On the back of ludicrous allegations of EVM tampering/hacking, the Election Commission chose to check random vote counts after the recently concluded Gujarat Elections. As reported, the Election Commission officials claimed that random vote counts in 1 booth each in 182 seats of the state and their corresponding VVPAT paper trail proved to be a 100% match, which means that there was no discrepancy in what button the voter had pressed and what the EVM consequently registered in its system.
Even after this thumbs-up to the EVMs, folks like Hardik Patel continued to cry wolf. Some in the Congress too repeated the EVM hacked line, even when others in the Congress accepted the results and in fact claimed a “moral victory” based on the results given by the same “hacked” EVMs. Now, AAP leading portal Janta Ka Reporter, which is famous for spreading fake news, has come up with another sensational report. The report is titled:
It is based on the part-statement of the Chief Electoral Officer B B Swain, who as per a PTI report, made the following statement:
There was a 100 percent match in the results of the EVM and the slips produced by the paper trail machines
Janta Ka Reporter, refers to the same report, and quotes another statement by him:
There was a mismatch of some votes on one booth each of these four seats. This occurred because the Returning Officer must have made the same mistake but it could not be detected earlier. So we took into account VVPAT slips for these booths during the counting and resolved the issue
Janta Ka Reporter then adds the “analysis” of a twitter user, who based on this statement, extrapolated that the quantum of mismatch was over 2%, which could impact 6 lakh votes. He further pontificated, as quoted by Janta Ka Reporter:
Accepting the premises set by EC that 1 booth per constituency is representative sample. number of booths = 50128. questionable booths = 4*50128/182 = 1102 EC accepted that there were 4 mismatch. 1102 possible mismatched booth in 182 constituency. so approximately 6 mismatched booth per constituency. One EVM can store around 3800 votes. assuming on average EVM were half filled, so total number of questionable votes per seat = 11400. In general, the average victory margin in state assembly is 5000-6000 votes… In conclusion, the entire result can change within the error margin of EVM system. This has nothing to do with hacking. this was simple calculation for variability from election commission data. EC must tell what were the votes in EVM and VVPAT. and how did EC conclude that the mismatch was because of old data? EVM do not record time, as far as I know
The crux of the argument is that “Old data” was existing on some of the EVMs, which after extrapolating, could have impacted 6 booths per constituency, and as a result, 11400 votes per seat. The argument further says that the average victory margin is generally 5000-6000 votes, so the entire result could change because of this old data.
Firstly, yes the Chief Electoral Officer B B Swain did say that there was some mismatch in 4 booths. But there are some facts which preceded this statement, which have been ignored by the Janta Ka Reporter. Quoting from the PTI report in full:
Besides, as announced earlier, VVPAT slips were taken into consideration during Monday’s counting at 10 booths across seven constituencies as the presiding officers of these booths had failed to wipe out the votes from the EVMs during the mock poll ahead of the voting on December 9 and 14.
In addition, four such new cases had emerged during Monday’s counting at Vagra, Dwarka, Ankleshwar and Bhavnagar- Rural seats, Swain said.
“There was a mismatch of some votes on one booth each of these four seats. This occurred because the Returning Officer must have made the same mistake but it could not be detected earlier. So we took into account VVPAT slips for these booths during the counting and resolved the issue,” he said.
The first two paras have been deliberately omitted by Janta Ka Reporter. The full picture reveals that the “old data” which remained on some of the EVMs was actually the “Mock Poll” data. The Election Commission (EC) had earlier ordered repolling in six booths, where first phase polling was held on 9 December as the presiding officers of these booths had forgotten to wipe out the mock drill poll results necessitating the re-election.
Before we move ahead, we need to understand the concept of a “mock-poll”. Mock polls are conducted on the election day to test the efficacy of the EVMs and VVPATs. An hour before the actual polling begins, a mock poll is conducted where around 50 votes are cast at every polling station by the presiding officer in presence of the polling agents. Polling agents are the authorized representatives of all the candidates at a particular booth. Thus, each booth has polling agents from all the contestant parties/individuals.
The polling agents of all parties are involved so that each of them can individually verify that the EVM and VVPAT are working fine. For example, if an EVM is “hacked” by BJP, the Congress poling agent will immediately detect that the EVM is not functioning properly during thr mock poll, and file a complaint. Hence, the mock poll acts as a test, and is done for each and every EVM. After the mock poll, the presiding officer ascertains the result in the control unit of the EVM, counts the VVPAT paper slips and confirms that the results tally for each candidate.
Once this is done, the EVM data is deleted and the VVPATs are emptied to let the polling agents verify the empty VVPAT drop boxes. The mock poll slips are stamped on the back with inscription “MOCK POLL SLIP” and sealed in a thick black paper envelop. Now, the EVM count is back at zero, for normal polling to begin.
Coming back to the case of the four booths, Mr Swain had stated that there was a mismatch at these booths, probably because the returning officers did not wipe out the mock poll data. It appears that in case of 14 booths (10 earlier and 4 now), out of total 50,128 booths, the officials failed to delete the data.
How was the mismatch in these 14 booths noticed? Because all polling booths maintain manual records of total votes polled. That’s how we get regular update of polling percentage on voting day, as polling officers keep sending this data regularly to ECI. This data is in fact cross-checked every few hours, i.e. EVM total is matched with manual count, and the tallied data is sent for reporting percentage.
This final count of polling is recorded on Form 17C. During counting, this total for each booth as per Form 17C is matched with total shown in EVM for that booth, and EVM data is accepted only when this data matches. The Form 17C also has a provision to account for mock polling data which was deleted from the EVM. Hence if there were errors in other booths, those would have been noticed by counting officials and authorized counting agents of candidates.
Janata Ka Reporter has wrongly assumed that 4 booths out of 182 booths had errors, and then used this to show that the error rate is 2% for entire state. This it totally wrong, 182 booths didn’t have any error. And 14 different (not 4) booths had errors in the entire state, so there is no case for expanding anything by 2% to get the figure of 1102 booths with possible mismatch. 14 is the final number of booths where mismatch was detected.
Even in such booths, the mismatch can be expected to be of around 50 votes, which were polled during mock-polling. Janta Ka Reporter claimed that these votes could be as many as 11400 per seat. Thus using half-cooked data, a half-baked analysis has been presented by Janta Ka Reporter.
Corporate Dropout, Freelance Translator