// At the last UK general election we were told by some polls that magic grandpa was heading for number 10 ! //
This isn't true, though. The last polls that showed any kind of Labour lead, prior to the 2019 election, were in July, when Boris Johnson had been PM for about three days. Before then, there was a two-month period where LibDems, The Brexit Party, Labour and the Tories had all led at some point -- which is actually quite reasonable, considering the chaos at the time. After that, there was a period of polling that saw the Tory lead steadily increase, so that by the time of the election, the polling average was something like a 10-point lead for the Tories, which matched the election result fairly well.
Why not stick to a factual reason to be dubious of polling, rather than a made-up one? You could, for example, cite the 2016 US election -- although, even then, the scale of the polling error has been greatly exaggerated.