Model edge is the difference between the Smashrs model's estimated win probability for an outcome and the book's implied probability for that same outcome. It is the single number that answers "does the market disagree with us, and by how much?"
How it's calculated
edge = model probability - book implied probability
Worked example. Suppose the model gives Player A a 60% chance to win. The book is offering odds on A that imply a 52% chance. The edge is +8 percentage points in A's favor: the market is pricing A as a worse bet than the model thinks they are.
A negative edge means the opposite. The book rates the outcome more likely than the model does, so there is nothing to back.
Important: edge is always measured against book odds
Edge compares the model to a real market price, the book odds from sources like the exchange or sportsbook. It is never derived from a "model-implied" line invented from our own probability. The whole point is to measure where our independent estimate departs from the price you can actually get.
A positive edge is the precondition for a value bet, and consistent edges across the model consensus are what we track in the public backtest.