Problem in SQS

Hi ATAT users,
I am trying SQS to build an alloy compound that is Li4Mo1O5 and has 10 atoms in its smallest supercell,i.e., primitive cell.
I have four interchangeable sites between Li and Ag as you can see from the rndstr.in file below. I could successfully generate the clusters.out file using the following command:

corrdump -l=rndstr.in -noe -nop -clus -2=6.0 -3=6.0

The problem is when I used the command "mcsqs -n=20", I got the following message
Impossible to match point correlations due to incompatible supercell size.

I assume, 20 should satisfy the general rule that composition ratio times the total number of atoms = integer, i.e.:

0.75x20 = 15 and 0.25x20 = 5.
Therefore, this should not be a problem. However, unless I increase the number of atoms in my supercell to 40, I will get the same error. I mean, I get the same error even for the -n 30 case.

This is somehow strange because both 20 and 30 satisfy the requirement, then why I am getting this error?

Here is my rndstr.in file:

4.2757319400 0.0000000000 0.0000000000
-2.1361339865 4.6652067104 0.0000000000
-2.1372943438 -0.0035571530 4.6661983844
1.0 0.0 0.0
0.0 1.0 0.0
0.0 0.0 1.0
0.8131070000 0.4120270000 0.2164800000 Li=0.25,Ag=0.75
0.5976290000 0.7867680000 0.4092130000 Li=0.25,Ag=0.75
0.3984640000 0.2110110000 0.5868300000 Li=0.25,Ag=0.75
0.1883590000 0.5892170000 0.7903630000 Li=0.25,Ag=0.75
0.9504540000 0.0015830000 0.9987240000 Mo
0.2885690000 0.3636240000 0.1858930000 O
0.1023680000 0.8161600000 0.3605530000 O
0.9267730000 0.1883120000 0.6375500000 O
0.7409660000 0.6405500000 0.8131770000 O
0.5192230000 0.0017430000 0.9989500000 O

Any help will be appreciated.

As answered many times on this forum, your randomly occupied sites are not symmetrically equivalent, so each site individually should have the right composition in order to have a proper SQS.

Look at the rndstrgrp.out file: each site is listed as a separate group, so your minimum cell size is
40.