Do you take the leftmost tail data at a certain point rather than the whole distribution data to get the correct distribution.
Gumbel_density <- exp(-ts_rSPY)
