Single Player bugs and crashes v38 plus (SVN) - After the 20th of February 2018

[JosEPh_II]

TD and WFL are the two greatest tools for ensuring this kind of balance. Playing without them and expecting the balancing efforts to pander to you is like shooting out your tyres and then complaining about how your truck handles.

No wrong! Absolutely wrong. For the system to be balanced you Must balance the Base 1st. Then you adjust the Options that pander to those that need crutches, AI or Player. And that is what TD and WFL were designed originally to do, be Player or AI crutches. If you Make or Set the modifiers for TD and WFL to be ON all the time then you have changed Base Play.

You statement shows a total Lack of understanding of this base play.

Raxo reduced the tech cost discount AI get from handicaps, and you still complain...

The way he worded it, it came across as an Increase. And I reacted to that. So get off my back about that misunderstanding.

It was a thing since overhaul in V38.5
For a year AIs were getting 5/10/15/20/25/35 reduction to tech cost on Prince/Monarch/Emperor/Immortal/Deity/Nightmare.

Not for years as you put it, but only after the new system was installed in C2C. It's these generalized statement that is causing the confusion.

In vanilla BtS techs costs 30% more for the human player on deity than it does for the AI.
A tech that costs 100 beakers for the AI will cost 130 beakers for the deity player.

And vanilla BtS has How Many Techs? And How Many Eras? Compare apples to apples please. And stop the damn nitpicking at what I'm trying to convey.

C2C was closer to vanilla BtS in this regard before rev 10410.

No it was not, again 84 BtS Techs vs 920+ C2C Techs, 6 Eras vs 15 Eras. Bad analogy and Bad comparison.

Look you can ***** at me for wanting a solid base of play that the player can use. A scale of Handicaps that make sense for using C2C's massive scale. Comparing something as Huge as C2C to Base BtS id flawed. And I thought you wanted correctness???

 

[Yudishtira]

No wrong! Absolutely wrong. For the system to be balanced you Must balance the Base 1st. Then you adjust the Options that pander to those that need crutches, AI or Player. And that is what TD and WFL were designed originally to do, be Player or AI crutches. If you Make or Set the modifiers for TD and WFL to be ON all the time then you have changed Base Play.

You statement shows a total Lack of understanding of this base play.

I suppose there's no point trying to get through to you, but the set of options you use don't automatically become "the base" by virtue of you using them. By working to address precisely what you're complaining about (when nothing else did), TD and WFL became part of "the base". If you don't believe me, why not check TB's reply to me above.

 

[raxo2222]

Not for years as you put it, but only after the new system was installed in C2C. It's these generalized statement that is causing the confusion.

I meant around year, not multiple years here.

 

[Thunderbrd]

Actually research was more expensive for players before V38.5
Techs used to be more expensive for player by 5/10/15/20/50/125 percent on Prince/Monarch/Emperor/Immortal/Deity/Nightmare.

True but the distinction is moot. If it's 30% more for a player then it's 30% less for an AI by the same equation.

One thing i have noticed lately about spawning is that the "natural" resources like horses, pigs.llama, etc etc are NOT spawning, i am into late Ancient era now and yet to have some spawn?? none in the area whatsoever (pic 1)

There's nothing to isolate those types into behaving differently in the code so aside from manipulating base rates on that, I'm not sure what the problem would be. I've seen spawning happen normally in all test games recently and the one change in that region would affect all rates of spawning, not just some, though it could appear to make less common spawns appear even less common.

No wrong! Absolutely wrong. For the system to be balanced you Must balance the Base 1st. Then you adjust the Options that pander to those that need crutches, AI or Player. And that is what TD and WFL were designed originally to do, be Player or AI crutches. If you Make or Set the modifiers for TD and WFL to be ON all the time then you have changed Base Play.

I get what you're saying from the perspective not just of playing but of trying to improve design. The handicaps should be an assist rather than a necessity for the mod to function reasonably well.

No it was not, again 84 BtS Techs vs 920+ C2C Techs, 6 Eras vs 15 Eras. Bad analogy and Bad comparison.

I don't see that as relevant. Sure there's a longer road but we're not talking about the distance to travel. We're talking about the speed of travel. Unless you're talking about VERY short distances and considering factors like acceleration and having to time slowing to stop before hitting a wall at the end, just talking about how well one player keeps up with another has nothing to do with the length of the path, only the speed they have. If we can set the AI handicap to be less than it was set on Vanilla, that purely means that our AI is better than the vanilla one at optimizing it's research path to obtain greater growth and more accomplishment in a smaller amount of time, and that such potential is enabled in our buildings and they are taking advantage of it better. Full stop. That has nothing to do with how many techs there are or how many eras or anything about that.

 

[raxo2222]

I don't see that as relevant. Sure there's a longer road but we're not talking about the distance to travel. We're talking about the speed of travel. Unless you're talking about VERY short distances and considering factors like acceleration and having to time slowing to stop before hitting a wall at the end, just talking about how well one player keeps up with another has nothing to do with the length of the path, only the speed they have. If we can set the AI handicap to be less than it was set on Vanilla, that purely means that our AI is better than the vanilla one at optimizing it's research path to obtain greater growth and more accomplishment in a smaller amount of time, and that such potential is enabled in our buildings and they are taking advantage of it better. Full stop. That has nothing to do with how many techs there are or how many eras or anything about that.

Buildings just for comparison:

We have 7762 buildings according to Notepad++ search of "<Type>BUILDING_"
Assuming some are repeating overrides and others are not buildable ones, still there is 3000 - 5000 buildings, that city could potentially own over course of game including 100 to 400 regular (not wonders not religious not special) buildings per era.

1110 buildings have AIweight of 1, this AI can build 1000 more buildings in city over course of whole game.
Hunting Instruction and Community Discussion, AI always build those now - @Dancing Hoskuld never could get AI to build Hunting Instructions before.
Categories that got AI weight are: National wonders giving resource or free buildings or being requirement for something, regular buildings being upgrades of older buildings or requirements for other buildings, smelters/smithies and Industrial+ resource producers.

This is what slows down game units aside and what boosts AI economy, so handicaps had to be flattened.

Civ4 BTS had 159 buildings of all types, some of them being unique to civilization.
That is there is much more to see during adventure

 

[Snofru1]

The Mathematics Department replaces the Mathematics Academy. The latter starts with 3 but gets more just when you research Geometry etc. so in the end you have 9 when its replacement can be built. But the Department only has 5 so in my case with all other effects on science included I get a reduction of 33 when I build the Department (although it also has a free scientist). The amount of should be either reduced for the Mathematics Academy or increased for the department (or both).

 

[Dancing Hoskuld]

If I remember correctly the Mathematics Academy is a Great Scientist building. It should not be built with nor should any building built with replace it. None of the GP buildings should it detracts from the GP.

 

[raxo2222]

If I remember correctly the Mathematics Academy is a Great Scientist building. It should not be built with nor should any building built with replace it. None of the GP buildings should it detracts from the GP.

Mathematician Academy is normal building and can't be placed by GP. Scientist can place building simply named Academy.
@pepper2000 didn't look at its tech boosts apparently.

 

[Thunderbrd]

Mathematician Academy is normal building and can't be placed by GP. Scientist can place building simply named Academy.
@pepper2000 didn't look at its tech boosts apparently.

10412
Mathematics Department no longer replaces Mathematics Academy

Report was here.

Not replacing it isn't right either. Either the original needs to be toned down or the replacement ramped up. Or a little of both. There's FAR too much science boosting taking place as the mod goes on past Medieval...

 

[raxo2222]

Not replacing it isn't right either. Either the original needs to be toned down or the replacement ramped up. Or a little of both. There's FAR too much science boosting taking place as the mod goes on past Medieval...

I guess Renaissance scientific revolution is simulated too strong .
Those buildings coexist for two eras and then Mathematics Academy is obsoleted.

Probably those buildings were designed by someone, who didn't beeline at all.
I don't beeline in my games too.

 

[Dancing Hoskuld]

Hunting Instruction and Community Discussion, AI always build those now - @Dancing Hoskuld never could get AI to build Hunting Instructions before.
Categories that got AI weight are: National wonders giving resource or free buildings or being requirement for something, regular buildings being upgrades of older buildings or requirements for other buildings, smelters/smithies and Industrial+ resource producers.

Not true I never tried. I have posted a number of suggestions and then there was the Hunting Review that went nowhere.

Putting an AI weight on a building is a last resort. In general it is bad practice.

edit part of the problem with many buildings is that they were designed by many people. Some felt that small affects were best whereas others kept increasing the effects until the AI would build their buildings first. There was no over all view of how they would work together and with the Traits. This means that some buildings that should be in keeping with some traits were change so much that traits have no effect.

 

[Thunderbrd]

There was no over all view of how they would work together and with the Traits. This means that some buildings that should be in keeping with some traits were change so much that traits have no effect.

You lost me a little here but interested me in what you were trying to say. I may have to clarify something but I'm not sure. In what way are you employing the term 'Traits' and how did Traits come up here?

 

[Dancing Hoskuld]

I probably meant Flavors which have the same values as the vanilla Traits. I am always getting those two mixed up.

edit values is the wrong word... Flavors allow you to say that this building is favored by these traits and by how much. They seem to have a sum of 10 across the flavors/traits with zero being normal.

edit 2 according to the Modiki

Flavors - Used by the AI to decide what buildings to build based on personality

where I assume personality is what is also called traits. However they don't seem to be named the same anyway...
eg

            <Flavors>
                <Flavor>
                    <FlavorType>FLAVOR_GOLD</FlavorType>
                    <iFlavor>10</iFlavor>
                </Flavor>
                <Flavor>
                    <FlavorType>FLAVOR_ESPIONAGE</FlavorType>
                    <iFlavor>2</iFlavor>
                </Flavor>
            </Flavors>

 

[Thunderbrd]

I probably meant Flavors which have the same values as the vanilla Traits. I am always getting those two mixed up.

edit values is the wrong word... Flavors allow you to say that this building is favored by these traits and by how much. They seem to have a sum of 10 across the flavors/traits with zero being normal.

edit 2 according to the Modiki

Flavors - Used by the AI to decide what buildings to build based on personality

where I assume personality is what is also called traits. However they don't seem to be named the same anyway...
eg

            <Flavors>
                <Flavor>
                    <FlavorType>FLAVOR_GOLD</FlavorType>
                    <iFlavor>10</iFlavor>
                </Flavor>
                <Flavor>
                    <FlavorType>FLAVOR_ESPIONAGE</FlavorType>
                    <iFlavor>2</iFlavor>
                </Flavor>
            </Flavors>

I use Flavors for the Trait selection AI so I'm very familiar with them, and also have seen how they work in the code.

There are some things that should be improved in the building selection code, particularly where flavor evaluation is concerned as well. The goal of flavors in any case they are used is to help the AI vary it's strategic approach based on it's leader's personality. It's common for us to think of Traits ('leader promotions') as being personalities but they really are more subordinate to personalities, in the best case scenario reflective of the personality.

A leader's personality is programmed across all the tags in its XML. So when you select the game option: random leader personalities, you're selecting that the leader name and avatar many not be at all like he is originally programmed but rather his XML personality settings are randomized to some other leader's XML values. All minor distinctions but can be very critical for us as modders to understand.

The Flavors of a leader are one of the tags defined for his Personality profile. When a building is considered during a later stage step of priority than those immediate high priority buildings that get optimized for all of them, like production buildings that are cheap, for example, the flavors of the building are multiplied by the flavors of the leader (default is 0) to get a number that adjusts the overall value of the building. It's in this step that AIWeight is also added into the mix. The only applications we're using right now, mostly, of AIWeight, is to add a little to ensure the buildings don't come up with a 0 value if they are critical enough to demand the AI consider them worthwhile to build. But if AIweight is OVER utilized then the value of a building based on it's flavor matchups to the leader's flavor preferences, the weight will overwhelm and drown out the personalization value that flavors produce. There are other ways buildings gain value as well, by a valuation system that assigns weight to the values on tags. It's possible for a building to become negative in value if it's more harmful than good as well.

All that said, there's even more complexity than that to how buildings are ultimately selected for construction but that's some basics to keep in mind.

 

[Thunderbrd]

I guess Renaissance scientific revolution is simulated too strong .
Those buildings coexist for two eras and then Mathematics Academy is obsoleted.

Probably those buildings were designed by someone, who didn't beeline at all.
I don't beeline in my games too.

No. Beelining has nothing to do with it.

When buildings replace each other, they do so before the one being replaced is obsoleted in case you just never got around to building it for a while. This wasn't an accident. The accident was that the replacement is worse than the original.

And THIS is a RAMPANT problem in the Modern Era. This is only one of some 50 buildings that have this issue. And the reason for the problem is always the same, the tech bonuses were not taken into account.

Personally, I feel the tech bonuses are too strong. I also feel that the progressions should be charted out to ramp things up with much more care.

But making one not replace the other is returning to the issue of the original building being obsoleted at some point and when it does get obsoleted your civ loses it in every city and takes a significant hit, making the tech a bad thing to get. The point of replacement is to eliminate this sensation and allowing for a smoother transition.

Furthermore, as I stated before, allowing both to be concurrently operating in your cities leads to an even greater glut of research and we're already needing to tone it down a bit. It's the # of sources for research * the potency of those sources that is a bit problematic. If we tone down the amount of research each of those sources gives a bit, we will introduce improvements in balance.

Not that this is the only problematic commerce and/or yield. I think they all ramp up a bit too high, too fast and should really be toned down across the board but do so with a charted and better measured sense of progression.

I think Hydro, Pepper, and Azure, would all admit that they were setting up initial values that would eventually have to be rebalanced and tweaked before we'd have a solid sense of game balance in these later eras.

 

[pepper2000]

I think Hydro, Pepper, and Azure, would all admit that they were setting up initial values that would eventually have to be rebalanced and tweaked before we'd have a solid sense of game balance in these later eras.

Yes indeed.

This conversation has made me think about how to deal with the issue. We have the big spreadsheet of buildings and their categories, though I don't know how outdated it is by now. Maybe we could write a script that programmatically resets some of the building stats depending on its X value. That would be rather drastic, and that would be yet more regulation that would have to be enforced for all future work.

 

[Thunderbrd]

Yes indeed.

This conversation has made me think about how to deal with the issue. We have the big spreadsheet of buildings and their categories, though I don't know how outdated it is by now. Maybe we could write a script that programmatically resets some of the building stats depending on its X value. That would be rather drastic, and that would be yet more regulation that would have to be enforced for all future work.

I don't think we should script it out because that becomes too restrictive and wouldn't thereafter allow for choices to defy the norms - also doesn't do very well at adapting to tech enhancements to existing buildings, which has been one of the biggest design bugbears we've hit. HOWEVER, charting it on spreadsheets and establishing NORMAL rates of increase is absolutely what we should be doing.

Unfortunately these spreadsheets haven't been well updated with all the adjustments made to the tech tree so they'll have to be updated/recreated after the next, and hopefully nearly last, major tech tree edit project going on now.

 

[Thunderbrd]

@Dancing Hoskuld @Toffer90
A problem needs to be resolved in python with the captive system. When a captive is earned and the captive shows up on a Sea domain plot, it must check to see if it can load onto a friendly transport there. IF it can, it must be loaded right away. If it can't, it must be destroyed.

Fairly simple I think but I'm not proficient enough in Python to do it myself at the moment.

 

[JosEPh_II]

Turn 508 in latest game (Emperor Normal) 7 AI (not 6 ) , Just finished City Planning Tech Classical Era; turn times on my old i7 2600K 3.4 GHz cpu 4 cores 8 threads is now up to 3:51 at EoT wait. Brennus, Wrub, and now maybe Darius are at war with each other, and I am fighting stacks of Barbs on my little continent. Darius is already sending Galleons loaded with units toward Brennus. Both Darius and Brennus now have 27 cities each.

I have climbed out of last place to 5th place as I have been back filling my Empire so that Brennus can not plop a city(s) on my continent...yet. So I now have 22 cities with 6 of them new.

At the rate of EoT wait increase this game may not make it to Med Era before I'm forced to stop. I guess my days of playing up into Ren Era are over.

 

[raxo2222]

Putting an AI weight on a building is a last resort. In general it is bad practice.

AI is so lethal now thanks to all those AI weight of 1 (and DLL fixes), so we had reduce tech discount for AI.

It seems like flavors in buildings were ignored A LOT - they were used 2150 times, while there is 4000 - 5000 buildable buildings.