“We got it right!” those four magical words that we can now proudly proclaim. These were difficult elections to predict and project, even those giant media companies and survey organizations couldn’t fully comprehend many aspects of these set of assembly elections, but we at 5Forty3 with near zero resources got most aspects of four state elections right;
- Rajasthan: 5Forty3 had clearly predicted that this is a wave election, much before the exit polls had even arrived into the horizon. Also, we had made it a point that seat projection for a wave election like in Rajasthan could be anybody’s guess since it could be anything above 120.
- Chhattisgarh was a state that we had all along projected as “touch and go” with slight edge for the BJP and we have been proven 100% accurate.
- In Madhya Pradesh, once again 5Forty3 had projected the return of the BJP government, but the only error was that we had predicted a better performance by the Congress and had over-estimated the strength of a party that is completely moribund.
- In Delhi again, we were almost accurate in our projection of BJP leading the pack, with AAP and Congress vying for the number 2 position. Our vote-share projections based on 3rd party exit poll data was nearly perfect, but for a minor over-estimation of Congress and under-estimation of the new entrant AAP, which has led to this humiliating defeat of the Congress.
At the end of the day we at 5Forty3 strongly believe that predicting seats in Indian election is nothing but astrology and actual psephology is about assessing sub-regional undercurrents and capturing political trends. To that extent, 5Forty3 has been a resounding success in these set of assembly elections. Let us try and elucidate this aspect by highlighting 6 specific points;
- Predicting the rout of Bastar for BJP in the phase 1 election of Chhattisgarh was unique because no other media house or opinion poll survey had made that projection at that point. We were the first to capture that signal clearly and even elucidated it with the specific examples of how many strongholds of BJP were vulnerable using Turnout Differential Factor
- 5Forty3 was the only place on earth which captured a very important political development of sending hundreds of Muslim families on a paid pilgrimage to Ajmer by some BJP supporting organizations in Raipur-Raigarh-Bilaspur belt. In the end analysis this aspect became extremely crucial for a closely contested state election of Chhattisgarh as BJP won at least 4 seats in this region by very small margins.
- Projecting a wave election in Rajasthan based on ground reports, Turnout Differential Factor and exit polls data
- In Madhya Pradesh, in an otherwise wave election where BJP has won almost everything, it is in the old stronghold of Baghelkhand that surprisingly BJP has lost many seats, including top leaders and ministers. We at 5Forty3 were almost the only place in this whole world who had projected this particular upheaval! (will write more about it in another detailed piece) It is just that Congress losses were huge in all other regions so it was decimated in the state.
- In Madhya Pradesh, clearly projecting that Jyotiraditya Scindhia would have no impact on Gwalior-Chambal region using TDF was another feather in our cap
- In Delhi, we specifically targeted New Delhi constituency and projected that the CM would lose the seat. Not only that, we projected clearly that the race is incredibly tight for the number two position between BJP and Congress. This has been borne out by the results.
Today we thank all our supporters and well-wishers for their continuous support. Those who have been ridiculing us, we thank even more, for it is their criticism that has provoked us to improve our mathematical models and electoral systems to achieve near perfection. We will continue to reinvent and improve our tools to change the way elections are analyzed in India. Our aim is to become the final word in Indian election analysis.
A Note on Turnout Differential Factor (TDF)
It was after the Himachal Pradesh election last year – wherein a few specific assembly seats of Mandi and Kangra had gone against our assessment – we started to work on a specific tool to determine these minor changes at the local level that can be captured using some available data points. It is now well established that BJP lost a few seats in Kangra and Mandi due to clear internal sabotage, but the challenge was how does an election analysis account for such shenanigans? After a lot of back-testing and reverse analysis, we zeroed in on what we now term as the Turnout Differential Factor (TDF).
TDF as a tool is simple in its output wherein we compare localized turnout trends to differentiate between general turnout trends based on different party representations. But the most vital part of TDF is to identify the specific assembly segments for which we use a mathematical model that assigns adequate weightage to various criteria – past margin, past 3 election results, demographics, current election ground reports, rebel factors etc. Thus TDF can be used as a leading indicator to determine the overall trends of an election with or without exit poll data.
TDF has been used openly for the first time in these assembly elections and we have shown how it clearly signaled a wave election in Rajasthan and Delhi on the one hand and the sub-regional disparities of Madhya Pradesh and Chhattisgarh on the other hand. We are now working on a newer version of TDF known as Modified Aggregate Turnout Differential Factor which is even more precise as our test results for internal assessments have shown in these set of elections. We hope to employ MATDF as a much superior leading indicator in the upcoming parliamentary elections.
Epilogue: 5Forty3 was launched as a special purpose vehicle just to analyze these 4 assembly elections. We are not sure if this platform should be continued after today. We thank all of those who visited our blog in the last one month and patronized us.