Accuracy
How accurate is the Opepen Edition model?
Every estimate is benchmarked against real sales the model never trained on (a seeded 80/20 hold-out).
Median accuracy
87.4%
Mean accuracy
81.2%
Within ±5%
20%
Within ±10%
41%
Within ±20%
68%
Bias
+3.8%
Weekly median accuracy
Hold-out accuracy by week
Error distribution
Signed % error on recent sales (0 = perfect)
Recent sales vs model
The 60 most recent sales, each valued as of its sale date
| Item | Edition Size | Date | Sale | Model | Error | Accuracy |
|---|---|---|---|---|---|---|
| #14968 | Forty | Jun 16, 2026 | 0.082Ξ | 0.087Ξ | -5.7% | 94.4% |
| #3822 | Twenty | Jun 16, 2026 | 0.072Ξ | 0.077Ξ | -7.2% | 92.9% |
| #9525 | Ten | Jun 16, 2026 | 0.072Ξ | 0.084Ξ | -14.2% | 85.8% |
| #9525 | Ten | Jun 16, 2026 | 0.066Ξ | 0.084Ξ | -21.0% | 79.0% |
| #3822 | Twenty | Jun 16, 2026 | 0.067Ξ | 0.077Ξ | -13.9% | 86.1% |
| #14058 | Forty | Jun 15, 2026 | 0.086Ξ | 0.087Ξ | -1.6% | 98.4% |
| #951 | Twenty | Jun 15, 2026 | 0.200Ξ | 0.194Ξ | +2.8% | 97.2% |
| #9655 | Five | Jun 15, 2026 | 0.169Ξ | 0.123Ξ | +37.1% | 62.9% |
| #9655 | Five | Jun 15, 2026 | 0.170Ξ | 0.123Ξ | +37.9% | 62.1% |
| #8268 | Five | Jun 15, 2026 | 0.170Ξ | 0.123Ξ | +37.9% | 62.1% |
| #15096 | Five | Jun 15, 2026 | 0.170Ξ | 0.123Ξ | +37.9% | 62.1% |
| #658 | Five | Jun 15, 2026 | 0.170Ξ | 0.123Ξ | +37.9% | 62.1% |
| #1625 | Five | Jun 15, 2026 | 0.170Ξ | 0.123Ξ | +37.9% | 62.1% |
| #12852 | Ten | Jun 15, 2026 | 0.118Ξ | 0.099Ξ | +19.5% | 80.5% |
| #14574 | Twenty | Jun 15, 2026 | 0.099Ξ | 0.096Ξ | +3.2% | 96.8% |
| #4281 | Forty | Jun 14, 2026 | 0.086Ξ | 0.087Ξ | -1.6% | 98.4% |
| #12655 | Forty | Jun 14, 2026 | 0.087Ξ | 0.087Ξ | -0.6% | 99.4% |
| #11322 | Ten | Jun 14, 2026 | 0.120Ξ | 0.099Ξ | +20.7% | 79.3% |
| #9594 | Five | Jun 14, 2026 | 0.178Ξ | 0.124Ξ | +44.0% | 56.0% |
| #8099 | Twenty | Jun 13, 2026 | 0.097Ξ | 0.097Ξ | +0.4% | 99.6% |
| #210 | Ten | Jun 13, 2026 | 0.138Ξ | 0.100Ξ | +37.7% | 62.3% |
| #9963 | Forty | Jun 12, 2026 | 0.085Ξ | 0.080Ξ | +6.2% | 93.8% |
| #2167 | Twenty | Jun 12, 2026 | 0.110Ξ | 0.169Ξ | -34.9% | 65.1% |
| #986 | Forty | Jun 12, 2026 | 0.074Ξ | 0.080Ξ | -7.4% | 92.6% |
| #8182 | Twenty | Jun 11, 2026 | 0.118Ξ | 0.092Ξ | +28.8% | 71.2% |
| #5695 | Forty | Jun 11, 2026 | 0.100Ξ | 0.090Ξ | +11.5% | 88.5% |
| #786 | Five | Jun 11, 2026 | 0.150Ξ | 0.112Ξ | +34.5% | 65.5% |
| #11177 | Forty | Jun 11, 2026 | 0.200Ξ | 0.228Ξ | -12.3% | 87.7% |
| #2872 | Ten | Jun 11, 2026 | 0.149Ξ | 0.096Ξ | +55.1% | 44.9% |
| #11590 | Forty | Jun 11, 2026 | 0.411Ξ | 0.271Ξ | +51.6% | 48.4% |
| #4694 | Twenty | Jun 11, 2026 | 0.360Ξ | 0.403Ξ | -10.8% | 89.2% |
| #951 | Twenty | Jun 11, 2026 | 0.122Ξ | 0.201Ξ | -39.2% | 60.8% |
| #4040 | Forty | Jun 10, 2026 | 0.100Ξ | 0.088Ξ | +13.0% | 87.0% |
| #2194 | Five | Jun 10, 2026 | 0.200Ξ | 0.127Ξ | +57.9% | 42.1% |
| #14366 | Forty | Jun 10, 2026 | 0.140Ξ | 0.090Ξ | +55.6% | 44.4% |
| #7811 | Forty | Jun 10, 2026 | 0.068Ξ | 0.084Ξ | -20.0% | 80.0% |
| #13093 | Forty | Jun 10, 2026 | 0.068Ξ | 0.084Ξ | -20.0% | 80.0% |
| #15347 | Forty | Jun 10, 2026 | 0.120Ξ | 0.118Ξ | +1.5% | 98.5% |
| #10988 | Ten | Jun 9, 2026 | 0.950Ξ | 0.435Ξ | +118.2% | -18.2% |
| #222 | Twenty | Jun 8, 2026 | 0.097Ξ | 0.099Ξ | -2.1% | 97.9% |
Method details on the methodology page.