A close look at the 2024 Andhra Pradesh assembly elections polling data uncovers what could be a ‘mischief’ behind the ‘miracle’ which led to the ruling alliance in the sunrise state winning 164 out of the 175 assembly seats.All the figures given here are taken from data released by the Election Commission of India, the Andhra Pradesh Chief Electoral Officer, their official notifications and press announcements and releases. On the evening of May 13, 2024, the CEO of Andhra Pradesh announced to the media that 68.04% votes were polled as of 5 pm. He said a “large number” of people were still waiting to cast their votes and the poll percentage could further increase. Screengrab of a report on The New Indian Express: https://www.newindianexpress.com/states/andhra-pradesh/2024/May/14/cliffhanger-in-andhra-pradesh-as-voter-turnout-may-touch-80As per the ECI press release No. ECI/PN/82/2024 dated May 13, 2024, the voter turnout was 68.12% at 8 pm.The Election Commission’s 8 pm press release on May 13, 2024. Thereafter, at 11:45 pm on the same day, May 13, 2024, the voter turnout rose to 76.50%, according to ECI press release No.ECI/PN/84/2024 dated May 13, 2024. The Election Commission’s 11.45 pm press note on May 13, 2024.On May 17, 2024, the Andhra Pradesh polling percentage was further revised upwards to 80.66% through the release ECI/PN/89/2024, dated May 17, 2024. The Election Commission’s press release on May 17, 2024.The CEO of AP put the figure at 81.86% and the ECI’s Statistical report put the final figure at 81.79%. This includes EVM votes and postal ballots in both the figures. The Election Commission’s ‘Statistical Highlights’.If we scrutinise these figures closely, the mischief behind the AP miracle unfolds clearly.A voter turnout of 68.04% (as per the CEO) till 5 pm on the polling day points to an average polling of 6.8% per hour. By then, 10 hours of polling had taken place.It increased to 68.12% by 8 pm. That means the increase until 8 pm, i.e. in three hours, was 0.08%.However, the polling percentage went up to 76.50% as of 11.45 pm on the same day, the day of polling. This was an increase of 8.38% between 5 pm and 11.45pm. And the CEO declared that polling was still in progress in 3,500 booths. The Election Commission’s 11.45 pm press note on May 13, 2024.Screengrab from The Hindu’s report: https://www.thehindu.com/elections/andhra-pradesh-assembly/andhra-pradesh-registers-a-polling-percentage-of-8186-the-highest-so-far-in-the-country-says-chief-electoral-officer/article68178365.eceVotes polled in these 3,500 booths after midnight are worked out to have raised the Voter Turnout (VTR) by 4.16% and brought it to 80.66% as per ECI press release No. ECI/PN/89/2024 dated May 17, 2024. This doesn’t include postal ballots.The Election Commission’s press note on May 17, 2024.If you convert that percentage into actual numbers, it means that in 3,500 booths, after midnight, a staggering 17,19,482 votes were cast. That means, on an average, a little over 491 votes per booth were cast after midnight. Obviously, it cannot be exactly 491 votes in every booth. But if the number was less than 491 votes in some booths, then it had to be more than 491 votes in some other booths for the polling to average at 491 votes per booth.The average number of votes cast between 7 am and 5 pm on the polling day in all the 46,389 booths across the state was 60.7 votes per hour. That itself can be termed as very brisk, considering that voters were casting two votes each – one for Lok Sabha and the second for the assembly. If, however, so many votes were yet to be cast after 11.45 pm in these 3,500 booths, then polling in those booths must have been extraordinarily slow between 7 am and 11.45 pm. It could not have been more than 24 votes per hour. If one does not consider this extraordinarily slow pace of voting but the state’s average pace of voting, polling in these 3,500 booths should have continued until 7 am or 8 am on the next day. And in a significant number of booths, where more than 491 votes were cast, it should have continued until 10 am or 11 am on the next day. But did it? It did not.The CEO of AP in his press conference of May 14, 2024, said clearly that the last vote was cast at around 2 am on May 14. This means that polling concluded in 2 hours and 15 minutes after the update at 11.45 pm on May 13, 2024. That means in 2 hours 15 minutes, as many as 17,19,482 votes were cast. Screengrab from The Hindu’s report: https://www.thehindu.com/elections/andhra-pradesh-assembly/andhra-pradesh-registers-a-polling-percentage-of-8186-the-highest-so-far-in-the-country-says-chief-electoral-officer/article68178365.eceIf we take the average of only one minute for one vote to be manually cast, for 491 votes to be cast in a booth, it would have taken 491 minutes. That would be at least about 8 hours. If it took 2 minutes per vote, that would be about 16 hours. It is highly probable that the 17,19,482 were not spread uniformly across all the booths. If in some booths it’s much below 491, it ought to be much higher than 491 in some other booths. In those booths that had 491 or more voters waiting to cast their votes, polling would have continued much longer. The booths that had less than 491 voters would have completed polling much earlier. But the CEO said on record that “the last vote was exercised at around 2:00 am on May 14.”That means any booths that had 491 or more waiting voters could not have concluded polling by 2 am. If, however, the polling indeed concluded by 2 am on May 14, that could only mean the highest number of voters waiting at any booth after the official closing time of polling ought to be much smaller than 491. That could only mean that the number of waiting voters at any booth could not have been more than, say, between 100 and 135. Then, how does that 2 am timeline announced by AP CEO square with the number 17,19,482 who were claimed to have voted between 11:45 pm on May 13, 2024 and 2 am on May 14, 2024?If we take the AP CEO’s timeline of 2 am, a booth that had 491 voters would have witnessed a staggeringly high rate of polling: nearly 3.6 voters per minute. That works out to be at least one voter voting every 20 seconds continuously. Remember that Andhra Pradesh had simultaneous polls for the assembly and the Lok Sabha. That means every voter had to get their name checked, signed or given their thumb impression, their finger inked, cast two votes, physically moving from one voting compartment to the other and come out – all in 20 seconds. In these 20 seconds the unavoidable dead time in the polling station is 14 seconds – after each vote is cast, 7 seconds have to pass in each compartment for the VVPAT slip to come out. That itself will take away 14 seconds. That means for the voter to get their name verified in the electoral roll, to sign the register or give a thumb impression, get their finger inked, and move from there to the first compartment, cast the first vote and then move to the next compartment, cast the second and come out – for all this only 6 seconds are left. This sheer process itself makes this extraordinary feat physically impossible for 17,19,482 votes to have been cast just in 2 hours and 15 minutes – before 2:00 am on 14th May. Only the ECI can explain how it made this extraordinary feat physically possible. ECI’s Handbook for Presiding Officers 2023 (Section 7.1.1) prescribes that voters who already joined the queue by the prescribed time of closing have the right to exercise their franchise. However, the following measures have to be taken by the polling officials: the gate of the polling station premises has to be closed by either a police officer or one of the polling staff. The waiting voters have to be given tokens – the voter at the tail of the line to be given token bearing number 1 and the one at the head of the queue to be given token bearing the last number. Those tokens have to be collected when voters carrying them come to vote and all the tokens to be preserved for future verification. In the guidelines on “Videography of Critical Events & in Polling Stations”, it is clearly laid down that among other things like positioning of voting compartment and presence of polling agents, video should be made of: “Voters waiting outside at the close of scheduled hour of poll and the last voter in queue.”That means the entire queue has to be video-graphed making sure that the faces of the waiting voters are clearly “discernible”. All these records need to be preserved for future verification. Equally important is that the Presiding Officer of the polling station has to record all this process in the Presiding Officer’s diary. The diary also has to be available for verification. Is the ECI ready to be challenged on these counts and open the records for inspection? It should, if it has nothing to hide and there’s no mischief. The AP miracle becomes clearer when one notices that the increase in polling was only 0.08% in three hours between 5 pm and 8 pm. But it increased by an extraordinary 8.38% as of 11.45 pm on the polling day between 8 pm and 11.45 pm, that is, in the next 3 hours and 45 minutes. But this extraordinary and miraculous increase doesn’t end there. On May 17, 2024, four full days after the polling, ECI released a new figure which increased the vote percentage by another 4.16%. The Election Commission’s press note on May 17, 2024.Thus bringing the difference between the polling percentage from 68.04% (as per the CEO) or 68.12% (as per ECI) to 81.86 (as per AP CEO) or 80.66% as per the ECI press release on May 17, 2025 or 81.79%, as per ECI’s Statistical Highlights.Here is the summary of the miraculous increases in voting percentage:On May 13, 2024 at 5 pm as per CEO of AP, voter turnout was 68.04%On May 13, 2024 at 8 pm as per ECI, voter turnout was 68.12%On May 13, 2024 at 11:45 pm as per ECI, voter turnout was 76.50%On May 17, 2024 as per CEO of AP, voter turnout was 81.86%. This includes 1.20% postal ballots.On May 17, 2024 as per ECI, voter turnout was 80.66%There are discrepancies between the figures of CEO of AP and those of the ECI, in both EVM votes as well as postal ballots. Strangely, the Statistical Highlights by ECI has a completely different number. It puts the total polling figure at 81.79%. It puts the postal ballots at 1.23% whereas the AP CEO’s figure for postal ballots is 1.2%. The contradictions between the data announced by the CEO of AP and the ECI is another saga that needs a detailed and separate narration. The Election Commission’s ‘Statistical Highlights’.Let’s now look at the time lines and the quantum of increases:Between 5 pm and 8 pm on May 13, 2024, that is in 3 hours, the increase was 0.08%, from 68.04% to 68.12%. In gross numbers, that is a mere 33,064 votes. Between 8 pm and 11:45 pm on the same day, that is, in 3 hours and 45 minutes, the increase was 8.38%, from 68.12% to 76.50%. In gross numbers, that is a huge 34,63,767. The average polling in this parcel of time works out to between 74 and 75 votes per polling station. Of course, understandably, the spread of the votes could not have been perfectly uniform. But do keep in mind each voter had cast two votes, one for Assembly and another for Loksabha. By May 17, 2024, four days after the close of polls, there was a further increase of either 5.16% or 4.16%, depending on whether you choose the AP CEO number or the ECI number. That is, from 76.50% to 81.66% according to the CEO of AP, and to 80.66%, if you opt for the ECI’s number, excluding postal ballots in both the figures. In gross numbers that works out to a further increase of 17,19,482, even using the lower figure of of 80.66% of the ECI, instead of that of the AP CEO, because ECI is indeed the final authority. All this brings the total increase of gross votes to 51,83,249 votes, with the extraordinary overall increase of 12.54% voting between the voting percentage figure announced at 5 pm on (the polling day) on May 13, 2024 and on May 17, 2024 on the day when the final polling percentage was announced. Therefore, the overall increase in gross voting comes to: 51,83,249 votes. Fifty one lakh eighty three thousand two hundred and forty nine votes, in words. If this figure is equally distributed among all the 175 constituencies in the state, it works out to be 29,618 votes per seat. A uniform spread of this increase is highly improbable. That can be ruled out. If the increases, however, are in select booths of targeted constituencies, then these increases would give a startling knockout advantage to some candidates/parties while mercilessly slaughtering the other candidates/parties. This increase, therefore, is behind the glorious victories of some candidates/parties and humiliating defeats of others in Andhra Pradesh assembly elections 2024.Interestingly, this extraordinary feat has been hiding all along in plain sight in the ECI/CEO-AP data and their press releases/announcements. Now the big questions are:(1) Are the victors in AP merely passive, unknowing recipients of this miracle’s bonanza? (2) Are the vanquished in the state just innocent victims of this extraordinarily staggering increase in poll percentage that led to their defeat?Clearly, the integrity of the election process in Andhra Pradesh appears to be severely compromised. What is not clear, however, is who all are complicit in it. The Andhra Pradesh election mandate is, therefore, highly questionable. The discrepancies listed above give room for suspicion that it could be a stolen mandate much in line with the doubtful mandate of 2024 Lok Sabha elections. Only proper and credible responses to the issues raised here could set these apprehensions at rest. Until then, doubts linger. Dr Parakala Prabhakar is a political economist and author of The Crooked Timber of New India.