Understanding Maximum Drawdown in Trading Backtests
Learn what maximum drawdown measures, why it matters for backtesting, how it is calculated, and how traders can use the metric to shape risk controls.
Produced with automation, then checked by deterministic quality rules and an independent source-grounded review before publication.
- 01The maximum drawdown metric quantifies the greatest loss from a historical peak to a subsequent trough.
- 02It highlights periods of capital erosion that may not be evident from average returns alone.
- 03High drawdown values can indicate vulnerability to market volatility or strategy over‑fitting.
- 04Comparing drawdown across strategies helps assess risk tolerance and capital allocation decisions.
- 05Maximum drawdown should be evaluated alongside other metrics such as Sharpe ratio, win rate, and expectancy.
Maximum drawdown is the largest percentage loss a trading strategy suffers from a historical high point to the next low point during a backtest. It is expressed as a single number that captures the worst‑case decline in account equity. Understanding this figure helps traders gauge the depth of potential losses before committing real capital. The metric is purely historical; it does not predict future outcomes but provides a concrete reference for the strategy’s most severe stress period.
Why is maximum drawdown important for backtesting?
Backtests aim to simulate how a strategy would have performed under past market conditions. While average returns can look attractive, they may mask periods of severe loss. Maximum drawdown reveals those hidden risks, allowing traders to evaluate whether a strategy’s risk profile aligns with their tolerance. It also serves as a sanity check for data quality, because unrealistic drawdowns often point to gaps or errors in the price series.
How is maximum drawdown calculated?
The calculation follows a simple sequence: track the equity curve, identify each new peak, measure the decline to the subsequent trough, and record the largest decline as a percentage of the peak. The result is the maximum drawdown for the examined period. Practically, the steps are performed on a time‑ordered list of portfolio values, ensuring that timestamps are accurate and that missing data points are handled explicitly.
Step‑by‑step example
- 01The equity curve rises to $10,000 - this is the first peak.
- 02The curve then falls to $7,500 - the drawdown is ($10,000‑$7,500)/$10,000 = 25%.
- 03The equity later climbs to $12,000 - a new peak is set.
- 04The next trough is $9,000 - drawdown is ($12,000‑$9,000)/$12,000 = 25%.
- 05The largest percentage observed (25%) is reported as the maximum drawdown.
What does a high maximum drawdown tell you?
A high drawdown suggests the strategy may be exposed to large adverse moves, possibly due to leverage, concentration in a few instruments, or sensitivity to market regime changes. It does not guarantee future loss, but it signals that the strategy has historically struggled to preserve capital during stress periods. Analysts often pair this insight with an examination of the underlying trade logic to identify structural weaknesses.
How should traders use maximum drawdown when selecting strategies?
Maximum drawdown should be compared against the trader’s capital buffer and risk appetite. A strategy with a 30% drawdown may be unsuitable for a small account but acceptable for a larger, diversified portfolio. Combining drawdown with other risk metrics provides a fuller picture, and it informs decisions about position sizing, stop‑loss placement, and capital allocation.
- Set personal drawdown limits based on how much capital you are willing to lose before re‑evaluating the strategy.
- Consider the frequency and duration of drawdown periods; prolonged declines can erode confidence and affect performance.
- Use drawdown to inform position sizing and stop‑loss design, acknowledging that controls cannot eliminate all risk.
How can you improve drawdown performance in a backtest?
Improving drawdown often involves adjusting risk controls, diversifying across uncorrelated assets, or refining entry/exit rules. However, each modification introduces uncertainty and may lead to over‑fitting if not validated on out‑of‑sample data. A disciplined workflow includes re‑running the backtest with a fresh data slice and checking that drawdown improvements persist.
A lower maximum drawdown does not automatically mean a better strategy; it must be weighed against expected returns and the trader’s risk capacity.
Where can I learn more about backtesting best practices?
For a deeper dive into backtesting methodology, see the Step‑by‑Step Guide to Backtesting a Trading Strategy. To understand which metrics matter most, read Backtesting Metrics That Truly Matter for a Trading Strategy. For guidance on setting risk limits, consult How to Set a Maximum Order Size for Your AI Trading Agent.
Frequently asked questions
Maximum drawdown records the single largest peak‑to‑trough loss, while average drawdown averages all observed declines. Maximum drawdown highlights worst‑case risk; average drawdown reflects typical loss depth.
Yes, by applying tighter position sizing, adding stop‑loss orders, or diversifying across assets. These controls lower exposure but may also reduce upside potential.
It is relevant for any time horizon because both short‑term and long‑term strategies can experience significant equity erosion. The acceptable drawdown level may differ based on the strategy’s intended holding period.
Maximum drawdown reflects historical performance under specific market conditions. While it offers insight into past vulnerability, future market dynamics may differ, so it should be used alongside forward‑looking risk assessments.
No. A high Sharpe ratio can mask large intermittent losses. Evaluating both return‑adjusted performance and drawdown ensures a balanced view of risk and reward.
Accurate timestamps, price integrity, and coverage are essential. Missing or erroneous data can artificially inflate or deflate drawdown figures, leading to misleading conclusions.
Sources and verification
Product claims in this article were checked against these first-party references. Runtime status remains authoritative for current availability.
- Felix documentationfirst party
- Felix machine referencefirst party
Build with Felix now.
Felix infrastructure is live through MCP and the API. The full trading app launches September 17.
AI trading and algorithmic trading both automate market actions, but they rely on distinct technologies and risk‑management approaches. This article clarifies their definitions, key contrasts, and practical considerations for operators.
Backtest overfitting occurs when a strategy is tuned too closely to historical data, capturing noise instead of genuine signals. This article explains the phenomenon and offers concrete methods to avoid it, helping traders build more reliable models.