In this video our new WillFull model is used for a flocking simulation of 500 agents in a confined arena of 1400 x 1400 m, with flocking speed of 8 m/s, under realistic environmental conditions. To visualize the response, we coded the color of agents according to their individual direction of motion. As can be seen, motion is much more coherent than in our previous attempts and there are no severe density fluctuations when the flock hits the corners, as collective decisions are driven by our concept of adaptive leadership.