New Hotfix Version (12-15)

Status
Not open for further replies.
As a general observation of AI performance, I'm actually noticing some underperformance from the AI, especially on difficulties below Immortal/Deity. I used to play King but upgraded to Emperor earlier this year, and tried Immortal a few times to see how the jump in difficulty felt. Then, with the first AI handicap adjustment I found that the boost in their performance knocked me back down to playing at King again (in the 12-6 patch). But now with the 12-13/12-15 patch I'm seeing really, really poor AI performance at King and upped the difficulty back to Emperor. It's really hard to get an objective read on this, but my subjective feeling is that in the 12-15 patch the AI at Emperor at the beginning of the game (pre-Ren) is noticeably weaker than the old handicapped AI at King. But then once a few eras have passed the AI really starts to pick up steam, a phenomenon I never really experienced with the old handicap model. But from what I'm hearing from the Deity players, with the current patch the early-game AI handicap is still really strong, and then gets stronger and stronger as the game progresses.

However, as it stands right now, even at Emperor the starting "power" of the AI Civs feels very weak. I'm so used to having to work hard early in the game to get as much of an early advantage, and now that feels even easier to achieve. The difference is that in previous patches once you got in the lead you pretty much never lost it as the AI got less and less effective as the game went on, but currently the AI feels like it can start to make a comeback in the later Eras is you don't constantly work to kneecap your strongest AI opponents. That's a really good thing, in my opinion, and I feel with a little more tweaking this new handicap curve could solve the issue of the game feeling like a forgone conclusion once you reach the Ren/Industrial Eras. From the perspective of a King/Emperor player, I feel the AI is still a little too weak in the Ancient-Medieval Eras. I'm not a mathematician, but if the current handicap curve could be adjusted to give them slightly more yields early on while not boosting them anymore than they currently get in the Ren+ Eras I think it would be perfect.

EDIT: I just wanted to clarify that I meant the early game AI bonuses need upward adjustment for difficulty levels below Immortal. Every tick down in difficulty level seems to have a huge effect on early game AI performance. In the 12-15 patch the early-game difference between King and Emperor feels huge. Maybe that's by design, but I feel like maybe the AI sees too much of a drop in early game power for each tick down the difficulty scale from Immortal to Emperor to King to Prince, etc.

It doesn't need to feel like old Emperor. It's a different concept. The goal is to forget that AI is being boosted. Suppose you are playing with other human players of your same skill. If you are behind not for strictly bad RNG reasons or your civ uniques, then AI is over handed.
You still need to play to your best, because AI will improve a lot later.
 
It doesn't need to feel like old Emperor. It's a different concept. The goal is to forget that AI is being boosted. Suppose you are playing with other human players of your same skill. If you are behind not for strictly bad RNG reasons or your civ uniques, then AI is over handed.
You still need to play to your best, because AI will improve a lot later.
Maybe it is time to make Deity really a challenge again? I’m not playing there, but when I read posts, I have a feeling that every 2nd person here plays and wins Deity. They certainly are all good players, no question about it, but maybe Deity even for them should be crazy difficult. It is freaking Deity after all...
 
Maybe it is time to make Deity really a challenge again? I’m not playing there, but when I read posts, I have a feeling that every 2nd person here plays and wins Deity. They certainly are all good players, no question about it, but maybe Deity even for them should be crazy difficult. It is freaking Deity after all...
Slightly impossible. Right now it is plain impossible.
 
Did you try to play to the end? In my Siam game i was 3rd out of 10 in tech in Renaissance, but ended up being 8th out of 9 players. And i was 10 policies behind leader. Industrial/Modern is where new bonuses start to work

I play most of my games to the end, actually.

My games have been more along the lines of Infixo and Mad Madigan. In my opinion, it's still too soon to be too specific about what needs to change. Like tu 79 says, if the game is competitive, then it's working. Speaking broadly, the AI does better culturally early on, which means it still has a lead in all my games. And isolated starts put me way back in culture, which is to be expected. But with a strong cultural civ, I came back -- and as usual, science came back right along with it. The AI does get stronger at the end, but if I'm on a roll, I still pull away incrementally in tech.This is why the game seems similar to prior versions (except tighter) to me.

On a philosophical side note, I agree with Infixo about Deity. That's what it was always suppposed to be. Which doesn't mean to make it tougher than it is today, but rather not too dilute it very much.
 
Emperor difficulty here, and my experience is largely the same as Mad Madigan’s. Early game is significantly easier than it was previously, and I find that now it’s perhaps too easy to get a significant lead in early/mid game. The AI does pick up later, but if the goal is for the player to be even with the AI for most of the game, right now the discrepancy in mid-game is quite high, albeit in the opposite direction from previous versions.
Personally I find the game less interesting when I’m significantly ahead, even if I can expect the AI to catch up later.
 
Emperor difficulty here, and my experience is largely the same as Mad Madigan’s. Early game is significantly easier than it was previously, and I find that now it’s perhaps too easy to get a significant lead in early/mid game. The AI does pick up later, but if the goal is for the player to be even with the AI for most of the game, right now the discrepancy in mid-game is quite high, albeit in the opposite direction from previous versions.
Personally I find the game less interesting when I’m significantly ahead, even if I can expect the AI to catch up later.
You were significantly ahead? Is it justified by your civ strengths? By a lucky start? By what age you were in such leading position? And what are your handicap values?
 
I think AIs get too much bonus culture. I think their handicap culture should be half of everything else. (Or maybe a number pushed to the xml so we can test it like ABC.) Even with Handicap values that feel correct for everything else AIs will speed ahead in culture.
Out of curiosity, I did some quick calculations. Idk if formula for calculating the base cost of a policy is different in Vox. If not (i.e. cost(k) = 25 + (6*k)^1.7) then:
policies vs techs.jpg


So, @ElliotS nailed it...
 
Out of curiosity, I did some quick calculations. Idk if formula for calculating the base cost of a policy is different in Vox. If not (i.e. cost(k) = 25 + (6*k)^1.7) then:
View attachment 483330

So, @ElliotS nailed it...
There's a reduction in tech cost for every other civ that knows the tech, and there's no such thing for culture. If one civ is too ahead, it doesn't matter, but if half world researches the upper side and half world researches the other side, then it speeds up a little.
 
You were significantly ahead? Is it justified by your civ strengths? By a lucky start? By what age you were in such leading position? And what are your handicap values?
Once for China, once for England. Medieval era both times. Using standard Emperor difficulty handicaps. Might raise the early game handicap a bit next time.
 
As a general observation of AI performance, I'm actually noticing some underperformance from the AI, especially on difficulties below Immortal/Deity. I used to play King but upgraded to Emperor earlier this year, and tried Immortal a few times to see how the jump in difficulty felt. Then, with the first AI handicap adjustment I found that the boost in their performance knocked me back down to playing at King again (in the 12-6 patch). But now with the 12-13/12-15 patch I'm seeing really, really poor AI performance at King and upped the difficulty back to Emperor. It's really hard to get an objective read on this, but my subjective feeling is that in the 12-15 patch the AI at Emperor at the beginning of the game (pre-Ren) is noticeably weaker than the old handicapped AI at King. But then once a few eras have passed the AI really starts to pick up steam, a phenomenon I never really experienced with the old handicap model. But from what I'm hearing from the Deity players, with the current patch the early-game AI handicap is still really strong, and then gets stronger and stronger as the game progresses.

However, as it stands right now, even at Emperor the starting "power" of the AI Civs feels very weak. I'm so used to having to work hard early in the game to get as much of an early advantage, and now that feels even easier to achieve. The difference is that in previous patches once you got in the lead you pretty much never lost it as the AI got less and less effective as the game went on, but currently the AI feels like it can start to make a comeback in the later Eras is you don't constantly work to kneecap your strongest AI opponents. That's a really good thing, in my opinion, and I feel with a little more tweaking this new handicap curve could solve the issue of the game feeling like a forgone conclusion once you reach the Ren/Industrial Eras. From the perspective of a King/Emperor player, I feel the AI is still a little too weak in the Ancient-Medieval Eras. I'm not a mathematician, but if the current handicap curve could be adjusted to give them slightly more yields early on while not boosting them anymore than they currently get in the Ren+ Eras I think it would be perfect.

EDIT: I just wanted to clarify that I meant the early game AI bonuses need upward adjustment for difficulty levels below Immortal. Every tick down in difficulty level seems to have a huge effect on early game AI performance. In the 12-15 patch the early-game difference between King and Emperor feels huge. Maybe that's by design, but I feel like maybe the AI sees too much of a drop in early game power for each tick down the difficulty scale from Immortal to Emperor to King to Prince, etc.

The goal is to have AIs generally even to players as mentioned. Part of the problem is that it can be easy to mistake yourself getting better with the AIs getting worse. Are you consistently getting ahead and staying ahead on Emperor? I'm not at all discounting what you're saying, just wanting more details. The formula is complex and needs to work on all difficulties.

Out of curiosity, I did some quick calculations. Idk if formula for calculating the base cost of a policy is different in Vox. If not (i.e. cost(k) = 25 + (6*k)^1.7) then:
View attachment 483330

So, @ElliotS nailed it...

Thanks. Culture is such an important yield this disparity can really matter.

There's a reduction in tech cost for every other civ that knows the tech, and there's no such thing for culture. If one civ is too ahead, it doesn't matter, but if half world researches the upper side and half world researches the other side, then it speeds up a little.

That's a good point.

I'd like to see formula adjusted so AIs get 50-60% of the current amount of culture. We'll see a downtick in difficulty, but can increase overall bonuses to get the 'right' ABC numbers so AIs are evenish with techs AND policies.
 
It doesn't need to feel like old Emperor. It's a different concept. The goal is to forget that AI is being boosted. Suppose you are playing with other human players of your same skill. If you are behind not for strictly bad RNG reasons or your civ uniques, then AI is over handed.
You still need to play to your best, because AI will improve a lot later.

I agree. My only concern with the current AI handicap formula is that the AI is too weak in the early game at a difficulty level where the challenge feels comfortable in the middle and late game, and that this gap increases the more you lower the difficulty. On Deity the AI starts off very challenging and then ramps up a bit as the game goes on, but turning the difficulty down to Emperor (from the perspective of someone comfortable at playing at Emperor) makes the early game feel way too easy, especially compared to how much more powerful the AI becomes later in the game when its handicap bonuses increase. And to be more specific, by "too easy" I mean that the AI seems to be trailing too much in technology, economics, and infrastructure.

As an example, in my most recent game as Germany (a Civ with no special bonuses until you can start consistently gaining CS allies) I started on a continent with Songhai, Carthage, and Portugal. Songhai and I went Authority and Carthage and Portugal went Progress and all of use were doing pretty well at expanding out and settling our lands. There were wars, and Songhai managed an early capture of Lisbon thanks to their special Horseman UU, but I quickly noticed that I was taking the lead in culture and science despite not really having any special bonuses (I only had 2 CS allies and a few more Friends in Classical). This lead got worse and worse even though I was only in defensive wars (although I did capture Carthage's Holy City as it ended up not being their capital and it was adjacent to my borders. Such luck!). By the time I got to Medieval Songhai hadn't even researched Writing, even though they were the strongest AI Civ on the continent. I figured once I got the ability to cross deep ocean and discover the other Civs on the opposing continent that there would be an overpowering Civ beating me in Techs or Policies. By the time I got Caravels and Explorers and discovered all the other Civs I found out that not only was I the tech leader, but I was leading by almost 10 techs! I'm not particularly skilled at this game, so for Emperor that felt like a really unusual lead, combined with the fact that I was playing as Germany. But sure enough once I got Hansa built in all my cities and got the Statecraft policy tree up and running so I could start allying all nearby CSs my slight lead turned into a tremendous one. In the end I was only ever leading in policies by 1-2, but I was a full Era and a half ahead of even the 2nd place Civ (America, it turned out). I reached Renaissance while most of the world was at the end of Classical, and I reached Industrial before a single other Civ could reach Renaissance. And I went Authority > Statecraft > Industry with a religion focused on Gold and Production (Synagogues, Tithes, Thrift, and Council of Elders), so it didn't make sense that the AI was having such a hard time keeping up science-wise. I did manage to spread my religion to my entire continent, so Council of Elders + Tithes was giving me some juicy yields, especially once I started spreading it to the oceanic CSs once Astronomy was researched. But that is still only 200-400 science per city converted, enough to chip a couple of turns off a Tech at Epic speed but nothing overwhelming. Frankly, the production yields from Council of Elders was the most obvious boon, and I was able to time my missionary conversions with Wonder production to grab some key Wonders as soon as they were available. As the AI reached the Renaissance I could really see them speeding up in science and culture production, plus they were all sending spies to Berlin to steal my techs (as they should). It just felt like too little, too late at the point, though. If I had actually focused on Science, or had a Civ with UA/UB focused on Science, how much further ahead in Techs would I have been?

I know, n=1 and all, but I still feel that the AI should be given a little more of a boost in the early game, at least for difficulties of Emperor and below.

EDIT: I guess I should mention that I started on a Sugar monopoly using Communitas, so all of my Sugar was on marshes, meaning I couldn't improve them til Machinery. The only culture-producing luxuries nearby was Perfume, but that was Songhai's territory so when I forward settled him for some Perfume he of course went to war. I only fought defensively and took only a single one of his cities, although I did later Liberate Lisbon in order to gain Maria I as an ally. Also, since my Sugar couldn't be improved for literal ages, I lucked out and got Purity as a Pantheon, so I could at least get good yields from the Sugar marsh hexes until I could improve them much later. BTW Purity is AWESOME with the right start and the Lake yields really did need the nerf that they got. The point, though, is that I did not have any natural sources of science or culture from luxuries, just food and gold (Sugar marshes w/ Purity) and production (scattered Marble, Silver, and Copper, but not enough for a monopoly), and yet I was still absolutely killing it with my Tech lead. Culturally I was only 1-2 policies ahead.
 
Last edited:
There's a reduction in tech cost for every other civ that knows the tech, and there's no such thing for culture. If one civ is too ahead, it doesn't matter, but if half world researches the upper side and half world researches the other side, then it speeds up a little.
Yes, but if you need 200% more science than culture, that 10-20% doesn't make a big difference. Also, with good handicaps and good play, you should be like in top 3 civs, so don't expect much from that effect :)
 
I know everyone hates rubberbanding, but after trying multiple tuning options with the difficulty level, what I most want is that the best AIs get no advantage, or a small one, and the worst ais get the better ones.

So instead of everyone getting 100 for example, the top 33% get 0, the middle get 100, and the bottom get 200.

I may try to edit in and run a version like that to test and see how it goes.
 
I made a mistake with my calculations, I counted Policy Finisher as a separate cost, but it is not. To finish a branch one needs to unlock 6 policies, not 7 ofc.
This makes a difference even bigger (3x):
View attachment 483332
Yeah the old formula took their current science into account. The problem is that made civs ahead snowball. This DID mean that they were getting more science than culture though. Now they get the same amount, which is way too much culture or too little science.

So 33%-50% culture vs science? Or 50% culture and 150% science? I guess it's a question of how valuable each yield is comparatively.
 
.
So 33%-50% culture vs science? Or 50% culture and 150% science? I guess it's a question of how valuable each yield is comparatively.
Depends how many yields are given by handicap. If more than science and culture, and only culture is off, then only culture should be adjusted. If only those two, then the proportion matters, because the main formula will adjust final output anyway. I dont know the details enough to suggest any approach.
 
See, that's where we need more playtesting. From my own experiences the AI Culture bonus is about right, but they need significantly more Science.
 
See, that's where we need more playtesting. From my own experiences the AI Culture bonus is about right, but they need significantly more Science.

Yes, and in my games they've been ahead slightly (vs a lot) early on, then get stronger late (but not enough). This is with Religion, however. Without, it's a different story, and that may be the single biggest factor to account for.
 
See, that's where we need more playtesting. From my own experiences the AI Culture bonus is about right, but they need significantly more Science.
Only a single data point, but in my China game, in medieval I was ahead by three techs and three policies. Granted China has an advantage in culture, but such a significant culture gap may indicate that the AI’s early culture may also need a bit of a boost.
 
Status
Not open for further replies.
Back
Top Bottom