Eugen Soloviov
Trading-systems engineer
Trading-systems engineer building bots since 2017: cross-exchange arbitrage (connected up to 30 venues), cointegration-based pairs arbitrage across spot and futures, scalping, news and sentiment-driven strategies, trend algorithms, and portfolio management and balancing algorithms. Also builds sub-millisecond order execution, big-data warehouses, backtesting engines, AI agents, and trading interfaces (incl. open-source profitmaker.cc). Stack: JS/TS, Python, Rust/Zig/Go, DevOps, backend, frontend, architecture.
Articles
The Deflated Sharpe Ratio: How Many of Your Backtest 'Winners' Survive Multiple Testing?
A parameter search is a machine for manufacturing luck. On pure noise — 1,000 strategies with zero true edge — the best annual Sharpe averages 1.63 and the naive significance test flags a discovery 100% of the time. We build controlled ground truth and show that the Deflated Sharpe Ratio, the Harvey-Liu haircut, and White's Reality Check restore honesty: false discoveries drop from 1.000 to 0.001-0.057, genuine edges above the noise ceiling are kept with power ~1 — and one real trap (correlated grids) where the raw DSR over-deflates and the verdict must be read across a whole band of effective-trial estimates, not one.
Objective-Function Design: The Metric You Optimize Secretly Picks Your Strategy
To search for the 'best' strategy you must first define 'best' — and that scalar silently chooses the winner. On synthetic data with a known edge (600 seeds, T=2000, 80 thresholds), a naive per-trade Sharpe crowns a lottery: it picks a sub-5%-exposure winner in 56% of seeds and degenerates in 57% — on the starkest seed, 8 trades posting an in-sample Sharpe of 21.09 that collapses to 0.13 out of sample. The honest repair is almost dull: measure on the full timeline, which never degenerates (out-of-sample 1.71). A trade-count (conf_k) shrinkage and an exposure floor can retrofit a per-trade metric, but even fully repaired they only match full-timeline Sharpe (1.70 vs 1.71) — never beat it. Goodhart's law, in a backtest, with controlled ground truth.
Look-Ahead Bias: How a One-Bar Mistake Manufactures a Sharpe of 15 From Pure Noise
A controlled study of the subtle look-ahead leaks that quietly inflate backtests. With zero real edge, a same-bar fill manufactures an annualized Sharpe of +14.8 out of pure noise; a one-bar indicator peek, +4.8. The taxonomy, the measured magnitudes, and how to detect each leak before it costs you money.
The Backtest Speed Ladder: 298x on a Laptop CPU, Identical PnL to the Last Trade
Five implementations of the same 80-combo parameter sweep, all verified to produce identical PnL: pandas rolling.apply takes 69.9 seconds, numpy 3.1, numba 2.0, parallel numba 0.23 — a measured 298x speedup on an Apple M2 Max with zero hardware changes, and still ~13x over a competent vectorized baseline. What each rung buys, why a GPU is not the missing piece, and where the real bottleneck in mass parameter search lives.
researcher: A Searchable Quant-Research Archive for Humans and AI Agents
How we built researcher.marketmaker.cc — a unified, full-text-searchable archive of quant research (arXiv papers, GitHub repos, quant blogs, TradingView Pine scripts) that humans browse and AI agents query over MCP.
algo-investor-skills: Claude Code Skills That Build a Scam-Proof Investor Proposal
A deep look at algo-investor-skills — a set of Claude Code skills that take an algotrading strategy from raw measured facts to an audited, honesty-forward investor proposal. Six composable skills, a financial-models engine, an independent-verification proof pack, and a mandatory skeptical-investor audit gate that never fabricates a number.
The Kelly Criterion for Strategies: How to Size Positions and Allocate Capital
A strategy with positive expected value can still blow up your account if you get the bet size wrong. We walk through the Kelly criterion from deriving the formula to a portfolio of strategies: why full Kelly is dangerous, how fractional Kelly captures 75% of the growth at half the volatility, and an interactive calculator that shows how the Kelly fraction moves return and risk.
Daily Stock Analysis: An AI System That Turns a Watchlist Into a Daily Decision Dashboard
A deep dive into daily_stock_analysis by ZhuLinsen — an open-source system that fetches market data across A-shares, HK, US, and more, runs technical and news analysis through an LLM, and pushes a structured 'decision dashboard' to your messenger every trading day. Architecture, data fallback, agent strategies, limitations.
Temporal Fusion Transformers for Multi-Horizon Portfolio Forecasting
How Google's Temporal Fusion Transformer brings interpretable multi-horizon forecasting to quantitative portfolio management, with attention-based variable selection, quantile outputs, and a worked pytorch-forecasting pipeline.
Conformal Prediction for Risk-Aware Position Sizing
Distribution-free prediction intervals with guaranteed coverage. We use split conformal, jackknife+, and adaptive conformal inference to calibrate trading risk and size positions without parametric assumptions.
Bid-Ask Spread Modeling and Prediction with Machine Learning
Decomposing and predicting bid-ask spreads with ML — from Roll's implicit estimator to gradient boosting and neural networks — with the units, leakage, and benchmarking pitfalls that bite in production.
DeepLOB: Deep Learning on Limit Order Books
How DeepLOB combines a CNN, an inception module, and an LSTM to predict mid-price moves from raw order book data — the architecture, the real FI-2010 numbers, and a working PyTorch reimplementation.