when I look at the cross-validation charts, there is a discontinuity between the last input step and the first predicted step, how can I mitigate this?

for cutoff in cv_df_drop['cutoff'].unique():StatsForecast.plot(

Y_df_drop,

cv_df_drop.query('cutoff == @cutoff').drop(columns=['y', 'cutoff']),

max_insample_length=cross_test_size,

unique_ids=['BTCUSDT'], #['ETHUSDT'], #!!! ['FTMUSDT']

level=[80],

engine='matplotlib'

)