In the default model, the team which maintains a higher population tends to be the team which begins the simulation with more strength among its agents. Due to the psuedorandom nature of the simulation, this is not always the case, regardless a typical run may look something like:
Placing the blue team agents closer to each other at initialization does not offer the team any leverage over the red team. Winning is still purely determined by the encountering agents’ strength.
Now say that if a battle occurs where blue has 2 neighbors within a certain distance, the red agent will become outnumbered and lose the fight. We see here that blue does substantially better.
Re-exploring strength, we now increase the initial strength of all the red team’s agents. Also changed in this run is that we now say blue must have 5 neighbors to guarantee overpowering a red agent. Here we find the red team outperforming blue.
The blue team is not sticking close enough together! This next run confines blue agent’s initial position and delta movements more than in the previous runs. Now the blue team is working together and defeating red again.
Having the blue team remain even closer together reduces their population loss. When the team sticks together the individuals are less likely to lose the fights.
Removing the restrictions on the blue team’s positions and delta movements causes the team to lose all their members nearly immediately.