Accuracy
How accurate is the mfers 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)
90.1%
Within ±5%
36%
Within ±10%
63%
Within ±20%
89%
Bias
+2.6%
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 | 1498 | 92.6% | 63% | +2.8% |
| middle | 24 | 72.8% | 25% | +0.6% |
| grail | 17 | 80.4% | 29% | -17.9% |
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 |
|---|---|---|---|---|---|---|
| #6643 | floor · high | Sep 19, 2026 | 0.141Ξ | 0.157Ξ | +11.4% | 88.7% |
| #4311 | floor · high | Sep 19, 2026 | 0.144Ξ | 0.154Ξ | +6.9% | 93.1% |
| #5211 | floor · high | Sep 19, 2026 | 0.144Ξ | 0.167Ξ | +16.0% | 84.0% |
| #5870 | floor · high | Sep 19, 2026 | 0.151Ξ | 0.162Ξ | +7.3% | 92.7% |
| #1727 | floor · high | Sep 19, 2026 | 0.165Ξ | 0.155Ξ | -5.9% | 94.1% |
| #5756 | floor · high | Sep 19, 2026 | 0.150Ξ | 0.160Ξ | +6.7% | 93.3% |
| #883 | floor · high | Sep 19, 2026 | 0.165Ξ | 0.152Ξ | -7.9% | 92.1% |
| #9254 | floor · high | Sep 19, 2026 | 0.149Ξ | 0.146Ξ | -2.0% | 98.0% |
| #883 | floor · high | Sep 18, 2026 | 0.152Ξ | 0.158Ξ | +4.0% | 96.1% |
| #1583 | floor · high | Sep 18, 2026 | 0.150Ξ | 0.157Ξ | +4.7% | 95.3% |
| #1056 | floor · high | Sep 18, 2026 | 0.150Ξ | 0.158Ξ | +5.3% | 94.7% |
| #8109 | floor · high | Sep 18, 2026 | 0.151Ξ | 0.159Ξ | +5.3% | 94.7% |
| #1727 | floor · high | Sep 18, 2026 | 0.155Ξ | 0.151Ξ | -2.6% | 97.4% |
| #9597 | floor · high | Sep 18, 2026 | 0.155Ξ | 0.164Ξ | +5.8% | 94.2% |
| #2294 | floor · high | Sep 18, 2026 | 0.157Ξ | 0.148Ξ | -5.7% | 94.3% |
| #8126 | floor · high | Sep 18, 2026 | 0.158Ξ | 0.154Ξ | -2.5% | 97.5% |
| #883 | floor · high | Sep 18, 2026 | 0.158Ξ | 0.150Ξ | -5.0% | 95.0% |
| #435 | floor · high | Sep 18, 2026 | 0.165Ξ | 0.153Ξ | -7.5% | 92.5% |
| #8889 | floor · high | Sep 18, 2026 | 0.164Ξ | 0.149Ξ | -9.2% | 90.8% |
| #2959 | floor · high | Sep 18, 2026 | 0.158Ξ | 0.149Ξ | -5.7% | 94.4% |
| #8652 | floor · high | Sep 18, 2026 | 0.165Ξ | 0.151Ξ | -8.7% | 91.3% |
| #9597 | floor · high | Sep 18, 2026 | 0.164Ξ | 0.159Ξ | -3.1% | 96.9% |
| #6403 | floor · high | Sep 18, 2026 | 0.158Ξ | 0.155Ξ | -1.9% | 98.1% |
| #8049 | floor · high | Sep 18, 2026 | 0.159Ξ | 0.154Ξ | -3.1% | 96.9% |
| #2959 | floor · high | Sep 18, 2026 | 0.148Ξ | 0.151Ξ | +2.0% | 98.0% |
| #883 | floor · high | Sep 18, 2026 | 0.149Ξ | 0.157Ξ | +5.4% | 94.6% |
| #8109 | floor · high | Sep 18, 2026 | 0.156Ξ | 0.148Ξ | -5.2% | 94.8% |
| #883 | floor · high | Sep 18, 2026 | 0.155Ξ | 0.148Ξ | -4.4% | 95.6% |
| #1583 | floor · high | Sep 18, 2026 | 0.154Ξ | 0.147Ξ | -4.6% | 95.5% |
| #6643 | floor · high | Sep 18, 2026 | 0.154Ξ | 0.152Ξ | -1.3% | 98.7% |
| #5870 | floor · high | Sep 18, 2026 | 0.156Ξ | 0.144Ξ | -7.7% | 92.3% |
| #1080 | floor · high | Sep 18, 2026 | 0.154Ξ | 0.148Ξ | -3.9% | 96.1% |
| #8858 | floor · high | Sep 18, 2026 | 0.156Ξ | 0.147Ξ | -5.8% | 94.2% |
| #6416 | floor · high | Sep 18, 2026 | 0.154Ξ | 0.150Ξ | -2.6% | 97.4% |
| #2239 | floor · high | Sep 18, 2026 | 0.154Ξ | 0.148Ξ | -3.9% | 96.1% |
| #6416 | floor · high | Sep 18, 2026 | 0.146Ξ | 0.157Ξ | +7.5% | 92.5% |
| #8858 | floor · high | Sep 18, 2026 | 0.145Ξ | 0.144Ξ | -0.7% | 99.3% |
| #3365 | floor · high | Sep 18, 2026 | 0.142Ξ | 0.152Ξ | +6.7% | 93.3% |
| #1727 | floor · high | Sep 18, 2026 | 0.149Ξ | 0.159Ξ | +6.7% | 93.3% |
| #2436 | floor · high | Sep 18, 2026 | 0.149Ξ | 0.152Ξ | +2.0% | 98.0% |
Model mfers-1789796135-2026-08-03.2 · evaluated through 9/19/2026. Method details on the methodology page.