Cut The Cheating AI

[Bjoring]

I understand that the game is not formally finished yet, but I'm still surprised that a strategy game would even go into early access with a strategy AI that so blatantly cheats. AI factions produce troops and respawn Wielders far faster than I can with even the most efficient possible map exploration and economy optimization. The only way this is possible is if the AI is getting significant buffs to resource and unit production. The AI doesn't usually combine all its new units into effective armies, but the sheer number of enemy units being spammed still puts the match into a dull stalemate.

The campaign adds to this imbalance by giving automatic, apparently escalating buffs to the AI in combat. For example: I collected all available XP and artifacts in Arleon missions 1 and 2, and most XP and artifacts in mission 3. By mission 4 a partial strength stack of Legionnaires was able to cleanly outfight my buffed and full strength stack of Queen's Guard. It's frustrating to lose a fight because AI units are arbitrarily more powerful than player units.

I'm sure there are already a few L33T PR0 GAMERS typing about how I ought to Git Gud and win fights by baiting out the AI or using other exploits. These people are thoroughly missing the point. Songs of Conquest is supposed to be a strategy fantasy game; the game mechanics exist to create the illusion of a fantasy world for me to play in. The role of the AI is to play the role of characters and factions within that world (and to do so without the bother of multiplayer). That experience is shattered when I have to treat the AI as an AI to be exploited rather than an in-universe adversary to be outwitted. If I wanted to look for flaws in AI I could just get a job as a programmer, and I wouldn't have to pay $30 for the privilege.

This game has a lot of potential and offers some interesting variations on the classic yet sadly rare HoMM formula. Cheating AI has too much effect on the entire rest of the game to ignore, though, so I'm going to set it aside for now. Fixing the AI might bring me back, and might even interest me in buying the supporter pack. Hint Hint.

[blaque_czar]

Have you tried turning the AI down to easy? Or Handicapping their growth in the Map Settings?

[Xajmai]

Stoutheart mission 4 is not a Good example of the ai cheating. The ai starts with basically half the map occupied and has more cities and resources than you do. It's possible to buy out a dead wielder 1 turn after it's gone too.

I'm in the later stages of that mission and now I own half the map instead, the ai has substantially smaller stacks because they control fewer cities.

Also I'm so tired of everyone comparing and wanting this to be homm. It's not. Tbh the battle and magic system is different and IMHO it's much harder to be a good spellcaster since you need to keep in mind what kind of units/buffs you have and which kind of spell power you're using.

[Bjoring]

Originally posted by blaque_czar:

Have you tried turning the AI down to easy? Or Handicapping their growth in the Map Settings?

I think "Hostile Growth Rate" in Map Settings refer to the growth of non-aligned armies found in the world map, not those controlled by players/AI. I don't think dropping the difficulty would improve the player experience either. The AI uses unit spam to compensate for its lack of strategy, which wouldn't be improved by removing the economic buffs. Skirmishes would go from stalemate to easy victory, which is just as boring in my opinion.

[Bjoring]

Originally posted by Xajmai:

Stoutheart mission 4 is not a Good example of the ai cheating. The ai starts with basically half the map occupied and has more cities and resources than you do. It's possible to buy out a dead wielder 1 turn after it's gone too.

I'm in the later stages of that mission and now I own half the map instead, the ai has substantially smaller stacks because they control fewer cities.

Also I'm so tired of everyone comparing and wanting this to be homm. It's not. Tbh the battle and magic system is different and IMHO it's much harder to be a good spellcaster since you need to keep in mind what kind of units/buffs you have and which kind of spell power you're using.

I cited Arleon mission 4 as an example of the AI's innate combat buffs in campaign, not the economic buffs. That was clear from the initial post.

Every SoC game mechanic is a variation of the Heroes of Might and Magic formula. The devs themselves have cited HoMM as one of their main inspirations. They also cited King's Bounty, another game by Jon Van Canagham, and the HoMM inspired Age of Wonders series. If you want something that's trying to be different from HoMM, then you're looking in the wrong place.

[Munin]

Thank you for the feedback on the Ai, Bjoring! The devs are aware of many of your points already, after all, the Ai is still in development. They're continuously working on improving it.

[Bjoring]

Originally posted by Munin:

Thank you for the feedback on the Ai, Bjoring! The devs are aware of many of your points already, after all, the Ai is still in development. They're continuously working on improving it.

Thanks for letting me know that the devs are aware of these issues. I'll be sure to keep tabs on any updates put out. I look forward to seeing what SoC grows into with a bit more time and polish.

[Exitarnium]

Agreed, the third mission is just absolutely messed up in terms of a difficulty spike where the game doesn't present you enough tools to deal with the doom stacks the AI send upon you.
Atleast as far as my experience went, no idea what the game expects you to do.

[HolyGrail]

Oh god AI sending doom stacks after doom stacks? Now I'm excited. Usually turnbased strategy with AI is just mopping the AI up after winning your first battle against them.. would be nice to see an "insane difficulty" for the AI that specifies that the AI will get buffs, I don't mind AI playing with different rules (afterall it is pve) but it might help to be transparent in the tooltip when choosing difficulty?

[EXEC_PAJA_/.]

About the AI

So, i'm going to do my best to keep this simple because i see that the devs are taking inspiration from Heroes of Might and Magic as well for their game (good for them! i'm not complaining, the style is good).

Here's the thing, i absolutely love playing against the AI in strategy games, but if you're giving the AI the Fairy Godmother's Wand of Wanding and have the AI get buffs to the economy from Nowhere as well as having them spawn full armies in the next turn or so, just like HoMM, this is gonna go bad real fast. If you've played HoMM against the AI, then you already know about the cheating AI's tricks and i'm sure the devs know about it too. It kills the enjoyment, the fun of the game reeeal fast and i'm not talking about the neutral army's growth, i'm talking about the opposing Player AI.

So, i'm just gonna make this suggestion and ask a simple question, hoping the devs might answer. I suggest you don't implement the same flawed system that HoMM has, it's okay for the AI to not be a massive, cheating fairy godmother and make mistakes. The game is still in early access and it can evolve the AI with time.

I love this game's style and design, and i can clearly see the love the devs are putting into the game, and i really want to buy it and show my support for games like these, but i still need to know if the devs are really planning on keeping the cheating AI buffs and army spawning on, even after development. I'm asking because i'm sick of games repeating the same thing over and over again with unfair advantage to the AI.

Thank you for reading and for your patience. I'm hoping for the great success of this project!

[Xajmai]

Originally posted by Bjoring:

Originally posted by Xajmai:

Stoutheart mission 4 is not a Good example of the ai cheating. The ai starts with basically half the map occupied and has more cities and resources than you do. It's possible to buy out a dead wielder 1 turn after it's gone too.

I'm in the later stages of that mission and now I own half the map instead, the ai has substantially smaller stacks because they control fewer cities.

Also I'm so tired of everyone comparing and wanting this to be homm. It's not. Tbh the battle and magic system is different and IMHO it's much harder to be a good spellcaster since you need to keep in mind what kind of units/buffs you have and which kind of spell power you're using.

I cited Arleon mission 4 as an example of the AI's innate combat buffs in campaign, not the economic buffs. That was clear from the initial post.

Every SoC game mechanic is a variation of the Heroes of Might and Magic formula. The devs themselves have cited HoMM as one of their main inspirations. They also cited King's Bounty, another game by Jon Van Canagham, and the HoMM inspired Age of Wonders series. If you want something that's trying to be different from HoMM, then you're looking in the wrong place.

I mean taking inspiration from and being the exact same are two different things. I think it's great that they take inspiration from homm. I also think It's great that the size of the stacks are caped. I enjoy the new magic system (it's not worthless). I guess what I'm trying to say is that I'm tired of people who are complain that the game is not EXACTLY like homm, it doesn't have to be, I think the new mechanics are fun and interesting. I guess a lot of homm fans see homm 3 (and 2?) through rose tinted glasses. For example coming from someone who has played a lot of casual homm I can say the biggest problem me and my Friends have had with that game has been addressed with the addition of the army size limitations, wielder limits and mustering grounds.

There are also other great additions like research, more adaptive cities where you can have econ and unit production centers, defensive locations (although I think the defense buildings could use a buff) and forward bases.

The magic system is very different from anything in any prior titles and I think that too is pretty neat. You actually need to think more about what kind of spells you're casting, sure the strategic aspect of saving mana is gone from the strategic map (now you can boost your in combat mana generation through capturing points etc) but it's replaced by in combat mana management, perhaps you wanna cast a spell to let one of your archers fire twice now but if you hold off from that you might be able to cast a more power spell later. I do think some of the magic needs some balance tweaks as there are some stuff that I think are underpowered in magic tree (especially damage spells)

I guess this is a bit off topic and the dev already replied about the ai but I started writing and couldn't help myself. Cheers

[Star Paladin]

Ya, even after update, I get slaughtered in mission 3,yet again. In skirmish, AI spams generals with so many units faster than I can defend from them. Instead of losing the game in distinct battles, I just get slowly wore down, which is not fun or strategic.

HoMM 1-7 vet here, and did XCOM on ironman, so it's not experience or that I can't stand losing. It's just not fun, any more than losing to a human that uses cheats would be fun.

[Bjoring]

Originally posted by Xajmai:

I guess what I'm trying to say is that I'm tired of people who are complain that the game is not EXACTLY like homm, it doesn't have to be, I think the new mechanics are fun and interesting. I guess a lot of homm fans see homm 3 (and 2?) through rose tinted glasses. For example coming from someone who has played a lot of casual homm I can say the biggest problem me and my Friends have had with that game has been addressed with the addition of the army size limitations, wielder limits and mustering grounds.

I get you. I also like most of the changes to mechanics compared to HoMM. I was just worried that the devs were too focused on fixing HoMM's imperfections and not paying enough attention to AI and other core game elements. It sounds like AI improvements are on the to-do list, though, which is promising.

[EXEC_PAJA_/.]

I guess i should also add in that, since i know that there are gamers that enjoy doing battle against cheating AI's, insane difficulty and the such, the developers should simple leave them as extra difficulty modes. I mean, see through this perspective for a second, i don't enjoy playing against opponents that ignore the game's rules. If they don't have to follow the rules, then why should i? I was recently watching another playthrough of the game and yes, the AI uses the same tricks as the cheating AI from HoMM does (they might be different, but to my eyes it's the same routine). But hey, some players love going against such difficulty modes, why not leave it in as extra difficulty modes?

I'm a firm believer of letting the player choose their experiences and content in games. Just because i hate going against opponents that don't follow the rules the game has set in the board and told me to follow doesn't mean that everyone is like that. I'm not here to say: "Hey! I hate that content, it's this this and that, remove it! It should be removed because of X, Y and Z reasons! Here's my dissertation of the why and how!". I mean, you could just leave it in to the "L33T PR0 G@M3RZzZ" and add in proper AI's that don't cheat for everyone else.

Hell, i'll even give you some examples: War for the Overworld has a nice variation of AI's. They're not perfect, they all have flaws and hell, they even have the cheating AI for players that like going against it. You wanna play against it? Hell yeah man, you can just pick it in the selection. Not everyone likes cheating AI, some like going against opponents that actually follow the rules and have to respect the natural generation of troops per round, the natural resource generation and spawn time. See what i mean? I'll even go one step ahead and say that you can even add in a handicap bar in the lobby and let the player decide if the ai gets, instead of a 1.5x multiplier buff to ecenomy and other characteristics, they can instead get a DEbuff of 0.5x multiplier.

Now, as an example for what i consider a "valid" scenario for an AI to cheat, is story mode. I'll use the game Sacrifice for this one. In the final battle, the player has to go against an AI that is almighty and does not have the same weaknesses as your character does, because of story reasons. So, you have to battle an actual demigod that cannot be defeated through normal means, so you have to use your wits and figure out an alternative way to defeat it. In this scenario, it's okay to give the AI extra abilities and buffs, because in this example we are in "Story time with Chad from accounting!" mode.

So you see, there are ways to deal with this situation. This game does NOT have to follow through the cliched flaws that HoMM followed through. I mean hell, even Total War Warhammer did it and people actually made tons of mods to get rid of SOME of the cheats that the AI used (sometimes unsuccessfully but well, it's an ATTEMPT at least).

I want to give a suggestion for this. In every community of strategy games like these, we have the L33T G@M3RzZz that are as active as the casual players, and that's awesome, more the success for the project. Have them design strategies that can be translated in algorithms that the AI can use. Get some of the L33T gamers to design strategies for the AI. Over time, i believe, you can have an AI that is able to use great strategies that came from the wits of harccore gamers. Take Forged Alliance Forever for example, the AI there uses several strategies from the expert gamers from the community.

[Star Paladin]

yes, having the AI cheat at higher difficulties is fine and understandable.

[Azimuth]

Donno if this is AI cheating of just imbalance, but I find 4th blue faction mission impossible. No matter what I do, no matter where I go, no matter how many times I reroll the map, I just get quickly outnumbered and attacked from all routes, with no time to build enough force even to defend ) Well, yes, balancing is a hard thing, especially for a small dev team.

Concerning difficulties. I recall Xcom director Jake Solomon said: "Why on earth did we limit game difficulties just to 4 levels?" Yes - game needs difficulties, the more the better. Most players aren't turn-based strategies veterans, not even close to that. With 1 "right" difficulty you are losing a lot of people.