Dynamic Position Sizer

Start conservative and adjust based on experience. Guidelines: Complete beginner: 0.5-1% max per trade. Intermediate: 1-1.5% typical. Experienced with proven system: 1.5-2%. Maximum for anyone: 3% (only with exceptional system and psychology). Factors to consider: your system's win rate and drawdown history, your emotional tolerance for losses, your account size (smaller accounts may need to accept slightly higher % to take meaningful positions), and your trading frequency (more trades = lower per-trade risk).

You have several options: 1) Step down to a Micro contract (best option in the US) - if 1 standard E-mini (ES, $50/pt) is too large, the Micro (MES, $5/pt) is 1/10 the risk per contract, so a 0.5-contract ES result becomes a clean 5-contract MES result. The same applies to MNQ vs NQ and MGC vs GC. 2) Skip the trade - if even 1 Micro exceeds your risk limit, the trade is too risky for your account size. 3) Use a wider stop - if stop distance is the issue, a wider stop reduces risk per contract, but only if technically valid. The availability of Micro contracts is the key US advantage - it largely solves the small-account granularity problem that exists in markets without micro-sized instruments.

Yes, automatically through the fixed percentage method. After wins: account grows, 2% of larger account = larger dollar risk = larger position. After losses: account shrinks, 2% of smaller account = smaller dollar risk = smaller position. This is automatic and correct - you don't manually increase or decrease. What to avoid: emotionally increasing size after wins ('I'm hot') or decreasing after losses ('I'm unlucky'). The math handles this correctly. Some advanced traders use equity curve methods to reduce during drawdowns, but basic automatic scaling is sufficient for most.

Portfolio risk limits: Conservative: 5-6% maximum total risk across all positions. Moderate: 8-10% maximum. Aggressive: 10-15% maximum (not recommended for most). Why limits matter: if you have 10 positions each with 2% risk, that's 20% total risk. If market crashes and all stop out, 20% drawdown. Most traders can't handle 20%+ drawdowns emotionally. Better to limit total exposure. Practical approach: if at your portfolio limit, don't take new trades until existing positions close or reduce.

Options have unique considerations (and in the US, 1 contract = 100 shares of the underlying). For long options: your maximum loss is the premium paid, so position size = account × risk% / (premium × 100). Simple - you can't lose more than premium. For short options: potential loss can exceed premium significantly. Use notional exposure or worst-case loss (to your stop) for sizing. For spreads: use max loss of the spread for sizing. Additional factor: options decay (theta), so even winning direction trades can lose money. Consider sizing more conservatively for options, especially short options. Many traders limit options positions to 50% of what their normal sizing would suggest.

Step-by-step implementation: 1) Calculate 14-day ATR for your instrument. 2) Decide ATR multiplier for stop distance (typically 2-3×). 3) Stop distance = ATR × multiplier. 4) Position size = (Account × Risk%) / (Stop distance × Point value). Example: E-mini S&P 500 at 6,000, ATR = 60 points. Stop distance = 2 × 60 = 120 points. Account $50,000, risk 2% = $1,000. Using the Micro (MES, $5/pt): Size = 1,000 / (120 × 5) = 1.67 → 1 contract. Key: recalculate ATR regularly (weekly is fine). Use consistent multiplier. The beauty is automatic adjustment - when market calms, ATR falls, stops tighten, size increases. When volatile, opposite happens. Consistent risk despite changing conditions.

Practical correlation tracking: 1) Simple method: categorize positions as 'index long', 'index short', 'sector X', etc. Positions in same category are highly correlated. 2) Calculation method: use 30-60 days of returns, calculate correlation coefficient in Excel (=CORREL function). 3) Approximation method: S&P 500-Nasdaq-100 ~0.9. S&P 500-large cap stocks ~0.6-0.8. S&P 500-gold ~-0.1 to +0.2. Different sectors ~0.3-0.6. Application: if adding position correlated >0.7 with existing, reduce new position by 30-50%. Or count as partial position for portfolio limits (e.g., 2 correlated positions = 1.5 effective positions). Don't overthink - rough categories work well enough.

Drawdown sizing strategies: Method 1 - Fixed percentage continues: keep risking 2% of current (smaller) equity. This automatically reduces dollar risk as account shrinks. Method 2 - Step-down: At 10% drawdown: reduce to 1.5%. At 15% drawdown: reduce to 1%. At 20% drawdown: reduce to 0.5% or stop trading. Method 3 - Equity curve: if equity below 20-day MA of equity, reduce by 50%. Recommendation: at minimum, continue fixed percentage (automatic reduction). Consider step-down if drawdown exceeds 15%. The goal: survive the drawdown, preserve capital for recovery. Don't try to 'make it back' with larger size - that's how accounts die.

Position sizing directly determines survival probability. Risk of ruin formula (simplified): depends on win rate, reward/risk, and position size. At 2% risk: even with poor 40% win rate and 1.5:1 R:R, risk of ruin is <1%. At 5% risk: same system has ~5% risk of ruin. At 10% risk: same system has ~20% risk of ruin. At 20% risk: same system has >40% risk of ruin. Key insight: doubling position size roughly triples risk of ruin. Conservative sizing (1-2%) makes ruin virtually impossible for any reasonable system. Aggressive sizing makes ruin likely even for good systems. This is why sizing is more important than strategy - good strategy with bad sizing still has high ruin risk.

Tiered sizing approach: A+ setups (all criteria met, high confidence): full 2% risk. A setups (most criteria met, good confidence): 1.5% risk. B setups (marginal, less confidence): 1% risk. C setups (barely meets criteria): 0.5% or skip. Implementation: define what makes A+, A, B, C for your strategy. Be honest about categorization. Track results by category to validate your assessment. Caution: don't let this become excuse for over-sizing 'can't miss' trades. Even A+ setups fail sometimes. The categorization should be objective, not emotional. If you don't have clear criteria for setup quality, just use fixed sizing.

Optimal f calculation: 1) Gather all trade P&L from system (minimum 50-100 trades). 2) For each f from 1% to 50% in increments of 1%, calculate: terminal wealth assuming that f was used for all trades (compound results). 3) Plot terminal wealth vs f. Find f that maximizes terminal wealth. 4) Practical use: never use full optimal f. Quarter optimal f (optimal f × 0.25) is aggressive but survivable. Half optimal f × 0.5 is common among quantitative traders. Track optimal f monthly: if optimal f increasing, system improving. If decreasing, system may be degrading. Warning: optimal f assumes known, stable probabilities. Real trading has estimation error and changing conditions. Use optimal f as upper bound, not target.

Risk parity implementation: 1) Decide total portfolio volatility target (e.g., 10% annual, or ~0.6% daily). 2) Decide number of positions (e.g., 5). 3) Each position contributes: target vol / positions = 2% annual each. 4) For each instrument, calculate daily volatility (std dev of returns or ATR%). 5) Position size = (Target contribution × Account) / (Instrument vol × Point value × Price). Example: Target 2% annual contribution = 0.13% daily. Instrument daily vol = 1.5%. Account $100,000. Size = (0.0013 × 100,000) / (0.015 × point value × price). Adjust for actual instrument characteristics, and use Micro contracts where standard contracts would overshoot the target. Result: each position contributes equal risk regardless of instrument volatility. Rebalance monthly as volatilities change.

Monte Carlo process: 1) Export all trade P&L (minimum 100 trades). 2) For each f level to test (1%, 2%, 3%, etc.): a) Randomly shuffle trade order. b) Apply position sizing to calculate account progression. c) Record: terminal value, max drawdown, whether ruin occurred (<50% of start). d) Repeat 10,000 times. 3) For each f, calculate: median terminal, 5th percentile terminal, max drawdown distribution, ruin probability. 4) Select f where: ruin probability <1%, 5th percentile outcome is acceptable, median outcome meets goals. Implementation: Python with numpy makes this straightforward. Excel can work for smaller simulations. Key insight: Monte Carlo reveals that same trades in different order produce vastly different outcomes. This shows the importance of conservative sizing.

System architecture: 1) Data layer: broker API for equity, positions; market data for ATR, VIX; database for historical trades. 2) Calculation layer: base sizing: (equity × risk%) / (stop × point value). ATR adjustment: multiply by (normal ATR / current ATR). Correlation adjustment: reduce if correlated with existing. Regime adjustment: multiply by regime factor. 3) Validation layer: round to whole contracts, choose standard vs Micro, check vs max position, check portfolio limit, verify margin. 4) Output layer: return final contract size, log all calculations, alert if trade skipped due to limits. Integration: function takes (entry, stop, instrument, direction) → returns contract size and reasoning. Call before every trade. Store all calculations for analysis. Build incrementally: start with basic percentage, add adjustments one at a time, validate each addition improves outcomes.

Institutional approaches: 1) Risk budgeting: allocate VaR (Value at Risk) budget across strategies/positions. Each strategy has VaR limit. Position sized to use VaR allocation. 2) Portfolio optimization: mean-variance optimization for capital allocation. Positions sized to target portfolio Sharpe ratio. 3) Factor-based: size based on factor exposure (momentum, value, etc.). Target specific factor exposures across portfolio. 4) Liquidity-adjusted: larger positions in liquid instruments, smaller in illiquid. Size limited by ability to exit within X days. 5) Regulatory capital: bank trading desks size based on regulatory capital requirements. Retail adaptation: apply risk budgeting concept (allocate risk % to each position). Use simplified portfolio vol targeting. Consider liquidity (avoid positions you can't exit quickly). Key principle: institutions think portfolio-first, not trade-first. Adopt this mindset.