AU: A MOD for the curiculum

[Theseus]

Friggin' awesome.

I'm getting scared of the PTW AU AI's (pronounced "pit-wowies"?).

[Catt]

Just downloaded the 1.14f patch (available at www.firaxis.com).

Haven;t even installed it yet, but I've read through the Readme and a couple things that jumped out at me:

* Added Hwach'a and Berzerk to upgrade path.
* Removed Hwach'a airlift ability.
* Spanish build preferences changed to building wealth often.

I'm glad the commercial Spanish will now take better advantage of their trait.

Catt

[alexman]

LOL! Pit-wowies!

Catt, thanks for the info! It's good I didn't start messing with the bix file yet!

Originally posted by Nor Me
Unlike Player1's, AU Bowmen have zero range bombard.

Good point.
Let's think about how to deal with that.

[player1]

Originally posted by Catt
Just downloaded the 1.14f patch (available at www.firaxis.com).

Haven;t even installed it yet, but I've read through the Readme and a couple things that jumped out at me:



I'm glad the commercial Spanish will now take better advantage of their trait.

Catt

By the way, I was the one who pointed these bugs/features to Firaxis (by sending a letter to Speed Bump, who manages all MOD donwload jobs for civ3.com).

But it's pretty possibile that they would find about that without me too.

[Catt]

Originally posted by player1


By the way, I was the one who pointed these bugs/features to Firaxis (by sending a letter to Speed Bump, who manages all MOD donwload jobs for civ3.com).

But it's pretty possibile that they would find about that without me too.

That's partly why I posted it. But Soren also browsed through these boards, commenting particularly on alexman's "Why, oh why does the poor AI . . . ." thread (in which numerous "bugs" or less than optimal AI play characteristics were highlighted) -- I know you contributed in that thread also. (I wonder if the AI will continue to build guerillas even when it has rubber ).

Catt

[Catt]

Is the AU Mod working?

Alexman, as always, great work

I've been "Civ-limited" a bit lately by RL and haven't been playing as much -- haven't even played AU 205. I play mostly standard games, with the occasional AU Mod game thrown in for comparison.

For those of you still getting a lot of playing time: How have the AU Mods performed to date? Have we succeeded in at least one of the early goals -- inducing the AI to make better gameplay choices and therefore providing a richer, more challenging SP experience? Do we have enough anecdotal evidence to say with confidence that the AU Mod produces "Killer AI's" with greater frequency than the standard game?

I am more interested in hearing if and to what extent the AU Mods seem to be inducing better AI gameplay (as opposed to simply a stronger AI because units that it uses frequently have better A/D/M stats, etc.) Penny for your thoughts!

Catt

[ducki]

I wish Philisophy needed a boost, because Science(Edit:Scientists) to Philosophy makes the ultimate amount of sense.

Philosophy was the original science.

Another case of realism and gameplay conflicting, but gameplay is more important, I think.

[alexman]

Re: Is the AU Mod working?

Originally posted by Catt
Alexman, as always, great work

Thanks Catt, but it's not my work, it's everyone's here.

How have the AU Mods performed to date?

Several of the changes have made the AI unquestionably tougher because we have changed the game to fit its strategy. (Infantry attack, double-effect entertainers, Communism boost, et cetera).

The tools that we have to control the AI's behavior are very limited, and the effects are not immediately visible. The build preferences and unit offense/defense flags are not a very powerful tool, but it definitely helps. In other words, something is better than nothing, but don't expect to have to drop down a level because the AI now builds factories before banks.

On the other hand, I expect the new AI research path to have a bigger effect, but this has not yet been tested in a real AU game.

The increase in OCN for the AI will also have a considerable effect, but this more of a cheat (or compensation for its inability to make good use of its FP).

[ducki]

Action: Added "Enables Military Alliances" ability to Polytheism
Reason: To help Religious AI civs B-line for Monarchy.
Comment: Why Alliances for Polytheism? "We believe in the same Gods, so together let's kill
all those who don't...". Writing still also allows Alliances.

I wonder...
...what would happen if you completely duplicated the "Allows.." from Writing to PolyT? I know it might seem illogical to some, but it just might generate a "killer" AI from those that start with any of the pre-reqs to the Monarchy chain.

I didn't originally like the "We believe in the same gods" logic, but I don't like separating Embassies from Alliances even more. I thought those two were inextricably tied, which may be tainting my view, but I don't think you need writing for "embassies" or alliances, just common ground, so maybe we could consider having two paths to embassies and alliances?

This might make it more likely to have different paths taken by each AI instead of all of them - well, most of them - going down the same road...
...we might actually end up with AIs taking Dominae's Monarchy beeline and raking in some cash from the ones taking the Republic beeline and vice versa.

What think you all?

[ducki]

Originally posted by Catt
Just downloaded the 1.14f patch (available at www.firaxis.com).
Catt

Excellent, but, as a web veteran, there's no reason it should be that hard or take a leap of logic that big to find that damn download.

That's just really poor information architecture.

[Theseus]

I totally agree with ducki... couldn't find it at Firaxis.

As to AI performance, it's looking good to me, but I have also been constrained by RL.

[ducki]

Oh, Theseus, don't follow any of the Civ3 links.
At Firaxis.com, click the nav link to Downloads->All downloads at the top nav bar.

First one on the list is 1.14

[Catt]

Re: Re: Is the AU Mod working?

Originally posted by alexman

Thanks Catt, but it's not my work, it's everyone's here.

True! And kudos to all who contribute. But without someone to occasionally give the process a swift kick in the a$$ by proposing something concrete and insisting on feedback, we would all descend into esoteric and philosophical discussions of this and that.

Several of the changes have made the AI unquestionably tougher because we have changed the game to fit its strategy. (Infantry attack, double-effect entertainers, Communism boost, et cetera).

The tools that we have to control the AI's behavior are very limited, and the effects are not immediately visible. The build preferences and unit offense/defense flags are not a very powerful tool, but it definitely helps. In other words, something is better than nothing, but don't expect to have to drop down a level because the AI now builds factories before banks.

On the other hand, I expect the new AI research path to have a bigger effect, but this has not yet been tested in a real AU game.

The increase in OCN for the AI will also have a considerable effect, but this more of a cheat (or compensation for its inability to make good use of its FP).

I'm not sure I like the OCN change for the very reasons you identify. I'd really like to see the results of the research path changes. And although there seems to be few levers available in the editor to truly improve AI gameplay, I am most interested in such levers (and remember being excited by the prospect that the "build often" flags might really have a notcieable effect).

The best work is of course catching Soren's eye with threads like "Why, oh why . . ." for nothing in the editor can compare to getting into the code.

Other views? How is the AU Mod doing at creating "Killer AIs?"

Catt

[Hutak]

Wow! I can't believe I missed this thread all this time. That'll teach me not to come to Apolyton often enough. But anyway... =)

1- Is there a reason why we're not giving communism the same trade bonus as republic and democracy? That would help the AI to stay more competitive in research, no?

2- Do we know *why* the AI switch to commie governements? Which is the biggest factor, war weariness or unit support? Did anyone make some tests on that? I think knowing why could help us tweak it to provide the AI with more incentive to stay demo/republic.

3- Increased OCN: I'm ambivalent on this one too. However, it should help the communist AIs quite a bit. If I understand the corruption calculations correctly, it will cut their corruption level by 1/3 under that governement.

4- Dominae: I believe the reason for not upgrading the attack value of the guerilla unit all the way to 8 is to try to prevent the AI from building it instead of infantry when it has access to rubber.

5- Alex: How about giving us a file with all those change implemented so we can test it (especially the new emphasis on certain techs for the AI)? I did implement AU 1.06 in PTW 1.04 yesterday, but then they released 1.14 and I don't want to redo it all again!

-Hutak

[alexman]

I have another question for feedback:

Player 1 has reduced the cost and attack factor of Berserks, because he thinks that having an axe-wielding warrior with equal attack to Cavalry with rifles is crazy.

I don't like this change because it will somewhat change Scandinavian strategy, as it would be easier to mass-upgrade archers.

Another solution would be to reduce the attack factor to 5, but give an extra HP to that unit, keeping the cost the same. The odds for a veteran Berserk beating a veteran pikeman fortified in a city would be approximately the same as in the stock version.

What do you all think?

Re: Alliances and Polytheism, I think the best solution would be to have Writing continue to allow alliances as well. It would change gameplay too much if we had to go all the way to Polytheism (in addition to Writing) just to form an alliance.

Re: Scientists, I think the best place for them in terms of both AI values and realism is in Mathematics.

Originally posted by Hutak
1- Is there a reason why we're not giving communism the same trade bonus as republic and democracy?

That would make the human choose Communism too often. We don't want to change the way the game is played that much. Perhaps we will try it another time.

2- Do we know *why* the AI switch to commie governements?

WW, definitely. I don't know any way to fix that. The AI just doesn't know when to stay out of wars and when to fight them.

5- Alex: How about giving us a file with all those change implemented so we can test it (especially the new emphasis on certain techs for the AI)?

OK, I'll have a version out today.