The calories per unit mass is about right. I think the mistakes are:
1: No batch time savings. This is a really easy fix that anybody can make for their own game. Simply navigate to blablabla\data\json\recipes\food.json, do a search for
result": "veggy_wild_cooked"
and under Autolearn type the line
"batch_time_factors": [ 80, 4 ],
or whatever you prefer (first number is time savings, second number is threshold) so you aren’t painstakingly cooking one root at a time while you starve to death.
2: Low number of units found per search, relatively high probability of finding units even at low survival skill. I don’t know how to edit this one myself, but I’d suggest returning about 2d6 units per successful search, making the success rate per survival skill much worse at low survival skill, and then applying a massive penalty to quantity found and a more mild penalty to success rate during winter.
My thoughts being that because plants tend to grow in large groups of the same plant, somebody searching for food during the growing season would be limited by knowing where to find a group of plants and how to recognize it as food if they did, not by the amount of that plant available in the patch.
Low Skill: “Oh look, a lone violet. How pretty. I wish I had food.”
Medium Skill: “Oh look, a lone violet. I could eat it, but it doesn’t even qualify as a snack. There must be another patch of fiddlehead ferns around here somewhere.”
High Skill: “Oh look, a lone violet. I know what kind of light conditions and soil conditions there are, and I’ve noticed the prevailing wind direction and where other plants are growing, so I bet there’s a good sized cluster of tasty plants right… around… here.”