Norman's method should work, it's a little wonky, in that it won't represent any rock / paper / scissor dynamics. Someone could theoretically find the cheapest ship with the highest combat strength and always auto resolve combat, and they would win out over someone who has a more expensive but more well rounded force.
then take it to the next level.
ships have 2 stats: combat strength, and speed/maneuverability.
or the next level:
attack, defense, range, and movement.
but no matter how simple or complex the model, balance will be required to prevent min-maxing from creating a dominant strategy - IE just building the ship with the best strength to cost ratio. but then you end up with all ships having the same cost/benefit ratio. Ie all choices become meaningless. perhaps have stronger ships cost progressively more, as they make it easier to bring a lot of force to bear at a single point - easier than gathering up a whole squadron of smaller ships to do the same job.
one of the big difficulties is the hands off nature of the game. to make fleet selection meaningful, you need in-depth combat rules. but the combat is hands off, its just the emperor watching the war on closed circuit tv from a surveilance satellite basically. and the emperor is not expected to know the nuances of plasma cannon usage in space combat. but they are expected to make informed meaningful decision as to what to build. its hard to add that gameplay feature and keep things high level.
The trick I intended is that you need/prefer different ships (fleet composition) vs different enemies. So, at the worst case there would be several dominant strategies (each applied on a separate border vs specific alien race).
Also, you can observe the batle and get a lot of detailed information how it works. So, while you can't affect the tactical battle you get tons of information how it works and you can base your decision on this.