In 2018, Chen et al. proposed the minimum first method (MFM) [1] to speed up the calculation of dynamic time warping (DTW), which is a well-known and essential step for solving the time series classification (TSC) problem. By simply rearranging the calculation order, MFM returns the optimal answer as the original DTW does with less calculation time. MFM is effective in most of experimental datasets, but it may be worse than the original DTW in some other datasets. In this paper, we present two quantitative indicators, including standard deviation of variations and wave oscillation, to automatically determine that which datasets are suitable for MFM. The most of suitability prediction accuracies are higher than 80%. Furthermore, we apply MFM with other DTW related methods to design hybrid methods, including DTW with band constraints and AWarp: warping distance for sparse time series and discuss their time efficiencies of those hybrid methods. In all experiments, our hybrid methods save different amount of time, from 4% to 62%.