The Search
Magpie Trading is a quantitative proprietary trading firm built around mathematics, data, and disciplined research.
We study markets, test ideas carefully, and deploy capital where the evidence supports an edge.
Our approach is systematic, exploratory, and built to adapt.
The Quest
Most markets are only slightly inefficient. Some are not.
We research both established strategies and less-crowded opportunities. Every idea is tested on its own merits before capital is put behind it.
We are especially interested in markets that larger firms may have less reason to pursue — because they are too small, too young, too operationally awkward, or simply unable to absorb enough capital to matter at institutional scale.
Today, Magpie trades equities, prediction markets, and digital assets. Research into futures and options is underway.
We are not trying to win a latency arms race.
Our growth comes from doing more research, covering more markets, and deploying more capital where the math supports it.
The Method
A pattern is not enough.
A strategy should have more than a convincing backtest. There should be a plausible reason the opportunity exists and evidence that supports it.
Our research process follows a simple progression:
Mechanism → Evidence → Validation → Paper Trading → Limited Deployment → Scaling → Continuous Monitoring
Ideas are tested out of sample, observed in live markets before capital is committed, introduced at limited scale, and monitored continuously after deployment.
Our research draws on statistics, machine learning, optimization, market structure, purchased datasets, and data we collect ourselves.
Retail participants can exhibit recurring behavioral patterns that contribute to market inefficiencies. We study those behaviors, along with incentives, liquidity, and the mechanics of each market.
AI accelerates research. It does not make trading decisions.
The Risk
Aggressive when justified. Disciplined always.
We are willing to concentrate capital, use leverage, and accept volatility when the expected return justifies the risk.
A difficult stretch does not automatically mean a strategy has stopped working. Trading involves variance. What matters is whether the results remain consistent with what the model expects and whether the conditions supporting the strategy still exist.
At the same time, we assume every strategy is temporary.
Markets change. Regimes change. Edges decay.
Alternative models can target the same opportunity, and capital rotates among strategies as the evidence changes. Risk controls are automated, systems are monitored around the clock, and critical infrastructure is redundant.
The objective is not to avoid risk.
It is to take risk deliberately — and only as far as the math allows.
The Record
Public metrics will appear here, drawn directly from our systems and updated regularly.
Cumulative Net Return
Since September 2026
Time-weighted return · Net of trading fees · As of Oct 12, 2026
Current Operations
- Markets & instruments monitored
- 227
- Strategies deployed
- 2
- Asset classes traded
- 1
- Proprietary data collected
- 190 GB
- Execution & risk automation
- 100%
Updated Sep 30, 2026 from internal systems.
We publish enough to show what we are building and how it is performing.
We do not publish individual strategies, positions, signals, or current exposures.