The background is: when there are slicers of the same table category and subcategory at the same time in the report environment, they will trigger the Auto-Exist mechanism. Then use the matrix visual object in the evaluation environment to accumulate the values of the subcategories, where the row labels are the subcategories.
3. For the measure [Cumulative-Correct], using Iterator functions within a measure will change the Filter Context to a Row Context of the Table defined in the DAX expression. For more information on how to determine the context of iterators, you can refer to the following thread that I have provided detailed description.
2. For the measure [Cumulative-WhyCorrect], it differs from [Cumulative-WhyError]. Because the measure [SumOfValue] is referenced in [Cumulative-WhyError]. Using the measure [SumOfValue] inside another Measure will apply a context transition when using filtering functions.
Because [Cumulative-WhyCorrect] and [Cumulative-WhyError] the value of the CurValue variable in the two measurement values has been fixed after the first calculation. The CurValue variable will not continue to be calculated during the calculation of the FILTER function, so It will not trigger a row context transition.
Finally, from the results of [Cumulative-WhyCorrect], it can be known that whether to remove the category filter is not the key, because the value of the measure value is 11, which already contains the results of the two categories