HN Debrief

Making difficulty curves in games

  • Gaming
  • Product Design
  • UX

The article is a broad primer on difficulty curves in games. It walks through patterns like steady ramps, sawtooth rises and releases, sudden spikes, and dynamic adjustment, with the aim of helping designers pace challenge across a full playthrough instead of tuning levels in isolation. That framing gave people a useful hook, but most of the useful signal landed on one point the article only touched briefly: dynamic difficulty is either invisible craft or a trust-destroying own goal.

If you build progression around leveling or accumulating power, avoid hidden scaling that erases gains or punishes efficient play. Adaptive systems work best when they preserve flow without being noticed, or when the game makes the rule legible and thematically grounded.

Discussion mood

Mostly positive about the article as a starter, but strongly skeptical of hidden level scaling and reactive balancing that cancels out player progress. People were much more favorable toward transparent or pacing-focused adaptation, and generally supportive of skill-based matchmaking in multiplayer despite its trade-offs.

Key insights

  1. 01

    Factorio makes scaling part of the world

    By tying enemy aggression to visible pollution, Factorio turns difficulty adjustment into a simulation rule instead of a backstage correction. That changes the emotional read completely. Players can predict the consequence, inspect it on the map, and choose trade-offs like preserving trees or expanding faster, so rising pressure feels earned rather than faked.

    If you need challenge to scale with player success, surface the mechanism in the fiction and UI. Give players levers they can understand and manipulate so adaptation reads as consequence, not interference.

      Attribution:
    • safwat-halaby #1
    • hlynurd #1
    • __alexs #1
  2. 02

    Pacing systems land better than stat scaling

    Left 4 Dead's AI Director is the clearest example of adaptive difficulty that players celebrate because it controls tempo, threat timing, and emotional peaks instead of rewriting player power. It keeps runs near the edge of failure while preserving the sense that your weapons, decisions, and skill still work the way you learned them.

    Use adaptation to shape encounters, downtime, and pressure swings before you use it to rewrite enemy stats or loot tables. Players usually forgive invisible pacing help far more readily than invisible progression math.

      Attribution:
    • sph #1
    • wingmanjd #1
    • rcxdude #1
  3. 03

    Half-Life Alyx shows the immersion trap

    Dynamic ammo refills in Half-Life Alyx kept many players from getting soft-locked, but once noticed they changed behavior immediately. Conserving ammo no longer feels smart if the game simply backfills scarcity after waste. The deeper point is that adaptation can solve unwinnable-state problems while still poisoning the resource economy that was supposed to create tension.

    If your game depends on scarcity to drive decisions, test what happens when players discover the refill logic. A safety net for unwinnable states should not train players to spend recklessly or second-guess every resource choice.

      Attribution:
    • snailmailman #1
    • qwery #1
    • gs17 #1
    • hnuser123456 #1
  4. 04

    Level scaling breaks the promise of progression

    Oblivion became the thread's canonical failure because leveling can make the world tougher even when your useful combat ability barely changed. The same complaint showed up in Grim Dawn, Tyranny, and Homeworld 2. When earlier areas never become easier and careful preparation makes future fights scale up, the game quietly invalidates one of RPG progression's core pleasures.

    If growth is a selling point, let players occasionally overpower old threats and recover from bad build choices in easier zones. Removing that release valve raises the odds of churn when someone hits a build or gear wall.

      Attribution:
    • ragequittah #1
    • namelosw #1
    • gs17 #1
    • nottorp #1
    • ThrowawayR2 #1
  5. 05

    Granular options help when challenge is multidimensional

    Comments around city builders, Ghost Recon Breakpoint, and puzzle hints made a useful distinction. Some games are not balancing one thing called difficulty. They are balancing combat pressure, survival friction, UI assistance, navigation burden, and puzzle help. In those cases, separate controls can respect very different player goals better than a single Easy or Hard switch.

    For systems-heavy or genre-mixing games, split accessibility and challenge into named components instead of one master difficulty label. That gives players cleaner control and reduces pressure to hide adaptation behind the curtain.

      Attribution:
    • nkrisc #1 #2
    • chihuahua #1
    • proc0 #1
    • Ukv #1
    • rcxdude #1

Against the grain

  1. 01

    Not every game should retain every player

    Several comments rejected the premise that a game must dynamically rescue everyone from friction. The argument was that strong, opinionated difficulty is part of the product, and broad accommodation can flatten the identity that makes a game worth buying in the first place. From this view, player drop-off is not automatically a design failure if the game is clearly targeting a narrower audience.

    Do not optimize retention metrics blindly if challenge and exclusion are central to your game's identity. Decide early whether you are building a broad funnel or a sharper product for a specific taste, then tune accordingly.

      Attribution:
    • orbital-decay #1 #2 #3
  2. 02

    Strict matchmaking can drain the social texture

    While many wanted stronger skill-based matchmaking, others argued that older server-based communities in Warcraft 3, Team Fortress 2, Battlefield, and PUBG produced a healthier mix of rivals, goofing around, and memorable skill gaps. Uniformly even matches can improve fairness while making every session feel samey and transactional.

    If you run PvP, consider whether ranked ladders should be only one lane rather than the whole game. Community servers, custom queues, or looser casual playlists can preserve social variety that pure matchmaking tends to erase.

      Attribution:
    • AussieWog93 #1
    • pphysch #1
    • orbital-decay #1

In plain english

AI Director ↩
An automated game system in Left 4 Dead that changes enemy spawns, pacing, and tension based on how the players are doing.
Diegetic ↩
Presented as part of the game world itself rather than as an external interface or hidden rule.
PvP ↩
Player versus player, meaning human players compete directly against each other.
UI ↩
User interface, the on-screen controls and visual elements players use to interact with a game.

Reference links

Game design references

Videos and informal references