vskokov/four-point-ai-trader

The 'four-point-ai-trader' is a representative example of a modern 'AI wrapper' for quantitative finance. It earns a low defensibility score (2) because it lacks any proprietary data, network effects, or novel algorithmic breakthroughs. The core components—Hidden Markov Models for regime detection, Kalman filters for pairs trading, and Kelly Criterion for sizing—are academic standards in quantitative finance. While the integration of LLM sentiment (via Ollama/Gemma) as a signal input is timely, it is a pattern being rapidly commoditized by both the open-source community (e.g., projects like FinGPT) and commercial platforms. With 0 stars and forks at the time of analysis, it functions more as a personal portfolio piece or a blueprint than a defensible software product. The displacement horizon is short because frontier labs (OpenAI/Anthropic) are increasingly capable of generating this exact boilerplate code on demand, and specialized trading platforms like Alpaca or QuantConnect are moving toward native 'agent' integrations that will render standalone execution wrappers like this obsolete.