AU: A MOD for the curiculum

[alexman]

Talk about radical!

If we remove the militaristic characteristic, it just means it won't be half price for militaristic civs. This is one of their better advantages. Let's not mess with the traits.

Right now the AI builds harbors mainly when there is not enough food to have the city grow. That often makes the AI start trading too late. I'm afraid that if we make coastal fortresses required for trade, the AI will start trading even later. They would build fotresses mainly in wartime when they want to build a navy. But that's exactly when they will have less civs to trade with!

[lockstep]

IIRC, Soren Johnson said in your 'AI' thread that at the start of a new game every improvement is pre-assessed by the AI according to its costs, flags etc. Shouldn't the AI know that it would need imp. no. 24 (for humans 'coastal fortress') instead of imp. no. 26 ('harbor') for trading overseas, and change its building priorities accordingly?

About leaving 'militaristic' for a +1 food only improvement - well, if it helps balancing the traits, I won't fiercely oppose it.

[player1]

Originally posted by lockstep
I knew you would say that.

Apart from moving an improvement into another age, what's your opinion on 'splitting' harbors?

EDIT: About the 'cannon graphic' ... walls don't look THAT different to me.

The only split, that looks logical to me, is taking out Veteran ability of Harbors.


P.S.
But walls become teared down when cities get pop of 7.
(look city view)

[lockstep]

No offense, player1. I just wanted to say that, IMO, the graphics for coastal fortresses don't look totally weird in the ancient age.

[player1]

No offense taken.

[Theseus]

Let's not stay too far... I support coastal fortresses for veteran ships, but that's about it.

Back to the build preferences: I'm not seeing the emergence of a killer AI civ in AU 103... if anything, they are all seemingly poor, and unable to achieve a tech lead. Admittedly, alexman created very artificial environments. Any thoughts?

[alexman]

The problems with the AI that I can see:

1) They build too many land units. This gets in the way of infrastructure. But if they don't do this, it will be too easy for the human to invade at will. So I don't see how to fix this. Perhaps to remove units from build preferences?

2) They build factories too late. Even with "production" flagged. Even with Soren emphasizing shield-producing improvements in 1.29. Not sure what else to do, but flagging production is definitely the right thing. The AI places a strong empasis on shield costs (he doesn't even build the spaceship if it's too expensive). If we make factories cheaper, the AI will build them more often. OTOH that will also make them a no-brainer for the human... The usual dilema: improve the AI or keep strategic decisions?

3) They are willing to give up all their income to get a tech. That means they have nothing leftover for research of their own. Once that happens, they're doomed. We can't fix this in the editor either.

4) They put their FP in stupid places. They don't use leaders to relocate their palace. Again, I can't see a solution apart from removing the FP alltogether.

Things they do right:

1) They now do build marketplaces. Not sure if this is 1.29f or the build preferences.

2) They build cultural improvements. In my AU103 game, I am actually behind in culture in mid-Industrial age.

In summary, I don't think we can dramatically affect the AI from the editor. Most of the decisions are hard-coded.

[player1]

The real probelm with AI is that it doesn't know concept of GOLD RESERVE.

He'll always sped his last gp for some rush buy or tech.

Unfortunately it can be only fixed by Soren.
(if he feels like)

[alexman]

Yes, number 3 above is the biggest, I agree...

Theseus: Another problem is that the AU103 was at Monarch. Perhaps you are just too 1337?

[lockstep]

Problem number 1 could be toned down by generally increasing shield costs of units and/or decreasing shield costs of improvements/wonders. But that's a HUGE task of rebalancing.

[Theseus]

I forgot that it's Monarch... I definitely feel the difference.

Another interesting aspect of AU 103 (SPOILER AHEAD), is that I have really been the only meaningful warmonger in the game, taking over America's island. I would have thought that would have tripped the aggression flag, which usually enhances the winner / loser feedback loop, but 1) I guess I was not enough of an imminent threat, and 2) I maintained polite or gracious relations with everyone.

My guess is that, in addition to starting environment, resource distribution, build prefs, etc., there needs to be at least two civs with aggression at +2 (inclusive of the human player) in order for worldwide aggression to contribute to the rise of a killer AI civ.

[=DrJambo=]

Originally posted by Theseus
My guess is that, in addition to starting environment, resource distribution, build prefs, etc., there needs to be at least two civs with aggression at +2 (inclusive of the human player) in order for worldwide aggression to contribute to the rise of a killer AI civ.

I was actually just going to suggest the opposite.

The AI is usually very good up until the late medieval, early industrial age. This is because many are at peace and expanding into the unsettled land. As soon as they fill up the empty spaces they look to picking a fight.

The reason the human player catches up in the industrial period is the AI spends most of the entire period in one war or another, signing MPPs or MAs with each other. Meanwhile the human sits in democracy, building infrastructure and speed-researching techs.

What i was going to suggest was to lower the balance of aggression across some of the civs. The balance is already way off centre as it stands...

Aggressive setting (number of civs):

1(2)
2(2)
3(4)
4(6)
5(2)

Therefore in the late period, once all land is settled, all you tend to get is a lot of AI civs fighting and falling behind. At least some of the civs should be encouraged to remain in peace and build infrastructure and research tech. Reducing the number of civs in the aggressive 4 setting should help contribute to a more balanced game, or so i think.

[lockstep]

Still thinking about the Coastal Fortress changes ...

Moving the 'veteran naval units' ability from Harbors to Coastal Fortresses is a good idea because it makes both improvements somewhat better balanced. However, the vet ability in fact becomes the new prime function of CF's; therefore the following changes are appropriate IMO:

• Make Coastal Fortresses available with Astronomy instead of Metallurgy. That way, you can build veteran naval units as soon as you can build Caravels (the second ship on the tech tree).
• Remove the ressource prerequisites (Iron, Saltpeter) from CF's. Barrracks (vet land units) and Airports (vet air units) also don't require ressources.
• Give CF's upkeep costs of 1 (same as Barracks). You also may want to increase their shield costs to 80, which is still only 50% of every other middle age improvement (Cathedral, University, Bank).

[alexman]

OK, the Coastal Fortress shouldn't be that big of an issue, but let's get it settled anyway. As I see it, we have three options within the scope of this mod:

1. Leave unchanged: The improvement stays useless and it gets built by neither the human nor the AI.
2. Make it produce 'elite' units: We would do this by removing the veteran flag from harbors, adding it to CF, and adding 1HP to all naval units. We would also add an upkeep cost, but not increase the shield cost, so the AI is encouraged to build them. If you don't build them, everything is as before, since most cities with enough production to build ships, already have a harbor. The AI will probably build them in wartime, just like barracks.
3. Make it a sea baracks: As above, but move it to astronomy and remove resource requirements. Add upkeep and increase cost. [Edit: and remove the HP bonus to naval units]

I vote for the second choice, because it's a good compromise of changing as little as possible. Elite units are less common for naval units (less battles), so the HP change is not critical, but it avoids naval battles between 3HP units, which are very unpredictable. (The difference in randomness of results between 3HP and 4HP is greater than that between 4HP and 5HP). Also, I'm not a big fan of the idea of having the fortress fire at ships before cannons are available!

What does everyone else think? (Lockstep, I know what you think )

[lockstep]

Originally posted by alexman
What does everyone else think? (Lockstep, I know what you think )

If no. 3 is what I think then it's without any hp bonus for naval units.

In the end, player1's comment on changing corruption (in Civ3 v1.17) may also apply for Coastal Fortresses: 'If you don't know how to fix something FOR GOOD, don't change it.'