Norm

I am in the process of learning more indices for FELT. I have a problem finding out the correct variance.
Here is an example with just 12's:

12V2 5 7
12V3 2 5
12V4 0 2
12V5 -3 -1
12V6 -7 -4

The first column of indices is given when I compute them using CVBJ. The second column is obtained from CVData.
I made sure that I used the same input to each program which is:

2D H17, DOA, DAS, RSA, Dealer peaks 10&A, No surrender - Asked for full indices.

What am I doing wrong and which is correct?