Could you beat DeepMind AI?

[xer0s]

I doubt it...

https://venturebeat.com/2019/05/30/deep ... flag-mode/

[scared?]

Yeah I played 4 on 1... I easily defeated the morons... Next...

[xer0s]

I wonder who these “strong” players were. I have my doubts it could beat a team of us old school Q3 players as fast as it beat these other punks...

[Don Carlos]

In the end, the DeepMind team found that its agents were, on average, able to beat teams of two human players by a margin of 16 captures. “In a separate study, we probed the exploitability of the FTW [For The Win] agent by allowing a team of two professional games testers with full communication to play continuously against a fixed pair of FTW agents,” the researchers wrote. “Even after 12 hours of practice, the human game testers were only able to win 25% (6.3% draw rate) of games against the agent team.”

Yeah...lets see how they get on against some pro Q3 players. I bet it would be a very different story.

[Eraser]

How aimbot-y are these agents? If they have a 100% Railgun accuracy then that's just cheating.

[MKJ]

the goal of the project was not to code a bot that's better than a pro player (or better accuracy).
It was to have the AI learn the rules of the game all by itself, strategize, and then be able to beat actual human players. And that's pf impressive.

[DooMer]

There's a lot of mechanical execution in quake esp ctf with movement, how do they separate that from tactics?