Weekly Trading Review: Metrics, Narratives, and Testable Hypotheses

A weekly review is the simplest structure that turns daily noise into cumulative insight. It sits between the immediacy of post-trade notes and the drift of monthly summaries, long enough to collect meaningful data yet short enough to adapt. When the review is anchored in a small set of metrics, a coherent narrative, and explicit hypotheses, performance discussion moves from vague reflection to directed learning. The goal is not only to judge outcomes but to refine behavior in a way that compounds.

Why the week is the right unit A week captures repeated conditions without overwhelming memory. Research on spaced reflection and pattern recognition shows that batching experiences improves signal detection and reduces recency bias. A weekly cadence also moderates emotional extremes that follow big wins or losses. The trader is able to hold both the statistical and the experiential: the numbers that describe the system and the story that explains context.

Three pillars: metrics, narratives, hypotheses A robust review stands on three pillars. Metrics quantify performance and process. Narratives organize experience into cause and effect. Hypotheses convert observations into tests that guide the next week. Treat these as a closed loop. Metrics reveal patterns, narratives propose mechanisms, and hypotheses specify the change to examine. Next week’s data confirms, revises, or rejects the idea, and the loop continues.

Metrics that matter Keep the metric set compact and durable. The aim is repeatable measurement that resists overfitting to a single week.

Risk metrics show whether capital is being put to work within tolerance. Track planned risk per trade as a fraction of equity, realized risk versus plan, and maximum intraday or weekly drawdown. These values indicate whether exposures drifted under stress or during streaks. A stable risk profile makes other signals interpretable.

Expectancy metrics express the quality of decisions independent of streaks. Use R multiples so trades can be compared on the same scale. Weekly expectancy is the average R across trades. It is heavily shaped by the right tail, so include the distribution of outcomes rather than focusing on win rate alone. Combine expectancy with variance to judge reliability; a volatile expectancy often reflects inconsistent execution or unstable market regimes.

Opportunity capture describes how well entries and exits harnessed available movement. Compare actual exit to a simple benchmark, such as a volatility-based target or trailing stop that could have been applied, and note the proportion captured. MFE and MAE, expressed in R, help distinguish timing errors from thesis errors. If MAE is small but exits are late, execution is the issue. If MAE is large and frequent, idea quality or invalidation rules need work.

Process adherence is the anchor. Count how often trades met pre-trade criteria, used the planned size, and respected mechanical stops. A simple 0 to 2 scale per trade works well. Zero if the rule was broken, one if partially met, two if fully met. Sum and normalize by trade count to produce a weekly process score. This number often correlates with emotional stability and reduces the temptation to overreact to short-term P&L.

A compact scorecard Convert metrics into a weekly scorecard that fits on a single screen. Include weekly expectancy in R, hit rate for qualified setups only, process score, and maximum drawdown. Add a short qualitative note about market regime, such as high volatility breakout conditions or mean-reverting ranges. The design principle is restraint. The scorecard should be stable across months so trends can be seen without moving the goalposts.

Narratives that explain, not justify Numbers signal where to look. Narrative explains why. Effective narratives are short and focused on mechanism. For each trade, a few sentences can capture the context, the trigger, the execution, and the decision points. Separate observation from interpretation. Write the facts first, such as time, level, liquidity cues, and order flow, then write the interpretation, such as why the breakout was legitimate or why the pullback failed. Cognitive research suggests that structured storytelling improves memory and reduces hindsight bias by preserving the uncertainty experienced in real time.

At the end of the week, synthesize across trades. Identify the market conditions you actually traded, not the ones you wished for. Note repeated decision errors, such as chasing after missing the first entry, or cutting winners during slow periods. Also record what worked even if P&L was flat, like clean risk control through a volatile session. The narrative should be honest and specific, avoiding global statements about discipline that cannot be acted upon.

From observations to hypotheses A hypothesis converts a pattern into a testable statement. It ties a condition to an action and a measurable expectation. The aim is to run small, fast experiments in live conditions without risking strategy integrity. The statement should include setup, context, behavioral change, and the metric to judge success.

For example, a recurring observation might be that midday trades underperform relative to the first hour. A testable hypothesis becomes: When volatility as measured by average true range compression occurs after the first hour, skipping new entries until a range break will raise weekly expectancy by at least 0.2 R compared to the prior week. The condition is ATR compression, the action is to skip entries, and the success metric is the change in weekly expectancy controlled for risk per trade.

Another example could involve position scaling. If partial exits are consistently cutting off the right tail of outcomes, the hypothesis might be: Keeping the first unit intact until the trailing stop is hit will increase the proportion of trades exceeding 2 R without raising weekly drawdown beyond prior levels. The measurement focuses on tail frequency and drawdown constraints, not only on average P&L.

Good hypotheses are small and isolated. Change one lever at a time, such as entry timing or exit management, and hold others constant. Set a minimum observation window that matches trade frequency. If the strategy produces ten to twenty trades per week, a one-week test is directional but fragile. Record the result and replicate for another week before cementing the change.

A Thursday checkpoint Because today is Thursday, use a light pre-review. Scan the week’s trades and tag anomalies that deserve a deeper look tomorrow. Mark any potential hypotheses that emerged. Check whether data collection is complete, including risk, R results, and process scores. A short Thursday checkpoint reduces the cognitive load of the full Friday review and protects against end-of-week fatigue.

A practical workflow Begin the week with one active hypothesis at most. During the week, record core metrics for each trade: planned risk, realized R, MFE and MAE in R, and a simple process score. Each evening, write two or three sentences for the day’s narrative, preserving details that can disappear by Friday. On Thursday, do the checkpoint. On Friday after the close, compute weekly metrics, complete the scorecard, integrate the narrative, and decide the status of the hypothesis. Over the weekend, turn accepted changes into playbook rules, schedule one new hypothesis if needed, and archive rejected ideas for future conditions.

An illustrative mini-case Consider a short-term breakout trader who notices that entries taken within the first 10 minutes after the open show higher adverse excursion and lower expectancy. The weekly metrics confirm that early trades have a larger MAE with no compensating tail. The narrative describes frequent slippage and crowded order flow on the first attempt. The hypothesis becomes: Delaying breakout entries until the first pullback after a 15-minute base will reduce average MAE by 0.3 R and maintain expectancy. The next week, the trader implements the delay and tracks MAE, R results, and process score. Expectancy improves slightly, MAE drops, and maximum drawdown remains stable. The hypothesis is accepted and written into the playbook with clear rules for the base and the pullback. Subsequent weeks confirm the improvement, and the change becomes a standard operating procedure.

Common pitfalls and how to avoid them The most frequent error is collecting too many metrics. Excess measurement dilutes attention and invites narrative gymnastics to reconcile contradictions. Focus on a handful that anchor risk, expectancy, opportunity capture, and process.

Another pitfall is confirmatory bias in narratives. It is tempting to write stories that justify entries in hindsight. Counter this by starting with raw facts and writing interpretations in a separate line. Review at least one losing trade in depth each week to build discomfort tolerance and to learn from the tails.

Overreacting to a single week is also common. If the week had unusual macro events or platform issues, label the week as atypical and avoid large playbook changes. Protect the integrity of the core system by testing changes in small increments and repeating the test in a different regime when possible.

Tooling that keeps the loop tight Any simple spreadsheet or journal can support the loop. The essential fields are date and session, setup name, regime tag, planned risk size, realized R, MFE and MAE in R, slippage estimate, process score, and a short note. A weekly tab can aggregate expectancy, hit rate for qualified setups, process average, and drawdown. Add a text box for the weekly narrative and the active hypothesis with its measurement plan. The format matters less than consistent use.

Turning review into a habit Consistency is a skill. Set a fixed window for the weekly review and protect it like a meeting with a client. Begin with the scorecard so the scope is defined, read the week’s narrative to restore context, and then evaluate the hypothesis. Finish by writing one concrete instruction for next week and one environmental tweak, such as adjusting the watchlist or predefining alerts. Habit research suggests that stable cues and a short, repeatable sequence make follow-through more likely.

A simple starting template Choose three metrics for the scorecard, commit to a two-paragraph weekly narrative, and run one hypothesis per week. That minimal structure is enough to begin a scientific cycle without friction. Over time, the review becomes a quiet competitive edge. It reveals what the market is offering, what the trader is actually doing, and what should change next week to close the gap.