SportIQ — Match Predictions

Every prediction runs through three layers that are combined into a final result.

Data collection Last 10 matches each team + H2H history fetched from the API

↓

Weighted team profiles Decay weighting (0.9ⁿ) — recent games count more

↓

Poisson model Expected goals → 6×6 score matrix → Win / Draw / Loss %

↓

Rule-based markets Goals, Corners, Cards, BTTS thresholds with confidence scores

↓

ML overlay (optional) Random Forest trained on real historical matches

What data is collected

Last 10 home team matches from the API
Last 10 away team matches
Last 10 head-to-head meetings between the two teams

Data providers (35+ leagues)

Provider	Leagues	Key required
ESPN (unofficial)	11 leagues — PL, La Liga, Bundesliga, Serie A, Ligue 1, MLS, Eredivisie, Primeira Liga, Super Lig, Champions League, Europa League	No
OpenLigaDB	3 German leagues — Bundesliga, 2. Bundesliga, 3. Liga	No
football-data.org	Top European leagues — PL, Bundesliga, Serie A, La Liga, Ligue 1, Eredivisie, Primeira Liga, Brasileirão, CL, EL	Free tier
API-Football.com	Championship, Ligue 2 + extras	Yes
TheSportsDB	MLS, Brasileirão, and others	Free tier

Exponential decay weighting

Matches are sorted oldest → newest. Each gets a weight of 0.9ⁿ, so the most recent game counts the most:

weights = [0.9⁴, 0.9³, 0.9², 0.9¹, 0.9⁰] = [0.656, 0.729, 0.810, 0.900, 1.000]

A goal last week is worth ~52% more than a goal from 5 games ago — the model reacts faster to streaks.

avgGoals = Σ(weight × goals) ÷ Σ(weights)

Daily Picks & Weekly Picks

The Today's Games landing page and Daily/Weekly Picks features automatically scan all 35+ leagues, predict every fixture, and surface the highest-confidence bets across all markets (goals, BTTS, corners, outcome). No manual league selection needed.

Step 1 — Expected goals (λ)

Using a Dixon-Coles / Maher inspired formula:

λ_home = homeAttack × awayDefence × 1.1
λ_away = awayAttack × homeDefence

The 1.1 multiplier captures home advantage (+10%). If venue-specific stats are available it drops to 1.03 since the advantage is already in the split averages.

Step 2 — Poisson formula

Probability of scoring exactly k goals given expected rate λ:

P(X = k) = e^−λ × λ^k / k!

Example — λ = 1.8 expected goals:

P(0 goals) = e⁻¹·⁸ × 1.8⁰ / 0! = 16.5% P(1 goal) = e⁻¹·⁸ × 1.8¹ / 1! = 29.8% P(2 goals) = e⁻¹·⁸ × 1.8² / 2! = 26.8% P(3 goals) = e⁻¹·⁸ × 1.8³ / 3! = 16.1%

Step 3 — Score matrix (6×6)

Home and away goals are treated as independent, so every scoreline probability is:

P(2–1) = P(home = 2) × P(away = 1)

This creates a 6×6 grid. The cell with the highest value = most likely score.

Step 4 — Sum the matrix

Home Win = Σ P(h,a) where h > a Draw = Σ P(h,a) where h = a (diagonal) Away Win = Σ P(h,a) where h < a

Over / Under 2.5 goals

Under 2.5 = P(total ≤ 2) = P(0) + P(1) + P(2)
Over 2.5 = 1 − Under 2.5

Both Teams to Score (BTTS)

BTTS Yes = P(home ≥ 1) × P(away ≥ 1)
= (1 − P(home=0)) × (1 − P(away=0))

Corners & Cards (rule-based)

Market	Rule	Confidence
Corners O/U 9.5	Over if combined avg > 9.5	min(│diff│ / 4, 1)
Cards O/U 3.5	Over if combined avg > 3.5	min(│diff│ / 2, 1)
Goals O/U 2.5	Over if combined avg > 2.5	min(│diff│ / 1.5, 1)

Confidence = how far from the line. Combined avg 4.5 → confidence 1.0. Combined avg 2.6 → confidence 0.07.

Overall confidence level

< 5 matches for either team → LOW 5 – 8 matches → MEDIUM > 8 matches for both teams → HIGH

An optional Python ML microservice runs alongside the Poisson engine and gives a second opinion on each market.

What it predicts

Over / Under 2.5 goals
Both Teams to Score (BTTS)
Corners Over / Under 9.5

How it's trained

Features fed into the model:

home_avg_goals_for / against away_avg_goals_for / against home_avg_corners away_avg_corners home_form_rating (0–1) away_form_rating (0–1) home_win_rate away_win_rate h2h_avg_total_goals h2h_played poisson_over25_prob poisson_btts_prob

Training data is built from real historical matches (PL, Bundesliga, Serie A, La Liga, Ligue 1 — 4 seasons). The rolling window ensures only pre-match information is used, so there is no data leakage.

Algorithm

Random Forest (200 trees, depth 8) — one classifier per target market. Evaluated with 5-fold cross-validation AUC. XGBoost is also available via the /train endpoint.

To enable

# Terminal 1 — start the ML service cd ml && uvicorn app:app --port 8000 # Terminal 2 — build real training data (first time only) node scripts/build-training-data.js # .env — enable ML predictions ML_SERVICE_URL=http://localhost:8000

Market Predictions Poisson-backed probabilities

⚽

Total Goals Over / Under 2.5

Over 2.5

Under 2.5

🎯

Both Teams Score BTTS

Yes

📐

Corners Over / Under 9.5

Over 9.5

Under 9.5

🟨

Yellow Cards Over / Under 3.5

Over 3.5

Under 3.5

Safe Bets All markets at a glance

Double Chance

1X —

X2 —

12 —

Total Goals

O 0.5

—

O 1.5

—

O 2.5

—

Team Goals

Home

0.5+

—

1.5+

—

Away

0.5+

—

1.5+

—

Match Outcome Poisson model

Most Likely Score

xG vs xG

Top 5 Correct Scores

Home Win

Draw

Away Win

Team Stats Comparison

Stat (per game)	Home	Away

Recent Form (last 5)

Head to Head

Date	Home	Score	Away	Result

🗂 Last 10 Matches

Date	H/A	Opponent	Score	Result	Corners	Shots

Date	H/A	Opponent	Score	Result	Corners	Shots

📅 Today's Games

Premier League

League Table

⚽ Top Scorers

📊 Bookmaker Odds

Market Predictions Poisson-backed probabilities

Safe Bets All markets at a glance

Match Outcome Poisson model

Team Stats Comparison

Recent Form (last 5)

Head to Head

🚑 Injuries & Suspensions

🗂 Last 10 Matches

📊 Bookmaker Odds

Predictions

Statistical Model

Recent Form (last 5)

Head to Head

🎟 Tickets