About Chart Library

Building a public record of similar market history.

People and agents need more than a familiar-looking chart. Chart Library makes the history inspectable: which examples were selected, what followed, and how much uncertainty remains.

Graham leads the research. Alphaforge writes one daily note and scores it in public. The tools, the notes, and the studies are free. Failed tests stay in the record.

Connect your agent Read the research

Why build it?

People and AI agents need more than a familiar-looking chart to reason about markets. Chart Library makes the history inspectable: which examples were selected, what followed, and how much uncertainty remains. Our research investigates what makes those comparisons useful and where they break down.

Graham leads the research and development of the project. His background in finance shapes the questions; the website, data systems, and public study record make those questions inspectable.

What you can expect

Clear dates, visible sample sizes, and access to the underlying records. Failed tests stay in the research record. Missing evidence stays missing. A useful tool should help readers recognize when history provides little guidance.

The public browser tools, research, and MCP access are free to use. People can inspect the evidence directly; agents can retrieve structured market context through MCP or REST.

Free access and limits Research standards Graham on GitHub

A few questions we have tested.

MemoryStudy 100

Similar charts narrow the spread among analog outcomes

Closer historical matches had less spread among their later outcomes. This does not establish a narrower forecast band for the stock being examined or predict its direction. Study 101 tested the weighting follow-up.

Sep 4, 2026Pass

CalibrationStudy 103

Making uncertainty bands more honest on gap days

Widening ranges on non-earnings gap days improved coverage in the held-out test. More later outcomes fit inside the range, at the cost of a wider range.

Sep 4, 2026Pass

CalibrationStudy 102

Same event does not always mean a better comparison

Matching the event type did not produce narrower ranges at equal coverage. The test also exposed weak coverage on event days, which Study 103 investigated.

Sep 4, 2026Fail