CVNA DEMA–SRPR Walk-Forward Trading Strategy (2020-2026)
CVNA DEMA–SRPR Walk-Forward Trading Strategy (VectorBT + Monte Carlo Validated)A fully reproducible, rule-based trading strategy for CVNA (Carvana) built with Python and VectorBT, combining trend filtering and momentum exhaustion signals with Walk-Forward Optimization and Monte Carlo robustness testing. This product is designed for systematic traders who want more than a simple backtest — it delivers a research-grade validation workflow.The strategy uses a Double Exponential Moving Average (DEMA) filter for trend-aligned entries and a Smoothed Percent Rank (SRPR) momentum model for statistically extreme exits. Parameters are not fixed blindly — they are selected through rolling walk-forward optimization windows and then tested on unseen data segments to reduce overfitting risk.Included is a complete research pipeline: parameter grid search, walk-forward testing, combined out-of-sample portfolio construction, benchmark comparison, statistical testing, and block-bootstrap Monte Carlo simulations that generate distributions of return, drawdown, and Sharpe ratio. You can see not just performance — but how fragile or robust that performance may be under randomized return paths.This product is ideal for quant learners, systematic traders, and developers who want a transparent, modifiable trading model with professional validation structure. All code is clean, editable, and ready to run.Walk-Forward Portfolio MetricsFrom provided WFO run: Total Return: ~1200% (Buy and Hold Only ~400%) Trades: 33 Win Rate: ~84% Max Drawdown: ~98% Profit Factor: ~3.26 Sharpe Ratio: ~1.08 What you get: Complete Python strategy code VectorBT backtest implementation Walk-Forward Optimization engine Parameter grid search framework Combined out-of-sample portfolio builder Buy-and-Hold benchmark comparison Monte Carlo block bootstrap simulator Risk distribution charts and statistics Strategy parameter sets (original + optimized) Built for education, research, and systematic strategy development. Not a black box — fully inspectable and extensible.
Get it → kryptera.gumroad.com