AI War Beta 7.025-7.028 "Extermination Protocol MkI" Released!

[keith.lamothe]

When carriers deploy their contents due to the AI having insufficient ships in the air at once, how do they prioritize which ships to deploy?  If they always deploy high strength ships first, that seems like it would help with carriers appearing to shrug off damage because you're beating on a starship and don't realize it.

I'm not sure how carriers unload when there's more than one carrier, though.  Patch notes suggest only one can auto-deploy every 2s per planet, so if the same one autodeploys repeatedly, then the other carriers might not get a chance to dump their starships for a while.

It always deploys in the order which the ships were initially added to the carrier.  For waves that means starships and such first, iirc, but I think all the others are no-particular-order.

[Draco18s]

And it damages the carrier in the same order, then I presume?

I almost think "next carrier unit being damaged" should be picked randomly.  Less odds of hitting that beefy starship, but also not-having-to-chew-through-starships-first.

[keith.lamothe]

And it damages the carrier in the same order, then I presume?

I almost think "next carrier unit being damaged" should be picked randomly.  Less odds of hitting that beefy starship, but also not-having-to-chew-through-starships-first.

The units inside the carrier don't exist; it just has a list of types (otherwise memory cost would be prohibitive in many cases).  So it tracks its list of types and the damage done to the next unit in line, but cannot track the damage done to each individual unit inside.

[Draco18s]

I know they don't actually exist, I meant the "carrier has taken damage, so it removes a unit of <type> from its list" calculation.

Could be I'm misunderstanding the details of that mechanic.

[Toranth]

When carriers deploy their contents due to the AI having insufficient ships in the air at once, how do they prioritize which ships to deploy?  If they always deploy high strength ships first, that seems like it would help with carriers appearing to shrug off damage because you're beating on a starship and don't realize it.

I'm not sure how carriers unload when there's more than one carrier, though.  Patch notes suggest only one can auto-deploy every 2s per planet, so if the same one autodeploys repeatedly, then the other carriers might not get a chance to dump their starships for a while.

Or could it simply prioritize starships when selecting what ships remain on the map, rather than get loaded into carriers?  If the starships never went into carriers in the first place, it would help avoid this sort of problem.  At least, until you had 1000 or more starships in a wave...
It'd also help prioritize the targeting of Raid, Plasma Siege, and Spire starships over Leech and Zenith when they beeline for my command station.

[Winge]

A thought on this to see what y'all think:  what if it showed a calculated or estimated strength of reprisal on the tooltip of the hacker itself?

You mean the text that pops up when you mouseover the hacker?  Just checking, but why would you want it there rather than the alert box?  Too many things in the alert box already?

So that you would know how much fun to expect from a hack before starting.  Would help a lot of the confusion, as well as providing a baseline to balance against (right now, it's just "too easy" or "too hard"...there's not much to go by...).  It could pop up in the alert box (Superterminal would probably have to...), or appear in both places.

[Winge]

When carriers deploy their contents due to the AI having insufficient ships in the air at once, how do they prioritize which ships to deploy?  If they always deploy high strength ships first, that seems like it would help with carriers appearing to shrug off damage because you're beating on a starship and don't realize it.

I'm not sure how carriers unload when there's more than one carrier, though.  Patch notes suggest only one can auto-deploy every 2s per planet, so if the same one autodeploys repeatedly, then the other carriers might not get a chance to dump their starships for a while.

Or could it simply prioritize starships when selecting what ships remain on the map, rather than get loaded into carriers?  If the starships never went into carriers in the first place, it would help avoid this sort of problem.  At least, until you had 1000 or more starships in a wave...
It'd also help prioritize the targeting of Raid, Plasma Siege, and Spire starships over Leech and Zenith when they beeline for my command station.

Agree completely.  I almost never think about leaving my starships in transport...they're too useful outside of it.  Autocannons, on the other hand...

[TechSY730]

I forgot to mention this last time, but I too am of the opinion that nothing of guardian tier or up (this includes starships) should be packed into carriers, for two reasons. One, they tend to have useful extra effects that are lost upon being put into a carrier (engine damage, aoe, vamperism, etc), especially among guardians. Two, guardians and starships rarely get to a very high count, and thus, rarely should packing them into carriers really help the situation much. And if there is a situation where it would help (there are hundreds or even thousands of guardians or starships), then that is probably a sign of a mechanic going awry anyways (IMO, even at 8 HWs with the AIs having +300% and being a type that gets extra starships, and even in the most extreme of exos or hacking responses or whatever, the AI should NEVER get over 1000 starships, (outside of cheats of course. ))


EDIT: The "what unit(s) to damage next" algorithm would still be an important part of carrier balance, but is no longer crucial (like it is now) if starships and guardians are not allowed to be in carriers.

[Faulty Logic]

Post-response strength for the superterminal halved.

Half (or a little less, without wild-rolls) is still far too much. The end-response would be the only important part of the hack.

20%, I think, is the highest reasonable strength, though I'd prefer no post-hack response at all.

[keith.lamothe]

Post-response strength for the superterminal halved.

Half (or a little less, without wild-rolls) is still far too much. The end-response would be the only important part of the hack.

Two things:

1) The midway pulses may be too small right now, yes.  But a lot of math has changed in 7.025 and will change in 7.026, and while Toranth's numbers were very helpful I haven't seen much gameplay-based feedback even on 7.025 (and obviously not on 7.026).  So I wanted to see how the math actually worked out in practice before doubling the superterminal's normal pulse size or whatever.

2) If, in practice, the end-response is too high it can be toned down.  If the end-response is ok and the midway pulses, in practice, are too low they can be increased and the end-response toned down by the same proportion (to maintain what seems to be an ok level.

So I apologize if I'm missing something obvious, but I don't see the logic of how the presence of a particular level of end-response trivializes the rest of the hack unless the rest of the hack was already trivial.  And if it is, that can be addressed.  But a lot's changed in that math, so I wanted to see before further addressing was done.

[Faulty Logic]

2) If, in practice, the end-response is too high it can be toned down.  If the end-response is ok and the midway pulses, in practice, are too low they can be increased and the end-response toned down by the same proportion (to maintain what seems to be an ok level.

Excellent. I'm pretty sure that's what's going to happen.

The problem is that if a single pulse is strong enough to be interesting, then forty of them will squash me flat.

Say instead of pulses, you were figuring out how many basic fighters you could handle. You fight them one at a time, and when you're done, half that total number appear all at once. The only important part of your calculation is the half-at-once part, because compared to that, the individual spawns are trivial.

The same logic applies with superterminal pulses. It's always going to be the half-at-once superspawn that I calculate around, because it's by far the most strength at once, so any survivable end-response will mean that the individuals are trivial, or the end-response wouldn't be survivable.

[keith.lamothe]

The same logic applies with superterminal pulses. It's always going to be the half-at-once superspawn that I calculate around, because it's by far the most strength at once, so any survivable end-response will mean that the individuals are trivial, or the end-response wouldn't be survivable.

I see what you're getting at, but there's a difference in how you can approach the tactical problems:

1) The enemy just threw some ships at me, and I know it's going to do it again in another 15 seconds or so, and again after that... and I need to hold this particular planet (having no control over superterminal seeding).

2) The enemy just threw a lot of ships at me, but there's not going to be a follow-up attack (necessarily), and I no longer have to hold this particular planet.


In other words, your options for dealing with the end-response are broader.  You could engage in the irony of spending AIP warheading it (which may still work out in favor, depending on how much extra AIP you got out of the ST for it).  Or, more reasonably, you could fall back on the distributed defense of your empire, which tends to greatly diminish the force of that kind of normal-threat attack.  Neither of those approaches really fits for the during-the-hack fights.  I think the variety is a good thing, in theory.

[Faulty Logic]

In other words, your options for dealing with the end-response are broader.

Yes, which is why I think a 20% response would be reasonable, as opposed to none.

But I think 50% removes the possibility of tuning the pulses to the proper level, because there won't be anyone left to record the data 

In other words, the increased flexibility in player response can't deal with such a huge burst of strength.

[keith.lamothe]

Update: 7.026 hotfix for several severe balance issues with the new hacking post-responses, and a few other fixes and changes that snuck in since 7.025.

because there won't be anyone left to record the data 

And the AI unlocks the "It Was The Only Way To Be Sure" achievement

[Faulty Logic]

Update: 7.026 hotfix for several severe balance issues with the new hacking post-responses, and a few other fixes and changes that snuck in since 7.025.

Awesome. I'll be back with some empirical data.