Page 5 of 7

Re: My Feedback

Posted: Wed Jun 08, 2016 5:54 pm
by zero
If shipcount is a significant contributor to lag, perhaps an option to reduce shipcount is for scrapping ships to refund the industry it cost to build them (provided it is done while in supply). Essentially one could salvage an older fleet to accelerate the production of modern hulls.
There are drawbacks, like being able to pump up industry at an isolated colony (if that's necessarily a drawback), needing the AI to be aware of it and perhaps having to track the original build cost of the ship. On the plus side the computer would be able to scrap its big stacks of laser robotic hulls in exchange for plasma grav or titan hulls, cutting shipcount and partcount without losing combat effectiveness. Same result for troops; I've sometimes found AIs with massive stacks of troop ships all the way back to medium hulls while their modern version is heavy asteroid or grav.

Re: My Feedback

Posted: Wed Jun 08, 2016 6:20 pm
by Geoff the Medio
zero wrote:There are drawbacks, like being able to pump up industry at an isolated colony...
That, and the resulting potential to abuse it via micromanagement-intensive shuttling, is a major reason this isn't done.

Re: My Feedback

Posted: Wed Jun 08, 2016 10:06 pm
by IamZeke
Geoff the Medio wrote:
zero wrote:There are drawbacks, like being able to pump up industry at an isolated colony...
That, and the resulting potential to abuse it via micromanagement-intensive shuttling, is a major reason this isn't done.
Agreed. I've seen this kind of thing abused before in games. Scrapping ships should only count towards a one shot bonus to build a ship in the dock where the scrapping is done, that is less than the amount used to create the ship in the first place. As a general one time bonus to to build a new ship that is a net loss of total PP used to build the scrapped ship in the first place then people won't scrap perfectly good ships that still have value just to boost another colony's economy.

Upgrading ships would be a great idea too. You can't change a hull type but upgrading components would keep ships viable without requiring too many extra ships built to increase a ship stack strength.

But the idea of a colony support ship has merit. Instead of suffering a backdoor micromanagement gimmick it wouldn't be farfetched in a space colony game to have special designated ship that deliver a support bonus to an isolated system. Instead of a colony or outpost pod you have a PP pod instead. Historical exploration used dedicated supply to get an outpost up to self sufficiency. Otherwise you end up having problems like the Jamestown colony had. By building something dedicated you stop players from trying to find a way to backdoor the idea. The more popular a game gets the more players hunt for exploits. It's just a fact of life for games. Better to channel an existing drive in the players than try to block them.

Re: My Feedback

Posted: Thu Jun 09, 2016 3:18 am
by Kassiopeija
What puzzles me is that fleeting ships together doesn't reduce lag. In other game it does so significantly. Given that there's no upper limit on the number of ships you can fleet in FO together, if a fleet would just be handled by the game like a single ship it would solve the issue at hand.

Salvage may become interesting or balanceably when ships aren't made out of an abstract production but need various resource-materials (which may be found & harvested in space or planets). These could be refunded but that would require more production (=working done).

How about Robotic Ships (with Robotic Crews or not) being able to combine their ships? Such a comination may not be able to fly but the structure, weapons & shields would stay instact. Basically you form a non-moving spacestation out of your early ships which may act as a defensive stronghold over critical systems.

Nevertheless, if the way how the combat targetting system works isn't changed - I see no reason at all to do away with multitudes of weaker ships - seriously, even lategame these are able to swallow so much enemy weaponsfire (lots of excess overkill here). Yeah, you see them dying regularily and may think they're not much worth but in truth they saved your precious main ships from destruction.
Build a single Scattered Asteroid barring only weapons & surround it by 10 Swarms barring only hitpoints and you'll see how immensely powerfull "decoys" are.

Re: My Feedback

Posted: Thu Jun 09, 2016 10:23 am
by IamZeke
I just finished turn 350 in a game tonight. The game was basically won by then due to my overwhelming production and research advantage, but the AI still had a big stack of about 250 ships parked in the system next to a critical choke point where a breakthrough would have made me vulnerable for a while. Lag was running well over a minute at that point. So I decided to send about 30 powerful solar hulls in that had about the same total combined weapon and armor points as the AI's 200+ stack.

My solar hulls slammed them and wiped the big stack out while only taking about 5 losses. The game was basically won with that move, but I decided to test the lag a couple more turns. It dropped to about 5 seconds.

Say whatever you want about the root causes but I've seen this before. Too many hulls will lag the system.

So any potential lag fix that works on the number of total hulls in the game is a potentially valid option.

I do see one big cause of hulls. The AI decoy ships they stack up for protection. While I understand the purpose of them, to soak up attacks, they are clearly part of the problem. That the 200+ stack was at least half decoys underscores that point. The AI needs that tactic taken from the code. I realize the defensive usefulness of the AI decoys but they also slow down the game too much.

Decoys could be replaced possibly with an old military concept called chaff. An active defense that makes the enemy waste shots on harmless decoy material. A single ship component that cuts the number of incoming attacks. I don't have a working mechanism in mind on how to code the ship part. Nor do I see it as easy as it would be a terrific component to abuse as a component. Limit it to not being stackable or perhaps make it a circle internal component. I could definitely see it being hard to implement and still keep game balance, but the concept still is worth looking at to replace decoy ships serving the same purpose.

Re: My Feedback

Posted: Thu Jun 09, 2016 2:28 pm
by MatGB
The main use of decoys is currently a UI fix, there's no way to force planetary defences to fire at a passive enemy unit (a floater or scout) unless oyu have an aggressive unit in the system. That's not idea and needs fixing.

Chaff is basically coming with the Fighters branch (which isn't going to be in the forthcoming Release but is playable now, the AI will likely die horribly and the tech/refinements haven't been balanced, but it's usable), where there'll be a LOT of extra small ships flying around soaking up shots.

I'm hopign that a) we can remove the need for decoys completely (I hates them I does) and b) we can give planets some form of fighter defence systems, perhaps to replace some of their existing defences.

But yeah, the AI does sometimes go overboard with decoys, and also never scraps troop drops if it's got too many, etc. Fortunately, they don't have a massive number of Effectsgroups each, so each doesn't contribute massively to the lag (it seriously is mostly caused by effectsgroups, and warships have a LOT of them from various techs, etc.

Re: My Feedback

Posted: Thu Jun 09, 2016 3:36 pm
by Geoff the Medio
Kassiopeija wrote:What puzzles me is that fleeting ships together doesn't reduce lag.
For most cases of effects being applied to ships, what fleet they are in doesn't have any impact on the processing time to determine what ship(s) (or other object(s)) each effect acts on or what the results of the effect is on the ship(s).
...if a fleet would just be handled by the game like a single ship it would solve the issue at hand.
That's not possible, given how effects work. Each ships has a variety of independent variables attached to it, and the set of effects that act on each ship and the resulting modifications to each ships state are not and can't be all the same due to them being in the same fleet. Fleets do have some state that is shared / used for all ships in them, but this is not the cause of effects processing taking a long time.

Re: My Feedback

Posted: Fri Jun 10, 2016 3:47 am
by IamZeke
MatGB wrote:The main use of decoys is currently a UI fix, there's no way to force planetary defences to fire at a passive enemy unit (a floater or scout) unless oyu have an aggressive unit in the system. That's not idea and needs fixing.

Chaff is basically coming with the Fighters branch (which isn't going to be in the forthcoming Release but is playable now, the AI will likely die horribly and the tech/refinements haven't been balanced, but it's usable), where there'll be a LOT of extra small ships flying around soaking up shots.

I'm hopign that a) we can remove the need for decoys completely (I hates them I does) and b) we can give planets some form of fighter defence systems, perhaps to replace some of their existing defences.

But yeah, the AI does sometimes go overboard with decoys, and also never scraps troop drops if it's got too many, etc. Fortunately, they don't have a massive number of Effectsgroups each, so each doesn't contribute massively to the lag (it seriously is mostly caused by effectsgroups, and warships have a LOT of them from various techs, etc.
I've even seen the AI make too many outpost base vessels in a system. I'm not even sure how that could happen. They aren't labelled as decoys and the number of planets never gets smaller. So I'm not sure how it makes a couple too many. Surely the computer can count planets as well as I do.

Of fighters being added I have two very opposing thoughts. First as a player not knowing about the excess ship problem because fighters from space ships sounds very cool. And then you have the great mass of ships causing lag now making me fret that lag is about to really jump. I guess I'll just have to see how it plays out.

As for needing systems to attack, there is always the standby of fixed gunboats (monitors, killsats,and space bases for alternate names). Instead of a lot of something, you just go bigger. Something that just floats can be much larger than something that deals with traveling inertia. And should upgrading ships ever come to pass then fixed gunboats is the ideal first test choice.

Re: My Feedback

Posted: Fri Jun 10, 2016 4:14 am
by Geoff the Medio
IamZeke wrote:Of fighters being added I have two very opposing thoughts. First as a player not knowing about the excess ship problem because fighters from space ships sounds very cool. And then you have the great mass of ships causing lag now making me fret that lag is about to really jump. I guess I'll just have to see how it plays out.
Fighters as implemented are not ships, and their number does not impact effect processing time.

Re: My Feedback

Posted: Fri Jun 10, 2016 8:23 am
by IamZeke
Geoff the Medio wrote:
IamZeke wrote:Of fighters being added I have two very opposing thoughts. First as a player not knowing about the excess ship problem because fighters from space ships sounds very cool. And then you have the great mass of ships causing lag now making me fret that lag is about to really jump. I guess I'll just have to see how it plays out.
Fighters as implemented are not ships, and their number does not impact effect processing time.
That's good to hear. Realizing it still needs testing and possibly more development before full release, the idea of fighters in a space war game sounds great.

Re: My Feedback

Posted: Fri Jun 10, 2016 10:02 am
by Kassiopeija
IamZeke wrote: I've even seen the AI make too many outpost base vessels in a system. I'm not even sure how that could happen. They aren't labelled as decoys and the number of planets never gets smaller. So I'm not sure how it makes a couple too many. Surely the computer can count planets as well as I do.
See this too in every game. My guess is that AI colonizes system --> enqueues cheap base --> in the meanwhile colony ship or outpost ship arrives to claim that planet.
Donno if the system blanks out planets which already see a form of colonization. Maybe requires some additional hickup to go bad, like colony ship had to be reroute because of previous target gone missing due to enemy etc
Nevertheless, all things considered I'd rather wait a turn or more for my cheaper base to claim that planet and have an extra mobile colonizer waiting in space.

Re: My Feedback

Posted: Fri Jun 10, 2016 4:55 pm
by wobbly
I'd be curious to take a little look at how the AI is making it's decisions even just for interest sake. How readable is that part of the game to a non coder & where should I be looking?

Re: My Feedback

Posted: Fri Jun 10, 2016 5:17 pm
by MatGB
Look in the AI logs, stored alongside the normal game logs wherever your system puts them, they're not complete but they do record a lot of the decision making in a way that's very human readable, deliberately, the team use them for debugging and trying to work out why it's doing that silly thing that makes little sense to an observer.

In the last few years the number of stupid things has been massively reduced, which is good. Of course, some of the stupid things it's doing now it wasn't doing two years ago but, well, c'est la vie.

Re: My Feedback

Posted: Fri Jun 10, 2016 5:59 pm
by defaultuser
Kassiopeija wrote:
IamZeke wrote: I've even seen the AI make too many outpost base vessels in a system.
See this too in every game. My guess is that AI colonizes system --> enqueues cheap base --> in the meanwhile colony ship or outpost ship arrives to claim that planet.
I've seen cases (not at all uncommon) where the AI has a colony start that can't finish because the planet is populated. That has to be something similar, like "built an outpost, started a colony build, then a colony ship showed up and colonized the planet."

Re: My Feedback

Posted: Fri Jun 10, 2016 7:27 pm
by MatGB
defaultuser wrote: I've seen cases (not at all uncommon) where the AI has a colony start that can't finish because the planet is populated. That has to be something similar, like "built an outpost, started a colony build, then a colony ship showed up and colonized the planet."
That would explain it, I've seen things like Egassem colonies half finished on an Egassem world and always wondered how they managed it. So, a fairly obvious logic hole (and a fairly big waste of resources), I'm guessing not an easy fix in the decision tree logic tho, anyone?