Risk Management
Risk of Ruin: The Hidden Math That Destroys Trading Accounts
Most traders focus on finding the right entry. Very few calculate the probability that their account hits zero before their edge plays out. That probability has a name: risk of ruin. And it explains why traders with profitable backtests still blow up their accounts.
Risk of ruin is not a character flaw or bad luck. It is a mathematical outcome driven almost entirely by one variable: how much capital you risk on each trade. Reduce that number, and ruin becomes vanishingly unlikely. Ignore it, and even a verified edge cannot save you.
Why Traders With an Edge Still Blow Up
A positive-expectancy system does not protect you from sequences of losses. Markets produce runs — sometimes long ones — that fall well outside what most traders mentally model.
Consider a system with a 55% win rate. The probability of hitting 10 consecutive losses is roughly 0.45^10 = 0.034%. That sounds negligible. But over 3,000 trades — a reasonable career volume — that sequence is nearly certain to occur at least once. At 10% risk per trade, 10 consecutive losses wipes 65% of your account. At 20% per trade, you have less than 11 cents on the dollar.
The math does not care that your backtest was profitable. Ruin does not wait for your edge to play out.
The Risk of Ruin Formula
The simplified risk of ruin formula for a fixed-fraction betting system:
Where:
- = probability of ruin
- = win rate minus loss rate (e.g. 55% win, 45% loss = edge of 0.10)
- = number of risk units in your bankroll (bankroll divided by amount risked per trade)
For a system with 55% win rate and 1:1 payoff (no R-multiple advantage), risking 2% per trade on a 10,000 USD account (500 risk units):
Change that to 10% per trade (100 risk units):
Change that to 20% per trade (50 risk units):
A 13% chance of blowing your account is not an extreme edge case. Run that scenario 10 times across a trading career, and ruin is essentially inevitable.
How Position Size Drives Ruin Probability
The relationship between position size and ruin probability is nonlinear. Small increases in risk per trade produce large increases in ruin probability once you cross a threshold.
| 1% Risk Per Trade | 2% Risk Per Trade | 5% Risk Per Trade | 10% Risk Per Trade | |
|---|---|---|---|---|
| Risk per trade | 1% | 2% | 5% | 10% |
| Risk units (10,000 USD account) | 1,000 | 500 | 200 | 100 |
| Ruin probability (55% win, 1:1) | Near zero | Near zero | 0.002% | 1.7% |
| Max drawdown on 10 consecutive losses | 9.6% | 18.3% | 40.1% | 65.1% |
| Recovery time from max drawdown | Fast | Manageable | Slow | Very slow |
| Verdict | Institutional standard | Recommended ceiling for retail | High variance — edge must be strong | Dangerous — avoid on live capital |
The difference between 2% and 10% risk per trade is not a 5x increase in ruin probability. It is the difference between near-zero risk and a coin flip that lands on ruin once in every 60 careers.
Three Ways to Calculate Your Risk of Ruin
Use when: quick sanity check, symmetrical win/loss system
Formula: R = ((1 - e) / (1 + e))^N — where e = win_rate - loss_rate, N = bankroll / risk_per_trade
Example: 58% win rate, 2% risk per trade on 25,000 USD account
e = 0.58 - 0.42 = 0.16
N = 25000 / 500 = 50
R = (0.84 / 1.16)^50 = 0.7241^50 ≈ 0.00003 (0.003%)
Result: Essentially zero ruin probability.
Use when: asymmetric payoffs, variable position sizing, more realistic modeling
Run 10,000 simulated sequences of N trades drawn from your actual win/loss/size distribution. Count how many sequences hit ruin level before reaching target.
Python sketch — see the full implementation below the tabs:
simulate(win_rate=0.58, r_ratio=1.5, risk_pct=0.02, bankroll=10000, target=20000, ruin_level=1000)
Returns the fraction of simulations that hit the ruin level before reaching target. Run this before sizing any live system.
Use when: your system has a known R-ratio and you want the mathematically precise threshold
The Kelly fraction f* = W - (1-W)/R gives the sizing that maximizes long-run growth with theoretically zero ruin (in the limit). Any sizing below Kelly produces finite ruin probability approaching zero as position count grows.
Sizing above Kelly guarantees eventual ruin regardless of edge.
Rule: If your position size exceeds your Kelly fraction, you are in the ruin zone. Compute Kelly first. Size below it. Always.
See Kelly Criterion position sizing for the full derivation and a live calculator.
Full Monte Carlo implementation in Python:
import random
def simulate(win_rate, r_ratio, risk_pct, bankroll, target, ruin_level, n_sims=10000):
ruin_count = 0
for _ in range(n_sims):
equity = bankroll
while ruin_level < equity < target:
risk = equity * risk_pct
if random.random() < win_rate:
equity += risk * r_ratio
else:
equity -= risk
if equity <= ruin_level:
ruin_count += 1
return ruin_count / n_simsThe Equity Curve When Ruin Approaches
The chart below simulates a 10% risk-per-trade system — a 58% win rate, 1.5R system — over 60 trades. Starting equity: 10,000 USD. The run of losses in trade 31–38 is not fabricated; it is within the normal variance of the win-rate distribution.
Equity Curve — 10% Risk Per Trade, 58% Win Rate (Simulated)
Peak equity reached 25,500 USD — a 155% gain. Then 7 consecutive losses pulled the account to 5,445 USD before recovery began. That is a 78% drawdown from peak. Most traders abandon the system at this point, locking in permanent loss.
Practical Ruin Thresholds
You do not need to hit zero to be ruined. Define your personal ruin threshold — the account level at which you can no longer afford to trade your strategy — and use that as your ruin target in calculations.
Common ruin thresholds:
- 50% drawdown — most systematic strategies require at least 50% remaining capital to size properly
- PDT minimum — 25,000 USD for US day traders; fall below and you are locked out of intraday execution
- Margin call level — broker-specific; varies by instrument and leverage
Whatever your threshold, calculate ruin probability against it, not against zero. A 50% drawdown that triggers a PDT lock is functionally ruin even if your account technically exists.
For a deeper look at sizing rules that prevent crossing these thresholds, see position sizing for algo traders.
Frequently Asked Questions
Can a high win rate eliminate risk of ruin?
A higher win rate reduces ruin probability, but it cannot eliminate it at aggressive position sizes. A 70% win rate system risking 15% per trade still carries non-trivial ruin probability over a long run.
The formula makes this clear: ruin probability scales with bankroll depth (N = bankroll / risk_per_trade). At 15% risk per trade you have just 6.7 risk units per 100% of account. A streak of 7 losses — statistically certain over thousands of trades even at 70% win rate — ends the account.
Win rate and position size both matter. Neither alone is sufficient.
Is a 20% drawdown really a risk of ruin concern?
It depends on your resilience — financial and psychological. A 20% drawdown is not ruin by definition, but it is a stress event that causes most retail traders to abandon their system, reduce size at the worst time, or make emotional decisions.
The goal of controlling risk of ruin is not just mathematical survival. It is behavioral survival: keeping position size small enough that losses are tolerable, so you stay in the system long enough for the edge to play out.
How often should I recalculate risk of ruin?
Recalculate whenever: your win rate estimate changes based on new live data, your account size changes materially (up or down), you adjust position sizing rules, or you add a new strategy to your portfolio.
A single calculation at the start of a system is not sufficient. Markets change, edge degrades, and accounts grow or shrink. Treat ruin probability as a live metric, not a one-time sanity check.