Some feedback for 0.45 after 3 games

Describe your experience with the latest version of FreeOrion to help us improve it.

Moderators: Oberlus, Oberlus

Forum rules
Always mention the exact version of FreeOrion you are testing.

When reporting an issue regarding the AI, if possible provide the relevant AI log file and a save game file that demonstrates the issue.
Message
Author
Damiac
Space Floater
Posts: 20
Joined: Mon Oct 05, 2015 4:47 pm

Some feedback for 0.45 after 3 games

#1 Post by Damiac » Mon Oct 19, 2015 5:39 pm

First off: This game is already pretty fun, and has that "One more turn" feeling, which is critical for 4x games. The micromanagement is kept to a minimum, which is really nice.

Now, let me focus on what I see as some negatives.

Xplosive Xpansion.
Expanding is easy, especially when you have a few growth techs under your belt. Too easy, in fact. And it only gets easier as you go, at least until you've taken up all the planets available to you. True, some production goes into expanding, but it pays for itself incredibly quickly.

The solution isn't too complicated in my eyes. I think the best way to slow down expansion, and make it something the player chooses to do, rather than a no-brainer, is to make it more of a long term investment. Make it so the population for the planet actually has to come from somewhere! I know right now you need to have the species on a planet with enough population to be able to colonize, but the new colony doesn't actually take population away from anywhere. And it should.

If the initial population for a planet were subtracted from the source planet, it would slow down expansion quite a bit, because you'd have to wait for population to build up to expand further. If you made the population cost higher than the initial population of the new planet, it would mean expanding would temporarily hurt your economy, although it would still pay off in the long run. This creates an interesting choice of when to expand, and creates the opportunity for a player to take advantage of another player's overexpansion.

AI Trooper Troubles
I've seen this one brought up before, but the dev response was... underwhelming. I just want to make sure it's clear to the devs how bad this problem is. The AI is generally fairly decent, and I know it's still a work in progress, but when it comes to troop ships they are flat out stupid.

The AI regularly sends unescorted troop ships to my planets. Sometimes, there's not even an attack force anywhere around. Other times, they're a planet behind the attack force, meaning the AI could have capped my planet if he'd just left an attack ship behind to keep the shields down. But most of the time, they're just all by themselves, milling around in enemy territory for no particular reason. I would guess that the AI throws away somewhere between 1/4 to 1/3 of their production budget in suiciding troop ships. On top of that, the AI misses the opportunity to capture planets because it doesn't leave an attack ship to keep shields down for the troop ships.

The easy solution would be to make sure the AI knows that troop ships should never, ever be without an escort outside of their own territory. They already know to group up attack ships so they can't be easily picked off one by one, so it seems to make sense to extend that logic to tell them to group their troop ships with their attack fleets. The nice extension of that fix is that it prevents troop ships from sitting around on planets with shields, doing nothing.

Tame space monsters are lame
My last game, I took over several kraken and juggernaut nests. Why not, right? It keeps them from spawning annoying enemies, and I'll get some free ships. The problem is, I didn't really get anything until about turn 120 or so, and they're so incredibly slow, and have such weak weapons for that time in the game, that they served no purpose whatsoever. This is a much more minor complaint, and I know work is being done on them, but I figured I'd throw my two cents in since I tried them out.

User avatar
Kassiopeija
Dyson Forest
Posts: 212
Joined: Wed Jul 16, 2014 6:14 pm
Location: Black Forest

Re: Some feedback for 0.45 after 3 games

#2 Post by Kassiopeija » Mon Oct 19, 2015 8:33 pm

I agree on some points you raise. First off, it's always bad esp. for a strategy game, if something is created out of nothing. A colony ship can create 3 population in one turn whereas a lowpop planet that starts with 1 pop needs the hell of a lot of turns to reach that....
So once a planet is colonized the new population should be subtracted from the general population of that species in the empire.
(taking pop away from an individual planet could be problematic because of the way how you can create colonies via outposts....)
That would stiffle the productive explosion a bit.

I also find that some production or research multiplicators are much too much, esp. when combined with an aim to go for high populated planets.
But tbh, nerfing something here and there won't make the game any more challenging. The thing is that, once you've got the proper research path figured out, you can practically overrun all AIs at any time of the game if you just focus on heavy colonization early enough. There are no real good alternative options available, at least not on maps created by an average/medium setting.
The point is that once I meet the AI I already have 10 times the amount of his production+research, and perhaps already more production than all AIs combined, which also enables me to be technologically so far ahead that I can beat his military quite lossless with just a few ships and therefore, bank most of my production into colonization/terraforming - which will only raise my production again and only make matters worse.

I've personally modded the game quite a bit in order to make these things much harder for me (for example, increased the shipcost multiplier, reduced production etc) but in all honesty, it didn't change anything because the AI also did suffer from these changes.

There needs to be some sort of penalty that grows exponentially with military or productive strength. That is, the more you grow the more you have to struggle to overcome this penalty, that would give weaker empires some more breathing room. That would make it harder for a player to win the game, although it would also make it harder for the AI to defeat the player^^ That's why the AI needs additional bonuses. Why aren't there any difficulty levels which give the AI hidden bonuses? That's not a cheat, every and any 4X game that I know of has it, some even give +50% bonuses as a normal level to reflect on the fact that no AI can make as good decisions as a human player....

just my 2c

edit:
to add to the space monsters, I actually find them quite useful, although I didn't see them in the couple last games perhaps something has been changed. I remember (somewhat smallish) games where most of my military were Krakens, which were invincible at 3rd stage, which I got quite early but that might be random. The Juggernaut seems quite OP compared with the other 2.

AndrewW
Juggernaut
Posts: 781
Joined: Mon Feb 04, 2013 10:15 pm

Re: Some feedback for 0.45 after 3 games

#3 Post by AndrewW » Tue Oct 20, 2015 1:07 am

Kassiopeija wrote:There needs to be some sort of penalty that grows exponentially with military or productive strength. That is, the more you grow the more you have to struggle to overcome this penalty, that would give weaker empires some more breathing room. That would make it harder for a player to win the game, although it would also make it harder for the AI to defeat the player^^ That's why the AI needs additional bonuses. Why aren't there any difficulty levels which give the AI hidden bonuses? That's not a cheat, every and any 4X game that I know of has it, some even give +50% bonuses as a normal level to reflect on the fact that no AI can make as good decisions as a human player....
Other games giving the AI bonus's unavailable to the human player does not make it any less of a cheat. This was discussed awhile back on the forum. One idea that surfaced was to make it optional, where you would specifically be giving the AI bonus's, rather then a simple difficulty that you really don't know what it is doing. So it would be the players choice if the AI cheats or not.

User avatar
Dilvish
AI Lead, Programmer
Posts: 4768
Joined: Sat Sep 22, 2012 6:25 pm

Re: Some feedback for 0.45 after 3 games

#4 Post by Dilvish » Tue Oct 20, 2015 1:31 am

Damiac wrote:AI Trooper Troubles
I've seen this one brought up before, but the dev response was... underwhelming.
Hehe, you may very well still find this response underwhelming, but I'll give it a go anyways...
I just want to make sure it's clear to the devs how bad this problem is. The AI is generally fairly decent, and I know it's still a work in progress, but when it comes to troop ships they are flat out stupid. The AI regularly sends unescorted troop ships to my planets. Sometimes, there's not even an attack force anywhere around.
After the previous post that you refer to, Cjkjvfnby (one of our other AI devs) offered to take a look at this, but he and I have both been busy and I haven't followed up with him to see how things are going. I myself have not noticed this problem in the recent test builds; that might just be coincidence (and I haven't had time to play much lately), or it could be that some tangentially related changes have serendipitously made for a big improvement in this behavior. Regardless, for anyone interested enough to post on these boards I would encourage you to play with the current weekly test builds rather than 0.4.5; there are a wide variety of ongoing improvements.
Other times, they're a planet behind the attack force, meaning the AI could have capped my planet if he'd just left an attack ship behind to keep the shields down.
I have sometimes still seen an issue with that. There is code in place that generally prevents that, but under some circumstances the warship could get called off for a higher priority task. It's something I'll be keeping in mind to pay attention to and look into a fix for (and perhaps CJ will make headway on that), but I think it won't be trivial and it won't be soon that I myself would be able to do much on it.
But most of the time, they're just all by themselves, milling around in enemy territory for no particular reason.
Hehe, it's an amusing image, but no, it's simply not like that.
The easy solution would be to make sure the AI knows that troop ships should never, ever be without an escort outside of their own territory.
The concept is simple, and is of course one I've contemplated when I first begain working on invasion planning. Some basic implementation of it would not require too huge of a modification to current code, I think, but actually doing it reasonably well would not be easy, and I don't think it's really the right answer.
If I provided any code, scripts or other content here, it's released under GPL 2.0 and CC-BY-SA 3.0

User avatar
MatGB
Creative Contributor
Posts: 3310
Joined: Fri Jun 28, 2013 11:45 pm

Re: Some feedback for 0.45 after 3 games

#5 Post by MatGB » Tue Oct 20, 2015 1:50 am

For what it's worth, I continue to observe it, and a few weeks back even saw a warship stay on a high tech natives system until the turn before the troops turned up and then moved off, the troops then arrived, the shields were back and they sat there until I killed them. That's an edge case but seeing troop ships lead and stay in very dangerous positions is very common.
Mat Bowles

Any code or patches in anything posted here is released under the CC and GPL licences in use for the FO project.

User avatar
Dilvish
AI Lead, Programmer
Posts: 4768
Joined: Sat Sep 22, 2012 6:25 pm

Re: Some feedback for 0.45 after 3 games

#6 Post by Dilvish » Tue Oct 20, 2015 3:18 am

MatGB wrote:For what it's worth, I continue to observe it, and a few weeks back even saw a warship stay on a high tech natives system until the turn before the troops turned up and then moved off, the troops then arrived, the shields were back and they sat there until I killed them.
Yes, that's exactly the thing I said above that I do sometimes still see and intend to follow up on. I have an idea of what might be causing it. If I'm right about the cause then I have an idea on fixing it as well, but neither confirming the problem nor my anticipated fix is simple enough for me to squeeze in anytime really soon.
If I provided any code, scripts or other content here, it's released under GPL 2.0 and CC-BY-SA 3.0

Damiac
Space Floater
Posts: 20
Joined: Mon Oct 05, 2015 4:47 pm

Re: Some feedback for 0.45 after 3 games

#7 Post by Damiac » Tue Oct 20, 2015 3:47 pm

As I said, I've only played 3 games through, so my observations are obviously limited, and I may be seeing patterns where there aren't any. I'll certainly try the latest test builds from now on, I can see how feedback for problems you may have already fixed might not be terribly useful. Plus I'm going off memory, I wasn't exactly taking detailed notes as I played, so it's possible I'm exaggerating in some places. I realize my suggestion to force the AI to keep troop ships with an escort was more of a patch over the problem than a real solution, so maybe that's not desirable, but I do think it would give good results in the short term.

My other suggestion about expansion wasn't really meant as a way to make the game harder vs an AI, just to make expansion more of a strategic option rather than the obvious thing that you should do in almost any circumstance. As it stands now, it's pretty hard to justify NOT taking any planets you can get a population on, ASAP. I feel like there's a really fast ROI for expanding, and I think it would be good for the meta strategy aspect of the game if that ROI was much slower. Taking population from the source planet(s) seemed like a fairly easy way to push that out a little bit, but I'd think when you guys get infrastructure more integrated into the game it will also be able to serve as a slowdown for expansion. Expansion should be a long term investment with higher up-front costs, in my opinion.

I'm thinking of the civ games here, which are obviously different in many ways, but the expansion parts of both games are quite similar. In Civ, however, building a settler has an immediate cost to your economy: the city supplying the settler actually loses population, thus losing production. In Civ, it wasn't the best strategy to just devote your entire economy to expansion until you were surrounded, but I feel that in freeorion that is the best strategy.

I appreciate you spending the time to read my scattered thoughts about this project. I think you guys are doing great work, and there have been plenty of games that sold at full retail price with worse AIs than this game has now.

User avatar
MatGB
Creative Contributor
Posts: 3310
Joined: Fri Jun 28, 2013 11:45 pm

Re: Some feedback for 0.45 after 3 games

#8 Post by MatGB » Tue Oct 20, 2015 5:10 pm

Damiac wrote: My other suggestion about expansion wasn't really meant as a way to make the game harder vs an AI, just to make expansion more of a strategic option rather than the obvious thing that you should do in almost any circumstance. As it stands now, it's pretty hard to justify NOT taking any planets you can get a population on, ASAP. I feel like there's a really fast ROI for expanding, and I think it would be good for the meta strategy aspect of the game if that ROI was much slower. Taking population from the source planet(s) seemed like a fairly easy way to push that out a little bit, but I'd think when you guys get infrastructure more integrated into the game it will also be able to serve as a slowdown for expansion. Expansion should be a long term investment with higher up-front costs, in my opinion.
I, mostly, agree with this, and am increasingly sold on the draft proposals for certain bonuses not kicking in until a planet has a minimum infrastructure which I'll be testing as soon as I'm done with current project. However, I am suspecting differing strategies and relative experience levels play a bigger part here.
I'm thinking of the civ games here, which are obviously different in many ways, but the expansion parts of both games are quite similar. In Civ, however, building a settler has an immediate cost to your economy: the city supplying the settler actually loses population, thus losing production. In Civ, it wasn't the best strategy to just devote your entire economy to expansion until you were surrounded, but I feel that in freeorion that is the best strategy.
The problem with this is twofold, 1) if buidling a colony ship reduced population by the same amount, the planet losing the population would likely be a high population planet (ie: your homeworld), and they can recover population incredibly quickly, basically with the current maths you would barely notice the effect as a 'full' large planet can grow at several population per turn.

2) I disagree, personally, that full on colonisation-expansion is the best strategy, I'm currently playing on High planets (something I rarely do but need to test some stuff with) and I'm finding myself ignoring at least 40% of habitable systems most of the time, sure, in the early stages it's grab what I can, but once I've got a nice species mix I only grab strategic systems or specific resources (then fill the system). Military conquest remains the most cost effective expansion method despite the recent changes making colonisation easier and quicker.

I still want to tone the latter down and a relatively soon project will be to redo fleet upkeep modifiers and increase various costs a fair bit in the mid to late game.

It's also possible that the game's still incredibly steep learning curve makes some strategies remarkably easier to learn, but not necessarily better, we do need to improve tutorials and have some sort of mission/quest system (or similar) to encourage going to war early (which is what the game's about afterall).
I appreciate you spending the time to read my scattered thoughts about this project. I think you guys are doing great work, and there have been plenty of games that sold at full retail price with worse AIs than this game has now.
Feedback is always good, and I agree completely about the AI, I'm constantly finding the flaws with it but it regularly impresses me how good it is despite being nowhere near 'complete', there are fully released games out there where the AI is massively more stupid.
Mat Bowles

Any code or patches in anything posted here is released under the CC and GPL licences in use for the FO project.

Damiac
Space Floater
Posts: 20
Joined: Mon Oct 05, 2015 4:47 pm

Re: Some feedback for 0.45 after 3 games

#9 Post by Damiac » Tue Oct 20, 2015 5:38 pm

You make a fair point of course, I have only played 3 games, so perhaps I'm wrong about mega expansion being the best strategy, however, aside from the scenario of impending invasion, or the opportunity to disrupt an opponent's expansion, it's hard to imagine a time when an outpost ship and colony isn't worth the investment. There's no time when you decide, it's better to hunker down with what I have now and ramp up production, because you never have to choose between expanding your territory or improving it.

Maybe that's something I should have brought up in my first post. In Civ, you have to divide your non-military production between expanding and improving what you already have. There's no such thing in freeorion, aside from occasional 1 off buildings you unlock when you first get a new tech. Maybe if a planet had to go to 'colonization' focus to colonize another planet it could cause a similar effect, although without some kind of automation that sounds like a micromanagement nightmare... kinda like civ!

I suppose it might be of some value for someone (maybe even me!) to play a series of games with the exact same map seed and settings, and try a few different overall strategies, meaning one game you'd just go all out expanding, the next take only x%, the next take only what's necessary to get your military to your enemies, and seeing how your overall status looks after some number of turns. That could at least give some idea of whether certain strategies truly are better than others.

Morlic
AI Contributor
Posts: 268
Joined: Tue Feb 17, 2015 11:54 am

Re: Some feedback for 0.45 after 3 games

#10 Post by Morlic » Tue Oct 20, 2015 6:05 pm

Damiac wrote: I'm thinking of the civ games here, which are obviously different in many ways, but the expansion parts of both games are quite similar. In Civ, however, building a settler has an immediate cost to your economy: the city supplying the settler actually loses population, thus losing production. In Civ, it wasn't the best strategy to just devote your entire economy to expansion until you were surrounded, but I feel that in freeorion that is the best strategy.
Guess you never played Civ II then where the best strategy was to literally spam cities on every single tile possible. Later games obviously changed that and especially Civ IV was very punishing any overexpanding if you couldn't back it up with an already strong economy.


Anyway, I mostly agree with Mat. I do not think that colonisation itself is too powerful. What I personally think is way too powerful at the moment are the flat boni you gain regardless of population (+2 research and +5 production from the relevant techs). Due to this, even 1 pop planets have resource output of >50% of your "average" planet. Also, the relatively slow growth of population would in principle delay the ROI (at least until cryo pods) but because the meters still grow with 1/turn and reach the 2res/5prod thrshold as early as 5 turns after colonisation, the entire mechanic is rendered completely redundant.

For the very same reason, early invasions of native species are completely broken. Often, the required troop ships cost significantly less than to colonize and you acquire a planet with pretty much full population so the resource output is a high multiple of your own freshly founded colony (at least until you get the 5 flat prod from Adaptive Automation). Also remember that you gain a homeworld, i.e. the max pop is increased significantly.

This means even without the other strategic advantages such as having more colonizable planets (due to different environmental preferences), getting complementary research/industry boni compared to your starting species or relevant piloting traits it is almost always the best idea to invade a native planet immediately (or focus your gameplan on building the necessary fleet to do so) instead of expanding peacefully.
This does not only limit the viable strategies but also means that players without concquerable natives nearby have a significant disadvantage.


About adding Boni to the AI: I would prefer seeing selectable handicaps for the human player (which could also be interesting in multiplayer). There should be a variety of different handicaps to be selectable to adjust for preference and finetuning of game experience/difficulty:
I) Starting Boni
- Reduced number of starting ships (or worse ones, e.g. outpost instead of colony ship)
- reduced planet size on starting planet
- no cultural archives
- no shipyard
- no drydock
- population of homeworld not maxed at game start
- resource meters start at zero

II) Overall Gameplay
- Increased Fleet upkeep (some different factors selectable)
- Increased Colonisation upkeep (also different factors selectable)
- Increased Construction cost for buildings
- Increased Research cost

(The resource output should probably remain the same considering meters and conquests)
If I provided any code, scripts or other content here, it's released under GPL 2.0 and CC-BY-SA 3.0

Damiac
Space Floater
Posts: 20
Joined: Mon Oct 05, 2015 4:47 pm

Re: Some feedback for 0.45 after 3 games

#11 Post by Damiac » Tue Oct 20, 2015 6:43 pm

Heh, well I know in civ even with the small costs of expansion, it was still worthwhile, my point wasn't that their system actually accomplished what they wanted, just that they imposed a slight up-front cost in population for settlers, and that sort of up-front cost is a good way to move the ROI point to wherever you want it to be.

Good point about natives, some of them really are freebies. Probably 0 defense natives ought to be rethought, at least the high tech natives demand some military investment.

Nexus
Space Floater
Posts: 47
Joined: Fri Aug 14, 2015 8:29 pm

Re: Some feedback for 0.45 after 3 games

#12 Post by Nexus » Tue Oct 20, 2015 9:32 pm

Why not defend useful native species with a fleet strong enough to be challenging in the early game? The really good specials are often defended by a powerful monster, and this situation is similar in some ways.

User avatar
Kassiopeija
Dyson Forest
Posts: 212
Joined: Wed Jul 16, 2014 6:14 pm
Location: Black Forest

Re: Some feedback for 0.45 after 3 games

#13 Post by Kassiopeija » Tue Oct 20, 2015 9:57 pm

Damiac wrote:In Civ, you have to divide your non-military production between expanding and improving what you already have. There's no such thing in freeorion, aside from occasional 1 off buildings you unlock when you first get a new tech. Maybe if a planet had to go to 'colonization' focus to colonize another planet it could cause a similar effect, although without some kind of automation that sounds like a micromanagement nightmare... kinda like civ!
To be fair, Terraforming + Gaia actually let you improve what you already have. However, considering the costs I find it more rewarding to spend the production into Military and conquer planets. Only if no additional military is needed, enough troop ships are en route, and all free planets have an outpost ship enqueued or en route to them, I bunker the remaining production into terraforming or gaia. There might be exceptions with Large or Huge planets, but on the other hand, you've also got Small or Tiny planets where it remains questionable if your investment even pays off...
If there wouldn't be so much excessive production capacity I would perhaps ignore them, but as of now I'm happy I find things to bunker production without creating additional stuff that creates lag (such as ships).

And it is quite a micro because you have to do all planets individually, once completed, backcheck if good for gaia or additional terraforming is needed, so sometimes you have to return 4 times to a planet until it is finally gaiatized^^

User avatar
Kassiopeija
Dyson Forest
Posts: 212
Joined: Wed Jul 16, 2014 6:14 pm
Location: Black Forest

Re: Some feedback for 0.45 after 3 games

#14 Post by Kassiopeija » Tue Oct 20, 2015 10:12 pm

MatGB wrote: The problem with this is twofold, 1) if buidling a colony ship reduced population by the same amount, the planet losing the population would likely be a high population planet (ie: your homeworld), and they can recover population incredibly quickly, basically with the current maths you would barely notice the effect as a 'full' large planet can grow at several population per turn.
But you can also create colonies via outposts, via outpost ships that can be from another species. So you either subtract from another species (which is quite illogical) or you have no real source planet...

But I think his point still stands, because no matter how fast the pop grows back it is still missing -3pop for a turn, which will decrease production... And perhaps, as an example, if the pop is taken from the nearest same-species planet, it might be one that also got colonized just prior a few turns earlier, and that loss might take 10-20 turns to grow back....

User avatar
Kassiopeija
Dyson Forest
Posts: 212
Joined: Wed Jul 16, 2014 6:14 pm
Location: Black Forest

Re: Some feedback for 0.45 after 3 games

#15 Post by Kassiopeija » Tue Oct 20, 2015 10:18 pm

Nexus wrote:Why not defend useful native species with a fleet strong enough to be challenging in the early game? The really good specials are often defended by a powerful monster, and this situation is similar in some ways.
Well, there's this natives special that gives them more troops and better defenses+shields, but, to be fair, troops are cheap to build and a good native planet early on is mostly worth the cost of perhaps loosing a warship and some troopers in the conquest battle. Esp. if that species allows the colonization of new planet types or have better pilots.

I really like the Acirema solution as they build their own ships. What actually do the other Natives with their production? Some have their focus even set to production but never release a ship? They can't impossibly go on and make buildings endlessly....

Post Reply