(1-VT) The AI should be more willing to buy the last copy of another player's luxury

[ma_kuh]

The recent shift in luxury resource valuation by the AI is refreshing, but has introduced some UI issues when attempting to engage the AI with deals.

The problem: It is cumbersome to know what the AI is willing to buy. It seems like there is an assumption by the deal AI that requesting the last copy of a resource from the player should never be attempted. This leads to a player with a happiness surplus needing to try and sell each of their 1-of luxury resources in turn to see which ones the AI might need.

Proposed solution: Remove restrictions on AI attempting to purchase the last copy of a player's resources when they would otherwise want to buy it. The AI is already prepared to give up their last copies of luxuries at higher rates, they should make the same assumption about the player and other AIs.

 

[Zanteogo]

I would only be ok with this, IF they give us some easy way to to know if the AI is asking for our last copy.

Currently the only way to do this is to remove resource and to see the total count. There are of course other ways, but it's the easiest way to do it from the trade screen. It would be a a pain to have to check this every time the AI asks.

 

[nekokon]

Personally I just hate the change to AI that made them refuse to buy extra luxury. One thing, it completely remove the option (instead of just increase/lower the price to discourage players from selling to them) which is a bad design in itself, and two, it's done for a very gamey reason "I don't want my ppl to be happy until they scream at me to buy a specific thing, so that when I bought them they would love me for 10 days"

 

[Zanteogo]

Personally I just hate the change to AI that made them refuse to buy extra luxury. One thing, it completely remove the option (instead of just increase/lower the price to discourage players from selling to them) which is a bad design in itself, and two, it's done for a very gamey reason "I don't want my ppl to be happy until they scream at me to buy a specific thing, so that when I bought them they would love me for 10 days"

There was talk about making the WLKD resource request check ALL luxury resources, including the ones you already have. I don't know what the current status of it is.

It would give a strong reason to collect as many luxuries as possible. Right now, there is a good reason not to, which like Nekokon said, is very gamey.

 

[ma_kuh]

I would only be ok with this, IF they give us some easy way to to know if the AI is asking for our last copy.

Currently the only way to do this is to remove resource and to see the total count. There are of course other ways, but it's the easiest way to do it from the trade screen. It would be a a pain to have to check this every time the AI asks.

That's a good point I hadn't considered. Is there an easy way where hovering over the luxury could say how many are remaining, similar to how gold/turn reads?

Personally I just hate the change to AI that made them refuse to buy extra luxury. One thing, it completely remove the option (instead of just increase/lower the price to discourage players from selling to them) which is a bad design in itself, and two, it's done for a very gamey reason "I don't want my ppl to be happy until they scream at me to buy a specific thing, so that when I bought them they would love me for 10 days"

Overall I think the direction is good; as a player you aren't generally going around spamming extra luxuries (I think?), so having the AI better match that "meta" seems fine. I think there's a strong case to be made that extra luxuries should be evaluated based on the value of happiness provided to the buyer, not just on whether they satisfy a strategic goal (WLTKD, City-state quest), but that would be beyond the scope of this proposal. I think there's some interesting design space around that.

 

[CppMaster]

There was talk about making the WLKD resource request check ALL luxury resources, including the ones you already have. I don't know what the current status of it is.

The last post of @Recursive about that was that he hadn't decided yet. I guess now it'd have to be a proposal.

 

[nekokon]

Overall I think the direction is good; as a player you aren't generally going around spamming extra luxuries (I think?), so having the AI better match that "meta" seems fine. I think there's a strong case to be made that extra luxuries should be evaluated based on the value of happiness provided to the buyer, not just on whether they satisfy a strategic goal (WLTKD, City-state quest), but that would be beyond the scope of this proposal. I think there's some interesting design space around that.

The design itself can be mitigated into "increasing/decreasing price to discourage/encourage the trade based on its need" which cover a much bigger picture than just outright disable it. Numbers can be tweaked, binary yes/no option can't.
I think the best option would be going for what Zanteogo said, make WLTKD check for all luxury resources so there's no more gamey reason to not do the best you can (grabbing extra luxuries when cheap/possible)

 

[Recursive]

The recent shift in luxury resource valuation by the AI is refreshing, but has introduced some UI issues when attempting to engage the AI with deals.

The problem: It is cumbersome to know what the AI is willing to buy. It seems like there is an assumption by the deal AI that requesting the last copy of a resource from the player should never be attempted. This leads to a player with a happiness surplus needing to try and sell each of their 1-of luxury resources in turn to see which ones the AI might need.

Proposed solution: Remove restrictions on AI attempting to purchase the last copy of a player's resources when they would otherwise want to buy it. The AI is already prepared to give up their last copies of luxuries at higher rates, they should make the same assumption about the player and other AIs.

I believe it would be better to solve the "AI is unwilling to buy luxuries because of a gamey reason" issue, as that's the root cause. I'll note that I plan to fix this for next version.

 

[Recursive]

Proposal vetoed.

Reason: Underlying AI issue has been fixed in the October 1 release.