This would be a matter of data modeling rather than a measure implementation.
When you present the difference between two versions, you first need to be able to place the two versions to compare on your report like the image below. Thus you need to have two different columns which you can select on slicers or matrices, etc.
My approach is to create a new table which duplicates each row with its successor versions. For instance, you will replicate the row of V110 into four successor rows: V111, V112, V113, and V115. In this new table, each row represents every possible pair of versions which you want to compare. Then you can calculate the difference of the two versions in each row. The new table would look like below.
This table could be created like below using calculated table for instance. You should also be able to do that using Power Query.
// DAX calculated table
// In this code snippet, the original table is referred to as "Versions".
Version Comparison = GENERATE (
SELECTCOLUMNS (
Versions,
"Version Before", [version],
"Points Before", [count_of_points]
),
SELECTCOLUMNS (
FILTER ( Versions, Versions[version] > [Version Before] ),
"Version After", [version],
"Points After", [count_of_points],
"Points Diff", [count_of_points] - [Points Before]
)
)
Then, it might be a good idea to define a measure for the diff, which is just referring to the "Point Diff" column value only when a single pair of versions is selected.
// DAX measure
Diff Measure = SELECTEDVALUE ( 'Version Comparison'[Points Diff] )
Output images
