A fixed opener changes the sample space
The named first speaker is assigned position 1 by rule. If m other people are entered, only their m positions are randomized, producing m! equally likely tail orders. The model assumes the opener was chosen before the shuffle and that each later name appears once. It does not allocate speaking time, decide attendance, or judge whose topic should be earlier.
Worked agenda: Chair plus four speakers
Enter Chair as the fixed opener and Avery, Blake, Casey, and Dev as the remaining roster. There are 4! = 24 possible tails. If the shuffle returns Casey, Avery, Dev, Blake, the displayed agenda is Chair → Casey → Avery → Dev → Blake. Avery has probability 1/4 of each later position—second, third, fourth, or fifth—and probability zero of first. Casey did not receive a preference; Casey occupied one of four symmetric tail slots.
Check a completed agenda from the visible names
Count one fixed name plus m later names, then compare the tail with the entered later roster. Each should occur exactly once. Sorting the two lists alphabetically is a quick coverage check, while the first displayed name verifies the procedural exception. The form rejects a name that appears both as the opener and in the tail. When two attendees share a name, use unambiguous labels such as Sam R. and Sam T.; otherwise the result cannot be audited after copying.
Absence and interruption need an announced policy
If a person leaves before their turn, decide before the draw whether they are skipped, moved to the end, or cause a new tail shuffle. These choices are not mathematically interchangeable. If a meeting ends after the third presentation, preserving the existing remaining order keeps the original allocation; reshuffling creates a second chance for an earlier slot. Timing limits, breaks, accessibility accommodations, and substitutions belong in the agenda policy, separate from the permutation.
Where this procedure is appropriate
A moderator may need to open a meeting, a host may need to give a safety notice, or a previous winner may have an agreed right to begin. In each case the reason for position 1 should be visible. When every participant must have the same chance of every position, leave the opener field out of the procedure and shuffle the whole roster instead. Random order is not a calendar: it gives no dates, durations, or guarantee that a person will be present when called.
Keep a useful record
For a group process, retain the fixed-first reason, the complete roster, the resulting sequence, and the absence rule. Those facts distinguish a transparent procedural exception from a hidden change after seeing who went late. The browser produces the order locally and does not create a permanent meeting record.
Position probabilities can be checked one person at a time
With Chair fixed and four names in the tail, Avery has four possible positions: second through fifth. Exactly 3! = 6 of the 24 tail permutations place Avery second, so the chance is 6/24 = 1/4. The same count applies to third, fourth, and fifth. That symmetry is the relevant fairness property here. It does not say Avery and the Chair receive equal treatment, because the Chair is outside the randomized positions by an explicit rule.
Use a separate mechanism for equal time
Order and duration are different variables. A six-person agenda with five-minute slots takes 30 minutes before questions; a six-person agenda with open-ended presentations has no predictable finish merely because its names were shuffled. If later speakers need equal preparation or a hard closing time, publish time limits, a question rule, and a continuation policy alongside the sequence. Randomising names cannot solve the scheduling problem created by unequal topic length or a participant who arrives late.
Substitutions create a new roster question
Imagine Blake is replaced by Blair after the tail has been made. Replacing the label in Blake’s existing position preserves the allocation of positions among the original attendance slots. Adding Blair and reshuffling all four tail names is a new draw. Moving Blair to the end is a third policy. None is universally correct, but calling them the same “random order” conceals a consequential choice. A chair can avoid that dispute by naming the substitute rule before the roster is final.
Make the output usable in the room
Copy the complete arrow-separated result where everyone can see it, then mark each completed turn rather than deleting the remaining names. The untouched tail proves which positions were allocated at the start. Use role labels when identities could collide: “Sam Lee — finance” and “Sam Lee — operations” are auditably different where two bare “Sam” entries are not. This local tool is appropriate for a meeting ritual, not for queues governed by statutory priority, accessibility rights, or emergency response.
Record a late change as a procedure change
If the host must insert a safety announcement after the shuffle, write that insertion separately rather than relabelling the randomized positions. The distinction preserves what was random and what was necessary administration. It also prevents an ordinary interruption from being misdescribed as someone losing their allotted place.