For the life of me I can't understand how JIRA counts the number of unestimated issues in order to calculate the % of issues that aren't estimated.
For example, I'm looking at my Release Burndown report for Release 1.0 and it says '12% unestimated issues'. I hover my mouse of that label and it says '12% of issues do not have an estimate (31 of 261)'. However, when I scroll down to see all of the stories in each sprint there are only 18 that don't have story points, not 31.
Why does JIRA think there are 31?
I've read the documentation (https://confluence.atlassian.com/display/AGILE/Viewing+the+Release+Burndown) 20 times but haven't found the answer.
Thanks.