Regarding size of production runs, keep in mind that Starfleet production capacity has exploded in the past decades. Consider the example of the Renaissance - we're well on our way to building over two dozen of them by 2325, a decade after the prototype finished.
However, this doesn't really impact the 'do we need a new heavy cruiser design' question, which revolves around "do we need much more than 15-20 of this class of ship, ever?"
So this is an interesting question, but as it stands right now, from a pure gameplay standpoint, we actually don't need a large production run for a new ship design. Because ship design and prototyping is in fact no longer a large investment due to:
a) huge increases in our pp and rp incomes
b) research teams pretty much available on demand (and if we don't have one available, purchase one, which would have other uses anyway)
c) prototype costs aren't that significant, and berth space for the prototype is extremely unlikely to be the blocker
Consider the example of a nominal garrison frigate that's simply a Centaur-A but 20sr cheaper and maybe a stat point increase or two (not completely sure this is possible, but should be close). Let's call this the Hippocentaur.
Research: Probably around 20pp to request the design. Research team may or may not be available, but taking into account that we might not need to purchase one, and if we do, it'll have other uses anyway, let's just say it'll cost another 10pp. So 30pp total. And with our rp budget plus the 30rp ship design bonus, rp is not a concern.
Prototype: This is a 80br 50sr 2yr ship, so the prototype costs an extra 40br 25sr and another 1yr of berth space. 40br 25sr is very generously worth about half of a Centaur-A. You just need to build two Hippocentaur instead of two Centaur-As to recoup the extra costs of that prototype. Berth space is hardly going to be an issue, and if it somehow is, 1kt berths are cheap, especially amortized over their expansive lifetimes.
Altogether? You don't need to build
that many ships to justify this new Hippocentaur class, even if it's barely better than the Centaur-A. PP and RP-wise, it's a relative drop in the bucket. Resource-wise, you end up saving them after just two builds. And meanwhile, there might be a Centaur-B refit, but that's orthogonal to this, because you'd expect the Hippocentaur to be competitive with the refit, and you'd still have a possible refit for this new class later on at the piddling cost of 20pp or so.
Of course, this is a specific example, but the point is that under the current ship design cost structure, small incremental improvements and small intended production runs could still technically warrant a new ship class.
I can think of narrative reasons either way for why this state of affairs should or should not exist. For ex, there's supposed to be economies of scale and training costs and so forth... but on the other hand, most Starfleet pattern starships share a lot of components anyway and berth/crew training is likewise bound to be fairly similar across Starfleet. Star Trek canon Starfleet is riddled with ship classes, but on the other hand, TBG Federation/Starfleet is a much smaller affair.
Ultimately, I think Oneiros will have to make a decision on whether he wants to encourage this sort of thing, or to restructure the ship design costs.
edit: oops I meant 2yr ship, not 3yr ship
