Accuracy
How accurate is the Moonbirds 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)
94.8%
Mean accuracy (100−APE)
92.4%
Within ±5%
49%
Within ±10%
75%
Within ±20%
93%
Bias
-0.4%
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 | 1547 | 94.9% | 75% | -0.4% |
| middle | 6 | 84.0% | 0% | +21.5% |
| grail | 13 | 85.4% | 15% | -13.2% |
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 |
|---|---|---|---|---|---|---|
| #6026 | floor · high | Sep 14, 2026 | 0.461Ξ | 0.516Ξ | +11.9% | 88.1% |
| #9370 | floor · high | Sep 14, 2026 | 0.468Ξ | 0.512Ξ | +9.4% | 90.6% |
| #4629 | floor · high | Sep 13, 2026 | 0.474Ξ | 0.513Ξ | +8.2% | 91.8% |
| #6258 | floor · high | Sep 13, 2026 | 0.482Ξ | 0.531Ξ | +10.2% | 89.8% |
| #6760 | floor · high | Sep 12, 2026 | 0.506Ξ | 0.487Ξ | -3.8% | 96.3% |
| #9471 | floor · high | Sep 12, 2026 | 0.505Ξ | 0.501Ξ | -0.8% | 99.2% |
| #4129 | floor · high | Sep 12, 2026 | 0.494Ξ | 0.516Ξ | +4.5% | 95.6% |
| #6765 | floor · high | Sep 12, 2026 | 0.510Ξ | 0.508Ξ | -0.4% | 99.6% |
| #4775 | floor · high | Sep 12, 2026 | 0.505Ξ | 0.508Ξ | +0.6% | 99.4% |
| #5369 | floor · high | Sep 12, 2026 | 0.505Ξ | 0.509Ξ | +0.8% | 99.2% |
| #2662 | floor · high | Sep 11, 2026 | 0.473Ξ | 0.526Ξ | +11.2% | 88.8% |
| #2842 | floor · high | Sep 11, 2026 | 0.471Ξ | 0.512Ξ | +8.7% | 91.3% |
| #8793 | floor · high | Sep 11, 2026 | 0.505Ξ | 0.501Ξ | -0.8% | 99.2% |
| #9952 | floor · high | Sep 11, 2026 | 0.550Ξ | 0.515Ξ | -6.4% | 93.6% |
| #3669 | floor · high | Sep 10, 2026 | 0.490Ξ | 0.503Ξ | +2.7% | 97.4% |
| #6444 | floor · high | Sep 10, 2026 | 0.508Ξ | 0.507Ξ | -0.2% | 99.8% |
| #332 | floor · high | Sep 10, 2026 | 0.540Ξ | 0.509Ξ | -5.7% | 94.3% |
| #8793 | floor · high | Sep 10, 2026 | 0.490Ξ | 0.517Ξ | +5.5% | 94.5% |
| #5954 | floor · high | Sep 10, 2026 | 0.490Ξ | 0.508Ξ | +3.7% | 96.3% |
| #3803 | floor · high | Sep 10, 2026 | 0.490Ξ | 0.504Ξ | +2.9% | 97.1% |
| #4389 | floor · high | Sep 10, 2026 | 0.492Ξ | 0.543Ξ | +10.4% | 89.6% |
| #4708 | floor · high | Sep 10, 2026 | 0.490Ξ | 0.566Ξ | +15.5% | 84.5% |
| #7291 | floor · high | Sep 10, 2026 | 0.490Ξ | 0.505Ξ | +3.1% | 96.9% |
| #1537 | floor · high | Sep 10, 2026 | 0.491Ξ | 0.538Ξ | +9.6% | 90.4% |
| #7453 | floor · high | Sep 10, 2026 | 0.490Ξ | 0.561Ξ | +14.5% | 85.5% |
| #9832 | floor · high | Sep 10, 2026 | 0.492Ξ | 0.514Ξ | +4.5% | 95.5% |
| #3184 | floor · high | Sep 10, 2026 | 0.490Ξ | 0.561Ξ | +14.5% | 85.5% |
| #3669 | floor · high | Sep 10, 2026 | 0.500Ξ | 0.494Ξ | -1.1% | 98.9% |
| #7795 | floor · high | Sep 10, 2026 | 0.500Ξ | 0.504Ξ | +0.8% | 99.2% |
| #9952 | floor · high | Sep 10, 2026 | 0.515Ξ | 0.599Ξ | +16.3% | 83.7% |
| #5369 | floor · high | Sep 10, 2026 | 0.509Ξ | 0.618Ξ | +21.4% | 78.6% |
| #2983 | floor · high | Sep 10, 2026 | 0.490Ξ | 0.639Ξ | +30.4% | 69.6% |
| #1673 | floor · high | Sep 10, 2026 | 0.490Ξ | 0.604Ξ | +23.3% | 76.7% |
| #6760 | floor · high | Sep 10, 2026 | 0.493Ξ | 0.589Ξ | +19.5% | 80.5% |
| #3164 | floor · high | Sep 9, 2026 | 0.523Ξ | 0.527Ξ | +0.8% | 99.2% |
| #3362 | floor · high | Sep 9, 2026 | 0.515Ξ | 0.521Ξ | +1.2% | 98.8% |
| #1306 | floor · high | Sep 9, 2026 | 0.517Ξ | 0.553Ξ | +7.0% | 93.0% |
| #3656 | floor · high | Sep 9, 2026 | 0.508Ξ | 0.528Ξ | +3.9% | 96.1% |
| #3300 | floor · high | Sep 9, 2026 | 0.509Ξ | 0.529Ξ | +3.9% | 96.1% |
| #2964 | floor · high | Sep 9, 2026 | 0.540Ξ | 0.519Ξ | -3.9% | 96.1% |
Model moonbirds-1789414127-2026-08-03.2 · evaluated through 9/14/2026. Method details on the methodology page.