A review answers a question, and the question has to match the size of the record. One row can show that a step failed. A handful can show that the charter does not fit how you work, such as a look that misses the candle close. Neither shows what a rule produces. That takes enough signals of one setup recorded in R, and this manual's floor for reading a setup's results is at least 100, a floor and not proof, read against pass criteria you wrote before the count started. Reviews go wrong when they ask what a rule produces after a week and answer it with the week's result. A broken step can be fixed today, and the fix is not evidence of a better setup. This chapter ends in an edit to your charter, version 2, made from the questions your log is big enough to answer.
Judge each closed trade twice, and keep the two apart. Did it follow the plan written before the order? Setting the ticket beside the plan, field by field, answers that. And what did it return in R, net of costs? A trade can close above zero with its stop placed late, and one can stop out at its planned price with every step followed. The first is a broken step. The second is a followed plan whose worth one trade cannot settle. A pass gets the same treatment: it is judged by the rule written before it, not by where price went. Direction alone does not give a feasible entry, an invalidation, costs or a net result, so a coin that moved after you passed is a reason to reread the rule, not proof the pass was wrong.
A change to a rule makes a new version of it, and its count starts again. Old rows stay in the log, labelled by version, and answer questions about the old version. A change to execution or timing, such as when you look or how the stop goes on, leaves a setup's signal rules as they were, but it can change entry prices and costs. Label the rows before and after it and read them apart. The limit: a log shows that you broke a rule and which one. The cause is yours to find.
| Line | What you write | Test it must pass |
|---|---|---|
| Record | The dates covered. The decisions by kind: trades, plans that never filled, passes by the step they ended at, and signals skipped with no rule behind them. Cumulative R net of costs, its high, and the fall from it against your pause trigger. | Is every shortlist coin, plan and pass in the log, each with its reason written before the order? |
| Process findings | Each rule break, with the rule and the step, judged from the original plan and written apart from the result. Passes counted by reason. | Can you point to the plan field each break departs from? |
| Charter, line by line | For Markets, Timeframes, Risk ceiling and Pass rule: stays or changes, the old text, the new text, and the reason. | Does every change name its reason, rows from the log or a fact about your situation, and does any change raise exposure? |
| Awaiting a test | Each change the rows cannot support: the test, the data held back, the pass criteria written before it starts, and whether it runs on paper or at the smallest size. | Were the pass criteria written before the test started? |
| Not yet answerable | For each setup, the signals taken and recorded in R, by version, against the floor of 100, and the question that stays open. | Does any reading of a setup's results rest on fewer rows than the floor? |
| Next review | The charter's version and date, which rules changed and so start a new count, and the first thing the next review will check. | Is the old charter kept beside the new one? |
Objectives
This step supports the review step of the decision cycle: what to change in the process, and what to leave alone. It comes last because it needs what the other chapters produce: your charter from chapter 1 with the cost test chapter 2 added, the budget and pause trigger from chapter 3, and the plans and tickets from chapters 8 and 9. The log is the trading journal page's two tables, one row per decision, passes included.
- Set a ticket beside its plan and tally the rule breaks, judging the plan apart from the result.
- Work out R net of costs, cumulative R and the fall from its high against your pause trigger.
- Count passes by reason, and signals skipped with no rule behind them.
- Sort a proposed change into supported now, test first, or not yet answerable.
- Write the charter again, line by line, as a new version.
The 1 to 2% planned loss and the 3x ceiling are house constraints, and a review does not move them: the sizing page says to keep the planned loss the same or lower after a good run, never higher. The 100-signal floor is this manual's choice, and the testing page shows how wide the range still is at 100. The pause trigger is yours, set in R before you trade, and the trading journal and risk limits pages use 6R as an example only. How often you review is yours too. With a log of only a few decisions, the review is a process check: tally the breaks, repair the step a break points to, and give no verdict on any setup.
Practice is the same loop on charts that have already printed. Reveal a chart only up to a decision point, write the plan or the pass, advance one closed candle at a time, and finish the log row before you look further. The testing page's rules apply: act only on a closed candle and enter at the next open or later, record a level as it stood then, count the stop when one candle touches both stop and target, and take every signal the rules give. Staying on paper is a complete next step. A smallest-size trade is an option only when your pass criteria are met and that size fits your planned loss.
Hypothetical walkthrough: a charter revised from six weeks of rows
Hypothetical walkthrough. Invented prices and an invented outcome, chosen to teach the decision. Not a recorded trade, and not evidence of how often this setup works.
Assume the reader from chapter 1. Charter version 1 names two coins on one exchange's perpetual list, decides on the daily candle with one look a day just after its close, plans a 1% loss at a 2x ceiling, and has a pass rule that chapter 2 extended with the cost test. Their pause trigger is 5R. After six weeks the log has 37 rows: 6 closed trades, 2 plans that never filled, 26 passes, and 3 signals that fired at the daily close with no look and no rule behind the skip. The passes ended at observe 8 times (a coin labelled unclear, so any plan that needs a trend passes), at the shortlist 9 times (the cost screen), at the invalidation 5 times (no reason on the chart) and at the risk budget 4 times (3 on the group cap, 1 on a size under the minimum order). Cumulative R is 1.8R under its high, short of the 5R trigger.
The tally comes first. On 3 of the 6 closed trades the stop went on minutes after the fill instead of being attached. That shows a step failed, and nothing about what the trades earned. The ticket's stop row gains "place the stop with the entry or before leaving the form".
The 3 signals nobody looked at all came at the daily close, which falls in the middle of the reader's working day. The Timeframes line promised a look just after it, and the log shows the promise was not kept. The look moves to the evening. A later look means a later entry, so the entry window is written in candles from the new look and the room test is rerun at the price they would get then.
All 9 cost-screen passes were one coin. Recomputed on its fact sheet, that contract's costs break the one-fifth line at the stop distances its maps give. That is arithmetic on the sheet, not a result. The coin leaves the Markets line, which now holds one coin until another has a fact sheet and a template (chapters 2 and 4). The 8 unclear passes show the rule was followed. Whether it screens usefully is open, and whether a daily look at that chart is worth the time is a scope question for the next review.
Twice, after a cost-screen pass, the coin moved the way the plan wanted, and the reader proposes dropping the one-fifth line. The rows do not support it. Those passes were judged by a rule written first, and a move alone settles no entry, invalidation or net result. The question is fair, so it goes under Awaiting a test: per setup, paper the plans the screen rejects, costs included, against pass criteria written first, and read it at the floor. That bears on the screen and does not by itself show the whole process improves. A result can change their own pass rule. It cannot move the 1 to 2% or the 3x.
The ceiling and the pause trigger stay. The 4 risk-budget passes are the budget doing its job, and no row suggests moving either number. The count is 5 trades of one setup and 1 of another, so no reading of either. Version 2 is dated and version 1 stays beside it. The stop edit changes execution and the look edit changes entry timing, so rows from before and after each are labelled and read apart. The setups' signal rules are unchanged.
The outcome was chosen to show edits the rows support, one they do not, and entries left unchanged.
Reading list
- Trading journal: the fields for a trade and a pass, and the three speeds of review.
- Expectancy and drawdown: what a run of stop-outs does, and the pause trigger in R.
- Testing a setup: how a changed rule is tested before you adopt it, and why 30 trades are not enough.
- Portfolio risk: the group and total caps, so the log can show which plans a cap turned into a pass.
- Risk limits: the limits and the pause trigger as a tick-box list.
Read the journal page first for the fields, then expectancy for why a short record misleads, then testing for what a changed rule has to go through.
This chapter cannot settle whether any setup works. A log of a few dozen decisions cannot separate a sound plan from luck, this manual cites no test of any setup, and the count takes as long as your shortlist takes to produce signals. The answer may be that a setup's record never reaches the floor. The path does not end here. The revision is version 2 of the charter, and your next action is to take the next eligible decision through it and record whether the repaired step happened. The question to carry forward: what does the first line of the new charter let into the cycle? Trading mindset is where the revision goes.
You should now be able to
- Set a ticket beside its plan and tally rule breaks, judging the plan apart from the result.
- Work out R net of costs for a closed trade, and the fall from cumulative R's high against your pause trigger.
- Say why a pass is judged by the rule written before it and not by what price did next.
- Say how many rows a reading of one setup needs, and what a changed rule does to its count.
- Sort a proposed charter change into supported now, test first, or not yet answerable.
- Produce: your decision log, passes included, and a one-page process revision that updates your charter, using the table at the top of this page.