And let me put it another way a completely artificial example or two is unconvincing.
A better way of thinking might be to imagine a final agricultural productivity A as the product of any number of factors A= X+Y+Z +T etc.
In this case using a seed drill requires another period of oxen/draft animal use. It also implies a larger labor input in weeding in return for proposed loss of crowding, and potentially better seed placement (although presumably bad handling of your dill negates that). Broadcasting allows labor minimization – i.e. faster application, less draft animal uses and if successful a dense field that minimizes weeding. The day a farmer spends seed drilling might also have been used to say rent his oxen out and raise cash to buy manure etc and thus improve yield in that manor. The time his family spent weeding rows might conceivably be spent weeding a much smaller plot of vegetables or Vines or tending animals, or doing something else that is a net positive on yield.
Note also harrowing would minimize your bird eating difficulty…and appears in classical agricultural texts.
I am lost it here can you restate this point.
Again even in a modern setting the difference in broadcast and drilling is not staggering. Note the study referenced here (bottom table):
http://ohioline.osu.edu/agf-fact/0105.html
Again I would note you allow (seemingly in your examples) a 75% germination rate not much worse than an expected 80%+ rate for a modern seed and certainly likely to high for even last years reserved seed. More so still for second and third years saved seed which would have much lower germination rates.
Overall my point is that while it is true a seed drill certainly has the potential to improve one particular aspect total agricultural productivity, it also has a cost in animal and manpower labor and is still just one factor. To say its effect is so large as to obviously overshadow all other aspects of production, let alone storage, transport, the nutrition of the nominal grain mix consumed, of other health factors (medicine, sanitation, etc), pests, climate, economic structure of agriculture, property laws, transport costs etc when considering the final sustainable population of either Han or Rome is silly.