市場の相関関係を単に認識するだけでなく、分析に真に統合する方法を考える。

I've found looking at rolling correlations over different timeframes can highlight shifts. Sometimes a simple scatter plot over a recent period can also make a change in relationship pretty obvious.

That's a good distinction to make. Beyond just observing general trends, I find dynamic correlation metrics like rolling correlations over various lookback periods quite useful. They help identify when asset relationships are shifting, which can signal regime changes or early warning signs for diversification breakdown.