Why dispatch scores drivers instead of broadcasting to everyone
Broadcast dispatch is simple, fast to build and quietly terrible for everybody.
The simplest way to assign a trip is to send it to every nearby driver and give it to whoever taps first. It is easy to build, and it is what many platforms start with.
It is also bad for almost everyone involved, and we do not do it.
What broadcast rewards
Speed of tapping, not suitability. The driver who wins is the one who happened to be looking at their phone, which correlates with nothing useful and quite badly with driving attentively.
It also produces a race: a dozen drivers see a job, one gets it, eleven were interrupted for nothing. Repeat that all day and you have trained your drivers to ignore offers.
What we do instead
Nearby drivers are scored, and the trip is offered based on that score. The driver who receives it is one the system believes is genuinely appropriate for the job: close enough for a sensible approach, in the right vehicle class, in a state to take it.
The driver still decides. They see the pickup, the destination and the pay before accepting, and declining is a legitimate use of the system.
Why this matters to riders
A shorter approach is a shorter wait, and a driver who was chosen rather than who won a race is more likely to be the right vehicle for what you booked.
Why it matters to drivers
Fewer irrelevant offers. An offer that arrives is more likely to be one worth taking, which makes reading them worthwhile rather than a distraction while driving.
The scheduled case
For a scheduled trip the same scoring runs, but it runs ten minutes before the pickup rather than at booking time. Assigning hours ahead would optimise against a world that no longer exists by the time the trip happens.
The honest limitation
No dispatch system is perfect, and at quiet hours or in thin coverage the best available driver may still be some distance away. Scoring makes the choice sensible; it cannot invent supply that is not there.