If you are saying that, say, 2x cap of a low cap ships is inherently more valuable than a 2x cap of high cap ships (even if cap stats are adjusted to factor in their different pros/cons properly), then that is a much more interesting "problem".
That was (basically) what I was saying. And that the AI should therefore pay more for a cap of SBSs than gatlings.
Hmm, well there is something that the current AI cost "model" (ship caps only right now) does not account for. Resource costs and build time costs (both cap and per unit). This means that the AI is at an advantage when getting a ship type with high build costs (it doesn't have to pay metal and crystal to churn them out, we do, thus removing one balance point) and build times (the AI can spawn the unit instantaneously the moment it decides it wants some, which includes replacement times, thus removing another balance point). For ships close to the standard cap, these things don't really skew the cost "model" too much. But in the case of extremes (very high cap and low cap ships), this lack of accounting for a very important part of the ship's balance starts becoming very noticable (cheaping the use of high cap ships in AI hands, and boosting the usefulness of low cap ships in AI hands)
It would be tricky, but a revised formula for AI "cost" that takes into account not only ship cap but also individual build times and cap build times (but much more focus on individual build times) I think would go great ways to alleviating this issue. (And thus, targeting AI abuse of low cap ships without penalizing low cap ships explicitly, and also buff the AI use of low per ship build time high cap ships as well)
(Notice, I am not calling for ship resource cost to be factored in. I am assuming that the AI has enough resources stockpiled such that they can really build anything, and indeed have a large stash of them built "elsewhere")
There are simply too many shots per second for overkill to be as significant a factor as damage decay.
Agreed, it is less of a factor (in large part because we have ships with 5/96 ~= .052 the standard cap (5 ships), but no ships with 96/5 = 19.2 times the cap (1843), aka, we are more skewed towards the lower end. Also a large part is the lack of physics/terrain effects (almost all ships can target and hit anything in their range, regardless of how much is in the way) and part of it is the "has my target died before my shot reached" refund)
Which is one of the reasons why lower cap ships should have a slight penalty to their cap stats (which ones, I am not sure yet) compared to what high cap ships get to their adjustment.
Also, many of the high cap ships (due to being older than most of the low cap ships) have not been updated recently, so many (but not all) of their cap stats for many (but not all) high cap ship types are currently just plain inferior in every way.
In other words:
I think this is more an issue with high-cap ships not being strong enough. Once high and low cap ships are equally deadly (roughly), it will be easier to judge if the AI is getting too many ships.
The fact that we see ~20 caps of low-capppers but not triangles also bears investigation.
Yea, I suspect that either there is a lurking bug somewhere (perhaps just insidious rounding error), or the "debt is canceled at the end of the galactic reinforcement cycle" is a no longer negligible effect now the planetary ship type "preferences" can be applied to some planets.
Sadly, neither rounding error nor the effect of the second thing I described are things easily gleaned from the log files in terms of long running game effects.
