Finding a weekly slot for your D&D group

Campaigns almost never end in a TPK. They end after three cancelled weeks in a row: the group chat goes quiet, the DM stops prepping, and the next message anyone sends is asking whether people still want to play.

Session zero sorts out the tone, the safety tools and the house rule on critical hits. It does not sort out scheduling, because scheduling is not a decision you take once. And the hard part is not finding one free evening: any five people can find one free evening. It is finding the same evening, repeatedly, in a group where a single absence cancels the night.

That changes the question you should be asking. Not “when are you free?”, which produces five different answers and no overlap, but “which of these evenings can you hold?”, over a block of weeks, with every answer landing in one place.

Decide this before you send anything out, because it changes what the results table means. If a session needs all five players, you are looking for a row where everybody said yes. If you plus three is a session, you are looking for a row with three yeses, and you can stop waiting for the fifth answer.

Whichever you pick, put it in the poll title: “D&D, Thursday evenings, we play at four” answers the question before anyone votes.

A weekday session does not start “on Thursday”. It starts at 19:30, and it ends when the first person has work in the morning. A poll made of whole days hides both ends of that, and you find out on the night itself that one player can only turn up after nine.

So use the “Time slot” type, which asks about precise hours instead of days: Thursday 19:30-23:00, Tuesday 20:00-23:30. Two slots on the same day are fine when the group is split between an early start and a late one; the table reads the same way either way.

An end time is not pedantry. It is what tells the DM how much material to prepare, and whether tonight’s dungeon is meant to finish at the boss or at its front door.

1. 1

   Propose four to six evenings, not one

   A single date tells you whether this Thursday works. A block of four to six tells you whether Thursdays work at all, which is the thing you actually need to know.

2. 2

   Stay inside two months

   Hardly anyone knows their calendar further out than that. A yes given for twelve weeks away is a guess, and it is the one that gets cancelled.

3. 3

   Say when you will read the answers

   “I’ll pick the evenings on Sunday night.” There is no closing date in Pollendar, so the deadline is the one you announce; without it, the last two players never answer at all.

4. 4

   Poll again when the block runs out

   One poll per block beats one poll with thirty evenings in it: the same logic as any recurring meeting, where the long list is what stops people answering.

You can also add dates to a poll that already exists rather than creating a new one. The link stays valid, voting reopens on the new rows, and nobody has to dig the old message out of the channel. That is the cheap way to extend a block by a fortnight when the first four evenings produced nothing usable.

Each player answers one of three things: I am available, I might be available, I’m not available. The middle one is the most useful answer in the list and the most abused, because the table cannot tell these three apart:

• “Probably yes, I just cannot promise.” The honest maybe. It means prepare the evening, and I will be there unless something breaks.
• “I could, but I would rather not.” A polite no. It costs the group two days of waiting and then a cancellation on the day.
• “I haven’t looked at my calendar.” Not an answer at all. Ask the person to look, and the amber turns into something you can count.

Agree out loud, once, on which of those your group means. It is the cheapest improvement you can make to a game night that keeps collapsing.

The evenings stay in calendar order, never sorted by score, so you keep your bearings. It is the highlight that points. Green marks an evening every voter said yes to. Amber marks the next best one after that, or the leading one when nobody reached unanimity, which, for five people on weekday evenings, is most weeks. Rows rank by most yeses, then fewest nos, and an evening nobody said yes to is never highlighted at all.

That amber row is not a consolation prize. With the rule you set before polling (the DM plus three), it is the session you are actually playing, and you can announce it on the spot instead of waiting two more days for a fifth answer that may never come.

Once it is announced, Add to calendar sits on the highlighted row and offers iCalendar (.ics), Google Calendar or Outlook Web. If you asked participants for their email when you created the poll, the calendar event can invite the players who said yes, and an entry in someone’s own calendar does more for attendance than a message that scrolled past yesterday.

Every group has one. They want to play, they open the link, and they tick nothing: every row then stays one answer short and five people wait on one. Three things work, in this order.

• Ask once, by name. A direct message beats a reminder dropped into a channel of fifteen people. Pollendar sends no automatic chase, which is just as well: the nudge that works is the one with a name on it.
• Send the invitation from the poll. You can email the link to participants directly, twenty addresses per send. A link in an inbox outlives a link in a chat thread, where it disappears in an afternoon.
• Decide without them, and say so. “Thursday the twelfth, four of us confirmed, you are welcome.” If they cannot make it, that is information too, and it arrives before the evening, not during it.

If somebody has genuinely left the campaign, remove them from the table rather than reading around them. A column of empty circles makes every single row look worse than it is, and it will talk you out of evenings that were fine.