The CS2 Major projected-points simulator re-simulated outcomes that had already happened, so its projections drifted from reality as a major played out. Specifically it ignored the Champions Stage bracket entirely (re-simulating the whole 8-team playoff each iteration), only locked in Swiss stages all-or-nothing once 8 eliminations were recorded, seeded the bracket by world rank rather than real Stage 3 performance, and read qualifying points only from fully-completed events. This conditions each simulation on whatever is already known and only randomizes the undecided remainder: - Champions Stage now honors the real bracket (playoffMatches): actual QF pairings supply the seeding, and any completed QF/SF/Final result is used verbatim instead of re-simulated. Falls back to Elo seeding when no bracket exists. - Swiss stages reconstruct each team's current W-L from real match results (seasonMatches) so partly-played stages lock in their decided games and only the remaining teams are simulated; generalize simulateSwiss to start from seeded records. - Provisional QP already written to event_results for settled (eliminated) teams is used verbatim. - reconcileAdvancers keeps each stage at exactly 8 advancers so ragged mid-stage snapshots can't produce malformed downstream stages. Adds unit coverage for bracket honoring (full and partial), partial-stage locking, and recorded-QP override. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018zMG6JFng8bMqZYweffVuE |
||
|---|---|---|
| .claude | ||
| .forgejo/workflows | ||
| .opencode/plugins | ||
| .production-info | ||
| .storybook | ||
| .vscode | ||
| app | ||
| cypress | ||
| database | ||
| docs | ||
| drizzle | ||
| plans | ||
| public | ||
| scripts | ||
| server | ||
| test-fixtures/baselines | ||
| .dockerignore | ||
| .env.example | ||
| .gitignore | ||
| .npmrc | ||
| .nvmrc | ||
| .oxlintrc.json | ||
| AGENTS.md | ||
| BETTERAUTH_MIGRATION.md | ||
| CLAUDE.md | ||
| components.json | ||
| cypress.config.js | ||
| docker-compose.yml | ||
| Dockerfile | ||
| drizzle.config.ts | ||
| instrument.server.mjs | ||
| oxlint.json | ||
| package-lock.json | ||
| package.json | ||
| plan.md | ||
| react-router.config.ts | ||
| README.md | ||
| server.ts | ||
| TESTING.md | ||
| tsconfig.json | ||
| tsconfig.node.json | ||
| tsconfig.server.json | ||
| tsconfig.vite.json | ||
| vite.config.ts | ||
| vitest.config.ts | ||
| vitest.shims.d.ts | ||
Welcome to React Router!
A modern, production-ready template for building full-stack React applications using React Router.
Features
- 🚀 Server-side rendering
- ⚡️ Hot Module Replacement (HMR)
- 📦 Asset bundling and optimization
- 🔄 Data loading and mutations
- 🔒 TypeScript by default
- 🎉 TailwindCSS for styling
- 💾 PostgreSQL + DrizzleORM
- 📖 React Router docs
Getting Started
Installation
Install the dependencies:
npm install
Development
Copy .env.example to .env and provide a DATABASE_URL with your connection string.
Run an initial database migration:
npm run db:migrate
Start the development server with HMR:
npm run dev
Your application will be available at http://localhost:5173.
Building for Production
Create a production build:
npm run build
Deployment
Docker Deployment
To build and run using Docker:
# For npm
docker build -t my-app .
# Run the container
docker run -p 3000:3000 my-app
The containerized application can be deployed to any platform that supports Docker, including:
- AWS ECS
- Google Cloud Run
- Azure Container Apps
- Digital Ocean App Platform
- Fly.io
- Railway
DIY Deployment
If you're familiar with deploying Node applications, the built-in app server is production-ready.
Make sure to deploy the output of npm run build
├── package.json
├── package-lock.json (or pnpm-lock.yaml, or bun.lockb)
├── server.js
├── build/
│ ├── client/ # Static assets
│ └── server/ # Server-side code
Styling
This template comes with Tailwind CSS already configured for a simple default starting experience. You can use whatever CSS framework you prefer.
Built with ❤️ using React Router.