Fractal Softworks Forum

Please login or register.

Login with username, password and session length
Advanced search  

News:

Starsector 0.97a is out! (02/02/24); New blog post: Simulator Enhancements (03/13/24)

Pages: [1] 2

Author Topic: What's a good fleet doctrine for AI system defense?  (Read 8705 times)

Vextor

  • Ensign
  • *
  • Posts: 39
    • View Profile
What's a good fleet doctrine for AI system defense?
« on: June 23, 2020, 01:50:10 PM »

My Paragon patrols keep getting steamrolled by huge invasion fleets, but I know for a fact if the battle would take place in real time, my fleets would emerge victorious, even if their size is 1/4th of the invading fleet (i keep getting invaded only by the LC and hegemony mostly, so only low-tech ships).

So I would assume that the battles are decided by different factors, and not by the ship's actual combat performance? Is it quantity > quality? Big > small? Warships > all? Carriers > all? Officers? Weapons and fighters?
How does the game decide which fleet emerges victorious? Does it compare which fleet has more total ordnance points and flags it for victory?
Logged

Terethall

  • Commander
  • ***
  • Posts: 157
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #1 on: June 23, 2020, 01:55:34 PM »

My understanding is that d-mods and quantity of ships have an outsized impact compared to size of ships -- and compared to real battles, d-mods aren't so bad at all and size of ships is king, so that's kind of counter-intuitive. And I don't think the auto-resolve logic considers at all how the composition would fare in a real battle, e.g. warships vs. carriers vs. phase ships. I'm not sure if aggression impacts it, or how much importance the game affords to officers.
Logged

Morrokain

  • Admiral
  • *****
  • Posts: 2143
  • Megalith Dreadnought - Archean Order
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #2 on: June 23, 2020, 05:20:27 PM »

My understanding is that d-mods and quantity of ships have an outsized impact compared to size of ships -- and compared to real battles, d-mods aren't so bad at all and size of ships is king, so that's kind of counter-intuitive. And I don't think the auto-resolve logic considers at all how the composition would fare in a real battle, e.g. warships vs. carriers vs. phase ships. I'm not sure if aggression impacts it, or how much importance the game affords to officers.

Not sure on the D-mods part, but I thought Alex said it's purely based on FP and officers/loadouts don't matter at all. Am I wrong on that? (It's the last thing FP is used for iirc.) That's vanilla, though, Nex could change that calculation idk.
Logged

Grievous69

  • Admiral
  • *****
  • Posts: 2980
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #3 on: June 24, 2020, 12:44:51 AM »

No way it works like that, I would be genuinely surprised. From my understanding, yes the loadouts don't mean absolutely anything but the officer level alongside d-mods do. Along with the FPs of course, no idea what the math actually is but since there's a slider on officer quality and ship quality, it's gotta be substantial right?
Logged
Please don't take me too seriously.

Morrokain

  • Admiral
  • *****
  • Posts: 2143
  • Megalith Dreadnought - Archean Order
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #4 on: June 24, 2020, 02:40:14 AM »

No way it works like that, I would be genuinely surprised. From my understanding, yes the loadouts don't mean absolutely anything but the officer level alongside d-mods do. Along with the FPs of course, no idea what the math actually is but since there's a slider on officer quality and ship quality, it's gotta be substantial right?

D-mods I could definitely see affecting things since they effect deployment cost- which is what FP kind of models.

I don't think there is ship quality at least from what I've seen - though there is variant quality - which I thought was just for autofit? I truly don't know though- variant quality could play a role in autoresolve too, I just thought I read in the past that FP was the main thing there and nothing else. But that could have been a generalization or could even have changed since then.

Officer quality? Where is that defined if you don't mind me asking? If it's in the faction file, I thought that was for patrol generation in regards to the officer level of patrols in order to challenge the player in live combat rather than used for autoresolve in any way. Again, however, it could be used in both. I mean it would certainly make sense for sure, but I thought I read otherwise.

I'd dig up the post if I could... but it's probably 100 pages deep in the misc modding questions thread at this point. XD
Logged

Grievous69

  • Admiral
  • *****
  • Posts: 2980
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #5 on: June 24, 2020, 03:14:53 AM »

You can set the officer quality in your faction doctrine. If it did absolutely nothing then why would it exist in the first place?

@Alex
You really gotta improve the wording somehow regarding DPs, cost of deployments and FPs (maybe there's something else similar).

I see people get this wrong soooo often. D-mods never reduce the deployment cost of ships, even with skills! They reduce the cost of deployment, which is actually recovery cost, which doesn't even matter for auto-resolve in the slightest. D-mods just make your ships CHEAPER to deploy, not WEAKER in terms of DP. Otherwise that would be kida broken and you'd just have a zerg swarm of d-mod ships.

EDIT: Oh yeah and there's also maintenance cost, which can be modified, but it doesn't change DP, which some people look at there to see what's the DP of each ship (I know base recovery cost is equal to DP). But everything should look a bit more, easier to digest let's say. Deployment cost is a very important stat and you can only see it pre-combat. There's nothing saying ''dp of this ship is x'' in campaign screen.
« Last Edit: June 24, 2020, 03:28:49 AM by Grievous69 »
Logged
Please don't take me too seriously.

Mondaymonkey

  • Admiral
  • *****
  • Posts: 777
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #6 on: June 24, 2020, 07:16:57 AM »

Quote
My Paragon patrols keep getting steamrolled by huge invasion fleets,

Is that nex? Because my vanilla Paragon patrols throws much larger fleets down into the hell.
Logged
I dislike human beings... or I just do not know how to cook them well.

Vextor

  • Ensign
  • *
  • Posts: 39
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #7 on: June 24, 2020, 09:19:56 AM »

Is that nex? Because my vanilla Paragon patrols throws much larger fleets down into the hell.

Yes, I do have Nexerelin installed. Haven't played for a year so I forgot what vanilla fleets were like, my bad then if the mod caused this. Still, I wish there was a more detailed list of changes on the mod page.

To be honest it would be fine that my basic system patrols couldn't defeat a big invasion, but the fact that I have to interact with any inhabited planet just to be able to access the Special Functions menu in order to request a defense fleet is a massive PITA. As soon as I step foot in the outer edges of the sector, boom, invasion, time to rush back. Then once the request is sent, by the time I'm about to explore again, another faction sends an invasion fleet to a different system, or in the meantime my ally made a peace treaty with the hostile faction and the invasion is cancelled and me going back has been for absolutely nothing. It's... frustrating.
Logged

Morrokain

  • Admiral
  • *****
  • Posts: 2143
  • Megalith Dreadnought - Archean Order
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #8 on: June 24, 2020, 12:31:53 PM »

You can set the officer quality in your faction doctrine. If it did absolutely nothing then why would it exist in the first place?

Just to clarify, I didn't say it did nothing I just said it didn't affect autoresolve. I thought it affects patrol spawn quality by how high of a level the officers in the patrol fleet spawn with.

What you are saying is that it likely takes that a step further by calculating officer level of combined officers in each fleet for the autoresolve calculation. It's that part I'm not so sure about.

@Alex
You really gotta improve the wording somehow regarding DPs, cost of deployments and FPs (maybe there's something else similar).

I see people get this wrong soooo often. D-mods never reduce the deployment cost of ships, even with skills! They reduce the cost of deployment, which is actually recovery cost, which doesn't even matter for auto-resolve in the slightest. D-mods just make your ships CHEAPER to deploy, not WEAKER in terms of DP. Otherwise that would be kida broken and you'd just have a zerg swarm of d-mod ships.

EDIT: Oh yeah and there's also maintenance cost, which can be modified, but it doesn't change DP, which some people look at there to see what's the DP of each ship (I know base recovery cost is equal to DP). But everything should look a bit more, easier to digest let's say. Deployment cost is a very important stat and you can only see it pre-combat. There's nothing saying ''dp of this ship is x'' in campaign screen.

*facepalm* Yeah I was under the impression it reduced/improved the actual deployment cost in the precombat screen that's... not ideal, sigh. I agree it's also confusing because technically it's the same stat in the csv so making the assumption that it would operate this way when modified is a natural conclusion. I can't believe I never noticed it before.

Guess part of my work today will be removing most of my faction skin files in favor of separate ship definitions since stuff like XIV ships are supposed to take up more deployment space (avoiding cost now) and prevent the player from sending as many of them into battle.

As far as D-mods go, that makes them a bit more of a crutch than I thought they were. Though that's not to say that's terribly awful considering they can be removed, it does make me think that restoration of D-mods shouldn't be so expensive. In fact, similar to repairing a car instead of buying a new one I'd argue it should equate to about 60% of the ship cost (so 15% per mod) and actually be a bargain. Otherwise, I can tell you that unless I was stupidly rich in the sector I would save scum each time I lost a ship just like before recovery was a thing - which somewhat reduces the appeal of the recovery system to me.

Idk, I get that there is more nuance than that as far as sustainability with the salvager playstyle. But if I'm going elite merc, for instance, I wouldn't even bother with recovery now. If I lost a ship, I would reload or buy another one and eat the cost that way. Of course the exception would be rare ships I can't find easily.
« Last Edit: June 24, 2020, 02:58:17 PM by Morrokain »
Logged

Morrokain

  • Admiral
  • *****
  • Posts: 2143
  • Megalith Dreadnought - Archean Order
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #9 on: June 27, 2020, 10:55:22 AM »

No way it works like that, I would be genuinely surprised. From my understanding, yes the loadouts don't mean absolutely anything but the officer level alongside d-mods do. Along with the FPs of course, no idea what the math actually is but since there's a slider on officer quality and ship quality, it's gotta be substantial right?

I don't think there is ship quality at least from what I've seen - though there is variant quality - which I thought was just for autofit? I truly don't know though- variant quality could play a role in autoresolve too, I just thought I read in the past that FP was the main thing there and nothing else. But that could have been a generalization or could even have changed since then.

Just an update for informational purposes - I asked Alex about variant quality in autoresolve:

This came up in another thread and I really don't know how this works, but, is the quality modifier of a variant used in the autoresolve calculation in any way?

It doesn't, and in fact it's not used for anything anymore.

So I think it is just FP.
Logged

Grievous69

  • Admiral
  • *****
  • Posts: 2980
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #10 on: June 27, 2020, 11:10:08 AM »

This came up in another thread and I really don't know how this works, but, is the quality modifier of a variant used in the autoresolve calculation in any way?

It doesn't, and in fact it's not used for anything anymore.

So I think it is just FP.
So the fleet quality stat is completely useless? And the slider for ship quality? Because you said variant quality, that can be understood both as a ship itself and specific loadout for that ship (variant). So I'm still a bit confused. If it's actually true d-mods don't change a thing then what's the point in installing nanoforges at your colonies?
Logged
Please don't take me too seriously.

Mondaymonkey

  • Admiral
  • *****
  • Posts: 777
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #11 on: June 27, 2020, 11:19:12 AM »

"variant quality" - is a value used in .variant files, that determine loadouts. As Alex said - no impact to autoresolve.

"Ship quality" - value that determine amount of D-mods your faction ships would have. Each D-mode is a penalty to a ship "fleet points". So eah, it should theoretically have impact on autoresolve. Theoretically.
Logged
I dislike human beings... or I just do not know how to cook them well.

Grievous69

  • Admiral
  • *****
  • Posts: 2980
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #12 on: June 27, 2020, 11:33:01 AM »

That's precisely what I assumed from the start. But life has taught me that it's dangerous to assume anything.
« Last Edit: June 27, 2020, 11:54:18 AM by Grievous69 »
Logged
Please don't take me too seriously.

Alex

  • Administrator
  • Admiral
  • *****
  • Posts: 23988
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #13 on: June 27, 2020, 11:52:53 AM »

"variant quality" - is a value used in .variant files, that determine loadouts. As Alex said - no impact to autoresolve.

"Ship quality" - value that determine amount of D-mods your faction ships would have. Each D-mode is a penalty to a ship "fleet points". So eah, it should theoretically have impact on autoresolve. Theoretically.

This is mostly correct, yeah. Well, aside from variant quality impacting loadouts - it doesn't, it's not used for anything. It's a holdover from before autofit and ship quality. It's in the .variant files but it's not read by the game.

Generally, the officers/ship quality/number of ships stats are tuned to be roughly equivalent as far as autoresolve is concerned - the idea is that you use these to configure the sort of allied fleets you want to *fight alongside in actual battles*, but that for autoresolve, it's a wash, so you don't feel forced to go for the "best" settings for that.
Logged

Morrokain

  • Admiral
  • *****
  • Posts: 2143
  • Megalith Dreadnought - Archean Order
    • View Profile
Re: What's a good fleet doctrine for AI system defense?
« Reply #14 on: June 27, 2020, 12:03:02 PM »

"variant quality" - is a value used in .variant files, that determine loadouts. As Alex said - no impact to autoresolve.

"Ship quality" - value that determine amount of D-mods your faction ships would have. Each D-mode is a penalty to a ship "fleet points". So eah, it should theoretically have impact on autoresolve. Theoretically.

This is mostly correct, yeah. Well, aside from variant quality impacting loadouts - it doesn't, it's not used for anything. It's a holdover from before autofit and ship quality. It's in the .variant files but it's not read by the game.

Generally, the officers/ship quality/number of ships stats are tuned to be roughly equivalent as far as autoresolve is concerned - the idea is that you use these to configure the sort of allied fleets you want to *fight alongside in actual battles*, but that for autoresolve, it's a wash, so you don't feel forced to go for the "best" settings for that.

Ah ok thanks for the clarification! Actually, I wonder if I have messed these up somehow. I'll have to do some comparisons to vanilla implementation for these values.
Logged
Pages: [1] 2