Personality Archtypes

[RidetheSpiral23]

Soo...

<ReligionArchetypeInfo>
<Type>Crusader_Archetype<Type>
Variables
Variables
Etc
<Type>Devout_Archetype<Type>
Etc...

 

[Thunderbrd]

Exactly

 

[RidetheSpiral23]

Okay thunderbrd I should have the alpha versions done to post them tomorrow. We need to discuss a couple things though...

ibaseattitude - should either go in game strategy, diplomacy or maybe even defined normally on each leaderhead.

VictoryWeights - either as own archetype type or defined by leader

Other major thing is fixing the memories, which I am working on. Also, anyone have any input on adding in new variables based on C2C features? A variable for likeliness to sack cities (like their is for raze city). Contact rands for new diplomacy options that aren't core. Anyone else have any ideas? I guess this would be more Koshling territory to answer since he is AI work right?

 

[Thunderbrd]

Okay thunderbrd I should have the alpha versions done to post them tomorrow. We need to discuss a couple things though...

ibaseattitude - should either go in game strategy, diplomacy or maybe even defined normally on each leaderhead.

VictoryWeights - either as own archetype type or defined by leader

Other major thing is fixing the memories, which I am working on. Also, anyone have any input on adding in new variables based on C2C features? A variable for likeliness to sack cities (like their is for raze city). Contact rands for new diplomacy options that aren't core. Anyone else have any ideas? I guess this would be more Koshling territory to answer since he is AI work right?

ibaseattitude: presuming this is a base setting for diplomacy modifiers towards other civs right? If so, should probably be in diplomacy.

VictoryWeights - As mentioned some tags may have cause to have their own archetype category entirely. However, this seems to me to be something that could work perfectly into the same category as flavors. This way we can ensure that the flavor values the leader plays with do have correlation to the types of victories they seek. How possible do you think it'd be to work this tag's values into that category then?

Without the 'blending' method, we'll need to put some effort into making leader qualities somewhat harmonious with their own goals. (One of the concepts behind the blending was to smooth over incongruous leadership settings.) I recall you mentioned we had a 'game strategy' category... How will this, as a distinct and separate category, work with the Victory Weights and Flavors in a separate category? Will they potentially clash, leading to an ineffective leader, or should they somehow be merged?

As for the last issue, yeah, probably mostly koshling's territory at the moment, but this gives us a much easier way to update with a new variable if we find one would be useful (which I'm sure there could certainly be more depth here for C2C!) I wanted this structure in place so that if I ever considered adding a new leader tag (which in a sense I am - FavoriteTrait & extending Flavor Categories) it'd make updating our leaders MUCH easier!

I'm open to some brainstorming on new leader tags and ways leader personalities can tweak their decision making in general, but it'd come it at fairly low priority for now to act on those as there's a number of pressing goals to attend to.

For one thing, I'd really like to work the AI a bit more on refining and diversifying overall game strategies but that'd take a lot of study of the current code that I have not yet committed to. Just one example: Currently the AI is programmed to plant cities close to neighbors to attempt to hedge them in, but as a player, I don't feel that's always the best decision and should be based on some more personality and evaluation. IF a leader is expansive or organized and tends to be peaceful and possibly protective, then this is a great strategy to hedge in neighbors, spark a fight the opponent started in response and let them walk into your superior defenses. Works to start a war without being the aggressor. But if the opponent civ is far off enough or not able to be hedged in enough or you aren't good with large empires due to your traits, then perhaps just steadily expanding out (which as a player is usually what I do anyhow) is a better method to keep things nice and tight.

The optimal distance between cities is a playstyle choice, the amount of military the AI would want to keep, stack volumes, air or naval reliance, economic development strategies... all these things could be improved to vary based on personalities so that no two leaders try the same strategies and it keeps the player all the more on their toes. When they get predictable is when they get beatable.

 

[RidetheSpiral23]

The game strategy category deals mostly with peace rating, wonder build rate, and espionage. So no having the victory weights in different categories won't really conflict. The Flavors would be more likely to conflict with the victory weights so maybe they should go in the same archetype.

I'm just about done and then I can post them for you. Are you going to run through the Flavors and favorite traits then? I'm not solid on how all the Flavors work. Either way, after that what would be the next thing to do? Dynamic leaders? So I guess we would need many more leaders for all the cultures?

 

[Thunderbrd]

The game strategy category deals mostly with peace rating, wonder build rate, and espionage. So no having the victory weights in different categories won't really conflict. The Flavors would be more likely to conflict with the victory weights so maybe they should go in the same archetype.

Agreed.

I'm just about done and then I can post them for you. Are you going to run through the Flavors and favorite traits then? I'm not solid on how all the Flavors work.

Yes, this will be based on the final outcomes of the Trait flavor settings - we'll make them pretty much a perfect match for the Archetype flavor values. Will take a bit to work out though so in the meantime, finding current leaders that seem to correspond with the 'idea' of the Trait concept the Archetype is named after would present a sufficient temporary model.

Either way, after that what would be the next thing to do? Dynamic leaders? So I guess we would need many more leaders for all the cultures?

I'm going to buy some time to develop the traits out at that juncture and in the meantime have you work on updating the leaders we have to the new Archetype system. Your task there may seem daunting but its going to take a bit to get the trait structure sorted out right so there's a time gap there before I can meet back with you for a new direction.

At that point I will need a few things to work in the Dynamic Leaders direction.
1)Civilizations that will fit the Prehistoric (or more if other eras are developed by then by JohnySmith)
2)To make sure each civilization has at least a couple leaders.
3)A long list of generically named leaderheads (call them by number in this case would work fine) that will be developed under somewhat unusual settings that can then be pulled upon for a randomized leaderhead mechanism.
4)A decent list of potential leaderhead names that correlates to all cultures we have
5)A civilization for every culture in use and a culture for every civilization in use.
6)And a good solid plan for how it's all going to be implemented which will probably include adding some more tags that will be played on for it all.

We'll also be needing to create correlating units for the leaders and I'm going to be looking for art help from Johny Smith on that too.

BUT before all that can really be done, the Dynamic Culture and Adopt Culture systems need to be worked out so that's a lot of XML work and some graphics in adding a Property for each Culture. So that'll be the next big effort there I think.

 

[RidetheSpiral23]

I'm going to buy some time to develop the traits out at that juncture and in the meantime have you work on updating the leaders we have to the new Archetype system. Your task there may seem daunting but its going to take a bit to get the trait structure sorted out right so there's a time gap there before I can meet back with you for a new direction.

That makes sense. Do you have a time table at all do how long you think all of that will take you! Out of curiosity. It'll take me a while to update the leaders but I could always began to work on some of things you mentioned if I finish before you.

We'll also be needing to create correlating units for the leaders and I'm going to be looking for art help from Johny Smith on that too.

I personally can't wait to see in game leaders! But couldn't we implement the feature before this gets done and have the leaders die at random from old age/disease/etc without in game models?

BUT before all that can really be done, the Dynamic Culture and Adopt Culture systems need to be worked out so that's a lot of XML work and some graphics in adding a Property for each Culture. So that'll be the next big effort there I think.

I seriously am going to cry when this feature gets put in. I have thought a lot about this and have a lot of ideas. We will really need to have a discussion about this when the time comes so we can hammer out eachother's ideas and you can explain (more in depth) the mechanics behind properties.

 

[Thunderbrd]

That makes sense. Do you have a time table at all do how long you think all of that will take you! Out of curiosity. It'll take me a while to update the leaders but I could always began to work on some of things you mentioned if I finish before you.

Hmm... I hope to have that all completed by the end of next month. Hard to say how long it will actually take and I'm usually known for being overly optimistic on my time estimates.

We could probably get you working on some of the Dynamic Culture stuff once done here though. We'll also have some further work on getting other areas that use Flavors updated to using the expanded list too.

I personally can't wait to see in game leaders! But couldn't we implement the feature before this gets done and have the leaders die at random from old age/disease/etc without in game models?

Yeah, but it might be easier achieved by going straight into the in-game leaders as it could skip some steps in programming to do so.

I seriously am going to cry when this feature gets put in. I have thought a lot about this and have a lot of ideas. We will really need to have a discussion about this when the time comes so we can hammer out eachother's ideas and you can explain (more in depth) the mechanics behind properties.

Cool! I do have one request of AIAndy on a minor little tweak in the Property system to do this like I'd like to but otherwise it should be fairly easily done. And that's also a juncture where I can pick up some of the places I've been stuck on the Combat Mod too. I think we've all been needing this... a lot, even if we don't yet see it.

If you've been reading Johny Smith's page you can see there that the Dynamic Culture system is really needing established to enable a lot of design ideas there too.

Dynamic Culture is also a pre-cursory step in doing a lot of the Combat Mod stuff I'd like to do too, stuff that will be based on units having a declared culture stemming from the strongest culture existing in the city in which they were trained.

It's a fairly simple project, conceptually, but it'll play into a LOT of things we want to do and will interact with a lot of added design to support it. A number of new buildings, building tweaks, civic tweaks will be a part of all that too.

 

[AIAndy]

Cool! I do have one request of AIAndy on a minor little tweak in the Property system to do this like I'd like to but otherwise it should be fairly easily done.

Please explain before you start implementing. There might be better or more general solutions.

 

[Thunderbrd]

Please explain before you start implementing. There might be better or more general solutions.

May not be the right thread for it but I'll reply here anyhow.

The overall goal of the project would be to make a Property for every culture. Utilize all the nice diffusion and spread mechanics taking place there to track the spreading cultural influences. The only tweak I'd ask of the Property system would be to give us an ability to 'categorize' properties into groupings (in this case 'Cultures') and enable a % calculation of influence within that particular grouping. Thus, if I have 120 of Culture A and 80 of Culture B in a city, we'd have an influence % number next to those volumes (and a # I would want to play off of for various other effects, like declaring a city under a Primary influence of Culture A) that would indicate that Culture A is showing a 60% influence while Culture B is showing a 40% influence.

I'd establish differing layers of influence intensity that call for differing percentages of influence. Units trained would be declared under a particular culture based on a simple majority that exists in the city they were trained in, but UU's and UB's may demand much less influence to exist in the city.

Nations (players) as a whole would also have a declaration of cultural influence percentages that would compare to the 'Declared Culture' of the nation and would play into happiness and revolution modifiers that stem from how in-line that current declared culture is. Leaders would also have a declared culture and may or may not be more or less accepted by the people depending on how strongly their cultural influences correlate to the Leader's culture.

To me, the Property system is so magnificent in the way it gives us such control over diffusion and processing of each property that it seems obvious to go about establishing the cultures along the Property lines. And the tweaks I'm asking for there would have further application in other Property mechanisms down the road too.

So do you feel there's a better way?

 

[ls612]

May not be the right thread for it but I'll reply here anyhow.

The overall goal of the project would be to make a Property for every culture. Utilize all the nice diffusion and spread mechanics taking place there to track the spreading cultural influences. The only tweak I'd ask of the Property system would be to give us an ability to 'categorize' properties into groupings (in this case 'Cultures') and enable a % calculation of influence within that particular grouping. Thus, if I have 120 of Culture A and 80 of Culture B in a city, we'd have an influence % number next to those volumes (and a # I would want to play off of for various other effects, like declaring a city under a Primary influence of Culture A) that would indicate that Culture A is showing a 60% influence while Culture B is showing a 40% influence.

I really dislike this idea. Hydro and myself have spent a lot of time making the current Culture system what it is, and I don't especially want to change that. Also, adding a hundered (!!) new properties would totally kill any turn time improvements we've seen in the past few months.

 

[AIAndy]

May not be the right thread for it but I'll reply here anyhow.

The overall goal of the project would be to make a Property for every culture. Utilize all the nice diffusion and spread mechanics taking place there to track the spreading cultural influences. The only tweak I'd ask of the Property system would be to give us an ability to 'categorize' properties into groupings (in this case 'Cultures') and enable a % calculation of influence within that particular grouping. Thus, if I have 120 of Culture A and 80 of Culture B in a city, we'd have an influence % number next to those volumes (and a # I would want to play off of for various other effects, like declaring a city under a Primary influence of Culture A) that would indicate that Culture A is showing a 60% influence while Culture B is showing a 40% influence.

I'd establish differing layers of influence intensity that call for differing percentages of influence. Units trained would be declared under a particular culture based on a simple majority that exists in the city they were trained in, but UU's and UB's may demand much less influence to exist in the city.

Nations (players) as a whole would also have a declaration of cultural influence percentages that would compare to the 'Declared Culture' of the nation and would play into happiness and revolution modifiers that stem from how in-line that current declared culture is. Leaders would also have a declared culture and may or may not be more or less accepted by the people depending on how strongly their cultural influences correlate to the Leader's culture.

To me, the Property system is so magnificent in the way it gives us such control over diffusion and processing of each property that it seems obvious to go about establishing the cultures along the Property lines. And the tweaks I'm asking for there would have further application in other Property mechanisms down the road too.

So do you feel there's a better way?

I was thinking about introducing 1D-properties. So instead of one value the property has an array of values (stored as sparse arrays / maps so if the value is not there it does not cost anything).
The property manipulators could then be applied on the bunch of values at once (unless they specify a single value).
In the property info you would specify the info class this is based on (like religions, corporations, buildings, ...).
Some generic functions could be added to calculate derived values from the 1D-property.

The big question is how to display 1D-properties properly. Might need some settings in the property info for different display variants.

 

[Thunderbrd]

I really dislike this idea. Hydro and myself have spent a lot of time making the current Culture system what it is, and I don't especially want to change that. Also, adding a hundered (!!) new properties would totally kill any turn time improvements we've seen in the past few months.

I doubt it would impact turn times much but I'd defer to AIAndy to answer that. This was one reluctance to leap on the project earlier.

As for 'what it is' - this would just take it another step in a deepening direction, not replace 'what it is now'. Without what has been done so far, this wouldn't be possible. I'm not talking about introducing a different way to have a culture emerge, just a way to track cultural influences and if there's a change to the current structure in that its just about changing prerequisites on the 'culture resources' over to local culture influence values.

We'd also have ways to push its spread if we wanted. For example, a Storyteller would, like any other unit, have a culture designation attached to it based on the city it was built in. As a part of spreading culture (base game term use) to a city, it would also add a chunk of its designated culture's property to that city. As another example, a line of more minor cultural buildings would be useful for aiding in spread or suppression of cultural properties.

I think AIAndy has a really good idea going here as mentioned below that could greatly improve the processing of such a mechanism.

I was thinking about introducing 1D-properties. So instead of one value the property has an array of values (stored as sparse arrays / maps so if the value is not there it does not cost anything).
The property manipulators could then be applied on the bunch of values at once (unless they specify a single value).
In the property info you would specify the info class this is based on (like religions, corporations, buildings, ...).
Some generic functions could be added to calculate derived values from the 1D-property.

The big question is how to display 1D-properties properly. Might need some settings in the property info for different display variants.

I imagine this would greatly improve the 'speed' of such a system. One property array for all the cultures as it were eh? Works for me. They'd all have the same spread factors. I was thinking the icons would be necessary for each 'culture' flag, but I'd suggest they display on a comma delineated line like 'Icon:126(36%), Icon: 84(x%) etc...'

 

[AIAndy]

I doubt it would impact turn times much but I'd defer to AIAndy to answer that. This was one reluctance to leap on the project earlier.

That depends. Having 100 separate properties will also mean a lot of property manipulator rules and every single one of those has to be instantiated at every possible position it can apply and that costs time.

I imagine this would greatly improve the 'speed' of such a system. One property array for all the cultures as it were eh? Works for me. They'd all have the same spread factors. I was thinking the icons would be necessary for each 'culture' flag, but I'd suggest they display on a comma delineated line like 'Icon:126(36%), Icon: 84(x%) etc...'

Yes, the speed would be considerably faster.
Still, the display might look bad if all are displayed and potentially a lot of cultures could end up in one place so maybe one rule could eliminate weak cultures.

 

[Thunderbrd]

That depends. Having 100 separate properties will also mean a lot of property manipulator rules and every single one of those has to be instantiated at every possible position it can apply and that costs time.


Yes, the speed would be considerably faster.
Still, the display might look bad if all are displayed and potentially a lot of cultures could end up in one place so maybe one rule could eliminate weak cultures.

So that sounds pretty awesome. We have a bit of time yet before we'd be ready to work on this but it is coming up. What kind of timetable do you think we'd have before this would be available then?

Also: maybe the icons themselves could be set with a hoverover popup that shows the % of influence and they don't even display unless they have some influence present? Would that be possible do you think?

 

[AIAndy]

So that sounds pretty awesome. We have a bit of time yet before we'd be ready to work on this but it is coming up. What kind of timetable do you think we'd have before this would be available then?

Also: maybe the icons themselves could be set with a hoverover popup that shows the % of influence and they don't even display unless they have some influence present? Would that be possible do you think?

We'll see how long it will take to implement. Lots of code pieces have to be made aware of 1D-properties.

 

[Thunderbrd]

We'll see how long it will take to implement. Lots of code pieces have to be made aware of 1D-properties.

Ok. Fair 'nuff. I know how that goes.

 

[RidetheSpiral23]

https://docs.google.com/spreadsheet/ccc?key=0AmRhOtK_Ac3ndHFIUkdqZElKM1BNTWs1bHpKLThJdXc&usp=sharing

There is a link to compare Julius Caesar between how he is listed now and with archetypes (archetypes on right). Please take a look and see what you think. Items listed in red are missing entries thus far, (in both, or in mine). If you could clear up if all of those memory decays/percents are needed for me I would appreciate it.

I also attached the folder with the archetype files. Thunderbrd if you could tell me what I need to do from here that would be great. Do i need to enter them into the Schema? How do I structure them? Do I leave them modular in a folder like I have it in there?

 

[Thunderbrd]

On first glance, this is looking very good. Yes, the memories should probably be fleshed out and I trust you'd be able to determine some good values for them based on the others. If you have any questions as to what each one is exactly for, I can try to find a moment to figure that out for you.

As for moving forward, I need to set things up so that can be done and I've been a bit delayed in getting to that, which I apologize for. I'd expected to have it ready by now knowing it sounded like you were on track for being ready for that step very soon. It isn't a tough thing to setup but I need to get this stuff I'm working on off the plate and we need to get past the freeze before I can put it in. I will work out the schema and create a template xml entry in the new xml files and then you can go forward from there by simply entering in each entry.

Your help has been tremendous on this so far and extremely appreciated - as a result I feel bad I'm not quite there on my end yet... but soon... soon. If you have more you can do on spreadsheets for now, that'd be a good way to set yourself up for once that's ready in the meantime.

 

[RidetheSpiral23]

Okay sounds good man. I actually already started on a spreadsheet for leaders.