Well, it is artificial, mostly because it gives the DM leverage to say that, say, dinosaurs don't exist in their setting, so no, you cannot turn into one. Its not really meant to ban turning into a grizzly bear when you play in the Sword Coast, though it is banned if you play in a Dark Suns setting (I would guess that bears don't exist there?).
Sure, it does give the DM leverage. But the fact that this then relies on DM fiat is poor design in my opinion. A wizard does not require DM fiat to learn new spells, because it is assumed that the wizard will be doing their own research. A wizard does not need to wait for a scroll of fireball before they can learn fireball (though if the DM uses that as a reward, that's a bonus). So why must a druid be only able to learn new forms if they encounter the animal? What if the DM doesn't use beasts very often? Also, what does it say about the design of beasts that a DM must step in to adjudicate whether a particular form is allowable? What I'm trying to do is bring the design of wildshape more into line with spells. Spells are (better) balanced against one another and it is much easier to eyeball them to be within a particular power range. It is also easier to pump up many lower level spells with higher level spell slots. The only way for a druid to do this with wildshape is to abandon one form in favor of another.
Now, I'm a fan of allowing just a single form that grows with time. Or just leaving the usual 1/8th or whatever level CR most druids can change into and adopting one "advanced" form that you can shift into with growing stats. The reason is because I like the idea of a spider-focused drow druid, or a were(-wolf, -panther, -cow, -fox) shifter, and so on. Single animal form you embrace. I really do like the idea of a single war-form.
But you are making things way over complicated here. Most of the time, a druid turning into a bird or frog or whatever don't really depend on stats. AS you mention elsewhere, its mostly as an exploratory thing. So, why not just make notes for said exploration bits and go from there?
Because even exploration bits require at least some mechanics to run effectively and be similar to what current wildshape forms are capable of achieving. For example, if you use an animal form to sneak or explore, you probably want to know how well your stealth and perception skills work and whether you have access to darkvision or blindsight, which even low level druid wildshape can have access to. In addition, if you need to get somewhere fast or are trying to escape something, you probably wanna know how fast you are or what the chances are that something might hit you. Even in exploration, as a squirrel you might encounter a hungry cat, or that hobgoblin you're trying to sneak past might spot you and may decide you make fun target practice on its long and boring watch duty. Better for the DM and player to have something in the design for consistency and fairness rather than rely on DM fiat. Not to mention, without adding in bonuses to stealth or other achievable abilities with wildshape or adjustments to stats, you are now weakening the class over all as it now has no access to things that traditional wildshape would allow.
Hmm... HP seems to be following 4e's philosophy. Most of the tankiness is supposed to come from sacrificing spell slots. You're standardizing speed and, well, almost everything. That's kind of boring. Part of the interesting things about picking your form is being the large lumbering bear versus the agile snake, versus stalking tiger. Sure, there's the abilities, but having it reflected in the stats is good too. You're better off with either giving a second standard array for the Moon druid to track, or start off with a base animal stat and letting it grow with level in some way.
I came to these stats by literally combing through a decent sample size of CR 1 beasts, finding the average of their physical stats, and using that as the base. I also built in a method for improving the forms based on a preferred play style for the player. In some way this allows more flexibility because you don't have to rely on existing beast stat blocks. But it does sacrifice the ability to adopt different battle forms with various offensive/defensive strategies. Personally, I think the advantages of streamlining the process of wildshape outweigh the potential clunkiness and time cost of pouring through various options each time you wildshape as exists with the current iteration.
As for HP, as a druid increases in level the forms they have access to increase in power, including HP. I took the average HP of a CR 1 beast and eyeballed the HP increase of higher forms. I may be off in my assumption for approximate HP of a given and appropriate CR beast option, but I think the formula (base HP + druid level * X) is solid.
As it is? You're lacking any real growth. Hells, there's not even any multiple attacks at level 5+. This means that you'll always be attackign with just 1d8+3 damage. That's not good.
I had forgotten to include Multiattack in there for combat forms (I don't think non-moon druids should gain multiattack). But otherwise you are incorrect. If you read the entry carefully, each form gains an ASI when the druid gains an ASI, and they can gain perks which include improving armor, improving the weapon damage dice, and various other attack options. I did not elaborate on those such as poison because right now this work is still being polished, and I have only really tracked the various abilities animal forms can benefit from without yet adapting them mechanically. So thank yo for pointing this out. This isn't a final product.
Plus, you're not really looking at the Moon Druid's level 6~14 subclass levels. just the apprentice levels. While level 6's primal strike really doesn't need adjustments (maybe put in Multi-Attack here), level 10 and 14 are basically completely and utterly unrealistic to leave alone.
You're right, and I appreciate you pointing this out. This is in part because I am still reviewing higher level beast form options. I have not even looked at how to incorporate things such as the elemental wildshape. But these are areas I'm looking to include and polish. Any suggestions would be welcome.
Curious minds inquiring. How much have they mattered in your games? I've never seen dice rolls really come up in my game related to animal shapes.
I mention examples above. Exploration, stealth, and speed forms all have the potential to require stats and mechanics because you never know what a player might encounter that might ask for an opposed roll or use of some additional ability. My philosophy is better to have stats so that the player has a reasonable expectation of their abilities rather than require DM fiat and risk inconsistency from session to session.