Imagine an AI War Board game

[KDR_11k]

Having gone to the Spiel 2011 today I can't help but wonder what an AI War Board game would be like. Since I'm typing this on my phone I can't really do much brainstorming in this post until tomorrow...

[Coppermantis]

I imagine that if that ever existed, it would be like a mix between Risk and D&D. The territories (planets) are laid out, and a 'dungeon master' of sorts decides what's on it. Each player gets a reinforce/attack turn, and then the AI Player gets one. As it loses targets that ordinarily increase AIP, it gets a boost to how many units it can deploy per turn, etc.

[Orelius]

As a board game, the most important thing comes to mind:  Who plays as the AI?  How should the AI act?

If the AI was a person, things could get tricky, because obviously, the AI can easily crush the human forces in the earlier stages of the game.  However, forcing the AI player to abide by a certain set of rules as to when he can or can not attack may be frustrating.

Another problem is fog of war and scouting - it would be hard to regulate scouting and fog of war in general unless you put some sort of partition between the players, and even that is difficult to manage when territories are scattered.  This may be solved by simply using cards that can be flipped over, I suppose, but that would make it hard to track AI fleet movements for the group in general (writing them down on a pad is too frustating).

Ironically, I think that the best way to make an AI war board game would be to make it a sort of turn-based strategy game for computers; as the game is far to complex for things to be easily set up in a board game.

[KDR_11k]

I'm tending towards having the AI "played" by the game mechanics. It couldn't be as smart as the real deal but the important thing is preserving the feel of the game, not the details. Since I'm kinda brainstorming the following I'll make up piles and cards and stuff as they come up, not describe them all upfront.

The board: I'd say it'd be a bunch of large tiles that get placed randomly and then rotated by dice rolls to build the galaxy map. In-system battlefields should probably not be in play unless we're going to heavily limit the number of systems in play or go for a gigantic map (which would have problems with the real world design). Unexplored systems wouldn't have any attributes to them, the special stuff would be drawn from a deck once the system gets scouted (lazy world generation ). AI actions outside the scouted area would be completely ignored.

Startup:
After the board is set up somehow, eight planets on one side of the map are chosen as possible starting positions and have a bonus unit card placed on them, face up. Players pick their starting positions and place a home command marker on those. They take the bonus unit cards from their home planet, the rest of the cards are shuffled back into the bonus unit pile. All unused planets receive a "not scouted" marker (these might have some details for the generation of the system on their back, might have letters that determine where to place them on the board like in Starfarers Of Catan) and all players get a set of fighter, bomber and frigate cards.

Various mechanics:
Fleets: A fleet is composed of a basic fleet pawn with several slots in it, these slots can be filled with ship types (I'm thinking big plastic pins like those used in Battleships) by paying resources when you have the necessary type card and of course the ship cap is still in place so you only get a certain number of each unit.

Waves, CPAs, etc: Every turn or so a card is drawn from the AI action pile, most of these will probably just be "reinforce something" or "minor border aggression" but some would be waves or CPAs. Such a card would be placed aside and start acting the next turn: For waves three cards from the AI unit type pile are drawn and a fleet with those units is placed according to some rules (e.g. closest to a player homeworld). A CPA would generate random fleets on scouted AI planets for the next few turns. Such AI fleets would follow a set of movement rules, mainly going towards your main base with high priority targets (factories, high value command posts, etc) allowing a detour of a fixed number of hops.

Turrets: A planet has some base turret presence that will wear any enemy fleet on it down slowly. Planets can be upgraded to have serious turret presence that poses much more of a threat (and can be destroyed individually). Fortresses are a separate matter.

Knowledge: Is measured in knowledge tokens roughly equivalent to 3k knowledge each. Players start with 3 and get one for every system they take from the AI. Mark upgrades are 1 token for the first, 2 tokens for the second upgrade. Turrets are passive upgrades, when a system is marked as having a serious turret presence then all researched turrets count. Upgraded ships get additional unit tokens to reflect the rising ship cap.

Special attributes: Of course ships and turrets can have special rules (e.g. "movement through this planet takes two steps instead of one" or "when an enemy fleet is destroyed take one token of the last destroyed unit type and add it to your own fleet"). These probably upgrade as the ships and turrets get upgraded.

Resources: Probably just one type, not two. You get three tokens for your homeworld and one for each command station you have (2 for econ commands).

Command stations: Mark levels should probably be dropped for these to keep it manageable. When a player colonizes a planet they pick between econ, mil and log. Econ gives 1 extra resource token from the system per turn, mil greatly increases the passive resistance in the system and log lets friendly units take only half a move point when moving through the planet.

Special AI structures: When an AI planet is seeded it receives a "com + gate" marker. This can be flipped to show just a com station if the planet is gate raided. When a not-scouted marker is lifted it says on the back what kind of mark level and how many special structures the planet will have. Special structures are drawn from a pile and placed on the planet, these have effects described on their cards. Advanced factories and research stations are marked separately on the scout token though maybe a few ARSes could be in the AI structure pile too for some random flavour.

Combat: That's the aspect I'm the least sure on so far. I'm thinking combat should consist of smaller attack rounds that can happen multiple times per round when a direct battle is waged and once if a fleet is just passing through and not engaged directly but still encountering smaller skirmishes. Unit types should be destroyed one by one on a fleet instead of the whole fleet being worn down equally (unless an attritioner is present ). I'm shying away from dice rolls so far since AIW combat doesn't have randomness like that.

Mega units: Of course massive units like golems or fortresses can be in play, these take a ton of damage and aren't part of a fleet. I'm not sure what to do with regular capital ships.

AI Progress: I'd say a token of AIP is worth around 7 points. Killing a warp gate is 1 token, killing the com station another two. These would determine the strength of AI fleets. Also some of the AI action cards (waves, CPAs) could require a minimum value of AIP before they have an effect (not all equally, e.g. two "wave at any AIP" cards, one "wave at medium AIP" card and one "wave at high AIP" card so the number of wave cards in the deck effectively goes up as the AIP rises). Probably also continuous AIP increases of some kind.

Minor AI unit tokens: Only large AI attacks should be represented by fleets, un-free ships and border aggression should be handled separately. For these I'd say there are special tokens that are just placed as reinforce steps. Such a token is considered a homogenous mixture of AI units, a few of this type, a few of that type, you know. Tokens can be converted to fleets if the ships are freed or a CPA happens/an alert is triggered. Individual tokens would get destroyed by the basic turret presence on a planet, bigger numbers might be enough to destroy a com station. These probably wouldn't get the advantages of the bonus ship types the AI has. They'll still wear down fleets and get slowly killed one by one when sharing a planet with enemies.

[Palarran]

First thing, it would definitely be a wargame rather than a euro, and wargaming is a small subset of the boardgaming hobby.  Solitaire would probably work best with the human(s) working against the game system.

A recent example of a space themed wargame is Space Empires: 4x by GMT games.  AI war would have to be drastically simplified and revamped to enable it to be played manually on the tabletop.  It is complicated, with a myriad of interactions and calculations being handled by the computer.  If you put it all out in the open, it has to be handled by a person manually.  Most of the special exceptions and interactions would have to be reduced to a maximum of a dozen or two interacting variables.

[KDR_11k]

I believe most of the calculations AI War is doing are for the AI and range detection and such things, stuff that a human brain would do instinctively. Well, the AI would need to be simplified a lot and I'd say there's no need to run at a detail level where even all ship types and marks are fully handled. I'm not sure it would really count as a wargame but that may be up to interpretation. The way bonus ships and such things vary is certainly highly reminiscent of the way Agricola dishes out the personal cards with different deck complexities and such. Since most special rules in AIW are tied to specific unit types that could just be written on the unit cards and any types that aren't in play (especially when playing with the simple deck) wouldn't complicate the rules.

If we're talking about mass appeal sure an AIW game would suffer but merely including combat is already a major issue in that regard. At least it's not player vs player and the time-based AIP increases would keep the play time limited so there's no endless porcing up. Of course the game would still take some time since the whole macro-level decision process in AIW is generally fairly long.

[keith.lamothe]

It might actually work well as a co-op board game.  I figure the combat could be abstracted to some kind of 3x3 or 5x5 or something grid or possibly no spatial relationships at all (within a single planet).

But making it a paying commercial proposition would be quite a challenge   Production costs and high list prices, etc.  It would presumably take a fairly big publisher/distributor with a fair bit of upfront money.

[Palarran]

If you guys think you can make a go of it, contact GMT Games.  I think they'd probably be one of the best companies to work with, due to the nature of their game catalog: many hardcore wargames and a few more abstract games.  Almost all of their games are done by outside designers.

[blastpop]

I would think instead of GMT, while I love their games, might not be the best choice. Other choices include Fantasy Flight, Z-Man and Rio Grande among a partial list of front runners.

[Echo35]

As a board game, the most important thing comes to mind:  Who plays as the AI?  How should the AI act?

If the AI was a person, things could get tricky, because obviously, the AI can easily crush the human forces in the earlier stages of the game.  However, forcing the AI player to abide by a certain set of rules as to when he can or can not attack may be frustrating.

Another problem is fog of war and scouting - it would be hard to regulate scouting and fog of war in general unless you put some sort of partition between the players, and even that is difficult to manage when territories are scattered.  This may be solved by simply using cards that can be flipped over, I suppose, but that would make it hard to track AI fleet movements for the group in general (writing them down on a pad is too frustating).

Ironically, I think that the best way to make an AI war board game would be to make it a sort of turn-based strategy game for computers; as the game is far to complex for things to be easily set up in a board game.

The board itself Look at games like Descent, Battlestar Galactica, Arkham Horror, Shadows over Camelot, Pandemic, etc. The board itself plays the "GM". Something like that would be apt for AI War.