Version 0.84 discussion thread

[jdog5000]

@ PieceOfMind:

Please check whether workers leave forests is on like Juju comments, that could certainly cause Ragnar to be way over doing it.

In your first example, I can understand the AIs decision: 2F 1H 1C can be a decent tile in many situations. Also, that tile is actually being used by the city just at the bottom of your screen shot, not the one you zoomed in on so the situation is a little different than you think. My bet is that lower city doesn't of any other good hammer sources besides the hill next to it and the forests, so that's probably the main motivation.

 

[r_rolo1]

But if the motivation is hammers and they can't chop they should build lumbermills and not preserves And even if the motivation is not hammers, IIRC the AI is coded to give more value to hammers than to commerce ( 6 to 4 , right ? ) so it might not be that... maybe the AI is overvalueing the possible from preserves?

 

[Fimbulvetr]

The issue here is that you get preserves long before lumbermills and watermills in that specific mod, workshop loses food and doesn't give much hammers yet, and I'm not sure if the AI correctly considers that a cottage can grow. I know it does, just likely not enough.
And features giving happiness can give a huge boost to improvement value. For a little light reading I recommend CvCityAI.cpp's CvCityAI::AI_getImprovementValue()
If there really is something broken, please find and fix it

 

[Munch]

I was pleased to see in my current game that AI Ragnar successfully performed an intercontinental invasion of AI Justinian, in the Renaissance. Anecdotal evidence for sure, but evidence of a mod well made, I think.

 

[PieceOfMind]

@ PieceOfMind:

Please check whether workers leave forests is on like Juju comments, that could certainly cause Ragnar to be way over doing it.

In your first example, I can understand the AIs decision: 2F 1H 1C can be a decent tile in many situations. Also, that tile is actually being used by the city just at the bottom of your screen shot, not the one you zoomed in on so the situation is a little different than you think. My bet is that lower city doesn't of any other good hammer sources besides the hill next to it and the forests, so that's probably the main motivation.

Hmm that would be a good explanation why it was always the active player who went crazy about the forests. I didn't pay attention to the options so there's a good chance that's what was happening. Anyway, you are right about the Khmer situation in that it's another city working the tile. It was not the best example and I realised a bit after posting, but in general my impression is that the AIs are using the preserve too often. If you know for a fact the way the forest preserve is valued is very solid code I'm happy to live with it, but given the values that the debug mode showed I thought it a bit suspicious.

I won't be able to check the code Fuyu mentioned or try without the "workers leave forests" options yet as I'm out of town for a few days.

I hadn't even realised the options could affect the workers under AI autoplay, by the way. Learn something new every day...

 

[Fimbulvetr]

The AI defines "having no trouble with happiness" a bit differently than a human player would, so a preserve as soon as monarchy will look AWSOME to the AI. More awsome than waiting for units to be built to make the city happier, definitely. You'll have to move it far back in the techtree again if you don't want that.

I believe the AI also correctly recognises when a feature is in the BFC of more than one city: in that case that improvement gets the happy value bonus from all cities (I believe, don't quote me on that), what could explain that high value.

 

[TheOnlyDJCat]

Just a small update - played two times multiplayer; now with BBAI .84 and BUG Mod 4.3 and absolutely no OOS. So far everything is perfect

 

[jdog5000]

Fantastic news, thank you for the update!

 

[PieceOfMind]

Just to confirm, I believe I probably did have "Workers leave old improvements" and "Workers leave forests" checked. I'll see if fixing this does the job, with some test games soon.

 

[Newcomer24]

I have a small issue with the latest version of the mod. The top 5 cities screen is messed up.. and I have never experienced anything like that before installing the latest ver of the mod. This is the only mod I use.

It's in italian but you should still be able to see how the ranking is all messed up and doesn't make any sense. It happens in every game I start.
It could be either BetterAi (unlikely) or the UP... I'm using the 48 civ dll.


Apart from that I have noticed some small changes which I guess are coming from the UP but still worth asking as I couldn't find them on the changelog... like having 3 food resources if you settle a city in a flood plain tile, which wasn't the case before.

 

[NotSoGood]

I have a small issue with the latest version of the mod. The top 5 cities screen is messed up.. and I have never experienced anything like that before installing the latest ver of the mod. This is the only mod I use.

It's in italian but you should still be able to see how the ranking is all messed up and doesn't make any sense. It happens in every game I start.
It could be either BetterAi (unlikely) or the UP... I'm using the 48 civ dll.


Apart from that I have noticed some small changes which I guess are coming from the UP but still worth asking as I couldn't find them on the changelog... like having 3 food resources if you settle a city in a flood plain tile, which wasn't the case before.

I can't see what's messed up. And the 3 food on floodplains is because settling on them doesn't remove the floodplains.

EDIT: Ah, all the sconosciutos. Now I noticed it.
EDIT2: But isn't that just Italian and means "Unknown"? So you don't know the city's name.

 

[Newcomer24]

I can't see what's messed up.

LOL just noticed I uploaded thre wrong screen, will upload a new one later today when I will get home.

And the 3 food on floodplains is because settling on them doesn't remove the floodplains.

But this doesn't happen in vannila civ, it only happens in Better Ai 0.84... I was under the impression that no gameplay changes are made in this mod (only Ai tweaks), so.. what's the strory?

 

[NotSoGood]

But this doesn't happen in vannila civ, it only happens in Better Ai 0.84... I was under the impression that no gameplay changes are made in this mod (only Ai tweaks), so.. what's the strory?

Better AI contains also Unofficial Patch. And UP is the one that changes that. IIRC that change was originally from Mongoose SDK.

 

[Newcomer24]

Better AI contains also Unofficial Patch. And UP is the one that changes that. IIRC that change was originally from Mongoose SDK.

I understand, I guess it was added in one of the latest versions of UP.
I find it quite unbalancing, giving a considerable advantage to those who start near flood plains, well.. probably not the right place to discuss this.

 

[PieceOfMind]

But this doesn't happen in vannila civ, it only happens in Better Ai 0.84... I was under the impression that no gameplay changes are made in this mod (only Ai tweaks), so.. what's the strory?

It was added with one of the versions of the UP (as NotSoGood noted). A couple of people have already questioned its suitability for the UP, including EF, Fuyu and myself. I'd suggest adding your "questioning" of it to the UP thread.

I agree it's a fairly significant gameplay change.

 

[Janov]

Just noticed some odd behaviour in my recent 0.84 game. A stack (landed by boat) was besieging my city. While the catapults would bombard the defence-value every turn, some "fast" units (chariots) would attack the city (against fortified pikemen in a walled city with 50ish defense bonus, you can guess the outcome). This obviously self-sacrificial behaviour seems even dumber than the general "lets attack, who cares about the odds?" behaviour, maybe it has to do with the recent change in 0.84 (fast units pillage first)?

Also noted that combat odds are displayed weird for me, if attacker´s strenght = defender´s strength shouldn´t it be like 50%? Yes, all multipliers factored in, no first strikes involved... For me it shows like 35% or so.

Jan

 

[Yakk]

I've witnessed multiple workers on the same tile, working on different improvements.

(AI empire workers, in particular).

 

[PieceOfMind]

Also noted that combat odds are displayed weird for me, if attacker´s strenght = defender´s strength shouldn´t it be like 50%? Yes, all multipliers factored in, no first strikes involved... For me it shows like 35% or so.

Jan

Can you provide a screenshot of that happening?

 

[Janov]

Can you provide a screenshot of that happening?

I will work on that - have to find out whats the hotkey, first.

Jan

 

[technomancer]

Just noticed some odd behaviour in my recent 0.84 game. A stack (landed by boat) was besieging my city. While the catapults would bombard the defence-value every turn, some "fast" units (chariots) would attack the city (against fortified pikemen in a walled city with 50ish defense bonus, you can guess the outcome). This obviously self-sacrificial behaviour seems even dumber than the general "lets attack, who cares about the odds?" behaviour, maybe it has to do with the recent change in 0.84 (fast units pillage first)?
Jan

Happened to me as well. A very small stack of knights attacked one of my heavily fortified cities (without pillaging first) while the main army (no fast units) appeared a couple of turns later. The knights were crushed of course.
This behaviour is indeed new and dumber than before, I also suspect it might have something to do with "fast units pillage first", never noticed this before.
Unfortunately I deleted the savegame, so no screenshot.

techno