[We the People] Bug reporting

[raystuttgart]

It's most likely triggering on the river.

It became valid just because of the Bonus Resource "Salt". <-- This is the actual problem.
(But still it should not trigger inland but just on the Feature "Salt Drainage" which is coastal only.)

The question is how to fix this.

It can be fixed by the XML of the Improvement.
I will take care of it.

Edit:
It is fixed.
Thanks for reporting.

 

[jooe]

[version 4.0 github latest]
The results of CvCity::bestTeachUnitClass seem to be off. I get a high frequency of "Expert Cochineal Collector" even in native cities that cannot produce the yield cochineal at all.
The formula is quite complicated and honestly I cannot make much sense out of it ... maybe someone with a better understanding what happens here could have a look at it?

 

[raystuttgart]

The results of CvCity::bestTeachUnitClass seem to be off. I get a high frequency of "Expert Cochineal Collector" even in native cities that cannot produce the yield cochineal at all.
The formula is quite complicated and honestly I cannot make much sense out of it ... maybe someone with a better understanding what happens here could have a look at it?

That is not a bug - that is pretty normal. E.g. you may end up having "Seal Hunters" (in the middle of the map) because the algorithm has some randomness in there. (Which is good.)
The algorithm uses "weighting" but it also has some "randomness". This guarantees to also have chances of getting some Experts that would otherwise be very unlikely.

So please let us not make a bug out of this just because you got results you do not understand.
Due to the high amount of Experts we have right now there sometimes may be results like this.

Trust me - the algorithm is not bugged - it is actually working nicely since 10 years.
It simply creates results the player may not always understand.

-----

By the way:

The algorithm also uses the information in

CIV4CivilizationInfos.xml
<TeachUnitClasses/>

While it uses the information of Yields available in the Native Village it also tries to create other results from that file.
(It e.g. counts how often it already created certain Experts and will then decrease their weighting.)

 

[jooe]

I know it is calculated by a weighted randomness. But since the introduction of the new units/professions, in all my games villages teaching "expert cochineal collector" (ECC) were very common and others are really hard to find.
In my current game, i have e.g. 9 villages teaching ECC, most of them having no plot with cochineal yield at all. On the other hand, I have just 1 village for each of expert farmer and expert fisherman.
Other experts are also associated to the surrounding terrain, but not ECC, which appears in all terrains - that's why I thought there might be something wrong.

// edit: could it be caused by the fact that most yields have -3 yield in terrains where they cannot be produced, while cochineal only has +3 in shrublands and is not mentioned otherwise?
pull request here: https://github.com/We-the-People-civ4col-mod/Mod/pull/681

 

[raystuttgart]

... in all my games villages teaching "expert cochineal collector" (ECC) ...

That is the problem with randomness.
Sometimes it creates results that seem to have a "systematic" that repeats.

But if you e.g. look into the XML:

UNITCLASS_COCHINEAL_COLLECTOR
iWeight = 2

UNITCLASS_TOBACCO_PLANTER
iWeight = 3

So e.g. the Tobacco Planter has a higher weight than the Cochineal Planter.

----

You are deriving a bug because randomness has created a result that seems odd to you.
Which may happen but it does not mean that there is a systematic issue.

 

[raystuttgart]

Ok, I found the bug with the Cochineal Collector that messed up CvCity::bestTeachUnitClass.
It is in XML and I will fix it right away. Then there should also be no Problem anymore.

It was simply a misconfiguration in the XML.
(It was not related to the algorithm but specifically to the Profession Cochineal Collector.)

------

Sorry for doubting the bug, but since you pointed to the algorithm and I was pretty sure it was correct I did not believe in that bug.
But the bug was simply in the XML of the Profession which in the beginning I did not consider at all.

So again sorry for doubting.
Thanks for reporting the bug.

------

Ok, here is the correct fix for the problem:
(Of course it requires a new game.)

Fixed bug in balancing of Profession Cochineal Collector that caused … · We-the-People-civ4col-mod/Mod@7117d91

…issues to Teach Units.

github.com

 

[jooe]

[version 4.0 github latest]
For some reason, the AI does not found their first settlements. It just sends the starting naval unit with the settler loaded along the coast.
// edit: Note I have the new movement cost algorithm activated. But the AI didn't have difficulties with that until last week. Maybe the update to pathfinding confused the AI?

 

[raystuttgart]

Maybe the update to pathfinding confused the AI?

I have seen several games in the past where AI took 5 or 6 turns to find a good location for the first settlement.
(Especially on small random maps it often does not want to settle too close to the Natives.)

Please try to play 10 turns of Autoplay.
Then if AI has still not settled let us know.

Maybe also open WB then and take a few Screenshots of the Coast.
(AI tries to find a place with Food and if there is e.g. no such good place it may continue to search.)

 

[jooe]

New Game (Americas Gigantic), Jan 1494: All AI players have moved their starting units, most of them are on coastal plots. None of them has unloaded the settler.

 

[raystuttgart]

Yeah, we also just had a report internally that the AIs do not embark anymore with "New Movement System".
(Due to a bugfix for something else it is broken again ...)

So yeah you are correct - the bug was confirmed - and it is specific to "New Movement System".
Please play "Classical Movement System" until we notify that we have fixed again.

But I will rather wait for @Nightinggale to check this, because he wrote the code for the Pathfinding of "New Movement System" and understands it best.
(If I would mess with it chances are high that I would break more that I would repair - and I do not want to create overhead effort.)

 

[jooe]

Yeah, we also just had a report internally that the AIs do not embark anymore with "New Movement System".
(Due to a bugfix for something else it is broken again ...)

So yeah you are correct - the bug was confirmed - and it is specific to "New Movement System".

I think I managed to fix this, but please double check my suggested fix:

Fix for path finder so AI can unload their units again by jo-oe · Pull Request #684 · We-the-People-civ4col-mod/Mod

Please double check this is the correct fix. It was a wild guess ...

github.com

 

[raystuttgart]

I think I managed to fix this, but please double check my suggested fix:

Your new version of the fix was confirmed to work nicely.
Thanks a lot for taking care of this.
It is highly aprreciated.

 

[jooe]

[4.0 github current, new movement system]
The path displayed by a mouse goto (right click) has something like an off-by-one error. The last move possible during one turn is already shown as belonging to the next turn.

I will try to look into this myself, but won't have time for a debugging session until at least thursday. Just wanted to post it already to not forget about it

 

[raystuttgart]

The last move possible during one turn is already shown as belonging to the next turn.

Yeah, I noticed that as well, it is a remainder I think of the originally showing 0 while you still had movement points and 1 at end turn.
And then after we had the "division by 0" bug it was all increased by +1 so it is now a bit misleading / wrong.

I will try to look into this myself, ...

That would be nice. I don't know if @Nightinggale currently has time.
Otherwise he surely knows what to do about this since he implemented it.

 

[Nightinggale]

Yeah, I noticed that as well, it is a remainder I think of the originally showing 0 while you still had movement points and 1 at end turn.
And then after we had the "division by 0" bug it was all increased by +1 so it is now a bit misleading / wrong.

Yeah since the game started crashing due to dividing by 0, I did the quick fix of just adding one to avoid the crashes. At some point I need to go back and figure out what precisely to do about this, but since it's not on by default in the next release, I'm currently focusing on what is included by default in the next release.

 

[codiac]

Hi, New Hope latest, if you have lush grass you can farm chickens, but you can't build a pen.

 

[raystuttgart]

Hi, New Hope latest, if you have lush grass you can farm chickens, but you can't build a pen.

It is just one of small inconsistencies that resulted from the many balancing changes.
That is easy to fix.

Edit:
It is corrected now.
Thanks for reporting.

 

[jooe]

[4.0 github current, new movement system]
The path displayed by a mouse goto (right click) has something like an off-by-one error. The last move possible during one turn is already shown as belonging to the next turn.

I will try to look into this myself, but won't have time for a debugging session until at least thursday. Just wanted to post it already to not forget about it

A pull request for this is currently in review: https://github.com/We-the-People-civ4col-mod/Mod/pull/690
btw I really like the new movement system

 

[raystuttgart]

A pull request for this is currently in review:

@Nightinggale knows about it, but I think he was busy yesterday.

btw I really like the new movement system

Most people that gave it a try really like it.

Getting that positve feedback also confirms my decision to take the risk to implement it.
Because it was kind of an "experiment" and I could have wasted my effort by having to remove it again.

Still is pretty funny to me actually how negatively people often react to most of my proposals.
And once I implement them anyways the overwhelming majority then suddenly loves them.

 

[Nightinggale]

@Nightinggale knows about it, but I think he was busy yesterday.

Sadly it looks like I have ended up with a fever (nothing serious) and will likely be out for a few days
I will get back to this ASAP, but it will be "when I'm ready".