by The Mathematical Company of California

Ideas in. Trade plans out.

Drop ideas as your day unfolds. By the close you know which trades to take, backed by a full backtest and one-click deploy.

See our tech

Trusted by traders from

What we do

Three quiet steps: capture, replay, deliver.

Based on X, seems like Starbucks is underperforming, can you check opportunities?

System Logged. Checking technicals & sentiment...

Step 1 · Intake

Capture the idea

Send the thought exactly as it comes out. We log the wording, tickers, and time so there is zero extra work.

The system breaks your idea and researches hypotheses, catalysts, contagion vectors, and relevant datasets so nothing is lost.

What you can do

Simple prompts work best.

Three common asks. Keep the wording short; we fill in the rest.

user@quant:~$ “Just saw on X there’s a new virus strain. Show me how to trade this information.”

Analyzing intent... [Done]

Detected: viral_outbreak_sentiment (High)

Correlations: PFE (+0.8), MRNA (+0.9), CCL (-0.7)

Strategy: Long Vaccine Basket / Short Travel & Leisure

Risk: High Volatility

Status: Ready for paper trade

user@quant:~$ _

Capabilities

Deterministic math core

Our proprietary model is fully deterministic—every decision is tied back to transparent math, with GenAI used only to keep the narrative clear.

Alternative data native

Satellite imagery, insider flows, private market prints, and niche feeds are on standby for any idea.

Deploy anywhere

Keep it on paper, send to your broker, or run everything on-premise if you need zero external footprint.

Full encryption, zero retention

Ideas are encrypted end to end, never retained for training, and can be wiped as soon as the drop lands.

Backtest & robustness

AI speed, deterministic backtests.

The engine reruns your idea across decades, swaps in alternative data, and flags when it overfits. Every report includes robustness notes so you know exactly why a strategy passes or fails.

Walk-forward and holdout sets keep the math honest.
Parameter sweeps punish complexity to avoid overfitting hero numbers.
Alternative data stress tests show how the idea behaves when reality disagrees.

Walk-forward run

We slice history into regimes, re-fit, and demand the strategy survives each segment before it ships.

Overfit guardrail

Parameter sweeps penalize complexity. If two settings tie, we pick the simpler one or reject the idea entirely.

Stress data injects

Adversarial alt-data, liquidity shocks, and gap scenarios are injected so the report shows how it fails, not just when it works.

Customer notes

“I speak a thought in the cab, there’s a finished trade waiting when I hit the office.”

Enmanuel Rodriguez

Managing Partner, Alphabeth

“I no longer need Financial Analysts to run my ideas, I can do it myself without needing to spend hours on it.”

Juan Gracia

CEO, Subdiagonal

“I do not have the tools as a retail trader to run my ideas, this is a game changer, I can operate like a Hedge Fund”

Pablo

Retail Trader (+20 years exp.)

Q&A

ChatGPT writes copy; it does not run deterministic math, blend alternative data, or produce deploy-ready code with risk baked in.