Spaces:

---

kaushikvr06

/

reasoning-simulator

like 1

Build error

Thank you for trying. The expectation for this demo/simulation is that the allocations will total 100%. If we either under-allocate (< 100%) or over-allocate (>100%), the game doesn't move forward and displays an error in the game status. Please see the snip attached for the under-allocation scenario.

I'll make changes later today to make this more intuitive/obvious (perhaps by adding a small demo video/GIF of the app in action). I appreciate the feedback!

Please see a short demo of me playing the simulation against the AI. I've just added a demo of the last 4 quarters to keep the video short (and within the 50 MB limit). The AI adapts its strategy throughout the rounds, depending on its budget and the percentage of market share captured, while also considering whether its product is superior. The AI is strong and adapts reasonably well. Still, since this is merely a simulation, the AI opponent uses a heuristic-based strategy model designed to mimic the emergent behaviors and outcomes of self-play reinforcement learning (detailed in the SPIRAL research paper).

Please review it and let me know if you have any additional feedback. Thanks for trying and engaging!