Calculate Mean Absolute Percent Error (MAPE) when all values are in singular column?

Does anyone know how to calculate MAPE if all the data values are in the same column? 

This is the MAPE Formula for reference: (1/n) * Σ(|actual – forecast| / |actual|) * 100

Here's what my dataset looks like:

I made these measures to try and get the MAPE value: 

7. Backtesting Actuals =
CALCULATE(
    SUM('7_Temp'[Value]),
    '7_Temp'[Prediction Type] = "Actual"
)

7. Backtesting Projections =
CALCULATE(
    SUM('7_Temp'[Value]),
    '7_Temp'[Prediction Type] = "Projected"
)

7. Backtesting MAPE =
VAR MapeElement =
ABS(([7. Backtesting Actuals]-[7. Backtesting Projections])/[7. Backtesting Actuals])

VAR totalYears =
CALCULATE(
     DISTINCTCOUNT('7_Temp'[Date])
)

RETURN

IF(
    HASONEVALUE('7_Temp'[Date]),
    MapeElement,
    AVERAGEX(VALUES('7_Temp'[Date]),
    MapeElement
    )
)

However, my output doesn't quite end up being right. It looks like I'm able to get the Absolute difference between forecast and actuals for each date. However, I'm not able to do a summation and then divide by N of all those absolute differences. This is what I'm getting at the moment:

Does anyone know what I'm doing wrong here? The MAPE of this dataset should actually be 63.8%, not 2%. I'd appreciate any insight on the issue. Thanks!