Strange frequencies reported on keyword list

[Pahor de Maiti, Kristina]

Dear noSkE team,

I compiled a keywords list where the focus is a subcorpus and the reference is the rest of the corpus (the option in the corpus). The keywords are given frequency numbers that are multiplied by 2, but only if I deselect the option A=a. Could you please help me understand why this happens? Please see the screenshots below for clarity.

The link to the list: https://www.clarin.si/skelog/#keywords?corpname=parlamint41_si&tab=advanced&icase=0&k_itemsPerPage=200&ktab=keywords&n_itemsPerPage=50&ref_corpname=parlamint41_si&ref_usesubcorp=%3D%3D%20the%20rest%20of%20the%20corpus%20%3D%3D&showavgstar=0&showcounts=1&showdocf=0&showarf=0&showaldf=0&showreldocf=0&showrelfrq=0&showresults=1&showscores=1&showwikisearch=0&t_itemsPerPage=50&usekeywords=1&usengrams=0&usesubcorp=ars_obstr&useterms=0&w_itemsPerPage=50&include_nonwords=0

IMAGE 1: A=a not checked --> the frequency of keywords is multiplied by 2 (Čatež = 306, glup=223). If I open the concordances for the first (or any other) keyword, the frequency is halved (Čatež=154, glup=113). See IMAGE 2.

IMAGE 2:

IMAGE 3: If I select the A=a option, the frequencies are correct.

Best regards,

Kristina Pahor de Maiti