If you are growing grains and harvesting by hand eat the seeds from shorter plants. Taller plants are easier to handle. And give more straw for animal bedding. Taller grains have been mostly bred out of the seed stock. Machines are more efficient on shorter plants.
Do you actively select and breed the plants you grow, or does the selection happen "by itself" since only plants that are somehow adapted to local condition reach maturity?
Starting with a new plant I will buy a few varieties whose descriptions suit my fancy. Then plant as many seeds as I have space for. Adapting plants is a numbers game. Some won't germinate, some will grow slowly, some will become diseased and some will be eaten by pests. None of those are suited for the local environment. Thin the planting to allow the strongest growers room. It should be obvious which ones to keep. Once you have the plants that started from seed best its time to start selecting for other traits. Traits depend on the plant but for food you should be selecting mostly for taste at first. For example, I am crossing one pea variety selected for wirey strong vines and purple pods and a couple others for taste. This is my third year with those and I am getting sweeter pods on wirey vines but the vines aren't quite as strong or tall. So, I planted some of the original wirey vine seeds to try pushing the line that way more.
Some trait selection is unintentional. The plants are going to adapt to you as well. They will adapt to how you plant, harvest, and save the seed. With my grains some of the plants are too short to harvest with a sickle comfortably. Those go in the eat or feed to chickens pile. Some also don't separate from the plants when I thresh by hand. Also to the chickens. What I am left with are seeds that were easier to process just by not being super careful to get everything. With biennials or clonal root crops you will also unintentionally select for plants that store well as you have to keep the roots over winter to plant them in the spring.
I started saving seed about 5 years ago so nothing has really stabilized into something that could be considered a new variety. But, I have seen the steps closer to my goals.
Purchased seeds are a narrow part of that plants gene pool. If they are open pollinated they were bred until most of the produced seeds make the same looking and tasting plant in a specific environment. Really, if you are gardening, you should learn how to save seeds unless you have a seed grower near you. If you save seeds you will also have more seed than you know what to do with.
Most home gardeners can much more easily support the maintenance of old strains by buying heirloom seeds from breeders.
However, if one has the space, time, skills, and interest to do what you are doing, then by all means go ahead.
All you really need space for is allowing plants to reach full maturity. Some plants, like carrots or radishes, get huge if left to seed. One fruit from most plants will give you more seed than you likely planted. Each seed in a fruit is a unique individual that was fertilized by a different pollen. Even with self fertilization the seeds will not be genetically identical.
But that's not really realistic anymore. If someone like the other poster, who has the space (both to maintain a large enough population as well as the distance from other gardeners who might be growing modern strains), the time, and the interest, they can indeed be doing the "best" thing from a genetic diversity perspective (as long as they make sure that their strain doesn't die with them). But those people are very, very rare. So the next-best option is to try and maintain the current genetic diversity we have, in a persistent way (as opposed to a static way in seed banks), by having breeders breed as many different varieties as possible. They will hopefully be following the dual mandate of maintaining unique traits while also allowing adaptation to changing conditions.
Your average gardener, even if they wanted to, can't maintain any real genetic diversity because A) they don't have the space for a viable population and B) their genetics will constantly be contaminated by their neighbor growing Early Girl or something. So making sure that the already extant genetic diversity doesn't dissapear by supporting heirloom breeding is not the "optimal" solution to genetic diversity, but it is, in my opinion, the best "realistic" solution.