George and Al are tied with 100 electronic votes each. You count a
sample of 20 paper receipts and Al is ahead 15 to 5. Other than Al won
the sample, you have learned absolutely nothing about the correctness of
the total vote.

"Peter Duniho" wrote in message
...
"sfb" wrote in message
news:sSXsf.9913$Q73.2199@trnddc03...
Hand counting a sample proves nothing

I never said it "proves" anything.

as you can't assume the identical distribution of votes in the
uncounted votes.

Who would make such an assumption? Why would you? What point in
making such an assumption would there be?