Just in case starting the sample with 100 women who are not pregnant somehow skewed the numbers, lets simply consider two individual women over their lifetimes. Lets say that each has 25 fertile years. One is on the pill the entire time. If we add up her 2% chance per year of having a zygote flushed out, we end up with a 50% chance over her lifetime. The other woman does not use birth control. For every menstrual cycle, there is a 16% chance that she will have a zygote flushed out. This means that out of six periods, on average she will have one zygote flushed out one fertilized egg that will fail to implant in her uterus. While without birth control she will likely spend much of her time pregnant, she will still have periods in between pregnancies. And since its just as likely that a zygote will fail to implant in between two pregnancies as it is that it will successfully implant, thus starting another pregnancy, she will probably end up with just as many dead zygotes as she ends up with pregnancies.
Thus a woman on the pill for her entire fertile period has a one in two chance of her body killing a zygote by flushing it out, but the body of a woman not on birth control will flush out and thus kill numerous zygotes over the course of her life. Sarahs numbers, then, are low. The actual numbers indicate that a woman on the pill for her entire life flushes out something like 95% fewer zygotes than a woman not using birth control.