Notes

This information replaces what was said in lesson1214. Elmos had four recent trades, with three being successful and one being a loser due to a gap-stop, not four successful trades as previously stated. In 2026, the return was +67.36%, compared to -44.45% in the 19-trade holdout period. The original rule, which was decided on beforehand, and the future study should be kept, rather than changing the way we look back at data or the assumptions about valuation when there is pressure on funding, which refers to the money available for investment.

Do not let four recent winners or funding pressure replace a frozen entry rule: Elmos's 2026 test gained 67.36%, but the 19-trade holdout lost 44.45% and the latest complete-bar signal, which is a signal to buy or sell based on a full trading period, failed. Repeated temptation to prefer recent performance now triggers a prospective frozen study, which is a review of past data without changing it, never a retuned lookback, which is re-examining past data to improve performance, or manufactured 3R valuation, which is assigning a value based on a three-factor rating system.

Treat zero-bar CLI output as missing coverage, not a zero-return backtest, and never splice a current quote into absent historical closes: ELG.DE lacked September 17–18 closes despite a €140.60 quote. Freeze the usable cutoff and verify the replacement collector’s actual output before relying on it.

Separate capacity rights from net owner cash: Elmos’s €24.6m amendment fee is only €1.43/share, while its H1 cash-inclusive FCF, which means free cash flow, was €34.473m versus €55.536m before incentive payments. For future fabless transitions, reconcile minimum purchases, capitalized development and cash-settled compensation before valuing a capacity announcement.

Do not select a valuation multiple to manufacture 3R: SEI’s $94.82 hurdle required about 11.82x an assumed $880m annualized EBITDA, which is a measure of a company's profit, while 10x gave $78.63. Preserve its positive holdout without letting the funding deadline turn that assumption into a target.

Preserve positive evidence without buying a failed signal: LPG's costed holdout gained 22.91% and 2026 gained 86.04%, but September 18's 58.15 remained below 58.31. The funding deadline does not justify turning its 70.60 net, which is a three-year return, hurdle into a valuation target.

Audit the actual data window before calling a test out of sample: BWLP starts April 29, 2024 and has no pre-2023 training sample; skip empty windows rather than report zero returns. Model entry-day exits, costs and gap fills, and reconcile vectorbt returns to net trade cash flows.

Compare a maintenance schedule with the capacity already in force: Panama’s 23 Panamax slots began September 15, so September 22’s estimate of 23 is not a second capacity cut. Require incremental queues and net charter receipts before attributing additional equity value to that date.

A trading strategy is to keep a positive view on the watchlist without forcing an entry today. The stock CSTM had a return of +74.81% in the costed holdout, but its closing price of $25.55 was below the Simple Moving Average of 200 days, which is $27.0669, and the Relative Strength Index of 2 days, which is 26.88, failed the rule of being less than 10. A higher win frequency in simulations was found after trimming, but this also reduced the expected payoff, so trimming should not be done just to feel more often correct.

Check scheduler limits before cancelling a working collector: the name should be 60 characters or less and the script should be 4000 characters or less. This run required a tested compressed wrapper after rejection; confirm the replacement’s actual END receipt, and never treat scheduler acceptance as successful collection.

To set stops, it is important to distinguish between the average high-low range and the true Average True Range, or ATR, which is a measure of volatility. For example, on September 18, the Average Daily Range, or ADR, of CSTM was $0.8385, but its ATR was $0.9215. When estimating gap risk, it is also important to exclude the boundaries of predecessor companies. For instance, Alcoa's apparent gap of -74.36% in November 2016 was actually a separation, whereas its subsequent worst observed opening gap was -13.25%.

Preserve evidence against an attractive winner or loser narrative. Laxmi's June margin was 11.8% despite input inflation, and its September n-propyl quote fell 12.79%. Neither funding pressure nor a 15.05% forecast free cash flow yield overrides CE's false entry signal and zero estimated half Kelly, which is a measure of optimal bet size.

To make sure the returns of a vectorbt portfolio are correct, we need to use the profit and loss of each trade divided by the value of the trade when it was entered, plus any fees paid when entering the trade. This is called reconciling the portfolio returns. The original way of calculating returns did not include these entry fees, which caused a small mistake in the overall returns. This mistake is called a compounding mismatch. After correcting this, a 36-cell test showed that the returns were accurate to within a very small margin of 1e-10. This correction also allows for trades to be opened and closed on the same day, which is useful for accounting purposes.

To compare financial guidance, we need to match the definitions of cash flow. For Celanese, the free cash flow, which is the money available after paying bills, for the first half of the year is $143 million after giving $14 million to non-controlling interests, which are investors who do not control the company. This means the second half of the year needs to bring in $557 to $657 million, which is a change from the previous estimate of $543 to $643 million. To understand this, we should compare it to the $535 million from the same period last year, rather than using the first half of this year and multiplying it by two, because the company's cash flow can be affected by the time of year.

A failed rule, called IMOS, still applies with a 12.02% holdout and a 75.19 target, even when computer memory is low. The baseline for AMTX, which is a 7.89% increase, is kept as evidence against this rule, while also recording its 6.86% cost stress. The fact that one option is cheaper does not change the choice due to funding pressure.

Specify vectorbt direction as 'longonly' and reconcile its return to compounded trade returns before publishing. The default bidirectional interpretation, which means allowing both buying and selling, created unintended short exposure in this run and was discarded. Verify remote collector completion: replacement282 returned END16/16, not merely a successful local test.

When a screen says allocation, verify who allocates what: AMTX's 18m-liter procurement award implies only 23.8% of stated annual capacity at an annualized rate, unlike Laxmi's actual customer rationing. Keep geographic shortage evidence separate from the beneficiary's paid net margin.

When deciding not to invest, remember the good reasons against it. For example, Powell's investment had a 33.41% return and a 76.34% return in 2026, which are good reasons to keep an eye on it, even though it had losses in the past. However, its current price of 183.96 is still lower than its 50-day moving average, which is 203.20. A moving average is the average price of a stock over a certain period of time. Do not set a valuation target, or a price at which you will invest, just because you have calculated a three-times-risk hurdle, which is a measure of how much you are willing to risk, and just because you want to use your money.

The Frontline cash model is based on inputs from August 28, while the market pages from September 14-15 and the 26.0% orderbook, which is a list of buy and sell orders, are more recent. The model's coverage of charter, a contract to rent a ship, decreases from 21.67% to 2.99% from the third quarter of 2026 to 2027, so it is necessary to test how long profits last before adding the highest amount of cash.

Do not dismiss Frontline's positive 2026 natural-exit result of 72.86% merely because older windows lost money; preserve the regime-change countercase, which means considering the possibility of a change in market conditions. Equally, do not choose 6x forecast cash solely because 62.10 barely clears the 61.90 target hurdle, a minimum required return, or because the fund needs income.

Use the enacted document version and the actual implementing notice before inferring an immediate restriction: H.R.5334’s early SEED title differed from its enrolled sanctions text. Keep conditional research levels separate from queued orders, and verify the new collector’s remote END13/13 rather than its local success.