This study introduces a refined approach to measuring dynamic connectedness using a time-varying parameter vector autoregressive (TVP-VAR) model, which captures changes in the underlying data structure more flexibly and robustly. Unlike traditional rolling-window VAR models, the TVP-VAR model does not require an arbitrary rolling-window size, avoids the loss of observations, and is less sensitive to outliers. The authors conduct Monte Carlo simulations to demonstrate the superiority of the TVP-VAR model over rolling-window VAR models in terms of parameter accuracy and robustness to outliers. They apply the TVP-VAR model to analyze dynamic connectedness in the foreign exchange market, comparing it with rolling-window VAR models for different window sizes. The results show that the TVP-VAR model adjusts more quickly to economic events and provides more accurate estimates. The study also proposes confidence intervals for dynamic connectedness measures and evaluates the forecasting performance of the TVP-VAR model, finding that it outperforms rolling-window VAR models. The empirical analysis reveals that the Euro and Swiss Franc are significant transmitters of shocks, while the British Pound and Japanese Yen are significant receivers. The study contributes to the literature by providing a more detailed and complete illustration of dynamic connectedness measures and enhancing the understanding of financial market interactions.This study introduces a refined approach to measuring dynamic connectedness using a time-varying parameter vector autoregressive (TVP-VAR) model, which captures changes in the underlying data structure more flexibly and robustly. Unlike traditional rolling-window VAR models, the TVP-VAR model does not require an arbitrary rolling-window size, avoids the loss of observations, and is less sensitive to outliers. The authors conduct Monte Carlo simulations to demonstrate the superiority of the TVP-VAR model over rolling-window VAR models in terms of parameter accuracy and robustness to outliers. They apply the TVP-VAR model to analyze dynamic connectedness in the foreign exchange market, comparing it with rolling-window VAR models for different window sizes. The results show that the TVP-VAR model adjusts more quickly to economic events and provides more accurate estimates. The study also proposes confidence intervals for dynamic connectedness measures and evaluates the forecasting performance of the TVP-VAR model, finding that it outperforms rolling-window VAR models. The empirical analysis reveals that the Euro and Swiss Franc are significant transmitters of shocks, while the British Pound and Japanese Yen are significant receivers. The study contributes to the literature by providing a more detailed and complete illustration of dynamic connectedness measures and enhancing the understanding of financial market interactions.