Dominions 3 Announced

[ioticus]

http://www.shrapnelgames.com/News/pr_01.htm

A random map generator in nice (but not necessary). What I'd like to see is better AI. I find multi-player games take too long without forced hosting and it's too hard to keep up with games that force hosting. I want a better single-player experience and the best way (and the hardest from the developer's standpoint) is to improve spell casting AI, strategic AI, research AI, etc. Did I mention I want better AI?

[johan osterman]

I assume you want a better AI as well, but I couldn't see you mention it.

[Soapyfrog]

I'd like battles to be resolved in real time, and perhaps a slightly more complex scripting for ordering your troops.

For example, being able to script conditional orders: e.g. mages will only cast the big artillery if the enemy army is of X size ro something like that.

[johan osterman]

Battles won't be resolved in realtime, and frankly I don't see why that would be an improvement.

[liga]

johan osterman said:
Battles won't be resolved in realtime, and frankly I don't see why that would be an improvement.

great JO!!! I really like turn based games! there are plenty of RTS ... I hope dom3 will be like dom2 but with some improvement .. so no needing of a super-mega-big computer to run, no need to be a spider moving on a keyboard to play well or to be able to see all the hundreds of things happening in the same time on your monitor hoping to have a 25" monitor to be able to see everything!

the only sad thing I can see in the coming out of dom3 is that probably I'll have to stop playing dom2 ... and dom2 is a so great game that I don't know how to stop playing it!

good play
Liga

[Soapyfrog]

johan osterman said:
Battles won't be resolved in realtime, and frankly I don't see why that would be an improvement.

Ah, an anti-RTS snob I would suggest since you only have scripted control of the battle, real-time resolution would certainly be much better.

Let me give you some examples:

1) battles resolve faster becuase missile/spells go off as they occur (possibly simultaneously, and in conjunction with movement)
2) no more instantaneous flying to any part of the battelfield
3) no more "optimal placement" problem where you can rob charging units of the charge bonus by timing it so your units get the first attack.
4) less weirdnesses due to turn breaks and more precision... instead of quickness giving you two attacks on your turn, it would allow you to attack twice as quickly (for example), which I think you'll agree is a little more accurate.
5) it would LOOK better, although I agree this is below a tertiary consideration

Remember the absolutely classic game Centurion in the earlier days of PC gaming? Scripted real-time with some very limited direct control... it was beautiful.

Or the one really high point of Paradox's Legion: the scripted real-time battles. Albeit, could have been done better, but it was an excellent start.


If you really hate real time How about this instead: Break turns into impulses, ala star fleet battles. That would certainly be an improvement.

I might point out that the battle is already resolved in "real-time" for the viewer i.e. the battles run without breaks, you need to pause to gather information, so there would be no impact in terms of additional "business" for the viewer.

[RadiantFleet]

I would like to see some modularity for the AI in Dom3, specifically the main program calling an AI "module" or script that can be tinkered with by players. I think Space Empires IV is a good start, but really anything that would let players tinker with the decision trees by nation and "flavor". I think this would be a huge benefit for both players and illwinter, since the players would get better AI's and we'd help keep Kristofer from getting programmer burn out. I think this strategy has worked extremely well for maps, and would be a big benefit in the AI arena.

Also, regarding RTS. Please don't do that to Dominions. RTS games are so much about reflexes and so little about strategy and cunning.

PS. Another great example of using modules for AI control is the old tank game "omega".

[johan osterman]

Soapyfrog said:

Ah, an anti-RTS snob I would suggest since you only have scripted control of the battle, real-time resolution would certainly be much better.

Not at all, I like several RTS games. I just don't think dominions would benefit much from it. I also agree with your points, they just seems like minor concerns.

[NTJedi]

johan osterman said:
I assume you want a better AI as well, but I couldn't see you mention it.

GIFTS FROM HEAVEN this is important.

One of the greatest features lacking in almost every game is a challenging AI. There are several important decisions which could definitely improve the Artificial Intelligence. Hopefully the following listed below will be considered for improving the existing AI opponents:

1) Multiple AI personalities available which are randomly selected at the start of each game. This would keep human players surprised and uncertain what to expect from their neighbors.

2) Providing option to disable/enable starving status for AI units. The AI opponents are not aware of food supply levels within provinces and thus should not suffer the penalties of starving which include diseased units and low morale.

3) Allow scripting or some other tool where Users can develop and customize an artificial intelligence which can then be added to the list of random multiple personalities.

4) Don't have computer opponents send their pretender into the death matches. The battle afflictions, horror marks, and curses are not worth the risk.

5) Higher difficulty setting should have the computer opponents be more aggressive against human opponents as well. This makes sense since its the human player looking for a greater challenge when increasing the difficulty level.


I'm sure there are others but these will greatly improve the AI for the game. Double Thanks !

[Vicious Love]

Simple suggestion: More feedback regarding stealth, or at least more goodies in the manual. I honestly have no idea how effective a +30 stealth modifier is, how effective a patroller is, or how said effectivities counterbalance. If you were to ask me which was more likely to be caught:
A) 40 stealthy(+30) units vs 20 patrollers
B) 5 stealthy(+0) units vs 20 patrollers
C) 1 stealthy(+30) unit vs 50 patrollers
D) And so on, and so forth,

I wouldn't have the vaguest idea. I'm not even sure if 20 patrollers are always twice as effective as 10 patrollers, or 20 sneaking troops twice as effective as 40.
This should not be. I'd like to see stealthy armies and agents work as a result of a well-conceived battle plan, not a shot in the dark and a stroke of luck.