Summary of Results
Aggregate outcomes of our agentic pipeline across every open problem extracted from the source journals/conferences. Figures are computed from the published dataset each time this page loads, so they always match the problem listings.
At a glance
Progress across all problems
Each problem falls into exactly one bucket. “Partial progress” means no full write-up exists yet, but at least one partial result does.
What the solutions look like
Solution write-ups are classified by how the problem was resolved: an explicit construction achieving the goal, an exact characterization or dichotomy, a direct analytic estimate, a reduction or duality certificate, a class-wide impossibility or lower bound, or a counterexample refuting the claim.
How far the attempts got
The verifier stage rates every solver attempt out of three, against the problem's main result. Problems attempted more than once are counted at their best score. MOR write-ups carry a post-hoc audit score: the published document was re-graded by an independent verifier pass against the source paper. Scores rate attempts; the outcome buckets above rate published write-ups — one EC problem carries an unpublished 3/3-rated attempt whose write-up did not pass review.
Progress by category
Research categories follow the site taxonomy, ordered by number of problems.
Results by venue
The catalogue draws on two venues — Mathematics of Operations Research and the ACM Conference on Economics and Computation (EC). Each venue gets its own outcome bar and solution-style breakdown.
Further breakdowns
Finer cuts of the same dataset, for readers who want the detail behind the headline figures.
Source papers by publication year
Full category table
Download the data
One row per problem, joining the problem metadata to its solution status. Useful for reproducing or extending any figure on this page.