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)
93.9%
Mean accuracy (100−APE)
91.2%
Within ±5%
40%
Within ±10%
72%
Within ±20%
92%
Bias
-2.0%
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 | 1577 | 94.0% | 73% | -1.7% |
| middle | 28 | 73.0% | 32% | -10.8% |
| grail | 16 | 81.8% | 31% | -21.5% |
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 |
|---|---|---|---|---|---|---|
| #6874 | floor · high | Nov 3, 2026 | 0.137Ξ | 0.141Ξ | +2.9% | 97.1% |
| #7434 | floor · high | Nov 3, 2026 | 0.129Ξ | 0.122Ξ | -5.4% | 94.6% |
| #7639 | floor · high | Nov 2, 2026 | 0.146Ξ | 0.134Ξ | -8.2% | 91.8% |
| #6328 | floor · high | Nov 2, 2026 | 0.143Ξ | 0.144Ξ | +0.7% | 99.3% |
| #4713 | floor · high | Nov 2, 2026 | 0.135Ξ | 0.139Ξ | +3.0% | 97.0% |
| #5376 | floor · high | Nov 2, 2026 | 0.139Ξ | 0.128Ξ | -7.9% | 92.1% |
| #7202 | floor · high | Nov 2, 2026 | 0.135Ξ | 0.132Ξ | -2.2% | 97.8% |
| #5591 | floor · high | Nov 2, 2026 | 0.137Ξ | 0.142Ξ | +3.7% | 96.4% |
| #8009 | floor · high | Nov 2, 2026 | 0.134Ξ | 0.126Ξ | -6.0% | 94.0% |
| #2861 | floor · high | Nov 2, 2026 | 0.135Ξ | 0.130Ξ | -3.7% | 96.3% |
| #4327 | floor · high | Nov 2, 2026 | 0.135Ξ | 0.131Ξ | -3.0% | 97.0% |
| #5962 | floor · high | Nov 2, 2026 | 0.136Ξ | 0.127Ξ | -6.6% | 93.4% |
| #4713 | floor · high | Nov 2, 2026 | 0.139Ξ | 0.130Ξ | -6.5% | 93.5% |
| #2861 | floor · high | Nov 2, 2026 | 0.129Ξ | 0.129Ξ | 0.0% | 100.0% |
| #7824 | floor · high | Nov 1, 2026 | 0.133Ξ | 0.126Ξ | -5.3% | 94.7% |
| #1531 | floor · high | Nov 1, 2026 | 0.125Ξ | 0.127Ξ | +1.6% | 98.4% |
| #8242 | floor · high | Nov 1, 2026 | 0.125Ξ | 0.129Ξ | +3.2% | 96.8% |
| #4000 | floor · high | Nov 1, 2026 | 0.124Ξ | 0.127Ξ | +2.4% | 97.6% |
| #4403 | floor · high | Nov 1, 2026 | 0.124Ξ | 0.128Ξ | +3.2% | 96.8% |
| #1282 | floor · high | Nov 1, 2026 | 0.129Ξ | 0.123Ξ | -4.7% | 95.4% |
| #8109 | floor · high | Oct 31, 2026 | 0.122Ξ | 0.121Ξ | -0.8% | 99.2% |
| #7061 | floor · high | Oct 31, 2026 | 0.123Ξ | 0.124Ξ | +0.8% | 99.2% |
| #4676 | floor · high | Oct 31, 2026 | 0.124Ξ | 0.123Ξ | -0.8% | 99.2% |
| #4000 | floor · high | Oct 31, 2026 | 0.124Ξ | 0.122Ξ | -1.6% | 98.4% |
| #8242 | floor · high | Oct 31, 2026 | 0.124Ξ | 0.122Ξ | -1.6% | 98.4% |
| #1282 | floor · high | Oct 31, 2026 | 0.123Ξ | 0.120Ξ | -2.4% | 97.6% |
| #1531 | floor · high | Oct 31, 2026 | 0.124Ξ | 0.125Ξ | +0.8% | 99.2% |
| #4403 | floor · high | Oct 31, 2026 | 0.124Ξ | 0.121Ξ | -2.4% | 97.6% |
| #5434 | floor · high | Oct 31, 2026 | 0.128Ξ | 0.130Ξ | +1.6% | 98.4% |
| #5591 | floor · high | Oct 31, 2026 | 0.138Ξ | 0.129Ξ | -6.5% | 93.5% |
| #5591 | floor · high | Oct 31, 2026 | 0.129Ξ | 0.126Ξ | -2.3% | 97.8% |
| #5434 | floor · high | Oct 31, 2026 | 0.125Ξ | 0.139Ξ | +11.2% | 88.8% |
| #435 | floor · high | Oct 31, 2026 | 0.124Ξ | 0.118Ξ | -4.8% | 95.2% |
| #4455 | floor · high | Oct 31, 2026 | 0.129Ξ | 0.118Ξ | -8.5% | 91.5% |
| #5320 | floor · high | Oct 31, 2026 | 0.129Ξ | 0.124Ξ | -3.9% | 96.1% |
| #8595 | floor · high | Oct 31, 2026 | 0.123Ξ | 0.123Ξ | -0.4% | 99.6% |
| #3931 | floor · high | Oct 31, 2026 | 0.124Ξ | 0.118Ξ | -4.7% | 95.3% |
| #5397 | floor · high | Oct 31, 2026 | 0.123Ξ | 0.124Ξ | +0.8% | 99.2% |
| #8074 | middle · high | Oct 30, 2026 | 0.226Ξ | 0.262Ξ | +15.9% | 84.1% |
| #3321 | floor · high | Oct 29, 2026 | 0.138Ξ | 0.122Ξ | -11.6% | 88.4% |
Model mfers-1793703479-2026-08-03.2 · evaluated through 11/3/2026. Method details on the methodology page.