Jeppetto
Emperor
- Joined
- Dec 23, 2018
- Messages
- 1,021
Tourist: The jeans and pop music was nice, but the forest fires proved to be minor yet non-ignoreable inconvinience. 5/10 stars, average experience.
Civ VI Funny/Strange Screenshots
- Thread starter Carazycool
- Start date
Tourist: The jeans and pop music was nice, but the forest fires proved to be minor yet non-ignoreable inconvinience. 5/10 stars, average experience.
Deadly Dog
Prince
That qualifies, in my mind, as an unnamed Natural Wonder of the World. It's certainly better than more than a few actual civ6 Natural wonders.
I had a start once with Kupe where I found an island with 5 Turtle reefs.....that inspired the concept in my mind of "unnamed" Natural Wonders....city locations so good they shape your game.
These barbs really do not like gilgamesh (they have already wiped one of his cities)
The rest of the area was full of bananas too...utterly crazy start
The rest of the area was full of bananas too...utterly crazy start
Hakan-i Cihan
Emperor
Afaik districts from other players do give adjecency bonus. But why it is +3 and not +4 ( +2 river, +1 2 districts, +1 G Plaza) I do not know.
Pfeffersack
Deity
Now one should say again that a flat snow tile is the worst you can get in civ...
(Game not played in Apocalypse mode, Standard speed, disaster setting 2)
(Game not played in Apocalypse mode, Standard speed, disaster setting 2)
Last edited:
Naeshar
Prince
- Joined
- Oct 22, 2016
- Messages
- 384
I think its the other way around, if it says +1 from Government Plaza, then it doesn't count it as a district for adjacency, but gets the bonus from any government plaza (check the bonus on far side of the plaza). If you check the description on any adjacency related district it tells you the bonus is connected to the receiving district not plaza giving the bonus. And it's a separate bonus, while giving it to another civ is probably an oversight bug.
I also thought the districts from CS count as adjacent but they don't. Only CS improvements help, like mines or quarries.
7daysofwar
Warlord
- Joined
- Nov 24, 2018
- Messages
- 141
Well, I guess can make use of those mountain tunnels now
kaspergm
Deity
- Joined
- Aug 19, 2012
- Messages
- 5,833
This screenshot is absolutely glorious. It says so much about the AI of this game.Spoiler :
Pedro showing off his Huey placement skills.
shaglio
The Prince of Dorkness
To be fair, it’s probably not as easy to fix as it sounds. The game already has a line of code to check if there is an available lake tile to build the Huey Teocalli (if there isn’t, it shows up grayed-out in the build queue even for human players). But then you’d also need a check to see if there is a multiple-tile lake within the city borders. But what if there is a nearby multiple-tile lake, but only one of its tiles is in the city borders and the rest belong to another of that civilization’s cities (That would still be a reasonable place to build HT)? Or what if there is a multiple-tile lake nearby, but only one tile has been acquired by border expansion yet? Even if there is a multiple-tile lake with more than one of its tiles within the city’s borders, there would need to be code to tell the AI which tile to build it on (random tile, perhaps?). But then what if there was both a multiple-tile lake and a single-tile lake within the city’s borders such as in the screenshot above? The randomly-chosen lake tile might wind up being the single-tile lake. So there would need to be code to prevent that as well.
We’re talking about a lot of lines of code just to tell the AI where to properly place one World Wonder. People love to crap all over the AI of this game, but I can’t imagine it’s easy to take into account every single variable of every aspect of a game on such a grand scale as this.
tedhebert
Emperor
Being an IT professional, I understand what your saying, but...
We're talking about a WONDER here... That is a very important and very costly endeavour in CIV; I feel that special attention should be applied to it
I usually tend to defend FXs actions and decisions, but this, to me, stinks of either lazyness, or more likely bad project management decisions during the development process ( a la 'we don't have the time or money to spend on THIS little detail').
anyways, sorry guys, this is not the right thread for this kind of comment ;-(
Jeppetto
Emperor
- Joined
- Dec 23, 2018
- Messages
- 1,021
My bet is that it's just bug with city build & plot order management.
It may not be the easiest task to do properly, but shaglio also exagerates a bit, especially saying it's all work for one wonder when every regular programmer would see most of it can be boiled down to generalizable value evaluation that can be used for all wonders, even districts etc - telling AI not to build it on isolated lake isn't "lines of code", it's literally just number value assigned to each plot based on adjacency, not a rocket science, not long code - the game already does similar evaluation when deciding which plot to take by Culture, it assigns Cost to plots (or at leats did so in V) and the Cost is based on what the plot holds, distance, rings, adjacency etc.
The hard part is to implement it in way that would be optimal. Pretty sure If they really wanted they could make hella smart AI that makes smart moves every turn by evaluating practically everything as things change, but who wants to wait 1+ minutes for their Turn. Someone in YT section already mentoined that they noticed that units in war seem to move based on strategy established prior to it, so they saw enemy civilian units casually walking through their territory because they were passing city that was not there before. It's possible some of these decision-locks had to be implemented because the tests showed the game can't handle doing too much decision-making too often. Or maybe they're lazy and went with limited implementations.
My fear would not be lines of code but process time. You could easily determine value of each placement-sensitive wonder. In case of Huey, you literally just count amount of lake plots say within 6 plots of each city, dynamically, increasing Attractivness of it which would determine whether AI will consider building it, then each lake plot would have value based on adjacent lakes so that AI would recognize most lucrative choice. There goes the part that all of this is calculations for one Wonder and they stretch far If one wants AI to plan into future (settle city specifically because of "lake value", planning to build Huey), ideally each AI would have to do many of these evaluations constantly. Not programmer work for one wonder, but process time for one wonder. Whether it's more possible than I imagine or not would be better said by optimization experts, optimization really never was my cup of tea, but my point generally is that it's question of what the machine can handle rather than lines of code. In terms of lines of code, the system can be boiled down into being pretty simple yet at least averagely effective by just plot values.
Back to explanation, I would guess that exactly the AI already has established plan for what it wants to do with the plots and so it considers the lake plots to be "taken". It that way may be matter of wrongly placed priorities. Or perhaps, ironically, the AI determined that it needs not the Amenities and prefers the Fisheries to be clumped for safety or plans to build TS adjacent to the Huey, but that may be me overestimating it.
I would note though that AI acting properly should not be put aside as not worth the invesment, since it's huge part of the game. The game seems more focus on Single-Player and it's not Cities Skylines, so decent AI should've been priority early in development with all it brings.
It may not be the easiest task to do properly, but shaglio also exagerates a bit, especially saying it's all work for one wonder when every regular programmer would see most of it can be boiled down to generalizable value evaluation that can be used for all wonders, even districts etc - telling AI not to build it on isolated lake isn't "lines of code", it's literally just number value assigned to each plot based on adjacency, not a rocket science, not long code - the game already does similar evaluation when deciding which plot to take by Culture, it assigns Cost to plots (or at leats did so in V) and the Cost is based on what the plot holds, distance, rings, adjacency etc.
The hard part is to implement it in way that would be optimal. Pretty sure If they really wanted they could make hella smart AI that makes smart moves every turn by evaluating practically everything as things change, but who wants to wait 1+ minutes for their Turn. Someone in YT section already mentoined that they noticed that units in war seem to move based on strategy established prior to it, so they saw enemy civilian units casually walking through their territory because they were passing city that was not there before. It's possible some of these decision-locks had to be implemented because the tests showed the game can't handle doing too much decision-making too often. Or maybe they're lazy and went with limited implementations.
My fear would not be lines of code but process time. You could easily determine value of each placement-sensitive wonder. In case of Huey, you literally just count amount of lake plots say within 6 plots of each city, dynamically, increasing Attractivness of it which would determine whether AI will consider building it, then each lake plot would have value based on adjacent lakes so that AI would recognize most lucrative choice. There goes the part that all of this is calculations for one Wonder and they stretch far If one wants AI to plan into future (settle city specifically because of "lake value", planning to build Huey), ideally each AI would have to do many of these evaluations constantly. Not programmer work for one wonder, but process time for one wonder. Whether it's more possible than I imagine or not would be better said by optimization experts, optimization really never was my cup of tea, but my point generally is that it's question of what the machine can handle rather than lines of code. In terms of lines of code, the system can be boiled down into being pretty simple yet at least averagely effective by just plot values.
Back to explanation, I would guess that exactly the AI already has established plan for what it wants to do with the plots and so it considers the lake plots to be "taken". It that way may be matter of wrongly placed priorities. Or perhaps, ironically, the AI determined that it needs not the Amenities and prefers the Fisheries to be clumped for safety or plans to build TS adjacent to the Huey, but that may be me overestimating it.
I would note though that AI acting properly should not be put aside as not worth the invesment, since it's huge part of the game. The game seems more focus on Single-Player and it's not Cities Skylines, so decent AI should've been priority early in development with all it brings.
Pfeffersack
Deity
Poor Peter probably doesn't like himself…
Spoiler :
Spoiler :
DWilson
Where am I? What turn is it?
No, no, Peter likes himself fine. He merely liberated Paris from corrupt French tyranny.
And here I thought Wales was a land of sheep.
At least the part of the English occupying the land is accurate.
