A read, a countdown, and a clean close

Each question gets a three-second read before its answer window opens, then five to fifteen seconds on the clock, and a second of grace at the buzzer.

Three seconds to read it

When you launch a question it appears with a short lead-in before the clock starts, three seconds, which is roughly how long it takes a table to look up from their drinks and read a line of text.

This is not decoration. The answer clock starts after the read, so a player who answers the instant the options become live earns the question's full value. Without it, every question would quietly punish the table sitting furthest from the screen.

Then the window

The answer window runs for however long that question is set to: between five and fifteen seconds, with fifteen as the default. The countdown is on the host view and on every player's phone at the same time, counting the same clock.

Fifteen seconds is the ceiling on purpose. A longer window does not make a quiz more thoughtful, it makes a room go quiet, and a pub that goes quiet for half a minute at a time loses the night. If a question genuinely needs longer than fifteen seconds to answer, it is usually a question that needs rewriting.

The buzzer, and the second after it

When the window ends the question closes and answers stop counting. There is a second of grace on top, because a player who tapped exactly at the buzzer on a slow connection should not lose the point to their signal rather than to their knowledge.

Because the close is on a clock rather than on your judgement, the argument about who answered before you said stop does not happen. Nobody has to adjudicate it, including you.

How the clock changes the scoring

On flat scoring the timer decides only whether an answer counted. On time-weighted scoring it decides what the answer was worth: the value decays across the window, steeply at first and then flattening, down to a fifth of the question's points at the buzzer.

That interaction is worth knowing when you set a window. A five-second question on time-weighted scoring is brutal, because the whole curve happens at once. Fifteen seconds gives a table long enough to confer and lose points for doing it, which is usually the effect you want.

You still control the pace

The timer governs one question. It does not govern your night: the next question goes up when you launch it, so the gap between them is as long as you need, long enough to take an order, refill a table, or let a round settle before the next one starts.

That is the difference between this and software written for a classroom, where the whole quiz runs on a single clock and the room has to keep up with it.