Algorithmic Trading with Python

Implement a simple momentum, trend-following model called Alexander's filter in Python, trading long or short based on price reversals and running highs and lows on the S&P 500.

Explore Alexander's asymmetric filter with two stop losses—long side at 10.5% and short side at 13.5%—demonstrating in-sample gains and out-of-sample risks, with SciPy optimization for robustness.

Alexander's filter uses dynamic, vix-based stop losses—long side at 0.7 of the vix and short side at 1.2 times the vix—enhancing momentum trades on the S&P 500.

Build a breakout model by defining a binary dependent variable for uninterrupted uptrends or downtrends using a 10% threshold, and compute it with Python in two parts using for loops.

Apply logistic regression to predict binary breakout signals using relative position regressors, set a 0.5 decision threshold, and test predictions on in-sample and out-of-sample data for a breakout trading rule.

Define two-rule trading system from logistic regression signals, entering long at 0.7 and short at 0.3, with exits at 0.3 and 0.7, and optimize in-sample and out-of-sample returns.

Learn a mean reversion pairs trading rule for Brent crude oil and West Texas Intermediate, using logarithmic returns, a 30-day moving average, and a two-sigma trading signal.

Learn a mean-reversion pairs trading rule using standard deviation thresholds to trade Brent and WTI in Python, evaluate wealth index and max drawdown, and test parameter sensitivity.

Compute maximum favorable and adverse excursions in the factor model data frame, derive adjusted u max/min, then calculate u edge and d edge to guide long or short trading signals.

Apply a factor model rule: enter long at 0.9 and short at -0.9, exit at 0.6 and -0.6, with 5% stop, and assess cumulative return and maximum drawdown out-of-sample 2018–2022.

Apply the Kelly criterion to set the optimal bet size in trading, using win probability, capital fraction, and fractional Kelly to avoid ruin.

Backtest a gold-silver pairs trading model using a 120-day moving average and two standard deviations, and assess performance with rolling correlations and trade-level results.

Demonstrates a Python-based pairs trading model using Brent and WTI prices, computing log returns, moving averages, and z-scores to generate trading signals and backtest performance.

Explore the factor model for algorithmic trading, detailing data prep with pandas, numpy, and plotly, forward-looking USD/CLP prices, and maximum excursions with percentiles and z-scores.