1. Start with the football projection
GIQ estimates expected fantasy production from a player's workload, efficiency, scoring opportunity, receiving or rushing contribution, team context and expected games. The model also keeps a likely floor and ceiling around the central projection so two similarly projected players can still have very different risk profiles.
2. Evaluate the profile, not only the point total
The ranking system uses a broader set of signals than projected points alone. The current framework includes forward projection, sustainable projection, production/output, role/volume, role stability/security, efficiency, passing quality, rushing value, team environment, health/availability, age/timeline and draft-capital insulation. Not every category matters equally to every position or format.
3. Translate value by league type
Redraft is dominated by the coming season. Dynasty gives more weight to multi-year durability, age curves and long-term role insulation while reducing the tendency for one season of positional scarcity to overwhelm asset value. Keeper is intentionally closer to Redraft, with a smaller future-value component because generic keeper rankings do not know each manager's specific keeper cost.
4. Translate value by scoring
PPR, Half PPR and Standard are not just three labels on one board. Reception scoring changes the value of catch volume. Four- versus six-point passing touchdowns changes quarterback scoring directly. League size changes replacement-level scarcity, which can alter how much value a scarce position adds relative to an alternative at another position.
5. Market is a QA alarm, not a score
GIQ can compare its output to external rankings after the board is built. A large disagreement is a reason to investigate the player's role, projection, availability and model treatment. It is not a reason to move the player automatically. This keeps the board independent while still using outside disagreement as a way to find possible mistakes.
6. Personalization stays separate from True GIQ
Quick Pick learns how a user breaks close ties. After at least 20 decisions it can build a personal board with movement guardrails. That personal board never rewrites True GIQ. The distinction matters: the model's opinion and the user's preference are both useful, but they are not the same thing.
What GIQ does not claim
GIQ does not claim that a ranking is certain, that a floor or ceiling is a confidence interval, or that one model can eliminate football uncertainty. The purpose is to make the assumptions visible and consistent enough that a user can understand why the board changes when the league changes.