Accuracy
How accurate is the Sappy Seals 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)
92.5%
Mean accuracy (100−APE)
89.9%
Within ±5%
37%
Within ±10%
64%
Within ±20%
88%
Bias
+0.5%
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 | 1773 | 92.7% | 65% | +0.6% |
| middle | 35 | 80.9% | 23% | -4.8% |
| grail | 7 | 67.7% | 14% | +12.5% |
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 |
|---|---|---|---|---|---|---|
| #9924 | floor · high | Aug 2, 2026 | 0.070Ξ | 0.076Ξ | +7.8% | 92.2% |
| #5332 | floor · high | Aug 2, 2026 | 0.071Ξ | 0.081Ξ | +14.1% | 85.9% |
| #2938 | floor · high | Aug 1, 2026 | 0.072Ξ | 0.079Ξ | +10.0% | 90.0% |
| #369 | floor · high | Aug 1, 2026 | 0.072Ξ | 0.080Ξ | +11.3% | 88.7% |
| #3049 | floor · high | Aug 1, 2026 | 0.077Ξ | 0.075Ξ | -2.3% | 97.7% |
| #5470 | floor · high | Aug 1, 2026 | 0.077Ξ | 0.076Ξ | -1.0% | 99.1% |
| #1273 | middle · high | Aug 1, 2026 | 0.085Ξ | 0.126Ξ | +48.6% | 51.4% |
| #5899 | floor · high | Jul 31, 2026 | 0.077Ξ | 0.082Ξ | +5.9% | 94.1% |
| #9177 | floor · high | Jul 31, 2026 | 0.072Ξ | 0.079Ξ | +10.0% | 90.0% |
| #3049 | floor · high | Jul 31, 2026 | 0.072Ξ | 0.081Ξ | +12.8% | 87.2% |
| #6153 | floor · high | Jul 30, 2026 | 0.073Ξ | 0.077Ξ | +5.6% | 94.4% |
| #4805 | floor · high | Jul 30, 2026 | 0.078Ξ | 0.099Ξ | +27.5% | 72.5% |
| #3472 | floor · high | Jul 29, 2026 | 0.091Ξ | 0.081Ξ | -11.0% | 89.0% |
| #3832 | floor · high | Jul 29, 2026 | 0.073Ξ | 0.079Ξ | +8.2% | 91.8% |
| #9315 | floor · high | Jul 29, 2026 | 0.073Ξ | 0.079Ξ | +7.5% | 92.5% |
| #9388 | floor · high | Jul 28, 2026 | 0.078Ξ | 0.075Ξ | -4.2% | 95.8% |
| #567 | floor · high | Jul 28, 2026 | 0.129Ξ | 0.106Ξ | -17.8% | 82.2% |
| #2951 | floor · high | Jul 28, 2026 | 0.079Ξ | 0.076Ξ | -3.6% | 96.4% |
| #7840 | floor · high | Jul 28, 2026 | 0.074Ξ | 0.085Ξ | +14.1% | 85.9% |
| #2449 | floor · high | Jul 27, 2026 | 0.079Ξ | 0.076Ξ | -4.4% | 95.7% |
| #2951 | floor · high | Jul 27, 2026 | 0.074Ξ | 0.079Ξ | +7.2% | 92.8% |
| #2198 | floor · high | Jul 26, 2026 | 0.080Ξ | 0.075Ξ | -5.7% | 94.3% |
| #8236 | floor · high | Jul 26, 2026 | 0.085Ξ | 0.077Ξ | -9.4% | 90.6% |
| #5101 | floor · high | Jul 26, 2026 | 0.080Ξ | 0.087Ξ | +9.4% | 90.7% |
| #2003 | floor · high | Jul 26, 2026 | 0.080Ξ | 0.074Ξ | -7.0% | 93.0% |
| #2003 | floor · high | Jul 26, 2026 | 0.074Ξ | 0.071Ξ | -3.7% | 96.3% |
| #886 | floor · high | Jul 26, 2026 | 0.080Ξ | 0.083Ξ | +3.8% | 96.3% |
| #5470 | floor · high | Jul 26, 2026 | 0.074Ξ | 0.080Ξ | +8.4% | 91.6% |
| #1865 | floor · high | Jul 26, 2026 | 0.075Ξ | 0.072Ξ | -3.6% | 96.4% |
| #2908 | floor · high | Jul 26, 2026 | 0.080Ξ | 0.071Ξ | -10.9% | 89.1% |
| #8895 | floor · high | Jul 26, 2026 | 0.135Ξ | 0.090Ξ | -33.3% | 66.7% |
| #2449 | floor · high | Jul 25, 2026 | 0.073Ξ | 0.075Ξ | +2.2% | 97.8% |
| #5448 | floor · high | Jul 25, 2026 | 0.075Ξ | 0.078Ξ | +4.0% | 96.0% |
| #7330 | floor · high | Jul 25, 2026 | 0.080Ξ | 0.076Ξ | -4.8% | 95.2% |
| #3696 | floor · high | Jul 25, 2026 | 0.080Ξ | 0.075Ξ | -6.0% | 94.1% |
| #60 | floor · high | Jul 25, 2026 | 0.080Ξ | 0.074Ξ | -7.3% | 92.7% |
| #5452 | floor · high | Jul 25, 2026 | 0.075Ξ | 0.072Ξ | -3.6% | 96.4% |
| #2198 | floor · high | Jul 24, 2026 | 0.074Ξ | 0.076Ξ | +2.3% | 97.7% |
| #8077 | floor · high | Jul 23, 2026 | 0.074Ξ | 0.065Ξ | -12.2% | 87.8% |
| #9843 | floor · high | Jul 23, 2026 | 0.074Ξ | 0.069Ξ | -6.8% | 93.2% |
Model sappy-seals-1785694103-2026-08-03.1 · evaluated through 8/2/2026. Method details on the methodology page.