May 10 election automation: beware of selective deployment of good and bad equipment

Election Day is fast approaching, yet we have gotten zero statistics so far on the PCOS machine error rates, in particular:

  • the average number of operating hours before a machine fails
  • the average number of valid ballots the machine rejects
  • the average scan accuracy rate of the machine
  • the average failure rate of the transmission equipment

These secrets appear to be so well-guarded that no media, political parties, election watchdogs, much less Comelec or Smartmatic has cited or released credible statistics about these test results.

Yet, they must exist. After all, due diligence requires that Smartmatic do these tests thoroughly before accepting the PCOS machines from Chinese factories that manufactured them. Due diligence also requires that the Comelec test these machines independently, before accepting them from the vendor Smartmatic. Where are the results of all these tests? We want to know if indeed the machines meet Comelec specifications, or if they are of poor quality, as some made-in-China products are. We do know that Smartmatic field tests have been plagued with quality problems: its machines reject too many valid ballots; its modems have transmission problems even in areas with strong signals, like Metro Manila.

Why do we want these tests made public?

  • We don’t want the Filipino taxpayers money wasted on lemons. If a machine does not meet Comelec specifications and deemed of poor quality, what’s the point shipping it to Batanes or Tawi-tawi, where it will only cause trouble? In fact, why should the Comelec pay for an off-spec machine, which is worse than useless, because it can disenfranchise voters or even misrepresent their intent?
  • If Smartmatic doesn’t have enough good quality machines (after all, they were made in China a hurry), we should know early enough, so we can take precautionary measures and prepare a manual backup system.
  • We don’t want Smartmatic or Comelec selectively deploying good or bad machines to certain regions or provinces in a way that can bias the election in favor of some candidates, against others.

A second chance to determine the accuracy of these machines would have been the system audit and source code review conducted by Systest Labs. Smartmatic and the Comelec claim that the audit/review is done, and they have obtained the necessary certification from Systest. However, they have released neither the proper certification documents nor the full text of the Systest report, leading us to believe, that the audit/review is either not complete yet, or it is unable/unwilling — perhaps due to some adverse finding — to give the full certification to the Smartmatic machines and software as required by law.

If Smartmatic and the Comelec continue to keep results of these machine tests and third-party audits/reviews secret, there would have still been a third chance to determine to accuracy of these machines — on election day itself. By law, voters have a right to see whether the machine interpreted their choices, as reflected in their marks, properly. The machine should therefore show on screen the names of candidates which the machine interprets to be the voter’s choices, based on its scan. This way, every voter would have had an idea if the machine was scanning the marks accurately. The Comelec, however, disabled this feature, depriving every voter of their lawful right, and keeping all of us in the dark, as of election day, regarding the true accuracy of the machines.

The last line of defense — our fourth and final chance — in trying to determine the accurate these optical scanners would have been the post-election random manual audit, where an audit team will manually count the votes and compare their results with the machine count. Should there be any discrepancy, the team has to determine whether the error is in the manual count or the machine count. Though it will occur after the elections, we would still get an idea how accurate the machines have been. But again, the Comelec now indicates that it wants the random manual audit conducted after the proclamation of winners, which makes the audit moot and academic.

Seeing that the Comelec and Smartmatic have done everything in their power to keep the public in the dark about the true accuracy of the PCOS machines, leads us to very strong suspicions that:

  • They simply do not want the public to know the true accuracy of the machines
  • Given our limited anecdotal experience with these machines, some of them do not meet the minimum Comelec specifications. Yet, by accepting them, the Comelec is binding the Philippine government (and Filipino taxpayer) to pay for them.
  • A sinister purpose drives this secrecy because it gives those who know which machines are good and which are bad the power to influence the outcome of the elections by selectively deploying good and bad machines to favor some candidates over others.

These are very bad signs indeed.

2 Comments

  1. Posted April 12, 2010 at 9:58 pm | Permalink

    Accurately said. And 30 days from the election this is all still present. CHEATING IS A CERTAINTY and not a probability.

    Instead of being critical, I think my approach would be to write about HOW TO CHEAT WITH Automated Election! We can even give detailed instructions on how to hack and have it on slides. Want to get a sponsor too to give a prize to one who can hack the tally centers and flood it with data since the digital signature will actually not be working? It would not even be a good challenge.

    How about trying to call on ONLINE PEOPLE POWER! How?
    1. How do we make sure that posted online should be Tally of every precint or PCOS? Lets do it.
    2. Have everyone verify that the precint results and posted results are the same.
    3. Use google maps to plot different PCOS machines and report incidents, tallys, etc.
    4. The above 3 will STOP Transmission and Tally cheating or so outrage the blatant cheating that will happen.
    5. Now, its just auditing the PCOS machine that is left behind. A 100% partial manual audit is the only remaining solution. All machines but not all names in the ballot will be audited. That would only extend the time to 3 hours (President, VP and Mayor)

    How can we get young IT savvy filipinos nationwide to band up, use the internet and google maps to get accurate information online during Election day?

  2. Roberto Verzola
    Posted April 13, 2010 at 6:50 pm | Permalink

    Regarding your proposal to audit only president, vice-president and mayor: what is the rationale for leaving out senators, governors, House representatives, party-list, etc.?

    For instance, based on most recent surveys, the vice-presidential contest doesn’t seem to need a manual audit as much as the senatorial contest. Or why not all chief executives only, which is at least more consistent?

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

%d bloggers like this: