Trading Review Cadence: How Often Should Traders Review Performance?

There is no single mandated frequency for reviewing trading performance, but there is a practical default architecture: assign a different, bounded question to each of four review horizons — daily, weekly, monthly, and quarterly — instead of asking the same question at every interval. Trading review cadence is that assignment. Daily review checks whether today’s decisions matched the rules active at the time. Weekly review checks whether a specific, already-defined deviation is recurring. Monthly review aggregates results, risk, and execution into a broader picture. Quarterly review asks whether the accumulated evidence, or a change in conditions, justifies reopening the plan’s own assumptions. These are practical review horizons, not automatic statistical sample thresholds — the calendar creates the checkpoint; it does not by itself supply the evidence.

Picking a cadence badly shows up in two opposite ways. Reviewing everything daily treats single-trade variance as a pattern and invites a rule change before enough comparable evidence exists. Reviewing only occasionally lets a recurring deviation run for weeks before any checkpoint is positioned to catch it. Neither failure is solved by reviewing more or less often on its own — it is solved by matching each question to the horizon built to answer it, and by remembering that the horizon does not manufacture the evidence a question needs.

What this adds that post-trade review and the daily routine don’t

Post-trade review defines how to reconstruct and classify one decision or one trade. The daily trading routine defines the pre-session, live, and post-session sequence for one day. Neither one owns the question this article answers: once individual reviews exist, on what schedule do you step back, and what question does each step-back interval get that the others don’t?

This is a scheduling and scope-assignment problem, not a reconstruction problem. It assumes post-trade review and daily review are already happening and asks how to layer weekly, monthly, and quarterly review on top of them without asking the same question four times at four speeds. It also assumes each review’s classification is trustworthy; how to test for cognitive bias in a trading review covers checking that assumption itself, independent of which horizon the check is scheduled on.

Why the same review doesn’t work at every horizon

A review question is bounded by four different things: the calendar interval it is assigned to, the number of comparable observations available inside that interval, how stable the rule or process being evaluated has stayed across those observations, and the scope of decision the question is entitled to make. Conflating the first with the other three is the core cadence mistake — a longer calendar interval does not by itself manufacture a larger or more comparable sample.

• Whether one decision followed a known rule is answerable from a single observation, the same day it happens.
• Whether a specific, defined deviation is recurring needs more than one comparable occurrence under a stable rule version. A weekly checkpoint may reveal that recurrence once enough comparable decisions have accumulated by then — not because a week carries some special evidentiary weight.
• Whether a broader pattern in results, risk, or execution is developing needs a wider body of comparable evidence than a single week ordinarily provides. A monthly review can aggregate that broader sample, but the calendar month alone does not make it sufficient if trading was infrequent or conditions changed partway through.
• Whether the strategy or plan assumptions themselves deserve reconsideration needs enough comparable evidence across varying conditions that a single month, and often a single quarter, may still be thin. The question stays open until the evidence closes it, not until the calendar does.

Running the third or fourth question daily produces false signals: a losing stretch under an intact process can look identical, across a handful of trades, to a genuinely deteriorating one. Running the first question only monthly produces the opposite failure: a rule violation that recurred for weeks gets diagnosed late, after it has already shaped habits and account balance. The fix is not a different calendar — it is assigning each question to a checkpoint likely, though never guaranteed, to have accumulated enough comparable evidence to answer it.

Assign one question to each horizon

Each horizon below has a distinct job, defined by the kind of evidence it can realistically have accumulated — not by an assumption that the calendar itself supplies a sufficient sample. A finding that belongs to one horizon should not be forced into another just because it happened to surface there.

Horizon	Review question	Evidence it needs	What it should not try to answer
Daily	Did today’s decisions follow the plan that was active today?	That day’s trades, the active rule version, and the session record	Whether a deviation is recurring, or whether the strategy or plan is still valid
Weekly	Has a specific, already-defined deviation happened more than once?	Enough comparable decisions under one stable rule version to compare — not a fixed trade count	Long-run strategy performance, or a single trade’s classification
Monthly	What do results, risk, and execution look like together across a broader body of evidence?	A wider, classified sample, viewed by results, risk, and execution separately — sufficiency depends on trading frequency, not the calendar month itself	Treating one month’s aggregate as proof the strategy is or isn’t working
Quarterly	Does the accumulated evidence, or a change in market, instrument, account, or execution conditions, justify reopening the plan’s own rules or assumptions?	Multiple months of comparable evidence, plus any material change in the conditions the plan was built around	Optimizing the plan around the past few weeks’ P&L in isolation

The horizons are cumulative, not redundant: a weekly review does not re-litigate every daily finding, and a quarterly review does not re-run twelve weekly reviews from scratch. Each one asks a question suited to the evidence the shorter horizons have had the chance to accumulate — and each is entitled to answer “not enough evidence yet” rather than forcing a conclusion the sample doesn’t support.

The specific failure of skipping a horizon

Skipping a horizon does not just remove a checkpoint — it removes the review layer specifically assigned to catching a particular class of problem.

• Skip the daily check and a single rule deviation has no immediate record; by the time a weekly review looks for a pattern, the contemporaneous evidence — the trigger, the exact time, the stated reason — may already be gone from memory.
• Skip the weekly check and a deviation that occurred three times gets buried inside a month of otherwise-aligned trades, where it is easy to average away rather than see as a specific, repeatable gap. Trading mistake prioritization depends on that gap being visible as a distinct, counted pattern — it cannot rank a mistake nobody isolated. Whether a mistake has accumulated enough comparable evidence to earn that attention, rather than a single vivid occurrence, is a separate question this cadence structure does not answer on its own.
• Skip the monthly check and results, risk, and execution get judged only by whichever trade is most recent or most memorable, rather than by the wider body of evidence needed to tell a developing pattern from ordinary variance.
• Skip the quarterly check and a rule that quietly stopped fitting current conditions — a changed instrument, a changed account size, a changed personal schedule — keeps being measured against a standard nobody has revisited.

The four horizons are a practical default architecture, not a universal requirement — a trader running a different but equally deliberate review structure is not automatically wrong. What matters is that each information-resolution role above is covered by something. A trader who reviews daily and quarterly but has nothing in between, for example, has strong single-trade discipline and an occasional structural reset, with no layer positioned to catch a recurring deviation before it accumulates into a quarter’s worth of evidence.

A worked example across review horizons

A discretionary trader keeps daily, weekly, and monthly reviews on a fixed schedule.

Daily: In week two, one trade is classified as a deviation — an entry taken slightly before the defined confirmation condition. It is recorded the same day, with the actual timestamp and the setup rule it conflicted with. This step’s only job is preserving that contemporaneous evidence before memory fades.

Weekly: By the end of week three, the same early-entry pattern has recurred twice more, each occurrence recorded with comparable evidence. With the same early-entry behavior now appearing across several comparable decisions under one stable rule version, recurrence becomes a defined question worth investigating rather than assuming the first occurrence was isolated — the weekly checkpoint is what makes that repetition visible, because no single day’s review sees more than that day’s trades.

Monthly: At month’s end, the trader reviews results, risk, and execution together across all four weeks. Execution adherence is lower than the prior month, and the early-entry pattern accounts for most of the gap — yet results for the month are still positive. Positive P&L coexisting with declining execution adherence is exactly the kind of finding a monthly view can surface, because it holds outcome and process on separate lines instead of letting one stand in for the other.

Quarterly: The confirmation rule the early entries violated has not itself become unsuitable — nothing about the instrument, account, or session conditions points to that — so the accumulated evidence argues for testing a response to the existing deviation, not for reopening the rule itself. The monthly finding becomes the input to a feedback-loop test: define what response prevents an early entry, apply it for a fixed observation period, and check whether the recurrence stops. Had the evidence instead pointed to a structural mismatch — the setup no longer fitting a changed instrument or session, for instance — the quarterly review’s job would have been to reopen that assumption instead of routing the finding to a process test.

Note what each horizon contributed: the daily record supplied the evidence, the weekly review supplied the recurrence, the monthly review supplied the broader picture and the process-versus-outcome distinction, and the quarterly review decided that the finding belonged inside the existing process rather than requiring the process itself to change. Collapsing this into a single monthly review would likely still have surfaced the recurring entry, but later, and without the contemporaneous evidence the daily record preserved or the explicit check on whether the underlying rule still fit.

Common trading review-cadence mistakes

Reviewing performance only after a loss

A review schedule triggered by outcome rather than by calendar means aligned losses get scrutinized and misaligned wins do not. Keep the schedule fixed regardless of the most recent result; a bad week does not by itself justify an extra unscheduled review, and a good week does not justify skipping the scheduled one. A predefined risk or operational trigger is a different matter — see below.

Running the same checklist at every horizon

Using the post-trade review checklist as the weekly, monthly, and quarterly review as well means every horizon answers the same narrow question and none of them ever ask about recurrence, aggregation, or structural fit. Match the question in the table above to the horizon, not the other way around.

Treating a short losing streak as a reason to change the plan mid-cycle

A losing streak inside a single week is exactly the kind of variance a weekly or monthly checkpoint exists to put in context, not proof by itself that the process has changed. Changing the plan before a checkpoint has accumulated enough comparable evidence to distinguish a recurring deviation from ordinary variance — see trading feedback loop for how to test a change deliberately once that evidence exists — risks reacting to noise the next checkpoint would have explained.

Confusing a diagnostic check with a strategy change

A risk-limit breach, an operational or data error, an execution failure, or a material change in account or instrument conditions can justify an unscheduled diagnostic review — confirming what happened and whether an existing control still worked. That is not the same action as changing the plan’s rules based on a recent result. The first responds to a predefined trigger; the second lets an outcome override the cadence. A diagnostic review can happen anytime its trigger is defined in advance, but it does not by itself authorize a rule change outside the review that has accumulated evidence to support one.

Letting the quarterly review re-litigate every daily finding

A quarterly review that re-opens each week’s individual deviations duplicates work the shorter horizons already did and buries the one question only the quarterly view can answer: does the accumulated evidence justify reopening the plan itself.

What if the review calendar and a real risk event collide?

A predefined risk trigger — a maximum-loss limit, a position-size breach, an operational or data error, or another boundary the trader’s own rules already treat as requiring an immediate response — is not a cadence question and should not wait for the next scheduled review. That kind of event calls for the in-session response your plan already defines, the same distinction the daily routine draws between a pressure event and an ordinary checkpoint. An unscheduled diagnostic review triggered this way confirms what happened and whether an existing control worked; it is not the same action as an unscheduled strategy change, which still needs the accumulated, comparable evidence a scheduled checkpoint is built to gather. Trading review cadence governs when process and pattern questions get asked on the calendar; it does not replace a risk boundary that already has its own trigger and response, and it does not forbid a diagnostic check outside that calendar.

Where Costante fits

Costante’s structured logging and behavioral review give each horizon something concrete to work from: daily records for the daily check, discipline trends and pattern detection across the stored history for the weekly and monthly checkpoints, and a consistent evidence trail for the quarterly review to draw on. None of that determines how much evidence is enough — it makes the evidence available.

Costante does not set a trader’s review schedule, decide which horizon a finding belongs to, judge whether a sample is statistically sufficient, or determine when a plan should change. The cadence, the classification of each finding, and every decision about the trading process remain the trader’s responsibility.

Frequently asked questions

How often should you review your trades?

There is no single mandated frequency, but a practical default architecture works well: capture evidence daily, check for recurrence weekly, aggregate results, risk, and execution monthly, and revisit the plan’s own assumptions on a periodic — often quarterly — basis. How much any single checkpoint can conclude still depends on how often you trade and how much comparable evidence has accumulated by then; an infrequent trader’s weekly checkpoint may still be waiting on its first repeat occurrence.

Is a daily trading review necessary, or is weekly enough?

Weekly review can detect a recurring pattern once enough comparable decisions exist, but it cannot recover evidence that was never recorded. Without a daily record, a weekly review is reconstructing from memory rather than from a contemporaneous account of what happened, when, and under which rule — the two serve different jobs, and neither substitutes for the other.

What’s the difference between a weekly and a monthly trading review?

A weekly review asks a narrow question: has a specific, already-defined deviation happened more than once. A monthly review asks a broader one: what does the accumulated body of results, risk, and execution evidence show together, once enough of it exists to look at as a whole. The weekly checkpoint isolates a repeat occurrence; the monthly checkpoint looks at the wider picture that occurrence sits inside.

Should you change your review cadence after a drawdown?

Not automatically, and not based on P&L alone — repeatedly moving the calendar or the rules in response to a recent result converts a fixed process into an outcome-triggered one. That’s separate from a predefined risk-limit breach, an operational failure, or another exceptional trigger, which can justify an unscheduled diagnostic review without redefining the cadence itself. The distinction is between checking what happened and changing the process because of how it turned out.

Does a monthly or quarterly checkpoint tell you whether a trader’s skill has actually improved?

Not by itself. A checkpoint arriving on schedule tells you when to look; it doesn’t tell you whether the sample held at that checkpoint is large and comparable enough to support a skill-development conclusion rather than ordinary variance. Which review horizon measures trading skill development? covers that separate evidence-sufficiency question.

Costante provides educational workflow tools, not financial advice. Trading involves risk.