Why pileup values of MACS2 output have decimal values?

Question

[Aspire]

I have ran macs2 using the command

macs2 callpeak -t ChIP.bam -c Control.bam -g mm --broad

The pileup values counts the number of reads at the peak summit (highest point). My treatment file has less reads the control file. As explained by Tao Liu, the value is after scaling the sample size. However, since in my case the larger file (control) was scaled toward the smaller file (treatment), why the decimal values in the pileup column of the excel file? The read counts of the treatment should not be affected by the scaling.