Poisson Regression (Football Score Prediction)
A statistical model that predicts football match scorelines using historical attack and defense rates.
Poisson regression models goal counts as a Poisson distribution where λ (expected goals) is derived from team attack rates × opponent defense rates × home/away factor. It generates a probability matrix for every possible scoreline, used to compute 1X2, Over/Under, and BTTS true probabilities. SpinBonus's Sports Engine uses Dixon-Coles-adjusted Poisson with rolling 38-match form weighting.
Formula:P(home goals = h, away goals = a) = (e^−λH × λH^h / h!) × (e^−λA × λA^a / a!)
Example:λH = 1.8, λA = 1.2 → P(2-1) = 12.4%, P(1-0) = 9.8%, P(home win) = 51%.
Value Bet
A bet where the implied probability from the bookmaker odds is lower than the true probability.
A value bet (or +EV bet) exists when (1 / decimal odds) < model-derived true probability. Professional bettors only place bets with confirmed value, ignoring the rest. Compounding +EV bets over time produces consistent profit despite individual losses.
Formula:Implied Probability = 1 / Decimal Odds. Value exists when True P > Implied P.
Example:Bookmaker offers 2.50 on a match (40% implied). Our Poisson model says 50%. EV = +25%.
No-Vig Fair Odds (Vigorish-Free)
The "true" probability extracted by removing the bookmaker's margin from posted odds.
Bookmakers add margin (vig) by setting odds whose implied probabilities sum to >100%. Removing the vig proportionally reveals the bookmaker's true price. Comparing your model's probability to no-vig odds is the cleanest value detection method.
Formula:Fair P_i = Implied P_i / (Sum of all Implied P)
Example:Odds 2.0/2.0 imply 50%/50% but real market hold is 4%, so true price for each side = 48% / 48%.
1X2 (Match Result Market)
The three-way outcome market for football matches: Home Win (1), Draw (X), Away Win (2).
The fundamental football betting market with three mutually exclusive outcomes. SpinBonus's Poisson model outputs P(home), P(draw), P(away) which sum to 1.0. Used as the basis for all derived markets (Over/Under, BTTS, Asian Handicap).
Over/Under (Total Goals)
A market predicting whether total goals scored will exceed (Over) or fall short (Under) of a line, typically 2.5.
Over/Under markets typically use 0.5, 1.5, 2.5, 3.5 lines. The 2.5 line is most popular. Probability is derived from Poisson by summing P(0-0), P(1-0), P(0-1), P(1-1), P(2-0), P(0-2) for Under 2.5 and the rest for Over 2.5.
BTTS (Both Teams To Score)
A market predicting whether both teams will score at least one goal in a match.
BTTS pays out if both teams score. Poisson-derived probability = (1 − P(home goals = 0)) × (1 − P(away goals = 0)). Typically 50–55% probability in major leagues; market overprices BTTS:Yes due to public bias toward goals.
Asian Handicap
A goal-handicap betting market that eliminates the draw, refunding stakes on quarter-line pushes.
Asian handicaps remove the draw by giving one team a virtual goal head-start (+0.5, +1.0) or deficit. Quarter handicaps (+0.25, +0.75) split the stake between two adjacent lines. Tighter spreads and lower vigorish than 1X2.
Hold (Bookmaker Margin)
The percentage of total handle a bookmaker keeps as profit, also called overround or vigorish.
A market with odds summing to 105% has 5% hold. Sharper books (Pinnacle, Betfair Exchange) hold 1–3%; recreational books (DraftKings, FanDuel) hold 5–8%. Lower hold means tighter prices and easier value finding.
Formula:Hold = (Sum of Implied Probabilities) − 100%
Example:Odds 1.91/1.91 → 52.4% + 52.4% = 104.8%, so hold = 4.8%.