It would be possible to get a sensible number for this for a much shorter span than 300 years, I think - and also, I suspect that the percentage chance of having any descendants at X amount of time probably doesn’t decrease very much from about the third generation onwards. The reason being - once you have any descendants at all, the probability is that you have more than one, and it’s very much less likely that multiple ancestors of yours will ALL die without descendants, than that you personally, a single individual, will do so.
I did a quick BOTE simulation, with vaguely plausible simple numbers.
Suppose in your society, the chances at birth of a person leaving descendants is:
20% have 0
20% have 1
20% have 2
20% have 3
20% have 4
nobody has more than 4
(those numbers are obviously not drawn from real life, but they’re not awful - a distribution like that would lead to a population that doubled every 50 years or so so it’s slightly on the high side, but not too bad)
So an average person has an 80% chance of having at least one descendant in the first generation.
IF they have 0 children, they always still have 0 children in the 2nd generation -
IF they have 1, they have a 20% chance of having 0 in the 2nd generation
IF they have 2, they have a 4% chance of having 0 in the 2nd generation
IF they have 3 they have a .8% chance of having 0 in the 2nd generation
IF they have 4, they have a .16% chance of having 0 in the 2nd generation.
When I add all those numbers up, I find that Pr(0 in the second generation)=.21 + .2.2 + .2*.04+.2*.008+.2*.0016 ~=.25
So the probability of your random person having a descendant in the second generation is about 75%. It’s gone down really very little. It would go down even less in the third generation.
The question “how many people born in (say) 1700 still had descendants in 1800” may or may not be empirically determinable. But whatever the answer is, I believe the answer to “how many people born in 1700 still have descendants today” is probably more or less the same.