We don’t predict markets. We build mechanical systems that survive them.
Mechanical systems for self-directed traders. Rules you can read, test and run yourself, as Python code on end-of-day data.
At a glance
- Data
- End of day
- Code
- Python 3.10+
- Broker link
- None
- Free systems
- Two
Researched, tested and written down
End-of-day systems for US stocks and funds. Two are free; two will be paid code releases.
- Free
Harrier_SPY
Long/flat time-series trend overlay
- Free
Owl_SPY
Overnight hold after a weak session
- $249
Crane_ETF
Monthly rotation across a fixed basket of ETFs
- $499
Stork_NDX
Cross-sectional momentum with a volatility target
Where our work stops and yours begins
We write and test the systems. Everything that touches your money stays with you.
-
01Your decisionsYou
Your decisions
Whether, when and how much to trade. Nobody here tells you what to do.
- Which systems to run, if any, and with how much money
- When to start, pause or stop a system
- How much risk your account can take
-
02Your data and your brokerYou
Your data and your broker
You run the code on market data you are licensed to use, and place every order in your own account.
- Market data from a source you are licensed to use
- A brokerage account in your own name
- Every order, placed by you
-
03Our code and documented rulesStratoQuant
Our code and documented rules
Readable Python and the rules in plain language, with every assumption stated.
- Python source code you can read and change
- The rules in plain language, beside the code
- Every assumption stated: costs, slippage, data and test period
-
04Our research processStratoQuant
Our research process
Rules written before they are tested, checked on data they were not built on, and left alone once released.
- Rules written down before they are tested
- Checked on data they were not built on
- Left alone once released; any change comes as a new version
No access to your account or your money crosses this line.
What you get with every system
- Readable PythonPlain, commented code you can read line by line and change.
- The rules in plain languageEvery entry, exit and position size written out in words, beside the code that runs it.
- Every assumption statedCosts, slippage, data and test period, so you can check the result yourself.
- End-of-day data, no broker linkIt reads daily price files. It never connects to a broker or places an order.
// FilesHarrier_SPY-Ver00.zip
README.md # the rules, in plain language harrier.py # the code that runs them LICENSE.md # what you may do with it requirements.txt
// RunPython 3.10+
pip install -r requirements.txt python harrier.py prices.csv python harrier.py prices.csv --status
See what happens when you run them together
Systems that fall at different times can soften each other's worst stretches. The combining tool on the systems page lets you pick systems and weights, and shows the mix, its drawdown and how closely each pair moved together. Every figure there is a hypothetical backtest.
Most systems fail on the human, not the math
A rule you abandon in a drawdown was never a rule. A backtest you fitted until it looked good was never evidence. A process you cannot write down is not a process. It is a mood.
Everything here works the other way round. Write the rules before you trade them. Validate them against data you did not use to build them. Then let the mechanical system do what you said it would do, especially on the days you would rather it didn’t.
Rules over feelings. Systems over stories. Build the mechanical system, trust the process.
For firms
Engineering and licensing
StratoQuant builds rule-based portfolio tooling and the research infrastructure it depends on (backtesting environments, data pipelines, applied AI work), and licenses its models and tools to firms that make their own decisions.
- 01System developmentPortfolio and risk tooling, handed over as documented code.
- 02Research infrastructureBacktesting environments, data pipelines, validation harnesses.
- 03Model and tool licensingOur research models and tools, for firms that make their own decisions.
- 04Applied AI and machine learningWhere it helps, and where it is an expensive way to overfit.
Writing
- O-01 · portfolio risk
The Holy Grail is uncorrelated systems
- B-05 · execution
Buy the close, sell the open
- B-03 · system design
Mean reversion is not a law of markets
Start with the free code
Ask for Harrier_SPY or Owl_SPY and we will email you both: the rules in plain language and the Python code that runs them.
Running a desk or a fund? Start a conversation