feature request: difficulty settings

[raymond]

example (for a current unlimited population, if there is a limit later (maybe 300?!) adjust the population cap may percental:

very easy: population cap for AI to 125, no upgrades of weapons and economy, no walls, no tower, no fortresses, max garrison send to enemy 10

easy: population cap for AI to 150, no upgrades for weapons and economy, max 1 towers of each civic center, no walls, no fortresses, max garrison send to enemy 20

medium: population cap for AI to 200, some upgrades for weapons and economy (not all!), max 2 towers of each civic center, wooden wall, max. 1 fortresses, max garrison send to enemy 50

hard: no population cap (or the max in current build), all upgrades, unlimited walls (inclusive city wall), towers and fortresses, max garrison send to enemy pop cap

Another possibility:

Just limit the amount of resources which workers can carry:

AI workers can carry:

very easy: 5

easy: 10

normal: 15

hard: 20 (same as you)

(maybe) very hard: 25

(also adjust fisher; useful if the AI can build ships in the near future)

Also posted here: http://trac.wildfire.../1586#comment:8

Edited September 10, 2012 by raymond

[eduh]

This is a nice and simple to implement difficulty...That said, we all would play on difficult xD

[alkazar-ipse]

honestly, I'm disapointed 'bout what you are posting: was expecting suggestions for future AIs for being better...

Harder bots are beeing worked on, and if you seriously think the demo bot is not easy enough, but should be even more limited to have only 5 ressource carring

then put player "unassigned" and you'll be served with a waste of time game where you can't be harmed.

On the other hand, making bot gather twice as fast as you, and having double of your HPs is not really making him play better, you can then give him a 300 pop army at start arround your TC with no units nor resources

and then you got ur unbeatable bot... Must say with sadness that this is what AoE 3 did: just make enemy gatheres faster, it's really tricking: the bot is not better, not more intelligent, just cheating...

Not meaning to hurt anyone, just being honest...

edit: to me ArtificialIntelligence means making a bot less predictable, less monotone... in otherwords more intelligent, humanlike.

and not letting him win on purpose by just giving him stronger units.

Edited September 11, 2012 by alkazar-ipse

[raymond]

This are only suggestions (the first word of this thread: "example") how to integrate a difficulty system into 0 A.D.

I think this is easy to implement.

If the AI are less predictable and more intelligent are much better...but very difficult to implement: this is the long term goal I think.

But if the AI can react "perfectly" in milliseconds to every enemy situation, there must be any adjustments that human players have the chance to win the game (but fixing the pathfinding problem is much more important).

My wish is only: don't forget to implement a difficulty system (in this or another way)

Edited September 11, 2012 by raymond

[alkazar-ipse]

haha, well it(difficulty system) is done until now with the different bots.

Dont know exactly but i think they will remove the useless bots for alpha 12, so if I understand well,

your request is they keep the demo and qbot. I respect your request even though I dont request it my self.

cheers.

edit: of course I agree with you that changing the capacities of bot units IS easier to implement, and therefore absolutly what they will do, if they fail making a perfect bot, which they will.

But then -> the bots capacities will be increased for MORE difficulty: easier bots already exist.

Edited September 11, 2012 by alkazar-ipse

[feneur]

My wish is only: don't forget to implement a difficulty system (in this or another way)

We won't We will try to focus on making the AIs harder/easier based on their tactics/range of actions rather than e.g. gather multipliers or something though It is hard to create a truly hard AI without it cheating in one way or another though, so at the very least the most difficult AI will most likely have some bonus to gather rates etc.

[raymond]

Demo and qbot? I use the new and I think the supported bot in the future: Aegis Bot

@feneur: thanks for your reply.

Edited September 11, 2012 by raymond

[alkazar-ipse]

@raymond: Yes, Aegis IS the most upto date bot. Demo was concieved as place holder (that will be removed once replaced) and jqbot is not beeing worked on anymore as I understood it in IRC chatroom.

Dont know exactly but i think they will remove the useless bots for alpha 12

I think Aegis is easy enough, and I fully agree: really hard bots will have to cheat with stronger units(gatherers ie)

I like Mythos' proposition for a conqueror, defender and rush bots.

[raymond]

Remove outdated, not supported any more and useless bots is the right way. So I fully agree.

But: min one bot is needed

Conqueror, defender and rush bots are also okay...but how will this be implemented in details?

Edited September 12, 2012 by raymond

[alkazar-ipse]

Im not the author of this idea so i dont know exactly, but as feneur said: "based on their tactics/range of actions"

A conqueror will try to occupy the map and attack you with bigger armies.

A defender will occupy territory more slowly but wall a lot to secure the space he holds

A rusher will not wait for a big army to attack you but just rush you as soon as he thinks he can kill all your units, later in the game as soon as he thinks he can destroy your buildings.

A big challenge is to make bots think long terms and not just get stupid or -even worse- idle after the map is empty of resources.

edit: I think markets are a good solution for longer games; they should get limited numb of markets and limited numb of traders per market, BUTmuch more efficient (right now from one end of the map to the other (on mapsize small(2players) ) a trader makes about 20-30 resource units)

Edited September 12, 2012 by alkazar-ipse

[feneur]

edit: I think markets are a good solution for longer games; they should get limited numb of markets and limited numb of traders per market, BUTmuch more efficient (right now from one end of the map to the other (on mapsize small(2players) ) a trader makes about 20-30 resource units)

While this is not really having anything to do with this topic I'll just briefly mention that I disagree. To me it's better to have the limit being "natural". I.e. they don't gather as much resources in the same amount of time as e.g. farming or fishing does (which they shouldn't as the other forms of resource gathering should be faster and trade only something for the later game), but on the other hand you can build a large number of them. That in turn is limited by how much population space you have available for them (as is of course the total number of gatherers in general, but on the other hand you have citizen soldiers who can both gather and fight, something the traders can't). Also, since the traders need to cover a significant distance to get a significant amount of resources there is also a greater risk involved. But at the same time, unlike the other ways of gathering resources, trade "doesn't run out", which in my opinion is an argument not to increase the resources they gather. The exact numbers can of course be tweaked, but I don't think they are very far off in more general terms. Especially since you do gain more if you trade with an ally + there needs to be room for benefits from trade techs, so to me the number seems fine.

[alkazar-ipse]

feneur, I FULLY agree with you

hunting, mining, farming, anything should be faster then trading.

My point was : once the map is empty of resources, the team with more territory (and chinese silk road WALL for protecting the caravanes from one side of the map to the other ^^) gets resources with it.

I was only admiring this aspect of 0ad. Its a genius solution for motivating to REALLY control the map (for trading) and not only to empty it of resources.

Was suggesting the capacity gets increased (i.e with a time factor, like trading effect doubles from 0 to 30 min and then remains constant...)

On the other hand You just informed me about the pop that caravanes take.

Was assuming it takes none, so that limitation problem is fixed too: each player must make tradeoff between army size and caravanes in longer games.

Edited September 12, 2012 by alkazar-ipse

[feneur]

feneur, I FULLY agree with you

Sweet

Was suggesting the capacity gets increased (i.e with a time factor, like trading effect doubles from 0 to 30 min and then remains constant...)

Could be one way to do it. Personally I would favor techs that increase the amount of resources gained as it's clearer what's going on then. And if the cost is high enough (and if the big improvement is only available in the City Phase, there could definitely be a tech in the Town Phase that has a smaller increase as well) it still has that time effect to it.

[Sighvatr]

I don't really think there should be a hard mode, or an easy mode to play. I think it would make a bot more challenging, and fun, if they switch up their strategies or tactics.

One bot can play defensively by building more walls than a usual bot. Such as constructing a lot of towers, and defensive structures to garrison troops in. An attack bot that plays very aggresively, will send in war parties every chance he gets while constructing a larger army. Another bot can become an expansionist, and attempt to expand his borders everywhere while maintaining his defensive force to protect the borders.

Other than different playing styles, you can create AI to focus on building up a "pure" force of certain units, and upgrade everything for that specific unit/s, or create a variety of different troops to help counter attack certain units a player might use.

Edited September 12, 2012 by Sighvatr

[alkazar-ipse]

good ideas there Sghvatr to make bots still variing witin their mode.

So you dont just know that a rusher goes for a specific unit ie, that he can randomly chose which units he upgrades.

So basicly an AI has a general behavior but not a predictable one. Rushing can still vary within a few minutes, so you cant calculate at what precise time ur units gather at a specific place.

edit: yes feneur, nice idea with making it by upgrade at city phase, enough expensif that one thinks twice about it before actually investing resources on THAT).

Edited September 12, 2012 by alkazar-ipse