COMPCOMPARE
How AI Score works
AI Score is our explainable, rules-based opportunity score. It is not a trained prediction or a probability of winning.
The score deliberately discounts distant draws: low sales today can rise considerably before next week's draw. The same Current Value Ratio can be more interesting when entries are about to close.
| Signal | Weight |
|---|---|
| Current ticket revenue versus prize value, weighted towards closing | 40% |
| Value using a time-adjusted sales scenario | 20% |
| Value at sellout | 20% |
| Ticket affordability | 10% |
| Entry-pool size in the scenario | 10% |
Free entries score 100; paid entries score up to 99 on a fixed scale. Adding free competitions does not lower paid scores. The score is saved at refresh time, alongside the valuation used to calculate it.
Exact current calculation
S = tickets sold; A = remaining; N = total; p = ticket price; V = selected prize value; h = non-negative hours until the entry deadline, or draw time if unavailable.
Q = min(N, max(1, S + 1) + A × h / (h + 24))
r = S × p / V
C = 100 − 30 × r³ if r ≤ 1; otherwise 70 / r
T = C / (1 + (h / 8)²)
B(x) = min(100, 150 × V / (V + x))
F = 100 / (1 + p / 5)
P = 100 / (1 + sqrt(Q / 1000))
Score = round(0.40T + 0.20B(Qp) + 0.20B(Np)
+ 0.10F + 0.10P), bounded to 0–99Q is a scenario, not a forecast. The calculation requires valid finite ticket counts; a missing value produces zero value points. Free entries use the separate 100-point rule.
Current and Sellout Value Ratios remain separate comparisons. We do not publish a Projected Draw Value Ratio: that requires enough historical evidence to validate a prediction.