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)
92.2%
Mean accuracy (100−APE)
88.8%
Within ±5%
33%
Within ±10%
60%
Within ±20%
84%
Bias
-3.9%
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 | 1917 | 92.4% | 61% | -3.8% |
| middle | 40 | 72.3% | 15% | -11.4% |
| grail | 7 | 68.1% | 29% | -1.4% |
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 |
|---|---|---|---|---|---|---|
| #26 | floor · high | Oct 30, 2026 | 0.059Ξ | 0.053Ξ | -10.0% | 90.0% |
| #153 | floor · high | Oct 30, 2026 | 0.055Ξ | 0.056Ξ | +1.8% | 98.2% |
| #8103 | floor · high | Oct 30, 2026 | 0.052Ξ | 0.055Ξ | +5.6% | 94.4% |
| #3535 | floor · high | Oct 30, 2026 | 0.053Ξ | 0.055Ξ | +4.6% | 95.4% |
| #5788 | floor · high | Oct 29, 2026 | 0.057Ξ | 0.054Ξ | -5.3% | 94.7% |
| #1400 | floor · high | Oct 29, 2026 | 0.053Ξ | 0.054Ξ | +2.5% | 97.5% |
| #6058 | floor · high | Oct 29, 2026 | 0.053Ξ | 0.055Ξ | +4.0% | 96.0% |
| #6814 | floor · high | Oct 28, 2026 | 0.053Ξ | 0.054Ξ | +1.9% | 98.1% |
| #2986 | floor · high | Oct 28, 2026 | 0.053Ξ | 0.054Ξ | +1.9% | 98.1% |
| #5788 | floor · high | Oct 28, 2026 | 0.054Ξ | 0.073Ξ | +35.9% | 64.1% |
| #5160 | floor · high | Oct 27, 2026 | 0.072Ξ | 0.054Ξ | -24.8% | 75.2% |
| #4976 | floor · high | Oct 27, 2026 | 0.057Ξ | 0.054Ξ | -6.0% | 94.0% |
| #5160 | floor · high | Oct 27, 2026 | 0.053Ξ | 0.055Ξ | +3.8% | 96.2% |
| #6244 | floor · high | Oct 27, 2026 | 0.053Ξ | 0.054Ξ | +1.1% | 98.9% |
| #732 | floor · high | Oct 26, 2026 | 0.055Ξ | 0.054Ξ | -2.7% | 97.3% |
| #1400 | floor · high | Oct 26, 2026 | 0.053Ξ | 0.059Ξ | +10.3% | 89.7% |
| #5654 | floor · high | Oct 26, 2026 | 0.054Ξ | 0.059Ξ | +10.1% | 89.9% |
| #732 | floor · high | Oct 26, 2026 | 0.054Ξ | 0.066Ξ | +22.9% | 77.1% |
| #7635 | floor · high | Oct 25, 2026 | 0.061Ξ | 0.056Ξ | -8.2% | 91.8% |
| #8905 | floor · high | Oct 25, 2026 | 0.059Ξ | 0.063Ξ | +7.2% | 92.8% |
| #6244 | floor · high | Oct 25, 2026 | 0.053Ξ | 0.070Ξ | +32.6% | 67.4% |
| #2050 | floor · high | Oct 25, 2026 | 0.058Ξ | 0.056Ξ | -2.8% | 97.2% |
| #2204 | floor · high | Oct 25, 2026 | 0.059Ξ | 0.056Ξ | -4.7% | 95.3% |
| #26 | floor · high | Oct 25, 2026 | 0.054Ξ | 0.057Ξ | +6.0% | 94.1% |
| #249 | floor · high | Oct 25, 2026 | 0.054Ξ | 0.061Ξ | +13.2% | 86.8% |
| #2050 | floor · high | Oct 25, 2026 | 0.056Ξ | 0.058Ξ | +3.8% | 96.2% |
| #9270 | floor · high | Oct 24, 2026 | 0.061Ξ | 0.067Ξ | +9.8% | 90.2% |
| #2743 | floor · high | Oct 24, 2026 | 0.056Ξ | 0.067Ξ | +19.6% | 80.4% |
| #249 | floor · high | Oct 24, 2026 | 0.061Ξ | 0.058Ξ | -4.2% | 95.8% |
| #249 | floor · high | Oct 24, 2026 | 0.056Ξ | 0.066Ξ | +17.7% | 82.4% |
| #7635 | floor · high | Oct 24, 2026 | 0.057Ξ | 0.066Ξ | +16.0% | 84.0% |
| #9448 | floor · high | Oct 23, 2026 | 0.066Ξ | 0.067Ξ | +2.1% | 97.9% |
| #9750 | floor · high | Oct 23, 2026 | 0.057Ξ | 0.059Ξ | +3.0% | 97.0% |
| #9388 | floor · high | Oct 23, 2026 | 0.059Ξ | 0.079Ξ | +33.0% | 67.0% |
| #4377 | floor · high | Oct 23, 2026 | 0.059Ξ | 0.062Ξ | +4.4% | 95.6% |
| #6169 | floor · high | Oct 22, 2026 | 0.060Ξ | 0.071Ξ | +18.7% | 81.3% |
| #2022 | floor · high | Oct 22, 2026 | 0.062Ξ | 0.075Ξ | +21.8% | 78.3% |
| #5698 | floor · high | Oct 22, 2026 | 0.060Ξ | 0.071Ξ | +18.5% | 81.5% |
| #2488 | floor · high | Oct 22, 2026 | 0.061Ξ | 0.071Ξ | +15.8% | 84.2% |
| #2204 | floor · high | Oct 22, 2026 | 0.060Ξ | 0.067Ξ | +12.0% | 88.0% |
Model sappy-seals-1793387219-2026-08-03.2 · evaluated through 10/30/2026. Method details on the methodology page.