Phase 3: Practice · 8 min
Replace a vague “strategy” with a frozen hypothesis, consistent demo observations and an honest review.
By the end, you can:
- Write a setup that another person could follow.
- Separate test rules from claimed performance.
- Run a consistent demo sample without changing rules mid-test.
A strategy starts as a hypothesis
A trading idea is not an edge because it sounds logical or produced a few wins. Treat it as a hypothesis: under clearly defined conditions, a repeatable action may produce a measurable distribution of outcomes after costs. The test can reject the idea, expose missing rules or justify more research. It cannot guarantee the next result.
The six-part rule card
Every term must be observable. “Strong trend,” “good candle” and “looks ready” are not reproducible until they are defined.
An example of rule syntax—not a recommendation
Suppose a learner wants to study pullbacks in demo mode. The card might specify one instrument, one timeframe, an objective definition of prior swing direction, a precise pullback condition, a closing-price trigger, a structural invalidation point and a fixed risk calculation. That language makes the idea testable. It does not show that the idea has an advantage.
If a rule changes after every loss, the observations no longer describe the same plan. Finish the pre-defined block, review it, then change one variable and label the next block as a new version.
Collect clean demo evidence
- Set the scope. Choose a realistic observation window and record every eligible setup.
- Use demo funds. Platform fills and emotions may differ from live conditions, so treat results as practice evidence only.
- Record costs and misses. Include spread, fees, slippage assumptions and valid setups you skipped.
- Score adherence. A rule-breaking win belongs in the process-error column, not as proof of the plan.
- Review the distribution. Look at average R, variation, drawdown, costs and execution errors—not win rate alone.
What demo results cannot tell you
Demo execution can differ from live execution, and a historical sample may not persist in another market condition. Selection bias, tiny samples and changing rules can make results look stronger than they are. “Promising enough to study further” is more honest than “proven.”
You can organize a practice block with the Demo Session Planner, but the written rule card remains the source of truth.
Demo exercise
Practice: create version 0.1
Write one six-part demo rule card. Ask someone else—or your future self—to identify every subjective word. Replace each with a visible condition. Save the card as version 0.1 and choose one variable you will deliberately leave unchanged during the first test block.
Knowledge check
Answer two questions
Choose an answer and feedback appears immediately. Both answers mark this lesson complete in this browser.