A Designer's Checklist — Randomness in Action Games

Eight posts in. I have been covering a lot of ground and I want to pull it together into something practical.

This final post in the series is for game designers — specifically those building action games where skill mastery is part of the promise. If you are wrestling with a randomness decision in your design right now, run it through this checklist.

But first, one recap.

## The Thread Running Through All Eight Posts

Every post in this series has been orbiting the same core idea: randomness is a design tool, and like any tool, its usefulness depends on where and how you apply it.

In mastery-focused action games, the fundamental promise is that players can get better. Randomness serves that promise when it enriches the experience for skilled players without destabilizing the foundations those players are building on. It breaks that promise when it makes the game's core loops unlearnable or obscures the feedback players need to improve.

The cases we explored:
- The GunVein technique: micro-variance inside learnable macro patterns (good)
- Core weapon randomness: variance at the loop players practice most (bad)
- The skill expression stack: layers 1-4 and where randomness fits at each layer
- Enemy spawning: fixed foundations with potential for behavioral micro-variance
- Roguelite structure: legitimate alternative, but a different game, not an upgrade
- Perceived randomness: communication design matters as much as actual randomness

Now the checklist.

## The Checklist

**1. Which layer of the skill stack does this touch?**

Before doing anything else, identify whether the randomness you are considering lives at Layer 1 (survival basics), Layer 2 (pattern recognition), Layer 3 (optimization), or Layer 4 (expert improvisation). The higher the layer, the more room randomness has. Layer 1 and 2 variance should be approached with significant caution.

**2. Is the thing being randomized something players need to build a mental model of?**

If yes — proceed very carefully. Mental model formation requires consistency. Players cannot internalize a system that behaves differently every time. If your core loop needs to be learned, it needs to be stable.

**3. Is the variance macro (structural) or micro (within-encounter)?**

Macro variance (stage order, enemy spawning, item drops) changes what players practice. Micro variance (slight bullet offsets, minor timing jitter) enriches practice of the same thing. Know which you are doing and be intentional about it.

**4. Can a skilled player predict approximately what will happen?**

Run the test. Imagine a player who has seen this encounter many times. Can they form a reliable prediction? If yes, the variance is probably enriching. If no — if even experienced players have no useful prediction — the randomness is chaos rather than texture.

**5. Does the randomness create meaningful decisions or just outcomes?**

Good randomness creates moments of choice or adaptation. "I saw the deviation early and adjusted" is a skill moment. "I died because the bullet went somewhere unexpected" is not — unless the system includes tells that a skilled player can learn to read.

**6. What is this communicating to the player about whether the game is learnable?**

Players who see randomness will often conclude the thing cannot be mastered. If you are adding variance to something you want players to go deep on, ask whether your visual and feedback design makes it clear that there is structure to discover. Learnable things should look learnable.

**7. Are you making a mastery game or a variety game?**

This is the meta-question. Mastery games reward depth. Variety games reward breadth. Both are valid. But the role of randomness is very different in each. Heavy structural randomness in a mastery game is usually a mistake. Low randomness in a variety game often makes runs feel samey. Know what you are building.

**8. Would you be comfortable removing this randomness entirely and seeing what breaks?**

This is a useful stress test. If removing the variance makes the game clearly worse in a specific way you can articulate — patterns become too autopilot-able, runs feel too samey, something is missing — then the variance is load-bearing. If the game seems essentially the same without it, the variance was not doing much work.

## Where IS2 Lands

I have run every major design decision in IS2 through some version of this checklist. The short version of where the game ends up:

No randomness in the core weapon system. Stages are fixed and learnable. Enemy formations are scripted. Behavioral micro-variance exists on specific enemy types to keep skilled play demanding. No roguelite structure — the game rewards knowing the game, not adapting to a new run.

The demo reflects all of this. If you have been reading this series and want to test whether the philosophy translates to the actual feel of the game, that is the fastest path: [Play the IS2 demo on Steam](https://store.steampowered.com/app/3486180/Interstellar_Sentinel_2/)

## Closing Thoughts

None of this is absolute. There are great games that break every principle I have outlined — games where core-loop randomness works, where heavy macro variance coexists with deep mastery, where the player never knows quite what is real and still has a fantastic experience.

The checklist is not a rulebook. It is a set of questions to ask. The answers will depend on what your game is promising and who it is for.

What I hope this series has done is make the *tradeoffs* visible. Every randomness decision has costs and benefits. Understanding them clearly lets you make the choices deliberately, rather than landing somewhere by accident and not knowing why it feels wrong.

Thanks for following along. If any of this was useful for something you are building, I would genuinely like to hear about it.

---
*Previous: Post 7 — Perceived Randomness: What Players Think They're Fighting*

*This concludes the Randomness in Shmup Game Design series (Posts 1–8).*