Accuracy
How accurate is the Goblintown model?
Every estimate is benchmarked against real sales the model never trained on (a seeded 80/20 hold-out).
Median accuracy
79.6%
Mean accuracy
70.6%
Within ±5%
13%
Within ±10%
24%
Within ±20%
49%
Bias
+8.1%
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 | Date | Sale | Model | Error | Accuracy |
|---|---|---|---|---|---|
| #6536 | Jun 13, 2026 | 0.076Ξ | 0.094Ξ | -19.5% | 80.6% |
| #6855 | Jun 11, 2026 | 0.079Ξ | 0.086Ξ | -8.1% | 91.9% |
| #9515 | Jun 11, 2026 | 0.080Ξ | 0.087Ξ | -7.9% | 92.1% |
| #4807 | Jun 10, 2026 | 0.080Ξ | 0.086Ξ | -7.1% | 92.9% |
| #1987 | Jun 10, 2026 | 0.081Ξ | 0.084Ξ | -3.3% | 96.7% |
| #2462 | Jun 10, 2026 | 0.082Ξ | 0.083Ξ | -1.5% | 98.5% |
| #4070 | Jun 10, 2026 | 0.082Ξ | 0.086Ξ | -5.3% | 94.7% |
| #7816 | Jun 10, 2026 | 0.082Ξ | 0.085Ξ | -4.2% | 95.8% |
| #4569 | Jun 9, 2026 | 0.102Ξ | 0.091Ξ | +12.4% | 87.6% |
| #2974 | Jun 9, 2026 | 0.075Ξ | 0.083Ξ | -10.0% | 90.0% |
| #8936 | Jun 9, 2026 | 0.086Ξ | 0.085Ξ | +0.4% | 99.6% |
| #7692 | Jun 9, 2026 | 0.081Ξ | 0.088Ξ | -8.2% | 91.8% |
| #5880 | Jun 9, 2026 | 0.086Ξ | 0.087Ξ | -1.7% | 98.3% |
| #5441 | Jun 9, 2026 | 0.084Ξ | 0.086Ξ | -2.4% | 97.6% |
| #1630 | Jun 9, 2026 | 0.081Ξ | 0.087Ξ | -6.5% | 93.5% |
| #1794 | Jun 9, 2026 | 0.085Ξ | 0.091Ξ | -6.9% | 93.1% |
| #4914 | Jun 9, 2026 | 0.083Ξ | 0.083Ξ | +0.2% | 99.8% |
| #4219 | Jun 9, 2026 | 0.084Ξ | 0.087Ξ | -3.9% | 96.2% |
| #1306 | Jun 9, 2026 | 0.085Ξ | 0.090Ξ | -5.7% | 94.3% |
| #7470 | Jun 9, 2026 | 0.083Ξ | 0.090Ξ | -8.0% | 92.0% |
| #2799 | Jun 9, 2026 | 0.087Ξ | 0.085Ξ | +3.3% | 96.7% |
| #6050 | Jun 8, 2026 | 0.088Ξ | 0.091Ξ | -2.9% | 97.1% |
| #8616 | Jun 8, 2026 | 0.130Ξ | 0.100Ξ | +30.7% | 69.4% |
| #6582 | Jun 7, 2026 | 0.103Ξ | 0.091Ξ | +12.3% | 87.7% |
| #3937 | Jun 6, 2026 | 0.248Ξ | 0.147Ξ | +69.0% | 31.0% |
| #8616 | Jun 6, 2026 | 0.102Ξ | 0.100Ξ | +2.4% | 97.6% |
| #162 | Jun 6, 2026 | 0.087Ξ | 0.088Ξ | -1.4% | 98.6% |
| #4114 | Jun 6, 2026 | 0.087Ξ | 0.088Ξ | -0.6% | 99.4% |
| #1809 | Jun 6, 2026 | 0.087Ξ | 0.086Ξ | +1.5% | 98.5% |
| #9109 | Jun 5, 2026 | 0.103Ξ | 0.094Ξ | +9.0% | 91.0% |
| #3782 | Jun 5, 2026 | 0.087Ξ | 0.081Ξ | +6.8% | 93.2% |
| #9491 | Jun 5, 2026 | 0.087Ξ | 0.084Ξ | +3.3% | 96.7% |
| #5975 | Jun 5, 2026 | 0.087Ξ | 0.089Ξ | -2.8% | 97.2% |
| #3935 | Jun 5, 2026 | 0.087Ξ | 0.087Ξ | +0.6% | 99.4% |
| #6288 | Jun 5, 2026 | 0.087Ξ | 0.084Ξ | +3.5% | 96.5% |
| #8964 | Jun 5, 2026 | 0.082Ξ | 0.091Ξ | -10.2% | 89.8% |
| #8845 | Jun 5, 2026 | 0.082Ξ | 0.081Ξ | +0.7% | 99.3% |
| #1171 | Jun 5, 2026 | 0.200Ξ | 0.123Ξ | +62.6% | 37.4% |
| #850 | Jun 4, 2026 | 0.081Ξ | 0.089Ξ | -9.3% | 90.7% |
| #4982 | Jun 4, 2026 | 0.090Ξ | 0.094Ξ | -4.4% | 95.6% |
Method details on the methodology page.