How the 82-Game Season Simulation Works
The season is not a live sports forecast and it is not a black-box score. Build A Player converts the ratings you selected into a bounded, deterministic entertainment model. The same mode, picks, team, seed, and simulation version reproduce the same output. This page publishes the important relationships so players can understand why a spectacular strength sometimes loses to a more balanced build.
From selected traits to overall
The simulator starts with every selected trait rating. It calculates their average, minimum, and maximum. Overall is the rounded result of 88 percent of the average, 8 percent of the maximum, a balance adjustment equal to 12 percent of the minimum rating above seventy, and seeded variation between roughly minus one and plus one. The final value is bounded from 68 to 99.
This weighting explains the central strategy. The average dominates. The best trait provides a smaller ceiling reward. The lowest trait affects the balance adjustment, so an extreme weakness can undo some of the benefit created by a perfect strength. Randomness is deliberately narrow: it adds texture without allowing a poor build to become elite by luck alone.
Seeds and reproducibility
A standard run creates a random identifier at the beginning. Daily runs derive a stable seed from the date and mode, unless the database supplies a configured seed. Candidate boards are shuffled from that seed, the round number, reroll count, and used player IDs. Team options use another deterministic shuffle.
The simulation combines the seed, mode, selected team code, and a version label before generating its bounded random values. Shared links encode the choices needed to reproduce the run. This does not mean two different builds under the same daily board receive identical results; the picks, team fit, and resulting payload distinguish the final simulation.
Wins, awards, and playoffs
Hooper wins begin near 34, then add about 1.5 wins for each overall point above 75, the balance adjustment, and up to five seeded wins. The range is capped between 24 and 78. Bucket begins near 32, adds about 1.45 wins per overall point above 75, the balance adjustment, and up to six seeded wins, with a 25-to-76 range.
Playoff labels follow win thresholds: below forty misses the playoffs; progressively stronger records reach the first round, semifinals, conference finals, and Finals. At the highest level, a deterministic branch decides between a Finals loss and championship. Hooper awards depend on combinations of overall and wins. Bucket can award a scoring title when projected PPG reaches 32, or an all-league result for a sufficiently strong winning scorer.
How the stat lines are produced
Hooper scoring is mainly related to overall. Rebounds and blocks directly use paint impact; assists directly use playmaking; steals directly use perimeter defense. Each line adds a small seeded component and remains inside a sensible game range. The formulas create comparisons inside this project, not predictions for a real athlete in a real team context.
Bucket points, field-goal percentage, three-point percentage, career high, and count of fifty-point games are driven largely by overall with bounded deterministic variation. Individual scoring traits shape those outputs through the seven-rating average and floor. The current model does not expose usage rate, shot-location attempts, turnovers, or teammate-level possessions.
What team fit and archetypes mean
Team fit participates in the deterministic simulation seed. It can change the bounded season outcome while preserving the drafted ratings. The current version does not simulate a real roster, coach, salary cap, or depth chart, so team choice should be understood as an environment selector rather than an official matchup model.
Archetypes are chosen from the highest drafted trait. Defensive best traits create a Two-Way Anchor; handles or playmaking can create an Offensive Conductor; athleticism can create a Positionless Force. Bucket recognizes three-point, finishing, and mid-range leaders before using Isolation Artist as its remaining branch. These labels describe the build and do not add a second hidden score.
A controlled experiment for your next run
Save a completed share link, then change only one selected trait in a comparable new run. Record the old and new averages, minimums, maximums, overall ratings, wins, and stat lines. Even when the boards cannot be perfectly identical, limiting the change makes the result easier to interpret. Large multi-pick revisions may improve the score, but they do not reveal which decision produced the improvement.