fragment-namd result

[佐藤紅介]

I did the tutorial of fragment-namd and found the timescale of electronic transfer is around 30fs under no decoherence effects. In the paper, http://pubs.acs.org/doi/abs/10.1021/acs.jctc.6b00955 , however, the timescale is around 200fs, which is larger than the result of the tutorial.
I didn't modify the default conditions of the script; so they seems to be epual.
Do I have to modify the script in order to reproduce the results in the paper?

note:
The results discussed are with adiabatic basis and obtained by MSSH algorithm.