What changed after the resolution of tournament 67?


Tournament 67 was a fiasco where only 28 people got a cash payout and only one person got returns from staking. Since then, the tournaments no longer reward cash to the people with the best tournament scores, instead skipping over models at what looks like random. What changed in the competition?


Nothing, round 67 was just some bad randomness of the markets that screwed everyone. The rounds after that had a lot of accounts that were banned (for various reasons) hence they were not paid anything.

The tournament is working as intended. Sometimes the markets just act strange and unpredictable.


Hi, @xmaslightguy

@themicon is right. AIs you see with $0.00 reward but are in the top 100 are users who have broken rules, like botting or making too many accounts. We’ll have something on the site indicating that soon.


How did NUMBERGAMESAI break the rules. Only 3 accounts, not bots. Shared Python Script on Slack Channel as encouraged by NUMERAI for stronger models. Curious to know.


Here’s from one of the blogs:

Proof of intelligence
Data scientists collaborate on Numerai already. They share code. They share ideas on Slack. They write blog posts and tutorials. Numerai already has the spirit of a collaborative open software project. But the system design isn’t perfect.

The script was shared on Slack around August 17th or so.


Not fair if all you did was share script. Sounds more like a ‘false positive’ system side.


Hi @NUMBERGAMESAI, none of your 3 accounts are banned. I assume you’re referring NUMBERGAMESAI3's 50th place finish in round 67. Even though your model was in the top 100, it was barely worse than random, and we don’t pay out if your log loss is greater than ln(2), as specified in the “Earnings” section here: https://numer.ai/help.


That works thanks for reply.