5-Factor Streak Algorithm

1. Regression to Mean (40% weight)

Statistical principle: longer streaks are more likely to break. The algorithm assigns continuation scores based on streak length:

• ≤5 games: 70% continuation score
• 6-7 games: 65% continuation score
• 8-10 games: 55% continuation score
• 11-12 games: 45% continuation score
• 13-15 games: 35% continuation score
• 16-20 games: 25% continuation score
• 20+ games: 15% continuation score

2. Opponent Strength (25% weight)

Based on the opponent's current league position. The algorithm calculates a normalized position score:

Formula: (1 - normalizedPosition) × 100

Example: Facing the 1st place team (out of 20) = 95 score (very hard opponent). Facing the 20th place team = 0 score (easy opponent).

3. Venue Context (15% weight)

Considers whether the streak was built at home or away, and where the next match is played:

• Home streak playing at home: 85 score (boost)
• Home streak playing away: 20 score (penalty)
• Away streak playing away: 85 score (boost)
• Away streak playing home: 20 score (penalty)
• "All" scope: 60 score (neutral)

4. Head-to-Head History (10% weight)

Analyzes the last 5 meetings between these two teams and calculates how often the streak type was maintained:

For example, if a team is on a win streak and won 4 of the last 5 meetings vs this opponent, the H2H score would be 80% (4/5 matches).

The algorithm checks the specific streak type (win, BTTS, over 2.5 goals, etc.) in those historical matches.

5. Recent Form (10% weight)

Evaluates the team's performance in their last 5 finished matches:

• Win = 3 points
• Draw = 1 point
• Loss = 0 points

Maximum score = 15 points (5 wins). Score = (actual points / 15) × 100