10% of Medieval Manuscripts survive to this day - Is that an overestimate?

[countingtls]

Here in East Asia and in China, there are very precise records for book collections sponsored and compiled by royal courts from different dynasties, and they had known dates and index pages and that showed how many were lost.

One of the first "類書" collections of books in different categories (類) from the Ming Dynasty is "You are not allowed to view links. Register or Login to view." compiled in 1408 (although translated as Encyclopedia, but they are actually not, since they were collections/copies from sources as a whole, not reedited, and in their original format as a collection in the "royal library" they would contain 底本 the source materials). And out of the original 10k+ volumes, less than 400 survived (less than 3.5%) to this day.

We even know how many were lost in different eras. The next royal book collection You are not allowed to view links. Register or Login to view. started in the next dynasty, Qing Dynasty, in 1773 (finished in 1782), they actually did a survey for the previous collection from 1408, and found that at the time (after 370 years), more than 10% of the collection was lost, and nearly 80% of the source materials (底本) associated with them were lost. 

As for the 1782 collection, just over 100+ years later in 1893, they did another survey and found more than 2/3 or the original materials were lost already. Forward another 100+ years to today, more than 2/3 of the remaining 1/3 were lost again. (like only about 10% remained in 250 years)
I seriously believe we've lost a lot more than we thought.