Typo Affecting AI Yields

[stealth_nsk]

Does that mean that there were no foreign key checks in the AI sql schema? That seems like a hilariously stupid thing to not do.

It's not real RDBMS, that would be too performance-expensive. A lot of games use relational data without having full RDBMS (Total War games, for example). Although I'd spend some time to develop data sanity check tools for developers and modders to be run outside the game.

 

[Siesta Guru]

I am reading this all with considerable interest. Hopefully catching and fixing this error will result in a more competitive AI.

But it does leave me with another question: what exactly should the AI be doing?

There are some very specific things I wish the AI would not do (prioritise religion, or build cities, districts or some wonders in bad spots) and a handful of very specific things I wish it would do (build its unique units, build cultural specific wonders). But otherwise I don't really know what specifically I should expect the AI to do other than just 'do things better'.

e.g. Does the AI need to build more mines and IZs? Just build more troops overall, or a particular mix of troops?

Ideally it should specialize and just based on in what happens in game.
The reason players are much better even at the peace game is because they do this. A player knows when he can ignore culture and gold for a bit to get all universities out asap. A player knows that if under attack, youll need a certain amount of troops. If you dont have that, you should probably build more. Players prepare for future events, like finishing a tech, by walking workers into position so that they can chop trees the moment the tech pops to get that essential wonder.
Meanwhile, the AI just has a flat desire for 'wonder', and doesn't even have key wonders in mind. This is why ultimately, I think the approach of just shifting flat desires around is futile. Yes better balances among yields make the game a little more challenging, but it's never going to make it feel like the AI is actually thinking and is competing with you. At any point the player is in full control and knows exactly what to expect

 

[Infixo]

It's not real RDBMS, that would be too performance-expensive.

Just out of curiosity, what is SQLite missing that would make it „real RDBMS”?

And, btw, the engine does constraint checks after DB is built. From what I see in schema, references are added whereever possible i.e. all fields referencing PKs are properly defined.

 

[Trav'ling Canuck]

But it does leave me with another question: what exactly should the AI be doing?

Ideally it should specialize and just based on in what happens in game.

Absent that, I'd argue that each civ should simply behave differently, i.e. specialize regardless of what's happening in game.

With a random map and a random assortment of 7 AI, I'd expect the player would find winning more difficult if those 7 AI took different approaches. Clearly it would be better if they selected their approach based on the in game circumstances, but even having two civs focused on generating science, two on generating culture/tourism, two on building and using their military aggressively, and one on spreading their religion, even if those choices are sub-optimal for their current circumstances, would create the potential to pose more of a challenge.

 

[acluewithout]

Agreed.

Honestly, if the AI just built by default Commercial Hubs and either a Campus or Theatre District in a triangle with its city centre, or Harbour, Commercial Hub City Centre triangles, that would be a vast improvement over what the AI usually builds.

 

[FearSunn]

Just to draw attention to another "typo" fixed already with this mod:
https://forums.civfanatics.com/threads/ai-strategy-conditions-fix.630326/

From what I see in the database this error has much more severe effects than "yeilds".

 

[Abaxial]

Absent that, I'd argue that each civ should simply behave differently, i.e. specialize regardless of what's happening in game.

With a random map and a random assortment of 7 AI, I'd expect the player would find winning more difficult if those 7 AI took different approaches. Clearly it would be better if they selected their approach based on the in game circumstances, but even having two civs focused on generating science, two on generating culture/tourism, two on building and using their military aggressively, and one on spreading their religion, even if those choices are sub-optimal for their current circumstances, would create the potential to pose more of a challenge.

But that happens, does it not? My experience is that if any AI is going to sneak in a science victory, it will be Brazil or China, while Greece and Russia spam great works and culture the whole time.

 

[Trav'ling Canuck]

But that happens, does it not? My experience is that if any AI is going to sneak in a science victory, it will be Brazil or China, while Greece and Russia spam great works and culture the whole time.

Yes, the AI does pick a victory condition and that does affect it's behaviour. And the AI can't specialize too much because so much is interconnected in Civ 6.

My sense, though, is that the AI doesn't go far enough in it's focus on different victory routes, or doesn't specialize early enough. But maybe it's simply the subsequent individual decisions that let it down.

 

[MrRadar]

Just to draw attention to another "typo" fixed already with this mod:
https://forums.civfanatics.com/threads/ai-strategy-conditions-fix.630326/

From what I see in the database this error has much more severe effects than "yeilds".

Oh my... This almost deserves a thread of its own or at least a report in the 'Bug reports' section.
Well done!

 

[FearSunn]

Oh my... This almost deserves a thread of its own or at least a report in the 'Bug reports' section.
Well done!

As suggested I've just started separate thread for this
https://forums.civfanatics.com/threads/another-error-affecting-ai-medieval-error.630469/

 

[Stringer1313]

Agreed.

Honestly, if the AI just built by default Commercial Hubs and either a Campus or Theatre District in a triangle with its city centre, or Harbour, Commercial Hub City Centre triangles, that would be a vast improvement over what the AI usually builds.

Someone should test this...