Thanks!
First off, you're blurring together a couple things here, so I'm going to try to be extra explicit and consistent with the terminology. Also, I'm just a hobbyist that has been geeking out on this stuff lately. I think what I'm saying is scientifically valid but I could have some minor details wrong.
I've made feminized seeds by reversing females with STS before, though currently I prefer to add a couple regular autos to my grows (otherwise feminized autos) and keep male(s) long enough to collect pollen. Absolutely nothing against feminized seeds, but the reversal process adds a couple extra weeks and some timing variability, and I like to start and harvest everything at the same time. Since I have a plant count of 12 here (MI) and tend to do SOG-ish grows in a 2x2 tent, it's not much trouble to start 8-10 or so plants (some fem, some reg) and cull a few males or less vigorous females if I'm planning to flower 6. I would probably put much more priority on making feminized seeds if I was following a smaller plant count.
If you grow feminized seed made with modern feminization techniques, it's just a female plant, not really different from growing non-feminized seeds and having some be female. A lot of people believe they're genetically damaged somehow, but Ryan Lee/Chimera, a cannabis breeder with much more scientific background than most, says that is nonsense -- other agriculture uses feminized seeds (but nobody besides cannabis growers call them that), and you don't see cucumber farmers ranting about how the process is destroying their genetics or whatever. Seed feminization doesn't automatically increase the odds of intersex traits in the offspring or somehow spontaneously add them when not present, but also doesn't remove them if already present either. When reversing a female plant with chemicals, it's harder to tell if they were also going to grow a few male flowers due to stress -- so the feminized seeds still need the stress testing that breeders should be doing anyway. There's a good Shaping Fire podcast episode with Ryan Lee about all-female breeding that clarified a lot of concepts for me.
The chart (Punnett square) you're talking about is for when you're crossing autos and photos. I've just been crossing autos with other autos, because I have neither the time, space, or desire to get into the higher plant counts that doing proper photo/auto breeding would require. I'm just chucking pollen and making F1 personal crosses. At some point I may grow out 10 or so of them plus one or two of the parent strains and make backcrosses (like Pink Panama x (Pink Panama x Blue Microverse)) and also F2s based on the F1s I like better. I can't do a whole lot of selection with just like ten plants (a hundred would be a good start), but it would be fun to see how they turn out.
When crossing autos with autos, since both parents are already homozygous for the recessive autoflowering trait, the offspring will be autos too. The Mephisto F1 crosses are between their (already) autoflowering lines, when they make crosses with photos they don't sell them until worked to at least F4, because then they can lock down the autoflowering trait.
Also, feminization has nothing to do with autoflowering, really. A lot of people seem to think that autos are inherently feminized, but regular (non-feminized) autos are just less common because of market demand, not biology. The seedbanks with menus like "regular, feminized, autoflowering" add to the confusion. Some auto breeders, like Mandalorian, only/mostly sell non-feminized autos.
Finally, I try to avoid using the term "stable" talking about this stuff, because there are like six different things people may mean by it, so it's clearer to use distinct terms.
