Warning: wall of analysis below
Or, you know, we can do new design that will have similar or better stat on the same crew and wouldn't obsolete as soon (and which in turn can be refittet if needed long after current ones will be obsolete).
Or, you know, we can scrap or mothball obsolete ships freeing said crew for newer ones. The only reason we didn't do it and gone for refits is because refits looked faster and not as straining on our shipbuilding capacity.
Spent some time thinking about this and did a more complete analysis:
First, I want to address your "future refitability" argument separately: Suppose we have ship designs A, B, and C, where A is an old ship class, B is the last possible refit of A, and C is a completely new ship class that has comparable stats/costs efficiency with B (and has more refit potential). Suppose we have a ship of class A, and we want to decide whether to refit it to class B, or replace it with the new build of class C. Let's also suppose that refitting to B vs scrap A + new build of C is equivalent in cost somehow (it's not in reality, especially in berth time) to eliminate another variable.
Now, if we always choose to replace with class C because of future refitability, then this situation will keep repeating itself, and we'll never exercise that future refitability to its fullest, since that last refit simply won't be done. So by itself, the lack of future refitability when determining whether to do a refit does not matter.
It
does matter when considering between new ship builds, or when the B refit is so outdated (relatively inferior in stats/cost efficiency compared to C) and cannot be redesigned to be comparable in stats/cost efficiency to C. The latter situation typically shouldn't happen within two tech tiers and with the upcoming ability to design custom refits.
Now when considering stats, there are indeed cases when scrapping/mothballing an older ship and building a newer ship is better than refitting that older ship,
but it's a higher bar than you might expect. Refits are still faster than new ship builds and are economical when we have sufficient berth capacity, which we do. So the replacement option needs to be sufficiently good to beat that advantage.
Let's look at the closest example we have right now, which is whether to scrap/mothball our 2280s-era Constellations for 2300s-era Renaissances or to refit those Constellations. I'll also try comparing with the Constellation => Centaur-A replacements due to similar capabilities, but this replacement isn't as clear to analyse, since escorts are relatively expensive for br/sr and cheap on crew, such that at same crew costs, BR and SR costs in Centaur-A become prohibitive.
If you're considering a completely new ship design, well, that will cost more pp (more than the 23pp of the Constellation refit), research time, and prototyping time, so refits are still far superior in terms of cost&time-to-production. And it still isn't necessarily mutually exclusive with refitting, especially if we could improve an earlier proposed refit design with newer tech (see "Constellation-alpha refit" below).
I'm going to show two types of stat-based analysis:
1) simply summing across ships per stat - relevant to garrison constraints, min science constraint, max combat constraint
2) estimating "fleet combat potential" ("CP"=(C+H+L)^1.5) and "event potential" ("EP"=(S+P+D)^1.5) in a similar way to what
Nix did when estimating overall effectiveness in fleet combat and events. I want to emphasize that this is still a simplification and an approximation - there will be events where one "fleet" can still have higher chance at succeeding than another "fleet" that has higher total EP if the first fleet has higher EP per ship (it depends on the specific stat checks and event DCs). I'll also simulate the Lone Ranger cruiser +1 response bonus as D+1 in the EP formula.
Ship stats and costs, assuming we schedule builds to always take advantage of Patricia Chen's bonus:
- Constellation: CP19 EP23 C3 S2 H2 L2 P2 D3(4)* -70br -40sr 9qtr O-2 E-4 T-2 (scrap: +35br +20sr)
- Constellation-A: CP23 EP32 C4 S3 H2 L2 P2 D4(5)* -70br -45sr 9qtr O-2 E-4 T-2 (refit: -20br -10sr 4qtr)
- Renaissance: CP52 EP47 C5 S3 H4 L5 P4 D5(6)* -100br -80sr 9qtr O-3 E-5 T-3
- Centaur-A: CP23 EP27 C3 S3 H2 L3 P3 D3 -80br -70sr 6qtr O-1 E-2 T-2
* for defense stat, listing both the raw stat value and the simulated stat value with the Lone Ranger cruiser +1 response bonus
I should also note that the Constellation-A refit is kinda mediocre, and by 2313, we should have the ability to customize the refit design. Given that we aren't planning to build more Constellations, I can make the refit more expensive (including refit time) while adding more stat increments. I'm not sure if this is allowed, but let's say it is for sake of example. I'll call this speculative refit the "Constellation-alpha" and estimate its stats and costs as:
- Constellation-alpha: CP27 EP36 C4 S3 H2 L3 P3 D4(5) 80br 60sr 9qtr O-2 E-4 T-2 (refit: -40br -40sr 6qtr) - basically a Centaur-A with C+1 D+1 and takes longer to build
This might seem overpowered compared to the similarly sized Centaur-A, but consider that this is using 2313 tech versus the Centaur-A's 2304 tech, and that the Centaur-A probably isn't an optimized design (on the old ship design spreadsheet with 2304 tech values, the Centaur-A requires no fudge factors).
Since crew is our primary bottleneck, I'll try to align multiples of ships with about the same crew costs for replacement options:
- 3 Constellation: CP57 EP69 C9 S6 H6 L6 P6 D9 210br 120sr 27qtr O-6 E-12 T-6* (scrap: +105br +60sr 0qtr)
- 3 Constellation-A: CP69 EP96 C12 S9 H6 L6 P6 D12 -210br -135sr 27qtr O-6 E-12 T-6 (refit: -60br -30sr 12qtr)
- 3 Constellation-alpha: CP81 EP108 C12 S9 H6 L9 P9 D12 -240br -180sr 27qtr O-6 E-12 (refit: -120br -120sr 18qtr)
- 2 Renaissance: CP104 EP94 C10 S6 H8 L10 P8 D10 -200br -160sr 18qtr O-6 E-10 T-6
- 3 Centaur-A: CP69 EP81 C9 S9 H6 L9 P9 D9 -240br -210sr 18qtr O-3 E-6 T-6**
- 4 Centaur-A: CP92 EP108 C12 S12 H9 L12 P12 D12 -320br -280sr 24qtr O-4 E-8 T-8**
- 5 Centaur-A: CP115 EP135 C15 S15 H10 L15 P15 D15 -400br -350sr 30qtr O-5 E-10 T-10**
* note: Constellation costs here already "paid" for so this is listed for coming up with replacements with about same crew costs
** note: Showing multiple Centaur-A options, since escorts are relatively expensive for br/sr and cheap on crew compared to cruisers
Possible build options to "replace" 3 Constellations, assuming Constellations are scrapped for non-refits:
=> 3 Constellation-A: CP+12 EP+27 C+3 S+3 H+0 L+0 P+0 D+3 -60br -30sr 12qtr O-0 E-0 T-0
=> 3 Constellation-alpha: CP+24 EP+39 C+3 S+3 H+0 L+3 P+3 D+3 -120br -120sr 18qtr O-0 E-0 T-0
=> 2 Renaissance: CP+47 EP+25 C+1 S+0 H+2 L+4 P+2 D+1 -95br -100sr 18qtr O-0 E+2 T-0
=> 3 Centaur-A: CP+12 EP+12 C+0 S+3 H+0 L+3 P+3 D+0 -135br -150sr 18qtr O+3 E+6 T-0
=> 4 Centaur-A: CP+35 EP+39 C+3 S+6 H+2 L+6 P+6 D+3 -215br -220sr 24qtr O+2 E+4 T-2
=> 5 Centaur-A: CP+58 EP+66 C+6 S+9 H+4 L+9 P+9 D+6 -295br -290sr 30qtr O+1 E+2 T-4
So:
- The Constellation-A refit is cheap and fast and mostly improves its event potential and garrisoning.
- The Renaissance replacement option provides better combat potential at the trade-off of higher cost and more berth time, but doesn't improve event potential much over the Constellation-A refit.
- The Constellation-alpha refit at its speculated stats and costs is comparable to the Renaissance replacement, costing a bit more and with less combat potential, but with better event potential and garrisoning, which is pretty good for a ship that's intended for garrison duties.
- The Centaur-A replacements are harder to compare because of the very different cost characteristics. At the low end of 3 Centaur-As, it has about the same combat potential as the Constellation-A refit, is poorer at event potential and garrisoning, costs a lot more in br/sr and takes more berth time, yet saves a lot on crew. At the high end of 5 Centaur-As, it provides the most combat and event potential (as expected) but the costs and berth time are exorbitant. It's a pick your poison thing.
Out of those options, for a ship that's intended for garrison duties instead of combat, I prefer the refit options. Even the mediocre Constellation-A refit option is decent since it's really cheap and fast. If we had another modern cruiser design that was focused on events and garrisoning, I'd be more tempted to replace the Constellations with new builds of that design.
