Balancing the game

I have on several occasions designed RPG and Open RPG maps, but I've never finished one. The main problem I seem to have is balancing the game! I have tried several approaches:

* Use all basic stats, max lvl 10 and all the standard creeps. Balanced, but pretty boring as it lack my unique mark.

* Really low stats, 1-1 dmg, 50 hp and weapons that boosts a little. Max lvl 100 and stats increase drastically on level up. Very hard to balance, and I have to create creeps for each level, which is an imense work.

* Balanced stats, but max lvl 1. The easiest one to make as I only need to think of creep type (small, normal or tough), but will get boring as the hero don't learn any new spells and won't encounter any harder enemies. Pretty one-sided.

So how do you balance your maps? what would you say is the best approach?

it really depends on what you want to make. just look at common games. sc, for example, gives relatively low health and damage to units; an ultralisk has 800 health, which is a lot. a marine has 40 health and does 6(?) damage. yet in wow you can get 20k hit points on one guy. higher numbers mean more variations while being more complicated.

I don't actually see much difference between the second and the third options. In both, you have a certain power level and a set of enemies that matches that level. Sure, in the second case you go through a hundred such sets while in the third you only have one, but neither gives a good sense of progression.

Instead, you should have fewer tiers of enemies, but play with the enemy group compositions. First encounters are always with a few basic opponents, then as you level up the number of enemies begins to grow, but they're still the same enemies; that way players can see how much more effective their hero has become, something that wouldn't be very apparent if there was instead the same number of opponents of a different type that has more life and damage. As heroes advance further, individual monsters from higher tiers are encountered in addition to larger and larger groups of basic enemies, then mixed groups of both basic enemies and higher units, until you start encountering multiple higher-tier monsters only and they become the new basic tier. Repeat the process as needed, although the goal is to accommodate many hero levels with as few tiers of enemies as possible.

This way, a hero can go up many levels without you having to design too many enemies and as a bonus, players will be able to intuitively feel how much stronger their hero has gotten as they will be able to compare his strength to that of a basic enemy for a long time while the game remains challenging.

As for actually balancing the game, there are multiple options. One approach is the self-balancing setup where the player may clear the same creep camp as many times as he likes, allowing him to farm an area until he is strong enough to move on. The only thing you need to watch for here is that the experience gain rate is fast enough that the player doesn't have to stay in one area long enough to get bored.

A more difficult approach is the linear RPG, where the enemies do not respawn and players are expected to only go through each area once. The game is less open, but players are spared the monotomy of grinding. In this case, you must plot carefully the amount of experience the players will have gained with killing each enemy you place and adapt the group compositions accordingly. Do some test battles to determine good enemy groups for a few specific levels like 5, 10, 15... then interpolate in between.

For the second approach you could always use the hero template for every creep, allowing their levels to increase as well. Final Fantasy does this for most of their games I'm pretty sure. Even if you only use 5-10% of the maximum levels for each unit it saves some work.

For the third option the only limiting factor is that new abilities are not achieved by leveling up, but you can always add newer and better abilities in other ways. You could substitute lumber for an experience-point resource required for purchasing better abilities or upgrading existing ones, for instance.
A lot of first-person shooters are more on the 3rd one as well, and generally keep things interesting with low health to damage ratios and a wide array of customization options for each character. Doing this approach also allows you to place in a wider array of weapons, so a warrior could use swords, axes, and maces, rather than just having a liner progression of 5-50 swords that are all slightly more powerful. With different customization you can make weapons good at different things as well.
I'm personally a fan of the third approach, it allows T-Rexs and dragons to always be intimidating while dire rats and boars won't be waiting to one-shot you in high level areas. Rather than feeling rage at those killing you you should only feel shame. considering only one instance of every monster needs to be made it offers greater leeway to individualizing them as well; monster X is heavily armored, monster Y is high damage low health, monster Z confuses characters, whatever.
My current projects usually use sort of a combo of your first and third approach, were items are used to upgrade the characters to be ~3 levels stronger while offering 10 different levels of creeps, fulfilling the needs of gray mobs, green mobs, yellow mobs, red mobs, elites, single-player bosses, multi-player bosses, and multi-faction bosses.
Just make sure all the abilities are really worth using and offer something unique to the game-play. If they are more instant-damage effects make their cooldown and mana cost considerably lower than default melee map spell ratios to merit using.