Step 1
Define what balanced means for this activity
The default team workflow balances headcount: when the participant total is not divisible by the team count, group sizes differ by at most one. That is measurable and transparent, but it does not automatically balance experience, language, confidence, mobility, or access needs.
Choose the number of teams when room layout or equipment fixes the number of stations. Choose a target team size when collaboration quality is the main concern, then review the resulting team count before generating.
Step 2
Verify every participant as a distinct row
Compare the input count with the real attendance count and keep repeated names as separate people. Add a stable suffix such as Maya A and Maya B rather than deleting one as a duplicate. Random placement cannot include a person whose row is missing.
Remove headers, notes, and empty placeholders that are not participants. If a support adult or facilitator should not be assigned, keep that person outside the randomized roster rather than deleting them from a generated result later.
Check before continuing
- Every present participant has one row.
- Duplicate names are distinguishable.
- Non-participant notes are removed.
- The displayed eligible count matches attendance.
Step 3
Use constraints for known requirements, not preferred outcomes
Fixed members can anchor a facilitator, interpreter, equipment lead, or other declared role. Keep-together and keep-apart pairs should represent a known operational or safeguarding requirement. Too many constraints can make a valid arrangement impossible, so the generator must report failure rather than quietly ignore one.
Write constraints with unambiguous participant identifiers and review them aloud or with the responsible organizer. A rule that is introduced only after seeing the first groups is a manual change, not part of the original random process.
Step 4
Generate once under the published rule
Run the generator after the roster, team count, and constraints are settled. The secure random shuffle establishes a candidate order, and the placement logic distributes participants while satisfying the supported rules. If the tool reports an impossible configuration, revise the constraints openly instead of repeatedly drawing until a preferred result appears.
When groups will be announced publicly, show the final complete set rather than revealing only selected groups. This makes missing names, duplicates, and uneven sizes easier to notice.
Step 5
Check inclusion before making a transparent adjustment
First confirm that every input identity appears exactly once, no unexpected person appears, size differences match the stated rule, and every declared constraint is satisfied. These are correctness checks, not optional polish.
A teacher or organizer may still need to adjust a group for information the browser does not have. State that the final arrangement was manually adjusted and give the practical reason when appropriate. Do not describe the edited result as untouched random assignment.
Run it in this order
- Match output count to input count.
- Check each participant exactly once.
- Verify group sizes and constraints.
- Record any human adjustment separately.
Step 6
Export only the information the group needs
Use a text, CSV, PNG, or print result according to the destination. CSV cells must be neutralized when a participant label begins like a spreadsheet formula. Review the downloaded file before distributing it, especially when names include commas or non-Latin characters.
Keep rosters and exports on approved devices. Random Wheel Tools processes names in the browser, but a downloaded file, printed page, screen share, or copied message is under the organizer's control and may expose the group list.
Use the right surface
Tools mentioned in this guide
Common questions
Questions about fair random groups
Does balanced mean every group has the same size?
Only when the participant total divides evenly. Otherwise a balanced-headcount result differs by at most one person between the largest and smallest groups.
Can random groups also be balanced by skill?
Not without accurate role or skill information and an explicit policy. A purely random roster balances chance and headcount, not hidden participant characteristics.
What if the constraints cannot all be satisfied?
The tool should reject the configuration visibly. Remove or revise a rule and explain the change; it must not silently drop a participant or ignore a constraint.
Is it acceptable to swap two people after generating?
Yes when an organizer has a legitimate reason, but label it as a manual adjustment. The final arrangement is then constrained and human-edited rather than purely random.
How should I handle an odd participant count for pairs?
Use the pair generator's declared bye or three-person policy. Do not silently omit the unpaired participant.