Your submission is now in Draft mode.

Once it's ready, please submit your draft for review by our team of Community Moderators. Thank you!

Submit Essay

Once you submit your essay, you can no longer edit it.

Pending

This content now needs to be approved by community moderators.

Submitted

This essay was submitted and is waiting for review.

Will human players find a winning strategy against the OpenAI Five bots?

Question

Given how few games the OpenAI bots have played against top-tier human players, it's unclear how robust the OpenAI bots are. The OG team felt confident that given more games they'd be able to adapt and find a winning strategy.

@OG_BDN0tail says that he thinks with 10 games, they’d be able to win; with 50 games, win consistently. Would be super interesting to see... - GDB

OpenAI announced the OpenAI Five Arena, which will allow anyone to sign up to play against OpenAIs best in class bots.

Will players in OpenAI's Arena find a winning strategy against the OpenAI Five?

This question will resolve positively if a team made up of all human players finds a "strategy" that reliably leads to victory against the OpenAI Five bots. This strategy needs to be reproducible by other players, and publicly disclosed.

  • An examples of this type of strategy is, at the International, OpenAI's earlier versions of the bot made simplistic errors when being "kited" by humans. Human players could exploit that again and again to do heavy damage to the AI heroes and lead to humans winning.
  • In one of the AlphaStar matches the bot seemed to glitch in its unit movements, allowing Mana to waltz in and destroy its base. If that is a reproducible exploit, it would count as a strategy.

Make a Prediction

Prediction

Note: this question resolved before its original close time. All of your predictions came after the resolution, so you did not gain (or lose) any points for it.

Note: this question resolved before its original close time. You earned points up until the question resolution, but not afterwards.

Current points depend on your prediction, the community's prediction, and the result. Your total earned points are averaged over the lifetime of the question, so predict early to get as many points as possible! See the FAQ.