Research

Hypothesis studies recurring market shape after a fixed decision time. The product estimates when turns can occur. It does not predict long or short direction.

What the model claims

• Timing patterns across a fixed market window
• Sign-invariant shape similarity
• Relative conviction at the decision time
• A defined forecast that can be graded later

What the model does not claim

• A reliable long or short signal
• An exact closing price
• Guaranteed turn times
• Guaranteed profit

How grading works

Recaps use absolute correlation between detrended forecast and observed shape. The score ranges from 0 to 1. A higher score means the shapes matched more closely.

Smooth random paths often score near 0.25 to 0.31. Pooled crypto daily research has scored near 0.41. Direction accuracy stays near a coin flip.

A public note shows the same lesson on a fake symbol: patterns on a random walk.

How to read a map

Start at the marked decision time. Read the horizontal axis as the market window. Treat peaks and troughs as timing structure. Do not treat the slope as a trade direction.

Use Recap mode after the window closes. It compares the stored forecast with observed market data.

Current limits

The research does not show reliable intraday direction. Turn locations remain approximate. Results vary by symbol, model, and market regime. Past results do not promise future results.