量化点评报告：传媒、电子进入超配区间，哑铃型配置仍是最优解

• The industry mainline model uses the Relative Strength Index (RSI) indicator to identify leading industries. The construction process involves calculating the past 20, 40, and 60 trading days' returns for 29 primary industry indices, normalizing the rankings, and averaging them to derive the final RSI value. Industries with RSI > 90% by April are likely to lead the market for the year[11][13][14] - The industry rotation model is based on the "Prosperity-Trend-Crowdedness" framework. It includes two sub-models: the industry prosperity model (high prosperity + strong trend, avoiding high crowdedness) and the industry trend model (strong trend + low crowdedness, avoiding low prosperity). Historical backtesting shows annualized excess returns of 14.4%, IR of 1.56, and a maximum drawdown of -7.4%[16][18][22] - The left-side inventory reversal model focuses on industries with low inventory pressure and potential for restocking. It identifies sectors undergoing a rebound from current or past difficulties. Historical backtesting shows absolute returns of 25.9% in 2024 and excess returns of 14.8% relative to equal-weighted industry benchmarks[28][30][29] - The industry ETF allocation model applies the prosperity-trend-crowdedness framework to ETFs. It achieves annualized excess returns of 15.5% against the CSI 800 benchmark, with an IR of 1.81. The model's excess returns were 6.0% in 2023, 5.3% in 2024, and 7.7% in 2025[22][27][16] - The industry prosperity stock selection model combines industry weights from the prosperity-trend-crowdedness framework with PB-ROE scoring to select high-value stocks within industries. Historical backtesting shows annualized excess returns of 20.0%, IR of 1.72, and a maximum drawdown of -15.4%[23][26][16] - The industry prosperity-trend model achieved excess returns of 3.9% in 2025, while the inventory reversal model showed absolute returns of 1.3% and excess returns of -2.1% relative to equal-weighted industry benchmarks[16][28][30]