Game designers have to spend a lot of time balancing games. Building an entertaining mechanism is a good start but a game without balance quickly turns away the players – too dull, frustratingly impossible, unfair, etc. Nerfing, and other after-the-fact balancing, is often dismissed by gamers but I think among game developers nerfing is assumed to be a fact of life. You cannot predict all the ways in which players will interact with your game, so the game must be adjusted.
Games which utilize procedurally generated content are attractive because there is a potentially unlimited quantity of content. They are also risky, because the game relies upon the quality of the algorithms generating the content. Will they generate scenarios which drive away players?
Can the two be successfully combined? A game world which is algorithmically generated, but in which the risks and rewards are also procedural? Definitely! In many ways, I think you could describe Minecraft as exactly that. But most games I’ve played utilize only small bits of procedural content or procedural balancing. I think it is unwise to jump in without considering the risk, but the benefits are particularly attractive to indie developers. Revenge of the Titans did apply some amount of algorithmic adjustment of difficulty, but personally I found it counterproductive – I didn’t enjoy the “punishment from above” that occurred when I was doing well.
I prefer a game experience which is consistent or which allows me to choose select the difficulty. In this respect, Revenge of the Titans got this right with the option to make each level easier. A player could select this option multiple times to reduce the difficulty. This feels more natural to me than the age-old “Easy, Medium, Hard” option presented at the very beginning of a game (before you’ve even played it to know how hard it is!).
But difficulty isn’t the only factor involved. Conceivably, a proper algorithm needs to balance all the relevant aspects of the game. The first step towards a successful implementation requires identifying the variables involved. At I high level, I believe these are:
- Risk – factors working against the player
- Reward – beneficial artifacts or status symbols
- Peril – distinct from Risk. In most games, a “death” implies the loss of status, loss of valuable items, etc. Multiple negative consequences stem from the event. The likelihood that a player will perish, or the difficulty of staying alive, must be carefully considered.
I’d also like to consider how procedural balancing might allow for “unit testing” of game mechanics. The balancing act may take place before or after the game is released. The goal should be to minimize jarring changes for players.