World Conquest: Limiting Strat versus AI

Share and discuss strategies for playing the game, and get help and tips from other players.

Moderators: Forum Moderators, Developers

User avatar
Des
Posts: 116
Joined: November 7th, 2007, 7:58 am
Contact:

World Conquest: Limiting Strat versus AI

Post by Des »

Let me start off by saying that I consider World Conquest one of the best add-ons currently on the server. It’s got plenty of replay value, largely due to its added randomness (maps, heroic units, treasure), it’s playable online as a campaign with 2 or 3 players, and it's just plain fun.

Recently, I’ve been playing Natasiel’s WC NM extension, which fixes some issues and makes some neat changes. Read about it on its own thread here.

I enjoy playing online, but I also like playing two sides by myself for a bit of variety. So when I’m talking about the strategies below, keep in mind that I haven’t just played my side online with that strat, but had the advantage of keeping two sides in sync with that strategy.

One of the best features of World Conquest is that its large maps, especially in later scenarios, require a much grander strategy to play. Not to mention, each map is randomly drawn, forcing you draw up mental plans on the best route to recruit troops to or send out scouts.

That said, my general strategy for WC has been refined. Initially, I would try to grab as many villages and good terrain as possible. But even in the first scenario, this can prove to be a costly mistake, as the AI will simply blind rush from its keep and ruthlessly rid your units one-by-one at any cost.

Starting this way in the first two scenarios proved to be unsustainable for the long-term. This was especially true when level 2’s were put on the front at the outset.

After learning this, I adopted a strategy of defense first, letting the AI come to me while sending out a small number of scouts to grab the middle villages. This resulted in those scouts getting killed off and my main army being preoccupied with their main army so long that I would lose too much money for the next scenario. This was particularly true in scenario 4, where both of my sides were sandwiched by two of theirs (totaling my 2 sides versus their 4).

What got me past scenario 4 was a strategy that involved sending many scouts to capture villages, killing their scout units and in turn drawing some of their main force away from the choke points where my army met them. The scouts I sent were mostly level 1’s, but did include some level 2’s to grab and hold the many villages in the middle of the map. Meanwhile, my army would hold off on the terrain very close to my keep, and was comprised mostly of new level 1 recruits to act as cannon fodder before the level 2’s and 3’s are recruited (to prevent critical losses). In time, this allowed a sustainable economy for reinforcement on the front and leftover gold for next scenario.

While defeating the AI this way is satisfying, I can’t help but think of how much more challenging it would be after the AI is reworked.

The nature of the AI behavior (massing a quick, head-on assault to the front, regardless of terrain or TOD) seems to limit the scope of strategic responses to “hold tight with cannon fodder, while scouting extensively.” Other than the former strategy, is there an alternative response that addresses the problem the AI poses?
Redrock Gulch (Winter 2009 Map Contest Submission)

To rely on rustics and not prepare is the greatest of crimes; to be prepared beforehand for any contingency is the greatest of Virtues. - Sun Tzu, The Art of War

User avatar
Des
Posts: 116
Joined: November 7th, 2007, 7:58 am
Contact:

Re: World Conquest: Limiting Strat versus AI

Post by Des »

For those who want to see what I'm referring to: Here is a replay of me playing two sides on scenario 6 as an example. It gets really exciting towards the end, as I had to rush killing the leaders with all the gold I acquired late game. I ended up winning with only 2 turns to spare out of 44 turns.

The other replay is the first scenario with another player (Psychic), where we employed the strategy above to take very minimal losses and still win rather quickly.
Attachments
Best_WC_game_with_psychic_map_1.gz
Scenario 1 of World Conquest with Psychic.
(56.35 KiB) Downloaded 139 times
MP_campaign_-_WC_NM_6_replay.gz
Scenario 6 of World Conquest - Me playing two sides.
(250.58 KiB) Downloaded 167 times
Redrock Gulch (Winter 2009 Map Contest Submission)

To rely on rustics and not prepare is the greatest of crimes; to be prepared beforehand for any contingency is the greatest of Virtues. - Sun Tzu, The Art of War

User avatar
Des
Posts: 116
Joined: November 7th, 2007, 7:58 am
Contact:

Re: World Conquest: Limiting Strat versus AI

Post by Des »

Given that nobody has answered this thread in the full week that it's been up, I'll take that to mean there is only one viable strategy to playing World Conquest.

That is: to hold a defensive position near your starting keep as the initial onslaught of reckless AI troops foolishly rush forward--all while scouting extensively in the other direction. Once you overwhelm the AI, cleanup for level advancements and go for the leader kill.
Redrock Gulch (Winter 2009 Map Contest Submission)

To rely on rustics and not prepare is the greatest of crimes; to be prepared beforehand for any contingency is the greatest of Virtues. - Sun Tzu, The Art of War

User avatar
Sapient
Inactive Developer
Posts: 4453
Joined: November 26th, 2005, 7:41 am
Contact:

Re: World Conquest: Limiting Strat versus AI

Post by Sapient »

If you want to play a two-front war, I suggest giving a teleportation ring to your most powerful heroes. It's a risky strategy, though, because sometimes they rack up a lot of counterattack damage faster than you can heal them.
http://www.wesnoth.org/wiki/User:Sapient... "Looks like your skills saved us again. Uh, well at least, they saved Soarin's apple pie."

User avatar
Des
Posts: 116
Joined: November 7th, 2007, 7:58 am
Contact:

Re: World Conquest: Limiting Strat versus AI

Post by Des »

Sapient wrote:If you want to play a two-front war, I suggest giving a teleportation ring to your most powerful heroes. It's a risky strategy, though, because sometimes they rack up a lot of counterattack damage faster than you can heal them.
I'm assuming by "two-front war" you're referring to my scenario 6 replay. If that's the case, it should be noted that there really isn't a choice to stage a two-front fight as you imply. The scenario is set up so that each player has at least two AI sides coming at it.

And while I appreciate the reply, I'm not sure it really answers my question. Giving certain leveled units a nifty way to teleport is more of a tactical trick than a strategy. At best, it can be used to provide a front with more help from a higher level unit (given that it's not about to die off), or to retreat that unit when its health is low.

But as I've mentioned in this thread, using level 1's as mere meat shields to stop the coming onslaught is probably just as crucial to your success than using level-advanced units.
Redrock Gulch (Winter 2009 Map Contest Submission)

To rely on rustics and not prepare is the greatest of crimes; to be prepared beforehand for any contingency is the greatest of Virtues. - Sun Tzu, The Art of War

User avatar
Sapient
Inactive Developer
Posts: 4453
Joined: November 26th, 2005, 7:41 am
Contact:

Re: World Conquest: Limiting Strat versus AI

Post by Sapient »

Des wrote: I'm assuming by "two-front war" you're referring to my scenario 6 replay. If that's the case, it should be noted that there really isn't a choice to stage a two-front fight as you imply. The scenario is set up so that each player has at least two AI sides coming at it.
No, I am not referring to any of your replays. I am referring to some of my own experiences with using teleportation. If used properly and with a bit of luck, you can hold more several more outlying villages than would otherwise have been possible.

If that amounts to a "tactical trick" then I am sorry it doesn't meet your expectations.
http://www.wesnoth.org/wiki/User:Sapient... "Looks like your skills saved us again. Uh, well at least, they saved Soarin's apple pie."

User avatar
Des
Posts: 116
Joined: November 7th, 2007, 7:58 am
Contact:

Re: World Conquest: Limiting Strat versus AI

Post by Des »

I played around with teleportation, and I agree that it is useful. I found it easier to hold my outlying villages, as you mentioned.

However, I still consider this technique on a smaller scale than what my original question was asking. It fits the construct of the same strategy that I see as the only viable way to go about winning the scenario. If there was another strategy, I wanted to know about it. I feel that because the AI plays the way it does, it forces you to respond in only one way--and that is too bad because it would have been fun to try multiple ways of completing the campaign.

I guess we'll just have to wait for after 1.8 to get a smarter AI.
Redrock Gulch (Winter 2009 Map Contest Submission)

To rely on rustics and not prepare is the greatest of crimes; to be prepared beforehand for any contingency is the greatest of Virtues. - Sun Tzu, The Art of War

Mabuse
Posts: 2130
Joined: November 6th, 2007, 1:38 pm

Re: World Conquest: Limiting Strat versus AI

Post by Mabuse »

Des wrote: I guess we'll just have to wait for after 1.8 to get a smarter AI.
a smarter ai would just try to destroy your current startegy without leaving any other holes.
(that is use more units to conquer outlying villages while keeping same raw pressure on the main frontline to squish you)

you wouldnt get a bunch of new ways/new options how to win the game versus an enemy that has so much more gold than you if the ai would be smarter.

the only strategy is to let the "dumb but rich ai" bleed out while you try to get as many villages as you can get to have a higher income in the very end.

in fact (and i can only tell from the perpective as a scn designer) the amount of gold and stuff are always balanced toward the current capacities of the ai to pose a threat.

since the ai is quite dumb as hell and can only attack with max strengh it gets a good bunch of extra money to make up for that.

when the ai is smarter it would just pay more attention to outlying villges while still attacking you wit its overwhelming force. the onyl new option would be probably that you COULD try to make more fancy tactics at the (now slightly weaker) frontline to destroy it. also here with smarter implemented tactics the ai could make up for it anyways. (or since some of the units are anyway not needed since thy stuck, they coudl be used on outlying villages anyway)

the thing that i would like to know from you:

what kinds of strategy do you EXSPECT would suddenly be possible/playable when the ai is smarter. imo, it would be just more tricky to beat the ai with your CURRENT strategy
The best bet is your own, good Taste.

Euthanatos93
Posts: 83
Joined: March 30th, 2009, 2:17 am

Re: World Conquest: Limiting Strat versus AI

Post by Euthanatos93 »

He's right, a 'smarter' AI won't increase your options. However, modding the scenario will. It's on my todo list after I learn WML. I plan to mod a lot of my favorite scenarios for many of the reasons you list.

User avatar
Des
Posts: 116
Joined: November 7th, 2007, 7:58 am
Contact:

Re: World Conquest: Limiting Strat versus AI

Post by Des »

I understand that the current limited AI is aided by massive gold in order to make it more competitive. The problem with that fix is that it really doesn't matter how many troops the AI has--if they use them rashly, they will ultimately lose against the strategy I have presented (which nobody thus far has argued against it being the sole strategy to overcoming the AI).

Admittedly, a smarter AI won't present a dizzying array of strategies, but surely it will open it up to something more than "buy a ton of meat shields to soak up blind attacks, keep leveled units close to hold the major fronts, and scout the map's villages extensively to bleed AI recruiting possibilities."

For one thing, an AI that knows it shouldn't be attacking during a certain time of day will create strategical situations that are several times more dynamic than they are now. Likewise, injured units that retreat to heal up instead of vehemently trying to press on would make for more interesting battles, in a tactical sense, and more satisfying scenario victories, in a broader sense.

If that kind of game is hard to imagine, it's probably because the AI needs a lot of work, which almost goes without saying. Maybe the overarching goal to overcome the AI wouldn't change much, but at least it would be more dynamic and less simplistic. As it currently stands, a player need not rethink the original game plan. They can stubbornly execute it with the knowledge that the AI will surely fall into the trap plainly laid out for it.
Redrock Gulch (Winter 2009 Map Contest Submission)

To rely on rustics and not prepare is the greatest of crimes; to be prepared beforehand for any contingency is the greatest of Virtues. - Sun Tzu, The Art of War

Caphriel
Posts: 994
Joined: April 21st, 2008, 4:10 pm

Re: World Conquest: Limiting Strat versus AI

Post by Caphriel »

Des wrote:I understand that the current limited AI is aided by massive gold in order to make it more competitive. The problem with that fix is that it really doesn't matter how many troops the AI has--if they use them rashly, they will ultimately lose against the strategy I have presented (which nobody thus far has argued against it being the sole strategy to overcoming the AI).

Admittedly, a smarter AI won't present a dizzying array of strategies, but surely it will open it up to something more than "buy a ton of meat shields to soak up blind attacks, keep leveled units close to hold the major fronts, and scout the map's villages extensively to bleed AI recruiting possibilities."
That's pretty much the general strategy against the AI, whether in multiplayer campaigns, single player campaigns, or normal multiplayer games.

User avatar
Des
Posts: 116
Joined: November 7th, 2007, 7:58 am
Contact:

Re: World Conquest: Limiting Strat versus AI

Post by Des »

Caphriel wrote:That's pretty much the general strategy against the AI, whether in multiplayer campaigns, single player campaigns, or normal multiplayer games.
Really? Your strategy in normal multiplayer games is to bank on your opponent to play at a sub-par level by rushing into your defense, ignoring time of day and their own units' survival?

Right.
Redrock Gulch (Winter 2009 Map Contest Submission)

To rely on rustics and not prepare is the greatest of crimes; to be prepared beforehand for any contingency is the greatest of Virtues. - Sun Tzu, The Art of War

Caphriel
Posts: 994
Joined: April 21st, 2008, 4:10 pm

Re: World Conquest: Limiting Strat versus AI

Post by Caphriel »

Des wrote:
Caphriel wrote:That's pretty much the general strategy against the AI, whether in multiplayer campaigns, single player campaigns, or normal multiplayer games.
Really? Your strategy in normal multiplayer games is to bank on your opponent to play at a sub-par level by rushing into your defense, ignoring time of day and their own units' survival?

Right.
Added emphasis. You can play normal multiplayer against the AI. By "normal multiplayer," I was excluding custom scenarios and the like.

User avatar
Des
Posts: 116
Joined: November 7th, 2007, 7:58 am
Contact:

Re: World Conquest: Limiting Strat versus AI

Post by Des »

Even so, my point is simply that the current AI limits play styles by anticipating poor attacks. Currently, and especially in World Conquest (i.e. larger scale games), hard rushes by the AI require a defensive position for the most part. There is a general consensus that mustering a successful attack requires greater skill than playing defense (this is found throughout the forum).

If that's the case, then the AI is really only challenging one half of gameplay for the majority of crucial turns: the defense. This, in my opinion, is unacceptable.

If the AI would require you to coordinate attacks by retreating during an unfavorable time of day, the game would be much more dynamic. It would make for a much more satisfying experience.
Redrock Gulch (Winter 2009 Map Contest Submission)

To rely on rustics and not prepare is the greatest of crimes; to be prepared beforehand for any contingency is the greatest of Virtues. - Sun Tzu, The Art of War

Caphriel
Posts: 994
Joined: April 21st, 2008, 4:10 pm

Re: World Conquest: Limiting Strat versus AI

Post by Caphriel »

I quite agree with you; I was just pointing out that the problem is not limited to World Conquest at all.

Post Reply