That's almost cheating using Tom Hill and Chimera as examples. They aren't literally all pocket filling idiots.
You mentioned a few other words that are key... now take a look at what Ace seeds is doing with Old Timer's haze.
I replied to a post at another site about a week ago in reply to a quote from Tom Hill. Essentially sometimes it's better to mate 1-1, sometimes it's better to breed populations and that is often turned around / mixed up. That last thing is pre-mendelian pre-modern breeding. It's what farmers did a very long time ago. It's what the pioneers and breeders of some of those well known cannabis strains did too, and what is necessary to acclimatize and adapt landraces for indoor growing for example.
A hobby breeder at another forum called it "putting the best with the best" which sums it up nicely. The result (after many generations) is uniformity, but with a wide profile, without inbreeding depression (given large numbers). Traits are bred true when they are homozygous, which mean they received the same gene from both parents, while heterozygous genotypes have room for two different genes.
AA = bred true, all off spring will result in AA
Aa = not bred true, heterozygous, one gene dominates partly or entirely, but the other is still there.
Unstable means more different genes hence more different gene combinations hence more phenotypes. The same thing for stable, less genes, less possible combinations hence less phenotypes.
Inbred depression is the result of the loss of those genes you don't see, but would have passed on only to SOME of the offspring of 1 plant but can be passed on entirely given large enough plant counts.
That 'a' can be a gene for a new taste, resistance against a new virus it needs to protect itself again, one of the many genes that makes up the cannabinoid profile, etc etc.
Such a uniform yet diverse population is good for survival, Mexican farmers, I forgot the crop name, put landrace plants amongst their own every few years to keep that gene pool wide. Just in case they unknowingly bred out genes that took thousands of years to evolve and may yet have to show their purpose.
That only works out with a large amount of plants. I'm not talking 12 or 24....
And does not lead to homozygous lines best for creating hybrids. Well, on the very long run it does. Many tools and formulas have been developed to predict how long that takes. The saying it takes many years to breed a new strain... comes from putting the best with the best outdoors, once a year.
Anyway, heterozygous genotypes can carry more 'different' DNA info than homozygous. But crossing heterozygous with heterozygous leads to 3 different outcomes, not just 1, so leads to genotype and hence phenotype variation. While homozygous (true bred) x homozygous is all homozygous, aka a stable strain. What if you can have both, both heterozygous but uniform (no pheno type variation)... that's what IBLs and F1 hybrids based on selective breeding is about. AA x BB = all AB.
The point of an IBL is to mate it with another. "A" plant of the resulting F1 is not suitable for breeding (clone only from someone's hybrid...). An F1 population could be stabilized in many generations but F2 is not more stable than F1, for example.
IBLs and hybrids is what non-cannabis breeders have been doing since around WWII, based on the rediscover of Mendel's work and the phenotype vs genotype distinction from some Johansen guy over a hundred year ago. Thanks to that, we can all eat. See green revolution:
http://www.fao.org/english/newsroom/focus/2003/gmo2.htm (that mexican crop was apparently wheat...)
There's a place for both old skool open population breeding and more selective breeding. Modern plant breeding is gene mapping. Which simply means trying to figure out the genotypes behind phenotypes by tracking how they inherit. The goal there is to do as little inbreed generations as necessary. Merely crossing the best phenotypes (which what small/medium scale putting the best with the best comes down to) narrows breeds out more genes than necessary.
Those early pioneers and breeders put all that work in it so breeders since have been able to create hundreds of strains, well, phenos, with all those great tastes and smells, through selective breeding.
Saying breeders who use open pollination know breeding extremely well is when it comes to cannabis in general contrary to reality.
I'm sure the examples you mentioned had good reasons to do what they did... which should be obvious by now. Think long term. Think evolution...
Plant breeding is not mimicking evolution (too slow). To create modern hybrids and breed and lock in new tastes there's this thing called crossbreeding and intense selection. Biodiverse populations make good breeding stock for IBLs and in turn hybrids. To create a hybrid (with no pheno variety) one would still have to create a homozygous line first (two actually...).
Some people get rare land races or old seeds and then pop them to see if they can find a keeper.... that the most ironic scenario in this context.