It's worse than that, the authors actually revised their calculation and admitted that how they calculated it was a mistake and they have since revised their initial table and done a follow-up, see here.

The authors reply: Sun appropriately raises questions about the proportion of women reporting spontaneous abortion in our recent article. We agree that the denominator used in that proportion — 827 completed pregnancies — is not an appropriate denominator for the calculation of a risk estimate or rate.

The number of spontaneous abortions (104) reflects data reported by the participants as of March 30, 2021, during telephone follow-up. In this preliminary report, follow-up information was missing for the majority of pregnancies in which exposure to vaccination occurred in early pregnancy. Among the 1224 women who had been vaccinated before conception or in the first trimester, follow-up through 20 weeks of gestation had been completed for only 204 pregnancies that were known to be ongoing and for 1 pregnancy that resulted in stillbirth. Among the pregnancies that had not yet reached 20 weeks of gestation, there were 10 pregnancies with other outcomes before 20 weeks of gestation, including 8 ectopic pregnancies and 2 induced abortions. For the other 905 pregnancies, follow-up had not occurred to establish whether these pregnancies were ongoing past 20 weeks of gestation. We have amended Table 4 in our earlier publication and have clarified the text.

Subsequently, we completed telephone follow-up for the 905 pregnancies and enrolled additional persons in the v-safe pregnancy registry. To determine the cumulative risk of spontaneous abortion from 6 to less than 20 weeks of gestation, we used life-table methods to perform an updated analysis, now reported in the Journal, involving 2456 women who received at least one dose of an mRNA Covid-19 vaccine before conception or before 20 weeks of gestation.1 The estimated risks (14.1% overall and 12.8% in age-standardized analyses) are consistent with the risks of spontaneous abortion reported in the general population.1

These numbers are much lower than that initial sample (104 of the 165 occurring in a 127 woman sample vs. 61 happening in a sample that is 8x larger) which leads me to believe that they likely miscarried sooner rather than later post-vaccination, which raises more questions. Specifically what proportion overall miscarried (i.e. perform a similar and rigorous follow-up as the initial data collection was), and how long after vaccination did the spontaneous abortions occur? This is where a placebo group being injected with water would be useful.