MÉTODO
65% average + 35% bottleneck
Each role uses a visible weight matrix. A negative-power mean penalizes weak links that a plain MZLive-style weighted average can hide. Set plays remain a separate specialist score.
0 → HERO LAB · CAIXA DE FERRAMENTAS DO MANAGER
Resposta direta: compare a player across seven roles with transparent contributions, a weak-link penalty and a youth-first next training lane.
MÉTODO
Each role uses a visible weight matrix. A negative-power mean penalizes weak links that a plain MZLive-style weighted average can hide. Set plays remain a separate specialist score.
EVIDENCE
The current club dataset has thousands of player-match rows but too few independent players per exact role to learn stable engine weights. Future promotion requires time-blocked and leave-one-player-out validation.