ADVANCED: COURSE 1 | LESSON 5
Building a repeatable weekly analysis routine
Learning objectives
Assemble the tools from A1.1–A1.4 into a fixed weekly and daily analysis routine with defined inputs, outputs and time budgets
Write scenario-based ("if/then") trade plans instead of single-outcome predictions
Keep the routine honest with a Friday review that scores the analysis, separately from the P&L
Why a routine beats talent
Ask a discretionary trader at a fund what they actually do and you will hear something unglamorous: the same preparation, at the same times, in the same format, every week. The routine is not decoration around the "real" skill — it is the skill, industrialised. It guarantees that macro context, intermarket checks, levels and event risk are all considered before capital is at stake, in a fixed order, so that a busy Tuesday or an exciting chart cannot quietly delete a step. It also creates a written record, which is the only thing that lets you debug your analysis later (A2.3 builds the metrics on top of it).
Everything below assembles the previous four lessons into two artefacts: a weekly bias sheet (built on the weekend, ~60–90 minutes) and a daily update (15–20 minutes before your trading session). Adjust instruments and timings to your life; do not adjust the principle that outputs are written and dated.
The weekend session: building the bias sheet
Work top-down, in this order, and write one or two lines per step — full paragraphs rot into essays nobody rereads.
1. Macro scan (15 min). For each currency you trade: what changed last week in rate expectations (2-year yields, central-bank speak)? Where does carry help or hurt (A1.1)? Classify the risk regime using your checklist — equities, VIX, yields, gold, JPY crosses — as risk-on, risk-off, or transitional.
2. Intermarket board (10 min). Run the dashboard from A1.2: DXY versus the broad dollar, the yield spreads for your pairs, oil for CAD, copper for AUD, and note which correlations are currently alive versus broken. One sentence: what is the dominant driver this week?
3. Calendar mapping (10 min). List the week's tier-1 events (CPI, central-bank decisions, labour reports) with day and time in your timezone, plus which of your pairs each one touches. Decide the policy per event now, while you are calm: flat through it, reduced size, or trade the reaction only. Never leave this decision to the minute before the release.
4. Levels and profile work (20 min). For each pair: weekly and daily structure, last week's value area and POC, any naked POCs, and the obvious liquidity pools — equal highs/lows and round numbers where stops cluster (A1.3–A1.4). Mark zones, not lines. Limit yourself to the handful of levels you would genuinely trade from; a chart with fourteen rectangles is a chart with none.
5. Scenario planning (15 min). This is the step that separates a bias sheet from a prediction. For each pair, write two or three if/then branches covering the plausible week, for example:
A prediction can only be right or wrong. A branch plan is useful under every outcome — that is the point. You are not paid for forecasting; you are paid for having a prepared response to whatever happens.
6. Self state (5 min). Last week's P&L effect on you, open risk carried into the week, and your size setting for the week (normal, reduced after drawdown per your rules — see A2.4). One honest line.
The daily update: 15 minutes, not a rebuild
Each trading day, before your session: check overnight moves against your branches (which scenario is developing?); re-check today's calendar entries; adjust intraday levels for the new session's developing value area; and confirm your risk state (open positions, remaining daily risk budget). The daily update edits the weekly sheet — it must not relitigate the bias from scratch, or you will flip-flop with every headline. The bias only changes when a written invalidation condition from step 5 actually triggers.
During the session, the routine's job is to shrink your decision space: you already know your zones, your branches, your event policy and your size. What remains in real time is execution quality — the lower-timeframe trigger and stop placement from A1.4, and the liquidity awareness from A1.3.
The Friday review: score the analysis, not the money
The loop closes with 30 minutes on Friday (or the weekend, before the new sheet). Crucially, review the analysis on its own scorecard, because a good week of analysis can coincide with a losing P&L and vice versa — and if you only audit money, you will learn the wrong lessons from noise (A2.2 goes deeper on this).
Score three things. Bias accuracy: for each pair, did the week develop within one of your written branches, or did something happen you had no branch for? Track the percentage of weeks "covered by a branch" — it is a direct measure of whether your scenario planning is imaginative enough. Level quality: which of your marked zones produced a reaction? Which were ignored? Persistent misses on one kind of level (say, your naked POCs work but your trendlines do not) is exactly the feedback that should reshape next week's step 4. Plan adherence: of the trades you took, how many were at planned zones under a planned branch? Unplanned trades get flagged regardless of outcome — a winning unplanned trade is still a process failure that will bill you later.
Keep every bias sheet. Twenty of them constitute a dataset about your own analytical strengths that no course, including this one, can give you.
Making it stick
Routines fail from friction, not from ignorance. Reduce the friction: a fixed template (copy last week's sheet and overwrite), a fixed calendar slot protected like a meeting, and a hard time cap — an unfinished-but-done-on-time sheet beats a perfect one abandoned by week three. Expect the first few weeks to feel mechanical; that is the feeling of a process becoming trustworthy. When markets get chaotic, the routine is what keeps you from becoming chaotic with them.
Key takeaways
The weekly bias sheet has six fixed steps — macro scan, intermarket board, calendar mapping, levels/profile, if/then scenarios, self state — each with a time cap and a written output
Scenario branches with explicit invalidation conditions replace predictions: a plan should be useful under every outcome, not correct under one
Decide your policy for each tier-1 news event on the weekend, never in the minute before the release
The daily update edits the weekly sheet in 15 minutes; the bias changes only when a written invalidation triggers, not with every headline
Friday's review scores bias coverage, level quality and plan adherence separately from P&L — analysis and outcomes are audited on different scorecards