Taefin
King
- Joined
- Nov 28, 2020
- Messages
- 825
Hi everyone, first time posting here. I've been glued to the discussions about Dramatic Ages. When FXS announced the mode, I was immediately excited, because I heard: "A way to for AI to actually take each others' cities!" But I have noticed most people left pretty disappointed in the mode.
In particular, I'd read about the Free-City-Blob problem, and AI not being able to manage losing their cities. It also sounds like the mode was not achieving what I had hoped would be its main feature, causing Golden Age AIs to eat up Dark Age AIs so that they could snowball through expansion, much like a human player.
In my first few games, I notice Dramatic Ages was doing nothing to help the AI take cities from one another. Admittedly it was fun to race against the clock to finish a conquest before hitting a Dark Age and having my free cities threaten to flip the rest, but if I survived the early game, I was far enough ahead that I stopped playing.
So I made some tweaks to the City Flipping and Loyalty mechanics (in DLC > Byzantium_Gaul > Data > Byzantium_Gaul_DramaticAges_MODE), with the goal of helping a few AI really pull ahead and create some large empires. This first post will be to report some promising results, in case anyone else has been looking for a fix to Dramatic Ages, and get everyone's input as I work to bring this vision to life.
Modifications: I made it so that 40% of human cities and 20% of AI cities flip on Dark Age, and I reduced the free city loyalty pressure to 1.0 per citizen (I set this up to be the same at all difficulty levels). It seems the free city blob happens because the free city loyalty is 3.5+ per citizen at higher difficulties.
I tested this as Gaul (my first time playing them) on a huge splintered fractal with max # AI, online speed, Emperor with Tech Shuffle but no other modes, and with three mods: Smoother Difficult, Real Strategy, and a variant of Radiant Stronger Cities that I use to boost AI damage against city defenses (this really adds to the danger of leaving cities without walls, to ease with recapturing, and also means that a single siege weapon can take down the wall).
No blob happened, and I noticed that free cities would flip to either neighboring AI in ~10 turns. So that part was a success. With this change, it appears you can play through a game with Dramatic Ages, with all the excitement of rushing to finish a conquest before all your hard fought territory flips back (and it has been pretty thrilling!) without living in the dread of knowing your AI rivals will fall one-by-one into an empire of anarchy.
However, these changes did not trigger the runaway of the leading AIs that I had hoped for. I noticed very few AI were getting dark ages, maybe 3-6 out of 15 each age, and so I imagine there just were too few opportunities for a city to flip AND have a Golden Age neighbor ready to accept them. Often the cities that flipped were on the outskirts, and simply flipped back to their original owner.
The 3-4 leading AI finished the game with 10-14 cities each, compared to 6-9 cities for most, and 2-4 cities for the 3-4 trailing AI. I did notice that the borders between AI tended to occur at geographic choke points (splintered fractal), where 1-2 cities flipped back and forth, and where a flipped city often went back to its Dark Age former-owner because there were two few Golden Age cities nearby to take it. However, this is a somewhat promising result, because it suggests that the dominant AI on each landmass successfully expanded their borders to these choke points, which I have observed does not happen ordinarily.
Unfortunately, I only had one dark age because the game started slowing down as I pulled ahead with Gaul UU, and production was just through the roof in every city. So I suspect that dark ages cased 6-8 AI to fall being on Tech enough to slowed the era advancement, allowing most of the AI (all the leaders and most of the middle) to amass enough Era Score for golden ages most eras.
So I have a few ideas I'll try in the next game:
1) Cause 35-40% of AI cities to rebel on Dark Ages. I suspect this needs to be >33% so that a 3-city empire simply collapses on itself (with rounding up), freeing it up to be eaten by its neighbors. This happened at the very end of the game to one AI, when a growing neighbor caused all but one city to flip. It would have given this neighbor a clear edge, but it was too late to matter. I might cause 45-50% of human cities to flip. That way, an early dark age isn't instant defeat, but I found that the few times I barely avoided a dark age, that i could play around with the 8 cities I knew would flip (monuments are now exclusively for controlling who flips and who doesn't, i.e., 21 vs 20 loyalty), but if I lost 2-3 more cities I'd start to really feel it, especially when that would include a few key walled cities that were instrumental for reclaiming the rest.
2) Reduce free city loyalty even further, to either 0.5 (or maybe even 0). I fear that increasing the AI flips to 35-40% might reintroduce the blob if I keep this at 1.0. My goal here is that a golden age AI can slowly eat free cities, even if they form a sizable cluster. Lower loyalty will also speed up the flipping, so that AI are deprived of their cities for less time. Human players will have an easier time of it too, but we will likely still want to use force to reclaim our borderlands, and there could be some meaningful time pressure if cities would flip to a neighbor in only 5-8 turns.
3) Increase Era score requirement(s), though I haven't decided which. My inclination is to increase the base threshold from 25 to 35 and the repeat-golden-age penalty from 5 to 10 (or even 15). This should allow a few snowballing AI to regularly get golden age, but cut down the middle and bottom of the pack more often.
4) Possibly increase the bonus loyalty per Era-Score-above-Threshold. Originally this is 0.05 per point, so it was pretty common to have 2.1-2.4 loyalty pressure. If this were increased to 0.1 or 0.15, I could imagine AI really cutting into the free cities and even each other. This might also help them break through geographic barriers more often, but I think the main benefit will be empires with large borders having the chance to make gains against each other.
Let me know your initial thoughts, and I'd love to hear how this goes if anyone else tries it (or if you already have).
Edit - Seems there is something buggy about changing the loyalty per excess score. I ended up in a dark age with ~2.4 loyalty per citizen when I set the excess to 0.15 per citizen. So I’ll be keeping this at 0.05 going forward. With golden/dark age flat thresholds set to 35 (instead of 25) the era score thresholds are quite high (28 in ancient) I barely pulled off golden ages through Renaissance before falling short by ~60. But every one of the AI hit dark age two eras in a row, so to play with the higher flat thresholds you likely need to set the previous dark age buff a bit higher, maybe -15. But the 0.5 free city loyalty really helped the AI flip their cities back, with France slowly taking over Scotland, even when both have had back-to-back dark ages, after France took their capital by force in Classical.
In particular, I'd read about the Free-City-Blob problem, and AI not being able to manage losing their cities. It also sounds like the mode was not achieving what I had hoped would be its main feature, causing Golden Age AIs to eat up Dark Age AIs so that they could snowball through expansion, much like a human player.
In my first few games, I notice Dramatic Ages was doing nothing to help the AI take cities from one another. Admittedly it was fun to race against the clock to finish a conquest before hitting a Dark Age and having my free cities threaten to flip the rest, but if I survived the early game, I was far enough ahead that I stopped playing.
So I made some tweaks to the City Flipping and Loyalty mechanics (in DLC > Byzantium_Gaul > Data > Byzantium_Gaul_DramaticAges_MODE), with the goal of helping a few AI really pull ahead and create some large empires. This first post will be to report some promising results, in case anyone else has been looking for a fix to Dramatic Ages, and get everyone's input as I work to bring this vision to life.
Modifications: I made it so that 40% of human cities and 20% of AI cities flip on Dark Age, and I reduced the free city loyalty pressure to 1.0 per citizen (I set this up to be the same at all difficulty levels). It seems the free city blob happens because the free city loyalty is 3.5+ per citizen at higher difficulties.
I tested this as Gaul (my first time playing them) on a huge splintered fractal with max # AI, online speed, Emperor with Tech Shuffle but no other modes, and with three mods: Smoother Difficult, Real Strategy, and a variant of Radiant Stronger Cities that I use to boost AI damage against city defenses (this really adds to the danger of leaving cities without walls, to ease with recapturing, and also means that a single siege weapon can take down the wall).
No blob happened, and I noticed that free cities would flip to either neighboring AI in ~10 turns. So that part was a success. With this change, it appears you can play through a game with Dramatic Ages, with all the excitement of rushing to finish a conquest before all your hard fought territory flips back (and it has been pretty thrilling!) without living in the dread of knowing your AI rivals will fall one-by-one into an empire of anarchy.
However, these changes did not trigger the runaway of the leading AIs that I had hoped for. I noticed very few AI were getting dark ages, maybe 3-6 out of 15 each age, and so I imagine there just were too few opportunities for a city to flip AND have a Golden Age neighbor ready to accept them. Often the cities that flipped were on the outskirts, and simply flipped back to their original owner.
The 3-4 leading AI finished the game with 10-14 cities each, compared to 6-9 cities for most, and 2-4 cities for the 3-4 trailing AI. I did notice that the borders between AI tended to occur at geographic choke points (splintered fractal), where 1-2 cities flipped back and forth, and where a flipped city often went back to its Dark Age former-owner because there were two few Golden Age cities nearby to take it. However, this is a somewhat promising result, because it suggests that the dominant AI on each landmass successfully expanded their borders to these choke points, which I have observed does not happen ordinarily.
Unfortunately, I only had one dark age because the game started slowing down as I pulled ahead with Gaul UU, and production was just through the roof in every city. So I suspect that dark ages cased 6-8 AI to fall being on Tech enough to slowed the era advancement, allowing most of the AI (all the leaders and most of the middle) to amass enough Era Score for golden ages most eras.
So I have a few ideas I'll try in the next game:
1) Cause 35-40% of AI cities to rebel on Dark Ages. I suspect this needs to be >33% so that a 3-city empire simply collapses on itself (with rounding up), freeing it up to be eaten by its neighbors. This happened at the very end of the game to one AI, when a growing neighbor caused all but one city to flip. It would have given this neighbor a clear edge, but it was too late to matter. I might cause 45-50% of human cities to flip. That way, an early dark age isn't instant defeat, but I found that the few times I barely avoided a dark age, that i could play around with the 8 cities I knew would flip (monuments are now exclusively for controlling who flips and who doesn't, i.e., 21 vs 20 loyalty), but if I lost 2-3 more cities I'd start to really feel it, especially when that would include a few key walled cities that were instrumental for reclaiming the rest.
2) Reduce free city loyalty even further, to either 0.5 (or maybe even 0). I fear that increasing the AI flips to 35-40% might reintroduce the blob if I keep this at 1.0. My goal here is that a golden age AI can slowly eat free cities, even if they form a sizable cluster. Lower loyalty will also speed up the flipping, so that AI are deprived of their cities for less time. Human players will have an easier time of it too, but we will likely still want to use force to reclaim our borderlands, and there could be some meaningful time pressure if cities would flip to a neighbor in only 5-8 turns.
3) Increase Era score requirement(s), though I haven't decided which. My inclination is to increase the base threshold from 25 to 35 and the repeat-golden-age penalty from 5 to 10 (or even 15). This should allow a few snowballing AI to regularly get golden age, but cut down the middle and bottom of the pack more often.
4) Possibly increase the bonus loyalty per Era-Score-above-Threshold. Originally this is 0.05 per point, so it was pretty common to have 2.1-2.4 loyalty pressure. If this were increased to 0.1 or 0.15, I could imagine AI really cutting into the free cities and even each other. This might also help them break through geographic barriers more often, but I think the main benefit will be empires with large borders having the chance to make gains against each other.
Let me know your initial thoughts, and I'd love to hear how this goes if anyone else tries it (or if you already have).
Edit - Seems there is something buggy about changing the loyalty per excess score. I ended up in a dark age with ~2.4 loyalty per citizen when I set the excess to 0.15 per citizen. So I’ll be keeping this at 0.05 going forward. With golden/dark age flat thresholds set to 35 (instead of 25) the era score thresholds are quite high (28 in ancient) I barely pulled off golden ages through Renaissance before falling short by ~60. But every one of the AI hit dark age two eras in a row, so to play with the higher flat thresholds you likely need to set the previous dark age buff a bit higher, maybe -15. But the 0.5 free city loyalty really helped the AI flip their cities back, with France slowly taking over Scotland, even when both have had back-to-back dark ages, after France took their capital by force in Classical.
Last edited:
