Aerial Defense in Football

This signal mirrors Aerial Dominance but focuses on the defensive vulnerability side of aerial contests.

• Aerial loss rate (70% weight): Percentage of aerial duels lost. Normalised between 35% (elite, rarely loses contested headers) and 60% (poor, loses more headers than wins). This is the quality component of aerial defence.
• Aerial losses per match (30% weight): Volume of headers lost per match. Normalised between 10 (few losses) and 25 (many losses). Captures how frequently the aerial weakness is exposed.
• Shared data layer: Uses the same data loading and uncertainty calculation as Aerial Dominance, ensuring consistency. The difference is purely in perspective: winning vs losing.
• Data quality filtering: Matches where aerials won equals zero are excluded as backfill placeholders. Aerial losses are only counted when won data exists in the same match, preventing partial data from skewing results.

Aerial defence follows the same reliability patterns as aerial dominance, with additional context about when vulnerability matters most.

• Best with 10+ matches: Uncertainty drops to 0.40 at this sample size. Aerial loss patterns stabilise well over 10+ matches as random variations smooth out.
• Most impactful against aerial opponents: Aerial defence vulnerability is most relevant when the opponent uses crosses, long balls, or set pieces frequently. Against a short-passing possession team, aerial weakness matters less.
• Less reliable early season: With only 5-7 matches, a single set-piece-heavy game can distort the loss rate. The elevated uncertainty (0.65) reflects this sample-size risk.
• Pairs well with: Aerial Dominance (the opponent’s aerial attacking strength), Set Piece Threat (whether the opponent can exploit aerial weakness from dead balls), and Defensive Reliability (whether aerial losses translate into goals).