Complexity and cognitive load

Complexity is what players must hold in their heads to play legally: rules, exceptions, icons, upkeep. Depth is the richness of the decisions the game offers. They are not the same thing, and the designer's ideal is high depth at low complexity, which is why Go, with a handful of rules and a lifetime of strategy, is the perennial example. Every rule you add pays rent in cognitive load, and it should buy depth to justify the cost.

Cognitive load shows up in three places. Learning load is the rules overhead before the first turn, and modern audiences expect to be playing within ten to fifteen minutes for family weight games. In-play load is what must be tracked each turn, remembering triggers, modifiers, and upkeep, and it is where exceptions hurt most, because a rule that applies only sometimes is far heavier than a rule that always applies. Planning load is the lookahead the game demands, which is depth when players enjoy it and paralysis when the tree is too wide.

The best reducers are consistency and visibility. Consistent rules with no exceptions compress beautifully: "cards are always drawn face up" is one fact, while "face up except at night, or when the storm track is past three" is a memory test. Put state on the table instead of in heads, with tracks, tokens, and icons showing whatever players would otherwise memorize. Good iconography, clear phase structure, and player aids can cut the perceived complexity of a game by a third without touching a single rule.

When developing, hunt for rules that playtesters repeatedly forget, since forgetting is data: the rule is either unintuitive or unnecessary. Try deleting it entirely before rewording it. Most designs can lose ten to twenty percent of their rules in development and get strictly better, because every rule removed makes every remaining rule easier to remember.