Icebreaker games to play online with someone new (that are not painful)

Culture · by Simone Rainieri · 7 min read

The best icebreaker games to play online with someone new are structured question exchanges, guessing games about each other, and ranked-list debates, because each one hands both people a turn, a topic, and permission to say something real. That last part is the actual purpose. An icebreaker is not entertainment; it is a device for making self-disclosure feel safe faster than unstructured chat ever manages.

And there is unusually good science behind that claim, including one of the most famous experiments in social psychology.

Why structured prompts beat winging it

In 1997, Arthur Aron's lab at Stony Brook paired strangers and gave them 36 questions that escalate gradually, from would you like to be famous, through what you are grateful for, to what you would not want to have to say to anyone. After 45 minutes, pairs who worked through the ladder reported feeling significantly closer than pairs given the same time for ordinary small talk ([Aron et al., Personality and Social Psychology Bulletin](https://journals.sagepub.com/doi/10.1177/0146167297234003)). One caveat worth stating honestly: the study measured closeness after a single lab session, not lifelong friendship, and follow-up conditions found the effect did not depend on matched attitudes or expecting to like each other. The escalating structure itself did the work.

That is the mechanism every good icebreaker copies. Structure removes the two taxes that make talking to a stranger tiring: deciding what to say next, and guessing whether going deeper is allowed. A game answers both. The prompt decides the topic; the format grants permission.

It also explains why so what do you do fails. It is a status question with a rehearsed answer, and rehearsed answers disclose nothing. Separate Harvard research on live conversations found that asking questions, especially follow-ups, reliably increases how much people like you, precisely because it invites unrehearsed disclosure ([Huang et al., 2017](https://www.hbs.edu/faculty/Pages/item.aspx?num=52115)).

Question games for chat

The question ladder is the 36-questions principle at friendship temperature: alternate asking, start light, and let each round go one notch deeper. Twenty questions about taste works as an opener, would-you-rather with a rule that you must explain your pick is better, and two truths and a lie remains the most efficient icebreaker ever designed, since the reveal generates three follow-up stories automatically. For any of them, the rule that keeps the game honest is that both people answer every question, including the asker. A quiz is not an icebreaker.

Guessing games run on a different fuel, the flattery of being studied. Guess my ranking, where one person lists five films or foods and the other predicts the order, produces immediate argument of the best kind. So does the assumptions game, where you each state three guesses about the other and get corrected. Being slightly wrong about someone is a faster route to their real self than asking directly, because people correct errors with detail they would never volunteer.

Games for video calls

Video adds a room. Show and tell, the childhood version, works embarrassingly well on adults: pick the nearest object with a story and tell it. Desert island debates, best-of brackets settled in real time, or a shared quiz with both of you on the same team against the internet all give hands and eyes something to do, which quiets the interview feeling that plain video calls slide into. We covered why face-to-face-only video is harder than it should be in [our guide to video chat apps](https://vairi.app/journal/best-video-chat-apps-for-meeting-new-people).

The other reliable format is a synced activity with commentary: watch a short film or three music videos together and rank them. Reacting to a third thing is the lowest-stakes disclosure there is, and it is why co-watching keeps showing up in [our reporting on watching movies with strangers](https://vairi.app/journal/watching-movies-with-a-stranger-online-does-it-work). Vairi builds prompt games and synced co-watching directly into its chats for exactly this reason: two matched strangers with a structured game between them get past the wall that so what do you do cannot.

How to run any of these without it being weird

Name the game plainly: random question game, you in? Framing something as a game is what grants the permission, so skipping the frame and just firing deep questions reads as intense. Start shallower than you think necessary, because the ladder needs bottom rungs. Answer your own questions with real answers, since the game collapses the first time someone banks a disclosure without paying one. And quit while it is fun. An icebreaker that ends mid-laugh leaves an open loop that starts the next conversation, which, as we argue in [how to keep a conversation going](https://vairi.app/journal/how-to-keep-a-conversation-going-with-a-new-friend), is most of the battle.

One more honest note: games are scaffolding, not the building. If six rounds in you have learned their rankings but nothing has snagged, that is data too. The point of lowering the cost of real conversation is to find out whether one is available.

A ready-made question ladder to steal

If you want a ladder without building one, here is a twelve-rung version tuned for new friends rather than lab conditions. Rungs one to four, the shallow end: what is your most defensible unpopular opinion about food; what app would you delete if nobody could judge you; what were you weirdly good at as a kid; what is the best purchase you have made under twenty pounds. Rungs five to eight, the middle: what would you do with a free year and no money worries; what is something you changed your mind about in the last few years; who in your life do you feel most yourself around; what are you avoiding at the moment. Rungs nine to twelve, the deep end, only if the middle went well: what do you miss; what are you more afraid of than you let on; what has shaped you that almost nobody knows about; what do you hope is true.

Run four rungs per session, not twelve. The ladder is a staircase across several conversations, and stopping early is what makes climbing feel safe.

Icebreakers that backfire

A few formats fail predictably, and knowing why sharpens the whole toolkit. Trivia fails because facts are not disclosure; you can play six rounds and learn nothing about each other except who knows capitals. Debate prompts about politics fail at the icebreaker stage because they sort people into positions before they are people to each other. Never have I ever fails outside drinking contexts because its whole engine is confession under social pressure, which is the opposite of the voluntary escalation that made Aron's ladder work. And any game one person clearly dreads fails by definition, since the safety that games create comes entirely from both players actually opting in. If a suggested game lands flat, drop it without ceremony. The willingness to abandon the game is itself a disclosure, and a good one.

The pattern across all the failures is the same: anything that extracts information instead of exchanging it stops being an icebreaker and becomes a test.

Timing matters more than people expect. The best moment to introduce a game is not the opening message but the first lull, usually somewhere in the second or third conversation, when the biographical material has run dry and both of you can feel the effort rising. Offered there, a game reads as generosity, someone taking responsibility for the energy, rather than as a gimmick. Offered as an opener, the same game can feel like a script. And on a first video call, have one game already picked before you dial, because deciding what to play is itself an awkwardness generator when two strangers are staring at each other in the selection menu of an evening.

Small logistics, but early friendship is made of small logistics.

Key takeaways

Aron's 1997 experiment showed 45 minutes of escalating questions produced more closeness than equivalent small talk, and the structure, not luck or matched attitudes, drove the effect. Question games work when both people answer everything and depth rises one notch per round. Guessing games exploit the fastest disclosure trigger there is, correcting someone who is slightly wrong about you. On video, add a shared object or synced activity so reactions can carry the load, and always end a game one round earlier than you want to.

Related reading: [What is a digital third space?](https://vairi.app/journal/what-is-a-digital-third-space), [How to make friends online that actually become real friends](https://vairi.app/journal/how-to-make-friends-online), [Best apps to watch movies or YouTube with someone new](https://vairi.app/journal/best-apps-to-watch-movies-or-youtube-with-someone-new)