NOVOSKY Has a Type
We built a trading bot. Somewhere in 62 features and three gradient-boosted models, it developed preferences. Here's the unfiltered personality profile — backed by real numbers from live trades.
Read articleInsights & Updates
Journal
Deep dives into algorithmic trading, machine learning architecture, and the engineering behind NOVOSKY.
The Quantitative Research Arsenal: 17 Advanced Methods NOVOSKY Should Explore
A deep-dive survey of 17 quantitative methods — from GARCH and HMM regime detection to Transformers, GANs, Copula models, and NLP sentiment — each mapped to a concrete NOVOSKY improvement. With the math, the code, and zero hand-waving.
Read articleEverything NOVOSKY Is Still Getting Wrong
The bot runs. It makes real trades on real money. And we know exactly where it breaks. This is the unfiltered list of problems NOVOSKY is still actively fighting — because transparency is how you actually improve.
Read articleThe Science of Predicting Bitcoin: Quantitative ML Trading, Explained
Feature engineering, ensemble models, and out-of-sample validation — the three research pillars behind NOVOSKY's autonomous BTC/USD scalping. A deep dive that doesn't skip the hard parts.
Read article