Loksabha Elections 2014: My Chinese arithmetic

Yesterday all the phases of polls closed in the country and all the television channels were out with their voodoo science that they call exit polls. Since everybody seem to be doing it, I thought I should also do it. So here it goes.

Exit polls have been wrong in pretty much all the elections in past. Here is a snipped of how exit polls have performed in past.

As we can see in the chart above, there has always been a significant difference in the exit poll results and actual results. Barring 1998, BJP has been always over-estimated in exit polls while congress has always been underestimated in exit polls. So, let's look at kind of error that has been seen in exit polls and actual results.

In the chart above, select Errors for Size, Actual for Y Axis and Time for X axis and click on Play and then we see that both BJP and congress seem to have increasingly higher errors over the years with BJP being over-estimated and congress underestimated.


So, I have decided that I am going to take the average underestimation for congress and BJP from 2004 and 2009 and try to predict the result for this year.

• Average BJP overestimation = 27%
• Average Congress underestimation = 21%
• Average seats predicted for BJP in this year's exit polls = 273
• Average seats perdicted for congress in this year's exit polls = 110

So my predictions for BJP and congress are as follows.