Basically, when using sexual reproduction, it _would_ be highly effective for the species to have many more females than males, since females are the bottleneck of population growth. However, if the sex ratio is skewed then any traits which increases the likelihood of male offspring will enjoy a reproductive advantage (because the male offspring will itself have more offspring than the females), spread through the next generation, and reduce the sex ratio, until it reaches the balance of 50/50. This applies whether the trait is dominant or regressive, etc.
So in this scenario, as the tuskless mutation spreads, the number of males decreases and the genetic advantage of having tusks grows, until it balances out the genetic disadvantage of losing females to poachers.
I think you are making a good case for that conclusion, and would that not imply that the mutation - or, more specifically, its most recent manifestation - arose fairly recently, i.e. subsequent to hunting for ivory becoming a serious problem for elephants?
So while a tuskless female may generate 2:1 females to males, that's (hopefully, depending on poaching) far from a stable birth ratio for the population, though active poaching may cause the female to male ratio for living specimens can be further skewed much higher than 2:1.