In a discussion over at Geneamusings, I contended that is was exceedingly rare to be able to document back, in all lines, ten generations. As I stated I know of only one person that can do it and he is 100% French Canadian. I know dozens of the best genealogists and they can't do it. I know that royalty can't do it, including the present Queen or Lady Diana, both of whom have had teams of genealogists work on their ancestries.
Using myself as generation #1 and going back ten generations, I, like so many other Americans, cross borders. In one historical aspect my ancestors come from two places: the British Empire and the Austrian Empire. However, it means I must collect records from the following modern countries: U.S., Canada, Ireland, Scotland, England, and Slovakia. Using Randy Seaver's rule of needing to know both given and surnames my numbers crunch out as follows:
Generations nos. 1-6 are 100%. I know all my 3 greats.
Generation no. 7: 61 out of 64 for 95% [missing three women, two Slovak, one Irish]
Generation no. 8: 77 out of 128 for 60%
Generation no. 9: 78 out of 256 for 30%
Generation no. 10: 81 out of 512 for 16%
In generation #9 I have my first (known) repeat ancestors, that is, nos. 398 and 399 are the same people as nos. 406 and 407. Generation no. 10 is on average born in the 1690s, some 270 years before my birth. The precipitous fall between generation 7, 8, and 9 is easily explained. Empress Maria Theresa only allowed freedom of religion in 1772. Before that only the Roman Catholic religion was allowed and my ancestors were all Lutheran. Therefore there are no parish records for any of my paternal ancestors prior to 1772. In addition, during those generations, my Scots-Irish forebears move from North America back into Ireland where the records are practically non-existent. One-eighth of my ancestry is pure Scottish and their parish records dwindle significantly in the years prior to 1780. Thus, only my colonial New England ancestors have a chance at being known at those generational levels.
This all arose from the television show Faces of America, and the DNA testing that said certain of the celebrities shared a common ancestor within 10 generations or 250 or so years back. It should be noted that not even the Genetic Genealogist had heard of this test nor could he find any article in the literature that discussed it.
My contention was that you can play just the probability odds and take the DNA out of the discussion. At my tenth generation, I can only document 16% of my ancestors and hence I don't know 84% of my own ancestry. It literally could be almost anything. If it takes only 23 random people for the probability of any two of them sharing a common birthday to be 50%, then common sense dictates that if I don't know 400 of my ancestors from 250 years ago, that anyone with a similar background, with a similar number of unknowns, should be my cousin at least 50% of the time (although I'm guessing greater). If you increase the number of people to 57 in the birthday paradox, the odds go up to 99%. And we are talking hundreds of unknown ancestors.
Whenever I meet someone with significant New England ancestry like I have, I assume that we have at least one common ancestor. Glancing at Randy's ancestry, I easily see a Whitney line in common. And I only know 16% of my ancestry.