A week can carry positive expectancy and still be graded red.

The bands

Three states, and the boundary between the worst two sits in positive territory.

The green band begins at a comfortably positive expectancy and means the system is in its healthy operating state — continue under normal rules, still subject to gate and brake authority. The yellow band covers the region between a thin positive figure and that healthy threshold, and its instruction is to trade selectively and not to expand risk. Below the thin threshold is red, and red begins while the arithmetic is still positive. The gap between zero and the red boundary is the part operators find counter-intuitive and it is the part doing the most work.

Why the boundary is where it is

A thin expectancy has to survive costs, slippage and estimation error before it becomes money.

The blended figure is computed from branch probabilities and expected values, and those inputs are estimates carrying their own uncertainty. A figure a few hundredths above zero is comfortably inside the range that ordinary sampling error could have produced from a system with no edge at all. Beyond the statistical point, the number is also being asked to absorb friction the model does not fully carry — spread variability, slippage on entries and exits, swap on held positions. An edge that thin is not a small edge. It is an edge that has already been spent, and treating it as tradeable means deploying risk against a figure that will not survive contact with execution.

What the grade is and is not

Red describes the evidence, not the outcome, and it does not know whether the week made money.

The status is computed from expectancy, which is a forward-looking property of the process rather than a record of what the account did. A week can be graded green and have lost money, because expectancy describes the average of a repeated process and a handful of trades is not that average. A week can be graded red and have been profitable, usually because one outsized result arrived in a week whose underlying probabilities were poor. Both of those are the workbook functioning correctly. The grade answers whether the system is still worth deploying; the balance answers what happened. Conflating them is the specific error the scorecard was built to remove.

• Green with a losing week is variance, and it is not an argument for changing anything.
• Red with a profitable week is the more dangerous combination, because it feels like a false alarm.
• The status is about the process; the account balance is about the sample.

Where the grade goes

A red week is respected rather than negotiated with, and the doctrine says so in one line.

The workbook's stated position is blunt: if the scorecard reads red, the operator respects the risk doctrine instead of negotiating with the result. The wording matters because negotiation is the natural response — the week can always be explained, and usually the explanation is true. There were three news days, the sample was small, one branch was misclassified, spreads were unusual. All of that may be correct and none of it changes the instruction, because the alternative is a status that carries authority only when the operator agrees with it, which is not authority. Diagnosis of why the week was red happens in parallel and does not gate the response.

One week is a reading, not a condition

The threshold governs the week; the pattern across weeks governs the system.

Treating a single red week as evidence of edge decay is the opposite error and it is equally expensive. Weekly samples are small and noisy, and a lone red week inside a month of green is ordinary variance that requires no structural response beyond the posture the status implies. What warrants a structural response is a distribution: repeated reds inside a period, or a status mix that has shifted from mostly green to mostly yellow across a month. That is what the monthly rollup counts exist to surface, and it is the level at which questions about weights, branch health and edge decay are properly asked.

The key idea

Setting the failing grade above zero is what stops break-even from feeling like success.

If red began at zero, then every marginally positive week would grade as acceptable, and a system slowly losing its edge would spend months in the acceptable band while the operator deployed full risk against nothing. Placing the boundary above break-even builds a deliberate margin into the grading, so that deterioration is flagged while there is still something left to protect. The cost is occasional weeks graded red that were, in the end, fine. That is the correct trade, and it is the same trade the gate makes with drawdown.