Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
Discussions Rules and Guidelines
Report this post
Some have proposed to bring back the difficulty scale adjuster. A part of me see why that would be helpful because if the devs nuance the differences between the levels of difficulty, the easiest might be too hard for some, while the hardest might be too easy for others.
Another part of me likes having a clearly defined difficulty levels. It adds some kind of an identity to the AI.
Some interesting solutions:
- Have both the AI difficulty levels and scale adjuster.
- Decrease the differences between the difficulty levels, perhaps only for the more difficult side? I don't mind having the Deadly AI nerfed just a bit. ...but what about the easier side? Is the easiest difficulty too easy for all players? What are your thoughts?
Although I like challenges, but I think, even playing as efficient as possible, it is too difficult on some templates (ex. 2 players - Corridor). You just don't have enough resources to tech up to some units that could win against deadly and it could just spawn right next to you.Also, as it is right now, only few strategies are viable against the deadly AIs, ex. double Aurelian Sanctum / Workshop / Chelun Sanctuary / Tavern (as many), going full power spell build could help you win. If deadly is easier, players could use a variety of other strategies. If the map fails to provide your with required resources, you're likely dead.
As for FFA games. I had the most fun playing against all deadly AIs on the new largest map: Branching Trails. If I play on smaller maps (ex. Lion's Den, Ancient Valley, etc.), I find myself battling enemies from all sides. It is quite tiring to do so. Around turn 14-18, each AI has 2x to 3x full-sized stacks. Playing against 1 already feels like playing against 2-3 AIs. You multiply that number in FFA.
I was talking about how good Branching Trails is... I now think that this is also due to the current difficulty level.
I only can say there's a massive discrepancy between those.
FFA also has a lot more variance in what's happening, which makes it more interesting most of the time. I think in FFA, with enough luck I can win on difficulty-levels that I don't think I could win in 1v1. Simply because the AI will also fight each other and thus keep each other from getting too far ahead.
Since you talk about beating deadly, I guess you should maybe work closer together with the AI-devs. I don't really have any explicit advice on how they could still improve the AI as I'm unable to still spot improvement-potential as in many cases the AI already does what I'd also do in its situation.
At least not without a proper debug-mode for AI where I can see exactly what they have and what they do with it.
I am thinking of a few things that make AI still weaker than human players.
1. Not fully optimized for meta: In the early game (around round 10-20), certain wielders are stronger than others with magic build. The AI does not prioritize playing with these wielders. Also, their skill tree is not fully optimized with the units they use. For example, if a human player goes Marjatta, expect to see more than 1 stack of Necromancers. As for the skill tree, also expect to see Destruction 3, Arcana 3, and channeling 3, etc.
2. Future prediction: I remember the devs saying that they won't be getting into this. Maybe it's too complicated?
A human player can play better because he/she can plan ahead and make a long term assessment. Say for example, I see full stacks of troops with almost max magic resist rushing toward my Necromancers.
While an AI might choose to cast some spells right now, a player may choose to delay casting to collect a lot of essence for the next turn. To ensure that the enemy would not reach me yet, I could Ice Bolt or Insect Swam some fast-moving units. When the next turn comes... I have way more essence and could start spamming Sabotage to get enemy resistance to -100% or more. Then I start spamming spells that do crazy amount of damage. If I were to cast my spells the turn before, or cast sabotage before, I would not be dealing this much damage and could not wipe out all enemy forces. This is an example of future prediction and long term planning.
3. Exploiting an enemy weakness: If we look at the end game graph, we can see the deadly AI graph shoots up really high as the game progresses. Seeing this, a human player knows that there are loopholes that we can exploit.
For example, if a map grants you a lot of camps and the enemy isn't far away. You could rush and kill the deadly AI early on. It doesn't start with extra units! Last patch, on Valley Low, I won with Peradine maybe around round 4-5.
Around round 8-10 onward, the AI will become more and more powerful. Human players should stay away from the AI until getting high tech units and/or reinforcements. Although the graph keeps shooting up... there is a certain weakness we can exploit.
That is the max unit cap. No matter how high the graph shoots up, you will never get to fight more than max stack size units. And it can take a while until the AI would upgrade the max stack size. This means that if you can reinforce or collect camp units until you have max stack or almost max, you increase a lot of winning chance against the AI, given that your army is more optimized for the meta.
Pavel won against the deadly AI going double workshop Artificers spells build. I tried with other factions: double sanctum Necros and double sanctuary Cheluns. It works well. I could then fight head on with the deadly AI.
Also, you already mentioned another exploit. That is a human player can let the AIs in FFA kill themselves and find the best opportunity to attack. Furthermore, when we play against the deadly AIs, they are more likely to attack us (I think) because we have the least amount of units. Their assessment is off because they are not aware of what we can do with spells and they certainly do not have the ability to future predict.
- Simple : didnt even tryed it
- Fair : after learning the basics of the game i steamroll it ez
- Risky : yeah the ai have more unit and is faster, but i still can beat it without difficulties.
- Worthy : what am i doing wrong ?... The ai got 2 full stack armies and imediatly resplendish them if i beat one of it by chance... I dont even have one good army... how ?
The final version of the difficulties is probably not the one we have now so we appreciate when we get feedback like this. The way we designed it we can alter and fix the different variables without having to do a full update so we can and will change things as we get a more deeper feel for it.
And the way we get that feeling is thanks to players like you, so thank you! I've shared this discussion with our lead Game designer!
May I ask where you are getting this info? I'm curious about the different difficulty settings too. If I find a good reliable source for info about the different AI levels I will post it to the SoC Wiki so everyone will understand it better.
As for me, I find the Challenging AI difficulty the sweet spot, but only on larger free-for-all maps where I have time to develop nicely and can play the first half of the game defensively. The same Challenging AI on a small 1v1 map will simply rush me by turn 20-30 with a strong army and a hero with lots of equipment and twice the XP levels. I haven't found a counter to this yet. But on the same 1v1 map a Worthy AI will be quite manageable.
So to some extent, players need to tune AI settings to suit the scenario. Might be helpful for Lavapotion to add this to the AI tool tip when you are setting map opponents.
I got this from https://www.songsofconquest.com/changelog-archive under Changelog v0.93.
And yeah, it's kinda the same. On a bigger FFA-map I could compete with Challenging. On smaller FFA or 1v1 I can't. I think the bigger map made it so that I'm more out of the way and before I even entered the center where everyone had to pass through the AIs had already met and fought each other and I got lucky that none of them had snowballed out of control yet. In lategame, once I have a fully stacked army their advantages don't matter as much anymore.
But it's also not like FFA is that much easier in general. It really depends on the map. If it's the kind of maps where there's 2 players in one pocket that usually meet and fight before meeting the others, it's just as hard as 1v1.
When I meet an unscathed wielder on Worthy it's usually about my strength but I win due to small differences in managament/combat.
On Challenging under similar conditions I'm far outclassed.
Tutorials & Codex -> AI -> AI Difficulty
https://steamcommunity.com/sharedfiles/filedetails/?id=3211113961
Note that something is different. Challenging AI has +25% income in this source, not +10%
But i play a lot of RTS and some RTS manage the difficulty by allowing to set "Handicaps" for each individual "player" (AI or human) from +100 to -100 (and ofcourse default 0).
That changes stuff like income rates or HPs.
So lets say a worker would bring 10 gold per run and i have a handicap of +30, i only get 7.
What that allows is, that you not only have the ability to set a general difficulty, but also tweak it individually to make the best and most customized experience.
So when i play the game regulary and am stronger (like AoE2/4) but my friend only plays occasionally, i can set the enemies to hard, but lower his handicap (lets say to -40), so all his income is increased by 40% (or hp on units or whatever).
Every RTS makes different use of such a handicap slider, but i prefer the idea over simple and generalistic difficulty settings, because it allows me to tailer the challenge to my needs.
Brutal is too strong but Hard too easy? Ok so i set it to Hard but he gets -40 handicap so hes slightly stronger.
Maybe a Handicap on top of the difficulty settings could help to allow more detailled adjustments. And Handicap sliders are relatively easy to add and work.
They can be as simple as just taking or giving 20% more xp and gold per income source when a handicap is set to +20 but that difference means a lot for the difficulty.
This doesnt obviously fix the snowballing. Snowballing generally needs some gameplay mechanics that mitigate the effect.
A "return" mechanic, last chance or simply ways to accomplish goals with less resources or through other ways.
For example in this game, a lot relies on you moving around with a hero. If that hero is dead, thats a big setback.
The game then might need more ways to still improve your economy, rather than just having to press "Next Round" until you get your hero back.
I think i also played a Hero 4X like this, where the neutral enemies returned in some spaces, but stronger.
Maybe a stronger player would attract more threats in his realm to deal with.
Maybe roaming enemies.
And finally maybe some ways to increase defenses, like being able to wall off segments of the map you captured with buying walls or Towers that need to be attacked in order to be able to bypass an area.
And if the balance is, that defenses are cheaper than offense forces, you have a catch back system where the Snowball might be strong, but gets slowed down significantly.
Awesome thanks!
I will keep tabs on these updates from now on.
Excellent thanks, I didn't know this was in the game!
But there's one thing that bothers me on top of many things said so far.
It seems the AI has map hack. I get ganked way to often if i win a battle with many loses. I tried with quick save to test it and it they just seem to know where your guy is even if they should not have vision of you. You have to get very close to them to actually know what they have but they just seem to know what units you have and if they have even the slightest of chances they will come for you in a straight line.
This was while playing FFA 8 players.
But that is not at all what happened. I spotted the first AI-movement quite early maybe in turn 15 or 16. They basically blocked me in. I want back to reinforce with more units. But I saw them to get supplies from another wielder and still were orange-level-threat. Then they came closer and just killed my main wielder with a massive level- and troop-lead.
So either they got lucky to find me so early or they knew where I started and kinda beelined for me.
Worthy feels challenging but in the end I usually get the upper hand. Mostly because the AI eventually suicides their main-wielder into mine despite having a disadvantage.
But the gap to challenging feels just ridiculous.
I don't really have a good difficulty-level to play at. Worthy just means I'll eventually win and Challenging just means I'll be killed in the early stages of the game.