Max drawdown is the largest peak-to-trough decline in account equity over a period, expressed as a percentage. If an account peaks at $50,000 and falls to $35,000 before reaching a new high, max drawdown is 30%. It is the single number that tells you how much pain a strategy delivered in its worst stretch, and it matters more than average return when you size positions or evaluate a backtest.
Most glossary articles stop at the formula. This one goes further: the recovery math that punishes deep drawdowns, why interim peaks do not reset the calculation, how prop firms enforce daily and overall drawdown rules, and how to compute MDD in Excel or Python on an equity curve.
What is max drawdown in simple terms?
Max drawdown (MDD) is the worst loss an account suffered from a previous high before recovering to a new high. The number is always expressed as a percentage of the peak value. It captures downside risk in a way average return cannot, because two strategies can post identical annual returns while delivering very different worst-case drops.
A quick example. Strategy A returns 20% per year with a 12% max drawdown. Strategy B returns 20% per year with a 38% max drawdown. Same return, very different experience. Strategy B will force you out at the worst possible moment unless your sizing accounts for that drop.
The max drawdown formula
The formula is simple:
MDD = (Trough Value − Peak Value) / Peak Value
The result is negative, and most traders report it as an absolute percentage. To find it on a real equity curve:
- List every account balance point in chronological order.
- For each point, find the highest value reached up to that point (the running peak).
- For each point, calculate the percentage drop from that running peak.
- The largest drop is your max drawdown.
The key detail: the peak must occur before the trough. A new high after the trough does not erase the prior drawdown.
Worked example: how to calculate max drawdown
Consider a crypto futures account that posts the following monthly balances:
| Month | Equity | Running Peak | Drawdown |
|---|---|---|---|
| Jan | $10,000 | $10,000 | 0% |
| Feb | $12,000 | $12,000 | 0% |
| Mar | $9,500 | $12,000 | -20.83% |
| Apr | $13,500 | $13,500 | 0% |
| May | $11,000 | $13,500 | -18.52% |
| Jun | $8,100 | $13,500 | -40.00% |
| Jul | $14,000 | $14,000 | 0% |
Max drawdown is 40%, the June trough measured against the April peak of $13,500. Notice that the account hit a new high in July. The recovery does not reduce the historical MDD. Once a 40% drawdown happens, it stays in the record.
This is where many traders misread their results. A backtest that ends at a new equity high can still have hidden a brutal 40% draw mid-run. If you would have stopped trading at -25%, the strategy is unusable to you, regardless of the final return.
The recovery math that punishes deep drawdowns
Drawdowns are asymmetric. The gain required to recover is always larger than the loss percentage. Here is the math every trader should memorize:
| Drawdown | Gain Required to Recover |
|---|---|
| 5% | 5.3% |
| 10% | 11.1% |
| 20% | 25.0% |
| 30% | 42.9% |
| 40% | 66.7% |
| 50% | 100.0% |
| 60% | 150.0% |
| 75% | 300.0% |
A 50% drawdown demands a 100% return just to break even. A 75% drawdown requires you to quadruple what is left. This is the math that wipes out leveraged crypto accounts in a single week of revenge trading. The deeper the hole, the steeper the climb, and emotionally most traders cannot execute calmly after a 40%+ drop.
What is a good max drawdown percentage?
There is no single answer, but practical thresholds exist:
- Under 10%: Conservative. Common for low-leverage systematic strategies and professionally managed portfolios.
- 10–20%: Reasonable for active discretionary traders with disciplined sizing.
- 20–30%: Aggressive. Acceptable for high-conviction strategies if the trader can stomach it.
- Above 30%: Most retail traders abandon a strategy here, even if the backtest eventually recovers. Behaviorally unsustainable.
Prop firms enforce these limits directly. A typical funded account has a 5% daily loss limit and a 10% overall drawdown limit. Breach either, account closed. That structure forces position sizing discipline most retail traders never apply to themselves.
Max drawdown vs volatility: not the same thing
Volatility (standard deviation of returns) measures how much returns bounce around the average in both directions. Max drawdown only measures downside, and only the worst case. A strategy can have low volatility and still suffer a catastrophic drawdown during a regime change. A strategy can have high volatility and a modest MDD if the upside moves are larger than the downside.
For risk-adjusted comparison, traders use the Calmar ratio: annualized return divided by max drawdown. A Calmar above 1.0 means the strategy returns more per year than its worst drop. Above 3.0 is excellent. The related RoMaD (Return over Maximum Drawdown) uses the same logic for total returns.
How to compute max drawdown in Excel and Python
In Excel, with equity values in column B starting at B2:
- Column C (Running Peak):
=MAX($B$2:B2)filled down - Column D (Drawdown):
=(B2-C2)/C2filled down - Max Drawdown:
=MIN(D:D)
In Python with a pandas equity series:
python running_peak = equity.cummax() drawdown = (equity - running_peak) / running_peak max_drawdown = drawdown.min()
Run this on any backtest equity curve or live trading log. If you pull trade data from Bybit, Binance, or Hyperliquid via API, the same logic applies to the cumulative P&L series.
Using max drawdown to size positions
MDD is not just a report card. It is an input for sizing. The practical rule: estimate your acceptable account drawdown, then work backwards to per-trade risk.
If you can tolerate a 20% account drawdown and your strategy historically goes through losing streaks of 10 consecutive trades, your per-trade risk should not exceed 2% of equity. Run higher per-trade risk and the math guarantees you blow through your tolerance during a normal losing streak.
This is why every serious trader needs a journal that tracks the equity curve continuously, not just the win rate. Win rate hides drawdown. A 60% win rate strategy can still produce a 35% drawdown if the losers cluster.
How TraderNest helps you control max drawdown
Drawdowns are rarely caused by the market. They are caused by behavior: oversizing after a loss, holding losers past stops, doubling down on conviction trades, trading outside planned hours. TraderNest auto-syncs your trades from 10 crypto exchanges (Bybit, Binance, OKX, Bitget, MEXC, KuCoin, Gate.io, Kraken, Deribit, Hyperliquid) plus stocks via Alpaca, then AI Hawk scans the data for 15 behavioral patterns that drive drawdowns.
The patterns that matter most for MDD control:
- Revenge Trading: position size spikes after a loss. The single biggest cause of catastrophic drawdowns.
- Tilt Escalation: each successive loss triggers a larger trade. AI Hawk flags it before the streak ends your account.
- Inconsistent Risk Management: per-trade risk drifts from 1% to 4% without you noticing. The drawdown formula compounds quickly.
- Overtrading: more trades than your plan allows, eroding equity through fees and low-conviction setups.
TraderNest also calculates your live max drawdown, time under water, and Calmar ratio across every connected account, so you see the real picture, not the win-rate illusion.
Limitations of max drawdown
MDD is backward-looking. The historical worst case is not a cap on future losses. A strategy with a 15% historical MDD can produce a 35% drawdown next month if market conditions change. MDD also ignores frequency: an account with one 20% drop is treated the same as an account with five 20% drops, even though the second is clearly more fragile.
Use MDD alongside time under water (how long it took to recover), Calmar ratio, and the distribution of all drawdowns, not just the worst. A single statistic is never the full risk picture.
If you want to see your real max drawdown across every exchange account, plus the behavioral patterns causing it, read our complete guide to risk management in trading and start measuring what actually drives your equity curve.
