Two scores for every decision
Two drivers reach their destination: one followed the rules and one ran a red light. The same outcome does not imply equal decision quality. Trading likewise has both a result and a process. We will build a journal that records them separately. The goal is not to excuse losses, but to distinguish a planned loss from a decision error. A favorable result can follow a poor decision, while a disciplined decision can still lose money.
Financial outcome ≠ decision quality
One setup, five definitions
Choose one idea and write context, prerequisites, entry trigger, invalidation and loss budget. Strong trend is not an observable rule until the data and timing are defined. Save the pre-decision screenshot. Do not overwrite the original explanation after seeing the outcome; an improvement belongs in a new dated version. Even one clear sentence is more testable than several paragraphs of vague description. Another reviewer should not have to guess what qualified as a valid decision.
Context • setup • trigger • invalidation • budget
Separate before and after
Before the decision, record an identifier, timestamp and timezone, symbol, rule version and checklist responses. Afterwards add execution prices, costs, result and errors. Feeling rushed can be useful context, but cannot replace measurable conditions. Do not record account credentials or unnecessary personal data. Each field should support decision review or error diagnosis. Keep the post-trade screenshot beside the original pre-trade image rather than replacing evidence of what was known at the time.
Before: inputs and rule After: execution and result
A bad win and a valid loss
The first imaginary trade earns two R but violates an entry prerequisite. It fails the process check. The second loses one R while following all written conditions; that loss is not automatically an execution mistake. Keep the financial results unchanged. The difference belongs in the adherence score. Rewarding an unplanned winner can reinforce the very error the journal should reveal. Separate outcome and process columns so the distinction remains visible during review.
+2R with violation: process failure -1R within rules: permitted loss
No trade is also data
A journal of executed trades alone omits rejected opportunities. For selected no-trade observations, record the time, missing condition and rejection reason. Review whether abstention followed the rule or had no explanation. This is not a demand to trade more; a sound decision to stay out should also be reviewable. More trades do not independently prove quality. In your exercise, include a rejection for high costs and another for insufficient data.
Rejected observation + reviewable reason
Review one change at a time
Eight compliant decisions out of ten means eighty-percent adherence, not win rate. Classify the two errors and choose one specific improvement. Keep the next version separate. Apply safety stops when needed, but do not rewrite history to improve appearances. If early entry is the recurring issue, clarify the closed-bar requirement and test that change next. Changing size, exits and every filter simultaneously makes it difficult to learn which modification actually affected the result.
8 / 10 = 80% adherence One controlled revision at a time
Lesson six assignment
Record two imaginary trades and one no-trade observation. Include a rule-breaking winner and a compliant loser. Separate before and after for each row. A good answer explains why profit cannot replace process quality. Then create a versioned one-page playbook. The next lesson takes this version into a demo experiment with a clear endpoint, safety stops and review criteria. The accompanying worksheet contains both examples and blank rows for your own practice.
2 trades + 1 no-trade observation Output: one-page playbook
