Starting Position: 4
Minor Feedback Mode: None Tailwind Headwind False Promise

Task Description

The agent moves along a 10-cell corridor. The goal is to reach the trophy 🏆 and collect the reward while avoiding the bomb 💣. Switch modes to change the feedback along the path and observe how different environments affect learning.

🎯 Reward

• Reach 🏆: +10, episode ends
• Hit 💣: -10, episode ends
• Tailwind mode: 🍬 small positive feedback along the path
• Headwind mode: 🔥 small negative feedback along the path
• False Promise mode: goal becomes 🍕, only gives +2

🧭 State

• playerPos: current cell index (0 ~ 9)

🕹️ Actions

• 0: none (stay)
• 1: right (move one cell right)
• 2: left (move one cell left)

⚙️ Suggested Parameters

Beginner

• bins = 10, ε = 0.3
• α = 0.5, γ = 0.9
• Recommended: None mode

Advanced

• Switch Tailwind / Headwind modes
• Observe how path feedback affects convergence
• Compare "False Promise" vs normal goal