[Bouncing ideas for the far future of the X series] Director AI

[pittlebelge]

Hey, so how would you all feel about having a Director AI in an X game (if there isn't one already to some extent)? For those who don't know, a director AI functions like a game master in a table top RPG. The Director will monitor the progress of the player and adjust loot, enemy encounters, tip the odds in random events and so forth. It's an invisible hand that pushes on the scale to create difficulty spikes and respite time.

In the X series, it could pay attention to manual trade order given by the player and order the AI fleets to compete harder on the same products or give the player a break if money is tight. It could tweak the cost of Xenon ships to have them overproduce or slow down on their conquest. I could target the player's asset more directly by sending fleets to attack the player stations and ships.

I'm not sure if having such an AI behind the scene would be a good idea or not. We can see the faction's economies are already quite fragile, and the AI's meddling could ruin them. I'm also certain that everyone would not necessarily enjoy the push and pull it could create. I know you guys and you'd find ways to use it to manipulate the market. But I also suspect that if handled correctly it could create late game challenges of different nature that could keep the experience interesting way past the point where money is no longer relevant to the player.

So, what are your thoughts ?

[rudi_pioneer]

I like that X universe run on its own, however I can see some players enjoying director AI.

[Feloidea]

I would be strongly against any such game master system to increase the game difficulty. What's the point in being punished for success?

I would however see it as one potential avenue to make factions behave more organically with changing relation to other factions by say one level of reputation (say from neutral to either friendly or belligerent, friendly to either neutral or allied or from belligerent to neutral or war) depending on how the relative power of the two matches up (if faction A is currently belligerent to faction B and the latter's military power declines sufficiently, faction A could seize the opportunity to declare war and grab territory).

[TroubledRabbit]

what you have described is quite easy to archieve by semaphores and pseudo-decision trees (which are present in some form in the game), my students usually do that learning how to make 'bluffs' in card games they write during elementary programming courses. They could use ofc. machine learning but that would be cheating at this level of their education and disproportionately costly. There is enough 'get faster CPU' bad attitude already around. What the game could use (existing or future one), at least for some players enjoyment, is contextual 'quest generator' which does need to have the 'quest form'.

But I would be more than happy if the actual game actually would realize its own potential as it is without strange (and tbh lame in execution) yogas of 'existential crises', which are still hidden and left unused or just half-baked - the 'data vaults' are maybe the simplest example.