“On the morning that we were to make our first-ever appearance covering election results on the national television network, Doordarshan, a senior editor of a leading newspaper wrote an article across the bottom spread of the front page ridiculing our 366-seat forecast published in India Today, scorning what he called our new-fangled statistical polling and election analysis. He went on to say there was no way that the Congress would get anything like 366 seats, and that it would not surpass 225 seats” writes Prannoy Roy about 1984 election predictions and the media reaction to the same in The Verdict, his otherwise pedestrian book with no new insights on Indian elections and psephology. 

What Prannoy is underling there is the fact that most of our punditry regarding election analysis is post-facto in nature. It is indeed true that the entire editorial class, including top journalists and political pundits did not expect even in their wildest dreams about a novice like Rajiv Gandhi getting a 400-seat majority in 1984. Yet, almost each one of them pontificates today about the ‘1984 wave election’ as if they knew it like the back of their hand.

There is something eerily similar about 2019 elections too. Most top journalists and editorial class have already called this a “wave-less” election. The punditry is all about opposition alliances, agrarian distress, jobs et al. but when we meet ordinary voters, most of them are quietly determined to give ‘Modi another chance’. Today’s political structure too is quite similar to 1984. The gap between the two main national parties is huge, even as the main opposition party is not even the position to occupy the space of the “leader of the opposition”. The main challenge to the Prime Minister is expected to come from regional parties, especially from the south – remember that NTR was the main opposition to Congress and Rajiv in 1984.

With opposition in disarray and Modi being the sole contender for the post of the Prime Minister, 2019 turning into landslide was always on the cards, but when we built our models for the 2019 elections, we deliberately tweaked them to be conservative with pre-poll numbers. India like all other parts of the world in this generation has a much smaller attention span than at any given point of time in the history, so it is always better to rely on real-time or post-poll data than the opinions which are even just a week old.


To be sure, our models are currently showing 343 seats for the NDA with only about 23 seats in the “tossup” category. In terms of seats, it is hardly an increase of about 7 seats from the 2014 peak. What is surprising though is the swing in favour of the NDA which currently stands at 7% overall. A 7% overall positive swing from the already high 39% in 2014 should give a lot more seats for the NDA in a normal course of political trend, but we are using two different models for both the exercises.

The vote-share model is the macro-model that just calculates the overall support for a political party based on the latest available data points, it is therefore quite stable and only changes with major underlying geographies changing course (Rajasthan and Madhya Pradesh in today’s case). The MAPi – Micro Analytics Projections (INTELLIGENCE) – model actually uses polling booth level data analytics to calculate the number of seats that each party is currently leading. MAPi also has a stack algorithm which redistributes numbers as newer and more data keeps coming in from analogous polling booths. For example, if in phase 3 the behavioral patterns of Jatav dominated polling booths in Etah or Firozabad assembly segments could not conclude decisively as to which way majority of them voted, then the algorithm checks for similar Jatav dominated polling booths in Kannauj in phase 4 and adds a factorial index to the previous data. This is why the “tossups” keep changing with each phase as more and more data is added on to our servers.

In simple layman terms, the logic behind these algorithms is that eventually both macro and micro projections should meet at the same point without losing the inherent sensitivity to shifting numbers. Therefore, if the BJP and NDA sustain this vote-share of around 45-46%, then we believe the overall seat tally will also eventually hit 370 elevating the 2019 elections to the exalted status of a mega-wave of 1984.

The one key political difference between 1984 and 2019 is that in 84 it was sympathy that was the driving force behind Rajiv Gandhi’s success, whereas today it would be the performance of the Modi government in the last 5 years that is driving the wave. If one were to ignore 1984 as an election built on a tidal wave of sympathy, then the last time a single party majority government was reelected was way back in 1971 – it has taken nearly 50 years for India to discover a true leader who can walk the talk and bring good governance as the main theme of politics rather than caste, secularism, socialism, dynasty and many such ill-conceived concepts.

This stark contrast is most visible in the state of Bengal in the 2019 round of elections. An almost failed state is slowly discovering the courage to stand up and be counted. Bengali voters grit and courage is praiseworthy, for entire villages have been threatened by ruling party goons and yet a large number of them are going to the polling booths and quietly but effectively doing their jobs. The problems though with political projections in Bengal are threefold.

  • It would be difficult to put a number to the voters who would be intimidated before and on polling day and miss out on exercising their franchise.
  • There are still about 9000 (nine thousand) polling booths across the state where neither BJP nor any of the opposition has booth-level infrastructure which enables TMC cadre to run a one-man show.
  • The silent Hindu anger is also a factor this time which is working against Mamata in a big way. For all we know, our models must be underestimating the support for the BJP.


Under these circumstances, it is best to remain conservative with numbers. Yet, our models are suggesting leads in 3 seats for BJP in today’s round of polling till about 6 PM – some parts of Bengal data come in a bit late, so we will need more time to fully analyze the same. Currently, there are also 2 seats classified as “close tossups” which should be clearer with more numbers. With today’s trends what is certain about Bengal is that BJP is surely heading for a double-digit tally notwithstanding the Mamata bullying tactics. How big will the haul be, we will know as we go along. 

Bihar is one state that is consistently showing an NDA sweep in almost every phase. It is also along expected lines, for even earlier when Nitish and BJP fought the elections together, they would end up sweeping 90-95% of the seats. Today’s data too (whatever is available till 6 PM, Bihar data reporting is again a bit slower) suggests a sweep for the NDA once again. Even in the much publicized Begusarai contest, Bhumihar strongman of BJP, Giriraj Singh is coasting along to victory and the much touted Kanhaiya Kumar is likely to end at the 3rd position. In the neighboring Jharkhand too BJP is continuing its same good run in all the 3 seats that went to polls today.


From the east to the west, Rajasthan and Maharashtra have shown remarkable saffron resilience in today’s polling, as we have already observed in the post-noon analysis. Both the states could end up showing a complete sweep for BJP and NDA, at least as far as today’s polling is concerned. This was not how the script should have unfolded for the Congress party, especially in Rajasthan, but the voter anger there may actually sink the party further beyond the 2014 abyss – there were enough signals of the same today. 

Chief Minister Ashok Gehlot’s son was trailing throughout the day in Jodhpur and that could be disastrous for the state Congress government which is surviving on a wafer thin majority. Even in deputy CM Sachin Piolt’s home district, his own community Gujjar voters voted overwhelmingly for the BJP today. The only two seats where there was some contest were Kota and Barmer, the former has shown a small lead for BJP while the latter has been classified as “tossup” by MAPi in Rajasthan.

In Maharashtra too, the situation was pretty similar. As per our numbers till 5 PM, even the Pawar scion, Parth is seen to be trailing in Maval LS seat. Among the 8 seats of greater Mumbai region too, the only seat which saw a keen contest was Mumbai South where Shiv Sena still enjoys a slight edge. Urmila Matondkar is likely to lose the Mumbai north seat by a record margin, despite all the theatrics and media buzz. The two Shiv Sena strongholds of Thane and Kalyan were virtual no-contests which again proved that Raj Thackeray is all noise and no substance.

In the most crucial state of Uttar Pradesh too, this was a good phase for the BJP. Late afternoon data was suggesting clear leads for BJP in 8 seats, while SP was ahead only in the one seat of Kannauj (this is subject to change as full data gets amalgamated to our servers). Four other seats have been currently classified as “Tossup”, and among them, two because we still have only about 25% of the targeted data. Also, our turnout differential models are not yet populated with adequate data (it takes almost 1 full day at times for the data to reach us), so it would be premature to analyze the state at this stage.

Phase 4 has lived up to its billing of being the crucial day in the 2019 season. We can now safely say that Modi is coming back with a bigger majority, for the opposition stands very little chance to reverse these trends now.

[We will be doing in depth and detailed analysis of this phase in the next 2-3 days]