Keeping a naval simulation tied to real designs

Why not optimise the design by filling it with six-inch guns?

That could be effective in the game, but it would no longer be a useful test of plausible capital-ship technology and design trade-offs.

A game-optimised ship could be built around a mass of six-inch guns and a token heavy weapon. That solution would answer a different question from the intended comparison of plausible historical design choices and technical change. The more formal method cross-checks real ship data against the game's values, then repeats each scenario dozens of times while recording the results. The game also lacks fine-grained target-priority and salvo-fire controls, which limits what the test can represent. Its assumptions, available controls, and repeated trials must still address the historical question under investigation.