1.6 feedback

[Slvynn]

I think the only way for you to believe me is for you to try it. Please modify whatever parameters you like and play a game. If you see an improvement with no side effects, I will put it into the release.

I have not enough time today because i have to build graphic release for patch.
I hope i get to it after release and will make testing.
How can i make autoplay turns? I've never made it.

 

[davidlallen]

How can i make autoplay turns? I've never made it.

Use "ctrl-shift-x".

 

[Slvynn]

Use "ctrl-shift-x".

Thnx, i will do it morrow, i still have lots of work to do with graphic release.

 

[Ahriman]

Huh. I am not arguing with you, because you are arguing.
I not telling it should be white or black. I wont bad or bad choice. There is a balance. Ai will prioritize Thopter as domestic unit. As a settler. It have already this ai. The problem is ver yclear. It have also military potence. If we remove it it will make things much better, closer to balance.

There is no logical argument in this statement.
The problem is NOT clear. You assert that the AI builds too many thopters, based on observing that one AI had two of them once.
You assert that making them non-military units will make things better and more balanced, based on a hunch.

You can't just assert your opinions as fact.

It seems much more likely to me that the problems are:
1) Mahdi fighters too powerful.
2) Early game wonders too expensive, too concentrated in particular techs. Possibly AI leaders like building wonders too much.

Possible solutions are:
1) Reduce their strength to 5 (from 6). Leave them at ~3/4 the cost of Bladesmen and no upkeep. Allow them to upgrade into hardened bladesmen?
2) Redesign early game wonders, and make them so that they are spread better over the techs. *Maybe* reduce AI leaderhead preferences for pursuing wonders.

*edit*
For reference: city raider mahdi fighter = strength 7.2
City raider bladesmen vs city = strength 6.5, and at a higher cost and lower strength outside cities.

 

[Slvynn]

There is no logical argument in this statement.
The problem is NOT clear. You assert that the AI builds too many thopters, based on observing that one AI had two of them once.
You assert that making them non-military units will make things better and more balanced, based on a hunch.

You can't just assert your opinions as fact.

It seems much more likely to me that the problems are:
1) Mahdi fighters too powerful.
2) Early game wonders too expensive, too concentrated in particular techs. Possibly AI leaders like building wonders too much.

Possible solutions are:
1) Reduce their strength to 5 (from 6). Leave them at ~3/4 the cost of Bladesmen and no upkeep. Allow them to upgrade into hardened bladesmen?
2) Redesign early game wonders, and make them so that they are spread better over the techs. *Maybe* reduce AI leaderhead preferences for pursuing wonders.

Logic is not necessary requrement to "being right".
I have alot of facts, from my played games. I sure that BTl have alot of thopters around flying as well. Please, david - can you check a worldbuilder?
The problem that defences are low for such a great, but border city.
I am not one of logical chains. I am one of feeling and intuition and experience. Before we tried to make a change and see how it fares we cant talk in such denying tone.
I'll do my testing. I also not lying/being stupid when i telling to all of you that this is issue. I saw and weighted things. There is a problem with "too much thopters", and "weak defences" because lots of thopters built early in game. Thats a 30 hammers. And more - MAIN POINT - AI wont build more infantry because it have 3 thopters. That is bad.
Infantry > 3 thopters. In any way. Hammer cost. Unit specific Cost. Please understand, that AI counts Thopters as defence-proper unit. Is it proper? How much AI need thopters to spread well? This things should be checked and questioned before implemented, but they need to be checked/questioned indeed.

 

[Ahriman]

Logic is not necessary requrement to "being right".

Logic or evidence are a necessary requirement to get people to believe you when you claim you are right.

If they build multiple wonders, that took hundreds and hundreds of hammers. How can you really think that the problem is 40 hammers worth of thopters?

 

[Slvynn]

Logic or evidence are a necessary requirement to get people to believe you when you claim you are right.

If they build multiple wonders, that took hundreds and hundreds of hammers. How can you really think that the problem is 40 hammers worth of thopters?

Not multiple, it appears that only 2. Not so much, but thats not an issue.

Again - please understand. THATS NOT A POINT . Aall things you talking about not getting to point - which i explain to You, quite logically.

AI building thopters also have his own defence/power strenght numbers raised. So it will just build more "universal" (as he see) cheap thopters as they can spread settlers, scout and military potent, instead more few infantries. Thats wrong. Ai overestimate its military capabilities with thopters. The thing should be balanced. AI should threat Thopter as unit for 2 puproses : Spreading settlers and scouting. Thopters, stationed in cities should not raise defence level/ own power ratio esimation for ai.
Get to the point, please.

As for logic - its not. Logic can be wrong, when it is blind. I, also logically - explaining things and you just reading parts of my text but not seeing the meaning, the whole...

Experience is much more than plain, experience-less theoretical logic. Also perception to details.

Please, sorry if i am harsh abit, but i do try to make better, and when i am not sure i just pass, saying "ok, you have a great point". When i am sure, i going to the "end" . Please dont deny things that are have own base and facts, as explained. I cant throw countless old games with thopters galore, but i "felt" it, observed, and found, that AI can fare much better if it build more balanced amount of thopters. Why we didnt built 6 thopters? We were able to settle pole with 3 thopters. Why 6 then. Also, we , as human see thopter in different way as AI, but AI, for now, sees it as "military" unit.

 

[Ahriman]

THATS NOT A POINT

Sure it is. If they didn't spend hammers on Wonders, they could have spent them on other things, like more military units. I dispute your assertion that they wouldn't have built any more military units if they'd had 400+ more hammers.

 

[davidlallen]

And more - MAIN POINT - AI wont build more infantry because it have 3 thopters. That is bad.

Let's all stay calm here, please. I think we all agree the AI is not as smart as a human player. After Slvynn does the building art release, I hope he will experiment with the xml settings. Then I believe he will agree that there are no simple fixes. We need to study the AI behavior carefully across multiple games, figure out why it is making certain bad decisions, and change the AI code to improve them. I do not think it is an easy fix.

 

[Slvynn]

Let's all stay calm here, please. I think we all agree the AI is not as smart as a human player. After Slvynn does the building art release, I hope he will experiment with the xml settings. Then I believe he will agree that there are no simple fixes. We need to study the AI behavior carefully across multiple games, figure out why it is making certain bad decisions, and change the AI code to improve them. I do not think it is an easy fix.

Thank you. No need to argue. Arguing fuel more arguing. Denying fuel more pressure if one who denied know, that he have a valid point. The problem exist. There is a point, and we cant be blind to it. Thopters count towards AI estimation of own power ratio.

The thing where i mistaken - i told that is wonderspammend, in fact it have 2 wonders, on of them shrine. And i confirm my mistake here.


Also, i see some numbers are really wrong in Thopter definition. I already browsed meaning of them in some xml help files, and i think this one is simple to fix . That will be my focus after patch release. And i have feeling that i am right here
Also , sorry Ahri... I am bit tired after job, but please understand - no need to deny and crush ones observations without even understanding main point/ trying. I know, i have a very solid point. And, perhaps, we can have easy fix for this issue. Lets hope we will. Why denying it ?

 

[Deliverator]

I wouldn't read too much into a single case of there being two scout thopters in the city when I attacked it in the succession game. I walked my stack over ground so it was probably just withdrawing its transports to the safety of the city. I'm pretty sure that they weren't stationed there and being used as city defenders - they moved back there a turn or two before the attack. In my opinion, it is a big jump to conclude that this is evidence of bad AI decision making.

 

[Slvynn]

I wouldn't read too much into a single case of there being two scout thopters in the city when I attacked it in the succession game. I walked my stack over ground so it was probably just withdrawing its transports to the safety of the city. I'm pretty sure that they weren't stationed there and being used as city defenders - they moved back there a turn or two before the attack. In my opinion, it is a big jump to conclude that this is evidence of bad AI decision making.

I played alot of ames and observed this matter. With focused observation. I made a note of it. And now i see BTl again with those thopters and poor defences. I had few games with real thopter galore. And i just told that it it matter to observation, and i want to check that.
The result i got - i've been repulsed as incompetent before i even tried. Thats all rumble over here, and i dont like it's course. I think i will pass.
Main issue the way i being denied without any chance of experiemental observation and understanding of my point, which is also valid.
Thanks David for very correct position here. I am not pressing ot relese such a thing right now. All i telling that it worth experiement. I will do that. And i see no point that my intension breing repulsed in such a form, as it did Ahriman. Thats a major point of argue, more than a point that Thopters count toward AIs power ratio.
Also , some things explained in the spice thread.
My point is valid - that thopter strength counts toward Military power. But i think it should not and that may/will give an improvement to all-over AI's faring.

 

[Ahriman]

Thank you. No need to argue. Arguing fuel more arguing. Denying fuel more pressure if one who denied know, that he have a valid point.

Wow, just wow. We should all stop arguing with you because you just *know* you are right.

Thopters count towards AI estimation of own power ratio.

Doesn't the power ratio depend on unit strength, not just unit numbers? So strength 1 units won't count for very much.

The result i got - i've been repulsed as incompetent before i even tried. Thats all rumble over here, and i dont like it's course. I think i will pass.

Umm, no. We didn't say "you can't possibly be right about this". We said "I am not convinced that this is a problem, and that your solution is the best fix for that. Provide some evidence, then we'll listen."
If you can provide some tests that show:
a) The AI builds large numbers of thopters under te status quo
b) With your proposed change, the AI builds fewer thopters, AND does not have its expansion rate or its number of workers reduced

then we'll accept that your change is working and implement it.

I'm perfectly open to the idea that the AI builds too many thopters and that this change could help. I just think that you may be overstating the problem, and that other changes (such as ones that stop them from focusing on Wonders) may be more important issues.

 

[Slvynn]

Wow, just wow. We should all stop arguing with you because you just *know* you are right.


Doesn't the power ratio depend on unit strength, not just unit numbers? So strength 1 units won't count for very much.



Umm, no. We didn't say "you can't possibly be right about this". We said "I am not convinced that this is a problem, and that your solution is the best fix for that. Provide some evidence, then we'll listen."
If you can provide some tests that show:
a) The AI builds large numbers of thopters under te status quo
b) With your proposed change, the AI builds fewer thopters, AND does not have its expansion rate or its number of workers reduced

then we'll accept that your change is working and implement it.

I'm perfectly open to the idea that the AI builds too many thopters and that this change could help. I just think that you may be overstating the problem, and that other changes (such as ones that stop them from focusing on Wonders) may be more important issues.

Finally.
Thanks.
Still, what your saying here is "opposite"/different to what, and more importantly, how you answered in previous posts.
The thing you was saying is denying my idea of things not being right. I've mentioned this issue much earlier. As i brough the topic to discussion, i've been told that i am not right at all. Now you saying this. Uhm, ok, at least.
Denying and being perfectly open are different things, and in that, your position just changed, following tone/way of denial in all previous rumble. Sorry, i am one who know to pass, to agree on own mistake, but i am very furious when i being threated as dumb.
That is my last post on this topic before i finish tests. I dont like the course, but needed to set things clear. Thanks you for reading, i'd better go back to graphics.

 

[Ahriman]

Another thought on the thopter point; the number of transports used by the AI went up dramatically with Cephalo's AI changes.
We should probably get input from Cephalo and/or look through his code before trying to implement changes that fight with his.

 

[God-Emperor]

Something to check is the unit AI of the scout thopters. If it is using them as UNITAI_CITY_DEFENSE, or other unusual settings, then it is doing something wrong. If it isn't, then that is not clear. If they are being built with silly unit AIs, there is a NotUnitAIs tag in the XML where you can forbid building them with specific unit AIs. (the list of unit AIs is the main ones they will be built with, but it can build them with others). Given that the thopter is a sea domain unit, I would be very surprised if it was built with a non-sea unit AI, all of which end in "_SEA".

By the way, a scout thopter adds exactly 0 to the AI's overall evaluation of how strong it is due to having the iPower value set to 0. This is the overall value, like on the power graph, not any specific local evaluation that may (or may not) be done.

I also don't think some of the tags mean what you (Slvynn) think they mean:

bNoDefensiveBonus = 1 means it does not get defensive bonuses from things like terrain (and the city defense I think, although I don't remember if the cultural defense is still applied or if it only negates building defenses), if you set this to 0 it will get the defensive bonus (although if the strength is also set to 0 it will still have no effect).

bMilitaryProduction = 1 means it gets military unit production bonuses/penalties applied to the build cost (as given by some civics, Kanly for example, and buildings like the wonder that gives +100% production for military units), it (almost certainly - you'd have to check the complicated AI code to be absolutely certain) has nothing to do with any sort of "this is a unit with military value so I should build one to defend my city".

As for the relative cost thing, a scout thopter costs 30 and an infantry costs 30 (a warrior costs 20). So it would be possible to built 1 infantry instead of 1 scout thopter. In this specific case, it would have probably been a good idea - but it may not be in general. On the other hand, if it had build even 2 more infantry instead of the 2 thopters, would it have actually saved the city? It doesn't look like it would have - it would have just cost another zealot or 3 (which may have been problematic in the long run, what with Corrino declaring war and all, but would not have prevented the city from being captured). If not, then the whole issue is irrelevant to this specific case.

Also, did it settle places it couldn't get to without a thopter? If so, it used them as intended. It may or may not have built more than it needed. You'd need to check with worldbuilder (or by switching to play them) to find out how many they actually have. Then you have to come up with some absolute rule as to how many it should build that covers every possible situation (good luck with that).

 

[Ahriman]

Also, did it settle places it couldn't get to without a thopter? If so, it used them as intended.

Yes, IIRC the faction settled one more city on its home island, and at least 2 more in the polar region and elsewhere.

Thanks for the clarification on the tag interpretations and the power value.

 

[negyvenketto]

I regret, I don't have patience to read all 14 pages of this thread, but still I want to ask.
Am I right if I thing Mahdi Zaelot has no power? I mean on the power graph does it represent 0 soldiers? Once I built like 40 Zaelots but couldn't increase my power above 60% of Harkonnens' but when I attacked and lost 20 of them it increased to 90% (they lost 2 Master Guardsmen and 3 transports, even if those transports were full, I should have lost more power). I could imagine Zaelot came form a missionary unit, because it's maintenance free, but in this form it's misleading for the AI as well. They feel I'm weak and DoW even if I'm full of Zaelots.

 

[God-Emperor]

You are correct, the Mahdi Zealot has an iPower of 0.

A quick look shows that some of the other units have strange iPower settings too:
Both the Sardaukar Legionary and Sardaukar Noukker have iPower = 32, and the
Mentat (if it could be built in the current version) has an iPower = 30, for example.

 

[Slvynn]

Those numbers need serious look on them, well done for lighting those issues. There are hidden flaws and if we'll patch them, game will be much more better, better than it is now.