If you avoid trending months, basing strike selection purely on implied probabilities is a valid approach IMO and not neccessarily restricted to large accounts.
For example, the last 4 months (including this one) have been excellent for probability based (double or pregnant) butterflies or...