Nemesis Response Scaling

[keith.lamothe]

I believe it's been suggested, but to re-iterate:

- Make the caps of nebula starships and modular forts (the three you get from the nebula, at least) be multiplied by either the number of champs in the game or the number of human homeworlds in the game, whichever is greater.

I believe that will make 1HW+MultiChamp similarly balanced to 1+1, right?

I'm not super concerned about MultiHW being crazy (though I try to keep it from getting too different-game-ish) but SingleHW+multiChamps should give similar rewards to 1+1.

On the multi-HW cases I could make nemesis responses go up by some factor (not linear, I imagine, since many champ rewards do not scale with HW count) if the humans have more than one HW.

[Bognor]

Just for convenience, here's a list I made earlier about which benefits of a single champ scale with multiple homeworlds and which do not:

These benefits scale in proportion to the number of player homeworlds:

• Modular Forts - you get one per type per player homeworld
• The nebula starship fleet - you get 7 starships per constructor per player homeworld

But these benefits are not affected by the number of homeworlds:

• Metal, crystal, and energy from the constructors you get in nebulae
• Knowledge from nebulae
• Special rewards from some nebulae, such as the Gatlings and friendly Neinzul Enclaves
• The firepower and abilities of the champion itself
• Allied starships that emerge from completed nebulae

[Cinth]

On the multi-HW cases I could make nemesis responses go up by some factor (not linear, I imagine, since many champ rewards do not scale with HW count) if the humans have more than one HW.

This isn't really needed.  The nemesis spawns on the AI HWs along with whatever nastiness the RNG gives you are plenty to keep even 8 BBs form walking all over it.  What is hitting me hardest now in 16 HW games are the carrier compression logic bits.  Having a 4K ship wave get compressed into 600 core guardians is very harsh and starts happening relatively early (mid to high end MK II progress). 

[keith.lamothe]

Having a 4K ship wave get compressed into 600 core guardians is very harsh and starts happening relatively early (mid to high end MK II progress).

Kill the little guys, then kill the carriers, rinse, repeat.  That should cut back on the compression.

Or do you mean 4K carriers?

[Cinth]

Having a 4K ship wave get compressed into 600 core guardians is very harsh and starts happening relatively early (mid to high end MK II progress).

Kill the little guys, then kill the carriers, rinse, repeat.  That should cut back on the compression.

Or do you mean 4K carriers?

It was just the carriers.  The world where this is happening doesn't have a BHG up and running yet so I pretty much have to auto-pop carriers as they come.  But even in other places (where I have fleet) I'm getting a good bit of compression.

I wonder if modular ships are contributing? My fleetball is a SS ball in this game. 8 Champ CAs, 64 Riots , 96 Corvettes ... That's a lot of modules to count up.

[Hearteater]

Should Carriers even be allow to retreat?  Seems like they never should.

[keith.lamothe]

It was just the carriers.  The world where this is happening doesn't have a BHG up and running yet so I pretty much have to auto-pop carriers as they come.  But even in other places (where I have fleet) I'm getting a good bit of compression.

I wonder if modular ships are contributing? My fleetball is a SS ball in this game. 8 Champ CAs, 64 Riots , 96 Corvettes ... That's a lot of modules to count up.

From what I can tell from the code it only cares about AI unit count on that planet for answering the question "to combine or not to combine", so it doesn't care about your unit count or the overall AI threat or whatever (though the latter plays into whether it will automatically deploy a carrier on its own initiative).

Anyway, iirc if there's fewer than 1400 AI ships on the planet the carrier pop will only go for the combination logic if another carrier deployed somewhere in the galaxy within the last 3 seconds.  The delay is to make sure that a bunch of carriers don't get queued up for "the death logic" and then all execute their non-combining dumps before the full impact of recent carrier pops factors into the counters it checks.  It's a bit paranoid in that, but the idea is that if you just open up on a big swarm of carriers you can expect pain

[keith.lamothe]

Should Carriers even be allow to retreat?  Seems like they never should.

Why not?  Generally they'll wind up in threatfleet soon, and if it just throws everything in a carrier at defenses without consideration the game gets a great deal easier.  Special Forces also uses them as a CPU expedient.

[Cinth]

It was just the carriers.  The world where this is happening doesn't have a BHG up and running yet so I pretty much have to auto-pop carriers as they come.  But even in other places (where I have fleet) I'm getting a good bit of compression.

I wonder if modular ships are contributing? My fleetball is a SS ball in this game. 8 Champ CAs, 64 Riots , 96 Corvettes ... That's a lot of modules to count up.

From what I can tell from the code it only cares about AI unit count on that planet for answering the question "to combine or not to combine", so it doesn't care about your unit count or the overall AI threat or whatever (though the latter plays into whether it will automatically deploy a carrier on its own initiative).

Anyway, iirc if there's fewer than 1400 AI ships on the planet the carrier pop will only go for the combination logic if another carrier deployed somewhere in the galaxy within the last 3 seconds.  The delay is to make sure that a bunch of carriers don't get queued up for "the death logic" and then all execute their non-combining dumps before the full impact of recent carrier pops factors into the counters it checks.  It's a bit paranoid in that, but the idea is that if you just open up on a big swarm of carriers you can expect pain

 3 seconds across 120 planets :/? That's a bit much if you ask me.  The system in question cleared 28K threat in about 15 minutes.  I'm pretty sure a full strength line could easily clear 1400 ships in under 3 seconds.
 Anyway, I've never cared for the AI "cheating" up its forces in the name of saving CPU.  It just made my botnet irrelevant before I even got it to the front line (and kills reclaimation in general).

[Radiant Phoenix]

Oh, is that why the Botnet Golem seemed useless to me?

[keith.lamothe]

Anyway, I've never cared for the AI "cheating" up its forces in the name of saving CPU.

The idea behind the combination is that the resulting forces spawned are at least roughly in the same ballpark power-wise as what would have spawned normally.

It also tries to simply upgrade the mark levels of the fleet ships popping out rather than jumping straight to core guardians.

Anyway, if the conversion rate is off it's because the relative strength evaluation is off.  Currently I believe a single Core Guardian is considered worth about 1.25 caps of mkI triangle ships.

[Cinth]

Anyway, I've never cared for the AI "cheating" up its forces in the name of saving CPU.

The idea behind the combination is that the resulting forces spawned are at least roughly in the same ballpark power-wise as what would have spawned normally.

It also tries to simply upgrade the mark levels of the fleet ships popping out rather than jumping straight to core guardians.

Anyway, if the conversion rate is off it's because the relative strength evaluation is off.  Currently I believe a single Core Guardian is considered worth about 1.25 caps of mkI triangle ships.

Yeah, I remember all that (and complaining about it when it went in

So a Core Guardian is only worth about 120 or 360 MK I triangle ships? Either way that is incredibly weak for something that should be among the strongest of the AIs forces.

[keith.lamothe]

So a Core Guardian is only worth about 120 or 360 MK I triangle ships? Either way that is incredibly weak for something that should be among the strongest of the AIs forces.

Why should it be among the individually strongest?  Conceptually they aren't particularly huge or superweapons.

Mechanically, remember that you can run into mkIV guardians very early in the game (mkIV planets 2 hops from your planet), and core guardians are only 25% more powerful than they are.  So if core guardians really were "all that" your early fleet would really be getting the hammer in that case.

[Cinth]

So a Core Guardian is only worth about 120 or 360 MK I triangle ships? Either way that is incredibly weak for something that should be among the strongest of the AIs forces.

Why should it be among the individually strongest?  Conceptually they aren't particularly huge or superweapons.

Mechanically, remember that you can run into mkIV guardians very early in the game (mkIV planets 2 hops from your planet), and core guardians are only 25% more powerful than they are.  So if core guardians really were "all that" your early fleet would really be getting the hammer in that case.

But wouldn't that make say a wave of 1000 MK III fleet ships comparably stronger than the compressed variant of Core Guardians?  And shouldn't a MK IV neighbor be avoided in the early game?  In the single HW game, you wouldn't be getting enough ships in a single wave to be seeing Core variants of anything unless AIP was pretty high (meaning you had a lot of planets and firepower anyway).

[keith.lamothe]

But wouldn't that make say a wave of 1000 MK III fleet ships comparably stronger than the compressed variant of Core Guardians?

It depends entirely on the ratio of compression   My understanding from what you said earlier, and what others have said, is that the compression-to-guardians case actually produces a greater threat than the original ships would have been.

And shouldn't a MK IV neighbor be avoided in the early game?

Generally, where possible, but we can't make it an absolute thing or you're up a creek if all the 2-hops-from-your-HW planets are mkIV.  So early-game mkIV planets have to be such that you can take them with an early fleet if you're smart about it.  But if the AI has 5-10 units that are 80% the strength of some of its strongest-individual-unit ships... I don't think that's going to go over very well.

For that matter, 5-10 units that are 60% of that, or even 40% of that, sounds remarkably unpleasant