AppraisrAppraisr
Accuracy

How accurate is the CryptoPunks 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)
94.1%
Mean accuracy (100−APE)
90.9%
Within ±5%
44%
Within ±10%
70%
Within ±20%
90%
Bias
+0.2%

Accuracy by market tier

Floor, middle-market, and grail sales are scored separately so aggregate accuracy cannot hide a weak segment

TierSalesMedian accuracyWithin ±10%Bias
floor173294.5%73%-0.2%
middle10075.8%30%+7.5%
grail1873.9%17%+2.5%

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

ItemtypeTier / evidenceDateSaleModelErrorAccuracy
#6648Malemiddle · highAug 3, 202642.00Ξ59.84Ξ+42.5%57.5%
#8949Malefloor · highJul 31, 202635.00Ξ32.65Ξ-6.7%93.3%
#5327Malefloor · highJul 28, 202635.40Ξ34.11Ξ-3.6%96.4%
#5631Malefloor · highJul 28, 202648.69Ξ37.51Ξ-23.0%77.0%
#9945Femalefloor · insufficientJul 27, 202638.00Ξ33.58Ξ-11.6%88.4%
#4325Malefloor · highJul 25, 202638.49Ξ33.81Ξ-12.2%87.8%
#5917Femalefloor · insufficientJul 25, 202631.20Ξ30.50Ξ-2.2%97.8%
#3498Femalefloor · lowJul 25, 202631.99Ξ30.50Ξ-4.7%95.3%
#7323Malefloor · highJul 23, 202632.90Ξ31.18Ξ-5.2%94.8%
#4325Malefloor · highJul 23, 202632.50Ξ35.45Ξ+9.1%90.9%
#4710Femalefloor · lowJul 22, 202636.25Ξ32.74Ξ-9.7%90.3%
#1824Malefloor · highJul 22, 202634.50Ξ34.74Ξ+0.7%99.3%
#4256Malefloor · highJul 21, 202633.40Ξ35.64Ξ+6.7%93.3%
#270Femalefloor · lowJul 21, 202631.99Ξ31.03Ξ-3.0%97.0%
#2154Femalefloor · lowJul 21, 202633.00Ξ32.22Ξ-2.4%97.6%
#8992Malefloor · highJul 19, 202630.60Ξ33.06Ξ+8.0%92.0%
#7588Malefloor · highJul 18, 202669.00Ξ43.31Ξ-37.2%62.8%
#4822Malemiddle · highJul 16, 202667.50Ξ58.55Ξ-13.3%86.7%
#4059Malefloor · highJul 14, 202630.10Ξ34.64Ξ+15.1%84.9%
#270Femalefloor · lowJul 13, 202630.38Ξ32.59Ξ+7.3%92.7%
#2957Malefloor · highJul 12, 202634.87Ξ34.30Ξ-1.6%98.4%
#7361Malefloor · highJul 12, 202637.50Ξ38.31Ξ+2.2%97.8%
#6603Femalefloor · lowJul 12, 202632.47Ξ32.04Ξ-1.3%98.7%
#4782Femalefloor · insufficientJul 11, 202632.00Ξ33.04Ξ+3.3%96.8%
#1470Femalefloor · lowJul 7, 202652.50Ξ42.62Ξ-18.8%81.2%
#5168Femalemiddle · lowJul 6, 202639.99Ξ66.74Ξ+66.9%33.1%
#3378Malefloor · highJul 6, 202635.00Ξ33.92Ξ-3.1%96.9%
#382Femalemiddle · lowJul 4, 2026123.0Ξ89.47Ξ-27.3%72.7%
#9710Femalefloor · lowJul 4, 202632.48Ξ31.92Ξ-1.7%98.3%
#1930Femalefloor · lowJul 4, 202632.00Ξ32.93Ξ+2.9%97.1%
#1607Malefloor · highJul 4, 202634.60Ξ34.32Ξ-0.8%99.2%
#8655Malefloor · highJul 4, 202635.00Ξ33.00Ξ-5.7%94.3%
#2957Malefloor · highJul 3, 202633.25Ξ33.51Ξ+0.8%99.2%
#4654Femalefloor · lowJul 3, 202678.00Ξ43.75Ξ-43.9%56.1%
#2833Malemiddle · highJul 3, 202676.00Ξ61.16Ξ-19.5%80.5%
#6492Malefloor · highJul 3, 202632.24Ξ32.08Ξ-0.5%99.5%
#7138Femalemiddle · lowJul 3, 202663.33Ξ66.23Ξ+4.6%95.4%
#7519Malemiddle · highJul 3, 202658.89Ξ61.31Ξ+4.1%95.9%
#8276Femalefloor · lowJul 3, 202636.44Ξ38.02Ξ+4.3%95.7%
#3860Malefloor · highJul 2, 202642.00Ξ38.42Ξ-8.5%91.5%

Model cryptopunks-1785764387-2026-08-03.2 · evaluated through 8/3/2026. Method details on the methodology page.