Notes

A news scan found the following data line: {"return_pct":-10.906390051642312,"drawdown_pct":64.86861057258156,"win_rate":41.37931034482759,"sample":29,"mean_pct":0.4662762728441721,"avg_win_pct":14.577298233356464,"avg_loss_pct":-9.494445111046858,"halfKelly":0.015993233635613474,"partial":11,"missing_checks":107,"trades_missing":14,"endpoint":0}; this is a raw-price fixed-path result from a trading simulation, specifically the m15_True_L20_train_0.0016 model, trained on data from 2020 to 2024 and tested on data from 2025 up to September 11. The simulation used split-normalized prices, without accounting for cash dividend credits, and adjusted the entry signal for split-only execution. It considered 15-minute opening prices, filtered by the regular trading hours calendar, and assumed a managed variant with a quarterly return of two times the benchmark return for each sample, although the actual manager traded less frequently, so this is not a live replication. The drawdown was approximated on a daily basis, not intraday. The Vectorbt long-only compounded return assertion was passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the following data line: {"return_pct":-5.586159583417416,"drawdown_pct":64.30198682586128,"win_rate":41.37931034482759,"sample":29,"mean_pct":0.6674101100774836,"avg_win_pct":14.80668238601396,"avg_loss_pct":-9.3132526729365,"halfKelly":0.022537462906204536,"partial":11,"missing_checks":107,"trades_missing":14,"endpoint":0}; this is a raw output from a trading fund manager, which includes return percentage, drawdown percentage, win rate, and other metrics. The data is from train2020-24 and knownholdout2025-Sep11. The manager used split-normalized prices, no cash dividend credit, and adjusted entry signal with split-only execution. The manager also used a 15-minute calendar filter and assumed a quarterly return. The drawdown is approximate and based on daily data, not intraday data. The Vectorbt long-only compounded return assertion passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

The fund had a return of 65.23% and a drawdown of 20.75%. The win rate was 60% based on a sample of 15 trades. The average win was 11.65% and the average loss was -7.59%. A calculation called the half Kelly criterion was 0.17. There were 5 partial trades and 12 missing checks, with 3 trades missing. The endpoint was 1. The fund was trained on data from 2020 to 2024 and tested on data from 2025 to September 11. The prices used were normalized and did not include cash dividend credits. The entry signal was adjusted and execution was based on stock splits only. The fund used 15-minute opening prices, filtered by a trading hours calendar. The results assume a managed variant with a quarterly return of 2 times the benchmark, but the actual manager traded less frequently, so the results are not a live replication. The drawdown was estimated based on daily prices, not intraday prices. A test using a method called vectorbt long only compounded return assertion was passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the following data line: {"return_pct":70.26568169361845,"drawdown_pct":20.43281391629357,"win_rate":60,"sample":15,"mean_pct":4.158112042931031,"avg_win_pct":11.869254522797396,"avg_loss_pct":-7.408601676868519,"halfKelly":0.17516315093523788,"partial":5,"missing_checks":12,"trades_missing":3,"endpoint":1}; this is a raw output from a trading system test. The test was run on data from 2020 to 2024, and the results are being checked against known data from 2025 up to September 11. The prices used were adjusted to account for stock splits, but did not include any cash dividend payments. The system's entry signals were also adjusted, and trades were only executed when a stock split occurred. The test used 15-minute opening prices, but only during regular trading hours, and assumed that the manager would increase the investment amount by two times the average return every quarter. However, the actual manager does not make changes this frequently, so the results are not a live replication. The drawdown, or largest loss, was estimated using daily prices, not prices from within the day. The test also checked that the returns were correctly calculated, and this check passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the headline with the following data: return percentage is 32.35269580623552, drawdown percentage is 67.41213908736012, win rate is 46.875, sample size is 32, mean percentage is 1.9099570532120833, average win percentage is 15.239549002274428, average loss percentage is -9.851447607725278, half Kelly is 0.06266448741124267, partial is 11, missing checks is 196, trades missing is 14, endpoint is 0. This data is from a train period of 2020 to 2024 and a known holdout period of 2025 to September 11. The prices were split and normalized, with no cash dividend credit. The entry signal was adjusted and execution was split-only. The data uses 15-minute opens, filtered by a regular trading hours calendar. A managed variant assumes a quarterly return of 2 times the benchmark return for each sample, but the actual manager trades less frequently, so this is not a live replication. The drawdown is approximate and based on daily data, not intraday data. A Vectorbt long-only compounded return assertion was passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the following data line: return percentage is 41.1002127375832, drawdown percentage is 66.82024589200381, win rate is 46.875, sample size is 32, mean percentage is 2.113981144427799, average win percentage is 15.470258983292476, average loss percentage is -9.670969889864567, half Kelly is 0.0683240386185793, partial is 11, missing checks is 196, trades missing is 14, endpoint is 0. This data is from a train period of 2020 to 2024 and a known holdout period of 2025 to September 11. The prices used were split normalized, without cash dividend credit, and had an adjusted entry signal with split-only execution. The data used 15-minute opens, filtered by a regular trading hours calendar. A managed variant assumed a plus 2R quarter per sample, but the actual manager traded less frequently, so this is not a live replication. The drawdown is a daily mark approximation, not an intraday one. A vector backtest long only compounded return assertion passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the following data line: the return percentage is 57.201586876475695, the drawdown percentage is 15.42569022653214, the win rate is 60, the sample size is 10, the mean percentage is 5.1885924636438245, the average win percentage is 13.32804728286742, and the average loss percentage is -7.020589765191571. The half Kelly, a measure of optimal bet size, is 0.19464938687280603. This data is from a training period of 2020 to 2024 and a known holdout period of 2025 to September 11. The prices used were split normalized, meaning they were adjusted for stock splits, and did not include cash dividend credits. The entry signal was also adjusted and execution was split-only. The data used 15-minute opens, filtered by regular trading hours, and assumed a managed variant with a quarterly return of 2 times the benchmark return for each sample, although the actual manager traded less frequently, so this is not a live replication. The drawdown was calculated daily, not intraday. A Vectorbt long only compounded return assertion, a type of backtesting method, was passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

The return percentage is 60.37727306055889, the drawdown percentage is 14.758470803522139, the win rate is 60, the sample size is 10, the mean percentage is 5.3991803944428725, the average win percentage is 13.55493043074341, and the average loss percentage is -6.834444660007932. The half Kelly criterion, a measure of bet size, is 0.19915928089889717. There are 11 missing checks and 3 trades are missing. The endpoint is 0. The data is from 2020 to 2024 for training and from 2025 to September 11 for the known holdout period. The prices are split-normalized and do not include cash dividend credits. The entry signal is adjusted and execution is split-only. The data is filtered by regular trading hours and 15-minute opens. The managed variant assumes a quarterly return of 2 times the benchmark return for each sample, but the actual manager trades less frequently, so this is not a live replication. The drawdown is approximated daily, not intraday. The Vectorbt long-only compounded return assertion has passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the following data line: {"return_pct":-12.759306856913977,"drawdown_pct":70.5825896976696,"win_rate":31.03448275862069,"sample":29,"mean_pct":0.6548787376451026,"avg_win_pct":22.088031175133214,"avg_loss_pct":-8.990039859224549,"halfKelly":0.014824289508934746,"partial":0,"missing_checks":108,"trades_missing":14,"endpoint":0}; this is a raw output from a trading system test, which includes return percentage, drawdown percentage, win rate, and other metrics, for the period from 2020 to 2024, with a known holdout period from 2025 to September 11. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the following data: the return percentage is -7.549723431144761, the drawdown percentage is 70.10812502331474, and the win rate is 31.03448275862069. This is based on a sample of 29, with a mean percentage of 0.8563901576739998. The average win percentage is 22.33245184146533 and the average loss percentage is -8.807837600032096. The half Kelly criterion, a betting strategy, is 0.019173670758441202. There are 108 missing checks and 14 trades are missing. The endpoint is 0. This data is from a train period of 2020 to 2024 and a known holdout period of 2025 to September 11. The prices are split normalized, without cash dividend credit, and the entry signal is adjusted. The execution is split only and the trades are filtered by a regular trading hours calendar, with 15 minute opens. The manager variant assumes a quarterly return of 2 times the benchmark, but the actual manager trades less frequently, so this is not a live replication. The drawdown is approximate and based on daily marks, not intraday. The vector backtest long only compounded return assertion has passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the following data line: return percentage is 55.62507696402639, drawdown percentage is 19.038335846390765, win rate is 54.54545454545454, sample size is 11, mean percentage is 4.687073072096452, average win percentage is 14.100899066515806, average loss percentage is -6.609518121206774, half Kelly is 0.1661976676092392, partial is 0, missing checks is 0, trades missing is 0, endpoint is 0. This data is from a train period of 2020 to 2024 and a known holdout period of 2025 to September 11. The prices used were split normalized, without cash dividend credit, and had an adjusted entry signal with split-only execution. The data used 15-minute opens, filtered by a regular trading hours calendar, and assumed a managed variant with a plus 2R quarter per sample, but the actual manager traded less frequently, so this is not a live replication. The drawdown is a daily mark approximation, not an intraday one. A vector backtest long only compounded return assertion was passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the headline 'TEN R91304 raw-price fixed-path daily_False_L20_holdout_0.0006'. The body of the message is a raw data output showing the performance of a trading strategy, including a return percentage of 59.08677142578305, a drawdown percentage of 18.71384027976719, and a win rate of 54.54545454545454. This data is based on 11 samples, with a mean percentage of 4.8966569593157905, an average win percentage of 14.32932946637151, and an average loss percentage of -6.422550049151073. The data also includes a 'half Kelly' value of 0.1708613431915083, which refers to a betting strategy, and other technical metrics. The trading strategy was trained on data from 2020 to 2024 and tested on known holdout data from 2025 to September 11. The strategy uses split-normalized prices, without cash dividend credits, and an adjusted entry signal with split-only execution. The strategy is based on 15-minute opens, filtered by a regular trading hours calendar, and assumes a managed variant with a quarterly return of 2 times the benchmark return, although the actual manager trades less frequently. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the following data line: the return percentage is -13.167339159792713, the drawdown percentage is 69.27441553628378, the win rate is 30.303030303030305, the sample size is 33, the mean percentage is 0.5647444482408144, the average win percentage is 21.291161961642135, and the average loss percentage is -8.446741427151064. The data also includes a half Kelly value of 0.013262414922639049. This data is from a training period of 2020 to 2024 and a known holdout period of 2025 to September 11. The prices were split-normalized and did not include cash dividend credits. The entry signal was adjusted and the execution was split-only. The data was filtered using a regular trading hours calendar and 15-minute opens. A managed variant assumed a plus 2R quarter per sample, but the actual manager traded less frequently, so this is not a live replication. The drawdown is a daily mark approximation, not an intraday one. A vector backtest long only compounded return assertion was passed. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the following data line: the return percentage is -7.2430251357442526, the drawdown percentage is 68.77885170938278, the win rate is 30.303030303030305, the sample size is 33, the mean percentage is 0.7660754191064002, the average win percentage is 21.533987293015812, and the average loss percentage is -8.263451482593345. The half Kelly, a betting strategy, is 0.017787588723869624. This data is from a train period of 2020 to 2024 and a known holdout period of 2025 to September 11. The prices were split and normalized, with no cash dividend credit, and the entry signal was adjusted. The execution was based on split-only data. The data used 15-minute opens, filtered by a regular trading hours calendar. The manager variant assumes a quarterly return of two times the benchmark, but the actual manager trades less frequently, so this is not a live replication. The drawdown is approximate and based on daily data, not intraday data. The vector backtest, a type of simulation, for long-only compounded returns was successful. https://runway.adityaperswal.workers.dev/data/stock-bars?symbols=TEN&feed=sip

A news scan found the headline about an independent broker's report on tanker supply. The report, which covers the period from August 10 to the end of July 2026, states that 96.4 million deadweight tons of tankers were ordered year-to-date, including 197 Very Large Crude Carriers and 103 Suezmax tankers. The order book is wider than 31% of the trading fleet, with specific types of tankers having the following percentages: LR1 at 13.6%, handy at 17.6%, and chemical at 12.4%. In terms of capacity, 47.5% of the fleet is more than 15 years old, and around 21% is more than 20 years old. Additionally, 2.16 million deadweight tons were removed year-to-date. The broker acts as an intermediary in the market and has a commercial interest in transactions. Although the broker is an independent company from TEN, there may be some overlap in their databases. TEN reported a 19.9% slide in ship count for vessels over 25,000 deadweight tons in late August, with a 31% capacity-weighted average, but the context and denominator used are not explicitly the same. https://ifchorgalbraiths.com/research-insight/focus-on-tanker-fleet-august-2026/

On September 10, a call was held, and a transcript was retrieved from MotleyFool and Yahoo. The Chief Financial Officer, or CFO, mentioned that the profit share for the first half of the year was 71 million dollars, and for the second quarter it was 30.5 million dollars, which is higher than the 10 million dollars from the same period last year, and the estimated profit share for the fiscal year 2025 is 46 million dollars. The company has 13 ships, including 7 Suezmax and 2 VLCC, which is a type of very large crude carrier, a large ship that transports oil. The CEO wants to pay a higher dividend in November, which is a payment to shareholders, and is considering redeeming, or buying back, around 120 million dollars in debt, and there are 19 new ships being built. The chairman said that making a part of the specialized fleet a separate company is not a priority for now. Management thinks the new ships are worth 30 percent more than the 3.1 billion dollar program, but this is not a verified appraisal, and it is inconsistent with what they said before about doubling the value. https://finance.yahoo.com/energy/articles/tsakos-energy-navigation-ten-q2-002827393.html

A news scan found the headline is not available, but a slide presentation from September 10, slide 14, relays information from Clarkson, which reported that 44 ships were scrapped from January to August 2026, and 53 ships were scrapped in the fiscal year 2025. According to slide 13, it is expected that 2889 ships will depart over the next 15 years, with a significant number departing in the next 3 years, specifically 2889 ships, compared to an order book of 1140 ships. The age of the ships is not a factor that determines contracted scrapping. The source of this information is an issuer relay of a vendor, rather than an independent direct fleet database. https://www.tenn.gr/wp-content/uploads/2026/09/TEN_Q2_2026.pdf

A news scan found the headline 'September 10 issuer slide 7'. According to this, for every $1000 per day increase in the spot rate, the annual earnings per share, or EPS, would increase by 0.11, based on the current spot rate of vessels and the number of common shares as of June 30. This is a management sensitivity analysis, which means it is an estimate of how changes in the spot rate might affect earnings, and it does not take into account the cash paid or the whole-fleet rate beta, which is a measure of how much the fleet's earnings might change when the spot rate changes. https://www.tenn.gr/wp-content/uploads/2026/09/TEN_Q2_2026.pdf

As of September 8, the fleet exposure is as follows: on September 10, a presentation on page 5 stated that 23 out of 62 vessels are on spot contracts with a profit share, which is 37 percent, and 52 out of 62 vessels are secured with time charter contracts plus time charter profit share, which is 84 percent. There is an overlap of 13 vessels with profit share, implying 10 vessels are on spot contracts and 39 are on fixed contracts, which is not contradictory at 121 percent. The operating day mix for the first half of the year has a different period and denominator. https://www.tenn.gr/wp-content/uploads/2026/09/TEN_Q2_2026.pdf

As of December 31, 2025, the contracted yard installments for the year-end are 437.258 million in 2026, 632.763 million in 2027, and 898.277 million in 2028, as noted in the April 6 20-F Note 12. The first LNG contract on February 25 is for 254.4 million, which is in addition to the previous amounts, and the second LNG contract in July further changes the scope. Note that the aggregate net investing for the first half of the year should not be subtracted to infer the current remaining obligations. https://www.sec.gov/Archives/edgar/data/1166663/000119312526144027/d49235d20f.htm

A news scan found the headline about TEN R91304 March 2026 preferred liquidation claims. As of March 30, there were 47,459,947 E shares and 47,471,147 F shares, which is equivalent to 25 USD and totals 287.32735 million. This breaks down to 118.648675 million for E and 168.678675 million for F. The annual rate for E is 9.25% and for F is 9.5%. The issuer of E expects to continue at a rate of 9.25% under the LIBOR fallback, which is a backup plan for setting interest rates. The shares can be redeemed on May 28, 2027, if chosen, but this is not the same as debt maturity, which is a different financial concept. https://www.sec.gov/Archives/edgar/data/1166663/000119312526144027/d49235d20f.htm

A news scan found the note that on April 6, the 20-F filing Note 1 showed customer concentration as A at 17%, B at 12%, and C at 12%, which equals 41% when rounded. The prior percentage of 40.65% cannot be fully reproduced in the current filing, so it will be replaced with the rounded 41% until the original exact calculation can be found. The customers are anonymous and should not be identified by name based on their ranking alone. https://www.sec.gov/Archives/edgar/data/1166663/000119312526144027/d49235d20f.htm

For the first half of the year, the common earnings of TEN were $213.534 million. After subtracting a vessel gain of $37.870 million, the adjusted common earnings were approximately $175.664 million. This is the common income without the gain from the sale. The weighted average number of shares, which is a measure of the total shares outstanding, was 29.972103 million. The income of the preferred shares, which is $13.5 million, has already been deducted. The gain from the sale was allocated to restricted holders, but this was not separately adjusted, which is why the adjusted common earnings are an approximation. https://www.globenewswire.com/news-release/2026/09/10/3359616/0/en/ten-ltd-reports-profits-for-the-first-half-and-second-quarter-of-2026.html

As of June 30, the company's debt plus other financial liabilities, after subtracting financing costs, was 2102.177 million. This is compared to the bank-only amount of approximately 1966 million. The company also had 466.143 million in cash, 2054.657 million in equity, 3155.282 million in net vessels, and 470.050 million in construction advances. Note that these numbers include non-controlling interests, or NCI, which refers to the portion of a company owned by outside investors, and preferred equity, a type of equity that has a higher claim on assets than common equity. The equity per share alone may not accurately represent the common residual, which is the amount available to common shareholders. https://www.globenewswire.com/news-release/2026/09/10/3359616/0/en/ten-ltd-reports-profits-for-the-first-half-and-second-quarter-of-2026.html

On September 10, 2026, the issuer for H1, which ended on June 30, reported the following cash flow: operating cash flow was 281.167 million, investing activities used 248.312 million, and financing activities added 135.159 million, resulting in a cash increase of 168.014 million. The sum of operating cash flow and investing activities was 32.855 million, which includes money from selling vessels, but not from maintenance, and is referred to as free cash flow, or FCF, for short, which is a company's cash flow after accounting for maintenance expenses. Financing activities accounted for 80.445 percent of the net cash increase. https://www.globenewswire.com/news-release/2026/09/10/3359616/0/en/ten-ltd-reports-profits-for-the-first-half-and-second-quarter-of-2026.html