"Resource Pool" to Restrict Units

I can't think of any specific games that function like this, but the basic "unit selection based on a fixed number of points" sounds like a typical miniatures based wargame (Warhammer or Warhammer 40K for example).

The idea of damage being applied to the player, which forces the player to select which units they are losing, is interesting.

Would this cause units with high defenses to be overly valued to the players?

Good luck with your design!

I think that would reduce the strategy, rather than help it. The attacker attacks and does a certain amount of damage, but then the defender gets to choose where that damage gets applied. This completely eliminates the concept of "high-value targets" because it doesn't matter what you target.

If the target matters because some have damage-mitigating effect, it just means that an attacker will do what he can not to attack those. If they are walls that have to be brought down before anything else can be attacked, then there this does introduce some strategy, but it seems somewhat artificial. As the defender, I have to decide whether to let my walls fall or to accept the damage among my offensive forces in order to keep my damage mitigation around for another turn.

But it's not a very interesting decision -- there's probably a formulaic approach and doing anything else is just going to be wrong. I assume that the defender strategy would always be to put the minimum into walls that forces the attacker to attack them, and gives them the best damage mitigation that can reasonably be afforded. After the first round, collapse all the walls that the enemy CAN NOT attack in the next round, take the remaining damage in my offense, and dump all the rest of the offense into him in hopes of crippling his attack.

The whole idea that he attacks my east walls, so I take the damage in my west walls just feels silly.

Your problem is hard to deal with. Your mechanic has openings for abuse. And is hard to balance with. But I think you can work with it.

A normal war game often has
Cost = Health part + Damage part

Where you choose a normal unit to have
Health part = Damage part

In a system like that, walls have only the Health part.
Support units have 75% Damage part
Meat units have 75% Health part

Now, with your point system. This is all a bit to hard to work with. These natural RPS additions cannot be used. After all, the defender removes ALL walls first, then ALL meat units, the ALL normal ones, only to conclude with ALL support.
However, it still can be done to some extend if you add grouping.

(You are not using a board with fields, right??)

Is it an option to allow groups of units instead? Let's say, 1 group has an equivalent of 100 points. So instead of one big 1000 point army, you have now 10 little armies.
This means that the attacker can choose a group. Then the defender can choose what he/she wants to loose in that group.

And you get some more strategy if you add a mechanical RPS system.

Lets say, a meat unit that is anti infantry and a support unit that is anti tank. Another group has a meat unit that is anti tank, yet the support is now anti infantry.
And something like a group of infantry with 1 tank. If the damage is overkill on one infantry unit, the infantry unit will serve as meat for the tank (fodder effect)

This way, the attacker might still take out something he/she wants to take out asap.

I think it is better to make a complete general plan. Then start filling in details in every corner.

Building a war game is very tiring if you don't know where to start.

First things first. What part of your game has to be IT!? Meaning, is there a mechanic that you certain want to use? It could be in the combat, or on the board, or something in the rules.

From there on out, build up the other parts. Some parts will be strong, other weak. A wobbly circle of these points. All you have to do then is to even it out and test if the circle is round.

Start with an easy concept. And expand there where you want to. Untill you are satisfied.

Of course there will be one point that is really hard to get correct. And once you think you have something that works, you find out you have to tweak something else to make it work. (Example: For me this was the type of terrain in combination with all the propulsion types and projectile pathways. That part took me 1 year and was the last thing to do for me)
All I can say to that is: take one step backwards and finish. Just say no to something you want, yet can't.

Of course, you are free to discuss on this forum.

When you are done, you could send me the rules. Then I can take a look if you want.