Test the idea
Start from a precise question, gather data, and fit the simplest model that could answer it.
[00] Independent trading firm
Steady principles.Open questions.
We develop and test ideas, deploy capital selectively, and evaluate every decision with discipline.
[01] Our view
Markets move. Narratives change. Much of what looks like certainty is noise. We built Stillgate around a still point: a process that holds steady, so our views are free to change.
[02] The name
A steady centre.
An open way forward.
Our symbol is an S built from two walls with a straight passage between them. At its exact centre sits a single still point.
It stands for how we want to work: keep your bearings, and keep moving. What we believe about evidence, risk and integrity stays fixed. What we believe about markets is always open to revision.
- A Two walls: discipline and risk control
- B The passage: an open way through
- C The still point: principles that don't move
[03] What we do
Four disciplines.
One process.
Questions before answers.
Every strategy begins as a precise, testable question. Researchers build models, look hardest for the reasons they might fail, and only then ask what they might be worth.
- Hypothesis-driven
- Out-of-sample by default
- Peer review on every idea
Built, not bought.
We build the research platform, simulation tools and execution systems ourselves, so the people asking the questions can control every assumption underneath the answers.
- Research and production share one codebase
- Reproducible experiments
- Monitoring built in, not added later
Signal is earned.
Data is where every idea is tested, and where most of them end. We treat datasets as instruments: understood, cleaned and versioned before anyone draws a conclusion from them.
- Provenance for every series
- Bias and leakage checks
- Clear limits on what data can say
Precise, patient execution.
Trading turns research into decisions. Positions are sized to the strength of the evidence, executed with care, and reviewed honestly afterwards, whatever the outcome.
- Risk limits set before entry
- Execution measured, not assumed
- Every decision reviewed
[04] Process
Inquiry. Judgment. Action.
The loop behind every position. The charts are illustrative, not real results.
Size the conviction
Weigh what the evidence shows against what it can't. Size to the strength of the work, not the appeal of the idea.
Execute and review
Execute within clear limits, measure the outcome, and feed what we learn back into the research.
[05] The Stillgate problem
Like puzzles?
So do we.
You flip a fair coin repeatedly. On average, how many flips does it take to see two heads in a row (HH)? And how many to see heads then tails (HT)?
Most people guess these are equal. Make your guess, then let the simulator check.
Monte Carlo0 trials
Why aren't they equal?
After a missed HH (you flip H then T), you're back to the start. After a missed HT (you flip H then H), you still have the H you need. The exact answers are 6 flips for HH and 4 for HT.
[06] Culture
“I don’t know yet” is a respectable answer here. “I’m sure” needs evidence.
V.01
Independent thought
Form your own view, then test it hard. Ideas are judged on their merits, not on who proposed them.
V.02
Small teams
Researchers, engineers and traders sit close to the problem and to each other.
V.03
Clear writing
If an idea can't be explained simply, it isn't ready to trade.
V.04
Ownership
Own the outcome, including what didn't work, and share what you learned.
V.05
Long horizons
We are building an institution, not a single trade.
V.06
Keep learning
Reading groups, internal talks and time to explore problems that don't pay off immediately.
[07] Careers
Work on hard problems with people who care about getting them right.
We hire for curiosity and rigour, not a specific background. These are the areas we hire in.
R.01Quantitative Research
+
You'll
- Turn questions about markets into testable hypotheses
- Build and stress-test models on large datasets
- Present and defend your work in open review
You might have
- A strong background in maths, statistics, physics or CS
- Comfort writing research code in Python
- A habit of asking how you could be wrong
R.02Software Engineering
+
You'll
- Build the research platform, simulation and execution systems
- Own reliability, performance and monitoring end to end
- Work directly with researchers and traders
You might have
- Depth in Python, C++, Rust or similar
- Experience with data-intensive or low-latency systems
- Taste for simple, well-tested code
R.03Data & Infrastructure
+
You'll
- Source, clean and version the datasets research depends on
- Build pipelines that are correct before they are fast
- Document what each dataset can and cannot tell us
You might have
- Experience with large-scale data engineering
- Strong SQL and a scripting language
- Obsessive attention to detail
[08] Contact
Get in touch.
For careers, partnerships or general enquiries, email us and we'll reply as soon as we can.