Skip to content
Research tools / Market memoryOpen access · Freely available
Market memory

One market state.
Many possible histories.

Start with a symbol and a date. Explore comparable situations from the historical record, then examine the distribution of what happened next. The purpose is context—not a buy or sell instruction.
Example queriesNVDASPY · Aug 5, 2024AAPL · Jun 14, 2024
How to read a comparison

History offers a distribution,
not a single answer.

01 / Define

Choose the situation.

A symbol, date, and timeframe identify the subject. Check the data's as-of date: a stored daily chart describes a completed session, not an instantaneous market feed.

02 / Compare

Inspect the analogs.

The tool retrieves historical comparisons. Look at sample size, market conditions, and event overlap. A similar chart is not necessarily the same economic situation.

03 / Question

Examine the uncertainty.

Read the full range, not just its median. Check whether the bands have held up, and whether the published evaluation matches the condition you are studying.

Inside the research · A conceptual guide
OBSERVED STATEWHAT FOLLOWED ELSEWHERESUBJECTHISTORICAL WINDOWDISTRIBUTION, NOT A TARGET
01 / The question

What happened in situations like this?

Find comparable historical states. Examine the range of outcomes that followed. Then ask how much confidence that comparison deserves.

A range of outcomes, not a price target
Schematic illustration · Not live market data, a measured projection, or a forecast. Published studies and the search tool contain the actual evidence.
What the evidence supports

Similarity and direction
are different questions.

Study 100 found that closer shape neighborhoods narrowed the dispersion of subsequent outcomes beyond a same-situation baseline. It did not establish a directional trading signal.

Read study 100
Before interpreting a result

Check the comparison.
Then check the tool.

Coverage can vary across conditions and over time. Use the calibration record alongside the comparison, and treat weak samples or failed checks as reasons to withhold a conclusion.

Review the coverage record

The web tool is free to use. For structured responses and agent integrations, see the developer documentation. Historical analogs are observational evidence, not causal proof or investment advice.