Football Statistics Glossary
Everything you need to understand football statistics, our mathematical prediction models and Game State engines.
Statistics Glossary
Definitions of key terms used across all MatchPulse statistics pages.
BTTS — Both Teams To Score
→ StatsMarket where both teams must score at least one goal. BTTS Yes = both scored; BTTS No = at least one kept a clean sheet.
BTTS No — Neither Both Teams Score
→ StatsAt least one team fails to score. Teams with low attacking averages and solid defences are targeted.
Over 2.5 Goals
→ StatsWins if the total number of goals in the match is 3 or more. The most popular goals market globally.
Over 1.5 Goals
→ StatsWins if 2 or more goals are scored. Lower threshold than Over 2.5, typically priced between 1.05 and 1.35.
CS — Clean Sheet
→ StatsA team keeps a clean sheet when they concede zero goals. High CS rates indicate strong defensive sides.
WBH — Won Both Halves
→ StatsA team wins both the first half and the second half. Rare outcome typically associated with dominant home sides.
PPG — Points Per Game
Average league points per match: (Wins × 3 + Draws) ÷ Matches Played. Never derived from standings.points to avoid phase-accumulated values.
GF / GA / GD
Goals For (scored), Goals Against (conceded), and Goal Difference (GF minus GA). Core metrics for offensive and defensive output.
SBH — Scored Both Halves
A team scores at least one goal in the first half AND at least one in the second half of the same match.
CFB — Come From Behind
→ StatsA team losing at half time goes on to draw or win by full time. Measured as recovery rate against total trailing games.
WtN — Win to Nil
→ StatsA team wins the match while conceding zero goals. Combines a win with a clean sheet in a single market.
HT / FT
HT = Half Time (after 45 minutes). FT = Full Time (after 90 minutes). Stats may be split per half to separate patterns.
Prediction Models
MatchPulse runs seven independent mathematical models for every match. All models are shown side by side so you can compare and identify consensus signals.
Poisson Model
Uses each team's average goals scored and conceded to model the probability of every possible scoreline. The foundation of most football prediction systems.
Dixon-Coles Model
Extends Poisson with a correction factor (ρ) for low-scoring results (0-0, 1-0, 0-1, 1-1), which basic Poisson tends to underestimate.
Weighted Form
Applies exponential decay to recent matches, giving more weight to the last 5 games (factor 1.5×). Captures teams on a run of form.
Normalized Goals
Adjusts each team's attack and defence as a multiple of the league average. More accurate for comparing teams across leagues.
Shots on Target (xG Proxy)
Uses shots on target × conversion rate as a proxy for expected goals (xG). Useful when official xG data is unavailable.
Elo Rating
Dynamic rating updated after every match (K=32, home advantage 65 points). Predicts win probability based on the rating gap.
Combined Ensemble
Simple average of all active models. Tends to be the most stable predictor when individual models disagree.
All models are visible on every match page. Consensus across all 7 models is a stronger signal than any single model alone.
Game State EnginesPRO
Six real-time and pre-match analytical engines, exclusive to Pro subscribers. Based on conditional probabilities from 675,000+ historical matches.
Over 2.5 Sniper — Storm Score
Calculates a Storm Score (0-100) using the harmonic mean of Spark (% of matches with a HT goal) and Fuel (Over 2.5 rate when there was a HT goal).
Late Goals Predictor
Analyses the 46-60 minute window given the current HT score. Returns conditional Over 2.5 FT probabilities based on whether a 2nd-half goal has occurred.
HT Reaction
Shows how teams typically perform in the second half given their specific HT scoreline. Useful for in-play betting at half time.
Final Strike
Identifies matches with the pattern of an early goal, quiet period, then a late goal. Targets teams known for this rhythm.
Red Card Reactor
Monitors live matches with a red card. Calculates an advantage score (0-100) for the team with numerical superiority based on historical outcomes.
Under 2.5 Edge
Finds matches still scoreless past minute 20, then reveals each team's history of staying under 2.5 goals from a cold start.
HT Sniper
Excludes matches where either team has a persistent history of goalless first halves, then ranks the rest by how reliably both sides find the net before half-time.
Lay The Draw HT
Estimates the chance a match will NOT be level at half-time, using each team's HT-draw history shrunk toward the league baseline.
Lay The Draw FT
Estimates the chance a match will NOT end level, using each team's FT-draw history shrunk toward the league baseline. Only shows 80%+ confidence picks.
Lay The Draw Breaker
Live signal reacting to matches tied at half-time, estimating the chance the tie breaks before full-time. Only shows 70%+ confidence picks.
Confidence Tiers
MatchPulse classifies teams and signals into tiers based on signal strength and number of qualifying matches. Applied to leagues, teams and Game State engines.
≥80% rate with ≥12 matches. Strongest signal — very high confidence.
≥70% rate with ≥10 matches. Strong signal — high confidence.
≥60% rate with ≥8 matches. Good signal — moderate-high confidence.
≥50% rate with ≥6 matches. Moderate signal — use alongside other indicators.
≥40% rate with ≥5 matches. Weak signal — treat with caution.
Below threshold with some data. Low confidence — informational only.
Insufficient data to classify. Fewer than 5 qualifying matches.