Select mode + asset
then click ▶ Train
then click ▶ Train
Reward curve
Legend
Ball path (greedy policy)
Bear ball (two-ball)
Nash equilibrium pt
Bullish candle
Bearish candle
How it works
Each candle body = obstacle weighted by volume.
The ball learns to navigate the price chart — paying a penalty when moving through high-volume resistance.
Two-Ball mode: Bull ball starts low, Bear ball starts high. Where they converge = Nash equilibrium = real S/R level.
The ball learns to navigate the price chart — paying a penalty when moving through high-volume resistance.
Two-Ball mode: Bull ball starts low, Bear ball starts high. Where they converge = Nash equilibrium = real S/R level.
Current stats
Candles
—
Bins
80
Path len
—
Nash pts
—