I'm looking for non-disk (at z = 0) subhaloes in two haloes of TNG50 which could have been disks in the past. To achieve this, I initially identify non-diks by using the circularity parameter CircAbove07Frac in the suplementary data catalogs at z=0. Then, I consider all those subhaloes with circulatiry fraction < 0.4 as non-disks. For the two identified haloes, by using this method I found 138 non-disks in one halo and 78 (both at z = 0) in the other one. Also, I'm neglecting all those subhaloes with M<1.0e8 solar masses. With all these I got the IDs of all these subhalos, because later on I like to use these IDs to follow the circularity of these non-disks backward in time. In essence, I'm using the IDs of the identified non-disk subhaloes in these two haloes to following the evolution of the circularity of these subhaloes back in time, as I consider that a non-disk was a disk when CircAbove07Frac > 0.4 for a specific redshift. The problem appears when I do this experiment I found that only 12 subhaloes were disks in the past for the halo with 138 non-disks at redshift = 0, and 1 non-disk subhalo for the halo with 78 non-disks at redshift = 0. This makes no sense to me as when I analyse the total number of discs vs redshift for the two haloes the number of disks I get is completely different (see the figure attached in the link).
Am I using the wrong parameter to estimate the 'diskiness' of a subhalo or am I understanding something in the wrong way?
Thanks in advance,
The approach you describe seems ok, i.e. following the CircAbove07Frac backwards along the main progenitor branch of satellite galaxies which are non-disky at z=0. Afraid I can't suggest much other than to double-check your codes for possible issues.
I will take a look again to double-check.