V30 Plans

[Hydromancerx]

Sounds cool! Is that mod turned on yet?

 

[Dancing Hoskuld]

Sounds cool! Is that mod turned on yet?

It is not even on the SVN. All I am doing is bug fixes and screens this release (yea sure!) I have an outline for a bandits mod (if I can figure out their upgrade path) and the basic design for a complete overhaul of the religion system to better fit with a mod that goes from nomadic start to the Galactic era. People will hate it because it sort of includes the "Choose religion" and "Limited religion" options while having many religions.

 

[strategyonly]

I have an outline for a bandits mod (if I can figure out their upgrade path)

Can you PM me on this and let me know what it is thx. It might come in handy in what i am doing also

 

[Koshling]

Having just discovered that the entire evaluation system is being crippled by over-use of AIWeights in the XML, I now need to retune from scratch after eliminating those. As such I do no now expect to add anything else before V30.

 

[strategyonly]

Having just discovered that the entire evaluation system is being crippled by over-use of AIWeights in the XML, I now need to retune from scratch after eliminating those. As such I do no now expect to add anything else before V30.

thats an easy fix IF its ALL AIWeights? Does anyone want to do this, or do you want me to take all the weights off of everything??

 

[Koshling]

thats an easy fix IF its ALL AIWeights? Does anyone want to do this, or do you want me to take all the weights off of everything??

It needs the following doing:

1) Remove all AIWeights (except perhaps a few we absolutely KNOW will be needed like Volkarya's Wonder ones)

2) Play a whole bunch of games to assess what the AI is doping know the distortions have been removed

3) Tweak the AI and/or add a few special-case where-needed AIWeights in response to (2)

4) Go back to step (2) and repeat until things get to a satisfactory state

It's the facts that:

i) It will need a number of games to be played (probably at various speeds) to get a good sample of AI behavior;
ii) Playing a game to sedentary lifestyle on marathon basically takes a day;
iii) It's unclear how many times we'll need to go round this loop

that mean the entire process will likely take weeks.

 

[civfan613]

Koshling
IF you cheat (e.g. using Cheat Engine) and make a super-city (I just love playing around with this) with 1-turn science and/or growth/build (just make all 3 of them some 10k, hehehe) - would it still be reliable enough to see how AI acts?
Cause with using such cheats, you can insta-grow, while still playing in the right order and basically still seeing the AI's reactions, including auto-build.
Sorry, if this actually sounds stupid.

 

[Koshling]

Koshling
IF you cheat (e.g. using Cheat Engine) and make a super-city (I just love playing around with this) with 1-turn science and/or growth/build (just make all 3 of them some 10k, hehehe) - would it still be reliable enough to see how AI acts?
Cause with using such cheats, you can insta-grow, while still playing in the right order and basically still seeing the AI's reactions, including auto-build.
Sorry, if this actually sounds stupid.

Not really - it is highly contextual (for example - a building that adds one production to a city that only has 1 total is HUGE, but the exact same building is almost irrelevant in a city that already has 100 production)

 

[ls612]

It needs the following doing:

1) Remove all AIWeights (except perhaps a few we absolutely KNOW will be needed like Volkarya's Wonder ones)

2) Play a whole bunch of games to assess what the AI is doping know the distortions have been removed

3) Tweak the AI and/or add a few special-case where-needed AIWeights in response to (2)

4) Go back to step (2) and repeat until things get to a satisfactory state

It's the facts that:

i) It will need a number of games to be played (probably at various speeds) to get a good sample of AI behavior;
ii) Playing a game to sedentary lifestyle on marathon basically takes a day;
iii) It's unclear how many times we'll need to go round this loop

that mean the entire process will likely take weeks.

#1 is done now on the SVN.

Also, does this mean that we won't see any performance improvements this version?

 

[Hydromancerx]

I put AI weight 100 for Culture (African), Culture (Asian), Culture (European), Culture (Middle Eastern), Culture (Oceanian), Culture (North American), Culture (South American) and Culture (Barbarian).

The reason these give only +1 so the AI probably won't build them. However they are very cheap (5 base) and are important to unlocking all culture wonders.

In addition its probably a good thing to be recommended to be built first in your new city before you forget.

 

[strategyonly]

I put AI weight 100 for Culture (African), Culture (Asian), Culture (European), Culture (Middle Eastern), Culture (Oceanian), Culture (North American), Culture (South American) and Culture (Barbarian).

The reason these give only +1 so the AI probably won't build them. However they are very cheap (5 base) and are important to unlocking all culture wonders.

In addition its probably a good thing to be recommended to be built first in your new city before you forget.

For me, if it wasn't for the Heroes, i wouldn't even bother with the Culture Civ Units, i personally dont use them anyways.

 

[Thunderbrd]

Some aren't much chop, but some of those UUs are really awesome. They could use some rebalancing to make sure they are almost always in some way superior to their contemporaries. I say 'in some way' because you have some units where their benefits are somewhat counter-intuitive in a good way, like the Pygmies, which appear to be intended to be a touch weaker than contemporaries but EXTREMELY quick to build so you can use them for a quick swarm. Great for defense.

 

[strategyonly]

Some aren't much chop, but some of those UUs are really awesome. They could use some rebalancing to make sure they are almost always in some way superior to their contemporaries. I say 'in some way' because you have some units where their benefits are somewhat counter-intuitive in a good way, like the Pygmies, which appear to be intended to be a touch weaker than contemporaries but EXTREMELY quick to build so you can use them for a quick swarm. Great for defense.

See that also depends on the SPEED you are playing on also, i mainly play on Marathon and its quite useless for culture units, you get better results using NORMAL units, and it doesn't have a cost also

 

[Hydromancerx]

I put AI weight 100 for Culture (African), Culture (Asian), Culture (European), Culture (Middle Eastern), Culture (Oceanian), Culture (North American), Culture (South American) and Culture (Barbarian).

The reason these give only +1 so the AI probably won't build them. However they are very cheap (5 base) and are important to unlocking all culture wonders.

In addition its probably a good thing to be recommended to be built first in your new city before you forget.

You know what/ The more I think about these building the more I think it makes sense for them to just be auto-built. I mean they are not harming anything and you still have to make the wonders. It just would take out a repetitive step for a pseudo-building.

In short i plan to undo the AI weight on them and just make them auto-build buildings.

 

[Thunderbrd]

See that also depends on the SPEED you are playing on also, i mainly play on Marathon and its quite useless for culture units, you get better results using NORMAL units, and it doesn't have a cost also

True... the costs make you want to overlook them usually.



@Hydro: That's a good idea. What about we create an option where which base culture that auto-builds is based on the location of the first city? Seems to the people were set on a cultural path from their initial global location and the physical adaptations they underwent as a result.

 

[Hydromancerx]

@Hydro: That's a good idea. What about we create an option where which base culture that auto-builds is based on the location of the first city? Seems to the people were set on a cultural path from their initial global location and the physical adaptations they underwent as a result.

I don't understand. These are base cultures based upon what civ you picked. Such as if you picked French you would get Native Culture (European) auto-built which in turn gives Culture (European) to every city. Which in turn is required to build the various European cultures such as French, English, German, Greek, Roman, etc.

Are you saying it should be based upon latitude/longitude or even terrain? Because terrain makes little sense. Look at North Americans, they range from the Inuits in the cold Arctic to the Navajo in the hot desert to the Sioux on the Great Plains to the Iroquois in the woodland forest.

 

[Thunderbrd]

it should be based upon latitude/longitude

Yes. We spawn animals with this sort of consideration. Why are we not doing this for the starting spot for civs? I'm really not at all against completely dissolving the basis of the civilization selection outside of flag, (depending on option) leader access, and (even this is adjustable) border color.

Plus, this would make the base cultures tend to blend a bit better with their environment, lending to more commonly accessible specific cultures, as the prereqs for those specific ones will be more likely to be present in the area.

 

[Thunderbrd]

ulp... double posted

 

[Hydromancerx]

Yes. We spawn animals with this sort of consideration. Why are we not doing this for the starting spot for civs? I'm really not at all against completely dissolving the basis of the civilization selection outside of flag, (depending on option) leader access, and (even this is adjustable) border color.

Plus, this would make the base cultures tend to blend a bit better with their environment, lending to more commonly accessible specific cultures, as the prereqs for those specific ones will be more likely to be present in the area.

I know you can set <iMinLatitude> and <iMaxLatitude> for buildings but I don't think they have longitude code. But even if we did I am not sure if we should. Half the fun is having civs in locations of the globe they normally would not be. And if its based on the native cultures then you would be basically limiting them to a part of the hemisphere which means longitude more than latitude.

I also think it would mess up scenario maps a lot. We already have a problem with animals and scenario maps. Imagine how much worse it would be for civs.

 

[Koshling]

#1 is done now on the SVN.

Also, does this mean that we won't see any performance improvements this version?

It all depends on time. My priorities are now in this order:

1. AI economy tuning for build/tech/religion founding etc.
2. S&D AI
3. Performance

Since (1) is open-ended I don't know if I will get to (2) or (3).

BTW - I'm away for 5 days from this Friday (well - I'll have iPad access to the boards, but no modding/playing access)