AI in strategy games always cheats I haven’t seen a single game where the AI wasn’t built around cheating. Once you figure out how it cheats it’s usually a combination of resource multipliers, build time multipliers and not having a fog of war it becomes much essier to beat at any difficulty.
I hate the term "AI" applied to games, since AI means so many things and usually implies something smart, "intelligent". But in reality, it is more like a "bot" or a "computer player". And the main goal is not to be super-smart, but to be plausible enough and provide an appropriate challenge to the human player.
There are some "fair" bots in games - like in my favorite turn-based Mechanized Assault and Exploration from the mid 90s. Computer players follow the same rules as the human ones - e.g., if something is not visible to the radar, the computer will not see it. The only "cheat" is the resource boost computer players can have on the higher difficulty settings, but it is totally optional. And as an experienced player, you always let the computer have it, since you want a challenge, and without that boost, it has no chance whatsoever.
Real-time strategy AI is absolutely AI in the standard Russell & Norvig sense of AI. There's nothing about the computer science concept of AI that implies "super-smart" or always trying to outsmart the player (rather than trying to be entertaining).
Continuously shifting the goalposts of what "AI" is is, of course, a well-known phenomenon, giving rise to what's called the AI effect or Tesler's theorem [1].
Maybe it is, given that the classical AI definition is so broad, it can mean almost anything. But for me, there's a fundamental difference between something that "tries to be intelligent" and something that "tries to appear to be intelligent".
That is why I prefer to call them "bots" or "computers" - just to separate them from a shifting mess of definitions of what "AI" actually means. It reminds me of "Destination Void" by Frank Herbert, where the main characters were trying to build artificial consciousness and were struggling to define what it actually means.
For a lot of games it can be surprisingly easy to make an AI which beats the median player even when limited to just basic strategies, simply by not getting distracted by the gut feelings that humans have.
There's a whole scene around making bots for Starcraft: Broodwar, using an API (BWAPI) that doesn't allow cheating. They're quite good now, better than most humans. But the top bots still can't beat a pro, or even a high ranked ladder player.
Pitching in on this with a tangent - how good are LLMs with RTS games these days? As someone without friends into that genre, it’d be pretty cool to play eg. AoE II against a capable computer that play like a real human…
Depends on what you mean, LLMs can probably _make_ pretty good AIs. It'll have all the AI scripts in the base game, including the three iterations (base, FE, DE) all the user generated ones ( including barbarian ) and then able to consume the language schema. Rig up a baby model that takes the matchup during loading and hot swaps one of your pregenerated AI scripts.
If you meant _playing_ raw based on LLM input - that's probably the wrong tool for the job. The latency for you to react to a mango shot is faster than a billion tok/s lol
Instead of driving the agent with an llm, it might work to use the agent to hard code heuristics, and use some kind of a simulation to benchmark its skills? Then feeding the results back to the agent so it can improve the heuristics?
When I was a teen I was mostly writing RA2 custom map scripts and rules/units for my friends and watch them battle with my rules in internet cafes. When that was not possible, I was creating custom RA2 AIs, but it was very hard.
These days, I'm having incredible fun developing good old AI scripts with LLMs, for my own vibe-coded RTS game. Just choose all AI players here to make them battle each other: https://egeozcan.github.io/unnamed_rts/game/
I even let the LLM generate a tournament script to make AI scripts from different LLMs battle (headless): https://github.com/egeozcan/unnamed_rts/blob/main/src/script... GPT-5.5 leaves all in the dust currently. I cannot beat most in the game I set the rules myself :)
If you are like me, you can just make LLMs create your personal RTS game and also develop custom AIs. It's so much fun.
Amazing! You're welcome. Yes it must've been around 2000, I was around 13-14 when I built Tiberian Sun Factory and then RA2 Factory.
To be honest I didn't play with mods much myself because I was consumed enough with the joy of building tools (such as for INI editing and voxels) and the websites, and connecting with the community and other site/tool builders. I did play the vanilla games a lot though.
I contacted the community manager at Westwood Studios and was granted the chance to visit once during a trip through Vegas. I remember him and others being very friendly, giving me a tour through the office, seeing the reality of the life of game testers, having some time with the Yuri's Revenge expansion before release, and being given some merch - a Dune II box, a C&C poster, which I still have in storage. Just to share some more of what I can recall from those days.
Huh. You know, I wonder. The API provided to AI scripts must have enough info for limited strategies, but I've never seen what's what. You have.
What are the possibilities for just giving Claude that each turn. Yes, insane over kill, but in a couple of years Claude level AI will run locally on laptops...
I'm really not sure about the timelines, but I can easily speak to the state of things for today: LLMs are very wasteful and slow for real-time decisions. The "thinking" ones have no chance and if you disable thinking, they really struggle with querying the context with the right calls. For the once in a moonshot scenarios that they have enough context and ample time to react, they are great!
Developing AI scripts, however? They are crazy good! I try to re-balance the game to give a little edge to the humans, then it takes a single iteration for your not-even-sota LLM to destroy me. I mean I'm not the fastest RTS player but for the lack of skill, I have the advantage of being the designer of the game :)
About the RA2 AI scripts: You can react to enemy faction, composition etc. but it's impossible to program it to pull back its tesla tanks when they are under threat from a bunch of rocketeers. Those things are hard-coded in the game engine. I think the only exception was the DeeZire mod which patched the game exe, if I'm not hallucinating.
I recently tried playing OpenRA with GPT-5. Didn’t have enough time to put into it properly, but the solution I reached would plan future actions in batches based on game state, including decisions to make based on conditions/thresholds being met.
Maybe something like Gemini Flash would do a much better job.
Even if it worked, it most likely wouldn't be fun. Game devs have talked about the fine line one walks with AI in strategy games: your goal is to challenge the player, not to beat him. It's not impossible to make AI which can beat the player, but that wouldn't actually be fun to play against. So the challenge becomes trying to create something which will hit the sweet spot of challenging but beatable for most people.
I think the issue is less winning and losing, and more feeling and seeming real. Game AI opponents do absurd things that break the immersion; suddenly, you realize you're just playing a dumb game on a machine.
I suspect LLMs might be able to do that better. Even their hallucinations are less frequent and more convincing, very generally speaking, than game AIs IME.
It's already in the works. NDA but the studio I gig at, their Ai love to send NPCs off cliffs.
Because there is no collision between the sky and floor it determines that this is the quickest route. Even with zoning it does something you'll never think of.
If you play the original and then OpenRA you will be amazed how well OpenRA is balanced.
As an example, while in the original game using allied artillery against soviet tesla coils was a dead sentence in OpenRA is great to be able to fire well beyond its range forcing you to come out of the base to defend it.
They also added a ton of features which make the game truly enjoyable and fun to play.
Tangential, but I got introduced to Red Alert C&C through various 'Hell March' videos by random fans of various militaries on YouTube. It's funny how it vibes with nearly every military you throw it over.
Total annihilation had an amazing score.. it was a young media giving talent a way to shine without first having to do a pilgrimage through some elder legends sweatshop.
Against all the evil that Hell can conjure, all the wickedness that mankind can produce, we will send unto them... only you. Rip and tear, until it is done.
Weird. I find the balance for player vs AI to be actually pretty horrible. AI can outrange artillery sight so you have no choice but to push forward always or micro manage units. I have a fork of OpenRA on my GitHub where I try to address this, along with pathfinding bugs, enabling Tiberian Sun and fixing bugs with that. I also updated it to cross platform .NET 10 and bumped the performance up about 6-10x fixing minor bugs. Debating if I'll roll any of the code over as pulls to the main project ever after I tried to fix a random bug before and they were not welcoming to me but that was like a decade ago.
The comparison was against the original game. I haven't played on the computer since ages, but that was my impression the times I played it. I'm not deep into all the nuances of the game just that it was considerably better that the 1995 release.
1995 was the original "Command and Conquer" not "Red Alert". In the original C&C the computer wouldn't attack sandbox walls which was basically godmode since you could just fence them off heh.
The computer player would always target your nothern-most structure for airstrikes or nuclear weapons. This persisted to at least Generals Zero Hour (I never played the campaign of C&C3 or 4 so can't comment on them).
The load/save is what kills me. It's a great project, but having to wait 2 hours for a game to load, fans blaring on my laptop, makes it less playable.
For context, I love huge, massive maps with loads of players. OpenRA replays the entire game to restore, it doesn't have a save-current-state routine.
So 20 hours of massive map + 8 players means 2 hours of pegged CPU to reload the save.
Starcraft 2 has the same overall design. If you disconnected and rejoined in an online match you could watch it scrub through the entire game to get back to the synchronization point (not theoretically necessary I guess, but probably avoids a lot of headaches). 10:1 for simulation to real-time feels pretty on-par with what I remember blizzard could achieve.
Why is saving the state hard for them? So many games and other software save their state .... Imagine Excel 'replaying' the spreadsheet to return to the latest state.
it sounds so stupid u know i am curious Why on earth people would do this. How it is not possible to dump some memory as save/load. is the game state not known?? :/
It should work but you'll need .net 10 and powershell. works on macos, win64 and winarm64 so far I just haven't uploaded binaries since it's just work between a friend and I on weekends to fine tune or fix things
based, been playing this for months with my friend, over anything else.
EDIT> My fav setup is to join a free empty server , set up 2 teams, 2 AI and 1 human vs 2 AI and 1 human. And then play with my friend. Great fun. The AI adds a bit of a randomness to the games. Easy smooth quick interface. Just perfect for a quick free RTS game with a friend.
When I was a teenager, I lived on a farm and our neighbour's were another adult couple. He was late 40s and she was late 50s, her name was Jane.
Jane had an infectious laugh. She was always baking. She died her hair bright red. She drank too much wine. She didnt know much about the details of technology but she was intrigued by it, she read books and she volunteered to help as a teachers aide at the local primary school.
And Jane had a secret, she was one of the best Red Alert 2 players I had ever seen. We'd have matches over dialup and she would totally wreck me in such a short amount of time. I couldn't figure out a strategy to beat her, it was different every time.
I still have very vivid memories of Jane sitting in the corner of her farm house, big thick glasses on, glass of red wine, leading the comrades into war, and laughing as she bombed the allies into submission.
If you met Jane on the street you would never ever guess that under that farmer's wife persona, lurked a dangerous and cunning war strategist. Totally unexpected and utterly fabulous.
We used to play RA on my friend's home network, which was thin net running IPX. The house rule was that if we'd collectively built enough units that the game started slowing down you had to attack. It was good times.
Shoutout to one of my favorite OpenRA podcaster - watching the game is a lot less stressful than actually playing it: https://www.youtube.com/@CovertFlobert
I have such fond memories of this game. Editing the .ini files was a delight - I distinctly remember giving Tanya (with her incredibly rapid-fire guns) cruiser shells and having basically the entire map blow up instantaneously.
It is funny how StarCraft Brood War runs circles around C&C games - people still play it online, there are still 20k dollar tournaments in Korea... the game is just more fun to play and to watch.
It received an official remaster too that was only a graphics update
Meamwhile people make threads about RA what was a bad game even when it came out - "strategy" was to made few overpowered towers, then mass tanks and flood the computer with them. Mutliplayer was tanks + dogs, so the first shot of the enemy tank was wasted on your dog.
People say the same about AoE 2 compared to 1, yet I enjoy 1 a ton (my partner and I regularly play it) and found 2 annoying and couldn't get into it. Might it not just be that some people like RA better than your favorite game?
SC having the status it does in Korea has other causes than only being a good game. I wouldn't argue football is better than some other sport because fifa holds a tournament with prize money that many people watch
Ah man I remember playing Red Alert against my friends online where you had to put the other persons phone number in and (I guess??) you made some kind of direct network connection. No idea if it was billed as internet or what!
As a developer on the game (0AD) I must say I find it way more relaxing than RA and Starcraft(n) and easier to manage with only 4 basic resources. As for the AI challenge on all of these games it comes down to how fast you analyze random terrains and the most optimized opening build order - something they (RTS) all share in common.
EA showed they had Red Alert 2 tapes in the archive when they did the C&C Remaster stuff back in 2018-2020
The Mental Omega folks are also known to have a complete archive of everything which is why their "mod" is such a technical achievement above what the game engine can normally do
Timely. I just downloaded Augustus last night and played a bunch of Caesar III with the assets for $5 from GOG.
Interestingly Augustus is itself a fork of Julius, where Julius intends to replicate original behavior including bugs whereas Augustus adds a number of QoL improvements like better parking control for the walkers.
Anyway all that to say, open-source engine remake projects are awesome.
I hate the term "AI" applied to games, since AI means so many things and usually implies something smart, "intelligent". But in reality, it is more like a "bot" or a "computer player". And the main goal is not to be super-smart, but to be plausible enough and provide an appropriate challenge to the human player.
There are some "fair" bots in games - like in my favorite turn-based Mechanized Assault and Exploration from the mid 90s. Computer players follow the same rules as the human ones - e.g., if something is not visible to the radar, the computer will not see it. The only "cheat" is the resource boost computer players can have on the higher difficulty settings, but it is totally optional. And as an experienced player, you always let the computer have it, since you want a challenge, and without that boost, it has no chance whatsoever.
Continuously shifting the goalposts of what "AI" is is, of course, a well-known phenomenon, giving rise to what's called the AI effect or Tesler's theorem [1].
[1] https://en.wikipedia.org/wiki/AI_effect
That is why I prefer to call them "bots" or "computers" - just to separate them from a shifting mess of definitions of what "AI" actually means. It reminds me of "Destination Void" by Frank Herbert, where the main characters were trying to build artificial consciousness and were struggling to define what it actually means.
For a lot of games it can be surprisingly easy to make an AI which beats the median player even when limited to just basic strategies, simply by not getting distracted by the gut feelings that humans have.
Even for more complex strategy games like say Starcraft II where that's not enough, there's this: https://en.wikipedia.org/wiki/AlphaStar_(software)
If you meant _playing_ raw based on LLM input - that's probably the wrong tool for the job. The latency for you to react to a mango shot is faster than a billion tok/s lol
These days, I'm having incredible fun developing good old AI scripts with LLMs, for my own vibe-coded RTS game. Just choose all AI players here to make them battle each other: https://egeozcan.github.io/unnamed_rts/game/
I even let the LLM generate a tournament script to make AI scripts from different LLMs battle (headless): https://github.com/egeozcan/unnamed_rts/blob/main/src/script... GPT-5.5 leaves all in the dust currently. I cannot beat most in the game I set the rules myself :)
If you are like me, you can just make LLMs create your personal RTS game and also develop custom AIs. It's so much fun.
To be honest I didn't play with mods much myself because I was consumed enough with the joy of building tools (such as for INI editing and voxels) and the websites, and connecting with the community and other site/tool builders. I did play the vanilla games a lot though.
I contacted the community manager at Westwood Studios and was granted the chance to visit once during a trip through Vegas. I remember him and others being very friendly, giving me a tour through the office, seeing the reality of the life of game testers, having some time with the Yuri's Revenge expansion before release, and being given some merch - a Dune II box, a C&C poster, which I still have in storage. Just to share some more of what I can recall from those days.
I left some more details here https://news.ycombinator.com/item?id=48703769
What are the possibilities for just giving Claude that each turn. Yes, insane over kill, but in a couple of years Claude level AI will run locally on laptops...
Developing AI scripts, however? They are crazy good! I try to re-balance the game to give a little edge to the humans, then it takes a single iteration for your not-even-sota LLM to destroy me. I mean I'm not the fastest RTS player but for the lack of skill, I have the advantage of being the designer of the game :)
About the RA2 AI scripts: You can react to enemy faction, composition etc. but it's impossible to program it to pull back its tesla tanks when they are under threat from a bunch of rocketeers. Those things are hard-coded in the game engine. I think the only exception was the DeeZire mod which patched the game exe, if I'm not hallucinating.
Maybe something like Gemini Flash would do a much better job.
I suspect LLMs might be able to do that better. Even their hallucinations are less frequent and more convincing, very generally speaking, than game AIs IME.
Because there is no collision between the sky and floor it determines that this is the quickest route. Even with zoning it does something you'll never think of.
Using LLMs as NPCs can be hilarious to watch.
As an example, while in the original game using allied artillery against soviet tesla coils was a dead sentence in OpenRA is great to be able to fire well beyond its range forcing you to come out of the base to defend it.
They also added a ton of features which make the game truly enjoyable and fun to play.
Well done OpenRA team!
It helps that this is a childhood favorite game of mine.
I didn't play enough doom to really love it in the same way.
And then I pay bills for eight hours.
For context, I love huge, massive maps with loads of players. OpenRA replays the entire game to restore, it doesn't have a save-current-state routine.
So 20 hours of massive map + 8 players means 2 hours of pegged CPU to reload the save.
Heartbreaking.
It's super impressive this works at all.
1. download current version of Linux
2. an MS-DOS/Minix dualboot VM starts with Linus beating Prince of Persia
3. fast-forward all the way through the history of Linux to him merging the relevant patchset
Since when is having to micro your units viewed as an issue in an RTS?!
EDIT> My fav setup is to join a free empty server , set up 2 teams, 2 AI and 1 human vs 2 AI and 1 human. And then play with my friend. Great fun. The AI adds a bit of a randomness to the games. Easy smooth quick interface. Just perfect for a quick free RTS game with a friend.
Every time I've tried to install this previously, this was my wall :(
EA released CnC and Red Alert as free downloads twenty years ago.
I'm able to get the mechanics down. But I can't get the graphics past RA1 level yet without killing the process power.
Whoever runs it, you're awesome!
The player base is only slightly lower than when I used to play RA2 on dial-up like 20 years ago.
I've boycotted EA ever since they ruined the franchise.
Jane had an infectious laugh. She was always baking. She died her hair bright red. She drank too much wine. She didnt know much about the details of technology but she was intrigued by it, she read books and she volunteered to help as a teachers aide at the local primary school.
And Jane had a secret, she was one of the best Red Alert 2 players I had ever seen. We'd have matches over dialup and she would totally wreck me in such a short amount of time. I couldn't figure out a strategy to beat her, it was different every time.
I still have very vivid memories of Jane sitting in the corner of her farm house, big thick glasses on, glass of red wine, leading the comrades into war, and laughing as she bombed the allies into submission.
If you met Jane on the street you would never ever guess that under that farmer's wife persona, lurked a dangerous and cunning war strategist. Totally unexpected and utterly fabulous.
Love you Jane
Moving everything to TCP/IP came with a lot of improvements, but we also lost important things. Reminds me of the move from Flash to HTML5.
The one thing I'm missing is C&C Generals and Zero Hour. Those were also really fun to play over LAN.
https://github.com/TheSuperHackers/GeneralsGameCode
Meamwhile people make threads about RA what was a bad game even when it came out - "strategy" was to made few overpowered towers, then mass tanks and flood the computer with them. Mutliplayer was tanks + dogs, so the first shot of the enemy tank was wasted on your dog.
SC having the status it does in Korea has other causes than only being a good game. I wouldn't argue football is better than some other sport because fifa holds a tournament with prize money that many people watch
https://ihhub.github.io/fheroes2/
https://github.com/vcmi/vcmi
https://www.youtube.com/watch?v=hfMw0NKqFek
The Mental Omega folks are also known to have a complete archive of everything which is why their "mod" is such a technical achievement above what the game engine can normally do
Interestingly Augustus is itself a fork of Julius, where Julius intends to replicate original behavior including bugs whereas Augustus adds a number of QoL improvements like better parking control for the walkers.
Anyway all that to say, open-source engine remake projects are awesome.
OpenRA – Classic strategy games rebuilt for the modern era - https://news.ycombinator.com/item?id=42823667 - Jan 2025 (122 comments)
OpenRA – Classic strategy games rebuilt for the modern era - https://news.ycombinator.com/item?id=37553193 - Sept 2023 (169 comments)
OpenRA: Red Alert, Command and Conquer, Dune 2000, Rebuilt for the Modern Era - https://news.ycombinator.com/item?id=28511076 - Sept 2021 (199 comments)
OpenRA Release 20210321 (and new website) - https://news.ycombinator.com/item?id=26533422 - March 2021 (10 comments)
OpenRA: An open, cross platform and expandable implementation of Command&Conquer - https://news.ycombinator.com/item?id=23342320 - May 2020 (3 comments)
OpenRA: Classic strategy games, rebuilt for the modern era, open source - https://news.ycombinator.com/item?id=14065021 - April 2017 (1 comment)
---
(p.s. reposts are fine after a year or so; links to past threads are just to satisfy extra-curious readers!)
EA Open Sources Command and Conquer: Red Alert, along with other games
https://news.ycombinator.com/item?id=43197131
I only complain because I put in over $1,000 USD as part of a crowd funding campaign to do just that
...13 years ago :(