Exit polls: Why do we need them

The tyranny of television rating points has meant that exit polls are now part of the great Indian election tamasha. Let me make an honest confession: I am a firm believer in the CSDS method of post polls, which are done at least a day after the votes have been cast and where the pollsters actually go to people's homes to get a more exhaustive survey done.

Ninety per cent of the polls we have done with this method have proven right. I am more sceptical of exit polls, especially when some of them give "results" with a ring of authority barely minutes after the voting has concluded. Can such polls accurately measure the final result in a country as large and diverse as India where voting percentages build up till the very last hour? Unlikely.

A good poll can capture a trend, but it is not a maths sum that can have hundred per cent accuracy. And yet, we present them with great enthusiasm in the belief that the viewer wants to know who has won at least 72 hours before the actual results.

Yes, I am pretty certain that the BJP is winning Haryana and Maharashtra. But that is my gut political instinct, based on my analysis and ground feedback, not necessarily on what the polls say. Then, there is the case of today's Chanakya, the rather intriguing nom de plume of a certain VK Bajaj, who I first met in 2003 ahead of the Rajasthan Assembly elections when he was an amateur enthusiast and who since then has built a rather formidable reputation as a pollster.

The fact is, today's Chanakya, like all pollsters, has got some elections wrong too (he still doesn't reveal his sample size or methodology). But in psephology you are only as good as your last poll and since he predicted the Lok Sabha election result accurately, he is the flavour of the season.

Who knows, come Sunday, he could be right again. Or maybe, Cicero or C Voter will be more accurate this time. Like so much else in life, there is an element of chance to all exit polls. A classic case of Jo Jeeta Wohi Sikandar!