percent. The first half annualization is for diagnostic purposes only and is not a forecast. Nutrinova has a conditional value of 152 million USD in the fourth quarter, which is not a recurring free cash flow. https://www.sec.gov/Archives/edgar/data/1306830/000130683026000117/ce-20260630.htm

A news scan found the following percentage changes in chemical prices from September 7 to 15: ethyl increased by 6.666666666666665 percent, anhydride increased by 11.764705882352944 percent, butyl increased by 14.049586776859503 percent, and propyl decreased by 12.790697674418606 percent over 8 days, compared to the prices listed on https://www.laxmi.com/documents/default.aspx?f=customer-info/price-announcement-1048.pdf. The payment terms remain the same, with no consideration for volume or invoice realization, so this should not be referred to as a weighted chemical price index. The decrease in propyl prices contradicts the claim of uniform inflation. https://www.laxmi.com/documents/default.aspx?f=customer-info/price-announcement-1049.pdf

A simulation of the stock symbol EMN, which is a recent regime, was run with 20,000 paths. The stop percentage was 70.215, the target percentage was 19.28, and the time percentage was 10.505. The average profit and loss per share was -0.8359320844684804, with the 5th percentile being -7.453366432644417 and the 95th percentile being 12.696650988684453. The loss of 2 shares at the 1st percentile was -28.5410203058067. This simulation, seeded with 92004, used 20,000 paths for each of 4 cells, totaling 80,000 paths. The simulation looked at blocks of closing returns starting in either 2017 or 2025, with fixed barriers of 20 basis points on either side, set on September 18, and included any overshoot at the close. The simulation did not consider intraday highs and lows, conditional signals, or news models, and the probabilities were not calibrated to trade odds. A stress diagnostic for 2 shares of EMN exceeded the 120 fund cap, but no such position was proposed. https://runway.adityaperswal.workers.dev/data/stock-bars