I’m pretty sure this is a known bug going back to the early days of knime - behind the scenes, the row indices are 0-based, but I think the thinking was that a 1-based numbering would be more intuitive to users, and so the displayed value in the node settings is actually the stored value + 1. You can see the offending source code here:
I think that is unfortunately one that has been there so long that it is not going to go away because it would break too many workflows
As @s.roughley said this is a known bug and I have added +1 to it. Additionally to mention new Row Filter node is planned that will also take care of this issue. In a mean time I will check about adding this info to node description.
Edit: Created a ticket to add this information into node description.