How Normalization Works

When multiple evaluators score the same player, their scores may differ. One evaluator might be a tough grader. Another might be generous. This does not mean one is right and the other is wrong. It means they have different baselines.

RosterForge normalizes for this automatically. Here is how it works in plain terms.

Step 1: Weighted average. If three evaluators score a player on shooting, the system computes a weighted average of their scores. Categories can be weighted differently (shooting might count more than hustle) based on what you configure.

Step 2: Per-evaluator adjustment. The system looks at each evaluator's scoring distribution across all the players they scored. If evaluator A consistently scores 2 points lower than the group average, the system adjusts for that tendency. This is inter-evaluator normalization.

Step 3: Composite score. Each player gets a single composite score that blends all their category scores, weighted by category importance.

Step 4: Division normalization. The composite score is converted to a z-score within the player's division. This tells you how far above or below average the player is relative to their peers. A z-score of 1.0 means the player is one standard deviation above the division mean.

What this means for you. You do not need to worry about evaluator bias. You do not need to train evaluators to grade on a consistent scale. The system handles it. Focus on getting evaluators who know the sport and can watch a player. The math takes care of the rest.