Training definitely helps produce a more consistent yield - less weaker buds in lower areas, more evenly-distributed mass of flower etc etc.
It probably doesn't do much for overall yield, and in some cases can actually reduce it. You lollipop too much and you definitely reduce yield. You train too much and limit your vertical growth, that'll also cause a reduction in yield. primarily user-error stuff...
with autos on a biological cock, rather than us changing dark cycle length, you could argue things like topping would not be wise as you cut off growth that cannot be recuperated, but if you only cut off 1-2 nodes, it's not a huge effect, either. Most people will probably cut off more larf when they lollipop and not recognize that as loss of yield while doing it. six in one, half dozen in the other...
autoflowers are simply marijuana plants that go into flower based on some biological timer. That can make for differences, but not nearly as many as people attribute tot eh plant.
e.g.
autoflowers are not 'faster' - total nonsense they take just as long to flower when comparing apples to apples.
and just to cover a false negative too -- they don't yield less per sq ft - **indoors only** outside they definitely have a lower yield potential per arbitrary area measurement.
dli and atmospheric carbon sets the ceiling, then user error reduces from there. uncontrolled environment or pruning shit off that is helpful etc etc. weird shit based on pseudo science doesn't always have a negative effect but definitely can.