Method · bah-original-bands-v1

How our game model works

The model keeps historical facts, third-party source metrics, Build A Hooper assessments, and fictional simulation output separate. It is entertainment—not an official model, professional scouting evaluation, betting guidance, coaching advice, or objective player value.

Independent & unofficial. Build A Hooper is not affiliated with, endorsed by, sponsored by, or authorized by the NBA, any NBA team, the NBPA, NBA 2K, Take-Two Interactive, or any player. Historical NBA inputs are used through the 2021–22 season; all ratings, fit labels, archetypes, and season outcomes are original site-created assessments and fictional simulation.

Four explicit layers

Historical facts: pinned Wikidata textual identity and approved structured fields. Missing values stay null and are never guessed.
Historical source metrics: selected 2022 values from a pinned FiveThirtyEight RAPTOR file. These inputs are not displayed as official or current ratings.
Our game assessment: versioned DNA contribution bands, fit, strengths, trade-offs, and archetype labels created by Build A Hooper.
Fictional simulation: scenario choices, deliberately wide fictional season ranges, and outcome labels. It is not a forecast.

Original categorical method

Build A Hooper maps normalized historical inputs into four broad contribution bands: developing, steady, impact, and standout. Ten editorial DNA slots combine selected inputs under method bah-original-bands-v1. A source metric is never presented as a Build A Hooper rating.

Determinism and authority

The server owns the version set, candidate order, state transitions, fit, call effects, and result. Equal pinned inputs and the same engine build yield the same canonical output.

Limits

The 33-player sample is not a complete roster or ranking. Historical RAPTOR stops at the 2021–22 season; Wikidata may be incomplete or disputed; internal position buckets are eligibility choices, not official positions. Qualitative bands deliberately avoid 0–99 ratings, OVR, percentages, and pseudo-precise player comparisons.

Frequently asked questions

Is the model designed to predict real basketball performance?

No. It is a fictional game model built on historical inputs through the 2021–22 season, designed to create understandable trade-offs and repeatable entertainment outcomes. It does not evaluate real players.

Why are results shown as bands and ranges instead of exact numbers?

Bands and ranges communicate uncertainty honestly and avoid implying a precision the inputs do not support. The game deliberately avoids 0–99 ratings, percentages, and pseudo-precise player comparisons.

Can the same inputs reproduce the same result?

Yes. The same ruleset version, board seed, ordered picks, and Big Call choices return the same canonical result. The server owns the version set and the state transitions.

Does the methodology reveal the optimal strategy?

It explains the layers, categories, interactions, and limits. It does not publish server-only weights that would turn the draft into a solved lookup table.