More detailed attack probabilities

[rusu24edward]

The attacking agent has an attack accuracy parameter that applies for all of its attacks. It would be interesting to define an attack accuracy between types of agents, something like:

attack_accuracy={
    1: {2: 0.8, 3: 0.5, 4: 1},
    2: {1: 0.5, 3:0.1},
    ...
}

which would be interpreted as "agents with encoding 1 attack agents with encoding 2 with accuracy 0.8, 3 with accuracy 0.5, etc.

Challenge: How would this play with the agent's intrinsic attack accuracy? The accuracy mapping is by encodings the attack accuracy is by agent...