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

Sizing, drawdown and cash-flow denominators

Keep account currency, position increments, original risk, recovery percentages and deposits separate before drawing conclusions from a risk statistic.

Original edition: 9 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.

Start with the account and instrument units

A percentage risk input needs a named account amount and currency. A price distance needs the instrument's value per point and permitted quantity increment. Shares, contracts and platform lots are not interchangeable. The existing sizing tool computes a planning quantity from entered units; it does not confirm the contract specification or account suitability. Add applicable extra costs to the per-unit budget under a stated convention. Check funding and other exposure separately. A position can fit a planned stop budget while exceeding available cash, margin or a portfolio limit that the simple formula does not model.

Round down without manufacturing capacity

Compute the unrounded planned quantity, divide by the permitted increment, take the integer floor and multiply back. Do not add a convenient epsilon before flooring if it can promote quantity above the allowed budget. Numeric boundary checks should confirm that the final planned loss does not exceed the stated budget. If the minimum increment is too large, zero units is a valid planning result rather than a reason to round up. Validate missing fields as missing information, not deliberate zeros. Invalid input must clear all metrics so an old attractive result cannot be mistaken for the current calculation.

Recovery uses the reduced balance

An account beginning at 100 and falling to 80 has a 20% decline. Recovering 20 units relative to the new balance requires 20 divided by 80, or 25%. In general, for decline d below 100%, the recovery percentage is d divided by 100 minus d, times 100. A complete loss has no finite percentage recovery from zero. These are arithmetic identities, not a predicted recovery path. The calculator reports the assumption and rejects a drawdown at or above 100% in the recovery lane rather than displaying an unexplained infinite or misleading value.

Cash flows can disguise account movement

A deposit can restore the balance without the strategy recovering its loss; a withdrawal can reduce balance without a trading drawdown. Record contributions, withdrawals, fees and valuation dates separately. If a performance measure adjusts for flows, declare its method and required timestamps rather than improvising an aggregate from end balances. The public tool intentionally does not calculate a time-weighted or money-weighted strategy return. Keep original-risk R multiples distinct from account-equity percentages as well. A revised stop or later account deposit must not silently change a metric that claims to use the frozen original risk denominator.

A losing streak calculation has narrow assumptions

With a fixed fraction r risked on each successive reduced balance and every hypothetical loss exactly equal to that fraction, the remaining balance multiplier after n losses is one minus r to the power n. This assumes resizing, exact losses and no cash flows. Gaps, fees, minimum units or a fixed-notional policy break those assumptions. Zero losses preserve the starting amount even if the entered fraction is 100%; one complete loss leaves none. A deterministic streak exercise does not assign a probability to experiencing the streak or independently estimate the strategy's true win rate.

Keep expectancy and size in their proper roles

An expectancy calculation combines entered win frequency, average gain, average loss and costs under its declared outcome definition. Trade count alone does not supply a confidence guarantee or a universal minimum sample threshold. An apparent positive expectancy does not decide an appropriate personal fraction of account risk. Preserve sampling, overlap, uncertainty and execution assumptions separately. The publication's tools help inspect arithmetic and reject missing data. They do not choose a reader's size, validate a seller's grades or certify a strategy. A complete report should tell readers which denominator every percentage uses and which real-world constraints remain outside the calculation.

An increment fixture

Use account 10,000, risk one percent, entry 100, stop 98, point value one, additional cost 0.10 and whole-unit increment one. The planned per-unit amount is 2.10, so forty-seven units fit the 100 budget while forty-eight exceed it. Change the increment only after saving the original example and check the new quantity stays within budget. This tests conservative arithmetic, not a recommended trade size. Cash funding and the appropriateness of the risk percentage remain separate questions.

Source documents

Documentation establishes the stated mechanism, not tested trading performance. Checked 9 October 2026; a later version can differ.

Continue the investigation

Two-position cash-funded gap stress

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 stated valuation path

The workbook assumes no flows during its terminal liquidation exercises. The original cash-flow protocol remains the appropriate place for between-segment deposit and withdrawal questions. A 100-to-80 decline and a 20 deposit can restore resources without demonstrating trading recovery. Preserve marks and exact flow timing before calculating a performance measure. The new scenarios do not claim GIPS compliance, time-weighted certification or an independently audited return from a simplified example.

A completed denominator note

For the invented 10,000 cash account, the 650 joint liquidation loss is 6.5% of that specified starting base. That is not a maximum drawdown along an observed path, because interim valuations are absent. A real path could differ and needs its own timestamps and marks. Keeping the percentage and its denominator together prevents a scenario amount from becoming a current subscriber drawdown or a provider-wide risk statistic when repeated in a search snippet or machine summary.