Consider RISK-style win / loss model

[bildzeitung]

This is with respect to the TODO about whether a winning navy can lose ships.

IMHO, there are two aspects of non-deterministic combat outcomes that could be stimulating in a live setting:

• the potential for upset
• combat can always carry (some) cost, driving the need to protect resources

Perhaps, instead of a deciding a winner and subtracting ships from the loser, subtract ships from all sides with a bias towards larger fleet.

e.g. all sides lose 25% of their ships, but if the ratio between between A & B is 2:1, then A's percentage drops by 1/4 and B's increases by 1/4. Then, finally, add a fuzz factor (proportional to # of fleets involved ?)

The upside is that this can scale in a multi-party scenario -- perform the above calc between A & B, A & C, etc. for a final percentage lost.