Assessing handwritten mathematics questions with multiple assessors is a daunting task regarding grading reliability. One of the primary sources of assessor variability arises when the grade resulting from the correction scheme does not align with their holistic appreciation of the student's answer. In this study, we developed a 'checkbox grading' approach that could possibly circumvent this variability source. In our approach, exam designers preset atomic feedback items with partial grades; next, assessors should tick the items relevant to a student's answer, allowing 'blind grading' when the underlying grades are not shown to the assessors. The study was executed in cooperation with the Flemish Exam Commission with 10 assessors and 30 students during a mathematics exam. This paper answers the question 'Does blind checkbox grading enhance inter-rater reliability?' In order to do so, we compared 'blind grading' with 'visible grading' and concluded that blind grading enhances the inter-rater reliability when the grading scheme is stringent.