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)
94.0%
Mean accuracy (100−APE)
92.0%
Within ±5%
44%
Within ±10%
73%
Within ±20%
94%
Bias
+2.3%
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 | 2104 | 94.3% | 74% | +2.7% |
| middle | 25 | 77.7% | 16% | +13.4% |
| grail | 90 | 89.8% | 49% | -11.1% |
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 |
|---|---|---|---|---|---|---|
| #7116 | floor · high | Nov 5, 2026 | 0.335Ξ | 0.302Ξ | -9.9% | 90.2% |
| #5791 | floor · high | Nov 4, 2026 | 0.261Ξ | 0.275Ξ | +5.4% | 94.6% |
| #807 | floor · high | Nov 4, 2026 | 0.296Ξ | 0.268Ξ | -9.5% | 90.5% |
| #2714 | floor · high | Nov 4, 2026 | 0.269Ξ | 0.294Ξ | +9.3% | 90.7% |
| #6374 | grail · medium | Nov 4, 2026 | 1.080Ξ | 1.098Ξ | +1.7% | 98.3% |
| #4681 | floor · high | Nov 4, 2026 | 0.370Ξ | 0.307Ξ | -17.0% | 83.0% |
| #1430 | floor · high | Nov 4, 2026 | 0.277Ξ | 0.294Ξ | +6.1% | 93.9% |
| #6374 | grail · medium | Nov 3, 2026 | 1.100Ξ | 1.018Ξ | -7.5% | 92.6% |
| #4061 | floor · high | Nov 3, 2026 | 0.276Ξ | 0.271Ξ | -1.8% | 98.2% |
| #4465 | floor · high | Nov 3, 2026 | 0.290Ξ | 0.265Ξ | -8.6% | 91.4% |
| #3032 | floor · high | Nov 3, 2026 | 0.260Ξ | 0.269Ξ | +3.5% | 96.5% |
| #3032 | floor · high | Nov 3, 2026 | 0.276Ξ | 0.260Ξ | -5.8% | 94.2% |
| #4813 | floor · high | Nov 2, 2026 | 0.309Ξ | 0.327Ξ | +5.8% | 94.2% |
| #3258 | floor · high | Nov 2, 2026 | 0.273Ξ | 0.266Ξ | -2.5% | 97.5% |
| #3007 | floor · high | Nov 2, 2026 | 0.273Ξ | 0.281Ξ | +3.0% | 97.1% |
| #3974 | floor · high | Nov 2, 2026 | 0.261Ξ | 0.259Ξ | -0.8% | 99.2% |
| #3258 | floor · high | Nov 2, 2026 | 0.261Ξ | 0.261Ξ | 0.0% | 100.0% |
| #3854 | floor · high | Nov 2, 2026 | 0.273Ξ | 0.256Ξ | -6.2% | 93.8% |
| #738 | floor · high | Nov 2, 2026 | 0.257Ξ | 0.270Ξ | +5.1% | 94.9% |
| #6879 | floor · high | Nov 2, 2026 | 0.290Ξ | 0.262Ξ | -9.7% | 90.3% |
| #1598 | floor · high | Nov 2, 2026 | 0.284Ξ | 0.248Ξ | -12.7% | 87.3% |
| #9885 | floor · high | Nov 2, 2026 | 0.269Ξ | 0.272Ξ | +1.1% | 98.9% |
| #3550 | floor · high | Nov 1, 2026 | 0.268Ξ | 0.262Ξ | -2.1% | 97.9% |
| #1369 | floor · high | Nov 1, 2026 | 0.257Ξ | 0.262Ξ | +2.0% | 98.1% |
| #8489 | floor · high | Nov 1, 2026 | 0.267Ξ | 0.264Ξ | -1.3% | 98.7% |
| #1369 | floor · high | Nov 1, 2026 | 0.262Ξ | 0.253Ξ | -3.5% | 96.5% |
| #1430 | floor · high | Nov 1, 2026 | 0.288Ξ | 0.281Ξ | -2.4% | 97.6% |
| #403 | floor · high | Nov 1, 2026 | 0.252Ξ | 0.258Ξ | +2.5% | 97.5% |
| #1605 | floor · high | Nov 1, 2026 | 0.255Ξ | 0.255Ξ | 0.0% | 100.0% |
| #4613 | floor · high | Nov 1, 2026 | 0.262Ξ | 0.270Ξ | +3.0% | 97.0% |
| #1369 | floor · high | Oct 31, 2026 | 0.249Ξ | 0.262Ξ | +5.2% | 94.8% |
| #3032 | floor · high | Oct 31, 2026 | 0.248Ξ | 0.262Ξ | +5.7% | 94.4% |
| #1482 | floor · high | Oct 31, 2026 | 0.252Ξ | 0.259Ξ | +2.8% | 97.2% |
| #3803 | floor · high | Oct 31, 2026 | 0.339Ξ | 0.348Ξ | +2.7% | 97.4% |
| #110 | floor · high | Oct 31, 2026 | 0.260Ξ | 0.281Ξ | +8.1% | 91.9% |
| #9667 | floor · high | Oct 31, 2026 | 0.348Ξ | 0.269Ξ | -22.7% | 77.3% |
| #3550 | floor · high | Oct 31, 2026 | 0.246Ξ | 0.264Ξ | +7.3% | 92.7% |
| #7588 | floor · high | Oct 31, 2026 | 0.257Ξ | 0.252Ξ | -1.9% | 98.1% |
| #3032 | floor · high | Oct 31, 2026 | 0.260Ξ | 0.253Ξ | -2.7% | 97.3% |
| #1482 | floor · high | Oct 31, 2026 | 0.260Ξ | 0.266Ξ | +2.3% | 97.7% |
Model doodles-1793850791-2026-08-03.2 · evaluated through 11/5/2026. Method details on the methodology page.