Update the attack meta

[evanpelle]

The current attack meta is unintuitive as larger players can be given too much of an advantage. It also makes it very difficult for multiple smaller nations to gang up on the crown.

We should keep the following:

1. Larger attacks should give a troop loss boost to reward & encourage a single large attack vs many tiny ones.
2. The meta must change slowly over time. The initial meta should be fairly similar to the current one, and we update it to the desired meta after several releases.

[binmogit]

If assigned, add a conservative two-tier bonus so large, committed attacks lose slightly fewer troops.

• Tier 1 (>= 50% of max troops): small efficiency bonus (default 3%).
• Tier 2 (< 50% of max troops): no bonus.

Rules:

• Implement in DefaultConfig.attackLogic() and apply the multiplier last (base → defense modifiers → efficiency).
• Attack executions capture the attacker's max troops at commit time and pass it to attackLogic()
  • Include troops on transports when computing max troops to avoid boat-stacking exploits (optional).

Config knobs (defaults):

• attackTierThresholds = [0.5]
• attackTierEfficiencyBonuses = [0.03]

Quick tests:

• Test values just below, at, and just above each configured threshold — assert the attackerTroopLoss multiplier is only applied when the committed fraction >= threshold.
• Deterministic before/after aggregate loss comparisons (verify variance ≤ 3%).