# Discretionary Strategy Playbooks Build a repeatable setup description with measurable exclusions. Use this workbook alongside the course. Write your answers before opening the solutions. Practical work is self-reviewed; scored knowledge checks are in the Academy. ## 1. Setup specification ### Context conditions Context describes the conditions under which a setup is considered, such as a defined range or trend state. Use observations that can be recorded at the time. Context should narrow the question, not become a vague justification added afterward. ### Entry trigger An entry trigger is a specific event that makes an order eligible. State the price stream, interval and whether confirmation requires a close. Separate signal time from order submission and fill time. ### Invalidation Invalidation describes when the idea no longer satisfies its premise. A protective order implements part of that plan but may fill differently from its trigger. Record both the analytical condition and execution behavior. ### Time-based expiry Expiry ends eligibility when a setup has not entered within a specified window or relevant conditions change. A stale working order can create exposure after the original context has disappeared, so cancellation rules belong in the playbook. ### Worked example A breakout setup requires a completed bar above a reference level. The bar temporarily trades above but closes below. Under that rule there is no confirmed trigger, even if an intrabar chart looked promising. ### Independent exercise Write the context, trigger, invalidation and expiry for a hypothetical setup without using words such as obvious or strong unless defined numerically or procedurally. My inputs and assumptions: My calculation or decision: Evidence that would change my conclusion: ## 2. Trade management ### Initial protection Initial protection should reflect the stated invalidation and affordable exposure. Compute quantity from the chosen assumptions before entry. Moving protection farther away solely to avoid recognizing a loss changes the original risk decision. ### Target logic A target is a planned exit condition, not a promised outcome. Define whether it is a fixed distance, a reference area or another observable rule. Costs and actual fill availability affect realized reward-to-risk. ### Trailing rules Trailing rules need a precise update schedule and monotonicity convention where applicable. A backtest must not use an intrabar high to tighten a stop before that high was observed. The price path matters when both updates and triggers occur within one bar. ### Partial exit accounting Partial exits change remaining quantity and realized P&L. Record every execution and cost, then calculate the whole trade consistently. A favorable first exit does not determine the result of the remaining exposure. ### Worked example Two units enter at 100. One exits at 102 and one at 99, multiplier one and no costs. Total gross P&L is 2−1 = 1, not 2. The first partial exit is only part of the trade. ### Independent exercise Repeat with total costs of 0.6 and explain what belongs in the final trade record. My inputs and assumptions: My calculation or decision: Evidence that would change my conclusion: ## 3. Exclusions ### Event conditions Event exclusions should specify which information or execution conditions invalidate the study assumptions. A blanket description such as avoid news is ambiguous unless the relevant events and time windows are defined. ### Spread limits as a hypothesis A spread threshold can be a research hypothesis about feasibility. Record how spread is observed and whether order size matters. Select the threshold before evaluating results rather than excluding costly losses afterward. ### Missing data Missing required data should produce an unavailable decision, not an invented input. A playbook must define whether it waits, cancels or remains inactive when the necessary observation cannot be established. ### Conflicting exposure Conflicting exposure includes multiple positions driven by the same risk factor. Assess the account-level plan before adding a setup that looks independent on a separate chart. Different labels do not guarantee diversified losses. ### Worked example A learner excludes every losing trade whose spread was high, but keeps winners from the same conditions. This is an outcome-dependent filter, not a valid implementation rule. ### Independent exercise Define a symmetric spread exclusion that can be applied before entry to winners and losers alike. State what data must be retained to verify it. My inputs and assumptions: My calculation or decision: Evidence that would change my conclusion: ## 4. Review ### Store before-and-after evidence Store a before-entry record with context, trigger and planned risk, followed by the actual result. A post-trade screenshot alone can hide what was visible when the decision was made. ### Classify valid losses A valid loss follows the rules and loses within the observed execution process. It belongs in the sample. Calling every loss a mistake prevents honest estimation of the strategy's outcome distribution. ### Separate rule changes from execution errors Separate strategy changes from execution errors. A rejected order or incorrect size is not the same issue as an eligible setup that loses. Different problems require different corrective actions and evaluation records. ### Compare playbooks net of costs Compare playbooks using the same accounting and sample conventions. Net results, exposure, drawdown and uncertainty matter alongside win rate. A method with more winners can still have worse average losses and lower net expectancy. ### Worked example Playbook A wins 60% with average win 1 and average loss 2. Before costs, expectancy is 0.6×1−0.4×2 = −0.2. A high win rate alone does not establish a favorable result. ### Independent exercise Calculate expectancy for a playbook winning 40% with average win 3 and average loss 1, then subtract cost of 0.3 per trade. My inputs and assumptions: My calculation or decision: Evidence that would change my conclusion: ## Course project Create two contrasting playbooks and replay them on a reserved chart sample. ### Self-review rubric - Concepts and reasoning: 25% - Calculations, data and evidence: 30% - Process and risk controls: 25% - Limitations and communication: 20% Record one correction and one next practice task. This rubric is not automatically graded. ## Worked solutions ### Exercise 1 A valid specification lets another learner classify the same observations. It identifies exactly what data is needed and when. The result is a testable hypothesis, not evidence that the setup will make money. ### Exercise 2 Net P&L is 0.4. The record includes both exit quantities and prices, their times and the allocated costs. Reporting only the favorable exit would overstate performance. ### Exercise 3 Specify the measurement source, time, size convention and threshold before outcomes are known. Retain the observed spread and eligibility decision for every candidate, including excluded cases. ### Exercise 4 Gross expectancy is 0.4×3−0.6×1 = 0.6. Net expectancy under the stated cost is 0.3. These are sample assumptions, not guaranteed future probabilities. ## Further reading - https://www.cmegroup.com/education