Reality check please
What seems to be being missed is the actual, genuine, real, mathematical change in votes and how disproportionately that has affected the final outcome.
The Tories gained just 0.8% increase in vote share (2010 to 2015), with Labour outperforming them with a 1.5% INCREASE in vote share. This could be summarised as - not a great deal of difference for the 2 main parties between 2010 and 2015. However, due to our utterly outdated FPTP system this means that there has been a major shift in final Seat distribution. This point is constantly missed. Ultimately, the system has created this result, not really the electorate.In short: Cameron failed to convince any meaningful number of additional voters that his programme of systematic destruction of the UK was a good idea, likewise Milliband was unable to convince any more voters that his version of austerity was much different to the current one. Lies, damn lies and statistics - sums up UK politics.
On this point:
"With 650 consituencies, the only way to get accurate polls is to massively increase the sample. If you ask 6500 people, that's only 10 per constituency (if you even bother to select people from each)."
Err no, you only need poll the 50 - 100 seats where any kind of change is likely. There is literally zero point in polling people in Sunderland or Guildford, for instance, about their voting habits as those seats will 'never' change. Focus on the few that matter. And Scotland.