Pardo R. 2008, Rules of Engagement: Blizzard’s Approach to Multiplayer Game Design
- Implement the multiplayer part of the game first, then the story and single-player components.
- In PVP, focus on balance, skill differentiation (e.g. reflexes for FPS, multitasking and strategic thinking for RTS, knowledge of the mechanics for both RTS and FPS, economic dominance), and ladders/ratings. For co-op games, focus on the communication between players and complementary classes. Ex: in Warcraft 3, a mine is considered 'full' when 5 peons work on it; the economic part is dumbed-down to encourage the players to focus on the micro-management of fighting units.
- Avoid differentiation on map knowledge: it's not really a skill. Instead, reveal the map but keep a fog of war (like in Starcraft 2), so that players know the flow of the map and where the resources are, and can pick their strategy accordingly.
- Everything should feel overpowered, not mediocre.
- Balance first for the expert, then for the novice.
- Balance is affected by the maths, but also by the UI (e.g. WoW's UI mods, or the possibility to select only up to 12 units in Starcraft 1 as opposed to an arbitrary large number of units in Starcraft 2), maps/level design, special effects (e.g. too much blurs the vision, cf the War of Emperium of RO set /mineffect by default to limit the visual flood of skill effects)
- Players hate loosing, hence make games shorter so that they can play more games per play session, and eventually win some.
- Reward the behaviors you want people to do/
make it a bonus
- Tie art and game design together. The appearance of Heavy of Team Fortress is explicit: tough, lots of HP, and lots of damage.
- Spectatorship enables empathy with the players, cf Poker became more popular when hole cameras were introduced because the audience understands better what's going on.
- It does not make sense for warriors to cast spells, therefore they don't have mana but rather they have rage.