T-Rex vs Tiger Simulator
티라노사우루스 N마리 vs 호랑이 M마리가 2D 오픈필드에서 싸우면 누가 이길까? 문헌 기반 파라미터 범위에서 불확실성을 샘플링하는 몬테카를로 에이전트 기반 시뮬레이션으로 승률을 추정합니다. 기본 시나리오는 T-Rex 10 vs 호랑이 100.
A Monte Carlo, agent-based battle simulator: N Tyrannosaurus rex vs M tigers on a 2D open field, with literature-based parameter uncertainty sampling. Runs entirely in your browser.
Live: https://trex-vs-tiger.vinylims.com/ · Releases: GitHub Releases · CHANGELOG
Features
- Species choice on each side (heavy: T. rex, elephant, buffalo, brown bear, woolly mammoth, Homo erectus; pack: tiger, lion, wolf, hyena, Homo erectus) or a matchup preset, plus your own counts, number of Monte Carlo runs, seed, arena size, time limit, pack tactic and morale options
- An editable
(min, mode, max) range for every model parameter
- Latin Hypercube or plain random sampling, plus per-individual variation
- Results:
- win / draw probabilities with Wilson 95% confidence intervals
- histograms of battle duration and survivors
- Spearman sensitivity (tornado chart) of each parameter
- Animated replay of a single battle with abstract glyphs, an event timeline, skip-to-contact, speeds 0.25×–16× and a reduced-motion overview
- Scenario sharing as a URL; the seed is included, so results reproduce exactly in the same browser engine (and statistically across engines)
- Settings export and import as a
TVT2-… code or a .json file (previewed before applying)
- Result sharing: a
/r/ result page with a snapshot and "run it yourself" check, a story or square result card image, and a Wordle-style emoji text; link previews show the result through a tiny /r/-only Pages Function
- Korean (
/) and English (/en/) UI, plus static guide, model, parameter, about, patch notes, privacy and license pages in both languages; generated sitemap, robots.txt, llms.txt and structured data
- Interactive guided tour (header 튜토리얼 / Tutorial button or
?tour=1)
- Release version badge that detects newer deployments and can reload the latest version after clearing this site's cache and storage
- No backend: every simulation runs in your browser (the only server code rewrites link-preview text on
/r/, ADR-0006)
Quick start
npm ci
npm run dev # http://localhost:5173
npm run check # typecheck + lint + format + unit tests
npm run build # → dist/
Node: see .node-version (24).
Documentation
| Doc | Content |
|---|
| ARCHITECTURE.md | System, modules, run sequence, state machines, pipeline (Mermaid) |
| docs/MODEL.md | Simulation model, parameters, sources, assumptions |
| docs/DEPLOYMENT.md | Cloudflare Pages setup and free-plan budget |
| docs/CONFIGURATION.md | Every config file explained |
| docs/adr/ | Architecture Decision Records |
| docs/SPECIES.md | Species roster, presets, bounds, sources, calibration |
| docs/PRESETS.md | Simple-mode matchup presets, measured fixture, rout check, banned words |
| docs/REPLAY.md | Replay renderer v2: glyphs, camera, timeline, accessibility |
| docs/SEO.md | URL languages, head tags, sitemap/robots/llms, IndexNow, owner actions |
| docs/PAGES.md | Static content pages, URL scheme, drift guard |
| docs/TUTORIAL.md |
Disclaimer
This is an educational toy model. T. rex and tigers never coexisted, and every interaction rule is an assumption. The output shows how the outcome depends on uncertain inputs. It is not a prediction of a real fight. See docs/MODEL.md.
License
MIT
