Matrix sub-total rows not seemingly numerically associated with underlying data

PBIX FILE: https://drive.google.com/file/d/1j_4CrUj0fPh0xK6hnO7OZ-uIWkoP5Kdg/view?usp=sharing

________________________________________________

***UPDATE***

Identified one of the offending measures:

When OBSERVED_DEPRECIATION is the Matrix output, and it's programmed as follows: 

VAR OBSERVED_DEPRECIATION = 
        SUMX(
            SUMMARIZE (
                'Helivalues Transaction History',
                'Helivalues Transaction History'[Sale Year]
            ),
            [Average Annualized Price change (weighted by Model Year Units)]
        )

The Matrix output looks like this (since I changed from AVERAGEX to SUMX, I would expect the Subtotals to accumulate: 

When OBSERVED_DEPRECIATION is the Matrix output, and it's programmed as follows: 

VAR OBSERVED_DEPRECIATION = 
        AVERAGEX(
            SUMMARIZE (
                'Helivalues Transaction History',
                'Helivalues Transaction History'[Sale Year]
            ),
            [Average Annualized Price change (weighted by Model Year Units)]
        )

When I change back to AVERAGEX(), the subtotals don't change. 

___________________________________________________________________

PROBLEM:  The values in the sub-total of the Matrix visual don't seem to be numerically associated with the underlying data that is supposed to be feeding them as highlighted in the red row below.

Intended output is ideally, the individual rows for the "Upside (Graph)" columns remain as shown, and the sub-total represents the average of the rows above it.  Alternatively, I could create a calculated measure where the "Upside (purposefully for graphing)" column represents the average of all rows in the column directly to its left.  

I'm trying to create a line chart from the data in the above Matrix.

The line on the line chart should be the average of the columns of "Upside (Graph) in the matrix.  So the data point for 2019 should be (roughly from eyeballing) around $15,000,000.  For reasons that are completely foreign to me, the datapoint on the line chart is pulling the $5,504,716 for the 2019 observation.

Even when I slice my PowerBI to individual rows associated with the Matrix, the row-subtotal changes slightly from what's shown above, but still isn't an average of a single datapoint = single datapoint.

The Measure to create the values for the Matrix table is relatively complex with VARs but the final calculation is:

MATRIX_OUTPUT = 

            (EARLIEST_OBSERVED_PRICE * (ABS(SIGN(SELECTED_DEPRECIATION)) * POWER((1 + SELECTED_DEPRECIATION),t))) 
            *   COEFF_UPSIDE 

EARLIEST_OBSERVED_PRICE = 

EARLIEST_OBSERVED_PRICE = 
    CALCULATE ( 
    [Average Unit Price Current Year (weighted by Model Year Units)], 
    'Helivalues Transaction History'[Sale Year] = EARLIEST_SALE_YEAR 
    ) 

EARLIEST_SALE_YEAR = 

EARLIEST_SALE_YEAR = 
        CALCULATE(
            MIN('Helivalues Transaction History'[Sale Year]),
            ALLSELECTED('Helivalues Transaction History'[Sale Year]) 
        )

SELECTED_DEPRECIATION = 

SELECTED_DEPRECIATION = 

    DIVIDE(
    CALCULATE(
        AVERAGE('Input - Depreciation'[Input - Depreciation]),
        ALLSELECTED('Input - Depreciation'[Input - Depreciation])
           
    ),
    100
    )

t = 

t = 
        CURRENT_SALE_YEAR - EARLIEST_SALE_YEAR

note, I thought/think this could be an issue, with the AVERAGE() not being the correct function to pull in what I'd want, but even if I used SELECTEDVALUE() the same output results

CURRENT_SALE_YEAR = 

CURRENT_SALE_YEAR = 
    AVERAGE('Helivalues Transaction History'[Sale Year])

COEFF_UPSIDE = 

COEFF_UPSIDE = 
    POWER(  ABS( SQRT( ABS (1 + SELECTED_DEPRECIATION + 2 * VOLATILITY) ) )   ,   t   )

VOLATILITY = 

VOLATILITY = 
    ROW_STANDARD_DEVIATION 
        * POWER(
                (1 + 
                    DIVIDE( ABS(SELECTED_DEPRECIATION - OBSERVED_DEPRECIATION) , 
                        ABS(OBSERVED_DEPRECIATION)
                    )
                ) , 
            2)

ROW_STANDARD_DEVIATION = 

ROW_STANDARD_DEVIATION = 
        STDEVX.P(
            SUMMARIZE (
                'Helivalues Transaction History',
                'Helivalues Transaction History'[Sale Year]
            ),
            [Average Annualized Price change (weighted by Model Year Units)]
        )

OBSERVED_DEPRECIATION = 

OBSERVED_DEPRECIATION = 
         AVERAGEX(
            SUMMARIZE (
                'Helivalues Transaction History',
                'Helivalues Transaction History'[Sale Year]
            ),
            [Average Annualized Price change (weighted by Model Year Units)]
        )