Draft Value is how many points a player is worth above a freely available replacement at his own position — the best player you could get for nothing if you skipped the position entirely. Higher is better. It is what orders the overall board, and it is why an elite tight end can outrank a quarterback who scores more raw points: the drop-off behind him is steeper.

Model, Expert and Blend are three separate ranks. Model is what the projection model says on its own. Expert is MetricMind's own ranking. Blend is the published order, combining the two. The ▲ and ▼ markers show how far the blend moved a player against the model alone — a large green number means we disagree with our own model and moved him up.

Floor and Ceiling are the middle half of ten thousand simulated seasons. A player finishes inside that range about half the time. It is deliberately not a best case and worst case: a 90% range is accurate and useless, running 200+ points wide for a mid-round back.

Scoring formats. The three formats differ only in what a catch is worth — one point, half a point, or nothing. Switching recalculates every projection and re-ranks the board, which is why running backs climb in standard leagues and pass-catchers fall. Model and Expert ranks are full-PPR rankings and do not change.

Overall rank is scarcity-adjusted across every position, so inside a position tab the numbers skip — the missing ranks belong to players at other positions.