Overkill : On only a 100% recount of VVPATs
Verification of a statistically significant sample of VVPAT slips should suffice
The introduction and use of the Voter Verifiable Paper Audit Trail (VVPAT), an adjunct system attached to the Electronic Voting Machine (EVM), and the provision of counting VVPAT tallies from five random polling booths in every Assembly constituency to be matched with the EVM vote-count, have not assuaged critics of the use of EVMs in Indian elections. Some suggest the fact that the process could be more transparent if a machine audit-trail of all the commands that are executed is maintained in the system, beyond just the votes recorded in the EVM’s ballot unit and the printed slips in the VVPATs, allowing for an audit to rule out any malicious code. This could indeed make the system more robust and be considered as an upgrade to the existing machines. Others suggest that the use of VVPATs has introduced potential vulnerabilities that did not exist with the standalone nature of EVMs and the technical and administrative safeguards that undergirded the legacy system . This too could be addressed by reworking the safeguards to ensure that the VVPAT-combined systems are as secure and foolproof as the standalone EVMs were. But incomprehensible is the critique from many, including political parties such as the Congress, that only a 100% recount of all VVPATs would suffice, instead of the current method of sampling the number of recounts, in order to have full transparency . The Supreme Court of India has now listed a series of petitions related to this demand.
Despite Cassandra-like pronouncements about malpractices and EVM-hacking, there has been no proof of any actual tampering of EVMs so far . While EVMs, as any machine is bound to, have suffered glitches, and promptly been replaced in the event of machine failures, the critique that they are prone to hacking or manipulation despite the existing technical and administrative safeguards has not been made with any actual proof. The sample counting of VVPATs, across both the general election in 2019 and several other Assembly elections for instance , which allow the voter to confirm whether the mandate matches the choice made on the EVM, have shown that the mismatch between the VVPAT recount and the EVM count has been minuscule — an outcome of trivial errors such as non-deletion of mock polls in the machine before the voting process or errors in manual recording of the final count from the machine. An increase in the recount sample to make it more statistically significant, by making the selected number of Assemblies specific to each State/Union Territory based on the size of the province , or, simply to increase the recount sample in seats where the margin of victory is narrow (say, less than 1% of the overall votes) could be solutions. But to insist on a full recount seems an overkill and a clear lack of trust in the EVM itself.