Originally Posted by Chris

It’s interesting how it was massive sample MRP polls, supposedly with a crystal ball capable of predicting results at constituency level, that were predicting Con/LD parity and have now apparently been shown to have fallen well short of their promised accuracy.