The Weekly Trading Review Routine That Finds Leaks

The daily log tells you what happened; the weekly review is where you actually decide to change. A week is the smallest window that holds enough trades to spot a pattern and few enough that you can still remember every one of them.

Why weekly is the review that changes behavior

Daily notes are too close to the action — one bad session feels like a crisis, one good one like a breakthrough, and neither is true. Monthly reviews are honest about the numbers but too far removed to fix anything specific; by the time you spot the leak, you’ve repeated it twenty times.

The weekly review sits in the sweet spot. You have five sessions’ worth of trades — enough to see a repeated mistake, recent enough that the fix lands before the habit hardens. This is the cadence at which behavior actually bends. Block a fixed 30–45 minutes, same time each week, and treat it as non-negotiable as the market open.

Aggregating the week

Start with the flat numbers, no interpretation yet. Pull the whole week and compute:

Total trades and net result in R, not just dollars.
Expectancy — your average R per trade across the week. Feed the week’s trades into an expectancy calculator so you’re working from one hard number instead of a vibe.
Win rate and the average winner vs average loser in R.
Rule breaks — count them. Every deviation you flagged in the daily log, tallied.

Report these to yourself honestly before you explain them. The temptation is to narrate (“but Wednesday was a weird session…”) before you’ve even looked. Numbers first, story second.

One caution: a single week is still a small sample. Five to twenty trades won’t give you a statistically firm edge estimate, and the expectancy concept explainer is worth a read on why one green week doesn’t prove an edge. Use the weekly number to spot direction and leaks, not to certify that your strategy works. This is precisely where Shibiki’s Wilson confidence interval earns its place — it shows the range your true edge likely sits in given how few trades you’ve logged, so a lucky week reads as “wide and uncertain” rather than “profitable, confirmed.”

Sorting trades by setup to find the leak

This is the core of the review and where the leaks hide. Group the week’s trades by the setup tag you assigned in your daily routine, then compute expectancy per group.

Almost every trader discovers the same uncomfortable truth: one or two setups carry the account, and one is quietly bleeding it. The blended weekly number hides this completely — a strong A-setup can mask a B-setup that loses money every time you touch it.

Which setup had the best expectancy this week? Do more of it, and ask what conditions it needed.
Which had the worst? That’s your leak. Before cutting it, check the sample — one bad trade isn’t a verdict.
Which setup did you trade most often, and does frequency match quality? Traders massively over-trade their weakest, most tempting setup.

Sorting by setup converts a vague “rough week” into a specific, fixable sentence: “My fade setup lost 4R across six trades and I keep taking it in trends.”

Separating process errors from bad luck

Not every loss is a mistake, and not every win is skill. Before you write next week’s rules, sort the week’s outcomes into two piles:

Process errors — you broke a rule, sized wrong, chased, moved a stop, took an untagged impulse trade. These are 100% controllable and are the only thing worth building rules around.
Bad luck — you followed the plan perfectly and the trade lost anyway. This is the cost of doing business. Punishing yourself for a well-executed loser teaches you to abandon good setups.

The distinction matters because the two demand opposite responses. A process error means change the behavior. A bad-luck loss means repeat the behavior — the edge plays out over sample size, not per trade. Conflating them is how traders “fix” a perfectly good strategy into the ground. Thinking in R-multiples keeps this clean, since a rule-perfect −1R and a sloppy −1R look identical on the P&L but are opposite lessons.

Writing one to three concrete rules for next week

End every weekly review with no more than three rules for the week ahead. One is fine. Ten is a fantasy you won’t follow.

A good rule is specific, observable, and tied to a leak you actually found:

Vague: “Be more disciplined.” Useless — unmeasurable.
Concrete: “No fade setups on trend days” or “Cut size on the B-setup until it shows 20 profitable trades.”

Write them where you’ll see them at the open. Next week, the first thing your review checks is whether last week’s rules held. That closed loop — find the leak, write the rule, verify adherence, find the next leak — is the entire engine of improvement. If those hard limits are enforced at the broker rather than left to willpower, the rule holds even on the day your discipline doesn’t, which is exactly when it matters.