market-mechanics-betting_skill

Overview

This skill translates probability estimates into concrete betting and decision actions, helping you decide whether to bet, how much to stake, when to hedge, and how to improve forecast scores. It combines edge calculation, Kelly sizing, extremizing aggregated forecasts, and Brier-score optimization to convert beliefs into optimal, accountable moves. Use it to turn a numeric belief into a defensible betting or allocation decision.

How this skill works

The skill inspects your probability, the market probability (converted from odds if needed), and computes edge = your_prob - market_prob to decide bet/pass. If edge is sufficient, it computes optimal stake using the Kelly formula and recommends a fractional Kelly (typically 1/4–1/2) with bankroll constraints. For aggregated forecasts it applies an extremizing multiplier to the average, and for scoring it calculates and decomposes Brier score into calibration and resolution with actionable fixes.

When to use it

• You have a probability and need a decision: bet, pass, hedge, or allocate capital.
• You want to calculate edge against market odds or convert odds into probabilities.
• You need an optimal stake recommendation (Kelly and fractional Kelly) with bankroll limits.
• You are aggregating multiple forecasts and want an extremized consensus.
• You want to measure or improve forecasting accuracy using Brier score and calibration.
• You manage a portfolio of correlated bets and need hedging or rebalancing guidance.

Best practices

• Require a minimum edge before betting (e.g., 3–10% depending on context) to cover fees and uncertainty.
• Use fractional Kelly (1/4 to 1/2) instead of full Kelly to reduce volatility and model-risk exposure.
• Define and use a dedicated betting bankroll and respect market minimums and liquidity limits.
• Extremize aggregated forecasts away from 50% with a factor (default ~1.3) and cap at sensible bounds.
• Track calibration over time with calibration plots and decompose Brier score into calibration and resolution.
• Account for correlations across positions; reduce sizes for highly correlated bets and rebalance regularly.

Example use cases

• You forecast a 70% chance on an event priced at 60% by the market; compute edge and Kelly stake.
• Combine ten expert forecasts, average them, and extremize to produce a sharper consensus for a market trade.
• Assess a series of past forecasts to calculate Brier score, identify calibration bias, and correct forecasting behavior.
• Manage a portfolio where multiple political bets are correlated; calculate portfolio Kelly and hedge exposures.
• Lock in profit by hedging when odds move in your favor, using paired bets to guarantee a return.