Too much knowledge could be a bad thing. I was lead to the SzemerГ©di theorem by proving a result, about squares, that Euler had already proven, and I relied on an 'obvious' fact, about arithmetical progressions, that was unproved at the time. But that lead me to try and prove that formerly unproved statement- about arithmetical progressions-and that ultimately lead to the SzemerГ©di Theorem.