In 1936 the Literary Digest mailed ten million ballots and predicted Alf Landon in a landslide. George Gallup, polling a few thousand chosen carefully, called it for Roosevelt — and was right. Size didn't save the Digest; its list did.
That list came from car registries and telephone books, which in the Depression skewed wealthy. That's coverage bias: a whole slice of the town was never in the drum.
A convenience sample — whoever walks past your booth — quietly asks "is downtown at noon the same as the town?" Choosing where to stand is a hypothesis about who counts.
Only a random draw gives every resident the same chance, so the sample's average tracks the town's. The margin of error then shrinks with √n — but bias never does, no matter how many people you ask.
Something in the simulation stopped unexpectedly — the lesson continues without it. Nothing you did was wrong; you can move on.