[Leternel]

The problem with AI is the same with human. How confident are you in your prediction.
Imagine two emperors of SE, A and B, who can see each other's weaponry.
A is making a fleet of missileship, and is now thinking about a smaller but hidden fleet to defend it. How will he choose?
Point-defense corvets, since B will build lot of fighter which are protected against missiles?
Or cruisers because B will not build fighters, as he knows A wants to build a defending fleet, so he will build cruisers with point-defense?

Thinking about it, A as another probleme: "must I continue building my missileships if B can counter them?".

In most case, B will never have to defend himself, because A will hesitate.

If B had only 2 choices, fighters or PDF-cruisers, A could crush him with high-range torpedoes, but as A and B are equals, they both stuck in the trap/coutner-trap/counter-counter-trap/...
Against a human, you will be like A and B.
Against an AI, you will encounter a quite different problem: what level of AI is it? trap? counter-trap? CCC-trap?