Accuracy
How accurate is the BEANZ Official 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)
94.3%
Mean accuracy (100−APE)
89.7%
Within ±5%
42%
Within ±10%
67%
Within ±20%
85%
Bias
-3.7%
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 | 2002 | 94.5% | 69% | -3.5% |
| middle | 48 | 80.0% | 29% | +4.6% |
| grail | 59 | 79.9% | 20% | -17.8% |
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 |
|---|---|---|---|---|---|---|
| #18845 | floor · high | Aug 6, 2026 | 0.060Ξ | 0.049Ξ | -18.2% | 81.8% |
| #13457 | floor · high | Aug 5, 2026 | 0.046Ξ | 0.047Ξ | +2.0% | 98.1% |
| #1788 | floor · high | Aug 5, 2026 | 0.050Ξ | 0.051Ξ | +2.6% | 97.4% |
| #18845 | floor · high | Aug 5, 2026 | 0.046Ξ | 0.054Ξ | +17.1% | 82.9% |
| #11215 | floor · high | Aug 5, 2026 | 0.045Ξ | 0.051Ξ | +13.1% | 86.9% |
| #9487 | floor · high | Aug 5, 2026 | 0.045Ξ | 0.050Ξ | +11.1% | 88.9% |
| #5366 | floor · high | Aug 5, 2026 | 0.045Ξ | 0.051Ξ | +13.8% | 86.2% |
| #8528 | floor · high | Aug 4, 2026 | 0.050Ξ | 0.048Ξ | -4.0% | 96.0% |
| #15256 | floor · high | Aug 4, 2026 | 0.047Ξ | 0.048Ξ | +2.1% | 97.9% |
| #15256 | floor · high | Aug 4, 2026 | 0.048Ξ | 0.051Ξ | +6.7% | 93.3% |
| #5021 | floor · high | Aug 4, 2026 | 0.048Ξ | 0.051Ξ | +6.7% | 93.3% |
| #11672 | floor · high | Aug 3, 2026 | 0.100Ξ | 0.057Ξ | -43.0% | 57.0% |
| #17176 | floor · high | Aug 3, 2026 | 0.050Ξ | 0.048Ξ | -3.4% | 96.6% |
| #8091 | grail · high | Aug 3, 2026 | 0.350Ξ | 0.285Ξ | -18.6% | 81.4% |
| #19773 | floor · high | Aug 3, 2026 | 0.047Ξ | 0.050Ξ | +6.4% | 93.6% |
| #3808 | floor · high | Aug 3, 2026 | 0.050Ξ | 0.052Ξ | +4.7% | 95.3% |
| #2053 | floor · high | Aug 2, 2026 | 0.047Ξ | 0.054Ξ | +14.9% | 85.1% |
| #7147 | floor · high | Aug 2, 2026 | 0.075Ξ | 0.057Ξ | -24.0% | 76.0% |
| #3448 | floor · high | Aug 2, 2026 | 0.090Ξ | 0.062Ξ | -31.1% | 68.9% |
| #17792 | floor · high | Aug 2, 2026 | 0.050Ξ | 0.053Ξ | +6.0% | 94.0% |
| #10838 | grail · high | Aug 2, 2026 | 0.231Ξ | 0.191Ξ | -17.3% | 82.7% |
| #5551 | floor · high | Aug 1, 2026 | 0.051Ξ | 0.048Ξ | -6.0% | 94.0% |
| #17176 | floor · high | Aug 1, 2026 | 0.047Ξ | 0.055Ξ | +16.5% | 83.5% |
| #17821 | middle · high | Jul 31, 2026 | 0.150Ξ | 0.079Ξ | -47.3% | 52.7% |
| #19745 | floor · high | Jul 31, 2026 | 0.049Ξ | 0.054Ξ | +9.3% | 90.7% |
| #983 | floor · high | Jul 31, 2026 | 0.128Ξ | 0.072Ξ | -43.8% | 56.3% |
| #1109 | floor · high | Jul 31, 2026 | 0.081Ξ | 0.053Ξ | -34.6% | 65.4% |
| #15377 | floor · high | Jul 31, 2026 | 0.048Ξ | 0.052Ξ | +9.0% | 91.0% |
| #3448 | floor · high | Jul 31, 2026 | 0.065Ξ | 0.062Ξ | -4.0% | 96.0% |
| #8093 | floor · high | Jul 31, 2026 | 0.060Ξ | 0.050Ξ | -16.7% | 83.3% |
| #5008 | floor · high | Jul 31, 2026 | 0.048Ξ | 0.051Ξ | +7.4% | 92.6% |
| #487 | floor · high | Jul 31, 2026 | 0.050Ξ | 0.057Ξ | +13.8% | 86.2% |
| #19105 | floor · high | Jul 31, 2026 | 0.110Ξ | 0.058Ξ | -47.3% | 52.7% |
| #13641 | floor · high | Jul 30, 2026 | 0.110Ξ | 0.057Ξ | -48.2% | 51.8% |
| #3950 | floor · high | Jul 30, 2026 | 0.075Ξ | 0.054Ξ | -28.0% | 72.0% |
| #5551 | floor · high | Jul 30, 2026 | 0.048Ξ | 0.052Ξ | +9.0% | 91.0% |
| #5142 | middle · medium | Jul 30, 2026 | 1.390Ξ | 0.117Ξ | -91.6% | 8.4% |
| #1252 | floor · high | Jul 30, 2026 | 0.052Ξ | 0.054Ξ | +3.9% | 96.2% |
| #2670 | floor · high | Jul 30, 2026 | 0.052Ξ | 0.050Ξ | -4.4% | 95.6% |
| #13457 | floor · high | Jul 30, 2026 | 0.048Ξ | 0.050Ξ | +3.3% | 96.7% |
Model beanz-1785993095-2026-08-03.2 · evaluated through 8/6/2026. Method details on the methodology page.