It's a sham.
"The final analysis was triggered when 170 cases of symptomatic COVID-19 developed among the 43,661 enrolled participants, of which 41,135 had received a second dose of the vaccine as of November 13.
Of the 170 cases, 162 were in the placebo group and eight were in the vaccinated group, yielding the 95 percent efficacy."
I don't know exact size of groups, but assuming it is 1/2 and 1/2
21,830 placebo group - 21,668 didn't get Covid - 99.26% didn't get Covid
21,831 vaccine group - 21,823 didn't get Covid - 99.96% didn't get Covid
I don't know how you could assume equal exposure across groups and difference in infection percent in groups would be easily within the margin of error for exposure differences.
Edit: This is how I think they calculate the "95% effective rate"
1 - .9926 = .0074 and 1 - .9996 = .0004 ||||| .0074 - .0004 = .0070 ||||| .0070/.0074 = .945946
Experts are eager to see the full data, but the top-line results are very positive.
arstechnica.com