Rank Function - by measure over multiple columns

I want to make a measure that returns the rank of the current Reason×Resource by [Qty_Sum], respecting page/report slicers.

    Qty_Sum =
[Qty_2025_W38] + [Qty_2025_W39] + [Qty_2025_W40]
DAXQty_2025_W37 =
SUMX (
    CALCULATETABLE (
        fct_stock_movements,
        KEEPFILTERS ( dim_date[calendar_week_year_label] = "2025-W37" )
    ),
    COALESCE ( fct_stock_movements[quantity], 0 )
        * COALESCE ( RELATED ( dim_cost[Total Cost] ), 0 )
)

• Working baseline (table expression): Built

  _tmpTbl = ADDCOLUMNS(SUMMARIZE(ALLSELECTED(fct_stock_movements), dim_reason[reason_code], dim_resource[resource_code]), "__Qty", CALCULATE([Qty_Sum])) and added "__Rank" = RANKX(_tmpTbl, [__Qty], , DESC, DENSE);

this produced correct, varied ranks inside the virtual table. But it wasn't affected by the context window.

• Current best attempt:

Rank Qty (Reason×Resource) = 
VAR _value = [Qty_Sum]
VAR _universe =
    ADDCOLUMNS (
        SUMMARIZE (
            CALCULATETABLE (
                ALLSELECTED ( fct_stock_movements ),
                REMOVEFILTERS ( dim_reason ),
                REMOVEFILTERS ( dim_resource )
            ),
            dim_reason[reason_code],
            dim_resource[resource_code]
        ),
        "__Qty", CALCULATE ( [Qty_Sum] )
    )
VAR _set = FILTER ( _universe, NOT ISBLANK ( [__Qty] ) )
RETURN
IF ( ISBLANK ( _value ), BLANK (), RANKX ( _set, [__Qty], _value, DESC, DENSE ) )

Obviously we want to see a clean context 1,2,3,4,5,etc but instead we get the strange 1,1,2,1,1 behaviour.

Any help would be much appreciated!