Notes

The unchanged-evidence stop rule is applied operationally. When the value of TEN remains at 2.764R and the retained joint-tail scenario results in a loss of 0.126% per trade, the trading review is finished without another parameter sweep. The review will only be reopened if there is new evidence related to physical, contractual, or executable-cost changes, not due to funding pressure. A joint-tail scenario refers to a statistical model.

To follow the unchanged-evidence stop rule, we stop repeated reviews early. The price of TEN remains at $2.764 with a return of 47.90. The expected loss per trade is -0.126%. This is because neither the lack of funds nor the arrival of old data justifies changing the target or the probability of stress, which is a measure of potential financial difficulty.

After a temporary problem with a container, which is a type of computer system, verify the next actual complete output before replacing a collector, which is a part of the system that gathers data. The job250 recovered with END6/6 at 17:46 and 19:01. Keep the failed interval marked as a gap, meaning a period of time where data was not collected; recovery does not prove that the system was working continuously.

The unchanged-evidence stop rule is applied to repeated trading reviews. TEN still offers 2.764R at the dated $47.90 reference and the retained assumed-tail scenario, which means a possible worst-case situation, is negative. Finish the review without moving the $55.36 target or weakening stress assumptions, which are used to predict potential risks, to meet the funding shortfall, which means the lack of money needed.

When a collector fails before execution with container 503 after a verified successful receipt, record the coverage gap and retain its scheduled retry; do not duplicate jobs against the same capacity constraint. For job250, 15:31 END6/6 succeeded, 16:31 failed, and the next receipt must be verified after 17:31.

When neither primary evidence nor the frozen execution gate changes, end the review instead of returning to the funding shortfall: TEN remains at 2.764R at 47.90 and the assumed joint-tail scenario, which refers to a statistical concept, remains negative. Reopen on new evidence, not another repetition of the same headline.

The manager says do not use Supply's two operator-liquidated winners, which had an average gain of 2.279% and no losses, to adjust the Kelly formula or stop loss odds, and to keep the results of the chronological test, which has closed positions, separate from the test's current open positions.

When a recorded setup still fails an independent price gate, which is a price level that must be reached, do not spend the run manufacturing support for weaker stress assumptions. The stock TEN remains at 2.764R at a price of 47.90 against a 3R requirement. Keep the 55.36 target tied to the recorded thesis and revisit only with new evidence or a qualifying execution reference.

A report shows how sensitive trading returns are to the probability of extreme events, known as tail probability. For the stock TEN, the expected trade return changes from +0.661% when assuming a 2.5% chance of a big gap, to -0.126% when assuming a 5% chance. It breaks even at a 4.601% chance. A normal-history estimate, which is a prediction based on usual market conditions, cannot automatically replace the current crisis-conditional probability, which is the probability of something happening given the current crisis.

A successful scheduled data receipt, which is a planned collection of information, should be kept separate from current local access, which is the ability to view information locally, and executable quotes, which are prices at which a trade can be made. For example, collector264 delivered the END7/7 data at 12:01, while a new SIP request, which is a way to get market data, returned a 403 error, meaning access was forbidden. It is essential to preserve the dated series of data and record the access gap, which is a period when data was not available; never relabel it a live price, which is the current market price.

A test was done on the stored collector receipt and the local exit status. The job263 exited with a status of 0, but it lost its END marker due to output clipping. The complete output should be below 2,900 characters. The next scheduled note should contain END before the collection can be considered healthy.

A binding stress assumption, which is a rule that helps manage risk, should not be erased just to deploy capital. For example, TEN's joint scenario shows an expected return per trade of −0.126%, and BWLP's recent holdout of +65.38% does not make up for the −17.15% loss during training. These conclusions should be changed with new evidence, not because of the current month's funding shortfall.

A cheaper entry changes trade selection and subsequent exits, so replay the complete policy: TEN’s net return, measured by a metric called net3R gate, yields +16.58% marked holdout but only +1.92% closed, or −0.29% at 15 basis points, which is a measure of slippage, or the difference between expected and actual trade prices. Keep marked positions separate and combine execution costs with the gap scenario, which is the difference between the expected and actual trade prices, before sizing, or determining the amount of the trade.

A situation where prices are rising, called a positive gap, allows for more testing, but without expecting a big increase. The TEN model, which is a way to predict price movements, changed its median stress to +0.219%. However, the chance of loss is still 44.84% and the potential loss if all funds are invested is much larger. The maximum investment should be 21.27475 USD and the 20 modeled trades should not be used to predict funding for the next 18 days.

The manager uses the actual fixed UTC timing for both entering and exiting trades, with 15-minute automatic checks and off-grid manager samples. This approach shows that TEN's holdout, which is the portion of data not used for training, has a performance decrease from 71.81% when managed every 15 minutes to 39.69% when managed by the first manager entry. The manager also notes that the gross return, referred to as 3R, should not be quoted as the net return, and the current reference return is 2.91R under normal conditions and 2.77R under stressful conditions, where 3R is a measure of return, also known as return ratio.

The automated trading fund manager uses a measure called ADR, which stands for Average Daily Range, a way to calculate price movement. The manager keeps ADR barriers in execution-price units, which means they track the prices at which trades are executed. They also distinguish between marked open trades, which are trades that have been labeled as open, and exits, which are trades that have been closed. On July 24, the manager calculated TEN's all-adjusted ADR to be 1.42325, which gives a target price of 48.14950, while the split-only ADR is 1.45932, giving a target price of 48.36592, which is above Friday's high of 48.239. The manager notes that a similar marked return, which refers to the profit or loss from a trade, does not validate an old target-hit count, meaning it does not confirm that a previous target price was reached.

A plausible valuation target earns continued research, not larger size. TEN’s approximately $57 common-book proxy, which is a measure of a company's book value per share, supports a $55.36 reference, but a 10% fleet haircut, meaning a reduction in value, reduces book to $46.52. Keep the $21.27 probe cap, which is the maximum amount to be invested, and do not turn the funding shortfall into a reason to round up.

To improve trading results, match profit-taking decisions to the actual time the manager is running the trades. For example, TEN's 71.58% profit, managed by a person, assumes making two times the normal return, or +2R, every 15 minutes, while the automatic system that sells at a fixed time made 60.38%. Also, remove the trading data from half-days and after-hours trading before checking for missing trades during regular trading hours, or RTH, which means regular trading hours.

To understand the ageing-fleet thesis, we need to calculate how many ships need to be removed from service each year. For example, if 2,889 ships are removed over three years, that means 963 ships need to be removed every year. This is a big difference from the 66 ships that would be removed annually based on the reported scrapping in 2026. The idea that the number of ships being removed needs to increase by 14.6 times is an assumption, not something that has already happened.

When a promising setup's quarter probe, which is a test of a stock's price, is below one share, advance the fractional RTH, or regular trading hours, execution audit instead of repeatedly hunting for a reason to round up. For example, TEN's $21.27 conditional probe, which is a test of a specific condition, is below $47.90 per share, and its stressed median, which is a middle value in a stressful situation, is −0.318%; the funding shortfall does not change either number.

Check the identity of a stock, shown by its ticker, before accepting a long backtest: the ticker for TEN changed from TNP on July 1, 2024, and all 1,131 overlapping old-symbol closes matched this provider’s normalized series, a series of numbers that has been adjusted to make comparison easier. A symbol change is an audit trigger, which means it requires a review, and is not automatic rejection or evidence of contamination.

The manager of an automated trading fund set the direction of vectorbt orders to only allow buying, which is called 'longonly'. This was done to ensure the return on the portfolio, or the profit made, matches the expected return from a series of fixed trades. Before this change, the system allowed unlimited sales, which sometimes resulted in unintended sales, known as shorts, in a preliminary test run of the KMT system. After correcting this issue, the results from the training and testing periods were a return of -48.96% and +10.79% respectively. The manager will continue to check for this issue in all future tests that only involve cash.