The Skill Expression Stack — What Players Are Actually Learning in Your Game

This is the fourth post in a series about randomness and skill in game design. If you have been following along, you know where I am coming from. If this is your first stop, here is the short version: randomness is a tool, its placement matters, and the best designs protect the things players can get good at while using variance to create interesting context for those skills.

This post is the one I have been building toward. I want to put a name to something I have been thinking about for a long time: the **skill expression stack**.

## What the Stack Is

When a player is getting better at an action game, they are not learning one thing. They are learning a layered set of things simultaneously — and those layers have a natural hierarchy.

Here is how I think about it for shmups specifically:

**Layer 1: Survival basics.** Where are the bullets? How do I not die? This is the floor. Every player starts here. The game's job is to make this layer learnable quickly enough that players do not quit before they reach the interesting parts.

**Layer 2: Pattern recognition.** Wait, this enemy always does this. This wave has a route. The stage is scripted — I can learn it. This is where shmup players live for most of their time with a game. It is engaging, deeply satisfying, and has a clear progression of competence.

**Layer 3: Exploitation and optimization.** I know the patterns. Now: how do I play them *well*? How do I maximize damage during safe windows? How do I position to handle multiple threats at once? How do I route through a level efficiently? This is where the game opens up for dedicated players.

**Layer 4: Improvisation within mastery.** I know the patterns, I know the routes, and now something unexpected happened — a bullet deviated, I made a mistake, I am in a position I have not trained for. Can I improvise successfully? This is the highest layer. It is where GunVein's light randomness technique lives. It is what separates good players from great ones.

## Where Randomness Belongs in the Stack

The reason randomness in the core weapon system is so damaging is that it collapses layers 2 and 3. If weapon behavior is random, you cannot build reliable pattern-recognition strategies. You cannot optimize exploitation windows. The stack has a hole in the middle.

Players can survive at Layer 1 through trial and error. They cannot reach Layer 4 because the foundation of Layers 2 and 3 is unstable.

The GunVein technique works because it operates *exclusively* at Layer 4 — and only for players who have already built solid foundations at Layers 2 and 3. It does not disrupt the stack. It adds texture at the top.

This is the principle: add randomness only at layers your players have already mastered the foundations below.

## What This Means for Design Decisions

When you are making a design decision about randomness, ask yourself: which layer of my skill stack does this touch?

- If it touches Layer 1 (basic survival), be very careful. Making basics unreliable will frustrate new players and may prevent them from reaching depth.
- If it touches Layer 2 (pattern recognition), be extremely careful. This is where most players spend most of their time. Disrupting this layer is disrupting the game's core loop.
- If it touches Layer 3 (optimization), consider carefully. Some high-level variance here can create interesting decision-making. Too much collapses the optimization game.
- If it touches Layer 4 (expert improvisation), this is where randomness has the most room to work. Players at this layer have the tools to handle variance. It enriches their experience without destabilizing anyone below them.

## IS2 and the Stack

Interstellar Sentinel 2's combat mechanic — the reason you don't just hold the fire button — lives primarily in Layers 3 and 4. Casual players can play through Layers 1 and 2 without deeply engaging with it. As skill grows, the mechanic becomes part of optimization. At the highest levels, timing it correctly under pressure is a Layer 4 expression.

The enemy patterns themselves are non-random at their core, giving players solid Layer 2 foundations to build on. I am experimenting with some GunVein-style micro-variance in specific late-game encounters — but only where the macro pattern is already well-established.

The demo has the early game. The skill stack is visible if you go looking for it. I would be curious whether it feels like a game that wants you to get better at it.

[Try the IS2 demo on Steam](https://store.steampowered.com/app/3486180/Interstellar_Sentinel_2/)

---

*This concludes the Randomness in Shmup Game Design series. Thanks for reading — and if any of this sparked thoughts about a game you are building or playing, I would love to hear about it.*

*Previous: Post 3 — Why I Don't Randomize IS2's Core Weapon*