Descriptions:
Michael Automates presents an updated version of the Gaussian Channel algorithmic trading strategy — version 3.3 — now featuring shorting logic developed in collaboration with Claude Opus 4.6 and a custom Python backtesting engine that replicates TradingView’s calculations precisely. The central claim is that adding short-selling signals improved risk-adjusted performance on select assets: Bitcoin’s profit-and-loss figure nearly doubled from 2,700 to 5,600 while maximum drawdown decreased from 30% to 25%, and AEX saw roughly double the profit at equivalent drawdown.
The development workflow involved Claude autonomously generating multiple strategy variants, running backtests against downloaded OHLCV data, and selecting the top performer across several rounds of iteration. The video compares V3.1 (long-only) versus V3.3 (shorts enabled) across Bitcoin, AEX, Fetchai, Solana, and ADA — with mixed results. ADA actually performed worse with shorts enabled, both in profit and maximum drawdown, illustrating that the upgrade is asset-dependent rather than universally beneficial.
Access to V3.3 is provided through the Signnum platform, which supports automated execution on Hyperliquid and Bybit futures exchanges. The video also introduces HIP-3 support on Hyperliquid, enabling the strategy to trade tokenized stocks and metals alongside crypto assets. All results are presented as educational content; the presenter explicitly notes this is not financial advice.
📺 Source: Michael Automates · Published February 19, 2026
🏷️ Format: Showcase
