Repeated effects with controlled seeds

This deck shows how repeated player effects can vary while remaining reproducible. A synthetic NixieFX burst creates 24 particles and samples them at 0.35 simulated seconds. Two runs with seed 42 return the same particle positions. A run with seed 99 returns a different position array and a visibly different pattern. The comparison uses nixie-fx 0.1.17 and Three.js 0.185.1, with a fixed camera and a shared effect definition. It compares positions before translating the instances for display. The slides explain which inputs to record, how camera-facing transforms can confuse a test, and what a seed policy belongs to the host game. This local test does not establish cross-backend pixel equality, device performance or completed human technical review. AI created the deck, source and synthetic demonstration for a marketing assignment. The NixieFX documentation provides the backend reference: https://nixiefx.com/vfx-runtime-docs/

Transcript

1. Repeated effects with controlled seeds A 24-particle burst at 0.35 simulated seconds Seed 42 repeats the same positions Seed 99 changes the pattern
2. The inputs held constant One exported seed-burst effect nixie-fx 0.1.17 and Three.js 0.185.1 Fixed time and camera for the comparison
3. Local positions match for seed 42 First sample: x -0.127671, y 0.073710 The repeated seed produces the same position array Layout translation follows the equality check 42 = 42
4. Seed 99 changes visible variation First sample: x 0.292530, y -0.328630 The sampled position arrays differ All variants retain the same count and lifetime 42 ≠ 99
5. A seed is an input to reproduction Record the seed with the effect revision Record the simulation time and update schedule Keep the original fixture with the issue report
6. Different origins are a separate variable Compare positions before moving instance roots Camera-facing matrices can differ by view A translated particle array needs normalized coordinates
7. Repeatability check in the host Create two instances with seed 42 Seek both to 0.35 seconds in the test Compare captured particle positions
8. Variation policy belongs to the game Choose a repeatable event-to-seed rule Keep that rule stable during debugging Avoid claiming distinct seeds guarantee unique images
9. The evidence has a narrow scope This test compares one effect in Three.js It does not compare cross-backend pixels It does not measure device frame time
10. Runtime review and ownership NixieFX documentation supplies backend context Review export warnings for the actual settings Destroy effects when the owning scene ends