Combat power is a matchup estimate, not a fixed conversion between a displayed number and ship count. Evidence
The active roster, attacker and defender cultures, light/heavy mix, infrastructure, directives, SPUs, tactics, and world modifiers can all change a preview. This is why a displayed power value cannot answer “how many ships?” without the exact scenario.
Reproducible example
The 11.75 built-in simulator was run with seed 1175 for 200 trials per culture pairing. The profile used 15 attacking light fighters against 10 defending destroyers with level-5 central infrastructure. In all nine culture pairings, the defender won every run and the attacking force was lost. Evidence
This result proves only that recorded fixture. It should not be generalized to a custom multiplayer roster or a different building, directive, SPU, or tactics configuration.
Composition sensitivity
The estimate changes with the opposing light/heavy share and current modifiers. A force can look strong in one fixture and weak in another without changing ship count.
Reproducible comparison
Keep every input fixed, change one variable, and use recorded seeds. Store exact fleets, cultures, infrastructure, mission, target, modifiers, roster provenance, and build.
Strategic interpretation
Combat output does not include cargo value, siege objective, travel exposure, replacement time, or defense left behind. Those belong beside the simulator result, not inside a universal conversion.