The Erdős discrepancy problem fascinated me because it was so simple to state yet so deep in its implications.
The art of invention consists in finding out the causes of known effects.