 |
| |
|
|
Deathray
|
|
High Wycombe
Oct 2001
time: 05:15
|
|
quote:
Originally posted by Rhysie
I hope that if it does, this only happens at higher difficulty levels - Otherwise a warlord game played by a very slow, new player who takes their time and looks at everything before even moving a unit might find the AI they face harder than someone who is very quick and judges the situation without too much delay, making snap decisions which allow them to have very short turns could be facing a weaker AI on a higher difficulty level.
I understand that many players want a challenging AI, but I, along with some others, would like an AI which I can play against in a relaxed manner, not micromanaging everything and occasionally doing things I want to rather than because they will improve my winning potential, and still be assured of an eventual win. |
This isn't chess. The AI doesn't have a time limit on how long it has to think. If the AI thinks in your turn then it will just make its turn shorter, not make the AI harder.
|
|
| |
|
|
|
Ironwood
|
|
Fresno, CA
Oct 2001
time: 21:15
|
|
quote:
Originally posted by Deathray
The Simtex games (Master of Orion & Master of Magic) had pretty damn good AI. Both beatable though, even without use of cheese. |
The Simtex games had good AI? I have heard that MOO's was good (or, at least, better than MOO2, which I have played), but I remember playing MOM. While the AI certainly knew how to kick ass, that was about all it could do. Alliances were useless: the AI would break them indiscriminately. The agression levels were pointless: even the most supposedly peaceful or honorable were backstabbing conquistadors. If you were too powerful, they attacked to knock you down. If you were too weak, they took advantage of the situation... and attacked. If you were of aproximately equal power... they attacked, regardless of any promises they had made.
The AI was certainly difficult to beat, no doubt there. Do we know if it cheated at higher levels? It just wasn't fun to play against. You weren't playing against the individual enemy Masters, you were playing against the game iteself, plain and simple.
|
|
| |
|
| |
|
|
|
albiedamned
|
|
Ellicott City, MD
Sep 2001
time: 00:15
|
|
It definitely makes sense to put roads on all your squares, since there is a trade bonus for roads (Andrew1999 just confirmed this on his strategy guide thread).
As for railroading everything, that may be a little bit of overkill. But if your workers run out of stuff to do, then why not?
By the way, I strongly suspect some sort of terraforming occurred on this map. Because Europe is not all grassland.
|
|
| |
|
| |
|
|
|
Deathray
|
|
High Wycombe
Oct 2001
time: 05:15
|
|
quote:
Originally posted by Ironwood
The Simtex games had good AI? I have heard that MOO's was good (or, at least, better than MOO2, which I have played), but I remember playing MOM. While the AI certainly knew how to kick ass, that was about all it could do. Alliances were useless: the AI would break them indiscriminately. The agression levels were pointless: even the most supposedly peaceful or honorable were backstabbing conquistadors. If you were too powerful, they attacked to knock you down. If you were too weak, they took advantage of the situation... and attacked. If you were of aproximately equal power... they attacked, regardless of any promises they had made.
The AI was certainly difficult to beat, no doubt there. Do we know if it cheated at higher levels? It just wasn't fun to play against. You weren't playing against the individual enemy Masters, you were playing against the game iteself, plain and simple. |
The Simtex AI didn't cheat. You're right. In both MOO and MOM the AI was very good at building a strong empire, defending and attacking. However, in both games the diplomacy wasn't up to much, with back stabbing being the order of the day. Also, the AI wasn't at all creative. I never once saw them using a regenerating army of black channeled war trolls or an invisible flying warship ;-)
|
|
| |
|
|
|
Admiral PJ
|
 |
Southeast England , UK
Dec 2000
time: 05:15
|
|
You shouldn't expect an AI as good as a human of course.. thats hopefully where a good Multiplayer game experience will come in, though humans will take longer on there moves.
The way i'm hoping to simulate intelligence in opponents in my coding is to use Fuzzy logic to simulate emotions/control logic routines to effectively create realistic unit movement tactics and general strategies.
Even pathfinding I find is hard to get perfect.. there are so many different things that affect it like enemy ambushes /attacks, units blocking others or obstacles .
I hope the new civ AI at least can transport an invasion party effectively across the whole map without just needing an airport.
I'm having to model how i think millitary strategies would work best. Its all dependant on the programmer - ideally all enemy player strategies should be able to be countered (which is theoretically impossible). AI on this scale is much more sophisticated than chess, where the chess AI of looking at all moves would make each turn take about 2 hours at least I'd predict and would make it very boring in so far as interesting competitors go.
One thing that us programmers often do is to use script based plots and action, where we can control and program all outcomes from a situation. This might be like in half life or Baldurs gate, where a script says for example to send a pack of ogres towards your location, and if only 1 is left then it pleads for mercy from the player.. otherwise it kills itself , or the ogres if victorious could become too confident and mutiny from their king , becoming mercenaries if their morale isn't high towards their king.
This would be better scripted , taking less time to make and allowing more interresting story twists.
I'm planning on using some scripted stories and plots in my own game Mantra, which is a similar genre to Civ3. You could with this recreate the US civil war, its like an advanced events.txt file but more flexible as can be hard coded. You could get given Robert Lee the confederate leader, and the option to negotiate with the french for aid if you manage to break the naval blockade. The talks with the french could be played out like a proper scripted plot, where the french ask for sole Trade rights to the new US when the confeds win, otherwise they could hold Lee's diplomats to ransom for example. My point is realistic and interesting dialogues could be made between characters, making God games seem more personal, and not just on a huge planetary scale where nations seem like a mere flag, not having cultural/political differentation.
AdmiralPJ
email peter@artpac.demon.co.uk for info or to work on
Project Mantra
ICQ : 41505662
|
|
| |
|
| |
|
|
|
Steve Clark
|
 |
Colorado Springs, CO
Oct 1999
time: 22:15
|
|
quote:
Originally posted by CapTVK
P.S
Those mountains still stick out like a sore thumb though. Why can't we have classic 'grey'-style mountains like in Civ2? They just don't seem to fit in with the rest. I hope we can come up with better mods for those. |
Just wait till next weekend, that'll be the first gif terrain replacement I'll use. I don't think there will be any need to create a new one, there are many better mountain terrain from custom terrain.gif files in Civ2 that I'll copy and paste from one of them (probably from Tim's HiRes file).
To get an idea, check out Blackcove's Huge Terrain file http://csc.apolyton.net/terrain.shtml (I wish I could show it to you here).
Last edited by Steve Clark on 29-10-2001 at 01:24
|
|
| |
|
|
|
Ray K
|
|
Allen, TX
Nov 1999
time: 23:15
|
|
quote:
Originally posted by Deathray
The Simtex games (Master of Orion & Master of Magic) had pretty damn good AI. Both beatable though, even without use of cheese. |
No way. You can nearly come back from the dead against Master of Orion's AI. I just did it the other day. Total war declared against me, and I was down about 20 colonies to 2. It took some time, but I won the game with just my 2 colonies. I never built a colony ship during the entire game... It's funny how I can decimate their colonies over and over again with Subspace Teleporters and Omega-V Bombs, but they can't figure out to do the same against me.
I'm with the guys that are saying "the AI will probably suck" until I actually see it do something smart. Guys, this is a very difficult task to program and no one has ever made a competent AI for a game yet. Ever.
The smart money says that we will all be back here talking about AI goofs in about two weeks. Without a 'cheat' menu to see what the AI is doing, it will take longer to find out how good or bad it is.
Last edited by Ray K on 29-10-2001 at 01:44
|
|
| |
|
|
|
Your.Master
|
|
Port Elgin, Ontario, Canada
Apr 2001
time: 05:15
|
|
quote:
Originally posted by The Rook
Has there ever been a really great AI ever? In any game? Computers are just not that capable of making decisions on such a large scale. What computers ARE good at doing is repitition. That's why they can be taught to play chess well. Chess has a limited number of moves, so the computer "brain" just goes through them all until it finds the best option. I think the same thing could be done with Civ, but on a lesser scale. During the human turn, the computer Civs can be planning, thinking ahead, and formulating a strategy, then on its turn, it acts on the strategy it came up with. Each turn it can adapt and improve on the strategy.
I just think during the long human turns, it could be doing number crunching and whatever. |
I agree that the AI civs should do a little of that (not too much, the map changes as you take your turn).
However, many computer games have had their AI reduced so that people could beat the game without dying of repetitive frustration syndrome. Those games end up easy. The reason they don't add it as a difficulty option? Well, sometimes people start at hardest and get frustrated (don't ask why), and more importantly, the AI tends to get overhauled when they find it too powerful.
I remember grade 10 (just 2 years ago) doing my final project for my computer programming class. It was a simple text-based Pac-Man type game, only with programmable levels (not the same environment every time). I thought up an AI as I was going to sleep the night I started, and the next day (the night I was mostly finished) I implemented it and, though simple, it was TOO good at first. For all my being a human who's really into games and knows EXACTLY how the game AI behaves, the little ghosts would always home in on me, surround me, and if I -did- escape somehow they would pursue me and take all the same turns as me...(I didn't even program that last part in...I suppose it must have been done by leprechauns). Anyhoo, I just had to dumb them up so I could make it to level 5.
Incidentally, I just beat level 7 on said program, and am making a level 8. Still no progress on the first level using the old AI.
|
|
| |
|
| |
|
|
|
Matthew
|
|
Manhattan, Kansas . USA
Jan 1970
time: 05:15
|
|
I'm not a computer programer by trade, but it doesn't seem to me that it should be all that difficult to design AI's to be a real pain in the butt even if they are basically as dumb as a doornail (which I guess they are; even the brightest of them). Six mobile units disembarking and attacking a city on the same turn could be very irritating, especially if the AI then destroys the city as soon as it is captured, so you can't get it back. As far as coordinating the attack, the AI just builds the whole force, boats, units, and all, in the same city, and sends it off to terrorize a coastal city when done. It moves as a single unit until arriving at the disembarcation point. Cheating of some sort may be involved in building the group in a reasonable ammount of time in one city, but the AI pretty much has to cheat anyway to be compeditive, if you call it cheating.
Another option: Say you play against 10 or more civs. Two of these, perhaps two in a shield rich area, and preferably a good distance away from the human, start building wonders that humans find useful the instant they build that first city, or at least start building a wonder and then do research towards the more valuable one, switching to it as soon as possible. The civs that do this may never ammount to much, but this still leaves 8 or more to compete against the human player. Imagine someone who likes an early republic-celebrate strategy in civ II seeing the Hanging Gardens built by the AI by 1800 B.C., or a Super Science city fan seeing the Colosus built that early? Neither fatal to the respective strategies, but certainly hurtful. Perhaps another civ beelines research towards another nice wonder, while preparing a city to build it, and trying to keep the Human from that. The rest of the civs can concentrate on gobbling up real estate, except for the one nearest the human, which builds about six crappy not worth capturing cities, then just sends units, in groups, to harass the human player. Basically the civs work together to make life difficult for the human. And it doesn't even have to appear coordinated. Not something stupid and simplistic like just having the civs all declare war on the human, or no one trading with the human.
|
|
| |
All times are GMT. The time now is 05:15.
Apolyton Time is 00:15. |
top of page
|
| archivepost |
|
Forum Rules:
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is ON
vB code is ON
Smilies are ON
[IMG] code is ON
|
|
|
|
|
|
28-10-2001 02:13
|