This seems an impossible question. How many things are there that we don’t know about that solve problems, including problems that we might not know are problems?
Gregor Mendel worked on breeding experiments 1856 to 1863 or so, died in 1884. His 1865 paper on plan hybridization was mostly ignored. Hugo de Vries and Carl Corren published articles describing duplication of Mendel’s work in 1900.
John James Waterston’s 1845 paper, which would have qualified as a historically important paper had it been published when it was written instead of much later, was mentioned in the Royal Society’s publication in 1845, but was only published in 1892 as a result of the efforts of Lord Rayleigh then Secretary of the Royal Society, well after Waterston’s death in 1883.
Suppose for example 10,000 hypotheses a year are proposed by unknown researchers and that one tenth of one per cent are valid, which would be 10 out of 10,000. Do you have a numerical guess?
There is a follow up question.