[dmoy]

> I'm wondering how do they test it, do they inject volunteers with the vacine, and they tell them to go infect themselves with a virus that 1 time out of 100 will kill them and an unknown percentage disable them?

Obviously no

They get a large set of people (tens of thousands), half placebo, and then check rates afterwards.

[loeg]

The larger the set the faster the results, but a large set is expensive. "Afterwards" is sort of a simplification. The study arms are compared regularly throughout in case the study needs to be ended one way or the other (Data and Safety Monitoring Committee). (If the vaccine is effective: great, we want to ramp up production and give it to everyone ASAP. If the vaccine is ineffective or has very bad side effects, we also want to conclude the study.)