How to Adjust Risk During a Trading Drawdown

To adjust risk during a trading drawdown, define the drawdown measure and trigger before the loss sequence, move to a pre-defined risk state when that trigger occurs, and apply that state’s risk multiplier to each new position-sizing calculation. Return to normal risk only through a restoration rule chosen in advance, not in response to the latest outcome.

The difficult part is not writing “trade smaller after losses.” A usable drawdown rule must answer three separate questions: When does risk change? By how much does it change? What evidence permits normal risk to return? The trader must choose and test those answers for their own method, instruments, risk capacity, and account conditions. There is no universal drawdown percentage or recovery schedule.

What does reducing risk during a drawdown mean?

In this article, a trading drawdown is a decline from a defined account or strategy-equity reference peak to a later value. Reducing risk means lowering the maximum planned loss permitted on each new trade, or removing permission for new exposure, while that drawdown state remains active.

That definition needs a consistent measurement rule. Decide whether the drawdown uses closed-trade equity, end-of-day equity, or another value your process can reproduce. State whether fees are included, how deposits and withdrawals are treated, and whether open P&L can activate the rule. Mixing definitions makes the trigger movable just when it becomes uncomfortable.

Drawdown risk is also different from a daily loss limit. A daily loss limit controls permission within a defined session. A drawdown rule can remain active across several sessions and govern the risk allowed on later trades. Both can apply at once; the more restrictive active boundary takes precedence.

Why reduce position risk in a drawdown?

Reducing planned loss changes the amount of capital exposed to the next uncertain outcome. It can slow the rate at which further losses compound while the trader investigates whether the drawdown reflects normal variation, a strategy problem, execution drift, or some combination. It does not repair the strategy or make the next trade more likely to win.

The arithmetic is asymmetric. A 10% decline from a reference value requires an 11.1% gain on the smaller value to return to the reference point. A 20% decline requires 25%. This is ordinary percentage math, not a forecast: the deeper the decline, the larger the percentage recovery required from the remaining base.

Changing size also changes exposure directly. CME Group’s education for individual traders notes that exposure rises with more contracts and falls with fewer, while its trade-plan guidance asks traders to define parameters such as maximum trade loss, maximum day loss, and maximum account exposure in advance.¹² Those sources do not prescribe a drawdown ladder. They support the underlying sequence: establish the risk plan first, then make quantity fit it.

Reducing risk has another practical function: it separates recovery pressure from the position-sizing input. Research on risky choice does not show that every person takes more risk after a loss. Imas found different responses after realized and paper losses in experimental settings.³ The defensible response is therefore not to diagnose yourself from the P&L. It is to compare the next proposed exposure with the rule that was active before the opportunity appeared.

Build a drawdown risk ladder before you need it

A drawdown risk ladder maps an observable account state to a permitted level of planned risk. Each row needs an entry trigger, a risk multiplier, and an exit rule. The trader must derive those variables from their own tested method and capacity for loss.

State	Entry condition	Permitted planned risk	Exit condition
Normal	The chosen drawdown measure is below the first trigger	Normal risk unit	First reduction trigger occurs
Reduced	The first trigger is active	Normal risk unit × reduced-state multiplier	A deeper trigger occurs or the restoration gate is passed
Minimum	The deeper trigger is active	Normal risk unit × minimum-state multiplier	Stop condition occurs or the restoration gate is passed
No new exposure	The stop condition is active	No new position risk	The separate review and reset rule is completed

“Risk unit” means the trader’s own maximum planned loss for a qualifying trade in the normal state. The multiplier changes that input; the normal position-sizing method still converts the revised planned loss into quantity.

active maximum planned loss = normal risk unit × state multiplier

position size = active maximum planned loss ÷ estimated risk per unit

The position-sizing guide covers the calculation in detail. A drawdown state should not change the setup’s invalidation merely to preserve a familiar quantity. If the invalidation distance is wider, the calculated position becomes smaller. If the smallest tradable unit would exceed the active maximum planned loss, the trade does not fit that state.

1. Choose one reproducible drawdown measure

Write the reference peak and current value beside every state decision:

drawdown percentage = (reference peak − current value) ÷ reference peak × 100

For example, if the stated reference peak is $20,000 and the consistently measured current value is $18,400, the measured drawdown is 8%. This does not say whether 8% should be a trigger. It only makes the measurement auditable.

Do not silently choose whichever equity definition produces the preferred state. If open P&L counts for activation but not for restoration, state that asymmetry before it matters. If a new deposit changes the account balance, do not let it masquerade as trading recovery; follow the adjustment method written into the rule.

2. Define triggers as state transitions

A trigger is more useful when it names the response it activates. “At the first drawdown boundary, risk becomes reduced” is clearer than “be cautious.” Include whether the condition activates when the measure touches or crosses the boundary and when the new state begins.

The trigger does not have to be purely financial. A trader may design a separate review trigger for repeated execution deviations, because a drawdown produced while following the method presents a different question from one accompanied by size drift or unplanned trades. Keep the evidence separate even if both triggers lead to the same temporary risk state.

3. Set the degree of reduction independently of the next setup

The multiplier should exist before a compelling setup appears. Otherwise, confidence in the next trade can become an unrecorded exception. The reduced state governs the risk input; it does not grade setup quality or determine whether a setup has an edge.

Evaluate a candidate multiplier across the whole ladder before using it. Project the cumulative exposure still possible under the state’s trade and session limits. Recalculate representative trades with the actual instruments, invalidation distances, unit values, and costs. Check whether the smallest tradable quantity can remain inside the reduced limit. Finally, decide whether the state permits enough comparable, rule-aligned decisions for the planned review; a nominal risk level that produces no usable decisions cannot generate evidence for restoration. This test does not identify the correct multiplier, but it can expose a ladder that is internally inconsistent.

Check all active boundaries before taking new exposure:

Does the setup qualify under the unchanged strategy rules?
Which drawdown state is active under the chosen measurement?
What maximum planned loss does that state’s multiplier permit?
What quantity follows from the normal sizing calculation?
Do session, account, and external constraints permit the trade?

A qualified setup can still be prohibited by the active risk state. Conversely, reduced size cannot make an unqualified setup valid.

4. Give the deepest state a stop decision

Repeated reductions can eventually create tiny positions without resolving the underlying question. Define what happens at the deepest boundary. The response may be no new exposure pending a strategy, execution, and risk review. It should not be an improvised promise to “win back enough first.”

This is where the broader trading risk management framework remains the parent process. The drawdown ladder governs changing trade-level permission across a loss sequence; account capacity, session limits, order mechanics, and external requirements still apply.

A hypothetical risk-reduction example

Assume a trader has already chosen these illustrative rules outside live trading. Drawdown uses end-of-day equity after trading costs from a $20,000 reference peak:

normal risk unit: $200 of maximum planned loss per qualifying trade;
at an 8% drawdown, enter the reduced state at 0.50 times the normal risk unit;
at a 12% drawdown, enter the minimum state at 0.25 times the normal risk unit;
at a 16% drawdown, permit no new exposure until the stated review is completed;
restore minimum to reduced only below a 10% drawdown after the defined aligned sample and scheduled review; and
restore reduced to normal only below a 6% drawdown after a new aligned sample and scheduled review.

These figures are examples of mechanics, not appropriate settings for another trader.

End-of-day equity reaches $18,400, an 8% drawdown, so the reduced state becomes active. On the next independently qualifying trade, estimated risk per share is $0.40:

active maximum planned loss = $200 × 0.50 = $100

calculated size = $100 ÷ $0.40 = 250 shares

If the same trade had appeared in the normal state, the formula would have produced 500 shares. The setup, entry, and invalidation did not improve or deteriorate because of the prior losses. Only the account-level permission changed.

Later, end-of-day equity reaches $17,600, a 12% drawdown. The minimum state now permits $50 of maximum planned loss. At the same $0.40 estimated risk per share, calculated size is 125 shares. If equity then improves to $17,800, the measured drawdown is 11%; the trader remains in the minimum state because the 10% restoration boundary has not been crossed.

When equity closes at $18,100, the measured drawdown is 9.5%. If the pre-defined aligned sample is complete and the scheduled review approves the transition, the next session begins in the reduced state—not the normal state. A later close at $18,900 is a 5.5% drawdown. Normal risk returns only after the new aligned sample and review gate also pass. This example shows the complete mechanism: a trigger moves risk down immediately, while a separate threshold and process evidence move it up one state at a time.

How should you return to normal risk after a drawdown?

Return to normal risk through staged, pre-defined gates that examine the account state and the process that produced it. Avoid making “back to break-even” the only gate. A financial threshold shows where equity stands; it does not show whether execution drift was corrected or whether aligned strategy performance has stabilized.

A restoration rule can use four kinds of evidence:

Restoration evidence	Question it answers
Financial state	Has the consistently measured drawdown moved through the stated restoration boundary?
Aligned sample	Under the same strategy and rule version and review window, did applicable trades and valid non-trades follow the rules, with unclassified decisions kept visible rather than counted as aligned?
Execution state	Are recent position sizes, invalidations, attempts, and session boundaries classifiable and aligned?
Review decision	Has the scheduled review concluded “restore,” rather than the live trader deciding mid-session?

The trader must define the required review window and number of applicable decisions before the drawdown. Winning trades, profitable days, and percentage recovery are not substitutes for that definition.

Restore one state at a time

If the ladder has minimum, reduced, and normal states, move through the stated sequence rather than jumping from minimum to normal because of one large win. A staged restoration makes each transition observable and limits the effect of one outlier result on the next risk decision. Restoring position size after a drawdown works through a step-based version of this sequence, gated by sample size, profit factor, and expectancy rather than a single result.

Use separate entry and exit thresholds when needed

If the same boundary both activates and deactivates a state, small changes around that value can cause frequent switching. A trader may choose separate, pre-defined entry and restoration thresholds so the state does not flip with every minor variation. This design choice must be tested outside live execution; it is not permission to move a threshold after seeing the next trade.

Restore from evidence, not urgency

A common restoration failure is “I am close enough.” Nearness to the prior peak can make a larger trade feel like an efficient way to complete the recovery. But if normal risk is restored to reach the threshold, the outcome has started governing the risk rule.

Use a written if-then transition instead: if the stated financial, aligned-sample, and review conditions are all met, then the next session begins in the next risk state. Research on implementation intentions finds that specifying the cue and response in advance can help translate intentions into action across studied goal domains.⁴ That evidence does not validate any trading threshold or guarantee adherence. It supports making the transition explicit before pressure arrives.

Common failure modes

Reducing quantity without recalculating planned loss

Half the quantity does not always mean half the risk. Invalidation distance, unit value, existing exposure, costs, and execution assumptions still matter. Recalculate through the same instrument-specific method.

Moving the trigger during the drawdown

Changing the equity definition, ignoring costs, or resetting the reference peak can keep normal risk active after the original rule says otherwise. Record any proposed rule change for later review; judge the current sequence against the version that was active when it began.

Treating reduced risk as permission for more attempts

Smaller per-trade loss can still create equal or greater cumulative exposure if trade count, adds, or correlated positions expand. The drawdown state must coexist with attempt and session limits.

Letting one outcome override the restoration gate

One outcome changes the account value but provides little evidence about a recurring execution or strategy problem. Apply the full restoration gate rather than using relief as a transition rule.

Staying permanently reduced without a review rule

Risk reduction without a restoration process can turn into indefinite avoidance. If the minimum tradable quantity no longer fits the active limit, or if the trader cannot define evidence for restoration, the next step is a review—not repeated ad hoc resizing.

Confusing planned reduction with risk escalation

Moving from normal to reduced risk under a pre-defined rule is a planned transition. Restoring full risk before the exit condition is met is risk escalation, even if the full-risk amount would have been valid earlier in the sequence.

Review the complete drawdown episode

Do not review only the deepest point or the trade that recovered the most. Reconstruct every state transition:

reference peak → drawdown measure → trigger → required risk state
→ calculated risk → actual exposure → rule status → outcome → next state

For the full episode, ask:

Was the same drawdown definition used throughout?
Did each trigger change risk before the next order?
Did submitted size match the active maximum planned loss?
Were strategy qualification and risk permission kept separate?
Did any trade count, add, or session boundary expand while unit risk fell?
Which restoration conditions were met, and who made the transition decision?
Did normal risk return at the planned time or after a persuasive outcome?

Then separate the diagnosis. Aligned trades losing under the intended exposure belong in a strategy and market-condition review. Size, invalidation, attempt, or restoration deviations belong in an execution and behavioral review. The trading performance framework shows how to keep results, exposure, and execution on separate lines.

Where Costante fits

Costante supports the behavioral layer around a trader’s own drawdown process. A trader can make the intended state visible before a decision, preserve the trigger and actual exposure through low-friction logging, and use structured review to inspect whether restoration followed the pre-defined gate. Session planning, self-defined behavioral guardrails, and pre-trade or in-session checks support that observable sequence.

Costante does not choose drawdown thresholds or multipliers, calculate position size, monitor live account equity, receive broker data, grade a strategy, decide when normal risk should return, place or block orders, or enforce account rules. The trader remains responsible for the method, risk design, interpretation, and every execution decision.

Sources

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

CME Group. Position and Risk Management. ↩
CME Group. Risk Management and Your Trade Plan. ↩
Imas, A. (2016). The Realization Effect: Risk-Taking after Realized versus Paper Losses. American Economic Review, 106(8), 2086–2109. ↩
Gollwitzer, P. M., & Sheeran, P. (2006). Implementation Intentions and Goal Achievement: A Meta-analysis of Effects and Processes. Advances in Experimental Social Psychology, 38, 69–119. ↩