Free & hands-on
Build it yourself, in Python.
The practical track: doing quant with a laptop and a brokerage account. Real, runnable Python for environments, data, backtests, and execution, with the focus on the parts that actually decide whether retail algo trading works: data hygiene, honest backtesting, and cost-aware execution.
Getting Started
The practical path into quant, done properly: a reproducible Python environment, honest free data, a backtest built from first principles, and the biases that quietly destroy most retail results.
- 1
Set Up a Python Quant Research Environment
Build a reproducible Python research setup, isolated environments, pinned lockfiles, a sane project layout, deterministic seeds, and vectorized code, so a backtest you run today produces the same number a year from now.
30 min
- 2
Getting Market Data for Free (and Its Traps)
Pull equities, macro, and crypto data from free sources with real code, store it efficiently in parquet, and, more importantly, learn the silent traps (adjusted vs raw close, survivorship, restated data, gaps, timezones) that quietly corrupt backtests.
40 min
- 3
Cleaning Financial Data in pandas
The pandas patterns that actually matter for quant research, computing returns correctly, resampling to bars, aligning multiple assets, handling missing data, and winsorizing, with a hard focus on the subtle lookahead bugs that hide inside innocent-looking preprocessing.
40 min
- 4
Your First Backtest, Done Right
Build a vectorized backtest in pandas from first principles, signal, position, costs, and metrics, then see exactly how a one-line lookahead bug manufactures fake alpha.
45 min
- 5
The Backtest Traps That Fool Everyone
The six systematic ways a backtest lies to you, look-ahead, survivorship, data-snooping, ignored costs, in-sample overfitting, and a p-hacked split, each demonstrated with runnable code that shows the bug inflating results, then the fix, then the size of the damage.
50 min
- 6
From Backtest to Paper to Live
The realistic path from a clean backtest to real money, why paper trading flatters, how to place orders through a broker sandbox without hardcoding keys, modeling slippage and execution lag, reconciliation and idempotency, implementation shortfall, risk of ruin, and the kill-switches that keep a live system from destroying your account.
45 min
Build It Yourself
Code-along tutorials that build the strategies from the database from scratch, signal, portfolio, costs, and evaluation, so you understand every line, not just the equity curve.
- 1
Build a Dual-Momentum Strategy in Python
Implement Gary Antonacci's dual momentum end-to-end, relative ranking plus an absolute-momentum cash filter, on a cross-section of ETFs, with monthly rebalancing, turnover costs, and honest metrics.
60 min
- 2
Build a Pairs Trade with Cointegration
Implement a statistical-arbitrage pair end-to-end, Engle-Granger cointegration test, OLS hedge ratio, a z-scored spread, a mean-reversion rule, and an OU half-life, all in pure numpy/pandas, then watch a structural break kill it.
60 min
- 3
Build a Trend-Following System
Implement a CTA-style time-series momentum system across a multi-asset universe, a trend signal, per-asset volatility scaling to a target risk, portfolio combination, costs, and see the positive skew and crisis-alpha convexity that make trend a portfolio diversifier.
60 min
- 4
Build a Short-Term Mean-Reversion Strategy
Implement a cross-sectional short-term reversal strategy on equities, rank by 5-day return, short the winners and buy the losers, market-neutral, daily, then confront the elephant: reversal's edge is real gross but is eaten alive by turnover costs.
55 min
- 5
Add a Volatility-Targeting Overlay
Take any strategy's return stream and scale exposure to hold volatility constant, estimate rolling risk with EWMA, lever and delever to a target, and measure the effect on Sharpe, drawdown, and turnover, plus the gap risk that leverage buys you.
45 min
- 6
Build a Cross-Sectional Factor Screen
Combine value and quality into a composite equity signal, z-score each factor across the cross-section, blend, form a long-short book, and evaluate it the way a quant actually does: with the information coefficient and the fundamental law of active management.
60 min
These guides are the practice to the learning tracks' theory, each links back to the concepts it uses and forward to the strategies you can build with it.