Skip to account-risk research
Risk Management for TradersUnits / Funding / Scenario limits

Common risk-management mistakes

Inspect stop budgets, rounding, cash flows and shared exposure with reproducible examples instead of generic risk slogans.

Original edition: 10 October 2026. Expanded edition: 13 October 2026.

Educational risk arithmetic, not personal sizing advice. Stops do not guarantee fills. Stress inputs are invented assumptions without assigned probabilities; margin, shorts and derivatives need their own contracts.

Calling a planned stop loss the maximum possible loss

A stop budget describes an intended loss under a particular quantity and exit-price assumption. It is not automatically a ceiling on an account's loss. Investor.gov distinguishes order mechanisms, including a stop order becoming a market order once triggered. A trigger, a market instruction and an execution report are separate events. Gaps, unavailable liquidity and partial quantity can separate the planned stop level from the actual realised price.

Write the scope next to every budget. An invented long position of fifty shares at 100, planned stop 98 and extra round-trip cost 0.10 per share has planned loss 105. If an assumed exit is 94 instead, the same quantity has scenario loss 305. These figures use one consistent currency and are not observed trades. Do not describe the second as a percentile or probability estimate without a separate empirical model.

Rounding up or accepting an incomplete unit contract

For invented budget 100, price distance 2 and additional per-unit cost 0.10, the permitted whole-unit quantity under this formula is floor(100/2.10), or forty-seven. Planned loss is 98.70. Rounding to forty-eight gives 100.80 and breaches the same declared budget. A display rounded to two decimals must not drive the acceptance calculation. Keep sufficient precision in the calculation and verify the accepted quantity afterwards.

A blank unit value is not zero. Reject an incomplete point-value, currency, increment or account-denominator field rather than retaining a previous valid result. Shares, futures contracts, option legs and FX base units require different contracts. A decimal quantity accepted by a generic form may be impossible for the actual instrument. Quantity below the minimum tradable increment means no accepted size under the entered assumptions, not permission to round upward.

Treating an external deposit as recovery

Consider invented starting equity 10,000, followed by equity 8,000 and then an external deposit of 3,000. The account value becomes 11,000 without any subsequent market gain. Calling that a ten-percent strategy return mistakes cash flow for performance. The earlier market decline was twenty percent in this simple no-prior-flow segment. A genuine recovery comparison requires a declared flow-adjusted convention and timing.

Closed balance can also hide unrealised losses or gains. Name the marked-equity series, valuation clock, prices and treatment of missing marks before computing a drawdown. Do not splice pre-flow and post-flow values into a performance curve without an accounting rule. A unitised curve and a raw account-value plot have different interpretations. Keep the original cash-flow ledger so a reader can reproduce the chosen adjustment instead of trusting a cosmetically smoother chart.

Ignoring the exposures that share one account

An individually acceptable instruction can be unacceptable when combined with existing positions and pending orders. Inspect outstanding quantity, common instrument exposure, funding and the ability to cancel or reduce orders. Two differently named strategies can hold the same underlying instrument. Net exposure may conceal substantial gross exposure or obligations if exits do not occur together.

Use scenarios with each position visible before the total. An assumed gain in one position can offset a loss in another arithmetically, but that does not prove a reliable hedge under stressed execution. Do not estimate correlation from one invented shock. Margin required by an account is not the amount a position can lose, and cash availability is not a complete derivative-risk assessment. Preserve the declared account lane and route unsupported instruments to a separate contract rather than stretching a long-share worksheet.

Changing a control after seeing the unfavourable outcome

A later stop amendment, holding extension or discretionary rescue changes the experiment's action history. Preserve who authorised it, what information was available and when it took effect. Moving a stop retrospectively to a price that avoided a loss cannot be folded into an unchanged systematic record. A control that was bypassed is an operational exception even when the final trade happens to win.

Keep policy revision and model-selection history together without letting exit-adjusted results redefine the original entry ranking. A new control can be researched as a separate candidate with its own evaluation boundary. Repeatedly inspecting the same reserved outcomes does not make a revision independently tested. For actual operations, a policy document needs deployment and route evidence; a file labelled live does not establish that the broker-facing path enforced that version.

Using precise-looking summaries instead of reconciled records

A risk score, warning badge or clean dashboard is useful only within its declared input coverage. Ask how the display handles missing prices, stale account state, duplicate instructions and failed cancellation. A green status can mean a request was accepted by software rather than that execution and position state were reconciled. Preserve unknowns and errors instead of converting them to neutral values.

The blank review record turns these mistakes into evidence requests. Reproduce the budget, gap and deposit examples with their exact assumptions, then test separate failure cases using synthetic inputs. Those tests establish accounting behaviour, not a validated risk strategy or personalised allocation. Where account terms or active controls remain unknown, leave the relevant finding open. A bounded record of what a system cannot yet establish is more useful than a universal claim that risk is fully managed.

Review checkpoints

Evidence requests, not completed study results
ObjectPreserveDo not infer
BudgetUnits, quantity and extra costsStop budget is an absolute loss ceiling
AccountMarked equity and external flowsA deposit proves strategy recovery
ControlLive route and reconciled exceptionsA clean dashboard proves enforcement

Download blank review record (CSV)

Headers only. No trades, subscriber outcomes or operating receipts are supplied.

Source documents

Documentation establishes the stated mechanism, not tested trading performance. Checked 10 October 2026; a later version can differ.

Detailed research

Joint gap risk and shared exposure

Compare planned stop loss with disclosed hypothetical account shocks without pretending to estimate correlation, tail probability or a guaranteed maximum.

A completed workbook cross-check

The existing protocol stays intact. The new specimen exposes an additional source of error: comparing isolated and joint stresses while changing positions, cost conventions or account bases. It holds A, B and C constant and shows all five scenario exits. The isolated-A result includes costs on unshocked positions, so it is not simply A's price-distance loss. The reviewer can reproduce the difference without inventing a historical occurrence rate or declaring the positions independent. This is an arithmetic correctness case, not a new risk percentage recommended for an account.

Do not infer certainty from completeness

A complete scenario table can still omit financing, partial fills, execution availability or actual quantities because its scope is intentionally narrow. That is acceptable when stated, but not when the narrow result is marketed as a complete risk certification. Keep the specimen beside the blank observation journal, with a separate source request for real records. A current price or qualification cannot supply an old ledger's missing cutoff, and a budget calculation cannot determine the exit policy's owner.