EDIT: This suggests another hidden benefit to having more big powerful ships. They're more likely to pile up the Event successes and get crew experience.
Honestly, I would've considered it a bit imbalanced if our smaller ships could gain experience as rapidly as our larger ships, because not only are they cheaper and more numerous, crew ratings are proportionally more effective on smaller ships.
Depends entirely on what you're replacing them with, a refit of the same ship should presumably keep the crew bonus, an entire new ship won't but it might still be better to replace the ship entirely if the bonuses are large enough.
On a related note, I really really hope that the new FYM Excelsior that's being crewed with the Miracht's old crew begins with blooded status. It's a entire new ship, yes, but it's the same ship class (even if they're probably of different flights).
Honestly maybe the way to do it is just kind of determine some "cutoff" points? Part of the issue with the
Constellations is that they're old; yes that's some of why they're not as awesome stat-wise but from how
@OneirosTheWriter has presented things, it seems like genuine age ends up a factor as well.
Crews that have only been on a ship for a couple years seem like idea "swap" candidates, as they're unlikely to be Blooded, whereas those with a decade of experience could probably get by with a refit, at least for a while.
It's more about tech tier and refit costs than age directly. The older and more obsolete a ship gets, the more it would cost to refit the ship up to some modern standard, which changes the cost efficiency between such a refit versus replacing it with a ship of a newer design. I mean, we probably could technically refit the Constellation into basically a Centaur-A (which is similarly sized) and save crew that way, but that's practically like deconstructing the whole ship and rebuilding it into a new design anyway, and thus I'd expect it would cost at least as much as directly scrapping the ship and building a completely new Centaur-A.
So rather than looking at age when determining to do a refit, we need to carefully analyze the trade-offs between refitting and replacing a ship among many possible designs.
The only mechanic directly related to age that I know if is when the ship starts having reliability problems, which happened to our Soyuz ships. (As an aside, it's not clear if this "reliability" is related to the ship annual reliability rolls that affect all new designs from the Renaissance forwards.) Furthermore, it's been shown that we can extend the longevity of ships by refits, whether they are maintenance refits (happens behind the scenes, or if some failure, costs resources to replace warp core or something) or new design refits (like the Constellation-A).
It's a good question and much like the old reliability discussion you aren't going to get a consistent answer, because risk tolerance varies. Unless maybe someone digs into the data and figures out a ship's event rate and therefore failure rate and therefore disablement or loss rate, which would require knowing a lot of checks that ended up hidden as we often don't know what went wrong with a failure.
Personally I don't want to fail more than 10 to 20% of checks. And it's front loaded. If we can reduce the S check fail rate to 10% from 25%, then the effective reduction in failed hull check totals is higher than if we doubled the hull check success rate.
I think the percent of failures that translate into a serious enough penalty is an important consideration here. If a ship responded to an event, failed at it, yet didn't suffer any penalty, then it's as if it didn't respond to the event at all (assuming that failed event response doesn't result in a penalty). It's not like such a ship would've taken the place of a better ship, because I'm pretty sure every available ship in the sector rolls for event response. (Otherwise, less event-capable ships like most escorts would result in severe sector event penalties if their very presence reduced the chances of better ships in the sector from responding to events, which I don't think Oneiros wants.)
Ideally, we'd have the data that could be used to determine the net value of a ship with regards to events. But that requires more knowledge about event response DCs, event resolution DCs, event success rewards, and event failure penalties (or smaller rewards), and distributions and correlations of all of these. There could be events where lack of response results in no reward/penalty, yet successful response and failed resolution result in a penalty. There could be events where lack of response results in a penalty, and successful response and failed resolution results in no reward/penalty. My gut feeling is that the former type is more common.
We do at least have some data on non-EC failure penalties. Here's what I tallied up from the event analysis spreadsheet:
Failures resulting in military-grade Starfleet ship* loss: 3
Failures resulting in other friendly ship loss: 3
Other failures that had at least some penalty: 6
Failures that had no penalties or actually had rewards: 12
Total failures: 24
* ships that we control fleet deployments for and count toward combat/science/garrisoning stuff
Unfortunately, I think this spreadsheet isn't making a distinction between event response failure and event resolution failure, and those need to be considered separately with different consequences for fleet deployments and ship design. Still, if we make the simplifying assumption that all event response failures result in no penalty and that all the above 24 failures are event resolution failures: then about half of failures are trivial failures, and thus could be treated as no-penalty response failures. So event failure rate is practically cut in half.
I'm also assuming that the % of failures that are trivial failures doesn't depend on the ship design (and its stats), but I haven't looked into doing a ship design-level analysis and I suspect we don't have a large enough sample size for it anyway.
