Arcen Games

Games => AI War II => Topic started by: keith.lamothe on April 03, 2018, 11:13:59 AM

Title: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: keith.lamothe on April 03, 2018, 11:13:59 AM
Release notes here (https://wiki.arcengames.com/index.php?title=AI_War_2:_Final_Pre-Beta_Checks#Version_0.718_A_Wild_GUI_Appears)!

This one is the first publicly-available release with part of the new UI that Chris has been working so hard on. Specifically, the UI you see in planet-view and galaxy-view in the game itself. There's still a lot of work-in-progress but, wow, that sidebar. MANY thanks to Eric T. Edwards in lending his mighty powers of UI/UX design. We've undoubtedly slaughtered that design in various ways, but the result is still way better than we could have come up with ourselves.

The other major advance is in the in-depth feedback from players like chemical_art and Magnus, which is tremendously helpful in our actually balancing this thing and resolving tensions in the design. Notable changes on that front this round:
- The AI planets you start next to are now MUCH easier to conquer, getting back towards how AIWC handled it.
- Starships are now a bigger deal: they're now so expensive in metal and fuel that you can't really even support _one_ at the start (though you generally can after conquering your first planet, it's just going to tie up most of that fuel). But they're now also 3x as strong, making even a single Mark 1 starship a significant presence in the early game, and they're now easily the most efficient way to spend Science to increase your mobile striking power.
-- So in many ways these aren't like Fleet Ships at all anymore, except that they're both part of your mobile fleet. You can choose how much to invest in each category (which may vary based on whether you've got more Fuel or Science available), but you're going to need both.

Also notable is the beginning of the integration of real shot graphics. Still a ways to go on that, but at least now you can see those menacing plasma torpedoes that are about to kill all your missile corvettes :)

Oh, and the model for the second Arkitect backer-reward Ark, the Orchid (https://youtu.be/6Jzh7yD9J0I), is now done. Maybe the AI will be moved to sympathy by this giant space flower... but I wouldn't count on it ;)

And there's a lot of various bugfixes and other improvements from Badger, Chris, and yours truly.

Enjoy!
Keith
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 03, 2018, 11:48:17 AM
This is such a satisfying one to have out the door. :)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 03, 2018, 12:01:00 PM
I've got to say that this is a keeper of a mine. 10 minutes in and I'm feeling the goosebumps of excitement that I was getting playing AIW 1 for in terms of excitement. This update is a huge step in the right direction. Grabbing worlds quickly just gives me such a good high and yet also makes balancing things earlier. If you hadn't made them easier to take I'd be a bit concerned at not having a starship at start due to fuel constraints. But I can take that first one quickly, then I pop out a starship and then seeing it with a new fleet then piledrive through a second world...so good. Since I don't feel the need to take four an individual one gets time to shine and I'm already eager to see the build up as I get more over time. Making fleetships the fuel and metal cheap yet science expensive and making starships a foil makes for great synergy.

Today I will finally hit mid game, with pleasure.  :)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 03, 2018, 03:07:35 PM
Sweeeeet!  :D
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: eRe4s3r on April 03, 2018, 04:39:54 PM
So is this the time to start playing this for first time in-depth impression for real? Or is still much in flux regarding mechanics and GUI menus setting and stuff?

Don't wanna waste the first impression ;)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 03, 2018, 04:51:26 PM
So is this the time to start playing this for first time in-depth impression for real? Or is still much in flux regarding mechanics and GUI menus setting and stuff?

Don't wanna waste the first impression ;)

I haven't put the mid game or late game through the ringer yet, so it is possible you can't finish. Also the GUI finally has a framework that works, but it will still need to be refined I think.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 03, 2018, 05:18:57 PM
At 70 AIP, the AI somehow got 3.5k threat. If they were smart enough to actually attack, they could wipe me. I have no idea were this threat comes from, I haven't moved my king unit once and have cleared every planet I have captured. It is frustrating.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: etheric42 on April 03, 2018, 05:28:44 PM
The GUI's initial setup for the HUD (for lack of a better word) is complete.  Some of the buttons have yet to be wired up and turned on (like rally) and features like wave/combat notifications and planet information on the galaxy map have not been completed yet.  I'd hold off until Chris can finish those if you want to wait for an extra special first impression.  It's playable and fun now, but a number of things have been discussed and planned but not implemented yet.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: BadgerBadger on April 03, 2018, 05:31:34 PM
@chemical_art, re: threat. Do you have any minor factions enabled?
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 03, 2018, 05:32:15 PM
Turrets feel very under powered. When you have to spread them out so you maybe 5 of each type covering the many wormholes and AI threat moves in groups of a few hundred they really don't do much. The AI able to destroy the turret controls at will because you cannot possibly defend them, there are too many and too spread out. Since my own fleet doesn't move faster on my planets I find myself chasing threat.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 03, 2018, 05:33:14 PM
@chemical_art, re: threat. Do you have any minor factions enabled?

I play with whatever the default settings are, I haven't glanced. Many, if not most, players will do their first game like this so I feel it the mode that gets the most testing / attention to balance.

Match dropped after 45 minutes. Threat was at 7500. Insane. They are growing faster then I can kill them at AIP 70.

EDIT: I played with sentinels, wardens, and hunters. The sentinels were the ones doing the thrashing.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: keith.lamothe on April 03, 2018, 06:45:31 PM
At 70 AIP, the AI somehow got 3.5k threat. If they were smart enough to actually attack, they could wipe me.
The threat numbers are in terms of strength (ship count is meaningless when it includes fleet ships and guardians and whatnot).

Your starting fleet (flagship + triangle) is 1.6k strength. A single Mark 1 starship is 600 strength. A single cap of Mark 1 turrets is 1000 strength.

So 3.5k isn't especially concerning.

Quote
I have no idea were this threat comes from, I haven't moved my king unit once and have cleared every planet I have captured. It is frustrating.
When you show up with overwhelming force the guardians defending a planet will often abandon the planet and retreat to come after you later as threat. Is that a possibility here?

Also, I don't know if waves are sometimes being launched simply as threat, as we were doing that for a bit, but I think Badger made it not-on-by-default.


Quote
Turrets feel very under powered.
I'm fine buffing them again, but it would help to understand what exactly you're expecting. Currently turrets cost half as much science as fleet ships and have 5x the HP and DPS. That's a pretty stark difference.

Quote
When you have to spread them out so you maybe 5 of each type covering the many wormholes
How many wormholes are you covering? Is there another way to structure your defenses to reduce the number of "hard" entry points by allowing certain planets to fall while still blunting the attack?

I don't understand what you're facing that would cause you to need to do even that, however. For 1000 Science (1 low-science planet's worth) you can research 4 Mark 1 turret types. Added to your starting Mark 1 Needler turrets that's 5,000 strength. Yes, you have to spread it around to some extent, but that's a really good deal on a strength-per-science basis.

Also bear in mind that you start with Mark 1 Tractor Arrays, which have a per-planet cap.

Quote
and AI threat moves in groups of a few hundred they really don't do much
A few hundred strength or a few hundred ships? If ships, 400 Mark 1 fighters is 200 strength, though for lower-cap stuff like missile corvettes it's 120 ships => 200 strength. Nonetheless, even that cheap 5,000 strength of turrets could stop quite a few groups of 500 ships each.

Quote
The AI able to destroy the turret controls at will because you cannot possibly defend them, there are too many and too spread out.
I did halve their number in 0.718. I could do so again, though it does beg the question of why-bother. The point of their existence is to have tactically-important terrain beyond "everything beelines from wormhole to wormhole or wormhole to controller". Metal/Fuel spots would be terrain, but they would not be tactically important most of the time. Is there a better approach we can take there?

Quote
Since my own fleet doesn't move faster on my planets I find myself chasing threat.
You can slow down enemy ships with a focused gravity generator (1000 science to get the mark 1). It has a planet-wide effect that's spread among all the enemy units on the planet (so the more enemy strength, the less the effect). It may need to be buffed again, as it received several nerfs due to the annoyance of the AI having it on so many planets (it's much rarer in AI hands now).
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 03, 2018, 07:45:08 PM
At 70 AIP, the AI somehow got 3.5k threat. If they were smart enough to actually attack, they could wipe me.
The threat numbers are in terms of strength (ship count is meaningless when it includes fleet ships and guardians and whatnot).

Your starting fleet (flagship + triangle) is 1.6k strength. A single Mark 1 starship is 600 strength. A single cap of Mark 1 turrets is 1000 strength.

So 3.5k isn't especially concerning.

I guess I'm going to have to disagree here. Having that many ships floating around just from taking border planets is too many. It is scary and confusing, the AI should just be waking up. It shouldn't have a fleet as strong as yours already buzzing from a very simple action. I would understand if I was doing something gamey but if I am doing what I am supposed to I feel like I'm being punished. It is not a good feeling.

Quote
I have no idea were this threat comes from, I haven't moved my king unit once and have cleared every planet I have captured. It is frustrating.
When you show up with overwhelming force the guardians defending a planet will often abandon the planet and retreat to come after you later as threat. Is that a possibility here?

Guardians have been very tenacious in my experience, they fight to the death unless they have been "freed" due to me popping the controller before they are dead. To prevent this I always ensure I kill them and everything else first. That is why I am so confused about the sky high threat. I haven't hacked, deep striked, failed an attack, or any of the other behaviors that would cause this. I was very good at managing threat in AIW 1 and considered it one of my strategic strengths. Right now I feel powerless to prevent it which is not a good feeling.

Threat hurts a lot more now because turrets are global instead of per planet. If I could put 1k strength on each planet then 3.5k isn't a problem. But with 4 planets I'm putting more like 250 strength on each planet. I know with practice and experience I can make it work but I shouldn't be getting these defensive headaches in the first hour. These headaches should be mid game issues where I have 200ish AIP and am now am really taking the time to plan strikes for taking objectives in preparation for the end game.

Also, I don't know if waves are sometimes being launched simply as threat, as we were doing that for a bit, but I think Badger made it not-on-by-default.

I may have gotten a wave at minute 5, but now that you bring it up I didn't get a wave since. That would explain why my threat would jump 2000 at a time, and how suddenly I have threat to have 2.2k infiltrators bum rushing my planets in a confused manner. DO NOT LIKE. Sorry for the shout, but this is the first time I had to shout in displeasure at even the possibility of it happening. Keep it like AIW 1 where if a player wanted that they would choose it, but it is not default. 

Quote
Turrets feel very under powered.
I'm fine buffing them again, but it would help to understand what exactly you're expecting. Currently turrets cost half as much science as fleet ships and have 5x the HP and DPS. That's a pretty stark difference.

Quote
When you have to spread them out so you maybe 5 of each type covering the many wormholes
How many wormholes are you covering? Is there another way to structure your defenses to reduce the number of "hard" entry points by allowing certain planets to fall while still blunting the attack?

I don't understand what you're facing that would cause you to need to do even that, however. For 1000 Science (1 low-science planet's worth) you can research 4 Mark 1 turret types. Added to your starting Mark 1 Needler turrets that's 5,000 strength. Yes, you have to spread it around to some extent, but that's a really good deal on a strength-per-science basis.

Also bear in mind that you start with Mark 1 Tractor Arrays, which have a per-planet cap.

Having Tractor arrays be per planet instead of per planet changes the whole defense game entire, so I reserve all judgement until I can see what I can do to make it work.
Quote
and AI threat moves in groups of a few hundred they really don't do much
A few hundred strength or a few hundred ships? If ships, 400 Mark 1 fighters is 200 strength, though for lower-cap stuff like missile corvettes it's 120 ships => 200 strength. Nonetheless, even that cheap 5,000 strength of turrets could stop quite a few groups of 500 ships each.

Quote
The AI able to destroy the turret controls at will because you cannot possibly defend them, there are too many and too spread out.
I did halve their number in 0.718. I could do so again, though it does beg the question of why-bother. The point of their existence is to have tactically-important terrain beyond "everything beelines from wormhole to wormhole or wormhole to controller". Metal/Fuel spots would be terrain, but they would not be tactically important most of the time. Is there a better approach we can take there?

All a player needs is a measure to fend off a half dozen raiders who keep killing the buildings without dragging in a mobile fleet. If this was AIW 1 I would say something like the minifort would be perfect solution for it had 1 per planet and could move to clean up loose threats. Something like that. Or provide a tech option which gives the structures extra HP and a light gun so it can win a fight with 5 or less standard fleetships.

Quote
Since my own fleet doesn't move faster on my planets I find myself chasing threat.
You can slow down enemy ships with a focused gravity generator (1000 science to get the mark 1). It has a planet-wide effect that's spread among all the enemy units on the planet (so the more enemy strength, the less the effect). It may need to be buffed again, as it received several nerfs due to the annoyance of the AI having it on so many planets (it's much rarer in AI hands now).

I see those are per planet two so shall certainly become a part of my defense tool kit. I will work with it and the tractor beams to see how defense feels.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 03, 2018, 07:56:29 PM
Personally, I feel like having to unlock key defensive structures is a big learning curve thing.  Having to build it is one thing, but balancing your ability to have basic defenses versus just Getting On With Things on the attack is not a good choice in my opinion.  It's likely to lead to players who want to just get on with it skipping some defensive tech and then taking losses simply because they were in a hurry.  And then feeling like they are being time-penalized if they do it the "right" way, later.

I'm not really trying to lay down an edict about any particular item -- gravity stuff, tractors, whatever.  But if something is considered a basic part of the defensive pattern for even non-turtles, then we ought to be giving that for science-free at the start.  And then let them spend science on getting more of those things, if need be, or better ones.

Having just a few of these things, so you can't do a really turtle-level defense, but you do have the basic pattern there, might do the trick; then letting the turtles choose to invest further in something everyone is using, and non-turtles focus elsewhere, sounds on paper better to me.

Also, I don't know if it's just me, but personally I think it would be simpler in every respect if we made all of the power-using stuff have a per-planet cap rather than a galaxy-wide one.  There's no telling what sort of galaxy-wide empire people might want, and having to choose HOW they spend their power at each planet, versus how they spend their limited turret cap between planets, is I think better.

The reason for upgrading to higher mark levels of a turret would be for the increased killing power, and maybe actually slightly reduced power costs per turret, even.  Who knows.  But in general, I'd want to be choosing which turrets I'm doing on a given planet, or if I want to increase my metal output instead, or if I want a fortress, or whatever.

I still also think that the idea of offensive science and defensive science has some merit.  Right now the tension between offense and defense seems a bit stark, with science being so tough to come by.  Science being tough to come by in itself is not bad, but people will be hard-pressed to spend on defenses in that scenario unless you really hammer them, and if you really hammer them, that's less fun and also makes getting more science harder...

I'm not sold on that idea, fully, but it keeps coming to mind.  There are a few things in chemical_art's general list of issues there that seem to me to be "we are training you to do this wrong with game mechanics, then telling you how to do it right verbally."  The game mechanics ought to be training enough for those things, ideally.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 03, 2018, 08:27:39 PM
Personally, I feel like having to unlock key defensive structures is a big learning curve thing.  Having to build it is one thing, but balancing your ability to have basic defenses versus just Getting On With Things on the attack is not a good choice in my opinion.  It's likely to lead to players who want to just get on with it skipping some defensive tech and then taking losses simply because they were in a hurry.  And then feeling like they are being time-penalized if they do it the "right" way, later.

I'm going to try to give a fuzzy feeling impression description:

In AIW 1 I start zoomed in on the king building. This building is guarded by a shield. As I zoom out, I see this building is far away from the wormholes. At very low AIP, for example if I did not attack at all, maybe 2 dozen enemy units will attack from the wormhole. I observe that is where the enemy starts and will observe the AI goes for the king building and the supporting civilians, the supporting "home base". But thanks to the shield, I am unlikely to die if I build anything at all. I bring this up because it provides an in game explanation of how waves work and provides the bread and butter core tactic: Stick the shield over the command building and keep it far away from wormholes. I figure out the rest and having a diverse tool kit to do so with many turrets so can experiment with what works. Organically I can figure out that the AI bum rushes a planet's command center and moves on until it eventual hits the king one, and builds a defense knowing this. I have access to all but the most exotic turrets so I can mess around with them, figure out which ones I like and upgrade them. I upgrade most of them anyway, but I can start figuring out which ones I like before I have to start dropping K.

In AIW2 The king is on a wormhole.  There is a controller too, but it isn't really important. I mean it *is* important for using your turrets but if it dies its OK, your king ship needs to live. Every planet gets a controller, they are like mobile command stations, although the AI doesn't seem as focused on them and sometimes ignores them, but we will get back to that. The King can move. But should it move? Why would it move? Oh! It is used for offense because it can build things and move. The brave King leads the way! That's why it is on the enemy wormhole! If it was meant to stay away from combat it would be tucked as far as possible from all wormholes, under a shield. Oof. Game is very angry I attacked with the King. So no attack. But it does move. So I guess I just hide it in a corner? Alright, so I need to build a defense around this. But I only got one turret type. What should my next turret be? I don't know, I don't know what the AI can throw at me? I don't have means to gather such intelligence. Turrets are now global instead of planet based, so I have to spread them out, attempt a defense in depth approach. But turrets can't chase those annoying raiders...

From a game mechanic standpoint AIW 1 was more straightforward in how to defend the King works at start. AIW 2 has this weird hybrid of flirting with the idea it can do things it really shouldn't. This is a huge ramble but I guess I'm saying is that because the ark is mobile and can produce things the player gets confused on what to do with the thing. 

I'm not really trying to lay down an edict about any particular item -- gravity stuff, tractors, whatever.  But if something is considered a basic part of the defensive pattern for even non-turtles, then we ought to be giving that for science-free at the start.  And then let them spend science on getting more of those things, if need be, or better ones.

Having just a few of these things, so you can't do a really turtle-level defense, but you do have the basic pattern there, might do the trick; then letting the turtles choose to invest further in something everyone is using, and non-turtles focus elsewhere, sounds on paper better to me.


Also, I don't know if it's just me, but personally I think it would be simpler in every respect if we made all of the power-using stuff have a per-planet cap rather than a galaxy-wide one.  There's no telling what sort of galaxy-wide empire people might want, and having to choose HOW they spend their power at each planet, versus how they spend their limited turret cap between planets, is I think better.

The reason for upgrading to higher mark levels of a turret would be for the increased killing power, and maybe actually slightly reduced power costs per turret, even.  Who knows.  But in general, I'd want to be choosing which turrets I'm doing on a given planet, or if I want to increase my metal output instead, or if I want a fortress, or whatever.

I still also think that the idea of offensive science and defensive science has some merit.  Right now the tension between offense and defense seems a bit stark, with science being so tough to come by.  Science being tough to come by in itself is not bad, but people will be hard-pressed to spend on defenses in that scenario unless you really hammer them, and if you really hammer them, that's less fun and also makes getting more science harder...

Since we have prototype starships, can we have prototype turrets? They can have a global cap and act in power as MK I turrets and the player starts with all of them unlocked. If a player unlocks a "proper" MK I turret, it gets turrets that have a planet cap instead. This gives the player an anchor to experiment defenses with early game. If they find a turret they like they can place it on all their worlds. Or, if you really want to go all out on the offense / defense science tree, you can have each turret have two branches: One branch has a per planet cap, the other branch a global cap but is individually stronger.

EDIT: This would be balance nightmare I could never wish on Keith. We got to pick on or another, not both. The checkpoint static defense using global caps or the defense in depth per planet caps. Having the mobile defenses be per planet while the turrets are global is crazy wrinkle I'll have to experiment with.

I'm not sold on that idea, fully, but it keeps coming to mind.  There are a few things in chemical_art's general list of issues there that seem to me to be "we are training you to do this wrong with game mechanics, then telling you how to do it right verbally."  The game mechanics ought to be training enough for those things, ideally.

This is a bit why I made that fuzzy feeling AIW 1 vs AIW 2 paragraphs. 1 seemed to in game convey what it wanted me to do more then in 2. It is wonderful that you and Keith can tell me what to do and over time I will pass this on to other players, but the goal really is for the game to do the teaching.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: keith.lamothe on April 03, 2018, 08:48:33 PM
Personally, I feel like having to unlock key defensive structures is a big learning curve thing.
We're still finding the middle-ground there, but not even AIWC gives grav-turrets for free, despite giving so much else. Here you get the three main things you can't defend without: tachyon (via each planet controller) to decloak, tractors to hold, and needler turrets to kill. The player will inevitably need more of each of those (including a wider variety of "kill") but those are the basics. Gravity is an advanced tool, as it was in AIWC.

I think the distribution nodes are skewing things relative to AIWC since it's such a different mechanic. They increase defensive surface area by quite a lot (that's kind of the definition of why we added them), whereas in AIWC players didn't need to care much about metal spots getting killed. I see two basic approaches to this problem:

1) Remove them. Planets will look pretty empty between fights, but at least defending them will be less annoying. Then maybe we add metal and fuel spots for decoration; kind of pointless, but possibly the option that causes the least overall annoyance.

2) Double down ;) The specific idea that comes to mind is: have the distribution nodes deploy defense drones like the carrier starship does. Like the carrier ones, they automatically go about their business of hunting and killing AI units on that particular planet. So if the AI is moving in any kind of strength they'll swat the drones like flies, but otherwise the nodes should be much less of a defensive liability.


Quote
Also, I don't know if it's just me, but personally I think it would be simpler in every respect if we made all of the power-using stuff have a per-planet cap rather than a galaxy-wide one.
It would be simpler from some angles, but it would put us back in an incoherent design: if turrets are under two separate per-planet caps (their ship-cap, and power), it makes unlocking higher marks pretty weird because you hit a saturation point where expanding the cap of that turret is irrelevant. You still get the stat buffs, but it creates cognitive dissonance to have the lower-mark upgrades increase both cap and stats and to have the upper-mark upgrades increase only stats (while still going through the motions of increasing cap, and charging Science as if that mattered).

An alternative would be to have turret upgrades only increase the stats and not the cap (power cost would stay the same). The stat increases would need to be exponential to match the current Science->Strength ratios (keeping them linear and reducing Science costs would be consistent, but would place a much harsher strength-ceiling on turtles), which would be a unique thing to learn about turrets. Also, it would mean that turret unit count would stay pretty low, though I'm not very concerned about that.

The biggest problem with that is that you still get to a point where there's no need to unlock more types of turrets because your existing unlocks saturate all possible available power. The really nice thing about galaxy-wide caps is that there's almost always a point in unlocking more turrets, because there's always some planet somewhere that you'd like to put some more turrets on.

Several versions ago I just removed power-costs on turrets so if you unlocked a turret type you could put the entire cap on every single planet and only pay a metal cost. It avoided the above incoherence, but ultimately it's not a good approach. It would be worse with the mark-upgrade changes we've made.

So basically there are three approaches that come to mind:
1) Turrets are per-planet capped and cost galaxy-wide fuel to support. This is what AIWC does.
2) Turrets are galaxy-wide capped and cost per-planet power to support. This is what AIW2 currently does.
3) Turrets are per-planet capped and cost per-planet power to support, but upgrades only increase the stats and not the cap (and therefore not the power cost). This has its own weirdness, but it would be different.


Quote
I still also think that the idea of offensive science and defensive science has some merit.  Right now the tension between offense and defense seems a bit stark, with science being so tough to come by.
I think AIWC's approach of simply giving enough science is preferable to splitting it into two resources.


I need to run, but to respond briefly to chemical_art's latest: the King isn't supposed to start on a wormhole. Your ability to move it is to enable more "bend but don't break" defense: if you need to fall back, you can. In AIWC if you couldn't defend that one specific planet you were done.


For 0.719 I think we can just go ahead and give mark 1 of most (maybe all) of the "kill" turrets, and see if that alleviates the early game sense of "I can't defend".
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: etheric42 on April 03, 2018, 09:13:30 PM
1) Turrets are per-planet capped and cost galaxy-wide fuel to support. This is what AIWC does.
2) Turrets are galaxy-wide capped and cost per-planet power to support. This is what AIW2 currently does.
3) Turrets are per-planet capped and cost per-planet power to support, but upgrades only increase the stats and not the cap (and therefore not the power cost). This has its own weirdness, but it would be different.

Or 4) Turrets have no caps and just consume planet power.

You pay science in order to either upgrade stats (more expensive) or to unlock more versus armor types (cheap).

2) Double down  The specific idea that comes to mind is: have the distribution nodes deploy defense drones like the carrier starship does. Like the carrier ones, they automatically go about their business of hunting and killing AI units on that particular planet. So if the AI is moving in any kind of strength they'll swat the drones like flies, but otherwise the nodes should be much less of a defensive liability.

I think drones are great to make battles look more impressive but not add to the amount of things a player has to control (which I feel fleet ships are on the edge of being even in early game).  I would like to see this.

I need to run, but to respond briefly to chemical_art's latest: the King isn't supposed to start on a wormhole. Your ability to move it is to enable more "bend but don't break" defense: if you need to fall back, you can. In AIWC if you couldn't defend that one specific planet you were done.

I get this, but Arks are beautiful large models that you get to pick when creating your profile.  The optimal situation for your Ark is for you to never see it since it's well protected.  I almost wish the model for the Ark and the prototype flagship were flipped since you see a lot more of the flagship.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: BadgerBadger on April 03, 2018, 09:21:47 PM
Well, there are going to be backer-designed models for Flagships; that was part of the kickstarter.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: keith.lamothe on April 03, 2018, 10:00:53 PM
Or 4) Turrets have no caps and just consume planet power.

You pay science in order to either upgrade stats (more expensive) or to unlock more versus armor types (cheap).
For that to work we'd basically need to have the player see:

1) 1 Mark 1 tech for each "type" of turret. Relatively cheap.

2) 3 "upgrade turrets" techs: Mark 2, Mark 3, and Mark 4. Very expensive.
- under the hood, each of these would be a bundle of _all_ the turrets of that mark level (but you couldn't use one you didn't have the Mark 1 tech for).

Is that what you had in mind?
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: BadgerBadger on April 03, 2018, 10:13:22 PM
What about giving the player all the Mark 1 turrets, but nerfing the caps significantly. Then have upgrade to Mark 2 just give a bigger cap bump than it does now so that new Mark 2 cap and old Mark 2 cap would be the same.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: etheric42 on April 03, 2018, 10:58:04 PM
1) 1 Mark 1 tech for each "type" of turret. Relatively cheap.

2) 3 "upgrade turrets" techs: Mark 2, Mark 3, and Mark 4. Very expensive.
- under the hood, each of these would be a bundle of _all_ the turrets of that mark level (but you couldn't use one you didn't have the Mark 1 tech for).

Is that what you had in mind?

That might also be very cool.  But what I had in mind was the same tech progression as now but messing with the costs.

Mk 1 laser turrets: 100 science (because it doesn't give you any extra "cap" just allows you to target a different defense), Mk 2 laser turrets 2k science (stat bonus but doesn't affect cap), Mk 3 laser turrets 6k science... etc.

But... your refinement of the idea may be even better.

Going your route:

Needlers, free, laser turrets 100, etc.
Upgrade turret power (all turrets to Mk2) 4k science (or something).
Upgrade turret quantity (upgrades all planets energy generation) 4k science.
Upgrade turret regeneration? (increases repair rate of turrets, or allows multiple turrets to be repaired at the same time, or decreases cost of repairing turrets) Xk science.

And I put those costs as high because turrets can be used defensively or in certain circumstances offensively.  If those costs are too high, consider splitting them:

Upgrade turret power (all turrets to Mk2) 1 or 2k science.
Upgrade defensive turret quantity (upgrades all planets energy generation) 2k science.
Upgrade offensive turret quantity (upgrades flagship and Ark energy generation) 4k science.

Edit: looking back these numbers may still be too high, but you get the idea.  Let the player go "Yay science, more offense.  Yay science, more offense.  Yay scie.... oh those waves are getting nasty and I'm tired of sending ships back to stop pushes... okay let's boost defense...."

Edit edit: Since I'm not as "in to" the turret defense part of the game, maybe I'm the wrong person to say this but... when I get Mk3 bombers I go "yeah, my bombers are gonna bomb them!" and when I get Mk3 needler turrets I go "Okay, my defenses are a bit better against a certain kind of attacker."
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: zeusalmighty on April 03, 2018, 11:04:02 PM
2) Double down  The specific idea that comes to mind is: have the distribution nodes deploy defense drones like the carrier starship does. Like the carrier ones, they automatically go about their business of hunting and killing AI units on that particular planet. So if the AI is moving in any kind of strength they'll swat the drones like flies, but otherwise the nodes should be much less of a defensive liability.

I think it would be a mistake to get rid of the nodes; their relationship with turrets/defense makes defense require good decision making.

I'm inclined towards doubling down. I would actually be interested in seeing a tech just for them, in line with giving players something akin to the "Guard Posts" from classic. Maybe the AI can have stuff like this on their higher mark planets as well
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: TheVampire100 on April 04, 2018, 03:55:29 AM
I played a little around with the new UI, so far it is very intuitive to use, easy to understand and a lot better readable than what we had before.
Some minor issues are still there with it.

First, all the tabs look the same. I like the idea of having different tabs that are dedicated to building, science, units and so on. This keeps things organized and no mater hwere I am, I can just look through my tabs when I need soemthing. No menu clutter anymore. In combination with the menu on the bottom for settings, saving and so on, this is the perfect solution for playing.
All I need for playing is right at my hand, no need to scroll through multiple menus and search for something.
Now back to what I was saying, the problem is however, the cons on all these tabs look the same (obviously because it are all ships/turrets and so on with your color). This lead for an example to upgrading a ship in science while all I wanted was to build it. Maybe you can somehow differentiate the icons from each other in soem way, add a lab flask icon on the ship symbols in science for an example, a hammer symbol in the contrsuction tab, so I see "allright, these are the ships to build not the ships to select".

On a side note, it was genius to merge the different mark ships in one icon instead of having a big ship clutter like in AIWC. See, this is the stuff that shows this is an upgrade to the older game not just the same with shinier graphics. These small quality of life changes make the game easier to understand for new players and make the game  alot better to read. I can see at one glane what ships I have aviable at hand and don't have to see the same ship icon 5 times and that for each ship type. AIWC had so many different ships and towards the end of a session you had a menu full of clutter.
This small thing alone, all mark ships merged into one icon, make the Ui such a big imrpovement. Thank you!

While were at this, add again the roman numbers to the marks in the science/construction tab. I don't actually need them in the unit screen because I can see the strentgh of my ships (and decide on that if I want to engage or not) but in cunstruction and science I want to see how far a ship has been upgraded. The colored name helps only a little, maybe it's just a matter of "get used to it" but  a number is a lot easier to process than a color. The human braon sees the number and thinks "okay, this is one, 2, 3, three is higher than two so this ship is better". However, we have first to associate colors to the number and THEN can understand what ship is higher tier.
I don't say remove the colors because the help too but you need both, colors and numbers, numbers are easier to read, colors make it easier to visually differentiate between things.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: Magnus on April 04, 2018, 09:10:03 AM
I played around a bit with the new build and my experience has been different. I conquered 4 planets (2 Mark I, 2 Mark II) in just under 40 minutes, with the AI never really countering me in any significant way. At the end I had 5 planets, a fully built fleet with Mark II fighters, one special unit type, two starships, and science to spare.

I understand the start is supposed to be easy but I found it too easy and what's worse, the main problem was the AI behaving tactically like an utter moron, completely unable to get any efficiency out of its local forces during any given battle. Let me elaborate a bit further.


Stupid AI behavior while attacking

I was busy mopping up AI forces on a nearby planet, when I get the notification that there's a wave on one of my planets. I switch to it and I see a bunch of weak units incoming... and promptly splitting themselves in several different mini-groups, one for each power node plus one for the planet controller.
The overall size of the wave was so small that splitting made them utterly ineffectual. I went back to my offensive battle, and after one minute or so the AI forces were still pounding on my power nodes. I built 5 turrets of different types (one of the initial science spendings) near my controller and the end result was: the turrets were quickly completed and started destroying the AI units which were coming to me one by one, often even taking detours whenever the controller reactivated a destroyed power node.
A second wave later on was bigger and required more turrets but the end result was pretty much the same. Now admittedly I could not have gone on forever like this, due to the turret cap if nothing else, but keep in mind I was intentionally playing suboptimally (no tractor beams at all, for one) and besides, given the speed at which turrets are built and the fact that starting AI forces are so small, I could always scrap old turrets and build new ones during a future attack.

I suspect this is an artifact of extremely small scale AI forces during the very early game, but you'd probably want to code something in to allow for higher focus on single targets (presumably the controller itself) especially if the AI has such a weak force available.


Stupid AI behavior while defending

Mark I adjacent planets had such small forces that the starting deathball with no starship is more than enough to brute force them. This is the expected outcome, I guess, but already from those I noticed the AI had a tendence to keep most of its guardians in place and have them escape once the planet was deemed lost.
The real problem came with Mark II planets. Here the AI forces would have been enough to give me a decent fight, but once again, the AI kept most of the guardians in place while launching the small ships toward me. Those quickly ran out of range from even friendly sniper starships, which thus couldn't play the role of artillery support, and were destroyed due to my far superior attrition resilience. Once I had mopped up the small fry, I could then sistematically proceed to focus-fire one starship at a time.
The end result was "deathball + focus fire = dead AI". Which is still better than "deathball and nothing else", but it gives a really bad initial impression to see such a stupid AI behavior in a game called  "AI War".


To be fair, I played with default settings, and haven't tried yet setting the AI slider to maximum. I'll try it and give you further information, but if the aim is to have an extremely easy starting game for default settings, you might want to place a popup or some other notification for the first game where you explain the expected challenge level so that if someone is looking for a challenge he/she knows that you're supposed to increase that slider right from the word "go".

Other assorted musings:

- there's a bug with the new UI for fleet constructors. If you reach the point where you have 7 different types of units (the 3 basic ones + 4 special ones), every constructor needs a second row in the UI to show all of them. However the second line is not shown until you force a redraw by collapsing/extending the constructor interface.
- fuel costs are much higher and you can't have the fleet plus one starship until you conquer at least one planet. This feels right.
- my initial planet was placed in the nexus of the local cluster (i.e. the planet sitting straight in the middle of the cluster). This was good for me (quicker to get more planets, and weaker ones to boot), but it can be harder for a starting player which finds himself "surrounded". Maybe you want to place a "beginner button" for the first game which forces the generator to place your starting planet "in a corner" and maybe also to have a small (4-5 planets) starting cluster with only one outgoing wormhole, so new players can more easily establish a base of operations.
- an extremely annoying UI feature is to keep units from multiple planets selected even while you're in the galaxy map. I understand why it's needed, but it very often results in me selecting a few units from a border planet to check something, forgetting to unselect them, switching to the main fleet, selecting it, moving it to attack... and having the faraway units also start moving toward the target. It would be useful to have some kind of notification, in the galaxy map, of all the planets containing currently selected units.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: etheric42 on April 04, 2018, 11:01:40 AM
- an extremely annoying UI feature is to keep units from multiple planets selected even while you're in the galaxy map. I understand why it's needed, but it very often results in me selecting a few units from a border planet to check something, forgetting to unselect them, switching to the main fleet, selecting it, moving it to attack... and having the faraway units also start moving toward the target. It would be useful to have some kind of notification, in the galaxy map, of all the planets containing currently selected units.

Are you saying that a left-click on another planet is adding ships to your selection instead of the expected replacing the selection, or are you shift-clicking?

In the planned galaxy map UI there will be an icon for your mobile military on each planet and if you have any MM on that planet selected it will get a selection halo/glow so you should be able to see that you have multiple planets selected.

First, all the tabs look the same. I like the idea of having different tabs that are dedicated to building, science, units and so on. This keeps things organized and no mater hwere I am, I can just look through my tabs when I need soemthing. No menu clutter anymore. In combination with the menu on the bottom for settings, saving and so on, this is the perfect solution for playing.
All I need for playing is right at my hand, no need to scroll through multiple menus and search for something.
Now back to what I was saying, the problem is however, the cons on all these tabs look the same (obviously because it are all ships/turrets and so on with your color). This lead for an example to upgrading a ship in science while all I wanted was to build it. Maybe you can somehow differentiate the icons from each other in soem way, add a lab flask icon on the ship symbols in science for an example, a hammer symbol in the contrsuction tab, so I see "allright, these are the ships to build not the ships to select".

I like the icon modifier idea.  How about the border of the sidebar/background of the sidebar have art/design elements that make it look like the appropriate tab?  The science tab has a sciency background/border.  The construction tab has yellow/black caution tape around the edges.  Etc.  Or do you think that would be too busy/not be the main thing looked at and the flask/hammer would be enough?  (My other concern is with both the flask and the roman numeral there may be too many little icons on top of your icons.)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 04, 2018, 11:09:53 AM
I like the icon modifier idea.  How about the border of the sidebar/background of the sidebar have art/design elements that make it look like the appropriate tab?  The science tab has a sciency background/border.  The construction tab has yellow/black caution tape around the edges.  Etc.  Or do you think that would be too busy/not be the main thing looked at and the flask/hammer would be enough?  (My other concern is with both the flask and the roman numeral there may be too many little icons on top of your icons.)

A colored border may be best, or even a light partially transparent background  of a certain over the whole tab. By associating colors with different tabs a player can passively know which tab they are on and that will avoid confusion.

If for example I associate the color blue with research and yellow with ship selection then it will lessen the odds that I get to two mixed up.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 04, 2018, 12:11:19 PM
Lots of stuff here, and a lot of it is unrelated to my areas (GUI, performance, visuals), so please let me know if I missed anything in those areas.  I am of course armchairing a few things outside of my areas, like anyone else, aimed at Keith. :)

- What if on the first planet of the player, there were fewer distribution nodes, so it was easier to defend?  I guess that would mean that players would always try to keep their Arks there, though, so that might not be a good idea. :/

- Starting the Ark right near the controller, rather than off on its own, seems like a good idea to me.

- As far as the Ark being custom but staying home, I see the point being made, but I can't think of a good solution.  It is the thing you need to take with you to go hacking, so it's not like it never leaves home.

- For the custom flagship paint jobs, that's actually really interesting.  Since people have paid for that, rather than having it be something that you discover out in the wild blue yonder, having those all be skins for the prototype flagship strikes me as a cool thing.

- Removing turret caps makes a certain amount of sense to me, I have to say.  Then they're not under two kinds of caps at all, as mentioned by Eric.
-- But I like Badger's idea of giving all of the Mark 1 turrets from the start, but with super low caps, then making mark 2 giving a bigger cap bump.

- Instead of having drones that just automatically pop out of turrets, what if we had "drone warheads" that the player started with and could deploy at will.  These are warheads that, when deployed, instantly "detonate" and become a random bunch of fighters, bombers, and missile corvettes (drone-type).  Having players able to find more of these semi-frequently would make it so that warheads are used a tad bit more.
-- Or we could tie that into the same tab, but under the "reinforcements" thing, and this is a form of calling in mercenaries for metal.  These particular mercenaries would be a third-party, and always low-mark, so you wouldn't want to use them past the early game, but they'd help get things off to a good start.

- TheVampire100, you make good points about the tabs all looking the same, and I know just how to fix that.
-- As to the numbers on those secondary tabs, I'm not sure where we'd put them without making the ship icons impossible to view sometimes.  There's very limited space there that we're working in.  It's possible that we COULD add in the numbers, but I worry about how scary that will look and how much space it will take.  I'm inclined to see if you get used to it, unless it bothers you really a lot.  Mainly I view this as being a usability (not overwhelming) thing for new players.  What mark you're building really isn't super important at a glance, since you can mouseover to find that out, the text color is there to help, and you can't choose between marks to build, anyway.

- Magnus, thanks for the bug note on the sidebar, I'm aware of that one but need to take care of it sooner than later.

- As for the stupidity of the AI on various situations, I'm wondering if the default settings are such that it's in the actively-dumb range?  If so, we need to up the default difficulty such that that's not the case.  We need those actively-dumb lower levels, but the default AI settings should be smart, but with them simply being numerically outmatched by you for the first part of the game.  Having them numerically outnumbered AND on stupid-mode is overkill and will make people question the name of the game. ;)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: Magnus on April 04, 2018, 12:34:18 PM
Are you saying that a left-click on another planet is adding ships to your selection instead of the expected replacing the selection, or are you shift-clicking?

I left click and use rectangle selection on the first planet. Then, without deselecting (by clicking on empty space), I tab to the galaxy map, click on another planet, and select units there.
Until I stay in the second planet map, the game behaves normally. As soon as I go back to the galaxy map, the game suddenly "remembers" I never deselected the units on the first planet, and if I right click on a third planet to move my fleet, the game will move the units from both planets to the target location.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: etheric42 on April 04, 2018, 01:11:49 PM
 That isn't expected behavior (or at least I wouldn't expect it). It should require shift clicking to add to a selection.  And if you have ships on multiple planets selected it should rememver that in planet view as well.

Chris/Keith, could you confirm if this is intentional or not?
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: Magnus on April 04, 2018, 01:23:31 PM
All right, I tried to restart the game at difficulty 10. And while the AI promptly killed me, it did it due to overwhelming numbers, not due to any smarter behavior.

This is its starting attack on my second planet which I just conquered:

http://i65.tinypic.com/s4cq5h.jpg

You'll notice it's still splitting the forces to go after nodes / clusters of turrets and tractors. While theoretically better in the long run (you starve the economy), this gives the player an age to react by either building more turrets or moving the fleet to intercept. The A.I. here has a strong enough force to take the controller quickly, so that's what I'd expect it to do at maximum difficulty.

In the end it did kill me, but it was due to ridiculous numbers:

http://i66.tinypic.com/2i1goew.jpg

And then it came to finish me on the starting planet, with a 15k wave which ended up as a 25k one (with a further 25k threat):

http://i63.tinypic.com/f22b1d.jpg

Needless to say, this was extremely unsatisfactory. I fully expect to die at difficulty 10, but not because there's a moron which just happens to have infinite fleets at AIP 30 and less than 30 minutes in.
Maybe I'm misinterpreting here, and the smarter behavior is on the strategical level (which this certainly shows). If that's the case, you should explain it beforehand to properly set expectations.

Oh btw. When you lose there's always a crash (console red messages appearing) and more importantly, the whole galaxy map is revealed. You might want to adjust this, otherwise players will just lose on purpose to study the map then reload a previous savegame.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 04, 2018, 02:32:13 PM
That isn't expected behavior (or at least I wouldn't expect it). It should require shift clicking to add to a selection.  And if you have ships on multiple planets selected it should rememver that in planet view as well.

Chris/Keith, could you confirm if this is intentional or not?

100% not intentional.  On mantis and trello now: https://bugtracker.arcengames.com/view.php?id=19563
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 04, 2018, 02:35:52 PM
@Magnus:

1. For the galaxy map being revealed, I've added a ticket: https://bugtracker.arcengames.com/view.php?id=19564

2. For the crash, I think I've either fixed it or mostly fixed it, for the next version.

3. For the AI complaints, those are longer-form and require more discussion with Keith.  I'll let him respond in this thread and work with you on it.  Suffice it to say, the more detail you can provide, the better.  Though from a skim of what you wrote (it's not directly relevant to me, so I just skimmed those bits), it looks like you did.  It also looks like Diff 10 is tuned up WAY too high in terms of unit counts.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: Magnus on April 04, 2018, 05:00:32 PM
Further impressions:

I kept playing the difficulty 5 game I started yesterday, and I take back the "it's too easy" remark.
I conquered another couple planets (bringing the total to 6 and the AIP to 130), and the A.I. finally reacted and hit me like a ton of bricks. The combined fleets strengths skyrocketed to 40k and in the following hour I got wiped off the galaxy.

The tactical intelligence is still nowhere to be seen (I think the speed of unit/turret/tractor beam rebuilding is throwing off your target selection algorithm; A.I. units flail ineffectually by throwing their dps at stuff which just keeps being replaced while ignoring the priority targets i.e. the rebuilder units) but the strategic reaction is definitely there.

I think my demise was appropriate, in the sense I was playing in a careless expansionistic mode; I awoke the A.I. and it crushed me like a bug. My only complaint would be that the AIP number is completely non-talking; I have no idea where I should have stopped to avoid awakening the A.I. There's also the distinct possibility that the current A.I. reaction is way overtuned, but without knowing how high is "130" supposed to be I can't really be sure.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 04, 2018, 05:13:38 PM
but without knowing how high is "130" supposed to be I can't really be sure.

130 in a standard AIW 1 game was meant to be a point strong enough a lucky (from the AI point of view) strike could occur, but if a player were to pause any offenses and focus on defensive actions they should comfortably fend off any attacks. It wasn't until around 300 that things would start to get dicey but if you still managed to keep control it wasn't a given you would lose.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 04, 2018, 08:25:46 PM
Ideally we can keep the sense of AIP values somewhat similar between the two games, with responses roughly on par within each.  But both have different flavors, for sure.

I've had a sense that the strategic game from the AI is really there, based on all the work Keith has been putting into the hunter and warden fleets, etc.  Definitely seems like your beef is with the tactical game.  Keep us on our toes with that, Magnus -- you just became my go-to person for criticisms of the AI tactical game.  If you're not genuinely happy with what you're seeing in the tactical game, then we're falling short.  And don't start softening your views just because things improve, if you can avoid it.

Basically I use different players as yardsticks for different things, mainly based on what they complain about, and how they articulate themselves when they do.  Some people are generalists, like Badger, or TheVampire, who I listen to on a host of things but don't have as a specific yardstick about any one thing.  chemical_art and Cyborg are my general "does this feel like AIWC" pair, and the former is my "is the learning curve okay" person.  Eraser is my "is the learning curve REALLY okay" and my "are the graphics not too offensive" guy. ;)  You just became my "is the AI a tactical dead fish" guy.  Other folks have also commented on it, but you were the first to seem genuinely offended and explain why, and I mean that in the best sense.

We're heading to Early Access on the 26th of this month, is the current plan.  With that in mind, I want to have the AI in better tactical shape by then, among all the other various improvements we have cooking and potentially don't even know we need to cook.  The plan is to then spend a number of months in EA so that we can get lots more feedback and have a rock-solid 1.0 in October, because frankly we just aren't getting enough feedback at present and I'm not confident that we will even when we say "the floodgates are open for beta!"  So we're back to the original plan from the kickstarter of having an Early Access, after all.  And a two-year development period -- wowzers, twice what we had planned.

Ah well, at least it's coming together well...
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: keith.lamothe on April 04, 2018, 08:55:23 PM
Much of the AI's tactical logic is in moddable code, and I can move the rest into it without too much effort. That way it's at least visible for review, and it's straightforward to define alternate logic to be used by a test AI type (or multiple different AI types), or something like that.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 04, 2018, 09:02:15 PM
Indeed -- if we have any modding-inclined folks who want to help us out in that area, we're always happy to have help.  There's a lot that needs doing right now.  Where is Red.Queen these days, incidentally? :)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: BadgerBadger on April 05, 2018, 12:02:05 AM
@Keith, if you want people to actively help out in the modding, you might want to include more comments to explain how all the pieces work ;-)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: TheVampire100 on April 05, 2018, 03:59:14 AM
You know, I came to think of something. The only reason we can attack the AI is because its attention is divided and it cannot focus too much on its home galaxy, mostly the small planets we capture at the start.
I like to describe this with ants. When ants grab some cherries from the bottom of the tree you won't bat an eye. Maybe you wated to take the cherries for yourself any maybe you try to safe them from the ants but there are still many cherries on top of the tree, so you mostly leav the ants alone if they aren't directly on your way to pick the cherries.
However, things will get more seriously when the ants start to climb the tree. You realize that they now have  ataste for cherries and want more, so they take your cherries from the tree. This is the moment when you activeley try to do something about them because you want to protect your cherries before they reach them.
And things go really downhill when the ants suddenly are on your doorstep and invade your home because they noticed there is way more food to be found. Suddenly you use everything you can to destroy them, call in extermintors. This has become personal.

The Ai has however one major advantage to a human player: It CAN be at multiple places at once. Even if dividing its attention means a loss in performance (maybe), it can still act at two things individually whiel the player can onl look at one pla et at a time.
My point is, the AI could notice when it is atacked at one place and use this to attack itself on another planet because it knows that you are currently occupied with something else maybe foricing you to retreat to save your planet or maybe simpl not noticing until you loose said planet.
The problem with this is however obvious: It could be too frustrating for players if this happens tool often, you attack soemthing,t he AI attacks in return and you don't want to split your attention all the time. But I still think this could make some interestint twists if this happens from time to time, the AI using your strategy against you. Would also help people understand the importance of turrets on your planets.


When it comes to attacking planets, the AI should have several priorit targets, maybe deciding on what units the AI uses and what defenses the player uses. If teh Ark is on the planet, the AI should always go directly for it (even if this sounds frustrating for players). This is the commanding station for players, the AI would be utterly stupid not to wipe out the one thing that commands all the ants to steal its cherries. Other valuable targets could be strong flasgships, golems and other stuff. Thse ships are a major thrteat when they land on your own planet, so better take them out before the player can use them.
Turrets should be really be downplayed as priority for multiple reasons. They are stationary. They have most of the time a very short range. You can simply fly past them (unless tractors). So unless htey are int he way and the AI has no better goal, it should try to fly past and reach more valuable targets. No one on this planet anymore? Take next planet.
I think theplayers should discuss, what objects should be significant for the AI, so it decides on that  what to attack first.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: Magnus on April 05, 2018, 06:05:08 AM
Well, things are a bit more complex than that but yeah, the bottom line is that in most practical situations it makes sense to ignore the turrets for an attacking A.I.

The problem is that the A.I. seems to assign more or less the same priority to every stationary target in the system. This includes:

- the planetary controller
- any static structure (starship dock etc.)
- power nodes
- turrets
- tractor beams

The thing is: by doing this, the A.I. splits its forces in several small groups which become (especially during the early game) too tiny to be effective. And the controller is rebuilding turrets and tractors so fast that as soon as one of them is destroyed, the units go "hey my target is back up... let's go kill it again!". So they keep "hovering" around the clusters and the smaller the attacking fleet is, the longer this lasts.

From an overall strategic standpoint, this might even make sense: forcing the controller to keep rebuilding stuff wrecks the player's economy if it keeps going for long enough. But in practice, all this achieves is to give the player several minutes to finish whatever he's doing and then moving to intercept and wipe out the attacking fleet. At that point the economy goes back to normal extremely quickly so the end result is a failed attack and the planet still in the player's hands. In fact, I just thought of an even better way of taking advantage of this: just build a lot of clusters composed of a single tractor beam, and watch the A.I. split its forces in dozens of useless micro-groups which keep going back and forth every time the tractor is rebuilt. This way you don't even need to use turrets except to build all of them on the fly in the planet under attack, kill the fleet, then scrap them all to be available for the next wave. I'll have to try the tactic later on.

My hunch is that the controller should always be the highest priority whenever there is little to no defending fleet. If the player has a big defending fleet instead, then things become more complex as the A.I. should factor the expected controller-time-to-kill given the units it has at its disposal, their travel time to bear their damage on it, and the continuously decreasing dps due to the attrition rate caused by the player's defenses (including the standing fleet). That's a system of differential equations, but it should be easy peasy to solve for a sentient A.I.  :P
If the A.I. determines it doesn't have enough strength, it should kill whatever target they can quickly reach and then retreat. Or (far more evil) check if there is a nearby wormhole which can be used to travel to another player-controlled planet thus "bypassing" the defending fleet, or at the very least getting the jump on it by forcing the player to chase it and thus gaining more time to kill the controller on the other planet.

The immediate problem I can see with this is: assuming it can be done, the game difficulty would instantly skyrocket. My hunch is something along the lines of:


As far as defending goes, the A.I. tactics simply do not make any sense to me. What I've observed it doing is:

- if the attacking force is too big, leg it with only a few of its starships
- no matter what, throw the small ships to the attacker force
- no matter what, several of the guardians stand in place doing nothing except passively fire on whatever reaches their range

This makes zero sense. If the attacker is too strong and the planet is considered lost no matter what, the logical thing to do is to have everything which can run away and go defend something else.
Alternatively, it can make sense to sacrifice the local fleet if it can buy enough time to significantly slow down the attack thus enabling the A.I. to better pursue offense with the next wave.
In any case, whenever the decision is made to stand and fight, the A.I. ought to employ its forces optimally, which most of the time means: park the small ships on top of the controller, have the starships hover near their maximum range, and focus fire the best target (which is in itself a complex decision, which depends on available dps given the target's defenses, target value, expected time to kill given continuosly decreasing dps). If the player then tries to "game it" by using e.g. sniper units then wait till they come within their maximum range (where they will presumably stop and fire) then "charge" them.

Is there any document explaining how to mod tactical A.I. behavior? I know a few programming languages myself and could give it a shot.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 05, 2018, 09:40:15 AM
Going to respond to a few things, Keith will I'm sure respond to more.

Well, things are a bit more complex than that but yeah, the bottom line is that in most practical situations it makes sense to ignore the turrets for an attacking A.I.

That's a good point.  Basically, the part of the AI that assigns where to move, anyhow.  When it comes to where to fire, then if turrets are in range, fire away.

The problem is that the A.I. seems to assign more or less the same priority to every stationary target in the system.

That does strike me as a problem, yes.

The thing is: by doing this, the A.I. splits its forces in several small groups which become (especially during the early game) too tiny to be effective.

I think there should be a rule that basically slows down reconstruction of stuff the more of an AI presence there is.  Maybe that would be too much of a nerf to static defenses compared to mobile ones, but then again maybe mobile fleet ships need to also construct slower when the AI presence on the planet is high.  Basically I'm thinking of a "construction penalty percentage based on AI strength at this planet" which would globally apply at that planet.

Fast refleeting is GREAT, don't get me wrong, but when it's during the middle of battle it gets into this strange territory you're talking about here.

Either that, or have a cooldown on turret remains and similar where they can't do any rebuilding for 2 minutes after being destroyed, or something like that.  A long and painful enough time for players to deploy other things elsewhere.  Then again, that would just make people want to scrap the remains and build fresh turrets, so that's really a no-go.

And the controller is rebuilding turrets and tractors so fast that as soon as one of them is destroyed, the units go "hey my target is back up... let's go kill it again!". So they keep "hovering" around the clusters and the smaller the attacking fleet is, the longer this lasts.

As you say, some of the AI's stupidity is based around these mechanics, so these also need to be addressed.

My hunch is that the controller should always be the highest priority whenever there is little to no defending fleet. If the player has a big defending fleet instead, then things become more complex as the A.I. should factor the expected controller-time-to-kill given the units it has at its disposal, their travel time to bear their damage on it, and the continuously decreasing dps due to the attrition rate caused by the player's defenses (including the standing fleet). That's a system of differential equations, but it should be easy peasy to solve for a sentient A.I.  :P
If the A.I. determines it doesn't have enough strength, it should kill whatever target they can quickly reach and then retreat. Or (far more evil) check if there is a nearby wormhole which can be used to travel to another player-controlled planet thus "bypassing" the defending fleet, or at the very least getting the jump on it by forcing the player to chase it and thus gaining more time to kill the controller on the other planet.

One of the things that makes an AI dumb, typically, is when it's too predictable.  Even if it's best to bumrush the controller, it shouldn't do that more than 60% of the time.  When the AI enters a planet, it should choose a tactic to use for a while.  What is its objective?  Kill the controller?  That's the most likely.  But other times it should randomly decide to kill a random power distribution node.  Perhaps with more desirable targets weighted more heavily, but still.  It's those surprising ones where the AI suddenly zips off toward the wrong end of your planet, accomplishes a semi-pointless objective, and is now in a different position and chooses a new objective from an unexpected angle that really make for a feeling of playing against something alive.

We're going to want to have as much of that logic as possible in place on lower difficulties, because that's what makes the AI feel like it's actually a cool opponent.  In my opinion, at the tactical level the AI should do almost as much as possible that is clever from difficulty 7 and down, and then keep a few truly devious tricks for higher levels.

Reducing the number of ships tends to also make things feel different (worse) for a variety of reasons, so simply having the AI get either lower-mark ships for a much longer time, or have a multiplier against their stuff that keeps them weaker in some other way, would be ideal for lower difficulties.  Perhaps from difficulties 1 through 6, the AI's damage per shot is reduced by some multiplier. 

Perhaps on difficulty 1-3, the AI's waves and reinforcements are always 2 marks lower than they normally would be, obviously minimum 1, so most of the game it's going to be mark 1 waves and reinforcements all over the place.  Some more nuance would be needed on that, but still.  And then difficulties 4-5 it's 1 mark lower than it normally would be.

Even on lower difficulties  (5 and down), we want people to go "wow, this AI is smart," but we want that to be followed up by "but it isn't enough, mwa ha ha."  Difficulty 7 is the default experience, and difficulty 6 is watered down just a bit to be something you can lean back and watch more calmly as you play.  And then on difficulties 8+, we want the AI to be pulling out extra tricks where the players go "WHAT?"

Difficulty 10, incidentally, is meant to basically be just extra curbstomping from more ships and higher-mark ships.  The last of the new tactics should kick in at difficulty 9, since 9 is meant to be incredibly hard but also fair.  Difficulty 10 is meant to be impossible to win, literally, but only barely so.  And that via pulling out all the stops as well as having unfair numbers.  Some people play exclusively difficulty 10, and try to find the chinks in the armor despite the overwhelming odds.  Whenever they win, they gleefully tell us how they did it, and we put in changes to "fix the bug" that they won. ;)  This tends to cause improvements for the AI at all levels, because we're not fixing the bug by just upping the numerical amounts more again.

As far as defending goes, the A.I. tactics simply do not make any sense to me. What I've observed it doing is:

- if the attacking force is too big, leg it with only a few of its starships
- no matter what, throw the small ships to the attacker force
- no matter what, several of the guardians stand in place doing nothing except passively fire on whatever reaches their range

This makes zero sense. If the attacker is too strong and the planet is considered lost no matter what, the logical thing to do is to have everything which can run away and go defend something else.
Alternatively, it can make sense to sacrifice the local fleet if it can buy enough time to significantly slow down the attack thus enabling the A.I. to better pursue offense with the next wave.
In any case, whenever the decision is made to stand and fight, the A.I. ought to employ its forces optimally, which most of the time means: park the small ships on top of the controller, have the starships hover near their maximum range, and focus fire the best target (which is in itself a complex decision, which depends on available dps given the target's defenses, target value, expected time to kill given continuosly decreasing dps). If the player then tries to "game it" by using e.g. sniper units then wait till they come within their maximum range (where they will presumably stop and fire) then "charge" them.

You know, in the first game the guard posts were immobile for a reason: you'd have to go around traveling to them, and each one was a bit different.  You had to be the one maneuvering your own ships for a change, rather than letting the AI come to you.  The AI was assumed to be pinned down to these positions, and the smaller ships were explicitly NOT ALLOWED to leave their post unless the post was destroyed.  Once the post was destroyed -- and it was not auto-targeted -- then they were free to flee or bumrush or whatever else.

I kind of miss that.  The guardians tend not to feel all that threatening to me, because I see them EVERYWHERE.  I kind of wish we had a really basic guard post that was just a passive thing like in the older AI War Classic 3.0 and below, and those were what you saw the most, with the above restrictions in place.

Then once AI ships are freed from a posting, they do the sort of tactics you're describing.  But before that, it's kind of reverse tower-defense.

And I feel like the guardians should be... I don't know.  Slightly more rare, maybe more powerful, to differentiate them more from starships.  They're like these roaming mini-fortresses on the AI worlds, and maybe they're really slow but also unable to go through wormholes at all.  So that way you have this one big scary thing (or more than one!) that you have to deal with on the planet no matter what, because it's the local goliath that just can't even leave if it wanted to.  It has to commit.

The other issue is how ships of the AI are stored in the guardians and then pop out.  That just feels bad.  I bet you didn't even know they were doing that; I knew from Keith and I talking about it, but otherwise I probably would have thought that new ships were just spawning.  Basically when reinforcements happen, they go inside the guardians and don't pop out until it's time to fight.  This keeps the targeting logic during large battles on AI worlds way simpler.  And it is a good idea, in theory, except that I can't see the entire battlefield properly.

For me, I think that we should keep the general mechanic in place...ish.  But basically make it so that the ships that would be "inside" the guardian are instead visible outside of it, in proper positions, but they aren't running their logic and can't be shot.  But then once you approach them, they come active and deactivate their shields... or something.  Right now the AI worlds keep feeling kind of empty to me.

Is there any document explaining how to mod tactical A.I. behavior? I know a few programming languages myself and could give it a shot.

Keith will have to fill you in.  I really appreciate the offer, though, and any help you can provide.  Already this thread is a goldmine of ideas.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: keith.lamothe on April 05, 2018, 10:33:08 AM
Short on time, but a few quick notes:

- The "won't move until you shoot them or get too close" guardians are the stationary guard posts. Even there, they will abandon the planet if it looks totally hopeless.

- The distribution nodes may well morph into guard posts, but they'd be guard posts that just revert to neutral when killed, and you can then reclaim. So we could have them be a variety, including stationary-carriers like I mentioned before, or turret-like, or planet-wide-buffers (I'd like them to all contribute to turret reload time as they currently do, but some could have a natural gravity effect, or a natural tachyon effect, ec).

I suspect that would solve several issues at once. Though no doubt it would raise others.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 05, 2018, 10:59:26 AM
10-4, makes good sense.  The guardians not moving are something that seems like a bug since they CAN move, though.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: etheric42 on April 05, 2018, 12:39:14 PM
I think there should be a rule that basically slows down reconstruction of stuff the more of an AI presence there is.  Maybe that would be too much of a nerf to static defenses compared to mobile ones, but then again maybe mobile fleet ships need to also construct slower when the AI presence on the planet is high.  Basically I'm thinking of a "construction penalty percentage based on AI strength at this planet" which would globally apply at that planet.

For current design of fleet ships it seems to make sense to allow them to be plentifully created in response to threat.  However this creates a weird optimal route where you scrap your fleet ships at the front line in order to rebuild them on the planet being attacked if you can stall the wave long enough since rebuilding can be faster than traveling.  I'm not sure how to fix that under the current design.

As far as repairing/rebuilding fixed defenses go, I think it would make sense to not allow place units (turrets, buildings, etc) be be built/rebuilt at all as long as there is even a single enemy ship on planet.  I think this would fix some of the bouncing around issue.

Another weird thing I noticed with repairs is when taking a Mk1 planet (this happened to me twice), I had full caps of two out of my three triangle classes (one of which was Mk2) attacking a single turret that they did not have attack advantage against, and they seemed unable to make headway against the planetary controller's repairs.  The solution was simply to move in my third triangle class (who had the bonus) or just take out the controller, but it felt very strange.  On one hand this could be seen as a teachable moment: use your attack multipliers.  On the other hand it did not appear to occur until the AI was down to its last turret and it was a strange circumstance that had my third triangle class elsewhere, so it would be hard to control for to ensure it was taught.  (Does repair energy get divided by the number of targets to repair or was it just a coincidence that both times i noticed it was the last turret?)

I kind of miss that.  The guardians tend not to feel all that threatening to me, because I see them EVERYWHERE.  I kind of wish we had a really basic guard post that was just a passive thing like in the older AI War Classic 3.0 and below, and those were what you saw the most, with the above restrictions in place.

On the other hand, guard posts weren't that threatening to me in AIWC, they were just a an AI turret type and one of the objectives on each enemy planet I had to take one at a time.  Wheras guardians might let me take them one or two at a time, or I could get swarmed.  When I micro in the first half of planet engagements, it's typically to ensure I burn down the guardians (because they feel like "permanent" damage, as opposed to disposable fleet ships).

The other issue is how ships of the AI are stored in the guardians and then pop out.  That just feels bad.  I bet you didn't even know they were doing that; I knew from Keith and I talking about it, but otherwise I probably would have thought that new ships were just spawning.  Basically when reinforcements happen, they go inside the guardians and don't pop out until it's time to fight.  This keeps the targeting logic during large battles on AI worlds way simpler.  And it is a good idea, in theory, except that I can't see the entire battlefield properly.

This just needs an audio cue: "They're launching fighters!" (with a few variations).  Make it clear (or even play up) that guardians are small carriers.  Maybe have the AI spend it's resource points on building new fleet ships on guardians while there is an ongoing attack (so they get reinforced just like the humans do with their flagship).
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: BadgerBadger on April 05, 2018, 12:48:15 PM
I don't think waiting till all the enemy ships are gone to start rebuilding makes sense. I think a "Your structure can't rebuild for 1 minute after being destroyed" restriction makes more sense.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: etheric42 on April 05, 2018, 01:18:34 PM
I don't think waiting till all the enemy ships are gone to start rebuilding makes sense. I think a "Your structure can't rebuild for 1 minute after being destroyed" restriction makes more sense.

Then the player's optimal move is to just build a new turret instead of waiting for the cooldown.  That could be counteracted by having destroyed turrets still count against cap/energy, but then the optimal move is to scrap the unit prior to destruction (or scrap the remains if possible) and build a fresh structure to skip the cooldown.  To stop this behavior, you'd need to require scrapped objects still count against cap/energy for the same cooldown period as if they were destroyed.  Then you have frustrated players who just want to rearrange their base because now they have to wait a minute or more after scrapping their turrets before they can replace them even when not under attack.

Okay, so to prevent that from occurring, then you could have the 1 minute cooldown but only have it apply when there are enemy ships present (and you probably want it set to longer than a minute, since an enemy attack wave might fight for longer and you don't want them sending a detachment back to the wormhole to deal with a tractor or turret they had destroyed a minute ago... plus the tractors/turrets that are about to come back online because they were destroyed seconds after the first one).

I agree that "while enemy ships are present" is annoying because what if the AI just sits an evasive/stealth ship (although planetary controllers have tachyons) on your planet to interfere with your construction, or what if you are in the middle of placing turrets during an attack.  I just don't see a way around it without incentivising annoying play.  And once you have "while enemy ships are present" clause, then you might as well drop the cooldown timer to keep it simple.

I am probably missing something, but that's what I see.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 05, 2018, 01:20:38 PM
I'd really prefer to avoid more audio cues like "they're launching ships from those guardians," because there already feels like there is way too much chatter from the lady on the Ark.  Not Metal Gear Survive levels, but she's way too up in my business, constantly.

A couple of remaining thoughts:

1. Yes, a number of times I've just scrapped a bunch of fleet ships to have them then pop out somewhere else.  This is... odd, as you say.

2. If turrets have to be rebuilt only after a certain amount of time, then it should be a planet-wide thing in some fashion.  Maybe... after a structure is destroyed via whatever means, if there are enemies on the planet, that Power doesn't get returned for 1 minute.  Because otherwise there are all sorts of "scrap the remains and micro the rebuild over here" workarounds.

3. Having fuel not be returned for a certain amount of time after a squad dies actually strikes me as interesting, too.  Basically logistical lags on power and fuel being returned whenever something that is consuming them dies strikes me as highly desirable.  We'd need to up the fuel caps in order for that to work with the current numbers, but it would basically be another form of throughput mediation instead of JUST a cross-class effective ship cap.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: etheric42 on April 05, 2018, 02:00:13 PM
1. Yes, a number of times I've just scrapped a bunch of fleet ships to have them then pop out somewhere else.  This is... odd, as you say.

Solved if you give players access to quick travel/microwarp to give players a legitimate way to get their fleet back to defend their planets.  Also solved by various other things like lingering in the cap, like you mentioned, but I thought I'd bring that up.

3. Having fuel not be returned for a certain amount of time after a squad dies actually strikes me as interesting, too.  Basically logistical lags on power and fuel being returned whenever something that is consuming them dies strikes me as highly desirable.  We'd need to up the fuel caps in order for that to work with the current numbers, but it would basically be another form of throughput mediation instead of JUST a cross-class effective ship cap.

Fuel was previously so high it wasn't really an issue until starships started costing megabucks (which is awesome), increasing fuel caps could cause a problem there.  And I'm afraid of too many throughput bottlenecks at one time.

Bottleneck 1: You're at cap, no building (this one is pretty trivial unless there is a cooldown before cap is returned) until ships are destroyed
Bottleneck 2: You're low on metal, slow building
Bottleneck 3: You're out of fuel, normally you have plenty but ships have been destroyed but you are waiting for the fuel to be returned, no building until the fuel is refunded

I feel like bottleneck 2 and 3 are double-dipping as a punishment for in-combat rebuilding.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 05, 2018, 02:33:13 PM
Fair points.  I'm out of ideas for now, then. :)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: keith.lamothe on April 05, 2018, 02:43:17 PM
Maybe have the AI spend it's resource points on building new fleet ships on guardians while there is an ongoing attack (so they get reinforced just like the humans do with their flagship).
It kind of does do that with its reinforcement budget, but I think there's 30 seconds or more between times it spends on that, and it prioritizes rebuilding turrets and guardians (iirc). It would certainly be easy to have it funnel its reinforcement budget (and maybe part/all of its other budgets) into such a thing, but that would probably counteract one of the key elements of guerilla warfare: "hit them where they aren't". Sure, it'd just be fleet ships, but it can pump out a lot of those on that kind of budget.

That said, surely there would be some acceptable middleground. And maybe then it would make sense that not all the AI guardians fly into the battle.


That brings to mind one really crazy idea, which I think will amuse Chris: don't have the human players build fleet ships at all, have them spawn automatically from starships just like the AI's fleet ships spawn from guardians :)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 05, 2018, 02:47:02 PM
With regards to turret building it is has already been discussed why having a single ship stop everything is bad and remains of any sorts causes headaches. For that reason I'd have build speed be slowed by enemy presence on a planet. Have the floor start at 100 enemy strength and have it cap out at maybe 3000 strength causing construction to be 4x as long (0.25 build speed). Doing this opens up new tactical possibilities. For one the ark has a use! If you are a gambler of a player you can bring in your ark and double your construction speeds to help mitigate this. Fleetship factory and prototype ships can roll out to aid with fleetship production, allowing the player to still make them as rapidly even at full malus.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: etheric42 on April 05, 2018, 03:03:36 PM
With regards to turret building it is has already been discussed why having a single ship stop everything is bad and remains of any sorts causes headaches. For that reason I'd have build speed be slowed by enemy presence on a planet. Have the floor start at 100 enemy strength and have it cap out at maybe 3000 strength causing construction to be 4x as long (0.25 build speed). Doing this opens up new tactical possibilities. For one the ark has a use! If you are a gambler of a player you can bring in your ark and double your construction speeds to help mitigate this. Fleetship factory and prototype ships can roll out to aid with fleetship production, allowing the player to still make them as rapidly even at full malus.

Yeah, I agree there are some headaches there... but even if build (not repair, but build) speed was at 25%, you would still have something that went from 0 to 1 HP and caused the AI (or player!) to break off ships to deal with it if it was important enough.  Maybe building is suppressed if there is a ship within X distance and the AI leaves a sentry fighter at each destroyed target?  Not a fan of that either.  If power distribution nodes produce drones (an idea Keith had) does that solve the "AI leaves one ship on your planet to annoy you" problem?  (Or the AI could just be told not to do that because it's too annoying....)

Edit: or maybe it takes a minimum amount of strength to cut off your new/re build ability.

That brings to mind one really crazy idea, which I think will amuse Chris: don't have the human players build fleet ships at all, have them spawn automatically from starships just like the AI's fleet ships spawn from guardians

Starships or just flaghips?

Would these be full, independant fleet ships or high-powered drones?

It would certainly be easy to have it funnel its reinforcement budget (and maybe part/all of its other budgets) into such a thing, but that would probably counteract one of the key elements of guerilla warfare: "hit them where they aren't". Sure, it'd just be fleet ships, but it can pump out a lot of those on that kind of budget.

The enemy's gate is down.  They have a (reasonably) infinite swarm of fighters.  You have a (reasonably) infinite swarm of fighters.  You just need to get hits on their guardians (which they can't replace easily).  It costs you an outsize portion of your resources to build starships.  It costs the AI an outsized portion of their resources to build guardians (if they can at all).  Your job is to hit key, critical infrastructure in the face of an infinite enemy.

Edit: each guardian probably has a max throughput that only allows the overlord to spend X resources per Y deploying ships to it, so as you destroy guardians you remove the world's ability to reinforce.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 05, 2018, 03:50:00 PM
I would find that amusing, but I think that strays into territory that is dangerously unlike AI War 1 in similar ways to how SupCom 2 was unlike SupCom 1.  A lot of us were angry about SupCom 2. ;)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: Magnus on April 05, 2018, 04:01:42 PM
A few more observations.

I agree about not having the A.I. always do the same thing, even if it looks like the optimal route. Predictability just means death against any human player.

I had no clue that Guardians were essentially self-defending slow but mobile carriers. Their behavior makes a lot more sense now, but you might want to somehow make this point clearer.

Slowing down building static structures during A.I. presence makes sense to solve the "bouncing target" problem, but it also makes turrets and tractors even weaker than what they are now. Given the numbers the A.I. is currently throwing around, you might not even bother building them.
The thing is: as soon as the A.I. has enough units to manage even a few escape from tractor beams, your turrets are nearly instantly lost. What's the point of having even 50-60 of them if the A.I. routinely attacks you with hundreds (and later thousands) of squads? There are simply way too many targets for them to inflict significant damage before getting destroyed.
The "rebuilding target" problem solves this by "forcing" the A.I. to split its forces, make them a lot less effective, and allow for the controller to rebuild fast enough to let the turrets inflict significant damage over time, but it does it in a perverse way which makes the A.I. look dumb.

For me, the best solution to the "scrap and rebuild somewhere else" problem is to allow fast travel. If I want to rebuild somewhere else, it's because I'm attacking planet X and the A.I. is attacking me on planet Y. And given how fast A.I. waves are, that's basically a constant situation. It becomes fairly frustrating because, as I said, it feels like a protracted game of whack-a-mole. My thought during that is "God, how much more time do I need to waste chasing these stupid A.I. waves".
The alternative is to allow for fleet-wide ship caps (in addition to the galaxy-wide ones)  to make multi-fleet operations manageable for a human player.


And now for a few more test results.
I tried another quick game, this time at difficulty 7, and around the 40-60 minutes mark the A.I. just starts sending in ever-escalating waves that overwhelm me (even if the AIP is not growing anymore, as I'm busy trying to survive rather than conquer new planets). They're so big (comparatively) that it takes longer to clear them than it takes it to send the next wave, at which point the game is lost.
I was starting to think the A.I. is seriously overtuned right now, so I did another test at default settings. I just built the starting fleet and a defense at my home planet, and let it run for 40 minutes or so. I got a few incoming waves all at strength 260, but it looked like the game was behaving correctly. Then I conquered one planet and things started getting weird. For one, as soon as the ark moved, I got the cross-planet attack notification and threat jumped from 0 to 3500+. I'm assuming it's the activation of the hunter fleet there, but after conquering the planet and sitting there a bit (savegame 1), if you click on Troria (the nearby Mark I planet) you can see Mark IV and V ships, which should not be possible (I think..?) on a Mark I planet at AIP 30.
Then things started getting even more weird. I've attached a second save game where there's a 980 strength wave attacking (from 260 to 980 with just one planet?). The wave will charge then run to Troria without engaging. Wait for them to be gone, then move the fleet (without the Ark) on Troria, let it sit on the wormhole, and look what happens in the next 3 minutes. Something is very wrong, I think.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: keith.lamothe on April 05, 2018, 06:25:24 PM
On the build speed thing, the idea I like is:

- If hostile strength on the planet is <= your strength on the planet, you build at full speed
- Otherwise, your build speed is divided by (hostile_strength / your_strength)

So if you're just overwhelmed the rebuilding shuts down, but if you're just in normal fight-for-your-life mode it's just a penalty. And if it's just some AI stragglers you get no penalty at all.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: etheric42 on April 05, 2018, 07:47:08 PM
That seems like double-tapping.  If you're overwhelmed they probably have enough power to just wipe you off regardless of you rebuilding (although this would help in the corner case that they have a tiny fleet versus a basically undefended planet to stop the planet from just rebuild-taunting them).

Another idea is to also have gravity affect build speed.  Plus then you can just use that element of the UI without having to add anything!  ;-)
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: chemical_art on April 05, 2018, 10:24:09 PM
On the build speed thing, the idea I like is:

- If hostile strength on the planet is <= your strength on the planet, you build at full speed
- Otherwise, your build speed is divided by (hostile_strength / your_strength)

So if you're just overwhelmed the rebuilding shuts down, but if you're just in normal fight-for-your-life mode it's just a penalty. And if it's just some AI stragglers you get no penalty at all.

I like the idea, with the caveat caps to both a minimal enemy strength to trigger any reduction and a cap that can be mitigated through risky and/or thoughtful play (for example, dragging in the King to rouse a desperate defense). Setting the construction to zero makes for binary play which shuts down the total number of available tactics. It is this reason why I'm also not a fan of having construction being shut down due to an arbitrary number: If the player is being overwhelmed and desperately rebuilding turrets buys time, why not? The game usually is decided by raw number slugging of who has the larger number of units so having a form of guerrilla/militia resistance is a welcome change. If higher level AI is adjusted so the tactic is less effective, fine. But don't just write it off all together for standard play. This mechanic is wrinkle to a game I feel is too straightforward in terms of gameplay.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: Magnus on April 06, 2018, 05:05:03 AM
What about having the power nodes also affect building speed? I've noticed the controller, for whatever reason, does not prioritize them over turrets/tractors when rebuilding. This would also make the current tactical behavior more appropriate (though it would still need changes).

My problem with the speed being affected by the strength ratio is: it will shut down rebuilding when it's not going to do much anyway, since the A.I. overpowers the defense and will destroy the controller no matter what.
I'm also generally against any system which incurs a "runaway winner" problem; if one side is much stronger, giving it an even bigger advantage usually makes for a lousy game. It also doesn't give any meaningful tactical option, further pushing the "OMG I need to move my fleet back ASAP" which is already excessive as far as I'm concerned. This is because, due to the galaxy-wide turret cap, you have no real way of changing strength ratio while defending, except by moving your fleet. In fact this would make it all but impossible to do anything else, since you couldn't even scrap other turrets elsewhere and rebuild them on the fly on the attacked planet. So, if you put this in place: what am I supposed to do as a player, apart from running back?

At least with power nodes being significant targets you can experiment with having turrets defend them instead of just the wormholes / the controller.
Title: Re: AI War 2 v0.718 Released! "A Wild GUI Appears"
Post by: x4000 on April 06, 2018, 09:59:41 AM
I split my response off to here: Brainstorming about offense vs defense, and refleeting speeds. (https://forums.arcengames.com/ai-war-ii/brainstorming-about-offense-vs-defense-and-refleeting-speeds/)