This generator creates preliminary quiz brackets among 3-team quizzes distributed among a certain number of rooms over a certain number of quizzes.

Teams

Quizzes

Rooms

Keep re-solving for better scores.

See the solver log in the Developer Window (F12 in Chrome)

This matrix shows the number of times a team (column) quizzes in a particular room (row). Ideally, teams will be balanced across all rooms. Cells are flagged red if they visit more than once. Visiting more than once is OK, this just highlights something to watch for.

Key Points: Each COLUMN (i.e. a team) should have nearly the same number of colored cells (i.e. teams should not visit a particular room more than anyone else. Also, each ROW should have about the same distribution of visits (i.e. about the same number of 0,1,2, etc as other rooms). Gray (i.e. phantom rooms) must be equal for all teams.

This matrix shows the number of times a team meets another team. Flagged if meets more than once.

Key Point: A good configuration will have this matrix consisting of a uniform low number, with zeros along the diagonal (team1 meeting itself doesn't count). For instance, a matrix of mostly ones means that every team meets every other team once. Teams that meet each other more than once are flagged red.

This matrix shows the number of times a team is on a bench. Flagged if meets more than once.

Key Point: A good configuration will have each team on different benchs the same number of times.

- Can I do better?
- Possibly! The solver uses Simulated annealing to optimize the configuration. Think of it as slow cooling down a system to get a large, perfect crystal. However, because there are random fluctations as its cooling down, it's possible that re-solving or adjusting the solver parameters will give a better solution (i.e. lower score).
- Colors
- Rooms are colored groups of teams (max of 3). If you don't have enough rooms (i.e. not all teams can quiz every round), teams are shuffled into gray "phantom" rooms.
- Assignment Scores
- The algorithm scores the team-assignments trying to minimize multiple match-ups with other teams and multiple visits to the same quiz room. For rooms, a configuration is penalized by room_penalty*(roomVisits-1)^2. For teams, a configuration is penalized by team_penalty*(teamMatchups-averageMatchups)^2. A lower score is better.