Main page

At turn 180, I have 130 pop, killed the Psilons, own all planets in all systems that i could access with the unstable warp tech and lead technology god knows by how much as well as have most military power, even though I barely built any ships, I earnd 500+ aturn,actually I just have 2 cruisers and 2 destroyers. Probably if the Council meeting would be coming up, I would probably win right away.

Now I researched the tech to allow travelling unstable warp points, only to see that the A.I. has done NOTHING.

 

The AI has not even colonized premium estate next to its homeworld. Next to Altair is a prime spot, nothing there!

 

I have not even started playing and the A.I. is just doing nothing. The game has no AI yet, period. 

 

Please devs, you need to plan A LOT of time to get this right. In all your wonderful roadmaps we only see additional races and all kinds of stuff, while the basics are not in even barely in place. You set yourselves a very agressive timeline and I just wonder who you will pull it off.

 

And btw, we do not want a cheating AI, we want a smart AI. I just dont see it.

 

 

Needless to say, I won the game super easily. and at any time i would have had the council victory, had I not abstained. The most difficult thing was pressing 'next'.and traveling.

The humans had expanded a little and the cats, but they neither had deep pockets, nor tech, nor strong fleet, nor correctly upgraded planets,

I felt bad for them, so I left the Titan in the Orion system and just sent one Battlecruiser which destroyed everything. If tactical view is abused correctly, then even one of my destroyers can take out huge fleets.

 

Note The tactical AI of course does not exist as well.

The battlecruiser on Simulated might have lost, some battles where it was outnumbered a lot, however in tactical battle you can flank/flee after shooting the first volley of missiles and then kiting along. This will happen:

- Some smaller ships will jump close to you with their displacmenet and insta die, what a loss for nothing, they should stay behind

- All ships follow you (you are fleeing into some corner), but you got superior speed

- your PD takes care of their missiles while you stay out of range of most stuff

- you torpedos hit, as they are following in a straight line

- you get to recharchge your shields

Of course this is all good, but the AI should be smarter and try to split instead of just mindlessly follow. and flank and not insta die. AI should never equip displacement until a good logic for this is in place., Same for the bases, 1 shot is enough and everyone will follow you.

If both parties stay out of range for long enough, a draw is called, after all the attacker should get close enough to bomb the planet, if he stays far away, why should fleets suicide into him, they should wait as well.

 

Some Dev please write that down, bombing and invasion should happen in tactical view (only if there are defensive forces ofc), requiring you to get close. That would open some nice tactical scenarios. 

The game is not complete but I for one was very supriced that the AI was not refitting ships util last patch. Lots of AI work ahead.

 

But for starters the Star bases does not yet refit them self. They still shoot nuclear missiles late game and die way too fast. They should automatically equip the newest weapons, shields, armors, computers and available specials you have researched (Just like in MoO2), so they will scale with the rest of the of the ships and not be the push overs they are now. And likewise will of course the missiles bases need newest missiles and the ground(?) batteries need newest beam weapons.

 

One more thing. The star bases, battle stations/fortresses should have different models for each race. It is a simple way of adding more feeling: Now I am attacking the Sakra or Now am I attacking the Psillons. As it is now all the different races bases look precisely the same and all are equipped with the same weapons, so you know on before hand the result of the fight and how it will play out. At the moment I automate all planet and star base fights that is not defended by big fleets.

 

Since you have chosen to fill the map with bases will most fights include bases and they should not be the same fight over and over again. That is boring. The worst thing with the Total War games is that most fights tend do be sieges and they are hard to play in different ways. The field battles are more fun since there the battleground and unit setup gives more influences on strategies. This could be a big problem in MoO too since I recon it is hard to create star bases where more than a few tactics are valid.

 

Side point: Star bases needs large guns that reach all of the battle field or its easy to move in range and fire a volley of missiles and then just move away out of range and wait for recharge and repeat. Missiles should also have limited ammo x2 x5 x10 influencing the space cost (just like in MoO2.).

Edited by Gamling_JB, 09 March 2016 - 10:10 AM.

@LiveLock_1

 

 

If you want to test the AI you have to stick with following rule at the moment:

 

Don't colonize more planets then the computer race with the most planets has. There are some other restrictions (build not too big fleets so that the AI is attacking you, don't force population growth too much to not win the election early)

 

I would not say that there is "no AI" as you can watch it if you stick to that rules. I have to say I was surprised!

 

Some races seek allies to attack you (Humans do that)

Some just mess around with all other races (Alkari and Mrrshan tend to do that)

Some are more forgiving (Bulrathi and Humans)

Some never forget (Sakkra and Psilon)

Some are easy offended (Sakkra) some try to give you strange trade options...

 

 

I think it is a balance thing. If you will not be capable of building such an empire so fast you would see that the AI is doing a lot of things and I know what they are talking about when they say "we try to create stories".

 

I think for the moment we just have to accept that if we want to test the AI we have to stick to some rules. I don't expect that it will be so easy to grow an empire in the finished game. The process of balancing that is very difficult.

 

But I think we all agree at the moment that they underestimate the time they will need for this. How could they look at their timetables and still stick with it?

 

What i find inadvertently funny is when a faction picks a fight with you, declares war - and then it takes you another 300 turns of expansion and colonisation before you even get close enough to have a fight.

 

I finished one game last night on a spiral map. I had the Mrrshan as neutral neighbours in the system next door, blocking one of the only two entrances to my part of the galaxy. I (humans) thus expanded the other way, and on, and on, and on - the galaxy was completely empty. Eventually the Sakkra declared war on me - apart from a single scout, I hadn't even met them. The Psilon, Alkari and Bulrathi had evidently all killed each other off before I even encountered them. Finally, after 400+ turns, I closed in on the Sakkra - who turned out to be tucked in tight behind the Mrrshan. So about 90% of the map turned out to be entirely uncontested....

An AI needs a strategic component and a heuristic operational local component.

 

I believe the devs should focus on a smart heuristic AI first, basically fixed build orders based on certain conditions, that gives the AI a strong eco and agressive expansion, even though it may not be smart yet.

1) planet in the interior, do not build defenses 

2) front planet & no active war - skip marines as long as possible

3) have certain types of planets, balanced, tech planet etc and treat them accordingly

 

Note: For the players, actually it would be great if we could define certain build orders and assign them to planets - including a black list, like 'never build space elevator here' 'never build (high maintenance building' - of course the AI also needs to use these heuristics, if the AI builds expensive buildings, it will always be at an disadvantage, if it prioritizes non-economy buildings first.

 

For the strategic component the AI will have to be goal oriented, like going towards certain tech, some predefined, some based on what enemies the AI encountered, how well it can defend etc. It will have to make trade off decisions and this smart AI behavior I did  not see anywhere. Even if it was there, the heuristic part is so weak, that even if it were there you wouldnt notice as the AI is bogged down buy all the space elevators etc it built.

 

Actually if the devs were smart, they would crowdsource the heuristic part. They would implment this build order thingy above and then people would experiment with builds order and they could 'steal' them and test them in their AI. They could also collect data on build orders from all the players, automatically correlate with game metrics and then select good ones for the AIs.

 

I hope they are not lazy and then simply rely on the fact that they can always have let the AI cheat. It will be a rich but dumb AI.

@live_lock_1

 

Would be nice if there were buildorders that could be saved/reloaded. No black list, simple a list of buildings. You can save them with names like "Planet Basic start", "Planet Industrial", "Planet Research"

 

Problem is that those lists will not fit for the next game so you have to remake them everytime: the making of the lists should be done quick and clever. So no Drag&Drop and no lists with butttons.

 

I am surprised how good the planet buldingsscreen is working since I got used to it and how easy it is to find the buildings. In fact I bet that I am a lot faster then searching in a list with this planet view. Diden't ecpext that. Maybe they have similar good ideas for buildlists.

 

Problem is: most people will refuse to learn how to use it (use the colour/symbol as "one view" filter around the planet and you will find what you are searching very fast).

how is the plant buildingscreen working, its THE WORST ever,  i want to be able to sort by maintenance cost, name, type whatever - i just want a simple list that shows cost vs  benefits. I think they can reuse it for ground invasion graphics and bombing, but for building management honestly, this terrible.

 

what would be best is a building list that is integrated with your building queue, so that you can see when something finished, how much overflow you will have and so and son on.

 

Well it somehow works.

Despite saying I would not I started a game again.

 

Medium circular galaxy, very strange seeding, I was in the sw corner with the Psilon just a system apart and the Bulrathi and the bird guys also close by. While the humans had the whole east half for themselves and the Sakkara the north. Worst off were the Bulrahti they were set on a dead end with only one spacelane out to explore and that let straight into my expansion area. But anyway due to that weird seeding I was not able to expand as I would normally and was stuck on 3-4 planets for quite a while. So the Ai took some swipes at me, the usual killing of scouts, but also attacking lone warships even when not at war, but only when the Ai had a numerical and technical advantage.

I do not remember why but for some reason the Sakkra declared war on me quite early on even as we were far apart and had no contact apart from scout ships meeting. So I set out with a fleet of a few cruisers through the galaxy for a preemtive strike to be sure. While the whole fleet drove through various occupied systems without any issues, when I tried to send a damage4s cruiser back the same route he got jumped and finally destroyed by being attacked from the AI even as I was not at war with them.

 

In this game that guard does not work as you would think it should was a real issue, as the AI did their ninja colonizing, you clear out a system, set a warship to guard a planet before you can get a colony ship there, and some AI colony ship just drops by and settles on "your" planet.  Same issue that you can not close your borders, and any hobo can fly through your systems but the AI is quite quick in closing borders and trying to box you in,  

 

But as soon as I got a bit stronger again, the AI reverted to the chicken mode and did not try to attack.

Also weird issue, I attacked a system and had a blocked the access blocked to it with a cruiser and for about 10 turns the Ai sent 1 Frigatte after the other at the cruiser all getting killed, instead of pooling them up with 10-15 they might at least have managed to scratch the paint on it.

 

And then the AI all declared war on each other, I watched the galaxy map no fleets moving around and no systems changing even after many turns so even against itself it seems to gets chicken when the odds are even.

 

 

@Boomer7

 

I really know what you are talking about when saying "chicken" mode!

 

Some suggestions to get rid of it: The AI could decide to sacrifice old ships or ships it doesen't need to give the player constant stitches even if the attack fleet is to small to have a real chance.

 

That would more feel like "We declare war" then "chicken" mode. For the AI at the moment it is a "all or nothing" thing.  But I can imagine if you are at war with 2 races then the "stitches" will cause problems to you sooner or later when the computerplayers are gnawing at your empires edges.

 

I am very disappointed about the AI not defending it's home planet "irrationally" as a human would do it. I would not expect this from a Meklar.. but I would expect this from Humans or a Bulrathi.

The humans just let my fleet through and positioned at the next system, knowing that they can't beat me, waiting until the fleet thy have gets big enough tho leave "chicken" mode.

 

That is very rational for an AI. But is that the spirit of the human race?

Edited by Idinyphe, 10 March 2016 - 10:45 AM.

Against better judgement I started another game, it was even more boring, I am really disappointed.

 

I ended it around turn 100, and could have ended it probably around turn 50. Less than two hours.

Interestingly if you look by the scores, population, even fleet, etc you do not see how much better I was until 10 turns before I won. Only at the end when I decided to attack all the scope, military etc charts go exponential. You also would need to take into consideration earned $$$ and tech for the score.

 

Killed everything with 2 cruisers, 1 would have been sufficient. Probably destroyers even. Its just sad. Altari had a fleet of 12 or so ships, frigs and destroyers. Only 1 fight. My cruisers just fired a volley of rockets then turned around, they followed only to be shot down by mercuiltes while I killed all their missiles. Of course they could not hit anything with their proton torpodos lol.

 

This time I stuck with the advice to keep less planets than AI, and GNN would report Altari as leading in military and pop etc, they 5 worlds, me 4 worlds. All my worlds high class planets, some further away, Altari grabbed all kind of crapclose by. So the explanation of an AI turtle mode when you appear to be stronger is nonsense. Also they declared early war on me when i desintegrated one of their socuts and refused to pay credits. No follow up at all after the declaration of war.

 

So yeah, I will not touch the game for another 3 months+. Even if they built a massively cheating AI that has all information from the map and lots of $$$, that would face me with many cruisers and pulson by turn 100, it would not matter. I

 

I have no idea why some people on the german board mentioned that they wanted a lower difficulty level, I was in tears, I mean it couldnt be easier, why would you need an easier level. Like if you smoke 1kg of weed a day and maybe you forget to do anything in 20 turns, and build no ships. IDK

 

Yeah big disappointment. Also the devs seem not to read this or not to care to answer. They are asking for feedback, but are not responding and everything we read through the official channels suggest they have not even remotely noticed that this is not a beta, but a very very early alpha. 

 

To summarize:

- AI builds wrong stuff on wrong planets, has subotimal economy

- AI equips ships with incorrect stuff (proton torpedo is really stupid)

- AI does not do anything interesting

- Tactical AI does stupid things (like rushing into enemry fire or missiles), needs to be way smarter

- AI does not colonize correctly

- AI does stupid diplomatic stuff without follow up

- Ai does not react to anything I do, like adapting its designs

- AI picks techs in wrong sequence, probably predefined way

 

All of that is basic AI stuff, then on top of that I would expect some surprises, teaming up on me, coordinated. Grabbing planets that I would want etc. really cool and smart and suprising stuff. But thats even further away.

 

Sorry there is nothing positive to be reported here. Obviously I am focusing on AI and any automation aspects. The game itself is pretty, fast and nicely made. But its just a shell still, a pltatform.

@livelock_1

 

There might be a missunderstanding of the word "chicken mode". If the AI has the slightest idea of not winning against you (even if it is ahead in planets and fleet) it will avoid a fight. They will declare "war" on you but not attack until you give it the chance to. I think we are different, I had fun testing the game. I would not actual call it "play the game" as the game is not ready by far.

 

I think of this alpha version as a chance to play with the AI and how it is reacting to actions I do. That for I test a lot of actions, even stupid ones. I have no intention at the moment to play and win "the game" as this would be to easy in this state.

 

I would not call it "wrong" to put protons on your ships. It depends on the enemy. At the moment the AI doesen't analyze the ships, they have a basic (bad) setup mix. 

 

So you are right with your points but I want to ask you what are your suggestions to improve this? Where is the small and clever change that gives maximum effect? You are of course right, this is a platform and that is why they gave us the chance to play with it. What did you expect?

 

Suboptimal economy: is a balance thing. I did not find a way at the moment to build a worse economy then the AI. (Talking about those players in german forum: I did read this too and was surprised as you are...) They found a way and I would like to hear HOW they did that...

 

I think the most players focus on productivity and quick increaese of population and don't focus on research as this seems not to be nessecary. The research passes to quick so that researching in a different path woulden't give the AI any advantage.

 

I really would like to experiment with the settings of the game how things can be influnenced like number of turns and basic multipliers.

 

Equip ships:

 

The ship design is bad at the moment so there are 2 possible ways: modify weapon stats or doing better in analyzing the enemy. But if it comes to this that the key point of the game is who got more money to change the fleet fast enough to keep up with the enemy I would not like it. Maybe the idea of refitting for money is bad and the old "Orion" approach is better.

 

AI does interesting things. But you have to stop playing for "winning the game" so that you can see that Think of it: if the game is finished you can not overrun it as you do it now. (I HOPE that they will take the time to improve that even IF that would delay the game for months!)

 

Tactical AI is stupid.

To be honest I don't think that the current approach of tactical fight is interesting enough. For me options for formations, flanking and protection is missing. If I can not protect my rocket ships with a wall of heavy armour ships...

 

Diplomatic stuff:

 

The AI has a greedy approach "give me everything or face war". I think the make or break of the game is the diplomatic approach, the "stories" those interaction create. "Geve me everything" is not a story.... A story is when the humans allied with the bulrathi after kissing my gluteal regions for 50 turns and then talked the Bulrathi into attacking me from the west during attacking from the south! More of that, less of "Give me everything"

 

The AI has no understanding where the keypoints of the map are. You can cut it off from resourceful planets very easy. It will show no interest at all.

 

I would really be interested in the point of view with practical suggestions what they could do. I have the feeling everybody is complaining but nobody digs into it. I have to admit this is hard work and I have no time for that. But without good feedback (and I don't ecpect the moderators or developers to show up here in the forum...) they will fail. 

 

 

 

 

 

 

Edited by Idinyphe, 12 March 2016 - 09:27 AM.

How they built a worse economy is simple, the build unneccesary expensive buildings on planets that do not need them, they built ships that they did not need. You need to projecxt into the future what building will benefit you most o na specific planet and then chose the best path. Probably also they do not change the taxation rate to optimize for certain objectives conditions (growth of small planets with lower morale, vs cash cowing from bigger planets. Like the Automatic Build mode always builds Marines which is often not necessary, I only build them when remotely in danger OR when i could raise the tax rate 1BC without losing workers to strike.  There is so many things, it pointless to mention all of it here.

 

Their latest post makes me very sceptic about that they understood how ineffective the AI is, same for the tactical battles

http://steamcommunit...49508280580905/

I dont think they have an idea what to fix here and how much work it is. The bullets listed are minor tweaks suggesting that everything is going great.

 

And also I base my criticism on that they said they were releasing a rather finished game into early access, not sure where I read that, pretty sure I did read it somehwere.

 

If they want to learn what AI they need to build they should watch some professional Starcraft 2 games. The very basics that you can observe is also the basics for an AI.

- Very strong and smart economic expansion

- Monitoring your environment and enemies, what they are doing, assesseing how storng they are, what their abilities are

- Based on that making trade off decisions, trading off risk dimensions (of being wiped by focusing on economy only vs building minimal fleet to defend while focusing on eco vs building strong fleet to counter attack or even to agressively attack)

... this is just based economy and there is so much more to be considered.

I realized that I bought Moo2 with this, I installed it, installed the ICE 11b mod and will be playing this then. Maybe they will fix the AI, introduce leaders and all the other cool stuff the original had. I am not a fanatic of replication the original, but right now the game has 0 depth.

@livelock:

 

I can build what I want, I build all buildings available... so ist must be the order of bulidings they are messing with.

 

There is a very simple straight order:

 

* Industry Improvment -> Terraforming -> Cash. I build all lower tier buildings before stepping to the higher ones.

 

Typical would be:

 

* Biosphere

* (Crust prospecting)

* Automated factory

* Government support

* Colonial revenue service

* Research

* Neutron collider

* Terraforming

* Orbital shipyard

* Space elevator

 

This is not the best quickest/best  way but it is enough to blast the other races out of space.

 

I think one part of the problem is that tech is outdated so quick. You don't have to struggle for it or use it for 50 turns. You switch so fast from tech to tech...

 

THIS is part of the problem, because the galaxy is built with far too many habitable planets ... by which I mean the "set down your colony ship and start breeding/building" type. This will result in ...

1. The AI (if it ever gets really smart) swarming opponents AND having to be held back for easier difficulties.
2. Human players getting overwhelmed (or even bored by having to go through 20 planets to tell them what to build) by the amount of stuff they MUST do to have a chance against a really good AI.

Galaxy generation should make "instantly habitable" planets really really REALLY rare (exception: racial tolerances should make a REAL difference in how hard it can be to colonize certain planets) AND for most other planets there should be a requirement of using the Space Factory in a "preliminary planet preparation". For instance: if we humans ever truly go to Mars to live there we will need an atmosphere, water and probably some plantlife. This would have to be engineered decades before colonists ever set foot on the planet but could range from "full planet wide" to "a large enough biosphere which works".

 

Fewer colonies will make the game "better defined" and it is important to settle the right planets instead of simply colonizing everywhere ... which is something you can easily do now due to the snowballing effect of population growth not being a limiting factor either.

Of course you should not build simply all the buildings. You only should build th emost effective building. A game that simple makes you build all buildings is taking away all the thinking. Once you know the best sequence you lose any interest and win.

 

Why would you build orbital and space elvator where you are not building ships, even neutron collider. There must be research planets, ship building planets, planets with def, planets without def.

There must be choices and trade offs and not just, simply click the next building in the list. Thats maybe what I expect in a browser game that retarded people play, but nothing that is close to Moo.

 

You are probably right in this. I just started a Moo2 game, so few good planets to chose from and many trade offs wrt Security, defendability. In this game you can simply take pretty much any reasonbly good planet and there is almost no consequence.

@livelock_1 and rabiator

 

 

You are both right but in my opinion that process is exactly the process of balance that they have to do at the moment. They did not claim that they already have a balanced game.

 

I just think that the time they have is not enough to do a good balance

I am afraid there is always to much pressure from people that have no idea how important this balance is for the fun of playing the game. Those people are calculators, they see some good looking pictures and force for release even if the most important part is not done

 

But this makes or breaks the whole game...

just completed Moo2 after many many years not touching, its great fun. Ofc I also beat the AI handily on `average`, but it gave me an interesting moment here and there, never too far behind me.

its now easy to see that many of the simplifications they made for Moo4 to be easier actually take away the thrill from the game.

 

with the choke points and the free military posts you know exactly whats coming, just move your ships there and thats your contribution. With Moo2 you dont know, they can come from everywhere, through a beach head etc. some things are weird. the lanes are probably a stupid idea. in Moo its open and while you attack you enemy on world x, he can destroy your outpost or even attack behind your lines at the same time.

 

sigh, if the developers had just learned from this :-(