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)
92.9%
Mean accuracy (100−APE)
88.4%
Within ±5%
37%
Within ±10%
61%
Within ±20%
83%
Bias
-5.3%
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 | 2056 | 93.3% | 64% | -5.1% |
| middle | 62 | 78.3% | 23% | +1.5% |
| grail | 76 | 80.0% | 22% | -17.7% |
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 |
|---|---|---|---|---|---|---|
| #3349 | floor · high | Sep 21, 2026 | 0.040Ξ | 0.046Ξ | +15.9% | 84.1% |
| #7109 | floor · high | Sep 21, 2026 | 0.040Ξ | 0.040Ξ | +1.2% | 98.8% |
| #6222 | floor · high | Sep 21, 2026 | 0.040Ξ | 0.042Ξ | +6.3% | 93.7% |
| #598 | grail · high | Sep 21, 2026 | 0.160Ξ | 0.334Ξ | +108.9% | 0.0% |
| #1910 | floor · high | Sep 21, 2026 | 0.042Ξ | 0.042Ξ | +1.2% | 98.8% |
| #17052 | floor · high | Sep 21, 2026 | 0.040Ξ | 0.045Ξ | +11.9% | 88.1% |
| #1531 | floor · high | Sep 21, 2026 | 0.042Ξ | 0.040Ξ | -3.7% | 96.3% |
| #9509 | floor · high | Sep 21, 2026 | 0.043Ξ | 0.040Ξ | -6.9% | 93.1% |
| #1170 | floor · high | Sep 20, 2026 | 0.044Ξ | 0.041Ξ | -6.2% | 93.8% |
| #16906 | floor · high | Sep 19, 2026 | 0.041Ξ | 0.042Ξ | +2.9% | 97.1% |
| #7872 | floor · high | Sep 19, 2026 | 0.044Ξ | 0.041Ξ | -6.2% | 93.8% |
| #7872 | floor · high | Sep 19, 2026 | 0.041Ξ | 0.040Ξ | -2.0% | 98.0% |
| #10649 | floor · high | Sep 19, 2026 | 0.050Ξ | 0.041Ξ | -17.2% | 82.8% |
| #12754 | floor · high | Sep 18, 2026 | 0.050Ξ | 0.042Ξ | -16.0% | 84.0% |
| #1170 | floor · high | Sep 18, 2026 | 0.040Ξ | 0.043Ξ | +8.0% | 92.0% |
| #13041 | floor · high | Sep 18, 2026 | 0.043Ξ | 0.043Ξ | 0.0% | 100.0% |
| #11070 | floor · high | Sep 18, 2026 | 0.042Ξ | 0.041Ξ | -2.4% | 97.6% |
| #18031 | floor · high | Sep 18, 2026 | 0.041Ξ | 0.040Ξ | -2.8% | 97.2% |
| #1171 | floor · high | Sep 18, 2026 | 0.040Ξ | 0.040Ξ | +0.8% | 99.2% |
| #5466 | floor · high | Sep 18, 2026 | 0.041Ξ | 0.040Ξ | -2.4% | 97.6% |
| #17584 | floor · high | Sep 18, 2026 | 0.051Ξ | 0.045Ξ | -12.1% | 87.9% |
| #7101 | floor · high | Sep 18, 2026 | 0.043Ξ | 0.042Ξ | -1.4% | 98.6% |
| #750 | floor · high | Sep 18, 2026 | 0.117Ξ | 0.053Ξ | -54.9% | 45.1% |
| #13564 | floor · high | Sep 18, 2026 | 0.040Ξ | 0.041Ξ | +3.5% | 96.5% |
| #515 | floor · high | Sep 18, 2026 | 0.100Ξ | 0.043Ξ | -57.0% | 43.0% |
| #10465 | floor · high | Sep 17, 2026 | 0.064Ξ | 0.045Ξ | -29.7% | 70.3% |
| #11961 | floor · high | Sep 17, 2026 | 0.039Ξ | 0.040Ξ | +2.3% | 97.7% |
| #9673 | middle · high | Sep 17, 2026 | 0.055Ξ | 0.116Ξ | +110.9% | 0.0% |
| #9682 | floor · high | Sep 17, 2026 | 0.040Ξ | 0.042Ξ | +5.0% | 95.0% |
| #3808 | floor · high | Sep 17, 2026 | 0.041Ξ | 0.040Ξ | -2.4% | 97.6% |
| #9463 | floor · high | Sep 17, 2026 | 0.039Ξ | 0.044Ξ | +13.4% | 86.6% |
| #13137 | floor · high | Sep 16, 2026 | 0.058Ξ | 0.046Ξ | -20.7% | 79.3% |
| #16411 | floor · high | Sep 16, 2026 | 0.070Ξ | 0.049Ξ | -30.0% | 70.0% |
| #8763 | floor · high | Sep 16, 2026 | 0.039Ξ | 0.049Ξ | +26.3% | 73.7% |
| #10649 | floor · high | Sep 16, 2026 | 0.040Ξ | 0.046Ξ | +15.9% | 84.1% |
| #16942 | floor · high | Sep 16, 2026 | 0.064Ξ | 0.054Ξ | -15.5% | 84.5% |
| #10743 | floor · high | Sep 16, 2026 | 0.042Ξ | 0.041Ξ | -2.4% | 97.6% |
| #18752 | floor · high | Sep 15, 2026 | 0.041Ξ | 0.044Ξ | +8.6% | 91.4% |
| #18031 | floor · high | Sep 15, 2026 | 0.041Ξ | 0.042Ξ | +3.5% | 96.6% |
| #17833 | middle · high | Sep 15, 2026 | 0.062Ξ | 0.082Ξ | +32.3% | 67.7% |
Model beanz-1789968779-2026-08-03.2 · evaluated through 9/21/2026. Method details on the methodology page.