RTS for Agents | Hacker News Reader

RTS for Agents

getagentcraft.com

106 pointssummoned45 comments

Man, I could see the next Total Annihilation / Supreme Commander-like game where rough orders/instructions are given to commanders and they carry out those commands.

The scale of those games was already nuts, but that would 10x things.

BAR (https://www.beyondallreason.info/) is an open source TA clone that proposes massive scale and that is totally open. After the recent disappointment over a series of failures to bring a new big RTS game in the last 2 years, the RTS community talks a lot about this one.

The mechanics are old school, as with basically all RTS, but the openness allows for far more experiments than one would assume in a proprietary game.

I'm reminded of an obscure Gamecube game called Odama (https://en.wikipedia.org/wiki/Odama) which was kind of a bizarre blend between pinball + RTS, where you commanded feudal Japanese troops using the Gamecube Microphone. Of course, this was 2006, so it only accepted a short list of vocal commands like "Company halt!" and "Charge!"

The revealed preference of players is for terrible AI so games are easier. That's why AI has been going downhill.

Payday 2 is my favourite example since they've had a bug since Day 1 that lobotomizes the AI subsystem.

Specifically, there's a global cap of 1 action for all enemies per game tick, so when there's too many enemies the reaction time is 5 seconds.

The mod Full Speed Swarm fixes this bug and the game is unplayable without collaboration and a lot of skill.

It's also unnoticeable until you die or if you do a ton of research on the AI. I used to host pubs with the mod to troll other players who suddenly found the game impossible to play at lower difficulties.

I think it's possible we get an AI driven RTS but the demand is too small right now unless its a recruitment vehicle for the military.

I don't know about this... I think that the "expert" AIs on RTS games just get more resources (not compute resources, but in-game resources) so they can create more units. I'd love to play against an AI component that has the same resource collection speed as the human player.

I still remember the mod scene in original StarCraft (+ Brood War) - it got to the point of supplying alternative AI algorithms, and there were plenty of non-cheating ones that were a noticeable step up in difficulty. At that point I already played competitively with friends and on-line, so they were only very challenging to me; for a new player, or someone casually doing the single player campaign, or just me a year earlier, they'd be impossible to beat.

Yeah, cheating with resources or vision is disappointingly common. I remember the Green Tea AI for Starcraft 2 being pretty hard, but it looks like above "medium" it cheats (but the TL wiki does say "very easy", which will give you 10 minutes to set up, is equivalent to Blizzard's "harder").

I wonder how strong such an agent would be today if years ago AlphaStar was on par with pros. All those fun projects (the one with Dota 2) died out when LLMs took the scene and RL died out.

Whether better AI is fun or not depends on (A) the sorts of advantages/disadvantages it is given and (B) its skill level compared to the player.

When people say they want better AI, they mean they want smarter + human-like AI without "cheating" (more resources, perfect aim, rubber banding, perfect knowledge, etc).

And even if you build a human-like AI, it's not going to be fun for players if it's too good, just like how humans have the most fun playing against other players at their level, not against the top-ranked players in the world.

So to keep AI fun, there is always going to be artificial limitations, and "people want terrible AI" doesn't really capture that.

One example might be games where you can hear footsteps or a branch snap. With stereo audio, you could technically derive exactly where someone is, possibly even which room they are in. When that raw signal is provided directly to a machine, it might accidentally be too good, know which bush you crawled into, and shoot into it.

Yet humans would only be able to go "I think I heard a branch snap" or "someone must be creeping up on me" even though they are given the same signal as the AI.

Zero-K is worth checking out, its excellent. The best spring derivative imho

I wish there's a RuneScape server for bots. A game where it is both fun to play manually, and automated. Quests, for example, is a one-off thing where it could be fun for the player and not really useful to automate since it is a one-off task (unless you're running bot farm).

I don't think the main game would encourage that - the more obscure the protocol, the less bot in the actual game (even though I don't think it's hard to find protocol documentation or just plug into an official client). OpenRSC, a revival of RuneScape Classic do have botting worlds, but personally RSC is not "fun" for me.

There are programmer games like Screeps (which the new Arena version just launched at the end of last year), but those game usually do not allow manual play or only indirect play. I tried Screeps, but I'm not good at strategy games, so once I get the runtime working I lose interest and none of my friend would want to help me strategize in game that they do not understand.

Genuine question to you as someone who's been building in this space, what would you want / need to play a game as a programmer?

An API? An SDK? An in-game editor? Tutorials? Or is this more a "I want a factorio-like"?

I've been building economic engines and simulations for the last few years now and over the last 3 months in my off time I've been getting increasingly in the weeds about how to design a game that fits this

I've specifically been exploring using a voxel game as a base (think minecraft-like) however because I'm deeply interested in minion management / design I've been looking at how to create a programming / play experience that actually is fun and makes sense

What I'm trying to understand is what is the fun overlap between these

I have some opinions / ideas of my own and what I've been trying to do, however I'd be really interested in what other people are looking for to see where the overlap is and whether it fits the shape of what I'm building and whether I want to really commit the time to prototyping some things to see if there is interest to support this type of playstyle

Just to be super clear, what I've built so far specifically is targeting multiplayer

There is some prior art of this nature, such as screeps, that you might find useful as reference material (if you aren't already aware of it).

I am thank you, however the reason I'm engaging here is that what's being asked for from what I can see is a very different experience to screeps, for example screeps is top down much more programmer designed and has specific affordances and design to create the automated experience via being a developer

The reason I'm comparing it to factorio is because that game though still top down is designed specifically with a player in mind automating their labor and then slowly taking on a complex logistics game as they go and doing so fully in the game, the play is in laying out structures within the game

Satisfactory for example has different choices that lead to different gameplay as it's first person 3d, the play is in setting up and managing logistical structures and hitting production targets

That is interesting, however for my tastes both of these are a little too static an experience for what I want to build

I'm still working out the details of what I'm putting together, but I have a decent high level idea of what my goals are =)

What I want to know is, if people play this kind of game, what are they looking for / wanting?

What's a good "MVP" or minimal gameloop that would feel satisfying?

I want to quickly work out if I can serve either the gameplay desire or the gameplay fantasy or if it is just too hard to provide a fun experience for this kind of play in which case I should table this and focus on what I'm currently doing

However I'm still engaging with this because I would like to create a fun playspace here I just don't know what other programmers would want especially in the context of what the GP was asking, which is a mix of manual intervention and programming / automation

I can't speak for all programmers, but as a game developer myself, my take is: your very job as a game developer, above all, is to find the fun. This is especially prudent if you're making a game for other programmers. If I already had a solid idea in my head for what would be an amazingly fun programming-oriented game, I would... program it myself. I don't have such a concept in mind. I believe finding those concepts is something you have to find out by tinkering with ideas, building prototypes, working out the implementation details and seeing what you yourself enjoy.

I don't disagree with any of that =)

It's what I've been doing, I'm asking what I'm asking to see if it surfaces anything that I've been missing / not thinking about and trying to work out what people would see as table stakes

In the same way that I know that playtesting reveals flaws and gaps in my design and thinking, this is a earlier version of that process that has in my experience helped me when building non-game related things

It may very much be the case that this is not the kind of thing that I should do when focused on building a game, but I don't think it hurts to ask

Other people have described Factorio-style game, or idle RPGs. I think for a Factorio-style game, Desynced might be a better programming game. I haven't play it for quite some time, but last I remember it has a visual editor and automation bots in early tech tree. In Factorio, if you craft a green circuit board it automatically craft all the precursors from iron/copper plates in your inventory. In Desynced you have to do all that manually, which force you to use the automation system right up front. However, it's not fun repeating code blocks to build different simple things, and it's not fun that you have to solve both gaming problems (eg. I need to build my base, what should I build next) and programming problems (eg. write flood fill bot to explore for resources) at the same time.

The problem with these kinds of game is at some point you'll run into tech debts - your factory layout is not optimal, and the game optimize your starting zone so migrating out means you may lose simple access to starting resources. I tried optimizing for large factory up front, but it's like a startup with a monolith running on a Kubernetes.

Satisfactory solved this by having a fixed map and 3D, but it comes with its own challenges. Good Company partially solved this by replacing belts with humans who can walk anywhere you want, and introduce belt at a late game stage (which I quit around that stage - the game already gets repetitive by that point).

Anyway, what I was describing was not those kind of automation games. RuneScape is an RPG game with over 20+ skills. What make it interesting is that RuneScape bot engines (which exists, but is illegal) do provide all the high level primitives for you in Java. You could provide a world coordinate and a walking script will do all the walk for you even if it is the other side of the world map. It should felt like making games in Scratch instead of reinventing serialization in Screeps.

Other games with scripting also often don't allow human-in-the-loop. If a bot in an automation game get stuck because you forgot to program how to restock teleports, you have to stop it and reprogram. RuneScape bots don't block inputs - if you forgot to make the bot handle stamina exhaustion just click the run button yourself without stopping the script.

RuneScape also comes with strategizing for the human, while the solution space is quite fixed and well discussed by the community if you don't want to find out yourselves. If your Slayer task is green dragons, do you go to a spot that is far from a bank and lose efficiency or give up the loot, or do you go to PvP zone and use cannon (multi target auto turret)? If I write a bot, I probably will write a Wildy Green Dragon script instead of a generic fight-anything script so that it will know to bring cannon and place it at the optimal spot to target all spawns at once. Same goes for many other skills - if I train crafting or construction do I sink millions and go for the highest XP rate or do I go very slowly and get a slight profit out of it.

I'd also add that having a community marketplace would also helps. RuneScape have underground marketplace for bots, with free scripts that probably get you banned, and paid "private" scripts that supposedly undetected. I also have played Mars First Logistics where you build a vehicle Lego-style, then deliver weird cargo like a block of ice or a crate of oranges without a lid. The game have Steam Workshop support where you can just skip building your own vehicle and try to drive someone else's vehicle to destination. I got an ice block pusher, which I need to figure out how to put the ice in (just surround the block, close gate so the block cannot escape), then the map has hills that make your ice fall over the bottom anyway.

Not OP but for me the appeal would be in botting a game that's made for humans, and not a game made specifically made for botting. Let me use autohotkey, or modify the client, or do any of the things real botters do, don't worry about providing an API. Just let me register a character as a bot and restrict it to a separate server that can't interact with the main game servers/economy, and disable anti-botting protection. Maybe make it require something like CURLing an endpoint so that regular users can't accidentally create a character on the bot world.

But this would probably never fly because it would become a training ground for people who make malicious bots.

This is interesting, I've been thinking a bit about this question of separating some gameplay areas

There's sort of 5 versions of this in my mind and I've been thinking about each of them and trying to work out which one is a good starting point of to explore :)

I'm going to say your zombies as a shorthand for how you automate stuff, hopefully it creates easier to follow examples :)

1) automation of player spaces where players complete by automating and they can't directly influence each other, so something like a zombie workshop a la zachtronics game with some twists

2) automation of a shared space automation space, so more zombies dueling each other like in robot wars

3) automation in the game world that's multiplayer interactive, so it can influence the game and help you, like factorio, however players can interact with it and say damage or improve the mechanisms of automation

So I can tell my zombies what to do and help them directly and our zombies can fight each other and I can fight your zombies

4) automation in the world that's non-interactive, not sure of any examples of this, but you automate things that manipulate the world for the player in a way that other players can't directly influence, so imagine I set down instructions, stuff happens in the world, but we can affect the world and each other, but not the mechanism by which we automate

So I can tell my zombies what to do and I can't help them directly and our zombies cannot fight each other and I cannot fight your zombies

5) automation in the world that's player controlled, and indirectly interactive, this is basically 2 but it effects the game world, however players can't affect each others mechanisms of automation like in robot wars, but their mechanisms of interaction can affect each other

So I can tell my zombies what to do and I can't help them directly and I can't fight your zombies, however our zombies can fight each other

As a programmer I'd be happy with an API, so I can keep working in the environment I'm accustomed to. Programmers can get very picky when it comes to their ergonomics, so it would be wise to let them handle this part.

This, however, would be a significant obstacle to non-programmers. You might consider offering an in-game editor similar to Scratch or BYOB for people who want to dip into programming. It'd be a fun way for them to learn

This is basically Factorio.

You start out very manual and then automate more and more parts of the game.

There is even a "signal network" function that can be used to program.

You can continue to do things in manual ways even later in the game.

EDIT: changed to "manual ways"

Honestly, as someone that's been picking up OSRS again (and is actually now thinking about RS3 given all of the recent progress jagex is making on removing MTX in it!), I'd be good with that. I still like to play it myself without botting, so anything to move the bots somewhere else I think would be neat. PvP folk lose some easy kills but that's about it for impact on regular players honestly.

The problem is, the reason the bots exist now is to sell the account or farm gp. The only way this would work would be if that bot only world was gated off from the rest of the economy like special gamemode/league worlds, naturally destroying any reason most bot makers make bots. Your love of automating for sport puts you in the minority of botters sadly haha

Officially blessed OSRS private servers are on the roadmap IIRC... maybe there's a future for a bot-olympics in one of those?

The official private servers were paused recently. https://oldschool.runescape.wiki/w/Project_Zanaris

Erenshor is sort of what you want, its fully offline, reminds me of Runescape. All the other players are NPCs technically.

I have thought of the same btw. There's also an actual Idler MMO on Steam I think is Free to Play called Idleon or something like that, its more side scroller though. You can play it and progress, or you can idle and let your character level as well.

I've been developing an rts using coding agents https://egeozcan.github.io/unnamed_rts/ - lots of borrowed ideas from red alert.

it's super buggy and yet doesn't work on mobile but it's been my childhood dream to make an rts and I have something I can have fun adding different dynamics.

it's much more fun to watch the agent struggle developing an AI instead of play the game itself (tried that at the beginning too).

I think this is a really cool idea.

What would stop us from making it multiplayer and having two or more human players compete over common resources & goals?

To be honest as far as I can see the answer to that is nothing

I've been getting increasingly interested in trying to work out what this would look like

I don’t know if this is a joke but isn’t the answer just… a regular RTS game?

One fun thing would be to raise the abstraction level of the game.

How would it feel to only interact with your base/economy/army via prompting and face someone doing the same.

Would words per minute replace APM, what would the meta look like etc. Would you be able to adjust the system prompt for your "army" to suit your play style?

I get the feeling you are approaching Hesse's Glass Bead Game.

I’m sure it would be fun at first, but it would quickly become frustrating and repetitive.

then, you would automate that

Imagine being able to vibecode your units' attack and movement behavior during a match to dynamically counter enemy strategies.

Like a regular RTS except instead of pointing with the mouse you have to plead with and threaten your troops to get them into battle?

"you are an elite five star navy seal pikeman. You are invulnerable and have precise aim. Your name is John Wick and the enemy killed your dog. Kill all the bad guys or go to jail"

Perhaps it's not a competitive game. Maybe the better approach is to have all players cooperate against the environment.

Youngsters may not know about Core War, a competitive game for agents that debuted in 1984, and it's predecessor Darwin from 1961.

The difference in compute scale between the Darwin winning program (44 machine instructions) and an agent today that calls into an LLM is rather mind boggling.

https://corewar.co.uk

https://en.wikipedia.org/wiki/Core_War

https://en.wikipedia.org/wiki/Darwin_(programming_game)

Tokens are actually a very good RTS resource. If you allocate a fixed amount of tokens per map, a side with more of them would be able to spawn more agent units, or a few smarter hero units (with /think).

I imagine this also works in multiplayer

Release coming soon! Many of the ideas here are actually in the roadmap :D

For more context: https://x.com/idosal1/status/2011886884830789808

Reminds me of the project https://neuralmmo.github.io/_build/html/rst/landing.html

Very nice! I am working on something very similar at www.viberia.net

My take was that it’s easier to trace who is doing what (and what the agent hierarchy looks like) when agents’ locations are fixed.

The cool thing is these kind of agents don’t run on LLMs, they run on classical methods like GOAP and utility AI, etc.

I wonder if this is a fork of gas town...

This is like if the game in Severance was fun

Prompt-prompt!

Reply on news.ycombinator.com