A link posted on Twitter led me to investigate this. This is the link. I don’t think the maths is accurate, or that Wikipedia is a good source for this kind of thing, but luckily, I was able to locate good sources.
Warning: I am no election expert, nor am I all that great at maths. Verify rebut or endorse at your own risk.
The vote share in Uttar Pradesh as per http://ceouttarpradesh.nic.in/GE2014/VoterTurnout.pdf
Scroll right down to the bottom for total votes cast. I read the number 80926378 .
Now head over to the election commission website for the UP page and scroll down to the piechart for vote share with list on left, with number of votes cast. The numbers with dots in front of them expand to show full number when you hover your mouse over them, and you get the following results:
| Party | Votes% | Vote Count |
| BJP | 42.30% | 34318576 |
| SP | 22.20% | 17988792 |
| BSP | 19.60% | 15914019 |
| INC | 7.50% | 6061236 |
| IND | 1.70% | 1414863 |
| AAAP | 1.00% | 821301 |
| AD | 1.00% | 812325 |
| RLD | 0.90% | 689409 |
| NOTA | 0.70% | 592406 |
| PECP | 0.60% | 500482 |
| D | 0.40% | 354578 |
| BMUP | 0.20% | 153821 |
| CPI | 0.20% | 129471 |
| Total | 79751279 |
That is looking like 1175099 LESS votes recorded.
Considering that several crucial victory margins can be explained by such numbers, if my maths is right, it looks like a re-poll, right?
Erm yes. EVM fraud is easily possible if you know how to hack it. Apparently EVM fraud likely happened in 2009 as well.
There are some explanations coming up, like the EC not providing vote share of “others” – probably independents, which could explain the difference. Others are saying that the difference could be the postal ballot. But going through the data per constituency disproves both.
The “others” being the difference gets disproved when you tally votes per constituency, where the number of votes each candidates gets is explicit and without room for omissions. We find that every constituency has a difference in number of votes cast and counted.
The postal ballot theory gets disproved because the votes counted don’t always exceed votes cast. A person voting by ballot can be understood. There is no explanation for votes already cast going missing – which is the difference in numbers on a state level, which goes above 11 lakh.
I will update data or all constituencies, but this initial sample is good enough to see what I mean.
| Constituency | Votes cast | Votes recorded | Difference |
| Saharanpur | 1193111 | 1194308 | 1197 |
| Kairana | 1117330 | 1119324 | 1994 |
| Muzaffarnagar | 1103916 | 1107434 | 3518 |
| Bijnor | 1058085 | 1060346 | 2261 |
| Nagina | 942326 | 942196 | -130 |
| Moradabad | 1126242 | 1127965 | 1723 |
| Rampur | 958585 | 956389 | -2196 |
| Sambhal | 1056322 | 1057002 | 680 |
| Amroha | 1095216 | 1095860 | 644 |
| Meerut | 1112424 | 1113384 | 960 |
| Baghpat | 1003533 | 1004263 | 730 |