Accuracy
How accurate is the Doodles 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)
95.3%
Mean accuracy (100−APE)
92.9%
Within ±5%
52%
Within ±10%
80%
Within ±20%
95%
Bias
-0.2%
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 | 1990 | 95.5% | 83% | +0.2% |
| middle | 29 | 51.9% | 3% | +10.5% |
| grail | 90 | 89.7% | 48% | -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 |
|---|---|---|---|---|---|---|
| #5244 | floor · high | Sep 18, 2026 | 0.311Ξ | 0.304Ξ | -2.3% | 97.8% |
| #5595 | floor · high | Sep 18, 2026 | 0.354Ξ | 0.322Ξ | -9.1% | 90.9% |
| #2411 | floor · high | Sep 17, 2026 | 0.310Ξ | 0.294Ξ | -5.2% | 94.8% |
| #5911 | floor · high | Sep 17, 2026 | 0.304Ξ | 0.294Ξ | -3.2% | 96.8% |
| #4959 | floor · high | Sep 17, 2026 | 0.310Ξ | 0.289Ξ | -6.8% | 93.2% |
| #1513 | floor · high | Sep 17, 2026 | 0.314Ξ | 0.287Ξ | -8.6% | 91.4% |
| #4094 | floor · high | Sep 17, 2026 | 0.309Ξ | 0.281Ξ | -9.1% | 90.9% |
| #4187 | floor · high | Sep 17, 2026 | 0.329Ξ | 0.294Ξ | -10.6% | 89.4% |
| #5244 | floor · high | Sep 17, 2026 | 0.304Ξ | 0.280Ξ | -7.8% | 92.2% |
| #5361 | floor · high | Sep 17, 2026 | 0.304Ξ | 0.301Ξ | -0.9% | 99.1% |
| #1313 | floor · high | Sep 17, 2026 | 0.298Ξ | 0.297Ξ | -0.2% | 99.8% |
| #7299 | floor · high | Sep 17, 2026 | 0.292Ξ | 0.300Ξ | +2.8% | 97.2% |
| #8266 | grail · high | Sep 17, 2026 | 0.450Ξ | 0.321Ξ | -28.7% | 71.3% |
| #2411 | floor · high | Sep 17, 2026 | 0.290Ξ | 0.316Ξ | +9.0% | 91.0% |
| #1313 | floor · high | Sep 17, 2026 | 0.296Ξ | 0.291Ξ | -1.7% | 98.3% |
| #7405 | floor · high | Sep 17, 2026 | 0.308Ξ | 0.299Ξ | -2.9% | 97.1% |
| #5361 | floor · high | Sep 17, 2026 | 0.300Ξ | 0.318Ξ | +6.0% | 94.0% |
| #5918 | floor · high | Sep 17, 2026 | 0.309Ξ | 0.303Ξ | -1.9% | 98.1% |
| #7299 | floor · high | Sep 17, 2026 | 0.299Ξ | 0.300Ξ | +0.3% | 99.7% |
| #6837 | floor · high | Sep 17, 2026 | 0.329Ξ | 0.301Ξ | -8.5% | 91.6% |
| #4432 | floor · high | Sep 17, 2026 | 0.310Ξ | 0.276Ξ | -11.0% | 89.0% |
| #1595 | floor · high | Sep 17, 2026 | 0.309Ξ | 0.279Ξ | -9.7% | 90.3% |
| #4981 | floor · high | Sep 17, 2026 | 0.298Ξ | 0.287Ξ | -3.6% | 96.4% |
| #1605 | floor · high | Sep 17, 2026 | 0.292Ξ | 0.284Ξ | -2.7% | 97.3% |
| #2946 | middle · medium | Sep 17, 2026 | 1.450Ξ | 0.801Ξ | -44.8% | 55.2% |
| #7930 | grail · low | Sep 17, 2026 | 1.270Ξ | 1.402Ξ | +10.4% | 89.6% |
| #7320 | grail · low | Sep 17, 2026 | 1.270Ξ | 1.401Ξ | +10.3% | 89.7% |
| #9279 | grail · low | Sep 17, 2026 | 1.180Ξ | 1.095Ξ | -7.2% | 92.8% |
| #2870 | floor · high | Sep 17, 2026 | 0.286Ξ | 0.288Ξ | +0.7% | 99.3% |
| #2069 | grail · medium | Sep 17, 2026 | 1.110Ξ | 0.895Ξ | -19.4% | 80.6% |
| #8033 | floor · high | Sep 17, 2026 | 0.304Ξ | 0.287Ξ | -5.5% | 94.5% |
| #1842 | floor · high | Sep 17, 2026 | 0.282Ξ | 0.276Ξ | -2.1% | 97.9% |
| #1605 | floor · high | Sep 17, 2026 | 0.278Ξ | 0.292Ξ | +5.0% | 95.0% |
| #3811 | floor · high | Sep 16, 2026 | 0.286Ξ | 0.291Ξ | +1.8% | 98.2% |
| #4981 | floor · high | Sep 16, 2026 | 0.284Ξ | 0.281Ξ | -1.1% | 98.9% |
| #4092 | floor · high | Sep 16, 2026 | 0.286Ξ | 0.285Ξ | -0.3% | 99.7% |
| #8033 | floor · high | Sep 16, 2026 | 0.282Ξ | 0.294Ξ | +4.3% | 95.7% |
| #2064 | floor · high | Sep 16, 2026 | 0.288Ξ | 0.284Ξ | -1.4% | 98.6% |
| #8291 | floor · high | Sep 16, 2026 | 0.286Ξ | 0.285Ξ | -0.3% | 99.7% |
| #4959 | floor · high | Sep 16, 2026 | 0.282Ξ | 0.290Ξ | +2.8% | 97.2% |
Model doodles-1789698203-2026-08-03.2 · evaluated through 9/18/2026. Method details on the methodology page.