Anti-Strategy - making the game harder

[Sten Sture]

Anyone that has played a few games of MP or read these boards for a while knows that the AI can't hold a candle to human players and won't anytime soon. However, I was wondering if there are some rules.txt mods that some of you have used to make the single player game vs the AI a little tougher.

Here are some things that I have been messing around with in my v2.42 (so as not to screw-up MGE, if I'm ever home long enough to play.)

Adjusting competitors. There are some civs that seem tougher than others depending on your style of play. For me, aggressive expansionist civs blow up my trade caravans and have small cities, so I don't get as much bonus money; and they have more units so are less likely to give tribute, and harass me early so I have to defend better. So I have adjusted the civ characteristics to make the opponents more Mongol and less Babylonian.

AI reseach priorities. I usually go for MC and JSB which the AI almost ignores, so I increased the research priority of these crucial techs.

Military unit adjustments. Howies are too effective when used by a human, I reduced their movement to 1 and dropped the ignore walls flag. Dips and Spys movement reduced by 1, and increase in cost. Settlers/Eng ignore ZOC to keep me from boxing in the AI expansion.

Terrain adjustments. We have all investigated an AI city and seen them prioritze food over trade, it will work fish over whales and a grassland square over mined wine. In order to get the AI to work the high trade squares I have added to the food production of big trade specials like gold, spice, gems etc.

I've done a bunch of other stuff too, but that is a start. What kind of recommendations do you guys have to make the AI more difficult?



------------------
If you can not think of a really good reason why you should build something other than a caravan, build a caravan. - jpk

[War4ever]

I have a military suggestion for you although it may sound hoky..... make the chariot an attack of four as the ai loves the wheel and the elephant an attack of 5.
Knights a 5. Now .... you are not allowed to build these units.... which could hurt you if you dont' race towards feudalism.....
and get pikeman.... then off to mono to get crusaders. this will make ancient times a little more difficult for you. Double the cost of caravans and dips.... drop the hit points in engineers down to 1. Limit caravan building for trade routes only.... no wonders. Disband obsolete units out side of a city.... no incremental rush building.
last but not least... do not destroy an ai city and race towards AC..... ok i am off to sniff more glue seriously though tell me what you think of some of these suggestions
'

[shamrock]

You bring up a good topic, Sten.

I love the rules.txt file and have several versions that I use depending on the type of game I want to play. In the one I use most often, I've made modifications along the lines you describe for the same reason as you (in order to make the AI more challenging).

I change the tech priorities a lot, increasing the value of the ones I perceive to be critical (i.e. Bronze Working, Monarchy, Conscription, Explosives, Computers, Monotheism, Invention, etc.) and decreasing the ones rated too highly (i.e. Map Making, Rocketry, Nuclear Fission, Genetic Engineering, Combined Arms, Chivalry, etc.).

I also change the number of shields required to build various city improvements and wonders, including changing Leo's to 600 shields rather than a measly 400.

But in the end, there's only so much you can do to combat an egregiously stupid ai.

[Carch]

I've noticed a lot of people complaining about how easy the AI is, then in the same breath saying they don't play the MGE. If you want tougher AIs, play that. They don't like to ally, they are constantly pissed at you once you get bigger than them, and they regularly tell you to bugger off when you ask for tribute.

If milking the AI is part of your strategy, I bet the MGE is more challenging without tweaking.

_/\ C

[Sten Sture]

Carch - you bring up a good point: the AI is substantially more aggressive in MGE than in v2.42. I remember the first OCC game I played in MGE, it really threw me for a loop. I was able to adjuste my game play, and now I don't find SP in MGE much more difficult than it was in v2.42, but some may find it otherwise.

I was looking for ideas that I could test in v2.42 for possible implementation in MGE, which is also the SP version that I usually play; though I will probably leave MGE alone for MP games.

[Aurelius]

War4ever,

I keep the sound on during my single player games. If you're going to add those bonuses to chariots and elephants and BigHorseys, enhance the sound too. That elepant charge still puts the fear of god in me comin' out of the dark lands--even when I have a phalanx in the mountain fortress.

My contribution to this discussion is to encourage the AI to use bribe money. Or to make bribery a bit more costly to the human dips/spies.

Hey Sten,

Howitzers would be much weaker if they had 10 attack with their 2 movement and ignoring city walls. Also, you could reduce their defense to 1--probably not though. I'd build tanks and not howies if we did too much to change them. Perhaps city walls should be a bit compromised from a historic sense in the late game.

Engineers are a big thing to change. Being able to build a railroad within the turn is huge. Same goes for fortresses. Is there a way to delay the fortress leaving the units exposed to attack? Maybe spies should be allowed to sabatage occupied fortresses

Of course, I know nothing about making scenarios or mods so my suggestions are based in wild speculation.

Another thing, make AI build a transport with substantial units in big numbers. They make so many stupid ships without ground troops. Make them use ships in conjuction with helicopters like they do with paratroopers. Damn! AI is senseless.

I have noticed that it stops attacking rifleman in the mountains when it has a few casualties with it's outdated weaponry

Your MC JS strategy is mine as well. I love those two wonders. So it makes sense to make it more shield cumbersome.

(Oh, thanks for the four whales. three wins of course, but not sterling results: 1922, 1906, 1890. Atleast I'm progressing! What surprises me is how different each game develops. And my memory for hut locations is seivish--I stay out of sewers and caves.)

Aurelius the Mad (--usual OCC title is not far off the mark.)

[Steve Clark]

Sten: Your list of suggested changes to the rules.txt is excellent, as usual. I'd be curious on your analysis of how these changes affect strategies and outcomes.

BTW, sounds like you got time to play this weekend. I thought someone in your position in the City would have to keep an eye on the money stuff. Anyway, let us know what happened in your games.


------------------
Also got to state this: If you can not think of a really good reason why you should build something other than a caravan, build a caravan. - jpk

[War4ever]

I have to agree with steve Clark here on this one Sten.... reducing the cost of wonders would only benefit you as the ai cheats when it builds them anyways......
I am wondering if the ai needs to reach a certain population before it is allowed to build a wonder... ie 500k or something.... does anyone know of this?

[shamrock]

I'm with War4Ever - I suspect that it would benefit the human. I always figured that the AI got its wonders for free at certain points in the game. I play MGE and it seems the AI can somehow build a wonder out of the blue when it isn't logistically feasible to do so (i.e. few/no roads, no caravans, small population). I think the timing of AI wonders depends upon a combination of factors, including the required tech, level of difficulty, game year and probably some random factor as well. Does anyone know THE TRUTH?
[This message has been edited by shamrock (edited December 31, 1999).]

[Sten Sture]

Hey I'm watching the money on two thirds of my monitors!

Would significantly reducing the build cost of early wonders make it more of less likely that a human would build them before the AIs. I was just thinking (while I was watching the money with my right eye) that very low cost early wonders might allow the AI to beat you to several of them. I think I'll change some of those first level wonders to 50 shields and see what happens.

On the tech path I also significantly raised the techs leading to monarchy to make all types of AIs try to get there quickly. And I brought Fundy forward to compensate the AI for my usual building of the SoL...

[Steve Clark]

Regarding building the early wonders, I've noticed that the AI doesn't undertake the project until too late. In some cases, I've completed 2-3 wonders before they start to build 1. Not sure if lowering the costs (which would obviously get them built faster) would get them to start sooner?

[Caesar the Great]

This is a very good topic. Many old civ 2 players have found diety too easy and don't enjoy challenges like OCC. MP is good but it takes a long time to play. This would be great for those people, and there enough of them out there. I am thining of a few other ways that one could help the AI but they don't seem to be coming to mind. I'll post something if I think of anything.

The changes that you have proposed in the rules.txt can to some extent be exploited by you. Since you already take advatange of high trade giving yourself more food would work to an advantage to you allowing you to grow your cities faster. Also by making Miches and Bachs more attractive you make other key wonders such as Adam Smiths etc less attractive allowing you to seize all of them while the AI desperatly tries to reach monotheism. Clearly since you know these changes they could benefit you more than the AI. Yes it would make the AI harder to beat, but you would know how to evade tactics you are trying to force the AI to use.

Thats my two cents.

___
CtG

[War4ever]

I personally really find Adam Smith to be a great wonder especially for warmongers as it means free libraries markets temples coastal fortress granary.... free of charge and for a warmonger that is great as i don't have the trade to keep up the maintenance

For the ics'er as well..... basically you double your science as the cities grow faster and your getting all the free gold and science to boot.....



------------------
They call me Mr. Fierce

[Mark_Everson]

Hi All:

I tried the same approach a couple years ago, and I thought what I had learned might help somebody out. First of all, I don't mean to be a wet blanket, but its Extremely difficult to get a good game out of the AI, even if you give it Lots of advantages, and Severely hobble yourself. So here's what I think I learned...

It is Much more balancing to hurt yourself than to try and help the AI. The AI is very rigid in its responses, and changing anything runs a fair risk of damaging it. The only thing I found I could do that would reasonably help the AI, was boosting the attractiveness of various technologies (Monarchy, and the techs leading to it were especially valuable).

There are lots of things you can do to Hurt yourself. Clearly if you add enough of these, you can finally make the game competitive. The problem is as you layer more and more of these each one makes the game a little less fun. I was never able to find a balance, where the game was difficult, And still fun. However, I'll list some of the things I did below, since they were mostly effective. Note that some of these are mutually exclusive, I tried different things at different times.

no trading tech
no stealing tech
upon conquering an enemy city, pick the most worthless tech
no extortion of other civs
no bribing cities or units
no diplomacy whatsoever with other civs
no great library (best in conjunction with various tech limitations)
double cost of all wonders (hurts you more than the ai, or you can just double the cost of your favorite ones)
no intentional we love x (sometimes you can't avoid it in an occasional city)
disband your second settler by 3900 b.c. (lets you build a few roads with him first)

There were some more things I did, but I think these where the big ones.

Good Luck!

[Sten Sture]

I should have known that the AI that switches from a 75% done wonder to a temple wouldn't be able to tell that a 50 shield wonder is a great deal. They still don't start building them early enough. Case closed.