New Contest - game like Risk

[mtobiass]

I really like this idea. Some ideas to add to simple "risk" are:

Line of supply.
For example, every player has a home base. Units that can not trace a line to the home base through own regions are out of supply. Units out of supply are more likely to loose in battle.

Simple diplomacy.
Maybe just "non agression pacts". Makes it more likely that smaller player will cooperate to defeat larger players.

Region improvements.
For example, instead of placing all armies, some armies might be changed to a "factory" that is placed in a region. The factory will then produce armies, making that region more valuable.

[GreenTea]

Line of supply.

This improvement add some realism, but i have doubts, that it can very simplify game.
In Risk one of the main goal is to not let enemy army penetrate into my territories. If one player was able to do this, than, even with much smaller income (but with safer territories), it can win. With "Line of supply" its imposiible to do such guerrilla war, because attacking army at once lose all reinforcements after entering eneimy territories.

Simple diplomacy.

Not bad idea.

Region improvements.

Also add some realism, but I consider its to simple way to increse income. Just alwais build factories everywere - non stop.. This will be another game. not Risk.

[RebelXT]

Excellent idea! I've just played a few games on warlight.net, and it's a fun game.

[Corniel]

Can't agree with you. On warlight.net (..)

So it is not so simple - with random or without random - skill is skill.

I have to admit I have no experience with 1 to 1. My experience is based on a total of 60 games I played (at the board) during my studies. Most of them were played with 4 players, some with 3, some with 5 ore 6. And our group stopped playing just because we got bored throwing dices all evening and founding out we ended up the same as we started. Only when a inexperienced (or stupid) player entered the group a decision was made. Mostly because that player went suicidal...

Most of the players were experienced board game players (including a Dutch junior champion chess). Risk seems to be to balanced to force a victory to me.

So I was addicted to playing Risk at some point.

[fizzer]

Hey guys,

I'm the creator of WarLight.net, and I love the idea of doing a Risk-like competition for the next AI challenge.

If you do decide to use WarLight-like rules, I can provide the code behind the game engine for use. I've already made a fully-functional demo of this you can see here: http://blog.warlight.net/index.php/2011 ... challenge/

In WarLight rules, each army can move to one adjacent node each turn. This is different from Risk, where armies can cross the entire map in a single turn. I think moving one node at a time provides a much better platform for an AI challenge as it emphasizes location. That is, which regions of the map you have access to is often more important than the number of armies you have.

I'm also a big fan of the 1v1 format. Having more than two players increases the complexity of the AI significantly, as these games are more decided by who attacks who than army-moving strategy.

If I can do anything to help out, I can be reached at fizzer@warlight.net. Thanks!

[GreenTea]

It be very interesting to see how best bots in risk play with top warlight.net players whether it is possible to make bot stronger than human, or not..

[Corniel]

I've already made a fully-functional demo of this (..)

I just did a quick read. But can I conclude that if I want to (yes GreenTea, maybe I will have try ) I can start already?

[lama3oid]

I think, this game could fit only if we remove random element from it, so it definetly should work without dices.
Whichever game would be chosen, the played, that's doing better moves should always win in duel games. Otherwise, the strongest bots will fight with the same skill and the luckiest player will win.
Also, I think, in the chosen game (whatever it will be) players should move simultaneously.

Btw, I don't like the idea. I played this game against AI few years ago and against real people in board game. It's interesting game, but not enough hard for such competition. Much easier than ants. Additionally, comparing to ants, it would be much easier to understand/copy someone's tactics just viewing replays.

[GreenTea]

I think, this game could fit only if we remove random element from it, so it definetly should work without dices.

Can add about randomness, that it can make bot programming more interesting. This is element of fuzzy logic. For example if I attack 4 enemy armies with 7 my armies I (for example) in 90% cases capture territory but also in calculations I should consider case when 10% - i not capture it. And in all game prediction, simulations everywere consider all possible cases of battle result.

Also, I think, in the chosen game (whatever it will be) players should move simultaneously.

Wrote about this earlier. With simultaneous move game will be to simple. And count of tactics will be much, and much less.

Btw, I don't like the idea. I played this game against AI few years ago and against real people in board game. It's interesting game, but not enough hard for such competition. Much easier than ants. Additionally, comparing to ants, it would be much easier to understand/copy someone's tactics just viewing replays.

Play on warlight.net. If you think it is very easy I can play standard 1x1 with you. May be you learn me how to play

[lama3oid]

in 90% cases capture territory but also in calculations I should consider case when 10% - i not capture it

90/10 is fine, but in original game you have bigger than 10% chance to win with 4 against 7 I think. Maybe I'm wrong.

Wrote about this earlier. With simultaneous move game will be to simple. And count of tactics will be much, and much less.

Again, it's just imho. I think, the only way to make sure is to try both variants. But try to imagine turn-based ants... The combat systems would be very easy.
With original rules of Risk there's no point in simultaneous movement. But I think, the rules could be changed somehow to make interesting simultaneous game.

Play on warlight.net. If you think it is very easy I can play standard 1x1 with you. May be you learn me how to play

I don't say it's easy, but it's not enough hard and my personal skill in this game doesn't matter. I just think, that finding the good strategy here would take much less time than in ants. At least with original rules.