Chapter 26 Key Takeaways

The core ideas from this chapter, distilled into points you should be able to recall and apply without re-reading.

  • Combat has the worst design-to-play ratio of any system in your game. A ninety-second fight may represent four designer-months of work. Budget accordingly, and accept that combat will be iterated on more than anything else in the project. Teams that ship great combat are teams that survived their own iteration count.

  • Pick a combat paradigm and commit. Action, strategy, or hybrid. Borrow from other paradigms only to sharpen your own. Middle-of-the-road hybrid combat that refuses to choose is the most common way bad combat ships. The player fantasy you wrote in Chapter 1 tells you which lane you belong in.

  • Fewer combat verbs produce deeper combat, not shallower. Hyper Light Drifter has three verbs. Hollow Knight has four. Sekiro's combat routes through one — deflect. Depth is the interaction between verbs, not the count. When you add a verb, defend it; if it does not do something no other verb does, cut it.

  • Frame data is your design dial. Every attack has startup, active, and recovery frames. When an attack "feels off," the fix is almost never more damage — it is retuning one of those three windows. The vocabulary comes from the fighting-game community and applies to every real-time combat system you will ever build.

  • Hit pause is the single highest-leverage game-feel technique in combat. Freeze the attacker and target for 4-10 frames on connection. Without it, attacks read as whiffs. With it, impacts read as impacts. If you implement only one game-feel effect in your combat, implement hit pause first.

  • Legibility is the designer's gift to the player. Every enemy attack needs a tell the player can read — an animation wind-up, a visual effect, or an audio cue. Aim for a 2:1 tell-to-active ratio minimum. The "I died but it was fair" feeling of Dark Souls is built entirely from readable telegraphs. Without them, combat becomes a stat-check instead of a skill-check.

  • Difficulty that only changes HP and damage numbers is a smell. Real difficulty changes what the player has to do — new attacks, coordinated AI, fewer affordances, new mechanical systems. Number-tuning makes fights longer, not harder. The Bloodborne BL4 community is evidence: Bloodborne survives its own numbers being stripped because the mechanics carry the difficulty.

  • Boss fights follow a three-act skeleton: intro, phase 1, transition, phase 2, (optional phase 3), kill moment. Phase 3 should almost never introduce new mechanics — it is an execution test on what the player already knows. The final phase is for mastery, not learning.

  • A good boss fight has a rhythm you can hear with your eyes closed. Playtest by recording 60 seconds of the fight and listening to the audio alone. A musical pulse with tasteful variation is the target. Metronomic is boring; arrhythmic is unfair; jazz is right.

  • Encounter design is enemy composition plus terrain, not HP inflation. A single strong enemy is a duel. Multiple enemies is a crowd-control problem that demands an area tool. Mixed enemy types (melee + ranged + flying) tax the full verb vocabulary — this is Halo's "30 seconds of fun" loop. Build arenas with elevation, cover, hazards, and affordances; flat featureless arenas are the weakest combat terrain.

  • Lock-on is for duels. Free camera is for crowds. If you support both, budget six months for camera work. The camera is combat's most ignored system until it breaks, at which point it is all the player complains about.

  • Data-drive your boss phases. Phase-specific movesets should live in a resource file, not hardcoded in script. This lets designers iterate without needing programmer intervention, and it is the single biggest factor in how fast your combat team can tune a boss.

  • Combat is never finished in one pass. You will extend your combat system across Chapters 27 (AI behavior), 30 (audio), and 32 (balancing). The CombatSystem.gd and BossFight.gd built in this chapter are skeletons; the meat is added over the rest of the book. Expect this. Plan for it.