The Kelly Criterion vs. Fixed Risk: Which Sizing Method Works for Prop Traders
The Kelly Criterion is a position sizing formula that tells traders the mathematically optimal percentage of their account to risk on each trade. For prop traders chasing funded accounts at firms like Apex or MyFundedFutures, position sizing isn't academic—it's the difference between passing your evaluation and blowing the account. But should you follow the math, or stick to a simpler rule?
The Kelly Criterion: Theory vs. Execution
Developed by John Kelly at Bell Labs in 1956, it balances two competing goals: growing capital as fast as possible while avoiding catastrophic drawdowns.
The formula is Kelly % = W - (1 - W) / R, where W is winning probability and R is the win/loss ratio.
Here's a concrete example. Assume you have a 58% win rate, your average win is $150, and your average loss is $100. Your Kelly percentage would be: (0.58 × $150 - 0.42 × $100) / $150 = $45 / $150 = 0.30, meaning Kelly is 30%. That's aggressive—it means risking 30% of your account per trade.
Full Kelly is mathematically optimal for maximizing long-term capital growth, but it comes with high volatility. Half Kelly (50% of the Kelly percentage) is often preferred by professional traders because it captures ~75% of the optimal growth rate while significantly reducing drawdown risk.
The appeal is real: Kelly adapts to your actual edge. A trader with 55% win rate and 2:1 gets a different % than 60% at 1:1, even if both are profitable.
The Problem: Prop Firms Don't Trade Real Money
This is where Kelly meets the wall. Prop firms don't care about your long-term capital growth curve—they care about three things: your drawdown relative to your daily/trailing limit, your compliance with the firm's rules, and whether you survive the evaluation period.
The single biggest lever for reducing risk of ruin is position size. The relationship is exponential — doubling position size doesn't double risk of ruin, it can increase it by 10x or more. On a 25K Apex evaluation with a 5K daily drawdown limit, even a 55% win rate strategy can annihilate you if you hit a seven-trade losing streak while sized for Kelly.
Fixed Risk: The Prop Trader's Standard
For most retail traders, risking 1-2% of your account per trade is considered prudent risk management. Prop traders typically go lower. Professional traders target risk of ruin below 5%. Balsara's research demonstrated that even strategies with positive expectancy face significant ruin probability when position sizing exceeds 3-5% per trade.
Fixed risk is mechanical: if you're on a 25K account and risk 1%, that's $250 per trade, every trade. In futures trading, position sizing is typically calculated backward from your maximum acceptable loss per trade. The goal is to determine how many contracts you can trade based on the distance to your stop and the dollar value of each point or tick movement.
Example: If trading an E-mini S&P 500 (ES) contract valued near 6000 points, a 20-point stop represents $1,000 risk (20 × $50 per point = $1,000). That same trader could therefore take one contract to stay within the chosen limit.
This simplicity is a feature, not a bug. Under a trailing drawdown rule, fixed sizing lets you predict your worst-case scenario. You know exactly how many consecutive losses you can survive.
When Kelly Sizing Breaks Down for Prop Traders
Betting more than the Kelly fraction guarantees eventual ruin; betting less is safer when probability estimates are uncertain. Prop traders deal with uncertainty constantly: your 100-trade backtest win rate may not hold during a live market regime shift. Your payoff ratio collapses on a news event. Or you're undercapitalized and a single whipsaw exceeds your firm's daily limit.
Futures can confuse newer traders because margin determines what you can hold, while position sizing determines what you should hold. You can often carry more contracts than your strategy can survive.
Add the prop firm's trailing drawdown and you face a binary outcome: stay small enough to weather the firm's variance window, or blow up trying to optimize growth.
The Hybrid Approach: Fractional Kelly for Funded Accounts
Half Kelly (50% of the Kelly percentage) is often preferred by professional traders because it captures ~75% of the optimal growth rate while significantly reducing drawdown risk. Quarter Kelly is extremely conservative, capturing ~50% of the optimal growth rate but with much lower volatility.
For prop traders, the math looks like this:
- Calculate your full Kelly % from real trades (minimum 100-trade sample)
- Apply a safety multiplier: 0.25x (quarter Kelly) or 0.50x (half Kelly)
- Use that as your maximum risk-per-trade percentage
- Recalibrate every 30 trades or when your win rate shifts
This preserves the personalization of Kelly while respecting the hard reality of prop firm drawdown limits.
| Sizing Method | Win Rate Required | Max Risk/Trade | Pros | Cons |
|---|---|---|---|---|
| Fixed 1% | 40%+ | 1% account | Simple, predictable, low ruin risk | Leaves money on table in trending periods |
| Fixed 2% | 50%+ | 2% account | Standard in industry, measurable | Higher variance, tight prop firm fit |
| Full Kelly | 55%+ | 30%+ account | Mathematically optimal growth | Psych-destroying drawdowns, likely exceeds firm limits |
| Quarter Kelly | 52%+ | 7-10% account | Good Kelly edge + prop firm safety | Requires accurate edge data |
What Actually Works for Your Account
Track your real metrics for 50+ trades, not backtested ones. Log every trade with full details — Record entry price, exit price, position size, and outcome for every trade without exception · Calculate your win rate over a rolling window — Use at least your last 100 trades to get a stable win rate estimate · Calculate your payoff ratio — Divide your average win by your average loss over the same sample · Run the Kelly formula — Plug in your current win rate and payoff ratio, then multiply by your chosen fraction (0.25 to 0.50)
Once you have real data, test both fixed 1.5% and half-Kelly against your firm's trailing drawdown rules. Run a Monte Carlo on your actual trade sequence—not theoretical parameters. The method that doesn't hit your firm's limit first is the one you use.
For most funded traders, fixed 1-1.5% will be safer. If you have a proven, multi-month edge with win rates above 55% and positive payoff ratios, quarter to half Kelly can unlock better returns. But if you're evaluating or within your first 6 months, fixed risk keeps you alive long enough to actually develop that edge.
Disclaimer: PropLedger is a trade-journaling tool, not financial advice. Prop firm rules change frequently - always confirm the current rules with your firm. Trading futures involves substantial risk of loss.
Sources
- Kelly Criterion - Optimize Your Position Sizing | JournalPlus
- Kelly Criterion Calculator | Position Sizing Tool | TradeSearcher
- Kelly Criterion Position Sizing: Maximize Long-Term Returns While Avoiding Ruin | Astute Investors Calculus
- Risk of Ruin Calculator - New Trader U
- Risk of Ruin Formula - New Trader U
- Risk of Ruin Calculator | Switch Markets
- The Risk of Ruin in Trading: Probability of Ruin and Loss | QuantifiedStrategies.com
- Risk of Ruin Calculator — Will Your Strategy Survive? | JournalPlus
- Risk of Ruin Calculator | GhostTraders
- Risk of Ruin Calculator for Trading | Free + Formula | BacktestBase
- Futures Position Size Calculator | Risk Management
- R-Multiple | CrossTrade
- Position Sizing for Risk-Adjusted Scaling | Damn Prop Firms
- Futures Position Sizing: Contracts Calculator + Examples | TradeVia
- Position Sizing Principles | Optimus Futures
- How to Calculate Position Size & Manage Risk Like a Pro Trader | Day Trading Profit Calculator
- Advanced Risk Management Techniques for Futures: Position Sizing and Stop Loss Optimization | QuantStrategy.io
Related research
Why Position Size Matters More Than Your Edge
Position sizing controls survival more than edge. Learn why doubling risk per trade can increase account destruction 10x, not 2x.
Probability Analytics for Traders: Reading Your Own Edge Honestly
Win rate by setup, R-multiple distribution, time-of-day performance, and why sample size and confidence intervals decide whether any of it means anything.