Am I the only one who never won a game?

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

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

Re: Am I the only one who never won a game?

#16 Post by MatGB »

Oooh...

*steps up plans to learn python ont eh priority chart*
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 and Programmer Emeritus
Posts: 4768
Joined: Sat Sep 22, 2012 6:25 pm

Re: Am I the only one who never won a game?

#17 Post by Dilvish »

Vezzra wrote:AFAIK our AI programmer (Dilvish) has nerfed the AI on easier settings already in response to complaints like those in this thread. But these changes are of course only in test builds released after 0.4.3.
Yes, there are some very significant newer nerfings, like limiting the number of planets the AI will attempt to conquer.
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
vincele
Space Dragon
Posts: 341
Joined: Sun Mar 23, 2014 6:10 pm

Re: Am I the only one who never won a game?

#18 Post by vincele »

I had the same thoughts about the game difficulty, in beginner mode, and found it incredibly difficult. Now I start to win games from time to time (still at beginner) and here's my ideas for taming down the aggressivity / difficulty:

* AI players should not be aggressive at all, they should not attack, only colonize, at level 2 they should only start to attack to retaliate, and then after a while forget about it if no further aggression, at level 3 they can attack ships but not invade planets, etc... i.e. GRADUAL difficulty levels
* Placement of initial colonies should be further apart, depending on game harshness level, f.e. a 2 players in a 2 arms spiral should get at least 1/3 of the galaxy between them, I always get to meet the AIs at beginning of the game sometimes as early as turn 15 (OK I'm playing smallish galaxies, but...)
* in easy mode, human player should get better planets to start with: starting system with one huge + one medium with starting colony, for example...
* disable xenoarcheology for AI, even without it we get wiped too easily, so imagine with...

I think newcomers (who choose beginner mode) may accept their first game to be defeated, but after that they will not think the game is cool if they cannot win at the easiest level.

BTW: where are specials added to systems / planets (in the code) ?
All the patches I'll provide for freeorion will be released under the GPL v2 or later license.
Let's unleash the dyson forest powa!

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

Re: Am I the only one who never won a game?

#19 Post by MatGB »

vincele wrote: * Placement of initial colonies should be further apart, depending on game harshness level, f.e. a 2 players in a 2 arms spiral should get at least 1/3 of the galaxy between them, I always get to meet the AIs at beginning of the game sometimes as early as turn 15 (OK I'm playing smallish galaxies, but...)
That should now be fixed in the most recent test build, Vezzra did something about it last week, it's been a gripe for ages. Run a few test setups.
* in easy mode, human player should get better planets to start with: starting system with one huge + one medium with starting colony, for example...
Disagree, I see where you're coming from, but the big problem in the game is the lack of explanation/ease of learning, giving a player a big advantage over the rather limited AIs is the wrong way to go (most other games of this nature I've played give the AIs the advantages to make them something of a challenge afterall).
* disable xenoarcheology for AI, even without it we get wiped too easily, so imagine with...
You can turn specials off if you want to, that might be something we recommend to new players, but if the feature is there it should be always available.
I think newcomers (who choose beginner mode) may accept their first game to be defeated, but after that they will not think the game is cool if they cannot win at the easiest level.
I mostly agree there, but as this is still an alpha release and the strategy within it is similar to other 4X games in what you need to do if not details, getting it balanced so that what you need to learn to do is fairly intuitive will help more.

Making the tech tree more user friendly so that obviously "good" techs are obvious might be an approach, I think the easiest mistakes to make are researching the wrong things early—to me, Subterraneam Habitation is such an obvious first take that I don't think too much about it, but it's not even a first tier tech displayed on the default view.
BTW: where are specials added to systems / planets (in the code) ?
[/quote][/quote]
Specials.txt in Default determines what they do and how common they are, and then specials.py in galaxy generation determines the specifics.
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 and Programmer Emeritus
Posts: 4768
Joined: Sat Sep 22, 2012 6:25 pm

Re: Am I the only one who never won a game?

#20 Post by Dilvish »

vincele wrote:I had the same thoughts about the game difficulty, in beginner mode, and found it incredibly difficult.
Hmm, it sounds like Beginner mode would be better nerfed a bit more. I think it is reasonable to still squeeze in some change for 0.4.4 if we decide quickly. Though I have some tentative family plans to go travelling the next couple days so... no promises.
[at Beginner level] AI players should not be aggressive at all, they should not attack, only colonize,
As in not build any attack ships at all, nor research planet defense tech? That would be possible, but really only suitable if Monster Density defaulted to None or Low, and could be reasonable done in time for 0.4.4, I think.
at level 2 they should only start to attack to retaliate, and then after a while forget about it if no further aggression, at level 3 they can attack ships but not invade planets, etc... i.e. GRADUAL difficulty levels
The whole retaliation-only thing is certainly not doable for me in time for 0.4.4, and I don't really like it anyway, sorry. Having levels 1 & 2 exclude invasions is very reasonable.

* Placement of initial colonies should be further apart, depending on game harshness level, f.e. a 2 players in a 2 arms spiral should get at least 1/3 of the galaxy between them, I always get to meet the AIs at beginning of the game sometimes as early as turn 15 (OK I'm playing smallish galaxies, but...)
I see now that Vezzra just recently did make a dynamic assessment of a starting min jump distance between homeworlds:

Code: Select all

min_jumps = max(int(float(len(systems)) / float(num_home_systems * 2)), 5)
, which he iterates down from if it doesn't work out.

I had also been planning on doing something, I'll go ahead and mention for consideration just because I already put the thought into it and it's a little fun :D : It can't be too close to the projected fully-distributed spacing , since the placement of the first few then might not be compatible with the required spacing for the remainder. But if we made a conservative estimate of densest possible packing and stepped it down from that it could work. Say, divide total number of systems by number of empires for an average allocation. if each allocation were distributed in a circle with hexagonal packing (sort of corresponding to high starlane density), then the number of systems that could be packed, tiered by jump radius 0, 1, 2, 3,...r, etc, would (I think) be 1, 7, 19, 37... (1+6r+r(r+1)/2). If we take the number of systems as S and the number of empires as E and solve the quadratic for r, we get

Code: Select all

r = ((161+8.0*S/E)**0.5 -13)/2.0
Theoretically with ideal placement you would be able to have about twice that radius between homeworlds, but because of the vagaries of starting placement we should just take something like 1.5 times radius, or

Code: Select all

r = math.ceil(0.75*((161+8.0*S/E)**0.5 -13))
For our default of 150 systems and 6 AI, this would work out to 5. If I run Vezzra's current approach with eliiptical shape and high starlanes, it starts with an attempted min jump distance of 10, but has to iterate down, seemingly every time to ... 5 :D So Vezzra's approach seems to work dandy but I just had to mention my idea :lol:
* in easy mode, human player should get better planets to start with: starting system with one huge + one medium with starting colony, for example...
Hmm, this wouldn't currently be possible without a bit more work on the C++ side. The server does know what client type all the players have, and with a small bit of work this information could be made available to the python script, which could then give all human-client empires a starting bonus. But, this is seeming like rather a bit of complication and hassle that just shouldn't be needed, and doesn't so much fit with our approach-- simply nerfing the operation of the AI should accomplish our goals just fine.
* disable xenoarcheology for AI, even without it we get wiped too easily, so imagine with...
That would be easy enough to do, and sounds reasonable for Beginner mode.
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
Dilvish
AI Lead and Programmer Emeritus
Posts: 4768
Joined: Sat Sep 22, 2012 6:25 pm

Re: Am I the only one who never won a game?

#21 Post by Dilvish »

Dilvish wrote:Having levels 1 & 2 exclude invasions is very reasonable.
In fact, code is already in place that should be preventing Beginner AIs from building any regular troopships until after turn 150. Though I suppose if they had a colony insystem they might still do invasion via 'SpaceInvaders' colony-base-troopers; I guess could restrict that as well. Had you been running into trouble with that? Or I suppose the 150 turn headstart may simply not always be enough.
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
Vezzra
Release Manager, Design
Posts: 6095
Joined: Wed Nov 16, 2011 12:56 pm
Location: Sol III

Re: Am I the only one who never won a game?

#22 Post by Vezzra »

Dilvish wrote:

Code: Select all

r7333 AIs now report their aggression level via the chat window at game start 
I've seen you committed this addition to trunk - might be useful in the 0.4.4 release too? Should we merge that into the release branch?

User avatar
vincele
Space Dragon
Posts: 341
Joined: Sun Mar 23, 2014 6:10 pm

Re: Am I the only one who never won a game?

#23 Post by vincele »

Hello,
Dilvish wrote:Hmm, it sounds like Beginner mode would be better nerfed a bit more. I think it is reasonable to still squeeze in some change for 0.4.4 if we decide quickly. Though I have some tentative family plans to go travelling the next couple days so... no promises.
I'm not advocating for those ideas to be implemented for 0.4.4, but if it's possible, I think the game will be more enjoyable by beginners, so why not...
As in not build any attack ships at all, nor research planet defense tech? That would be possible, but really only suitable if Monster Density defaulted to None or Low, and could be reasonable done in time for 0.4.4, I think.
I think they can research planet defense, maybe starting at level 2, to keep the beginner mode absolutely easy. My idea about attack ships was more that they can build them but won't order them to attack human players, maybe level 0: not build any attack, level 1 build but defense only (against human players), level 2 retaliation with forgetting after a while, level 3 attack / no invasion of planets, then they can invade and kick your ass... I think the difficulty levels should be the way human players level up to the "fair" game where AI is treated the same and where the game become interesting, but we have to keep them until they reach that level, so we have to be careful with the gradual difficulty levels.
The whole retaliation-only thing is certainly not doable for me in time for 0.4.4, and I don't really like it anyway, sorry. Having levels 1 & 2 exclude invasions is very reasonable.
I remember playing some game whose AI did this and it was quite enjoyable at intermediary levels, to get some heat from the AI without ruining the pleasure to eventually win. Imagine in a game with a lot of AI against you, if the levels are rightly chosen you'll have a bunch of low aggression AIs, then a few with more teethes, one maybe one of the harder ones, so that you get to defeat some of the easy ones first and finish the game with the toughest, and you'll have enjoyed some wins even if you finally get wiped away. Loose with pleasure -> more need/excitement to play a new game and win completely this one... I really think only one (or maybe 2) AI level really needs to be fair, the others are only here to make the learning curve more interesting.
* in easy mode, human player should get better planets to start with
Hmm, this wouldn't currently be possible without a bit more work on the C++ side. The server does know what client type all the players have, and with a small bit of work this information could be made available to the python script, which could then give all human-client empires a starting bonus. But, this is seeming like rather a bit of complication and hassle that just shouldn't be needed, and doesn't so much fit with our approach-- simply nerfing the operation of the AI should accomplish our goals just fine.
I don't think it's unreasonable for the real easy levels. It's just another way to keep difficulty low, the game is not only about battles, even if you'll eventually always get there. the empire building side of things should also get its learning curve tuned so that we don't need to write a howto, the game should be self teaching, and I think that can help achieve this goal.
* disable xenoarcheology for AI, even without it we get wiped too easily, so imagine with...
That would be easy enough to do, and sounds reasonable for Beginner mode.
Xenoarchaeology is really a difficulty breaker for beginners, its such a surprise to see in sitrep that someone that's not you (i.e. an AI) got an ultimate tech, when you already struggle to fight/resist with/against normal tech. I generally stop playing and start again, because that the sign I'll get wiped pretty quick... It's only about xenoarch, not the other "normal" specials that are more limited in effect and don't make a big surprise to human players. We can keep those, maybe starting at level 2, to keep beginner more abolute easy...
All the patches I'll provide for freeorion will be released under the GPL v2 or later license.
Let's unleash the dyson forest powa!

User avatar
Vezzra
Release Manager, Design
Posts: 6095
Joined: Wed Nov 16, 2011 12:56 pm
Location: Sol III

Re: Am I the only one who never won a game?

#24 Post by Vezzra »

Dilvish wrote:I see now that Vezzra just recently did make a dynamic assessment of a starting min jump distance between homeworlds:

Code: Select all

min_jumps = max(int(float(len(systems)) / float(num_home_systems * 2)), 5)
, which he iterates down from if it doesn't work out.

I had also been planning on doing something [...] So Vezzra's approach seems to work dandy but I just had to mention my idea
Feel free to rip my solution out if you have a better one :D My approach can be best decribed as "I had a hammer, I had a problem I could solve with a hammer, so..." anyway. It might accomplish it's purpose, but it's not very effective in doing so, just try a 1000 systems 3 arm spiral galaxy with 3 players total. Then check the log and see how many iterations the homeworld selection function runs through before it succeeds. And it is noticeably time-consuming, on my system the process takes ~half a minute with a 500 systems map IIRC.

If your math provides a better initial estimation of the minimum jump distance, we should implement it (maybe not for 0.4.4, but afterwards). Just make sure it not only works on elliptical maps, but also on completely different shapes like spiral galaxies.

User avatar
vincele
Space Dragon
Posts: 341
Joined: Sun Mar 23, 2014 6:10 pm

Re: Am I the only one who never won a game?

#25 Post by vincele »

Hello,

thanks to dilvish-fo and that patch, we, the LOO (loosers of Orion) folks, should not get wiped from the universe so quickly anymore !

Very good

LOO, please test the next test builds, and report your victories.
All the patches I'll provide for freeorion will be released under the GPL v2 or later license.
Let's unleash the dyson forest powa!

User avatar
winter leaf
Space Floater
Posts: 41
Joined: Tue Apr 15, 2014 4:28 pm

Re: Am I the only one who never won a game?

#26 Post by winter leaf »

Okay!
This seems cool,and given that I am also a LOO then I might as well.I haven't played FO in a while,so my latest game should be rather interesting.

So far in the game,I made some really cool ships (not really though,compared to what everyone else does...),Organic hull,Lazer 3,2 zortrium armours.cool?i have a reasonable amount of ships and a kraken nest.so far have been battling the AI and doing ok, stole one of their planets (hehe) and successfully kept it.they have started to surround me,but no problems so far.also have lots of defence techs.I have defended my kraken nest from an attack from a griffin,of which my two larval kraken a and defences blew up.AI is spreading though,getting worried...but its looking positive so far.
|I enclose this message in a text box|
Also( YES,not in a text box ) I release graphics,code,and other things under the licences used by this forum. :)

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

Re: Am I the only one who never won a game?

#27 Post by Dilvish »

winter leaf wrote:I haven't played FO in a while,so my latest game should be rather interesting.
Good Luck Winterleaf! :D
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
winter leaf
Space Floater
Posts: 41
Joined: Tue Apr 15, 2014 4:28 pm

Re: Am I the only one who never won a game?

#28 Post by winter leaf »

Thanks,something hilarious just happened...
Ok,so on turn 200,i got a message in the sit rep:WARNING: 5 Black Krackens have been spotted around the experimentors outpost!

...i had no idea that the experimentors were in my game.A few turns later,and i got attacked by something on my planet that green couldn't take that I stole.Underneath the combat message,it said a Black kraken was damaged.great,I dont have the tech to see it on the map.Next turn,green invades...or,they try.they send a reasonable fleet my ships would have had trouble with,along with troop ships.they suddenly dissapeared.The black kracken is defending my planet! :mrgreen:
|I enclose this message in a text box|
Also( YES,not in a text box ) I release graphics,code,and other things under the licences used by this forum. :)

User avatar
Vezzra
Release Manager, Design
Posts: 6095
Joined: Wed Nov 16, 2011 12:56 pm
Location: Sol III

Re: Am I the only one who never won a game?

#29 Post by Vezzra »

winter leaf wrote:The black kracken is defending my planet! :mrgreen:
I think it's just defending the yummy candy it wants to have all to itself... :mrgreen:

User avatar
winter leaf
Space Floater
Posts: 41
Joined: Tue Apr 15, 2014 4:28 pm

Re: Am I the only one who never won a game?

#30 Post by winter leaf »

Lol,the George on that planet are already dead from concentration camps(I'm playing Eaxaw) but if it wants to stay,its fine by me :D
|I enclose this message in a text box|
Also( YES,not in a text box ) I release graphics,code,and other things under the licences used by this forum. :)

Post Reply