This is a gacha simulation and planning tool prepared for Arknights: Endfield. Its core goal is not to simply recreate the entertainment loop of "single pull, ten-pull, and success animation", but to help players answer a more practical question: when resources are limited, goals differ, and banners keep rotating, which pulling strategy is closer to their own optimum?
The project is divided into two main pages. The simulator page recreates the banner pulling process and supports past, current, and possible future banner configurations. The planner page focuses more on strategy analysis. Players can enter their resources, target characters, expected savings, and other conditions; the system then uses Monte Carlo simulation to estimate the risk and expected value of different plans, producing a score and recommendation.
The backend is built with Python Flask, with Waitress used as the production server. It handles banner data, probability calculation, and planning logic. The frontend uses vanilla JavaScript and HTML to implement the interactive interface, with attention to responsive layout, dark mode, and a low-dependency structure. Rather than trying to become a complex platform immediately, the project first aims to make the core loop calculable, explainable, and iteratable.
The project is still in progress, and is a reserved address for now. Later, I plan to add richer data visualisation, user data import, plan comparison, and banner version management. At its core, this is also a system design exercise: breaking the uncertainty of a game into data, probability, weights, and decisions, then making those decisions understandable to the player.