I don't thing that the latest explainer from the Met Office they've published to Facebook and Twitter is particularly accurate or fair to the NWP scientists when it describes how difficult it is to forecast thunderstorms.
I'm not arguing that it's possible to do with any degree of accuracy, it's not, but if the pan in the analogy represents the British Isles, we can like yesterday know fairly accurately the area that they will occur in. So taking the analogy one step forward we know the bubbles will only appear in the area below 'EXACTLY?' as they did yesterday. Phil Avery may be able to say we can't forecast were thunderstorm will occur, but from observations we know exactly where they are occurring, and just how much rain is falling from them.
I think the problem in being slow to identify thunderstorms and flash flooding events is as much to do weather presenters not monitoring observational data than it is the NWP they seem to have implicit faith in. I noticed that a number of times this week how in the BBC lunchtime forecast the first graphic was for 6 pm that evening!
You could argue that it's ultimately the Met Office responsibility to do this monitoring and not the BBC's, but flash flooding is a news event, and the presenters at the BBC these days are as much journalists as they are meteorologists, there to issue advice about the latest ongoing severe weather event across the country.
I have a feeling if another Boscastle or Lynmouth happened again today it still would be a number of hours before it was discovered, watching last years flash flooding at Coverack brought this home to me.