Live data from Hacker News

DeepMind and Blizzard Open StarCraft II as an AI Research Environment

deepmind.com

81–90 of 283 posts

Re: DeepMind and Blizzard Open StarCraft II as an AI Research Environment

#81

This seems all in good fun but I wonder if it's come too late. Starcraft 2 is at its twilight. The biggest leagues of South Korea have disbanded. [1] The prolific progamers who transitioned to Starcraft 2 have gone back to Broodwar. [2] Blizzard itself has scrubbed all references to Starcraft 2 on the very home page of Starcraft. [3] Except for the twitter embed, it has only only one "2" character... in the copyright…

I disagree I got into Starcraft recently and find it very much vibrant, both in the pro scene and casual. But that’s irrelevant. The point is it’s still a great ai challenge

Re: DeepMind and Blizzard Open StarCraft II as an AI Research Environment

#82

Earlier quoted context omitted.

Awhile. This just isn't like Go or Chess. The gap from perfect information to imperfect information is quite a chasm, and from turn-based to real-time is even more vast. I play Age of Empires 2 semi-competitively, and I just can't imagine the research progress that would have to be made for a pro to lose to an APM-limited AI agent. So much of the game comes down to intuiting what your opponent is planning without bei…

On the other hand, bots are starting to beat professionals at (thousands of hands of repeated) Poker, so I think we can't say that imperfect information is something that's especially intractable for maching learning algorithms. http://spectrum.ieee.org/automaton/robotics/artificial-intel...

Yeah but compare the search space of Poker vs Starcraft.

Re: DeepMind and Blizzard Open StarCraft II as an AI Research Environment

#84

Earlier quoted context omitted.

Awhile. This just isn't like Go or Chess. The gap from perfect information to imperfect information is quite a chasm, and from turn-based to real-time is even more vast. I play Age of Empires 2 semi-competitively, and I just can't imagine the research progress that would have to be made for a pro to lose to an APM-limited AI agent. So much of the game comes down to intuiting what your opponent is planning without bei…

On the other hand, bots are starting to beat professionals at (thousands of hands of repeated) Poker, so I think we can't say that imperfect information is something that's especially intractable for maching learning algorithms. http://spectrum.ieee.org/automaton/robotics/artificial-intel...

This wasn't an issue for Bridge either.

As soon as you can build a probability distribution over possible states, you can use Monte Carlo like methods.

Re: DeepMind and Blizzard Open StarCraft II as an AI Research Environment

#85
post #78

"so agents must interact with the game within limits of human dexterity in terms of “Actions Per Minute”." I am really glad they are limiting APM because otherwise things just get stupid.

Would not actions per second be a better limit tho?

Why?

Speed is in MPH or KPH, would it be better to go by m/s?

The convention in gaming is APM, so they're just using the nomenclature that is already understood.

Re: DeepMind and Blizzard Open StarCraft II as an AI Research Environment

#87

Earlier quoted context omitted.

> SC2 got rid of a lot of that You think so? My impression is that SC2 had a lot more of repetitive tasks you had to do. E.g. wall off the ramp, send a worker scouting, ... and you have to perform certain actions every X seconds (like using chrono boost). A lot or mastering the game is rote learning, and polishing a build order. Another big part is constantly scouting and reacting to what the enemy is doing. Due to t…

I really like the idea of a helper AI that takes high level commands and handles the manual dexterity aspect of the game. That would make RTS games a lot more attractive for someone like me who prefers simple, thoughtful games rather than complex games with a steep learning curve.

There are actually 2 parts to the complexity of a modern RTS game like starcraft:

1. Memorizing certain well known strategies and counters very well and recalling them immediately.

2. Having decent speed with the mouse/keyboard to actually execute those strategies within a very short period of time.

I think what you're talking about it automating 2)... which I completely agree with. How to do it... that is more complex though...

Re: DeepMind and Blizzard Open StarCraft II as an AI Research Environment

#88
post #85
post #78

Earlier quoted context omitted.

Would not actions per second be a better limit tho?

Why? Speed is in MPH or KPH, would it be better to go by m/s? The convention in gaming is APM, so they're just using the nomenclature that is already understood.

Maybe what usaphp is getting at is that the AI could still gain an advantage by doing a set of actions much faster than humanly possible in just a fraction of a second as long as it kept its total number of actions that minute below the cap.

Re: DeepMind and Blizzard Open StarCraft II as an AI Research Environment

#89

The API Blizzard is exposing is really nice. Sadly most of the advantages AI had in SC1 were just due to the fact that an automated process could micro-manage the tasks the game didn't automate for you (a lot of boring, repetitive work). SC2 got rid of a lot of that while still allowing room for innovative and overpowered tactics to be discovered (MarineKing's insane marine micro, SlayerS killing everyone with blue f…

> SC2 got rid of a lot of that You think so? My impression is that SC2 had a lot more of repetitive tasks you had to do. E.g. wall off the ramp, send a worker scouting, ... and you have to perform certain actions every X seconds (like using chrono boost). A lot or mastering the game is rote learning, and polishing a build order. Another big part is constantly scouting and reacting to what the enemy is doing. Due to t…

>You think so?

Brood War is the most mechanically demanding game I have ever played, and certainly the most demanding that has ever been an esport.

[This clip](https://www.youtube.com/watch?v=UXH8eCcvQMI) of Flash playing SK Terran style is illustrative of what I mean.

Let's go down the list of things Flash needs to do:

He needs to click on every one of his production structures every 15-19 seconds, and click M or C. Otherwise, no army.

He needs to click on every one of his command centers every 13 seconds, and click s. Otherwise, no workers.

He then needs to tell every newly built worker to go mine, otherwise they just stand around doing nothing.

He needs to build supply depots roughly every 20 seconds, or his entire production grinds to a halt.

His army consists of well over a hundred marines and medics, which Flash needs to stim, split, and maneuver to take on lurkers and defilers who will crush him in an unmicro'd fight.

His science vessels need to be irradiating defilers constantly to prevent the Zerg reaching critical mass, while dodging scourge using the Chinese triangle technique.

If Flash just clicks his science vessel hotkey and casts irradiate, every ship on that control group will waste its irradiate on the same target - he needs to manually select each vessel before casting spells.

His entire army, and his scans, must be controlled with just 10 group hotkeys of at most 12 units each. Except for buildings, of which you can have at most 1 hotkeyed.

To maximize speed of his army going up and down ramps, he needs to spam-click the move command. A single instruction causes the units at the back to spaz out and take a long time to do anything.

As the above implies, Brood War unit AI is terribad.

https://www.youtube.com/watch?v=YCpUsmuQG3E

Which means everything Flash does above needs to be double and triple checked to ensure the AI doesn't decide to go off and pick its nose. Building supply depots was especially bad for this.

Brood War is a game that no human being will ever play perfectly. Even the top tier professionals, like Flash, Bisu, or Jaedong, can't do all of the above all the time - they have to prioritize some activities over others and treat their actions per minute like a resource the same as minerals and gas.

>Granted, this is maybe because I played SC1 more on LAN, and there wasn't all the metagame going on. But I think SC2 really does focus on "grinding" and rote learning to get better, probably this was chosen to make it more "eSports"-like.

Starcraft 2 is a game I could get to masters league in, despite only playing make a few hours a week for a month or two. In Brood War, that much practice wouldn't get you D ranking on ICCUP. Everything about SC2 is designed from the ground up to lower the grinding, remove the muscle memory requirement (you don't get to "play" Protoss in BW until you can hit your P key blindfolded), and encourage new or unskilled players to get on and play. At that it succeeded, albeit kind of at the cost of its professional scene.

Re: DeepMind and Blizzard Open StarCraft II as an AI Research Environment

#90

This seems all in good fun but I wonder if it's come too late. Starcraft 2 is at its twilight. The biggest leagues of South Korea have disbanded. [1] The prolific progamers who transitioned to Starcraft 2 have gone back to Broodwar. [2] Blizzard itself has scrubbed all references to Starcraft 2 on the very home page of Starcraft. [3] Except for the twitter embed, it has only only one "2" character... in the copyright…

I don't quite agree, FWIW. SC2 does seem to be at its twilight in Korea , and I agree progamers and fans there are super interested in Remastered. But I don't think Remastered will be very popular outside KR. The SC2 "war chest" promo appears to have made more money than expected, as measured by hitting its funding ceiling within a few days. So I don't think it's "Remastered replaces SC2", I think it's a divergence i…

As an esport / spectator sport, SC2 has been waning for a long time, and similarly the War Chest was capped at 200.000 for prize pool money and an unknown amount after that (which, compared to Dota 2 or League of Legends prize pools, is not a lot).

I think, given the matchmaking update for remastered, that SC1 will see a resurgence both inside and outside of KR, but I am not sure either SC1 or SC2 will stay competitive in the long run.

Post reply on HN