How to use the student group generator
- Paste the names, one per line (commas also work). Duplicate names are refused, so add an initial to tell two people apart.
- Enter the largest group size and the number of rounds.
- Read the groups for each round, the count of repeated pairings and the upper bound on repeat-free rounds.
- Change the seed under Advanced options for a different draw. The same seed always gives the same groups.
Formula
groups per round = ceil(people / largest group size)
group sizes differ by at most one, larger groups first
repeat pairings = pairs in the same group who were together in an earlier round
The page starts each round with a seeded random split, then swaps people between groups of the same round whenever a swap does not increase the number of repeat meetings. A pair that meets three times costs more than one that meets twice. The counts are recomputed from the final groups.
Twelve people in groups of three for three rounds
Twelve names in groups of at most three make four groups of three each round. The search finds three rounds in which nobody meets the same person twice, so the repeated pairings are 0. Each person meets 6 different people across the three rounds out of the 11 others.
Some shapes cannot avoid repeats. Twelve people in groups of four can be split once without repeats, but the second round must put four people into groups using only three earlier groups, so at least two of them were together before. The page reports the repeats it could not avoid.
How many repeat-free rounds are possible
In every round a person meets (group size - 1) other people, and there are only (people - 1) others to meet. So with 12 people in groups of 3, no more than floor(11 / 2) = 5 rounds can be free of repeats, and the page shows this limit. It is an upper bound: some shapes reach it and some do not, because the groups must also fit together.
The search is heuristic. It often finds a perfect answer when one exists, but it does not prove that none exists, so try another seed if the repeats look avoidable.
What it does not consider
The groups are random and balanced for repeats only. They do not consider skills, friendships, language, support needs or anything else a teacher knows. Adjust by hand where that matters. Names stay in your browser. For teams balanced by rating use the Balanced Team Generator, for picking winners the Random Name Picker, and for a classroom layout the Seating Chart Maker.
More planning tools
For exams, carpools and invoices, try the Carpool Rotation Schedule Generator, Time Log to Invoice Lines and Exam Time Budget Calculator.
Assumptions and limits
- Everyone is treated the same; groups are formed by chance with repeats as the only aim.
- Group sizes are as even as possible, with no group larger than the size you set.
- Up to 80 names, 20 rounds and a group size of 40 are accepted.
- The search is seeded and repeatable but is not guaranteed to find the best possible schedule.
- The example names are an illustration. Replace them with your own.
Frequently asked questions
Why do some groups have fewer people?
If the number of people does not divide evenly, the groups are balanced so sizes differ by at most one. Ten people in groups of four become groups of 4, 3 and 3.
Is it truly random?
It is pseudo-random from the seed you give, so the same seed repeats the same result. Change the seed for a new draw. It is meant for classroom groups, not for anything where unpredictability is a safety matter.
What does a repeated pairing count?
Each pair of people who share a group after already sharing one earlier counts once per extra meeting. Three people meeting again in the same group count as three pairs.
Can I keep two people apart or together?
Not directly. This page only avoids repeat pairings, so adjust by hand or run it again with another seed.