metrics

# Forecasting Metrics & Formulas

## Core Prediction Metrics

### Prediction Error

```

Error = |Predicted Demand - Actual Demand|

Error % = (Error / Actual Demand) × 100

```

### Overall Accuracy

```

Overall Accuracy = 100 - Average Error (MAPE)

```

- Calculated across all watches and all 12 months

- Higher is better (100% = perfect predictions)

---

## Inventory & Sales Flow

### Available Units

```

Available = Inventory Start + Production

Production = Predicted Demand (student's forecast)

```

### Units Sold

```

Units Sold = min(Actual Demand, Available)

```

- Can't sell more than you have in stock

### Ending Inventory

```

Inventory End = Available - Units Sold

```

### Stockouts (Lost Sales)

```

Stockout = max(0, Actual Demand - Units Sold)

```

- Unmet demand due to insufficient inventory

---

## Financial Metrics

### Revenue

```

Revenue = Units Sold × Sell Price

```

### Production Costs

```

Production Cost = Production × Base Cost

Labor Cost = Production × 20 CHF

Total Production Cost = Production Cost + Labor Cost

```

### Holding Costs

```

Holding Cost = Inventory End × Base Cost × 0.02

```

- 2% of product value per month for storage

### Lost Revenue (Stockout Penalty)

```

Lost Revenue = Stockout × Sell Price

```

- Revenue you missed from under-forecasting

### Excess Costs (Overstock Penalty)

```

Excess Inventory = max(0, Inventory End - 50)

Excess Cost = Excess Inventory × Base Cost × 0.05

```

- 50 units = healthy safety stock threshold

- 5% waste/obsolescence cost on excess inventory

### Total Costs & Profit

```

Total Costs = Production Cost + Labor Cost + Holding Cost + Excess Cost

Profit = Revenue - Total Costs

```

---

## Key Insights

- **Under-forecasting** → Stockouts → Lost Revenue

- **Over-forecasting** → Excess Inventory → Excess Costs

- **Goal**: Minimize both errors to maximize profit

 No newline at end of file