FORECAST ERROR: ORIGINAL FICTIONAL LEDGER
Convention: actual minus forecast. Equal row weights. Same predefined horizon.
Case | Actual | Forecast | Error | Absolute error | Squared error
A | 14 | 20 | -6 | 6 | 36
B | 18 | 20 | -2 | 2 | 4
C | 22 | 20 | 2 | 2 | 4
D | 26 | 20 | 6 | 6 | 36
Mean error: 0. MAE: 4. MSE: 20. RMSE: sqrt(20), about 4.4721.
Alternative forecast22 each: errors -8,-4,0,4; MAE4; MSE24;
RMSE sqrt(24), about4.8990. These rows do not establish future performance.

BLANK REVIEW FIELDS
Target / units / observation source:
Forecast issue date / target date / horizon:
Model version / information cutoff:
Actual / forecast / signed error / absolute error / squared error:
Training residual or held-out forecast error:
Missing values / zero denominators / exclusions / weighting:
Comparison method using same targets and cutoffs:
For binary probability forecasts use a probability score, not count-unit error.