If we put aside the protocol failure (accidentally giving half dose then full dose), and ignore the fact that they tried to not discuss this for some time until

If we put aside the protocol failure (accidentally giving half dose then full dose), and ignore the fact that they tried to not discuss this for some time until called out on it, of course, you can't stress-test (injection of virus to actually test for immunity) as it's not ethical. So the summary metric is the Vaccine Efficacy, calculated as 1-relative risk, where relative risk is the incidence of infection in the vaccine group relative to the incidence of infection in the control group. So with N = 75 the targeted number of post-dose infections, and V = # vaccinated out of the N, P = # placebo out of the N, are they saying that 100*(1-V/N) is the efficacy ? Then if it's 50%, it actually means you have no benefit over the placebo.