99.97% PCOS accuracy ‘practically perfect’

“Practically perfect.”

So said Commission on Elections (Comelec) Chair Sixto Brillantes Jr., of the 99.9747 percent accuracy rate of the precinct count optical scan (PCOS) machines, a rate 0.0203 percent short of the 99.995 percent accuracy rate required by the election body.

The results of the random manual audit (RMA) of the 2013 midterm elections were released by the Comelec on Friday and showed that for the second time, the PCOS machines had failed to meet the accuracy rate required by the Comelec itself.

During the 2010 elections, the PCOS machines’ accuracy rate of 99.96 percent also fell short of the Comelec requirement.

Definitely accurate

But the results did not dampen the mood of the Comelec chair, who described the results as “practically perfect.”

“With 99.97 percent, it’s practically perfect. It’s not perfect but it’s definitely accurate,” Brillantes said during a press conference at the Comelec main office in Manila.

“This validates the results of the 2013 elections,” he added.

It was the Comelec Special Bids and Awards Committee that set the 99.995 percent accuracy rate requirement before it selected in 2009 the automated election system (AES) that the country was going to use.

The requirement was included in the Terms of Reference used in the bidding process.

‘60-30-10’ pattern

Comelec critics have claimed massive fraud in the May mid-term elections, citing what it described as a “60-30-10 pattern.”

According to the critics, Team PNoy senatorial candidates consistently obtained 60 percent of the votes cast, while those from the United Nationalist Alliance got 30 percent, and the rest 10 percent.

The 2013 RMA report showed that the PCOS machine’s accuracy rate for the Senate race was 99.9775 percent; 99.9719 percent for the congressional elections, and 99.9748 percent for the mayoralty races.

A member of the RMA committee said that the “variances” found in the different accuracy rates were not enough to affect the elections.

“Not anymore. We don’t usually say it’s immaterial… it’s insignificant,” said National Statistics Office administrator and RMA committee member Carmelita Ericta.

Allowable margin

According to the RMA report, the variances—or the difference between the manual audit and the PCOS counting of the votes—were either “mathematical or clerical,” but still within the “allowable margin of variance” based on a “99 percent accuracy rate.”

The report noted that there were 184 variances found in the Senate race (the allowable is 8,177), 25 in the congressional races (962 allowable), and 28 in the mayoralty elections (1,030 allowable).

“Since most of the ‘variance’ can be attributed to human error or clerical error, aiming for a higher accuracy rate to as high as 99.995 percent could be statistically improbable,” the report said.

The RMA results were based on reports from 212 out of the 234 randomly chosen precincts because 22 of those precincts had problems, including 11 that had variances which “could not be explained,” said former Ambassador to the Vatican Henrietta de Villa, head of the RMA committee.

Almost impossible

Comelec Commissioner Christian Robert Lim said it was almost “impossible” to get a 99.995 percent or 100 percent accuracy rate from the PCOS machines, and that he would be worried only if the variance were “too big.”

“At the end of the day, I have to accept that the machine and human appreciation of the votes will never really have a one to one correspondence,” Lim said.

He added that the 99.995 percent accuracy rate required by the Comelec might be achieved if another machine had audited the PCOS results.

“As the chairman said, if we get a match of 99.995 percent or 100 percent, (people might think that) we had rigged (the results),” Lim said.