Can't decide where to eat?

Group deadlock is a process failure, not a food failure

When everyone says “anything is fine” and nothing gets chosen, the group is missing a decision ritual. Hong Kong’s restaurant density makes polite flexibility expensive: infinite acceptable options mean zero commitment. Deadlock thrives on open-ended chat. It breaks when someone introduces a shared shortlist and a time box.

Random Eats HK is built for that break. Shared filters create common constraints. Three random licensed picks create common objects to accept or reject. Reshuffle creates a second chance without restarting the whole debate. The food was never the hard part; agreeing how to choose was.

Name the failure mode out loud once: “We are not bad at food; we are bad at closing.” Then run the picker. Humour helps, but structure helps more. Groups that reuse the same ritual spend less emotional energy each night because nobody has to invent fairness from scratch while standing on a crowded pavement.

Chip presets that end arguments early

Before anyone names a restaurant, name constraints with chips. District first if you are meeting in person. Budget band next if price fights are common in your group. Then at most one style chip: quick meal for weekday nights, group dining when the table is large, skip chains when someone is tired of logos.

Presets work because they move conflict upstream into categories people can negotiate faster than specific shops. Arguing about whether Mong Kok plus $ is acceptable is shorter than arguing about eight candidate restaurants with overlapping reviews. Once presets are set, randomness does the remaining work.

• Agree district before opening the picker
• Add one budget band if money is sensitive
• Add one style chip only if it is a hard need

A time-boxed decision ritual you can reuse

Set a two-minute timer. Generate three picks. Everyone gets one veto total across the three cards — not unlimited complaints. If a card survives, open the map and go. If all cards die, reshuffle once before the timer ends. If the second round also fails, widen one filter and take the first acceptable card without further debate.

Write the ritual in the group chat once so you are not renegotiating rules while hungry. Solo users can compress it: one pick, one optional reshuffle, leave. The ritual’s power is the ending condition. Without an ending condition, “can’t decide” simply finds a new interface to live in.

If remote friends are joining later, still lock the meeting-point district now. Do not leave geography open “until everyone arrives.” Arrival is when you need a table, not when you need a debate. Generate the shortlist early, share screenshots, and keep one reshuffle in reserve for the person who shows up last with a veto.

Why “you choose” is a trap in this city

Handing the decision to one person sounds kind and often feels like pressure. That person becomes responsible for everyone’s mild disappointment. A random shortlist from shared filters spreads responsibility: the group chose the constraints; chance chose the candidates; the group only accepts or rejects.

This is why coin flips feel too blunt and review threads feel too long. Filters-plus-random sits in the middle. You keep agency where it matters — where and how much — and outsource the final shortlist to a fair draw over licensed restaurants.

Daily lunch indecision and evening group chats

Office lunch deadlock is usually geography plus fatigue. Lock the office district, enable a quick-meal chip, and reshuffle until one card feels easy. Evening group chats are usually preference noise. Lock the meeting point district, add budget, and enforce the veto limit so the thread cannot expand forever.

In both cases, the win condition is leaving the phone and walking toward a licensed address. If you finish the ritual with another bookmark and no movement, you did not finish. Can’t decide where to eat is solved by commitment technology, not by more information.