Rolling average across a Distinct calculation showing incorrect results

Hi Everyone,

hopefully this should be a simple one...

Some background: my dataset is for a logistics company, where there is one huge fact table (shipments) and many dimentions tables including 'type' of cargo, loading location, delivery location, Load date, Delivery date, Vehicle size etc.

I'm trying to create a rolling 3months avarage on the number of unique vehicles used to make deliveries per month.

note: A single vehicle will make many deliverys each month and the specific requirement here is that a unique count (distinctcount) of the number of vehicles used per month is charted. 

Its really simple to chart the distinct number of vehicles used per month, however when coming to create the rolling average if this value across a 3 month period i am having issues with the filter context and instead of taking each individual monthly value and avarageing them, my measure is looking across the 3 month window and doing a distinct count on that figure instead, creating an incorrect dataset. heres the example data:

To highlight the issue - if we look at Aug-Oct the total no. of Vehicles across these 3 months = 82, therefore the avarage for that period should be 82/3 = 27.3333333 - however as you can see in the 3rd column (the measure i am trying to write) it is slightly out for October (i.e. the 3 month period aug-oct) at a value 26.67. This is the case for most of the values in the table, and i have tracked the issue to this incorrect filtering.

Here's my Dax, it references the following tables:

- Shipments - the main fact table that holds all the shipment records as mentioned above

- PlacesD - list of countries and depots for Destination
- PlacesL - list of countries and depots for Loading

- DateTable - my Calendar table

m_rAvg12_NoVehicles_Month = 
var month = 3

return
CALCULATE(
	  DIVIDE(
                 CALCULATE(
			   CALCULATE(
                                     DISTINCTCOUNT(Shipments[ImoNo]),//distinct count on the vehicle id
                                     filter(PlacesD, PlacesD[CountryCodeD] <> DISTINCT(PlacesL[CountryCode]))//filter to remove domestic shipments
                             ),
			all(Shipments[LoadDate_DateOnly]), //Make sure any other filtering on other measures that use this is removed
					
			FILTER (
				all ( DateTable ),//use the calendar table
								
			        DateTable[firstDayOfMonth] >= EOMONTH(MAX(Shipments[LoadDate_DateOnly]), -month) //create window of time
				&& DateTable[firstDayOfMonth] <= MAX(Shipments[LoadDate_DateOnly])//Cap the window at most recent date 
								
			      )
		),
                month),
		
            DATESBETWEEN( //trimming off incomplete months from the beginning
			Shipments[LoadDate_DateOnly],
			DATEADD(FIRSTDATE(all(Shipments[LoadDate_DateOnly])),month-1,MONTH), //start date
			LASTDATE(DateTable[Date]) //end date
	    )
				
)

 It's worth mentioning that the result will be filtered by many of the dimentions tables, i attempted to solve this problem with a 'summarizecolumns' into a new table hower the relationships felt like an issue when trying to use the slicers.

I feel like i have made a stupid error somewhere - if ihave missed any detail then please do say and ill happily fill in the blanks.

Please help me see the light!!
Hoping you can help

Mat