So, who're you gonna vote for?

In their recently released, neatly summarised textbook, The Verdict: Decoding India's Elections, authors Prannoy Roy (top right) and Dorab Sopariwala (bottom right) speak of what they call the universal smile

Perhaps like you, for the past fortnight since dates for 2019 General Elections were announced, whenever in a chatty mood in the back of a rick/cab, I've casually straw-polled the driver on which way he's currently swinging at the hustings. Let alone what route he's taking me through that I have little control over. Either way, the metaphor is complete.

Also, while my sample size belongs to Mumbai, the metropolis probably holds half its roots in several parts of India, notably Uttar Pradesh, which for better or worse, inevitably determines the hawa, or wind/wave, carrying a party/coalition to power at the Centre.

In their recently released, neatly summarised textbook, The Verdict: Decoding India's Elections, authors Prannoy Roy and Dorab Sopariwala -- who, along with data-scientist Ashok Lahiri, pioneered election opinion polls in India, with the first independent, in-depth survey carried out in 1980 -- speak of what they call the "universal smile".

That's what mysteriously greets pollsters when they privately pry into people's poll preferences, which is primarily done in four ways-at the "street corner," "door-to-door" (before, or after elections), outside polling booths (exit polls). Or now, more easily -- but therefore largely shadily -- on social media!

Surely opinion polls would've taken a global beating especially after US elections, when Hillary Clinton's victory was a given, but the Donald trumped her in a complex web of what American elections are, anyway!

That said, as Roy-Sopariwala point out, of 833 electoral opinion polls conducted over 40 years in India, 75 per cent have remarkably managed to predict the winner -- this being the "strike rate", as against accurately forecasting number of seats, which is toughest, given statistical errors, when the sample size is unavoidably low -- running into tens of thousands, even within which, people's spot-responses may not match final outcome.

Furthermore though, at the Lok Sabha, if you set aside one election, the strike-rate of opinion polls is a phenomenal 97 per cent! That one outlier election, defying analysis, that everyone got wrong, was in 2004. This is when NDA (led by BJP) had finished its maiden full-term in power. The current election draws to a close NDA's second complete-stint.

Since the first in 1952 to the last in December, 2018, India has had 392 elections -- 16 to Lok Sabha, and 376 to state assemblies. Plodding through results in all, Roy-Sopariwala reveal three distinctive phases to explain voting behaviour/trends.

The first phase, 1952-77, was about 'pro-incumbency', where the electorate pretty much uniformly voted governments back to power. The second, between 1977 and 2002, was marked by "anti-incumbency" when the voters preferred to rotate victors. The next, or the current one, 2002 onwards, is what the authors call '50-50', which is the sort of chance the voter gives to a government to retain power, presumably if it's performed well.

Frankly, I read through Roy-Sopariwala's book, slowly, word for word, like a syllabus for an exam to crack. And that's because, these two gentlemen, through their television analyses (on NDTV, and even before that, on DD), are entirely responsible for making elections (at any rate, the results day) at once feverishly exciting, and phenomenally insightful for generations, who've rightly deemed electoral democracy as India's most-loved sport!

And why's that? Because India isn't a fair, fail-safe electoral democracy in the real sense. Its first-past-the-post system makes every voting pattern/polarisation/personality/issue open to interpretation, until the final word is in. For, rather than a majority, you need only one vote more than the candidate placed second, to rule/represent the rest.

Extrapolate that single constituency into a deeply heterogeneous population/nation, and so 31 per cent votes for BJP translates into 52 per cent seats; likewise, 38 per cent votes for BJP-led NDA means 61 per cent, or 336/545 seats (unimaginably comfortable majority), in the 2014 parliamentary elections. With 19 per cent votes, Congress enters with 8 per cent (or 44) Lok Sabha members!

Much hence depends on terms like "swing", "Index of Opposition Unity" (the strength of the rival coalition), booth-management, "dark art of voter suppression" that Roy-Sopariwala introduce and clinically examine in their book.

Wherein currently, they say, 1 per cent swing (in favour of a party, and so subtracting the same from the rivals) can lead to 15 Lok Sabha seats, straight. And it can always go either way, although empirically tabulated, there are 77 per cent chances of a landslide win for a party in a state, at the Lok Sabha elections, which when totaled, decides who leads India for the next five years.

I've been getting a mix of smiles and indifferent scowls during my stray, cab-ride straw-polls at the moment. But here's one thing we know. And this is from David Mamet, legendary Hollywood screenwriter, since that's another place that works singularly on profiting from gauging diverse public moods/opinions. Of people, Mamet is convinced: "Individually, they're idiots. But collectively, they're genius -- can't fool them; never met a dumb audience." Yup, think so, too. Okay, hope so for sure!

Mayank Shekhar attempts to make sense of mass culture. He tweets @mayankw14 Send your feedback to mailbag@mid-day.com

Catch up on all the latest Mumbai news, crime news, current affairs, and also a complete guide on Mumbai from food to things to do and events across the city here. Also download the new mid-day Android and iOS apps to get latest updates