Protecting The Votes in Venezuela Part II: Insuring the Integrity Of The Voting System

August 20, 2012

The Comando Venezuela will have to work with the same voting system used in the previous four elections. This system has been audited extensively and there is no proof that via this system any votes have been modified since the 2005 election. The only part that is different in this election, is that a new fingerprint software has been introduced, which implies that there will be extensive checks in this new part of the system.

The voting system process is like this:

The voter shows his or her ID card and register the fingerprint. If it is that in the database (which is local) the President of the “Mesa” unblocks the voting machine and the voter can cast the vote. Should this fail, the voter fills out a form and registers a new fingerprint. Thus, one of the audits that will be performed is that of the quality of the fingerprint database to insure that it is uniform across the country in both geographic and socio-economic terms. This are checks on the fingerprints which have international standards of quality.

The Comando Venezuela knows that the only purpose of the fingerprint system is to intimidate, it will thus audit and check the system to insure that voters can cast their vote with confidence that it will be secret.

The vote itself only contains information about the vote, it has no information about the voter and it is all encrypted. When the vote is stored in the machine, it goes into a temporary memory:

Each time a new vote comes into this temporary memory, another is moved at random to the permanent file with the record of the votes. Thus, theer is no sequencing in the memory. The national ID number (cedula) also goes into the machine, also into five memories and one is moved at random, to record that this voter already cast the vote and can not do it again. The two registers and files are separate and they are both encrypted. Recall that the opposition did not participate in the Parliamentary elections in 2005, precisely because the tests showed that there was some sequencing which was later corrected.

Once all of the votes are cast, the machine prints the total tally, with enough copies for all of the witnesses at each “Mesa”. After this printing has taken place and only after, the machine is connected to the telephone line and the results transmitted to the CNE. At the CNE, all political parties have representation in the totaling room. This has been the case since 2005. (Note that 2004 is not included)

After all of the mesas in one center close, 54% will be audited at random. That is, 54% will be counted by hand and insure that the manual count agrees with the machine count.

Thus, if we have witnesses at all Centers and Tables, it is very unlikely that the vote is not secret and/or that the vote can be changed. In fact, there is no evidence that any vote cast by a voter has been changed since 2005 or that the machine count has been altered after the vote was completed.

However, we need witnesses everywhere, as we need to insure that the votes are truly cast in each mesa and center by real people. In part 3, I will describe what the problem is and why the Comando Venezuela has made of this the centerpiece of its strategy to protect our votes.

35 Responses to “Protecting The Votes in Venezuela Part II: Insuring the Integrity Of The Voting System”

  1. extorres Says:

    I like the bowl analogy of balls being picked at random, but I think it may fail to represent what goes on inside a computer when one states that the computer picks a vote randomly from 5, or reorders cedulas from amongst the most recent 5. A computer’s choice is only random in that the results seem random to one that does not know the algorithm being used. For someone who does know the algorithm, the inputs are quite repeatable from reversing the outputs.

    Even without knowing the algorithm, however, there ways to determine a non negligible percentage of people’s votes. For example, if 6 votes in a row cast are for candidate A, then it is easy to associate at least one cedula to candidate that candidate with certainty. The percentage of people’s votes can reach worrisome levels if the probabilistic information coming from this system is then linked to the certainty of the tascon/maisanta information.

    If this seems farfetched, read up on services on the internet that offer having your IP be anonymous. Their server would strip all your traceable information before then requesting information for you from the Web. The FBI quickly came up with measuring the time it took for an IP request to come in to these servers and the time the anonymous requests came out. As these servers began countering with delays and randomnizing of the order of requests, the FBI began countercountering with more sophisticated statistics to at least have likely candidate IP numbers, that, with time and repeat behavior would allow them to remove the noise and pinpoint IP information with certainty. These servers are now working in groups sending information between each other before sending requests to the Web, but with each added measure, countermeasures are developed. This voting system is tantamount to child’s play, nowadays.

    • Tomaz V Says:

      Sure, but the question is why bother? If they wanted to retaliate for voting for the opposition they can do so much more easily by going against entire constituencies where their candidate lost. Running sophisticated vote determination schemes and thus compromising electoral integrity to the extreme to go against individual voters is simply not worth it, imho.

      • extorres Says:

        They would bother because they’ve learned from the master, castro. Keeping track of friends versus foes, is a huge deal. They need to know who they can trust to advance in their ranks and who they need to keep out of their ranks. They will pull no stops to obtain this list the way they got the tascon list. Even if they lose the election, this list is the foundation for getting control back. And if they win the election, it’s about cementing themselves for good.

        • Tomaz V Says:

          Hm, putting it that way gives it some sense. It still seems a bit excessive to me, but then again, they are an excessive bunch.

          At this point I’m not sure if Chavez can actually win. He can score a Pyrrhic victory at best, with a large drop in popularity as the hangover from the binge spending comes to play. Unfortunately the same could be true for Caprilles.

          • extorres Says:

            The real problem is that to most people something like this would seem like excess. The truth is that it only takes a single computer analyst to reverse engineer the data using the two databases. It’s a simple puzzle that would attract countless of candidates.

            It is for this reason, we should not let the data exist, at all. What I’m pointing out is that, as described, the 5 number jumble just doesn’t cut it.

            • Tomaz V Says:

              Yes, that’s true, but I was referring to going after every individual voter afterwards. It is true though that it is doable – perhaps even easy – if you have a widely decentralized system of harassment and abuse.

              So yeah, I can now see why this could be a major problem.

  2. extorres Says:

    As to vote changing, the question remains, *if* something odd is detected, then what guarantees that it will not go ignored. For example, what if the 54% of randomly chosen tables turns out not to pass a randomness test? Will a new 54% sample be chosen? Or do we have to accept the original 54% just because it was chosen by random procedure, even if it is well known that randomly selected samples do not always produce samples that pass randomness tests?

    For those who don’t remember, this example is exactly what happened with the RR, where the random sample did not pass a randomness test. To make matters worse, CC then determined that it was acceptable using correlation, which does not only does not detect lack of randomness, it’s of ignorants to think that a correlation tool would have shown lack of correlation in such circumstances. So, the question remains, what if something screwy is detected again?

  3. guest Says:

    “there is no evidence that any vote cast by a voter has been changed since 2005″

    I remember quite clearly that Chavez’s dad, Tarek Wiliam Saab, and Aristobulo had their votes changed (to null) at some election (and were allowed to vote again, despite this being extremely against the rules). And there have been plenty of reports of the machines changing the votes. Though I guess you meant “changed” after it was printed and deposited in the box, but that’s not necessary for a fraud to work.

    And I quite remember in one election where CaracasChronicles had a guy working at a voting center and the CNE guys were rather adamant about not letting anyone see the signature sheets, which only makes sense if the machines were allowing a voter to vote more than once and the vote total would contradict the number of signatures (e.g., the voting machine had a secret button that when pressed, allowed the voter to vote again in that machine).

  4. Ira Says:

    Stealing an election is the easiest trick in the book. Happens all the time, and happens everywhere.

    It’s not ONE method that does it, but a combination of “small cheating” techniques that added up together, guarantee a win.

    Just ask Joseph Kennedy. (Well, if he wasn’t dead, you could ask him.)

    Chavez has already cheated by using state resources for his reelection campaign, and making available resources to ferry his dimwit supporters to the polls. Combine this with a MINOR 1% to 2% ballot box cheating advantage and game over.

  5. extorres Says:

    By the way, supposing no problem with the 54%, what if later CNE reports different tallies for the 46% than what the loser has in his tallies? What’s going to be the process for determining the truth, a recount? The supreme court? or just the CNE’s word?

  6. moctavio Says:

    You cant change the printed record from the machines. `period.

    • Leon Says:

      What happens if the manual count of the printed ballots dont match the acta report that the smartmatic machine prints at the end of the day?

  7. moses Says:

    Good evening Miguel,

    When you say that 2004 is not included it is because previous elections were not 100% clean ?

  8. moctavio Says:

    Yes, most of the audits and controls were imposed after the 2004 recall vote. For the recall vote, there was only suèrficial checks of the vote. <there were no audits either.

  9. Alek Boyd Says:

    Miguel, I am going to take issue with some of your claims here:

    “This system has been audited extensively and there is no proof that via this system any votes have been modified since the 2005 election.”

    It is exactly the same claim that RG Aveledo has been making all along.

    Last time the REP was independently audited was in 2005. It was determined that it had numerous inconsistencies.

    Last time an opposition technician was allowed to play with the system was in a mock election in November 2005 (Fila de Mariches). Let us not forget that said mock election was suddenly stopped by Jorge Rodriguez, when the oppo technician (Leopoldo Gonzalez) was able to call out exactly how each of the participants (EU and OAS observers) had voted.

    Therefore, of what extensive auditing are we talking about here?

    You know as well as anyone, that the opposition has never been able to get its hands on a voting machine, and its software.

    You know as well as anyone, that the opposition has been signing audits to a REP done by the CNE, in which it neither participates, as witness / coauditors, nor it has any input as per methodology to be employed.

    So, in all frankness, I fail to understand how can you claim that the system has been extensively audited. For meaningful auditing means that all parties have access to all parts, I repeat, unimpeded access to all parts, of a voting system. You know as well as anyone that that is not the case / has not been the case in Venezuela, but please do correct me, with examples of dates, places, observers, etc., of audits done to date.

    Continue…

    • moctavio Says:

      Let’s not mix subjects. The REP was the subject of the previous post. The REP is being audited and all political parties get updates of the REP regularly. There are insconsitencies left but a large majority have been removed.

      There have been two mock elections like the 2005 since then, there will be a thir one in two weeks, so it is incorrect to say nobody has been allowed to play. In fact, they have looked at the software and played with the communications.

      There has never been a case of someone saying the machine counted less or more votes than the acta said.

      • Alek Boyd Says:

        “The REP is being audited and all political parties get updates of the REP regularly.”

        Correct, the REP is being audited by the CNE, in audits in which the MUD does not participate, nor does it have any input as to methodology of audits. It then signs in agreement with the CNE, ergo they are vouching for something they have no control over. And just to be clear: art. 120 of the LOSPP says that CNE must publish 60 days prior to an election the entire REP, and art. 97 established that REP is public. And yet, last independent audit of the REP was done in 2005…

        Two mock elections since 2005? Did oppo technicians get to “jurungar” the whole thing, or where they just invited to sit and watch what the CNE did? If the system is so trustworthy and secrecy is guaranteed, why did the oppo burned the cuadernos after the primary and refused the captahuellas?

        • moctavio Says:

          Again, I wrote a previous post on the REP, the discussion should be there. There is no detailed audit of the individuals of the REP, but what they have done is described in that post is a study of the changes in the REP since 2010. And they seem to be logical.

  10. Alek Boyd Says:

    “The only part that is different in this election, is that a new fingerprint software has been introduced, which implies that there will be extensive checks in this new part of the system.”

    It has already been established that the intended objective of the new fingerprint machine -to guarantee one person one vote- will not be attained.

    The MUD has already accepted as much.

    The MUD was also fastidious about not using said mechanism in the primaries, with the precise intention of encouraging massive participation. This, you will admit, is an explicit acceptance by the MUD that the captahuellas are indeed a fear factor thrown in for good measure.

    So the question is, why were the captahuellas a deterrent to participation in the primaries, but not so in the presidential vote?

    The MUD already knows that the captahuellas are not fit for purpose, why is it accepting them then?

    Continue…

    • moctavio Says:

      Because fighting the fingerprint machine with a CNE with 4 to 1 was a losing battle. Accept them and make sure they dont do any harm. Concentrate in the important matters.

  11. Alek Boyd Says:

    “Thus, one of the audits that will be performed is that of the quality of the fingerprint database to insure that it is uniform across the country in both geographic and socio-economic terms. This are checks on the fingerprints which have international standards of quality.”

    What exactly do you mean by uniformity of fingerprints across the country in geographic and socio-economic terms?

    What international standards of quality are those the MUD refers to?

    And then:

    “The Comando Venezuela knows that the only purpose of the fingerprint system is to intimidate, it will thus audit and check the system to insure that voters can cast their vote with confidence that it will be secret.”

    When, where, how, and by whom will the system be audited and checked?

    And then:

    “At the CNE, all political parties have representation in the totaling room. This has been the case since 2005. (Note that 2004 is not included)”

    This is incorrect. In 2005’s legislative elections the opposition did not participate, after it was demonstrated that the secrecy of the vote was compromised. The opposition did not have witnesses in the totalling room in 2005.

    Continue…

  12. Alek Boyd Says:

    “Thus, if we have witnesses at all Centers and Tables, it is very unlikely that the vote is not secret and/or that the vote can be changed. In fact, there is no evidence that any vote cast by a voter has been changed since 2005 or that the machine count has been altered after the vote was completed.”

    I am glad that, finally, the MUD has realised how to force Chavez to show the world whether or not he’s got the votes. As you know, I have been advocating for that for years.

    You are right in saying that there is no evidence that any vote cast has been changed and that the machine count has been altered. In my opinion however, that claim needs to be qualified. 2005 does not count for the reason mentioned. In 2006, the opposition failed to place witnesses in 40% of centers. The results of the vote in 2007 are, still to this day, unknown, therefore no party can claim that no tampering took place (Baduel, students, etc., would surely take issue with the “no tampering” bit that year). Arguably, since only 2008, the opposition can claim that results have not been tampered with.

    “However, we need witnesses everywhere, as we need to insure that the votes are truly cast in each mesa and center by real people. In part 3, I will describe what the problem is and why the Comando Venezuela has made of this the centerpiece of its strategy to protect our votes.”

    In the absolute absence of meaningful audits to REP and voting system, witnesses everywhere are the only solution, so I very much look forward to your next instalment about how the MUD is going to ensure presence in 100% of centers.

    • moctavio Says:

      Again , the witneses and “fake voters” is a different problem which has nothing to do with the machines, it has to do with placing witnesses in the right places. That will be post 3.

      • Alek Boyd Says:

        I am not arguing about “fake voters”, frankly I can’t tell one way or the other, without a thorough audit of the REP. However, I am extremely suspicious of the many parroquias where Chavez gets close to 100% of the votes.

        • moctavio Says:

          But that is probably the biggest problem, in mesas where we have no presence, before the close, those managing the mesa voting for those that did not show up.

          • Alek Boyd Says:

            “But that is probably the biggest problem, in mesas where we have no presence…”

            Indeed, last time round, it was 40% of mesas. This is why I am ever so worried, and look forward to hear what the geniuses of MUD -many of whom are the same ones that were calling the shots in 06- are going to do this time.

  13. moctavio Says:

    Is not the MUD, its Capriles’ command, there is a difference.

  14. Alek Boyd Says:

    It also worries me the stonewalling and insensible approach of Mario Torre, Roberto Picon, et al to very legitimate requests made by ESDATA, Salas, et al about how the monitoring is being done.

    And business relationships between both Torre and Picon with the Chavez regime most certainly does not add to their fitness, as thorough and objective fellas in whose hands the opposition’s hopes are placed.

    It’s as if the MUD/Comando Capriles/whatever were keen on adding to our many electoral doubts, rather than dispelling them. Are there no other technically capable people in Venezuela, without rabo de paja or suspect connections, to take on the CNE?

    And lastly, what bothers me the most is that, to MUD/Comando Capriles/Petkoff-and-his-sycophants/etc, a questioning attitude towards the CNE is akin to radicalism, when we have all been witnessing the obscenely illegal way in which it has behaved since 2004. Mind you, demanding for our legitimate rights to be respected is equal to radicalism. That is how fucked up the whole debate in Venezuela has become…

  15. Nitpicker Says:

    Nitpicking: “insuring” is the act of getting insurance (as in “insurance policy” backed by a financial institution) which is not the case. The proper verb would be “assuring” or “ensuring”.

  16. extorres Says:

    I’m still pondering the reasons for the temporary memory of 5 votes. Why is it 5? Why don’t they keep randomnizing the total of votes? Why are they even keeping a list of votes in the computer, instead of just a tally?

    This reminds me of accounting systems that actually “move” data from a libro diaro to a cierre de mes. To the computer, it is equivalent to simply mark as closed any item in the libro diario, without having to repeat the data on a different part of the disk. It’s just a waste.

    So, why keep a pseudo random sequence of the votes, instead of just not keeping them, or just randomnizing them in total rather than in groups of 5? And what they call temporary and final memory, are they talking ram and harddisk?

    The only explanation to which I keep coming back for this method is to have a way of reversing it. Otherwise, it makes no sense.

    • moctavio Says:

      Easy, so that audits could be performed of the system and whether theer was a trail or not. Cedulas have to be kept, so as to block people from voting twice.

      • extorres Says:

        That doesn’t quite address the issue I’m raising. Perhaps this question will hit the nail: does paper trail trump the computer trail, if there is a discrepancy between the two?

        If you can bear with the boring logic tree, below, you’ll see that the recorded list of votes, in all cases, is useless. The list has to to add up to the sum, or the machines are disqualified. So, there is no reason to keep the list of votes in any order. Heck, the computer could just keep all the recorded votes in alphabetical order for all we care. There is no reason at all, for the random bowl simulation. It’s pure computing waste, and achieves zero voting audit security. The only use I can see is backtracking to pair votes to cedulas.

        In fact, why do the cedulas have to be kept in any order, either. They should be in numerical order, to make it easier to spot duplicates or to find any single one of them. There is no reason to have a bowl simulator that risks reversing the process to pair up with votes.

        Logic tree: at the end of the day, assume a computer is one of the audited ones. One of two things can happen:

        A) the computer vote tally matches the paper ballot trail, or
        B) it doesn’t.

        If A, then there is no point in looking at the recorded votes in the computer.

        If B, then one of two things can happen:

        B1) the recorded votes in the computer do not add up to the computer tally of votes, or
        B2) it does.

        If B1, then the machine has to be considered faulty, so the paper trail must trump the machine. In fact, all the machines with the same hardware and software should be in doubt, which means all of them.

        If B2, which is assumed, so really they are redundant values, then one of two things can happen:

        B2a) the paper ballots are recounted and agree, which takes us back to A, or

        B2b) the paper ballots are recounted and still disagree, which takes us to the question, do the paper ballots trump the recorded votes in the computer?

        • moctavio Says:

          Yes, paper ballots trump if there is an audit.

          It is an incredibly kludgy system. All these checks have made it so.

          • extorres Says:

            But that’s my point, this bowl simulation thing is not a check. It serves to make the vote identification possible, while it adds nothing to the *audit* process. So under what justification was it added to begin with?

            It seems like a con pretending that this would help the audit, but that’s a lie. What’s left is an intentional attempt at reconstructing the order of voters and their votes.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 11,711 other followers

%d bloggers like this: