Cambrian Labs · Systematic Research

The market
always
changes.

TriloTrader is an automated trading system built around a single conviction: any strategy that only works in one market environment isn't a strategy. It's a coincidence.

Request Early Access How it works
Return
+29.5%
13 months · vs BTC −3.5%
Deployment
Live
Real capital · active
Engine
Evolutionary
Proprietary · in-house
On the nature of trading systems
Most systems are built to win in the present. They're optimized on the past, tuned for current conditions, and quietly obsolete by the time markets move. TriloTrader is built for a different question: what survives when the environment changes?
The Standard Approach
—Optimize one strategy against historical data
—Perform well in backtests, fail in live markets
—Single behavioral pattern — single point of failure
—Break silently when volatility or trend regimes shift
—Require constant manual rebuilding each cycle
TriloTrader
◆Searches across thousands of distinct strategy behaviors simultaneously
◆Selects strategies that generalize — not just fit historical data
◆Maintains a diverse population — structural resilience by design
◆Stress-tested across varied market regimes before deployment
◆The system adapts. No manual rebuilding required.
System Architecture

Three operations.
One robust engine.

The system runs continuously — analyzing, searching, selecting. Not a fixed set of rules, but a living process.

I
Analysis
Read the market honestly
The system ingests price behavior — volatility regimes, trend structures, microstructure patterns — without assuming the future will resemble the past.
II
Exploration
Search the full space of possibilities
Rather than refining one approach, the evolutionary engine explores thousands of distinct strategy behaviors — maintaining genetic diversity across the entire candidate population.
III
Selection
Reward robustness, not performance
Candidates are evaluated not on how well they fit historical data, but on how consistently they hold up across conditions they've never seen. Only what generalizes survives.
The Method

A speciating
evolutionary
algorithm.

Evolutionary algorithms are a known class of optimization. Most implementations converge — collapsing to a single dominant solution, losing all diversity. In markets, that monoculture is catastrophically fragile.

"We developed a proprietary speciating variant. Speciation maintains distinct behavioral niches across the strategy population — the same mechanism that produces biodiversity in biological systems."

Different strategy "species" survive by exploiting different market conditions. The population as a whole is more resilient than any individual strategy could be.

Proprietary · Cambrian Labs · In-house development
Strategy Population — Behavioral Landscape
I
Diversity is the risk management layer
A population of distinct strategies is structurally more resilient than any single "optimal" solution. When one niche underperforms, others compensate. Monocultures have no recovery mechanism.
II
Generalization over optimization
Maximizing backtest performance is the wrong objective function. We optimize for strategies that remain stable under conditions the system has never encountered — the only conditions that matter.
III
Evolution, not engineering
Rather than designing rules by hand, we create selection pressure for robustness and let the evolutionary process discover what works — and precisely where it stops working.
IV
Robustness first. Performance follows.
A system that survives regime changes is worth more than one that dominates a single environment. History is full of the latter. We build for the former.
Live Results

Real capital.
Honest data.

Early-stage live deployment over approximately twelve months. Small scale. Unpolished. We share it because transparency builds more trust than a polished pitch deck.

Cumulative Return · ~12 Month Period
Live deployment · real capital
TriloTrader
BTC Buy & Hold
TriloTrader Return
+29.5%
Compounded · Apr 2025 – May 2026
BTC Buy & Hold
−3.5%
Same window · passive benchmark
Outperformance
+32.9 pp
Through a full BTC cycle ($83k → $120k → $67k → $80k)
A note on honesty: This is early-stage, small-scale data — not a track record. Past performance does not predict future results. We present this as a proof-of-concept signal, not a performance guarantee. This is not investment advice.
Early Access · Limited

Built for those
who think
differently
about markets.

We're opening access to a small group of systematic researchers, quantitative traders, and technically-minded investors. No pitch decks. Just the real system.

Request Access Already have an account? Log in →
No marketing. No guarantees. Genuine research in progress.