Jump to content

Smurfing by AI in Starcraft 2?

Guest
Just now, floofer said:

It sees a camera view of the section, so not the whole map, like a normal player, but I suppose with automated hotkeys, it would switch between views quicker - but it does have to work within a limited apm. 

It can switch instantly, action a process instantly and can burst actions upto 1500APM's  All things a human can't do, but I really need to read/view those links to understand what method they implemented to mitigate that.  On the surface it seems there are just problems with testing to get qualifiable data.

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

26 minutes ago, mr moose said:

It can switch instantly, action a process instantly and can burst actions upto 1500APM's  All things a human can't do, but I really need to read/view those links to understand what method they implemented to mitigate that.  On the surface it seems there are just problems with testing to get qualifiable data.

1500APM is a bit much tbh. Burst APM I've seen is ~1100 / ~1200 for a human professional player. Around 400APM is the average, and the AI is meant to model that. Heres a quote from an article. [https://www.alexirpan.com/2019/02/22/alphastar.html

Quote

On average, AlphaStar acts at 280 actions per minute, less than pro play, but this isn’t the full picture. According to the Reddit AMA, the limitation is at most 600 APM every 5 seconds, 400 APM every 15 seconds, and 300 APM every 60 seconds. This was done to model both average pro APM and burst APM, since humans can often reach high peak APM in micro-intensive situations. During the match itself, viewers spotted that AlphaStar’s burst APM sometimes reached 900 or even 1500 APM, far above what we’ve seen from any human.

So it really depends on how they calculate the average, I have personally seen the figure above (1100 etc) from humans in casted recs, so I'm not sure where they are getting that those figures are above any human. The more prominent issue with the AI is the reaction times (switching builds to counter, such as building detectors for Dark Templar early). 

 

Not to say that APM is an issue - AlphaStar has perfect micro, and that is a huge issue, what I was saying in the OP about the Stalkers, effectively they (With AlphaStar's micro) have a counter. Beatable, but very very hard, because their in-game counter, the Immortal is nowhere near as effective. 

 

We can see the results, so the authors of AlphaStar will release a paper documenting the games. So yes it might be unfair, we don't really know for sure, but this is an opt-out approach to testing on the multiplayer, so players can simply choose not to play when they login - and we can see the results in the paper at the end of the testing, and the authors analyses.

Link to comment
Share on other sites

Link to post
Share on other sites

15 minutes ago, floofer said:

1500APM is a bit much tbh. Burst APM I've seen is ~1100 / ~1200 for a human professional player. Around 400APM is the average, and the AI is meant to model that. Heres a quote from an article. [https://www.alexirpan.com/2019/02/22/alphastar.html

So it really depends on how they calculate the average, I have personally seen the figure above (1100 etc) from humans in casted recs, so I'm not sure where they are getting that those figures are above any human. The more prominent issue with the AI is the reaction times (switching builds to counter, such as building detectors for Dark Templar early). 

 

Not to say that APM is an issue - AlphaStar has perfect micro, and that is a huge issue, what I was saying in the OP about the Stalkers, effectively they (With AlphaStar's micro) have a counter. Beatable, but very very hard, because their in-game counter, the Immortal is nowhere near as effective. 

 

We can see the results, so the authors of AlphaStar will release a paper documenting the games. So yes it might be unfair, we don't really know for sure, but this is an opt-out approach to testing on the multiplayer, so players can simply choose not to play when they login - and we can see the results in the paper at the end of the testing, and the authors analyses.

I don't actually see it as an issue of fairness,  but as condition of trying to gather useful information.  If building an AI that can think faster than a human is the goal, then accounting for reaction time is very important, because many little advantages the AI has that impede a humans progress could mask areas in need of improvement. 

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

couldn't they run sc2 on a different pc and use a capture card and a otg mode on something and run the ai on another pc so as to have fog of war like us meatbags?

I live in misery USA. my timezone is central daylight time which is either UTC -5 or -4 because the government hates everyone.

into trains? here's the model railroad thread!

Link to comment
Share on other sites

Link to post
Share on other sites

13 hours ago, will4623 said:

couldn't they run sc2 on a different pc and use a capture card and a otg mode on something and run the ai on another pc so as to have fog of war like us meatbags?

this latest version of DeepMind is exactly that, they have fog of war and can only via the screen only instead of whole map vision

Link to comment
Share on other sites

Link to post
Share on other sites

8 hours ago, RyomaSJibenG said:

this latest version of DeepMind is exactly that, they have fog of war and can only via the screen only instead of whole map vision

thx for clarifying!

I live in misery USA. my timezone is central daylight time which is either UTC -5 or -4 because the government hates everyone.

into trains? here's the model railroad thread!

Link to comment
Share on other sites

Link to post
Share on other sites

On 7/11/2019 at 7:40 PM, SC2Mitch said:

The bot is such a cheater, hope he gets banned by Blizzard. 

I don't really know anything about Starcraft, but I do have a question regarding the AI's "fairness":

 

A human player is limited by how fast they can execute commands / click units / etc. How is this regulated when playing against an AI? Can the AI not give commands to all units at once, without a delay that would be needed in a human player to select / click units etc? Or is the number of commands per second that the AI can execute limited somehow?

Link to comment
Share on other sites

Link to post
Share on other sites

On 7/15/2019 at 10:04 AM, maartendc said:

I don't really know anything about Starcraft, but I do have a question regarding the AI's "fairness":

 

A human player is limited by how fast they can execute commands / click units / etc. How is this regulated when playing against an AI? Can the AI not give commands to all units at once, without a delay that would be needed in a human player to select / click units etc? Or is the number of commands per second that the AI can execute limited somehow?

that is part of how they made it more fair.

I live in misery USA. my timezone is central daylight time which is either UTC -5 or -4 because the government hates everyone.

into trains? here's the model railroad thread!

Link to comment
Share on other sites

Link to post
Share on other sites

first wintrading is ok and now botting is also what is happening to online video games

image.png.bd206f31bc5d53dcba24772470d1ef28.png

Kappa

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×