Accuracy
How accurate is the DeGods model?
Every score comes from the newest chronological hold-out; the target sale is excluded from training, comparables, and its causal floor proxy.
Median accuracy (100−APE)
93.6%
Mean accuracy (100−APE)
92.0%
Within ±5%
41%
Within ±10%
72%
Within ±20%
94%
Bias
-2.8%
Accuracy by market tier
Floor, middle-market, and grail sales are scored separately so aggregate accuracy cannot hide a weak segment
| Tier | Sales | Median accuracy | Within ±10% | Bias |
|---|---|---|---|---|
| floor | 525 | 93.6% | 72% | -2.9% |
| middle | 1 | 51.4% | 0% | +48.6% |
Weekly median accuracy
Chronological hold-out accuracy by sale week
Error distribution
Signed hold-out error; positive means over-predicted
Scored hold-out sales
60 out-of-sample predictions frozen at evaluation time
| Item | Tier / evidence | Date | Sale | Model | Error | Accuracy |
|---|---|---|---|---|---|---|
| #1505 | floor · high | Sep 18, 2026 | 0.212Ξ | 0.160Ξ | -24.5% | 75.5% |
| #6434 | floor · high | Sep 18, 2026 | 0.287Ξ | 0.179Ξ | -37.6% | 62.4% |
| #7039 | floor · high | Sep 16, 2026 | 0.165Ξ | 0.152Ξ | -7.7% | 92.3% |
| #8770 | floor · high | Sep 16, 2026 | 0.125Ξ | 0.164Ξ | +31.2% | 68.8% |
| #1625 | floor · high | Sep 15, 2026 | 0.150Ξ | 0.148Ξ | -1.3% | 98.7% |
| #3517 | floor · high | Sep 14, 2026 | 0.148Ξ | 0.168Ξ | +13.5% | 86.5% |
| #1625 | floor · high | Sep 14, 2026 | 0.139Ξ | 0.166Ξ | +19.4% | 80.6% |
| #1251 | floor · high | Sep 14, 2026 | 0.148Ξ | 0.168Ξ | +13.5% | 86.5% |
| #4549 | floor · high | Sep 12, 2026 | 0.167Ξ | 0.170Ξ | +1.8% | 98.2% |
| #7494 | floor · high | Sep 12, 2026 | 0.168Ξ | 0.172Ξ | +2.4% | 97.6% |
| #7039 | floor · high | Sep 10, 2026 | 0.164Ξ | 0.195Ξ | +18.9% | 81.1% |
| #7494 | floor · high | Sep 10, 2026 | 0.170Ξ | 0.195Ξ | +14.7% | 85.3% |
| #277 | floor · high | Sep 10, 2026 | 0.163Ξ | 0.176Ξ | +8.0% | 92.0% |
| #4549 | floor · high | Sep 10, 2026 | 0.166Ξ | 0.200Ξ | +20.5% | 79.5% |
| #2540 | floor · high | Sep 10, 2026 | 0.161Ξ | 0.176Ξ | +9.3% | 90.7% |
| #4549 | floor · high | Sep 10, 2026 | 0.200Ξ | 0.199Ξ | -0.5% | 99.5% |
| #7039 | floor · high | Sep 10, 2026 | 0.195Ξ | 0.171Ξ | -12.3% | 87.7% |
| #7494 | floor · high | Sep 10, 2026 | 0.195Ξ | 0.161Ξ | -17.4% | 82.6% |
| #277 | floor · high | Sep 10, 2026 | 0.176Ξ | 0.161Ξ | -8.6% | 91.4% |
| #2540 | floor · high | Sep 10, 2026 | 0.176Ξ | 0.162Ξ | -8.0% | 92.0% |
| #5887 | floor · high | Sep 7, 2026 | 0.158Ξ | 0.204Ξ | +29.1% | 70.9% |
| #7039 | floor · high | Sep 5, 2026 | 0.171Ξ | 0.164Ξ | -4.1% | 95.9% |
| #5887 | floor · high | Sep 5, 2026 | 0.200Ξ | 0.174Ξ | -13.0% | 87.0% |
| #2540 | floor · high | Sep 5, 2026 | 0.160Ξ | 0.193Ξ | +20.6% | 79.4% |
| #7494 | floor · high | Sep 5, 2026 | 0.160Ξ | 0.234Ξ | +46.3% | 53.8% |
| #277 | floor · high | Sep 5, 2026 | 0.170Ξ | 0.192Ξ | +12.9% | 87.1% |
| #7124 | floor · high | Sep 4, 2026 | 0.273Ξ | 0.195Ξ | -28.6% | 71.4% |
| #9482 | floor · high | Sep 2, 2026 | 0.190Ξ | 0.196Ξ | +3.2% | 96.8% |
| #7204 | floor · high | Sep 1, 2026 | 0.190Ξ | 0.192Ξ | +1.1% | 99.0% |
| #5887 | middle · high | Sep 1, 2026 | 0.181Ξ | 0.269Ξ | +48.6% | 51.4% |
| #5887 | floor · high | Aug 31, 2026 | 0.269Ξ | 0.190Ξ | -29.4% | 70.6% |
| #277 | floor · high | Aug 30, 2026 | 0.210Ξ | 0.187Ξ | -11.0% | 89.1% |
| #7494 | floor · high | Aug 30, 2026 | 0.236Ξ | 0.184Ξ | -22.2% | 77.8% |
| #2540 | floor · high | Aug 30, 2026 | 0.196Ξ | 0.181Ξ | -7.7% | 92.4% |
| #1872 | floor · high | Aug 29, 2026 | 0.202Ξ | 0.181Ξ | -10.4% | 89.6% |
| #7111 | floor · high | Aug 29, 2026 | 0.191Ξ | 0.169Ξ | -11.5% | 88.5% |
| #190 | floor · high | Aug 24, 2026 | 0.200Ξ | 0.193Ξ | -3.5% | 96.5% |
| #4058 | floor · high | Aug 24, 2026 | 0.167Ξ | 0.202Ξ | +21.0% | 79.0% |
| #7111 | floor · high | Aug 22, 2026 | 0.170Ξ | 0.206Ξ | +21.2% | 78.8% |
| #7113 | floor · high | Aug 19, 2026 | 0.191Ξ | 0.206Ξ | +7.9% | 92.2% |
Model degods-1789712387-2026-08-03.2 · evaluated through 9/18/2026. Method details on the methodology page.