Reference Class Forecasting

## The Outside View When making predictions, humans naturally default to the "inside view" — focusing on the specific details of the case at hand. Reference class forecasting uses the "outside view" instead: what is the base rate for this class of event? ## The Classic Example An analyst is asked to predict whether a new company will be profitable within 3 years. The inside view focuses on this company's specific business plan, management team, and market opportunity. The outside view asks: what percentage of companies in this industry with similar starting conditions became profitable within 3 years? If the historical base rate is 25%, that is your starting probability — regardless of how compelling the specific business case looks. ## Application to Betting **The underdog upset rate:** Rather than analysing whether this specific underdog can win, ask: what is the historical win rate for underdogs in this specific odds range against this type of favourite in this competition? If underdogs at 4.00–5.00 win 22% of matches (implied probability 20–25%), and the bookmaker prices this underdog at 4.50 (22% implied): no value. The base rate and the price are aligned. **The clean sheet rate:** Rather than assessing whether this specific defence will keep a clean sheet, ask: what percentage of matches between teams at this quality differential produce a clean sheet for the better defence? ## The Base Rate Table Build a personal base rate reference table from historical data in your target leagues: - Win rate by odds range (1.20–1.40, 1.40–1.60, etc.) - Draw rate by match quality differential - Clean sheet rate by defensive ranking - Over/Under hit rate by expected goals range These base rates are the empirical outside view against which inside-view adjustments should be modest and specific.