IPL Oracle is an intelligent, interactive guessing game inspired by "Akinator," specifically designed for the Indian Premier League (IPL) ecosystem. Unlike traditional games that use static decision trees, IPL Oracle leverages Generative AI and Probabilistic Reasoning to identify any IPL player (past or present) through a sequence of dynamic, high-impact questions.
- Dynamic Questioning: No hardcoded logic or static if/else flows. The AI generates context-aware questions in real-time based on your previous answers.
- Information Maximization: Every question is designed to maximize entropy reduction, narrowing down the pool of all IPL players from 2008 to the present.
- Confidence-Based Logic: The engine triggers a final guess only when it reaches a confidence threshold of ≥ 80%.
- Adaptive Learning: The system incorporates user feedback ("No, that wasn't my player") to improve future predictions and knowledge.
- Efficiency: Designed to identify the target player in a strict maximum of 8 questions.
The project uses a modern, high-performance stack to ensure a smooth and engaging user experience64:
- Frontend: Built with Tailwind CSS for a clean, "classic and aesthetic" UI that is fully responsive.
- Backend: Powered by Node.js, handling the asynchronous AI engine loop and candidate probability updates.
- AI Engine: Integrated with the Gemini API for dynamic question generation and reasoning logic.
- Database: Utilizing Google Sheets as the foundational database for historical IPL player stats.
The system operates on a sophisticated AI-driven deduction loop:
- First Question: AI asks a broad, high-impact question.
- User Input: User responds with Yes, No, Maybe, or Don't Know.
- Probability Update: The system updates candidate probabilities across the entire IPL player pool.
- Optimal Selection: AI selects the next question based on the updated pool to minimize steps.
- Final Prediction: Once confidence is high, the AI makes its guess: *"You're thinking of [Player Name], right?.
The project is optimized for the following criteria:
- AI Reasoning (Very High): Ensuring the system is genuinely dynamic and avoids hidden hardcoded scripts.
- Question Intelligence (High): Questions are contextually relevant and information-rich.
- Innovation & Learning (High): The ability to adapt from mistakes and user feedback.
- Accuracy (High): Consistently guessing correctly within the question limit.