Designing Game Difficulty: Smooth the Curve, Don't Just Crank It Up

A common misconception in game design is that levels should get relentlessly harder as players progress. The real goal is to make players feel they’re improving. Rather than treating difficulty as a single rising slope, the piece argues for introducing mechanics—new enemies, abilities, puzzle types, or novel combinations of existing systems—each with its own gentle learning curve, and smoothing out the outliers that make sections feel abruptly too hard or too boring. The core hazard is that developers, who play their own games constantly, lose the ability to judge what a first-time player finds difficult.

The author lays out practical do’s and don’ts. Good practices include optional collectibles and alternate routes, teaching mechanics multiple ways (visual cues, prompts, environmental hints à la Mario’s coin trails), and dynamic difficulty that quietly adapts to how a player is doing. The don’ts are mostly about fairness and clarity: don’t reward strong players by locking upgrades behind expert-only areas, don’t punish failure by making retries harder, don’t let players skip content they haven’t actually learned, and don’t bury unskippable cutscenes in high-death zones. A worked example—teaching a wall-jump through a safe intro, escalating pressure, then an optional showcase—illustrates the pacing.

Crucially, the author stresses there’s no universal curve: a casual mobile title and a punishing hardcore game want opposite gradients, so developers must pick a target audience rather than chase everyone. To move beyond gut feel, the piece recommends empirical telemetry—death-location heatmaps (as in Mario Maker), per-level completion rates that expose difficulty or boredom spikes, structured playtest questions, and recorded or streamed sessions that reveal how players actually reason through a puzzle.