I also posted this under Game Programming, but perhaps here is a more appropriate section. Me and a group of friends are developing a simple tactics RPG. While we have mostly finished programming the game engine, and have an idea of how actual attack calculations should be worked out, we have absolutely no idea how to start developing a system of stat progression. Our tactics game will have multiple classes, such as your typical Warrior, Archer, Mage, etc. We will have the normal statistics - Strength, Defense, Agility, Critical, etc. While obviously different classes will begin with stats accordingly, we have no idea how, mathematically, each stat should advance as the characters level up in such a way that things remain reasonably balanced. If anybody has any experiencing in programming tactical role playing games or in stat progression for similar games, any advice or links to useful resources would be greatly appreciated :)

No experience actually making these, but I've played a lot. A few ideas.

1. Make sure your dudes don't miss all the time. RPGs where your attacks miss half the time get sort of boring. See: Wizardry 8.

2. In terms of stats and stat progression, simpler tends to work better. I'm not saying to limit yourself to only three or four stats or something, but what you should do is focus on making sure every stat is meaningful. In other words, don't have two stats that control roughly the same thing. Don't have any stats that could be totally neglected.

3. In terms of stat progression, I'd avoid random stat increases. They might be okay in a regular RPG, but in strategy games a higher degree of determinism is better for the user figuring out... well, what strategies to use.

4. I'd also consider finding a way of removing the ability to grind and making sure that characters never lag too far behind the rest of the party. Lots of SRPGs have issues, for instance, where late game characters are useless because their level is too low. I think maybe devising a system where everyone in the battle gets experience/level ups is a good idea, but maybe give some sort of bonuses to characters that get more kills/hits/attacks/whatever in.

5. Variation on the above, average the party's level and make that a threshold thing. Characters beneath the threshhold get a lot of bonus experience. Characters above it get an experience reduction.

6. Base the threshold on game progression so that you have more control over what levels the party is expected to be at when, so you can balance fights towards that.

Thank you for your suggestions, we will be taking them into consideration.

Perhaps I should outline a very basic concept we had in mind originally. We were in fact intending to stick to only a few basic stats - Strength, Defense, Agility, Magic Defense, Intelligence and some sort of Luck stat for determining critical hits.

The idea was that each class would have each stat labelled as "weak", "neutral" or "strong". Each would start off with unique base stats, which would increase each level according to these labels. Strong stats would increase by 1 every other level, and 2 for the remaining levels, neutral by 1, and weak by 1 point every 2 levels (these numbers could obviously be scaled, such as making it 4 for strong, 2 for neutral and 1 for weak).

For example, a Warrior would have a strong Strength and Defense stat, and weak Magic Defense and Intelligence. Other stats would be neutral. So progression would go like this:

Level 1
STR 2
DEF 2
INT 1
MAG D. 1
AGI 1
LUK 1

Level 2
STR 3 <-- Up by 1
DEF 3 <-- Up by 1
INT 1 <-- Doesn't go up yet
MAG D. 1 <-- Doesn't go up yet
AGI 2 <-- Up by 1
LUK 2 <-- Up by 1

Level 3
STR 5 <-- Up by 2
DEF 5 <-- Up by 2
INT 2 <-- Up by 1 (second level up)
MAG D. 2 <-- Up by 1 (second level up)
AGI 3 <-- Up by 1
LUK 3 <-- Up by 1

And so on...

Problem is, I can see this becoming incredibly unbalanced in the higher levels, with different classes having such a wide range of stats that everybody one-hits everybody. I figure we'll have to develop something less simplistic.

Sorry if this explanation was a little long-winded, but any thoughts or ideas?

[Edited by - code_punk on April 22, 2010 10:56:01 PM]