Momentum,
systematized.
Quant-Tech-Rotator ranks ten mega-cap tech stocks by momentum every month, keeps the top three, and steps into cash when the market turns. Five rounds of research went into it before a single trade went live.
Five phases to a live signal.
01 / Awareness
Retail indicators lag by design.
RSI and MACD react to price action that has already happened — useful for reading a chart, not for building an edge.
02 / Adaptive volatility
Markets don't hold still.
They shift between regimes, and a static strategy tuned for one regime tends to break in the next.
03 / Alpha decay
Public formulas get arbitraged away.
Institutional multi-factor models — like the GTJA191 dataset — lose their edge over time as high-frequency trading prices them in.
04 / Multi-asset architecture
A custom cross-sectional model.
Backtested on the S&P 500: a 1.21 Sharpe ratio and +40% annual return in 2024, with minimal drawdown.
05 / Paper tradingCurrent
Live data, simulated funds.
Real-time market data through the Alpaca API — testing execution, slippage, and latency before any real capital is involved.
The logic, monthly.
Universe
Ten mega-cap tech tickers, scanned every month.
Rank
Sorted by 60-day rate of change — strongest momentum first.
Filter
Negative-momentum names are dropped. If all ten are negative, the bot holds 100% cash.
Select & allocate
The top 3 survivors split the portfolio equally — about 33.3% each, via fractional shares.
A real rebalance.
Actual output from a monthly run.
Run it yourself.
Clone the repository
git clone https://github.com/mirconegri/Quant-Tech-Rotator.git
Set up the environment
python -m venv venv then pip install -r requirements.txt
Add your Alpaca keys
Set ALPACA_API_KEY and ALPACA_SECRET_KEY as environment variables. Never commit them.
Run it
python monthly_executor.py — once a month, ideally an hour before market close.
This is paper trading — simulated funds, not real capital. The script is also stateless: every run closes existing positions and rebuys from scratch, so running it more than once a day will bleed money to spread and slippage. Once a month, by design.
Built in the open, tested in public.
Look at the code, check the math, or fork it and run your own universe.