Skip to Content
 

Solitaire "urgency" track to make enemy AI launch increasingly desperate missions?

3 replies [Last post]
Broadsword56
Offline
Joined: 02/05/2015

I'm making a solitaire card-assisted naval age of sail game (historical).
This mechanism is needed to generate the enemy AI missions for each 2-week turn period.

Sorry for the long post, but here's the wider context that I think you need to know:

Historical situation
There were 2 sides. Each side vied for naval supremacy to control of a body of water. But the balance of power was so fragile that the loss of even one major ship on either side could instantly lose the entire campaign.
This posed a dilemma for the opposing admirals: Neither one could afford to sit in the safety of the homeport and let the enemy control the map. But contesting control meant seeking a decisive battle -- and risking everything in a single afternoon.
If one side gained a strategic advantage, then the other side usually avoided battle and got blockaded in their homeport until they could build a new ship and restore the balance. So it was also very difficult to bring a reluctant enemy to battle. The campaign was marked by cat-and-mouse actions, where one side would try to draw the other out into a vulnerable situation, catch a smaller part of the enemy force in an inferior position, or tit-for-tat raids on each other's homeport/shipyard while the enemy fleet was discovered to be off at sea.

Available Missions:
Each turn, the human player assigns each friendly ship an area of the map, and assigns each ship to any of 6 possible missions: 1. Refit in port, 2. Training cruise in the homeport zone, 3. Escort supply convoy, 4. Intercept (available only in the homeport zone, but this also gives a force the ability to engage enemy in an adjacent zone), 5. Patrol (engage the enemy in any zone), and 6. Support a land attack.

The game uses chit draws to secretly assign the enemy AI ships their areas and missions. But an entirely random mission draw won't work. I want the AI mission selections to be somewhat intelligent, and to respond to the evolving game situation.

I'm thinking of having some kind of "Urgency Track" mechanism, which the AI selects more aggressive/risky missions or less aggressive/risky missions based on the points representing the level of urgency in the AI strategic situation.

I want something that can force the AI to risk decisive battle. Since neither side would intentionally seek battle unless they have superiority, battles should result either when both sides are evenly matched, by accident, by miscalculation, or absolute desperation on the part of the inferior side due to some massive external pressure. (Example: In 1813 on Lake Erie, the British knew they had a disadvantage in firepower, but sought battle anyway because time had run out and they had no choice -- their supplies were exhausted and their Native American allies were on the verge of revolt)

Any ideas? Do you know of existing games that use tracks that influence (and get influenced by) piece behavior, changing board situations, and card-drawn events?

Orangebeard
Offline
Joined: 10/13/2011
Track or Pendulum?

I think the track idea would work, but it sounds like the track needs to allow for movement both ways; possibly starting in the middle and moving one direction for "strategy" and moving the opposite direction for "urgency". Perhaps the ends of the tracks would be 100% training on the strategy side and 100% all out attack on the urgency side.

Based on your description, it might be possible to manage the available missions by simply removing the possible strategy missions during urgent times and adding them back if the track ever returns to the middle.

Good luck with your design!

Dagar
Offline
Joined: 01/23/2015
Having a somewhat fitting

Having a somewhat fitting behaviour of the opponent in a board game is pretty difficult. In your scenario there are so many factors that should be taken into account, intelligent behaviour could not possibly be reflected.

In general, the more intelligent you want the board (in solitaire you play essentially against the board) to behave the more material and rules you will need and the more time and effort the player will have to put into board setup. So your balancing act as a designer is a battle you very likely will not be able to win.

That said I would try the idea of the escalation track. It should depend on number of ships, maybe just on the number of ships currently at sea. Maybe it should depend on the outcome of the last round instead, you will know this best. If you use more than one factor for determining the position on this track, keep it as simple as possible: add two integral numbers or use 'OR' or 'AND'.

The position on the track should determine how many cards from which behavioral deck you get. If every ship has one action, you should of course draw one card for each ship and assign them accordingly somehow. There could be a defensive deck, a balanced deck and an aggressive deck. Or you could divert it into production, training and maneuver. Or whatever else seems fit. Just don't use overly many decks as the variety of each deck will decrease with increasing deck number. Also, make each deck tendential; a card from defensive pile may also represent a more balanced or, very rarely, aggressive move. The production deck may also contain some maneuver to somehow supplement production. Just be sure that the decks are diverse, but in general still get their tendency across.

Broadsword56
Offline
Joined: 02/05/2015
Great replies -

Great replies - thanks!
Actually, I had a breakthrough in the meantime. I came upon John Butterfield's outstanding and much-loved solitaire game RAF (Victory Games, 1986) and the revamped version he did for Decision Games in 2009:
http://boardgamegeek.com/boardgame/40209/raf-battle-britain-1940
The "bones" of certain RAF mechanics can be adapted and modified to make a strategically intelligent and adaptive AI that will generate the AI ship missions and locations I need, much as they generated the raids heading for England in the other game.
And my version can be simpler than the one in RAF, because the entire British side had at most 6 ships and there's no need for all the early-warning, radar, replacements, and force-size mechanics that were so important to the 1940 Battle of Britain.
Once I finish writing the rules and have some test materials made, maybe I can post a design blog to illustrate how it would work.
In essence, a single control card drawn each turn sets the AI's low, medium, and high target priorities for the turn. Then a deck of Target cards sets the specific target for each AI mission and a "level of effort" number that cross-references with the current priority level of that target to determine the size of the force (in my case, the number of AI ship chits to draw from the deployable ships cup). The same Target cards also contain a possible en route event for the force, and a possible target event for the force, if it reaches its destination.
The system is able to shift AI strategy by changing its target priorities at specific times in the campaign, and with events that can be triggered by a certain Victory Points level on a track (- VP for AI British, +VP for human Americans). For example, a very low VP level might favor the AI British so much that it shifts the AI target priority from more cautious patrols to coastal raids on the US supply lines and depots.

Syndicate content


forum | by Dr. Radut