Accuracy
How accurate is the Quirkies 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)
92.6%
Mean accuracy (100−APE)
90.5%
Within ±5%
33%
Within ±10%
67%
Within ±20%
92%
Bias
-1.3%
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 | 1196 | 92.7% | 68% | -1.1% |
| middle | 14 | 61.8% | 0% | -11.5% |
| grail | 1 | 31.5% | 0% | -68.5% |
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 |
|---|---|---|---|---|---|---|
| #17 | floor · high | Nov 3, 2026 | 0.910Ξ | 1.009Ξ | +10.9% | 89.1% |
| #4849 | floor · medium | Nov 2, 2026 | 1.039Ξ | 1.004Ξ | -3.4% | 96.6% |
| #1913 | floor · high | Nov 2, 2026 | 0.916Ξ | 1.089Ξ | +18.9% | 81.1% |
| #3011 | floor · medium | Nov 2, 2026 | 1.039Ξ | 0.997Ξ | -4.0% | 96.0% |
| #1029 | floor · medium | Nov 2, 2026 | 1.039Ξ | 1.046Ξ | +0.7% | 99.3% |
| #2304 | floor · high | Nov 2, 2026 | 1.068Ξ | 1.088Ξ | +1.9% | 98.1% |
| #196 | floor · medium | Nov 2, 2026 | 1.010Ξ | 0.928Ξ | -8.1% | 91.9% |
| #979 | floor · medium | Oct 31, 2026 | 1.300Ξ | 1.036Ξ | -20.3% | 79.7% |
| #867 | floor · medium | Oct 31, 2026 | 7.500Ξ | 1.191Ξ | -84.1% | 15.9% |
| #4773 | floor · high | Oct 30, 2026 | 1.027Ξ | 1.022Ξ | -0.5% | 99.6% |
| #3789 | floor · high | Oct 29, 2026 | 1.200Ξ | 1.002Ξ | -16.5% | 83.5% |
| #2005 | floor · medium | Oct 28, 2026 | 1.010Ξ | 0.947Ξ | -6.2% | 93.8% |
| #4363 | floor · medium | Oct 27, 2026 | 1.050Ξ | 0.983Ξ | -6.4% | 93.6% |
| #1257 | floor · high | Oct 26, 2026 | 1.045Ξ | 0.915Ξ | -12.4% | 87.6% |
| #1632 | floor · high | Oct 26, 2026 | 1.040Ξ | 1.052Ξ | +1.2% | 98.9% |
| #3745 | floor · medium | Oct 26, 2026 | 1.040Ξ | 0.969Ξ | -6.8% | 93.2% |
| #1428 | floor · high | Oct 26, 2026 | 1.044Ξ | 0.972Ξ | -6.9% | 93.1% |
| #2200 | floor · high | Oct 26, 2026 | 1.040Ξ | 0.961Ξ | -7.6% | 92.4% |
| #2519 | floor · medium | Oct 26, 2026 | 1.050Ξ | 0.969Ξ | -7.7% | 92.3% |
| #1053 | floor · medium | Oct 26, 2026 | 1.070Ξ | 0.990Ξ | -7.5% | 92.5% |
| #2631 | floor · medium | Oct 26, 2026 | 0.915Ξ | 0.925Ξ | +1.1% | 98.9% |
| #2834 | floor · medium | Oct 26, 2026 | 1.030Ξ | 1.013Ξ | -1.7% | 98.4% |
| #125 | floor · high | Oct 26, 2026 | 1.028Ξ | 0.919Ξ | -10.6% | 89.4% |
| #4177 | floor · high | Oct 26, 2026 | 1.030Ξ | 1.040Ξ | +1.0% | 99.0% |
| #807 | floor · high | Oct 26, 2026 | 1.028Ξ | 0.911Ξ | -11.4% | 88.6% |
| #1143 | floor · medium | Oct 26, 2026 | 1.000Ξ | 0.954Ξ | -4.6% | 95.4% |
| #1279 | floor · high | Oct 26, 2026 | 0.999Ξ | 0.900Ξ | -9.9% | 90.1% |
| #1175 | floor · high | Oct 26, 2026 | 1.000Ξ | 0.881Ξ | -11.9% | 88.1% |
| #2486 | floor · medium | Oct 25, 2026 | 0.886Ξ | 0.993Ξ | +12.1% | 87.9% |
| #4517 | floor · high | Oct 25, 2026 | 0.972Ξ | 0.916Ξ | -5.8% | 94.3% |
| #533 | floor · high | Oct 24, 2026 | 0.978Ξ | 0.922Ξ | -5.7% | 94.3% |
| #1313 | floor · high | Oct 24, 2026 | 0.981Ξ | 0.875Ξ | -10.8% | 89.2% |
| #4593 | floor · medium | Oct 24, 2026 | 0.850Ξ | 0.956Ξ | +12.5% | 87.5% |
| #1313 | floor · high | Oct 24, 2026 | 0.850Ξ | 0.953Ξ | +12.1% | 87.9% |
| #3840 | floor · high | Oct 24, 2026 | 0.983Ξ | 0.906Ξ | -7.8% | 92.2% |
| #533 | floor · medium | Oct 23, 2026 | 0.879Ξ | 1.002Ξ | +14.0% | 86.0% |
| #4517 | floor · medium | Oct 23, 2026 | 0.878Ξ | 1.000Ξ | +13.9% | 86.1% |
| #3840 | floor · medium | Oct 23, 2026 | 0.870Ξ | 0.982Ξ | +12.9% | 87.1% |
| #2625 | floor · medium | Oct 23, 2026 | 0.990Ξ | 0.945Ξ | -4.6% | 95.5% |
| #3359 | floor · high | Oct 22, 2026 | 1.010Ξ | 0.916Ξ | -9.3% | 90.7% |
Model quirkies-1792597919-2026-08-03.2 · evaluated through 10/21/2026. Method details on the methodology page.