Trading

A trading idea answered in a day, by a tool the trader trusts

A futures trader with twenty years of ideas and no code waited two weeks for each backtest, then had to decide whether to believe it. We built a research harness that first matched his own platform's results, then ran any variant of his strategy in one command. An idea became an answer in a day, and one he could act on.

Trusted by

matching his platform's own results first

Idea to answer

two weeks to a day

Code written by the trader

none

What was wrong

A trader with two decades of market judgment and a working strategy on his platform had plenty of ideas and no way to test them quickly. Each variant meant a slow backtest on the platform itself, and the result still had to be taken on faith. The bottleneck was never knowledge. It was the pace of the loop from idea to answer, and whether the answer could be trusted.

What we did

We rebuilt his strategy in a research harness of our own and refused to call it done until it matched his platform's results. Only then did we make it fast: any variant, any filter, any sizing rule, in one command, with the platform kept as the final check before anything goes live. He never touches code; he says what to try and reads the answer.

What it got them

An idea becomes a result he believes in a day instead of two weeks. The limiting factor is now his judgment, which is the only part that should limit it.

A fast loop that lies is worth less than no loop at all.

Anonymized. A private futures trader; name, platform and strategy withheld. Engagement led by Ryan Luccioni, principal.

← All work