How to Run Game Sign-ups in a Group Chat With 10+ Players

The tenth player is where it stops working

"Who's playing Saturday?" A pinned message, a wall of replies — "+1", "in", "+2, me and a friend", "out, can't make it" — and somewhere around the tenth or eleventh reply you stop being able to hold the count in your head. So you start over from the top: scrolling, tallying on your fingers, matching whoever dropped mid-thread against whoever joined late. Under ten players this is mildly annoying. Over ten it's a recurring job, and it comes back every week the team plays.

That isn't a hypothetical. In one team's chat logged from April 6 to July 15, 2026, there were 145 bare "+"/"-" replies across the period, and the median roster for a game was about 16 people. At that size a thread of pluses and minuses stops being a list and becomes something you reconstruct from scratch, every single time.

Four ways teams keep the list by hand — and where each one breaks

Almost every pickup team lands on one of these. All four work fine at six players. Here's where each gives out.

The pinned message the organizer edits. One numbered list, kept current by hand. The most readable option and the most expensive one: every "+1" and every drop-out is an edit, and the edits arrive while you're at work. Miss two of them and the pinned list quietly stops matching reality — which is worse than having no list at all, because people trust it.

The plus/minus thread. Zero effort to write, all the effort at read time. Nobody can answer "how many are we?" without scrolling the whole thread and doing the arithmetic, so in practice one person does it and everyone else asks that person.

A poll. Good for a yes/no headcount, weak for everything else: a poll can't take "+2, me and a friend" as two people with names, can't hold a waiting list once the game fills up, and can't tell you which of the two guys named Max actually voted.

A spreadsheet. Correct, auditable, and the only one of the four that survives an argument about who owes what. The cost is that it lives outside the chat: someone has to copy sign-ups into it by hand, and most players will not open a link to add themselves. That friction is exactly what the group chat existed to avoid.

What breaks isn't the "+", it's the recount

The plus itself is a perfectly good way to say "I'm in". What stops scaling is the counting around it.

• A guest isn't a number. "+2, me and a friend" eventually needs a name attached to the friend, or nobody knows who's actually showing up. Under ten people everyone already knows who "the friend" is. Over ten, they don't.
• Drop-outs land mid-thread. Someone writes "+1" on Tuesday and "out, can't make it" on Thursday. Whoever counts by scrolling has to reconcile those two in order, and not miss the second one.
• People stop counting and start asking. Once the thread is too long to read end to end, "who's on the list?" becomes one of the most common messages in the chat. In that same corpus, one question shows up in at least fifteen different phrasings — "roster?", "list?", "who's in?", "what's the final lineup", "give me the roster for the game" — simply because reading forty replies to answer it yourself stopped being realistic.
• People stop trusting the tally and start checking themselves in it. Under a hand-kept column it's routine to see "wait, am I even on here?", "Alex, did you actually add me?", "is Chris on the list?", "didn't he drop out?" Not because anyone made a mistake — because a hand-kept list can't be verified except by asking the person who keeps it.

None of this is a knock on any particular organizer. It's what happens to any manually maintained list once it crosses the point where a person can hold it in working memory, which for most of us is somewhere between ten and fifteen items.

Registering someone else is the real bottleneck

Here's the detail that's easy to miss: most of the traffic on a real roster isn't people signing themselves up. In the same corpus, registering someone else was the most frequent assistant action of all — 125 calls, ahead of every other command the team used. "+2, me and Max" isn't one sign-up, it's two, and one of them belongs to a person who never typed a word in the chat. The second most common input form isn't a "+" at all: it's someone pasting the whole finished lineup as a block of text, sometimes two people pasting competing versions.

A keyword parser that counts "+1" and "-1" can't handle either case. It has no way to know that "add Max too, he's coming" means registering one specific extra person, named mid-sentence. The TeamUp assistant reads the message instead of pattern-matching it: "+1 for Max" registers Max specifically, gives him his own line and his own number, and — this is the part that matters for teams with regulars — searches the chat for an existing Max before creating a new one, asking you which one you meant only when it genuinely can't tell. Two people with the same first name in one roster isn't an exotic case; it happened twice in four months in that one chat.

What holds a 10+ roster together

A demo: in a real chat the assistant's reply is written by the model in the chat's own language, so the wording will differ.

The mechanism is deliberately unglamorous: people keep writing in the chat exactly the way they already do. "+1", "in", "add my brother", "take me off" — none of that changes. What changes is what happens after someone types it.

• Every registration gets a badge number at the moment it's made, and that number doesn't shift when other people join or drop. "#7" means the same person on Tuesday and on Saturday.
• A guest gets their own line, separate from whoever brought them, instead of vanishing into an unlabeled "+2" nobody can unpack later. If they went in under a placeholder name, renaming them later keeps both their number and their spot.
• "Who's on the list?" is answered from the current state, not from someone's memory of the thread. Ask it however you'd ask a person; what comes back is the roster as it stands.
• If the organizer set a headcount limit for the game, sign-ups past that number go to a waiting list automatically instead of piling on top of a full game, and the organizer can promote someone from it when a spot opens.
• A team that already has a finished lineup — pasted from last week, or grouped under team names — can hand it over as one block instead of one name at a time. It registers as an ordered list, badge numbers following the order given. Nobody gets removed silently: anyone currently registered who isn't in the pasted list is shown to you, and dropped only if you confirm.

From inside the chat the effect is that a roster of twenty behaves like a roster of six. Ask, get an answer, done.

Honest limits

Worth saying plainly, because a tool that quietly fails at the edges is worse than one that's upfront about them.

None of this replaces announcing the game — an event has to exist before anyone can sign up for it. A waiting list only exists if a headcount limit was set; without one, sign-ups are simply open, however many people show up. Splitting a confirmed roster into teams is a separate step with its own trade-offs, and it isn't a strength rating — teams are assigned randomly or by hand, with the organizer moving individual players around afterwards.

And if your team's chat isn't on Telegram or VK at all — a WhatsApp group, say — a roster can still run through a public sign-up link instead of a bot in the chat. But that link asks people to log in with Telegram, Apple or VK before they can sign up. "No install" is true; "no account at all" isn't.

The bottom line

The plus/minus habit isn't the problem — it's the recount underneath it that doesn't survive a long list. Past ten players that's not an edge case, it's the normal week. Whoever keeps the roster — a person, a spreadsheet, or an assistant reading the same chat — the test is the same: does "who's on the list?" get answered from a fact, or from a scroll? Every other surface is really just the group chat wearing different clothes, and if a bot is already in yours, the setup takes about a minute.