[compsciphd]

people haven't explained why this is the case, it's the simple genetics we learn in biology.

the first generation has a many sets of genes with "Highly Productive Dominant (HPD, my terms for this post) genes, Not so Productive Recessive Gene (NPR)". By inbreeding 2 lines for a while, one can end up with a strain that has both genes being one and thereby when you cross it with something else, you know the offspring will get its gene (as it really only has one). Therefore the first generation, will have many sets as what I described above and will 100% express the HPD of those pairs.

The 2nd generation though of seeds will have a randomized assortment of HPD/HPD, HPD/NPR, NPR/NPR genes. Therefore while 75% of the seeds expressed genes will be HPD on average, that's obviously less than 100% that one gets from first generation.

[Retric]

That’s a reasonable mental model for why this kind of thing can happen, but it’s worth remembering that other possibilities exists. Suppose strain R is adding RR, and strain K is adding KK, and you want RK because the combination is more useful than either RR or KK. There’s no way to get that into a population without some lucky mutations or genetic engineering.

Further you’re likely looking at many such combinations not just a single one.