Should you let math size your bets or stick to a steady rule?
Both approaches control risk, but they push your account in very different directions.
Kelly uses your win rate and payoff to chase maximum compound growth, often increasing position size when the edge looks strong.
Fixed fractional just risks the same share of equity every trade for smoother equity curves and easier psychology.
Thesis: Kelly can maximize returns if your edge estimates are solid and you tolerate big drawdowns; fixed fractional often produces better real-world results when estimates are noisy.
Core Comparison of Kelly Criterion vs Fixed Fractional Sizing
Kelly Criterion calculates how much to risk based on your actual edge and the variance of what you’re trading. The formula looks like this: fraction = (win probability × payoff ratio − loss probability) ÷ payoff ratio. When your edge is strong, Kelly sizes bigger. When it weakens, it pulls back. The whole point is maximizing long-term growth by putting more money into high-probability trades with good payoffs and less into marginal setups.
Fixed Fractional just picks a percentage and sticks with it. Every single trade gets the same treatment. Position risk = fixed percentage × current account balance. You might risk 2% whether the setup has a 55% win rate or 65%. Doesn’t matter. This thing prioritizes stability and controlling drawdowns over chasing maximum growth. Equity curves end up smoother because position sizes scale with your account in a predictable way, never adjusting for how good the trade actually looks.
Kelly wins when you’ve got solid stats and can handle some volatility in exchange for better compounding. Fixed Fractional works better when your edge estimates are shaky, when you need psychological comfort, or when market conditions shift enough that old win rates stop being reliable.
Key differences:
- Risk: Kelly adjusts based on edge quality. Fixed Fractional keeps it constant.
- Volatility: Kelly swings harder. Fixed Fractional stays steady.
- Growth rate: Kelly maxes out theoretical growth if your inputs are right. Fixed Fractional trades growth for stability.
- Error tolerance: Kelly breaks down fast if you’re wrong about your edge. Fixed Fractional doesn’t care about input errors.
- Complexity: Kelly needs ongoing calculations. Fixed Fractional just needs your account balance and a percentage.
Detailed Formulas and How Each Method Calculates Position Size
Understanding the actual math lets you implement these correctly without screwing up. Kelly uses probability-weighted outcomes. Fixed Fractional is basic arithmetic.
Kelly Criterion calculation:
- Figure out your win probability (p) from past trades. Winning trades divided by total trades.
- Loss probability (q) is just 1 − p.
- Get your average payoff ratio (b). Average winner divided by average loser.
- Plug it in: f* = (b × p − q) ÷ b.
- Multiply that fraction by your current equity to get position size in dollars.
- If you’re using fractional Kelly, adjust down (half-Kelly = 0.5 × f, quarter-Kelly = 0.25 × f).
Fixed Fractional calculation:
- Pick your fixed risk percentage. Most people use 1% to 3%.
- Multiply current equity by that percentage. That’s your dollar risk per trade.
- Figure out your stop distance (entry price minus stop price).
- Divide dollar risk by stop distance. That’s your share count or contract size.
The big difference in what you need is that Kelly demands accurate future estimates of win probability and payoff ratios, which change as markets evolve. Fixed Fractional only needs your current balance and a single policy decision about risk percentage. Kelly’s powerful when you have good data and stable conditions. It’s fragile when your estimates drift.
Side‑By‑Side Characteristics and Performance Behavior
Comparing how these actually behave helps match method to trading style and risk appetite.
| Method | Volatility | Drawdown Expectation | Sensitivity to Input Error | Growth Potential |
|---|---|---|---|---|
| Full Kelly | Very high | Can exceed 40–50% in adverse sequences | Extreme. Small estimation errors cause large position sizing mistakes | Theoretical maximum geometric growth |
| Half-Kelly | Moderate-high | Typically 20–30% range | Moderate. Errors have reduced impact | High growth with improved stability |
| Fixed Fractional 2% | Low | Generally under 15% | Very low. Method is input-agnostic | Steady but suboptimal when edge is strong |
| Fixed Fractional 1% | Very low | Typically under 10% | Very low | Conservative, slow compounding |
These performance differences mean aggressive traders with high-confidence statistical edges should lean toward fractional Kelly. Traders managing uncertain edges or prioritizing capital preservation should stick with Fixed Fractional in the 1–2% range. Full Kelly is rarely practical because the drawdowns destroy discipline. Most traders abandon the method during a 30% drawdown even when the math says keep going. Fixed Fractional’s predictable risk makes it easier to stick with the plan during losing streaks, which is why it dominates among discretionary traders and anyone without deep backtested edge measurements.
Numerical Examples Demonstrating Position Size Calculations
Example 1: Kelly Criterion with 55% win rate and 1.2 reward-to-risk ratio
You’ve got a $50,000 account. Your strategy wins 55% of trades (p = 0.55), and your average winner is 1.2 times your average loser (b = 1.2). Loss probability: q = 1 − 0.55 = 0.45. Kelly formula: f* = (1.2 × 0.55 − 0.45) ÷ 1.2 = (0.66 − 0.45) ÷ 1.2 = 0.21 ÷ 1.2 = 0.175. Full Kelly says risk 17.5% of your account per trade, which equals $50,000 × 0.175 = $8,750 per trade. Because full Kelly produces crazy volatility, most traders use half-Kelly: 0.175 ÷ 2 = 0.0875, so half-Kelly position size = $50,000 × 0.0875 = $4,375. If your stop is $1 per share below entry, you’d buy $4,375 ÷ $1 = 4,375 shares at half-Kelly sizing.
Example 2: Fixed Fractional using 2% risk per trade with a $50,000 account
You decide to risk 2% of equity on every trade no matter what the setup looks like. Dollar risk per trade = $50,000 × 0.02 = $1,000. You enter a stock at $50 and place a stop at $49. Stop distance is $1. Position size in shares = $1,000 ÷ $1 = 1,000 shares. Total position value is 1,000 × $50 = $50,000, which is 100% of your account, but your actual risk is capped at $1,000 (2%) if the stop gets hit. If the stock moves favorably and you capture a $2 gain per share, profit is 1,000 × $2 = $2,000, a 4% account gain. Fixed Fractional keeps the same 2% risk structure for the next trade even if your edge estimate changes, providing consistent risk exposure across all setups.
When Kelly Criterion or Fixed Fractional Works Best
Kelly Criterion works best when you’ve got a quantified, stable statistical edge from a large sample of historical trades (ideally 300 or more) and can accurately estimate win probability and average payoff ratios. It suits algorithmic traders, market makers, and systematic strategies where edge persists across time and you have high confidence in parameter estimates. Kelly also fits traders who can psychologically tolerate large drawdowns (20–40%) in exchange for maximum long-term compounding, and who operate in markets where probabilities stay relatively constant.
Fixed Fractional works best when your trading edge is uncertain, your strategy relies on discretionary judgment, or you trade in markets with shifting regimes where historical win rates become unreliable. It’s perfect for conservative traders, beginners building track records, and anyone prioritizing smooth equity curves over maximum growth. Fixed Fractional also suits strategies with many simultaneous positions because it prevents aggregate exposure from spiraling during favorable runs. Kelly can accidentally concentrate risk when multiple positions all show strong edges at once.
Market conditions where each method excels:
Kelly: Low-volatility trending markets with stable correlations, algorithmic execution with minimal slippage, and statistically validated edges updated regularly. Single-instrument specialists with deep historical data and consistent market microstructure.
Fixed Fractional: High-volatility or choppy markets, discretionary setups with subjective entry criteria, early-stage strategies with fewer than 100 completed trades, and portfolios holding correlated positions. Multi-strategy portfolios, traders scaling up capital, and anyone who’s previously abandoned a sizing method due to psychological stress during drawdowns.
Limitations and Drawbacks of Both Methods
Both position sizing methods carry hidden risks that only show up after implementation in live markets.
Kelly Criterion limitations:
Requires highly accurate estimates of win probability and payoff ratio. Small errors produce large position-sizing mistakes, often leading to overbetting and accelerated drawdowns. Assumes trade outcomes are independent. Correlated losses (common in trending markets) can cause simultaneous hits across positions, violating Kelly assumptions. Full Kelly produces psychologically unbearable drawdowns (30–50%) that cause traders to abandon the method mid-drawdown, locking in losses and destroying the long-term compounding advantage.
Fixed Fractional limitations:
Ignores edge quality entirely. Allocates the same risk to high-probability setups and marginal ones, underutilizing strong edges and potentially overexposing weak ones. Grows accounts slowly compared to Kelly when a genuine statistical advantage exists, leaving money on the table. Doesn’t adapt to changing market conditions. A fixed 2% risk might be too aggressive in high-volatility regimes and too conservative when volatility compresses.
Method selection depends on your tolerance for estimation error and psychological resilience. If you overestimate your edge by even 10%, full Kelly can double your intended risk exposure, turning a 20% expected drawdown into a 40% realized one. Fixed Fractional’s error tolerance is near-zero because it doesn’t rely on probabilistic inputs. You might grow slower than optimal, but you won’t accidentally over-leverage. Traders who struggle with discipline during losing streaks typically achieve better long-term results with Fixed Fractional’s predictability, even though the math favors Kelly when followed perfectly.
Practical Implementation Steps for Traders
Successful implementation starts with honest assessment of your edge quality and risk tolerance, not just picking a formula. Collect at least 100 completed trades (preferably 300+) to estimate win rate and average payoff ratio with reasonable confidence. Calculate your current win probability by dividing winning trades by total trades, then compute your reward-to-risk ratio by dividing the average size of wins by the average size of losses. Determine your maximum acceptable drawdown. Most retail traders can psychologically tolerate 15–25% before discipline breaks down. Use these measurements to decide whether your data supports Kelly or whether Fixed Fractional’s simplicity better matches your edge uncertainty.
Implementation sequence:
- For Kelly: Calculate full Kelly fraction using f* = (b × p − q) ÷ b, then immediately reduce to half-Kelly or quarter-Kelly by multiplying by 0.5 or 0.25 to control volatility and estimation error.
- For Fixed Fractional: Choose your risk percentage (1% for conservative, 2% for moderate, 3% for aggressive) and commit to that number for at least 50 trades before adjusting.
- Convert your chosen fraction or percentage into dollar risk per trade by multiplying by current account equity. Recalculate this after every trade or weekly to compound gains and scale down during losses.
- Translate dollar risk into position size by dividing by your stop-loss distance in price terms (entry minus stop for longs, stop minus entry for shorts).
- Set hard position caps regardless of method. Never risk more than 5% of equity on a single trade, and limit total portfolio exposure to avoid concentration risk when multiple positions trigger simultaneously.
- Review and update edge estimates every 50–100 trades. If your win rate or payoff ratio shifts by more than 10%, recalculate Kelly fractions or reassess whether your Fixed Fractional percentage remains appropriate for current market conditions.
Final Words
We walked through what each method is, the core formulas, step‑by‑step calculations, numeric examples, and when each approach performs best.
Bottom line: Kelly seeks maximum long‑term growth but needs precise edge estimates and brings higher volatility. Fixed fractional uses a steady percent of equity, delivering smoother drawdowns and more error tolerance.
Use position sizing Kelly criterion vs fixed fractional method as a practical framework: pick Kelly with high‑confidence edges, choose fixed fractional for uncertainty, and prefer partial‑Kelly when in doubt. Trade with clearer risk rules and confidence.
FAQ
Q: What is the Kelly Criterion and how does it compare to Fixed Fractional sizing?
A: The Kelly Criterion maximizes long‑term growth using estimated edge and variance; Fixed Fractional sizes trades by a constant percent of equity. Kelly grows faster but is more aggressive and input‑sensitive; Fixed Fractional is steadier and safer.
Q: How does the Kelly formula calculate position size?
A: Position size under the Kelly Criterion is calculated from your expected edge and variance (fraction ≈ edge/variance). For discrete trades use f* = (b·p − q)/b, which needs win probability and payoff ratio.
Q: How does Fixed Fractional position sizing work?
A: Fixed Fractional sizing uses a predefined percent of account equity per trade (for example 1% or 2%). Position size = account balance × chosen percentage, independent of estimated edge quality.
Q: When does Kelly outperform Fixed Fractional?
A: Kelly outperforms when your edge is stable, well‑estimated, and backed by ample data—producing higher geometric growth, though it brings larger drawdowns and more volatility.
Q: When is Fixed Fractional better?
A: Fixed Fractional is better when edges are uncertain, markets change, or you want smoother equity curves and smaller drawdowns; it suits rule‑based systems and conservative traders.
Q: What are the main risks and limitations of Kelly and Fixed Fractional?
A: Kelly can severely overbet if estimates are wrong and amplifies drawdowns; Fixed Fractional may underutilize high‑quality edges and slow account growth, though it’s more robust to estimation error.
Q: Can you show quick numerical examples for Kelly and Fixed Fractional?
A: A Kelly example: 55% win rate and 1.2 reward-to-risk → f* ≈ (1.2×0.55 − 0.45)/1.2 = 17.5% of equity. Fixed Fractional: 2% of $50,000 = $1,000 risk per trade.
Q: How sensitive are Kelly and Fixed Fractional to input errors?
A: Kelly is highly sensitive—small estimation errors can cause large overbets and volatility. Fixed Fractional is far more error‑tolerant, trading consistency over theoretical maximum growth.
Q: How should a trader choose between Kelly and Fixed Fractional?
A: Choose Kelly if you have reliable statistical edge, sufficient sample size, and high volatility tolerance. Choose Fixed Fractional if you prefer steadier growth, lower drawdowns, and simpler risk controls.
Q: What practical steps should traders follow to implement either method?
A: To implement: gather trade history, estimate win rate and payoff, set acceptable drawdown, compute full or partial Kelly (or fixed percent), cap sizing with a max percent, and monitor performance.
