New Beta Version - August 16th (8/16)

Status
Not open for further replies.
Regarding the above two posts about gameplay speed and the challenge.

One should not be discouraged by the fact that Epic Speeds is easier than Standard Speed. It does not mean that you are bad player at Standard v/s Epic.

On the contrary. It has nothing to do with your human skills. It has to do entirely with AI skills on those two alternative speeds.

All it means is that the human player outperforms the AI strategically. The longer the game the greater the strategical impact decisions weigh in. The AI has no concept of "the future", it cannot see the future, it can only evaluate the present turn. The human player has a concept of future and can execute strategy with great impact into the future. Therefore on Epic speed the human player outplays the AI because humans can feel the future and act in a forward looking way executing long-term strategies 50-100 turns ahead in time to which the AI is blind as it is stuck in the present time only.

That is why Standard speed is considered more fair challenge against the AI. The AI is good in tactical situations of the present but is very poor in long term strategy as like I said it has no concept of the future.

And that is why I believe new advances in AI technology is being researched, like machine learning. Aware that the AI cannot see the future naturally like the humans, the attempt is to teach the AI to look into the past as a best approximation to the future. Learning from the past and predicting the future from those past events. These new advances seem to work well because in games like chess and go (long term strategy games) the AI outperforms the human. And it does that by looking into past games and analyzing identical positions and choosing among the successful ones. The AI beats the human not by thinking ability but by sheer memory/databases availability. While Civ5 doesn't use these kind of AI technology, it uses functions and parameters to evaluate on every turn and therefore it lives only in the present. While machine learning can "see the future: because it has seen identical situation in the past in its database.
 
Last edited:
Even from my experience game on Epic is easier. I can pull a military victory (domination) on Emperor 8 on 10 playthroughs. On Standard pace is down to early aggression, if I can secure a vassal it's snowball from there. If not, AIs can piling up faster.
 
a3kov said:
Yes I remember about the division part, but I missed the increase. Also this is the first game where the unfairness of their bonus is so obvious. I'm ok when another science civ is on par, or even beating me to a few techs, but Brazil ? Come on.
Click to expand...

If you like me (and several others) play on epic speed there was a fix regarding ABC was not taking game speed into consideration.
So epic is almost 1 difficulty harder now than before.
 
HixBozo said:
All it means is that the human player outperforms the AI strategically. The longer the game the greater the strategical impact decisions weigh in. The AI has no concept of "the future", it cannot see the future, it can only evaluate the present turn.
Click to expand...

Not to nitpick, but the AI does have some limited ability to plan for future events.

Examples are when it wants to expand, it looks for targets and takes things like city power and relations with target city. It will move it's troops along it's boarder to setup a structured attack. Once the troops are setup it will declare war. Not much in the way of "future thinking", but still.

The AI will also sometimes forward settle future threats. Military AIs are big on this, sometimes early game they will place a city right next to mine attempting to block me in when much "better" locations are available. It does this with the idea that it plans on building up it's military so if I try to take this city out it will be ready.

The AI is actually quite decent in VP. Not to take away from G and team accomplishments but vanilla Civ 5's AI was alright as well, with a couple flaws. (like being unable to move and shoot in the same turn) It's a shame Civ 6 has such horrible AI. Unlike Civ 5 where the dev team actually patched the game to try to make it better, the Civ 6 has from what I have seen, has done nothing to improve it, not by much anyways. They keep releasing new civs and new features and game modes though. Gotta nickel and dime and all that.
 
Brief update from me: the scouting AI feels better than the old version. I think it might have been a little while ago, but the AI didn't used to heal their scouts when they were damaged. Now they do, and scouts will even head home to find somewhere safe to heal up sometimes :).

Also, I'm running some tests to get some better data on the City-State unique unit thing. I previously said I thought that it was just a mistake but that was a one-off unfortunately. So far my tests suggest that there is a preference for which unique units a military city-state will provide. I.e. if there are only two, they will be the same two (Janissary and Minuteman). There are several other UUs that turn up, but I think it's mostly only where those 2 UUs are already offered by other city-states, or the civ that they belong to is playing in the game. Anyways, I'm going to put my results on Github in a bug report after a couple more tests.

Edit: you can find the issue here (https://github.com/LoneGazebo/Community-Patch-DLL/issues/7042).
 
Last edited:
JamesNinelives said:
Also, I'm running some tests to get some better data on the City-State unique unit thing. I previously said I thought that it was just a mistake but that was a one-off unfortunately. So far my tests suggest that there is a preference for which unique units a military city-state will provide. I.e. if there are only two, they will be the same two (Janissary and Minuteman). There are several other UUs that turn up, but I think it's mostly only where those 2 UUs are already offered by other city-states, or the civ that they belong to is playing in the game. Anyways, I'm going to put my results on Github in a bug report after a couple more tests.
Click to expand...
Supreme Leader was intriguingly satisfied after being gifted a GDR during the late-game of a previous victory. :clap:
 
@Zanteogo

Sure the AI is not dumb, lots of improvements are done. The AI actually beats me on the field often, moves its troops better than I move mine. I mean I give serious casualties.

I just wanted to say that the reason for Standard being more difficult than Epic is due to the AI. It is not due to the human player. It is not because the human player suddenly forgets how to play on Standard.

The human player has the same knowledge of the game whether he/she plays Epic or Standard. The AI likewise has the same logic instructions in both speeds. But humans beat the AI on Epic much more easily than on Standard, because Epic and Marathon introduce time dilation where strategical errors are punished harshly. No time to react and fix quickly long-term strategical mistakes. On standard though such mistakes can be quickly fixed to a certain degree.

Since both the human knowledge and the AI instructions remain the same on these speeds, the fact that Epic players beat the AI more easily means to me that the AI copes worse strategically, but the fact that the AI more easily beats humans on Standard means the AI is good tactician.

Two differences I notice on these two speeds:

1, Standard/Quick - emphasizes tactics and the AI performs good. Time dilation does not effect it too much.
2. Epic/Marathon - emphasizes strategy and long-term forward thinking

This also has to do with space-time considerations. The space is limited (a map with fixed size) and when the time is dilated on Epic if you lose a major battle in the outskirts of your empire then it is over because you cannot quickly make new units to push the invader back. It takes much longer time to replace an army while the enemy is approaching steadily. Remember the AI is not moving dilated in fixed space, it is moving at the same rate and will reach you in a few turns. But your production is dilated and cannot counter the approaching army on time. Marathon/Quick being extreme cases, one favors the AI and the other favors the human.
 
Here is a very simplified example of the time-space dilation in Civ 5. Very simplified but it illustrates this important difference between Epic and Standard.

Imagine very simplistic example, an enemy capitol is two turns away from being captured in a sneak attack. You have a tank 2 turns away from an enemy capitol. The enemy has a tank too which can spoil the plan, but assume the tank is for some reason too far (busy with something).

The only way to save the capitol is to construct a military base. If can do it in 2 turns then spoils the plan.

Standard - takes 2 turns to construct it and my plan fails
Epic - it takes 4 turns to construct it and the capitol falls. Because space is not dilated, my unit moves the same rate of 2 moves, but the military base construction time is doubled.

Although the example is very simplified and elementary, on the grand scale this is what happens and the human exploits these situations better than the AI.

But on the battlefield itself, when I pit my army to the AI's army I give stupid casualties. Too often I either rush or misplace a unit in the formation, don't think it well over and the AI exploits the unit weakness very well. The reason is not that I am more stupid than the AI, but because so many units I get overwhelmed by the logistics and do mistakes. The more units you control on the battlefield the higher the chance of making mistake with a few of them due to the sheer number and complexity of logistics involved and even tediousness and loss of concentration, while the AI moves and attacks optimally all of them.
 
Hey, first time, long time, etc. Huge thanks to G and the whole crew for sinking so much time and genius into such a phenomenal product that's given me such joy over the last several years.

To follow up on what others have been saying about peaceful-wide happiness, I too have been really struggling here. I am usually an Emperor player for tradition and authority civs, have won a few Immortal games as Egypt. But since at least the March version when I started trying to improve with Progress, I have barely been able to make happiness work on King and have never won a game.

Very obviously my own inexperience and lack of skill play a part here, but it's definitely a challenging and non-intuitive path for a newcomer. I always play Communitu_79. Seems all of these games I see close, good land and take advantage of early progress synergies to set up a good wide base early, maybe pick a strategic fight or two with threatening neighbors and build infrastructure in classical. By the time I'm exploring nearby offshore lands in Medieval and Ren, I obviously want the good ones I've been priming to grab, and then I also want the nearby C- and B-grade spots as well that will allow me to continued easy access and defense for these A+ colonies. And then my happiness craters as I clear 10 cities, I fail to grow, I lose influence over city states and thus in the world congress down the road, and I waffle relative to 1 or 2 warmonger or turtle-tradition runaways.

I kind of agree with the folks saying that progress/ peaceful-wide SHOULD have these kind of challenges, it just seems like currently they are out of balance with the obvious strengths of tradition and authority. (I certainly can't think of many historical civilizations that have easily expanded far-and-wide without in fact having tremendous issues keeping their farthest subjects from revolting for their own independence). But, if you're going to grab those earlier tossup territories in the first place, it just seems easier to go full warmonger, forward settle and fight, then get your rich colonies from the choice pickings of your enemies' heartland in the mid-game. And if you want to be peaceful, easier to just not make that challenging reach for territory and make your big military plays in ancient and late-game only, no? Just seems like there's a lot more synergies in both extremes than in the middle, atm.

I've been having a blast with the new version, but it definitely has not helped me crack the happiness puzzle of peaceful-wide play.
 
Status
Not open for further replies.

Similar threads

Recursive
  • Locked
New BETA Version - 4.17.8 (December 16, 2024)
5 6 7
Replies
131
Views
26K
Recursive
Recursive
Recursive
  • Locked
New Version - 4.19.1 (February 8, 2025)
3 4 5
Replies
95
Views
20K
Recursive
Recursive
A
  • Locked
New BETA Version – 4.18.4 (January 21, 2025)
6 7 8
Replies
156
Views
28K
Recursive
Recursive
Recursive
  • Locked
New Version - 4.20.2 (Temporary Thread)
Replies
14
Views
4K
crobartie
C
A
  • Locked
New Version – 4.21.1 (May 3, 2025)
2 3 4
Replies
66
Views
14K
Recursive
Recursive
Back
Top Bottom