Comment by contravariant

Typically 1 million samples of anything is reasonable enough to get a good estimate, provided your sample is random.

The probability that you've picked the exact 0.0001% of galaxies that had a specific pattern is vanishingly small.