When Does Volatility-Informed Position Sizing Matter Most? Policy-Rate Dependence and Diversification Arithmetic of Multi-Strategy Trading in Moex Index Stocks

Abstract

This paper decomposes the performance of volatility-informed multi-strategy trading in constituents of the MOEX index, using six systematic strategies on 17 liquid stocks that account for about 80% of Russian equity market capitalization, evaluated walk-forward over 2022–2025. Two results emerge. First, diversification arithmetic: with an average pairwise return correlation of 0.09 and negatively correlated trend and counter-trend blocks (−0.24 to −0.11), the six strategies deliver approximately 4.2 effective independent bets and a diversification ratio near 2.0, so the choice of weighting scheme is second-order, and equal weights are best for four of six strategy portfolios. Second, policy-rate dependence: as the Bank of Russia key rate rose from 9.9% to 21.0%, the annual Sharpe ratio of the unassisted system fell from 2.99 to 0.40 while the regime-filtered system never fell below 1.87, widening the relative advantage of forecast-based position sizing from about 1.7 to 4.7 times. Volatility-informed sizing matters most when a high policy rate makes unassisted active trading hardest.