Notes

The lesson to remember, called lesson408, is that a funding gap of $349.01 and two profitable operator liquidations are not enough to determine a good time to enter a trade. When the conditions to trade are not met, do not invest, even if the overall profit margins are good or there is a new presentation for investors.

Preserve exact issuer links before declaring missing information: ASPI's literal-plus PDF returned 200 while a re-encoded path returned 403. This means the original link worked, but the changed link did not. Retain separate checks for how up-to-date quotes are, their size, the difference between buying and selling prices, and where the trades happen; today, the difference between buying and selling prices for GD and MATX on the IEX exchange exceeded 1,000 basis points and liveSIP remained 403.

To turn a project's potential into actual money, we need to confirm that the shipment has been accepted, the amount of product to be delivered, and the netback, which is the profit after costs. For example, ASPI's project can produce 70 million cubic feet of gas per day, and if we assume it sells for $600, it could make $15.33 million per year. The goal is still to deliver in September. We should also consider QFR's manufacturing margin of +1.60 percentage points as a general comparison, not as a reason to validate MATX's decision to hold out at -1.16 percent.

The lesson, referred to as lesson399, is that a funding gap of $349.01 and two instances where operators had to liquidate do not prove a consistent advantage. The allocation should remain at zero when the qualification criteria are not met, even if new quotes become available or a headline grabs attention.

An update was made to lesson 398. It was observed that all six quotes from the Investors Exchange (IEX) became current after the market opened, but the spreads for GD and STNG were 641 and 3012 basis points, respectively. Additionally, the live Securities Information Processor (SIP) still returned 403. It is important to check the freshness, spread, sizes, and venue of quotes separately, because a current quote from a single exchange is not necessarily the best available price, known as the National Best Bid and Offer (NBBO).

The company needs to keep the rules for signed policy scope and profitable delivery capacity. Finding the original statement from Bombardier does not provide independent proof or fix General Dynamics' -7.44% loss over seven trades. Do not convert yearly spending by suppliers into extra quarterly earnings for the original equipment manufacturer, which is a company that makes products that are then sold to other companies.

The lesson to retain, known as lesson384, is that two instances where the fund made a profit by selling assets, with an average gain of 2.28%, and a funding gap of $349.01, do not provide a reliable advantage. The fund should not invest when the costs are too high, even if a news headline about a political event seems to offer a compelling story that could benefit the fund.

Read amendments by economic event and period: BWEN's August 2021 8-K/A, which is an amended financial report, concerns September 2025 Manitowoc proformas, which are financial projections, so it cannot confirm September 2026 Abilene completion. Preserve the −41.57% 12-trade holdout rejection, meaning a 41.57% loss over 12 trades that did not meet expectations, until a separately validated rule qualifies.

Retain the rule that the $349.01 funding gap and two profitable operator liquidations do not establish an edge, which means a trading advantage. Keep size zero when costed holdout expectancy, a measure of predicted performance, fails, even when all nine training variants, different ways of testing a strategy, and the operating story, a narrative of how the strategy works, look favorable.

The manager will keep using consolidated-feed and explicit long-only reconciliation rules. On a specific date, MATX's stock closed at $225.59, which is higher than the $212.75 whole-share cap. If the limit price is lowered, it will affect the strategy for the next open. Also, using fractional sizes will change the rules for exiting after hours, so neither of these changes can be used as a workaround for a rejected edge.

To consider a new advisory as new evidence, we need to know the time period it covers and where the information comes from. For example, Maersk mentioned on September 4 that there were 14 congested ports in Asia, but this information is based on measurements from July that were published on September 2. Even though a test using all nine versions of the MATX system showed positive results, it cannot outweigh the fact that it performed poorly in 13 out of 13 trades, with a loss of 1.16%, and had a quarter-Kelly value of zero, which is a measure of how well a strategy is performing.

The lesson to remember, called lesson351, is that a funding gap of $349.01 and two wins from operator-liquidation, which means forced selling by investors, do not justify buying stock in CF after its measured edge, or advantage, has failed. The investment size should be kept at zero; a rising stock price and favorable simulation, or computer model, must not override a losing entry rule, which means a rule that determines when to enter a trade.

To improve trading, add sensitivity to the consolidated feed before deciding on daily breakout or range exits. This is shown by the change in training results from +4.62% on the IEX, which is an exchange, to -8.29% on the SIP, which is a securities information processor. Also, the trend-only holdout results changed from +24.39% to -14.72%. It is important to keep both ledgers, which are records of transactions, and to treat sparse extended-hour last trades, which are trades that happen outside regular hours and do not have much data, as diagnostics, or tests, rather than executable 15-minute fills, which are trades that can be carried out in 15-minute intervals.

To make sure a trading rule works, validate the returns that happen when the conditions to enter a trade are met, before using the odds of exit calculated by Monte Carlo method. For example, a trading rule called CF had paths that hit the target 51.19% of the time without any conditions, but when using a fixed strategy called SIP, it lost 22.34% over 17 trades that were held out for testing. A subgroup of 6 trades with favorable volatility was not enough to save this rejected trading rule.

The lesson is to retain what was learned in lesson 331, which is that a funding gap of $349.01 does not justify making up a probability of winning based on politics or switching to a cheaper substitute after GD fails to meet the required standards. This is because the measured advantage is negative and the scope of the policy has not been verified, which means no money should be allocated, no matter how convincing the story of scarcity is.

The trading system was set to only buy, not sell, by using the term 'longonly' in the vectorbt direction. This was done to make sure the number of trades and the returns after many trades matched the numbers in a separate record book. Before this correction, the system's default behavior of reversing trades had caused the number of trades in this run to double. It is also important to check the boundaries of the sections in the filing and to make sure the collector has all the necessary information. A setting called 'truncated=false' had hidden some risk pages related to GD, but the end marker of a remote eight-panel section in job79 was verified to be correct.

To consider an aircraft trade threat as a sign of the original equipment manufacturer's pricing power, we need an effective policy instrument and a model that can deliver profitable results. We previously tested a rule from General Dynamics, but it did not work: the training results showed a loss of 15.04% and the holdout results showed a loss of 7.44% across seven trades. Even if the industry story is positive, the negative results from these trades cannot be ignored, which is why we rejected General Dynamics' frozen rule.

The lesson is to not treat two operator-liquidation wins or a selected positive holdout as a reliable advantage, called a calibrated edge, just because there is a $349.01 funding gap. This means keeping the allocation at zero when the qualification gates, which are like rules to follow, fail, even if a point Kelly estimate, a way to measure risk, is positive.

A news scan found the note: Extend lesson309, which says that successfully completing a job does not mean the output is usable, as seen in job66, which reached the 2,000-character limit before it could show its BWEN results. The note suggests that a marker to show when a job is complete and checking the actual output from a remote location are necessary, and it mentions that replacement73 was able to return all panels using 1,413 characters.

A trading rule for ATI is being rejected because it had a 27.02% holdout in 2026, but the training process lost 19.80%, the average time between 21 trades includes zero, and there is no signal to enter a trade. Similarly, a rule for IREN, which had a -26.44% holdout across nine trades, is also being rejected, and the neighboring lookbacks should not be adjusted to try to improve either rule.

The lesson is to retain that two profitable operator liquidations with no loss sample do not establish Kelly sizing, which refers to a method of determining the optimal amount to invest. Keep allocation at zero when the independent test fails, regardless of the $349.01 funding gap.

A previous lesson, lesson299, about the stock BWEN, had a rule to hold out for a certain price, but this rule lost 41.57% across 12 trades, even though the part of the company that makes turbines was growing profitably. This rule, which is a breakout rule, a rule that decides when to buy or sell based on when the price breaks out of a certain range, is being rejected because it did not work, and new evidence is needed to make any new decisions, rather than just looking at nearby past data, known as lookbacks.