Before Feb 2026, will a transformer based reasoning model >1800 elo be able to explain 3+ chess lines at any position?
➕
Plus
5
Ṁ435
2026
53%
chance

Date selected to be approximately one year from prediction statement, and approximately two years from this paper: https://huggingface.co/papers/2402.04494

1800 elo selected as the approximated elo for GPT-3.5-Turbo-Instruct, rounded up to the nearest 100 elo points.

Lines defines as any series of moves that when viewed collectively could easily be explained to a human observer through human understandable heuretics, material exchanges, centipawn differences, or predicted win rates.

Elo requirement necessary so that explanation of potential lines is not alone sufficient, they must be competitive. For sake of rigor, the model must be able to provide three potential lines, if that many exist for the given position, within a single prompt without any further conversation turns, messages, or prompts from a human user.

Get
Ṁ1,000
and
S3.00
Sort by:
filled a Ṁ30 YES at 57% order

Hang on, which elo are we talking about here?

© Manifold Markets, Inc.Terms + Mana-only TermsPrivacyRules