Aside from the potential very large size of the early game reprisal waves, I haven't identified anything I actual have a problem with right now. Sorry if I sounded like I was complaining.
Oh, didn't sound like complaining at all.
Maybe I forgot to turn off the "ominously repressive" switch on the voice amplifier. Hmm. *click* Let's try "oddly disconcerting".I'm very grateful for feedback, especially on new mechanics, and especially in a manner focused on "what's actually happening" as in the case of your post. "What might possibly happen?" is also valuable in proportion, but tends to be more of a doubt than a question, and thus not susceptible to answers.
Changing the conversion ratio means "More Metal-More AI Ships" or "Less Metal-Fewer AI Ships". Right now, using the Bomber as a basis, you get the 2nd worst (meaning fewest AI ships) of all fleetship conversions. Only the TDL gets a worse ratio. But that also means the AI gets the 2nd fewest possible ships out of a reprisal. Since my only real concern so far is the added strength of the reprisal waves, I'm not yet in favor of upping that even for a mometary reward.
Well, bear in mind that the switch need not be a net increase in resulting pain (or resulting metal). The AI's % "take" (which, as you mentioned, is in a mathematically-similar role to the unit-to-salvage conversion function) could be adjusted downward by whatever factor seemed appropriate to achieve the desired level.
In that light, the only question (I think) that bears on whether to use convert-by-metal or convert-by-strength is which metric gives the most "accurate" results. In other words, how much _should_ the AI be throwing back in response to X of Y?
But you bring up a good point on the convert-by-strength proposal: it would make younglings count the same as normal triangle ships. That's... a bit too harsh. I'm all for the concept that Salvage brings of "casualties matter", even with things built around the concept of high-casualties (as those tend to be more effective than is particularly balanced), but we don't need to nail the pendulum to the other wall.
Even with convert-by-metal, the player can wind up getting the short end (in the short-term reprisal sense) if the AI converts it into a wave of younglings. It's getting full strength (if you threw 1,000 bombers at it on diff 10, it's going to convert that to something like 1,500 younglings, though the mark-level multipliers probably skew that substantially) but you'll get peanuts in salvage. On the other hand it could spend it all on bombers, and you'd get back a large chunk of change. That one point isn't a huge problem, but it's more widely-variable than I'd like.
Another approach, with non-trivial but non-extreme complexity cost, would be to have each AI player remember how much strength of
which types it had salvaged. Then a reprisal-wave would send back the types it could (below diff 10 it wouldn't be sending back the same quantity, but you get the idea), and only do the "melt down and spend on whatever I randomly pick" bit on ships that it can't legally send in waves (because it hadn't unlocked that type or whatever). That would make triangle casualties give much more consistent results, at least. And perhaps make high bomber casualties psychologically different for the player. Thoughts on that approach?
