Risk Budgeting Across Strategies

The Multi-Edge Problem

You've done the work. You have three validated edges, each with positive expectancy. Each edge specifies position sizing. But how much capital do you allocate to each? And what happens when they all want to trade at once?

This is the risk budgeting problem - allocating limited risk capacity across competing opportunities.

Total Risk Budget

Start with your total risk tolerance. This is the maximum drawdown you can sustain financially and psychologically.

Most professional traders use 1-2% maximum risk per trade and 6-10% maximum total risk exposure at any time. Conservative? Yes. That's the point.

Choose based on your survival requirement, not your return desire.

Strategy-Level Allocation

Given a total budget, how do you divide it among strategies?

Method 1: Equal Weight Give each strategy the same allocation. Simple but ignores that some edges are stronger than others.

Method 2: Performance-Weighted Allocate more to strategies with better historical performance. Risk: you're chasing past returns that may not continue.

Method 3: Sharpe-Weighted Allocate based on risk-adjusted returns (Sharpe ratio). Higher Sharpe = more consistent = deserves more capital.

Method 4: Confidence-Weighted Allocate based on statistical confidence in the edge. More samples, longer track record = higher allocation.

We prefer a blend of Sharpe-weighted and confidence-weighted approaches.

The Overlap Problem

Your three strategies all want to go long BTC right now. Do you: A) Take all three positions (3× normal exposure) B) Take only one position (ignore redundant signals) C) Take a sized-up single position (acknowledge convergence)

The answer depends on whether the strategies are truly independent.

Independent edges: Each strategy has different entry logic, different exit logic, different time horizons. Convergence is meaningful information - multiple independent analyses agree. Scale up, but not 3×. Maybe 1.5-2×.

Correlated edges: The strategies use similar data or similar logic. Convergence is expected, not informative. Treat as single position.

Mixed: Some overlap, some independence. Partially scale up with caps.

Correlation Between Strategies

Just as assets can be correlated, strategies can be correlated. Two momentum strategies will both win in trends and both lose in ranges. Allocating 50% to each gives you 100% trend exposure.

Calculate strategy correlation using daily returns:

# Simplified approach
strategy_returns = [strategy1_daily_returns, strategy2_daily_returns, ...]
correlation_matrix = calculate_correlation(strategy_returns)

High correlation (>0.6) means you're concentrated in one type of edge. True diversification requires strategies that perform differently in different conditions.

The Portfolio of Strategies

Think of your strategies as assets in a portfolio. Apply the same portfolio theory:

Diversification benefit: Uncorrelated strategies reduce portfolio drawdowns even if individual strategies have similar expected drawdowns.

Risk parity: Allocate so each strategy contributes equally to portfolio risk. Higher volatility strategies get smaller allocations.

Rebalancing: Periodically adjust allocations as strategy performance diverges. Don't let winners become dangerously concentrated.

Dynamic Risk Allocation

Static allocations are a starting point. Dynamic allocation improves results:

Performance-Based Adjustment:

• •Strategy in drawdown → reduce allocation
• •Strategy at new highs → consider increasing allocation
• •The logic: good times may continue, bad times may continue

Regime-Based Adjustment:

• •Trend regime → favor trend-following strategies
• •Range regime → favor mean-reversion strategies
• •High volatility → reduce all allocations

Opportunity-Based Adjustment:

• •More signals than usual → signals may be lower quality
• •Fewer signals than usual → signals may be higher quality

Practical Implementation

Here's how we implement multi-strategy risk budgeting:

Step 1: Define total risk budget. Maximum 8% of account at risk at any time.

Step 2: Allocate to strategy categories.

• •Trend-following: 40% (3.2%)
• •Mean-reversion: 30% (2.4%)
• •Volatility: 20% (1.6%)
• •Reserve: 10% (0.8% for discretionary/new strategies)

Step 3: Within categories, allocate to specific edges. Based on confidence, Sharpe, and current regime appropriateness.

Step 4: Handle signal conflicts. When strategies conflict (one long, one short), reduce both or sit out. Conflicting edges mean uncertainty.

Step 5: Enforce caps. No single position exceeds 2× normal size regardless of signal convergence. No total exposure exceeds 10% regardless of opportunity.

When Edges Compete

Your capital is finite. When multiple edges signal simultaneously, you must prioritize:

Prioritize by expected value. Edge A has 55% win rate × 2R = 1.1 expected value. Edge B has 60% × 1.5R = 0.9 expected value. Trade A.

Prioritize by sample size. Edge A has 2,000 historical signals. Edge B has 200. Trade A (more statistical confidence).

Prioritize by recency. Edge A worked last month. Edge B hasn't worked in 3 months. Maybe the market has shifted. Trade A (with caution).

Prioritize by capacity. Some edges have limited capacity (only work in low volume). Others scale. If you have excess capital, favor scalable edges.

The Reserve Allocation

Always keep a reserve - unallocated risk capacity held for:

New opportunities: An edge you're testing in small size. A one-time market dislocation.

Mistakes: You will have execution errors, unexpected correlations, bugs. Reserve absorbs these.

Psychological buffer: Knowing you have unused capacity reduces pressure during drawdowns.

Never be 100% allocated. 10-20% reserve is standard.

Risk Budgeting in Practice

Here's a sample allocation for a $100,000 account:

Current usage of 1.7% means plenty of capacity for more signals. If usage approached 6-7%, we'd consider reducing position sizes on marginal signals.

Rebalancing Strategy Allocations

When should you rebalance?

Calendar rebalancing: Monthly or quarterly, reset to target allocations. Simple, mechanical, prevents drift.

Threshold rebalancing: When any strategy's allocation drifts >5% from target, rebalance. More responsive but more trading.

Performance rebalancing: After strategy drawdown exceeds X%, reduce allocation. After recovery, increase. Responds to actual results.

Avoid rebalancing too frequently - you'll just add trading costs and whipsaw between strategies.

Takeaway

Risk budgeting is where individual edges become a trading business. A collection of edges without allocation rules is just a collection. A risk-budgeted portfolio of edges is a system.

Key principles:

• •Set total risk budget before anything else
• •Allocate to strategies based on confidence and risk-adjusted returns
• •Account for correlation between strategies
• •Keep reserves for opportunities and mistakes
• •Rebalance periodically but not obsessively

Congratulations on completing the Risk level. You now understand position sizing, drawdown management, correlation, stop losses, and multi-strategy allocation. These aren't the exciting parts of trading - they're the parts that determine whether you survive long enough to benefit from your edges.

Next level: Automation. We'll build the actual infrastructure that turns edges into trades.