[E-voting] The VVAT Debate - Technical or Not ?

David GLAUDE dglaude at gmx.net
Wed Mar 10 00:20:36 GMT 2004


I am not a statistician (actually I might confuse statistics and 
probabilities and that alone explain I am not an expert).

But what do we want to proof?

That the result of the election was not modified by an "error" in the 
system.

I guess if A win against B by 55% against 45%... the size of the sample 
must be bigger than if it is by 90% against 10%.

So if you want to garantee with a error margin of 0.1% that the result 
of the election is OK... how many vote you need to verify in a VVAT way.
Now assuming that durring this verification you find mismatch between 
eVoting and paper voting, wich level of mismatch would be acceptable 
(assuming they don't change the result).

But remember that for Ireland, you don't need to recount to verify. The 
sorting and random check is enough to catch most problem.

David GLAUDE

-- 
Don't let computer expert control election...
Endorse: http://www.free-project.org/resolution/
For Belgium: http://www.poureva.be/







More information about the E-voting mailing list