Most game design advice is prescriptive: "add juice," "give feedback," "respect the player's time." Useful rules of thumb — but rules without axioms don't tell you why they work, or what to do when they conflict.
This is a working theory. Three axioms, derived deductively from neuroscience and motivation research, that together explain where intrinsic fun comes from. It's hypothesis-stage — no original experimental data yet — but each axiom has independent empirical backing from existing literature. The goal is to give game designers a reasoning foundation, not just a checklist.
Epistemic status: Literature-validated hypothesis. Cross-referenced with Schultz (1997), Csikszentmihalyi's Flow Theory, Ryan & Przybylski (SDT, 2006/2010), Nunes & Drèze (2006), and Murayama et al. (2017). Not yet validated through original playtest data.
The Three Axioms
These three axioms are independent — each draws from a separate neurological or psychological mechanism, and each can produce fun without the others. But they also amplify each other when combined.
AX-001 — Prediction Error (Dopamine Loop)
Fun emerges from the gap between expected and actual outcomes. When your prediction is perfectly accurate, there's no signal — you're bored. When the gap is too large, there's no pattern to latch onto — you're confused. The sweet spot is calibrated surprise: outcomes that violate your prediction just enough to generate a dopamine response.
This is the neuroscience of Schultz's 1997 dopamine prediction error research, and it maps precisely to Csikszentmihalyi's Flow Theory at the game design level: too easy (complete prediction) → boredom; too hard (prediction fails entirely) → anxiety; flow zone → calibrated error → fun.
The design implication: difficulty curves aren't about "challenge." They're about maintaining the prediction error in the productive range as the player's model improves.
AX-002 — Agency (Choice → Consequence)
Fun requires the player's perception that their choices affect outcomes. Remove this and the same events become hollow — you're watching, not playing. Ryan & Przybylski's Self-Determination Theory identifies autonomy as an independent motivator, separate from competence or relatedness.
The key word is perception. Players don't need full control — they need to believe their decisions matter. This is why Hades' Heat System is so effective: it lets players directly configure their own difficulty curve, turning AX-001 calibration into an act of AX-002 agency.
The design implication: every player action should produce a visible, traceable consequence. The moment players suspect their choices are cosmetic, agency collapses.
AX-003 — Competence Growth (Mastery Trajectory)
Fun also comes from perceiving your own growth over time — independent of moment-to-moment surprise. This isn't the same mechanism as AX-001. Nunes & Drèze (2006) demonstrated this directly: endowed progress toward a fully predictable goal doubled completion rates even when the outcome was certain. There's no prediction error involved. The satisfaction of growth is its own circuit.
Murayama et al. (2017) confirmed this at the neural level: the dopamine/prediction-error circuit and the learning-progression meta-monitoring circuit are neurologically distinct. Growth feels good for a different reason than surprise.
The design implication: players need a visible model of their past self to perceive growth. Games that obscure progression (no stat history, no skill comparison) rob players of AX-003 even when the growth is real.
Why Three Axioms Matter More Than One
The common mistake in game design theory is collapsing everything into one variable — "engagement," "fun score," "retention." The framework above resists this for a reason: the three axioms fail differently.
| Axiom Missing | Symptom | Common Misdiagnosis |
|---|---|---|
| AX-001 (No prediction error) | Game feels "flat" — technically fine but unstimulating | "Needs more content" |
| AX-002 (No agency) | Players disengage despite enjoying the visual/audio | "Needs better UX" |
| AX-003 (No growth arc) | Strong early hook, sharp drop-off after ~30 min | "Needs a tutorial" or "needs content pacing" |
Diagnosing which axiom is missing is more useful than measuring aggregate "fun" — because the fix for each is completely different.
Applied: Designing the First 5 Minutes
A concrete test: what does a first-time player need to experience in their first 5 minutes to receive all three axioms?
If any of these three moments is missing or delayed past the 5-minute mark, you're likely to see session-end data reflect exactly which axiom failed. This is an actively testable hypothesis — it's what we're doing with ProjectPocket's AutoPlayer validation pipeline.
What This Framework Doesn't Explain
Social fun (cooperation, competition, status) is outside this framework. So is narrative immersion and aesthetic pleasure. These aren't less important — they're just different mechanisms that likely have their own axiom structures.
The claim here is specifically about intrinsic, single-player, mechanic-driven fun. It's the kind most directly under the control of systems designers, and the kind most likely to be lost when a game "has everything" on paper but still doesn't feel good to play.
Evolution Log
- 2026-04-02 — Initial publication. AX-001/002/003 framework derived from literature synthesis. Confidence: AX-001 (0.75), AX-002 (0.65), AX-003 (0.70). Active validation in progress via ProjectPocket AutoPlayer pipeline.