Soma Fractal | Predictive Analysis Engine

Systematic alpha from patterns invisible to the naked eye.
Markets contain hundreds of valid, repeatable price structures. Traditional analysis reveals a few. Our system detects them exhaustively.

Website: https://somafractal.com
Log in: https://app.somafractal.com
Contact: alpha@somafractal.com

The Problem

Markets are not efficient — they're crowded. Most funds trade the same indicators, signals, and factor models.

Different charts.
Same assumptions.
Same trades.

Our Edge

Markets behave like complex puzzles unfolding in time. While analysts search locally, we search the entire solution space.

Fractal search.
Mathematical uniqueness.
Structural equivalence.

Core Insight: The Soma Cube Analogy

A Soma Cube has 240 mathematically unique solutions. Even with dedicated effort, most people find fewer than 5.

Markets behave the same way — except the puzzle unfolds in time. Humans rely on heuristics and visible indicators. We built a machine to find the other 235 solutions.

< 5: found through manual effort.
240: total valid solutions.

You can try the puzzle yourself: https://somafractal.com/cube/

What We Built

A proprietary fractal search engine that makes zero market assumptions.

Deep Scan: Scans millions of historical price sequences.
Fractal: Identifies mathematically unique, non-obvious patterns.
Filters: Checks for repeatability, symmetry, and structural equivalence.
Soma Patterns: The output: Validated structures ready for deployment.

Example analysis shown on the page

The section shows a sample run of the engine on BTC-USD, searched against SPY, NVDA, IWM, BTC-USD, GLD and QQQ (150 days of past price action, 30-day forecast window, current phase: Volatile Bearish). It found 7 historical fractals:

BTC-USD, daily, 2018-08-28: 96.38% similarity (Volatile Sideways → Crash).
SPY, daily, 2008-06-05: 94.03% (Volatile Sideways → Volatile Bearish).
QQQ, daily, 2008-06-03: 92.99% (Volatile Bullish → Volatile Bearish).
IWM, daily, 2008-06-05: 92.33% (Volatile Bullish → Volatile Bearish).
NVDA, daily, 2010-01-07: 91.82% (Volatile Bullish → Volatile Bearish).
NVDA, daily, 2008-02-28: 91.68% (Volatile Bearish → Volatile Bearish).
NVDA, daily, 2021-12-22: 90.77% (Volatile Bullish → Volatile Bearish).

From the current price, the chart projects three paths over the next 30 days: Soma Fractal (built from the matched fractals, with its fractal zone), Soma Oracle, and a Monte Carlo baseline (inertia).

Concrete Examples

Volatility Compression (VOL_COMPRESSION): A multi-week compression that looks random on indicators → matches a rare historical structure preceding directional expansion.
Topological Identity (TOPOLOGY_MATCH): Two price series years apart, different assets, different regimes → identical mathematical topology, same forward resolution.
Hidden Setups (NON_INDICATOR): Setups that do not trigger RSI, MACD, or factors → yet recur with statistically consistent outcomes.

"This is not history rhymes. This is history matches."

Why this works

Reasoning: Heuristic (rules of thumb) → Fractal (exhaustive math).
Data handling: Indicators smooth data → We preserve structure.
Search scope: Analysts search locally → Entire solution space.

Positioning

Not prediction by opinion.
Not AI narrative.
Not alternative data theater.
Pure pattern discovery at scale.

Turn hidden market structure into systematic, repeatable alpha.

We are building the engine funds plug into — or compete against.

Send a message from the contact form at https://somafractal.com/#contact or write to alpha@somafractal.com.