This records changes based on gameplay observations from 2026-09-27. Reconstructing existing records checks arithmetic; it is not independent prediction validation.
What was checked
The old one-item-per-customer and simple average-price assumptions missed purchases across multiple products. The model now preserves per-product purchase rates for each observed scenario.
Recorded observations
| Scenario | Customers in 7 days | Recorded total quantity | Recorded revenue | Quantity reconstructed from rates |
|---|---|---|---|---|
| Coffee only · 800 customers | 800 | 1,812 | $9,060.00 | 1,812 |
| Coffee + tea · 915 customers | 915 | 3,343 | $16,565.00 | 3,339.75 |
| Coffee + tea + cupcake + ice cream · 900 customers | 900 | 4,131 | $23,457.00 | 4,122 |
Units per customer
Coffee only · 800 customers
Game day: ~249 · Observed promotion: 87%
- Cup Of Coffee: 2.265
Observed demand: Cup Of Coffee: 47→43%
Coffee + tea · 915 customers
Game day: 249–256 · Observed promotion: 87%
- Cup Of Coffee: 1.89
- Cup of Tea: 1.76
Observed demand: Cup Of Coffee: 43→43%, Cup of Tea: 58→43%
Coffee + tea + cupcake + ice cream · 900 customers
Game day: ~240 · Observed promotion: 87%
- Cup Of Coffee: 1.79
- Cup of Tea: 1.72
- Cupcake: 0.91
- Ice Cream: 0.16
How it calculates
Product quantity = customers × its purchase rate in the selected scenario. The calculator sums quantities times current data prices and costs, then adds paper-bag costs separately. Customer counts can be predicted or taken from the record.
Reconstructing the coffee-only record
The price is derived from recorded revenue divided by quantity. It can differ from the current calculator default. Reusing this record does not establish prediction accuracy.
800 × (1,812 ÷ 800) × $5.00 = $9,060.00
Limits of the evidence
Two- and four-product rates were recorded rounded, so reconstructed quantities differ. Two-product revenue is inconsistent: $16,565 recorded versus $16,715 from price × quantity. Four-product demand, exact hours, some prices and the game version were not recorded; gaps were not filled with estimates.
Next validation
Observe coffee + tea + cupcake and coffee + ice cream independently. Record starting/ending demand, prices, sales, paper bags, stockouts, staff, opening hours and promotion. Compare customer, product-quantity and revenue errors separately.