You don't know. The point was even if you did get there it could still be an overfit model you don't want, since it's based on a training data set, not the true

You don't know. The point was even if you did get there it could still be an overfit model you don't want, since it's based on a training data set, not the true statistics of the distribution the samples come from.