Overview

Trading Risk is an MCP server that performs risk calculations for trading strategies using only user-provided inputs such as account balance, win probability, payoff ratio, and portfolio positions. It implements position sizing via Kelly criterion and its fractional variants, computes portfolio heat to gauge total risk, and enforces guardrails to cap exposure. This enables offline risk assessment without real-time market dependencies.

Key Capabilities

kelly_position_sizing: Applies full Kelly formula (position = (p * (b+1) - 1) / b) to compute optimal fraction of capital to risk per trade, where p is win probability and b is payoff odds.
fractional_kelly: Scales Kelly output by a user-defined fraction (e.g., 0.5) to reduce volatility while retaining edge.
portfolio_heat: Aggregates risk across positions, expressing total exposure as percentage of capital at risk if all trades lose.
risk_guardrails: Validates proposed positions against limits like max position size, sector concentration, or total heat thresholds, rejecting violations.

Use Cases

A day trader inputs win rate (60%), average win/loss ratio (1.5), and $100k balance; kelly_position_sizing outputs 12% allocation per trade, while fractional_kelly (half-Kelly) limits to 6%.
Portfolio manager enters positions in stocks and options; portfolio_heat reveals 25% total heat, prompting rebalancing via guardrails.
Algorithmic trader tests strategy offline: supplies backtest stats to kelly_position_sizing for sizing rules before live deployment.
Risk team audits trades: risk_guardrails flags a 15% single-position exceeding 10% limit.

Who This Is For

Quantitative traders, portfolio managers, and risk analysts who need precise, input-driven risk metrics. Developers integrating risk controls into trading bots benefit from its API simplicity, especially in air-gapped or backtesting environments.