Enter Forecast Data
Use period cards below. Large screens show three cards per row.
Example Data Table
| Period | Actual | Forecast | Weight |
|---|---|---|---|
| Jan | 120 | 118 | 1 |
| Feb | 135 | 140 | 1 |
| Mar | 128 | 130 | 1 |
| Apr | 142 | 145 | 1 |
| May | 155 | 150 | 1 |
| Jun | 149 | 152 | 1 |
Formula Used
Error: Error = Forecast − Actual
Mean Error: ME = Σ(Error) ÷ n
Mean Absolute Error: MAE = Σ|Error| ÷ n
Mean Squared Error: MSE = Σ(Error²) ÷ n
Root Mean Squared Error: RMSE = √MSE
Mean Percentage Error: MPE = Σ[(Error ÷ Actual) × 100] ÷ k
Mean Absolute Percentage Error: MAPE = Σ[(|Error| ÷ |Actual|) × 100] ÷ k
Symmetric MAPE: sMAPE = Σ[200 × |Forecast − Actual| ÷ (|Actual| + |Forecast|)] ÷ s
Weighted Absolute Percentage Error: WAPE = [Σ|Error| ÷ Σ|Actual|] × 100
Bias Percentage: Bias % = [Σ(Error) ÷ Σ|Actual|] × 100
Weighted MAE: Weighted MAE = Σ(Weight × |Error|) ÷ ΣWeight
Weighted RMSE: Weighted RMSE = √[Σ(Weight × Error²) ÷ ΣWeight]
Tracking Signal: Tracking Signal = Cumulative Error ÷ MAD
MASE: MASE = MAE ÷ seasonal naive MAE
Note: k counts rows with nonzero actual values. s counts rows where |Actual| + |Forecast| is positive.
How to Use This Calculator
- Enter a period label, actual value, forecast value, and optional weight.
- Set the seasonality period if you want a meaningful MASE result.
- Choose decimal places and decide how zero actual rows should behave.
- Click Calculate Metrics to show the summary above the form.
- Review the detailed row metrics table for period-level error patterns.
- Download the summary and detailed rows as CSV or PDF.
FAQs
1. Which metric is best for everyday forecasting?
MAE is often easiest to explain because it stays in original units. RMSE is useful when larger misses deserve heavier penalties. Use both together for balanced judgment.
2. Why can MAPE show N/A?
MAPE divides by actual values. When any actual value is zero, that row cannot produce a valid percentage error. The calculator skips such rows for percentage metrics.
3. What does a positive bias percentage mean?
A positive bias percentage means forecasts are generally higher than actual results. That pattern suggests over-forecasting, which may inflate inventory, staffing, or revenue expectations.
4. When should I focus on RMSE instead of MAE?
Use RMSE when large forecast misses are especially costly. Because errors are squared first, RMSE grows faster than MAE and highlights extreme deviations more clearly.
5. What is WAPE good for?
WAPE expresses total absolute error relative to total actual volume. It is practical for comparing performance across datasets with different scales or for summarizing business-level accuracy.
6. Why include weights in the rows?
Weights let important periods influence weighted MAE and weighted RMSE more strongly. This helps when some periods carry more revenue, volume, cost, or planning impact.
7. How should I choose the seasonality period?
Pick the repeating cycle length in your data. For monthly data with yearly seasonality, use 12. For weekly patterns inside daily data, use 7.
8. What tracking signal range is usually acceptable?
Many teams monitor whether tracking signal stays between −4 and +4. Values outside that band can indicate persistent bias and may justify model review.