Slashdot Mirror
Finnish E-Voting System Loses 2% of Votes
kaip writes "Finland piloted a fully electronic voting system in municipal elections last weekend. Due to a usability glitch, 232 votes, or about 2% of all electronic votes were lost. The results of the election may have been affected, because the seats in municipal assemblies are often decided by margins of a few votes. Unfortunately, nobody knows for sure, because the Ministry of Justice didn't see any need to implement a voter-verified paper record.
The ministry was, of course, duly warned about a fully electronic voting system, but the critique was debunked as 'science fiction.'
There is now discussion about re-arranging the affected elections. Thanks go to the voting system providers, Scytl and TietoEnator, for the experience."
http://www.arkko.com/evotingfailure
For information, I am a citizen of one of the three small places where the system was tested. I have already sent out an appeal of the decision to the voting board; if necessary, I will also appeal to the administrative court. Lets see how this plays out. I think we have a good chance of overturning the election results.
The summary has more data than the article. This was a pilot in three (smallish) municipalities, involving the 2% of the voters.
Of the e-votes cast in these three municipalities, 2% were not accounted for. So both statements are correct.
Municipality / Number of votes given / number of lost votes / lowest number of votes for elected person
Vihti: 7087 / 122 / 77
Kauniainen: 2982 / 61 / 49
Karkkila: 2165 / 49 / 35
The original Ministry of Justice announcement (in Finnish) states: "A total of 12234 electronic votes vere cast in the electronic voting pilot of the 2008 municipal elections. - -"
232 is about 2% of 12234 and therefore the summary is correct.
According to the same announcement the total number of votes in the three municipalities in which the voting system was piloted was 21073 (Karkkila 4251, Kauniainen 4843, Vihti 11979), i.e., 8839 of all voters cast a paper ballot. (The voters could choose between the traditional paper ballot and trying the new electronic system.)
Actually ministry of justice itself described 2% failure rate as "very high" compared to ordinary paper ballot. In Finland an ordinary failure rate for paper ballots cast would afaik be around 0,5% and that includes Donald Duck and offensive drawings, which are not available to evoters.
One of the pro-evoting arguments was that we get significantly _lower_ failure rates compared to paper ballots. Apparently that was not the case...
For the whole country the failure rate was 0.7%, so much less than with the electronic machine. And usually big part of them are voting Donald Duck etc.
About 0.5% of votes are ignored in the traditional voting system.
Ministry of justice itself described 2% as being "very high" figure compared to that of (afaik around 0,5% or so for) paper ballots.
In finland we get a pencil and a ballot (a piece of cardboard, about the size of big postcard) where we write the number of candidate. If there are several elections conducted at once (which is pretty rare), we get several ballots.
And yes, people old or clueless enough can screw that up too, but the screw-up-rate for evotes was expected to be way _lower_ than for paper ballots.
More about voting here:
http://www.vaalit.fi/17098.htm
As an election vote counter I can assure that out of the approximately 7000 votes that went thru my hands during the counting, only 9 or 10 were that ambiguous that it couldn't be reliably placed to one single candidate. Those ambiguous votes go to the board of election officials that will ultimately decide whether it's a valid vote (and who has the voter voted for) or not. Other invalid votes were maybe 5 times as common. Most of the time it's a question of whether the number is "1 or 7?" and other common problems are "6 or 0?" and "5 or 6?"
The Finnish counting system was developed during times of great distress and has stood the test of time. It was good right after the civil war and therefore it's good for peaceful times too:
The votes are first grouped by candidate, then counted twice by separate persons and invalid or ambiguous votes taken aside. If the numbers differ, they're counted again by two separate persons. Then the count is recorded on two separate forms held by secretaries and those forms are cross-validated against each other.
After this, the votes are given to second counting group selected at random (obviously different from the first group) and counted again, with a possibility to take aside votes they found invalid that were accepted previously but not vice versa. If this verification count differs at all from the first count, the number of votes for candidate will be verified by counting again the number of votes for that particular candidate and if the first count seems to have been erroneous it'll be counted for the third time by a third group. Finally the invalid votes will be considered and decided whether it is an acceptable vote or not by higher election officials. Each party attending the elections have a right to set observators to the counting procedures but at times like these I saw none personally.
This whole procedure makes it really hard to cheat in the vote counting unless you're using e-voting where officials just download the XML, turn it into a PDF and print it. Then they tell us that this is the result. I'd love to link to the news video where they did that but unfortunately I'm unable to find it right now.
?SYNTAX ERROR
This sounds like a nice feature to keep stupid people from voting.
Spoken like a true, arrogant techie.
Why are only special people privileged to counting? Can they not be bought?
There are no special people. Counting the votes has to be done in public, you can go there and watch.
Life is just nature's way of keeping meat fresh.
If there was a visible feedback that a vote had or hadn't been placed (say a printed paper record) then the voter could immediately see that they hadn't pressed a final OK button and correct the issue.
As it is it appears there was no feedback or indication that there was a final step needed after selecting the correct candidate.
I finally found the video: http://www.hs.fi/videot/1135240559892?kategoria=Uutiset&sivu=1
?SYNTAX ERROR
In Spain the polls close at 8pm and typically 90% of the votes have been counted by 11:30pm, 95% one hour later and 98% by 3am. This is a country with 45 million inhabitants.
Your head a splode
I belive the AEC are counting what are known as donkey votes, from the same site the summary in their report on electronic counting after studying it during the last US elections and elsewhere is quoted below...
"Electronic voting has received significant recent media coverage, and, with the Internet becoming more pervasive, the topic will continue to receive much attention. It must be recognised that a lot of the hype being generated is by the vendors of electronic voting systems.
There are currently a range of issues associated with the introduction of electronic voting and vote counting. Each of these needs to be identified and strategies put in place to resolve them.
The possible starting points within Australia, recommended in this report, have significant business cases for providing alternative technical options to voters in order to strengthen the democratic process.
This paper does not suggest that Australian electoral authorities should at this stage embark on a program to fully replace the easily understood, publicly and politically accepted efficient, transparent paper ballot system that currently exists."
Translation for Aussies: "Tell Diebold they're dreaminn...". Further skimming of the report shows that electronic voting has been used as a successfull option in certain circumstances, such as assisting blind people to vote in secret.
And did you exchange a walk on part in the war for a lead role in a cage? - Pink Floyd.
Even if the risk that 1 person is making a mistake is 0.005% the risk is increased a if you have 5000 people counting votes.
So the probability of one person making no mistakes is 1 - 0.005% = 1 - 0.00005 = 0.99995. The probability of all 5000 people all making no mistakes is 0.99995 to the power of 5000, or 0.778796. The probability of at least one mistake is 1 - 0.778796 = 0.221204.
Or roughly one in four.
That leaves the question of what the impact of one mistake is. If it's dropping, inserting, or changing one vote, then that's probably acceptable unless the race is that close.
If it's misreading a digit when you report the number up the tree, it might have a much larger impact.
E-voting is a long term investment and staring at the initial costs is useless.
For the sake of the argument, I accept that. What does it do? Count faster and save money.
What are the costs? People are less able to generate correct ballots. Almost no one understands how the technology works, and thus are not truly able to trust the results. It becomes very easy for the machine makers to manipulate the outcome.
Voting is such an important part of democracy that doing it right is worth almost any amount of money. And speed? I'd rather have the people's candidate in a week than the machine makers' candidate in an hour. Even if I disagree with the people and agree with the machine makers.
Just like science, it's the process that's important, not so much the results.
--Jonas K