mtgarb · buy globally → sell AU

v1 vs v2 — side-by-sideOperator gut-check. v1 = existing engine (point-estimate ROI rank); v2 = inference rebuild (tiered: deterministic buylist + Bayesian regression log-price posterior). Where they disagree most. Read large absolute deltas as "v2 thinks v1's $X estimate is off by Δ". Shadow-window discipline retired 2026-05-16 (operator amendment) — p50 + Δ visible again; width column kept as calibration health metric.
game:allmagicpokemonsort:abs Δv2 P(profit)Tier 1 firstv2 E[U]
magic · scan_run db9aef89· started 2026-05-16 14:07· v1 opps: 200 · v2 preds: 400 · Tier 1 eligible: 2800· model: v1.1-magic-tier3-2026-05-16T0320Z
Auditslatest per (agent, slice). Operator-evidence only — engine + portfolio do not read.
agentslicerecommendationfindingsn_predn_pairedconfwhen
comparable-sales-enumeratorprediction 1aaccd4dPROCEED_DEEP_DIVE121200.8210h agoSummary:
6 high-similarity (>=0.85) AU sold comparables for Fifth Dawn #114 nonfoil over 180d, ranging A$24.18 to A$47.12. Top-5 by similarity median ~A$45.53 (range A$24.18-A$47.12). Two LP-condition exact matches around A$24-30 (881d50a9, bda00fe0), three LP/MP exact matches A$45-47. Operator landed cost A$31.63 sits between the soft-floor LP cluster and the harder MP/no-#-but-named-set cluster. Cross-printing references cluster A$36-52 and support a A$40+ ceiling. EU CM EUR 16.57 (~A$27.34 floor), TCG USD 30.67 (~A$47.54 ceiling). 17 cross-printing rows and 1 SLD foil / 1 signed / 1 proxy excluded - verdict: deep dive justified.
anomaly-watchermagic tier=0 1dINVESTIGATE_AND_DERATE(high)3000.9010h agoSummary:
RED. Three HIGH-severity findings. (1) eBay AU listings 20% of 7d baseline (58 vs 294) - likely scraper outage or missing EBAY_APP_ID/CERT on VPS. (2)(3) Pokemon foil cm_eur 100% NULL for pre-2014 AND 2014-2020 eras (n=243 combined predictions) - matches documented foil-column inversion bug, landed costs understated 3-5x. Operator next: SSH to VPS to verify EBAY_APP_ID/CERT and re-run eBay scrape; concurrently derate or suppress Pokemon foil picks in pre-2020 eras until column-mapping fix lands.
calibration-auditormagic tier=3 90dINVESTIGATE_AND_DERATE(high)74413225740.6510h agoSummary:
Model exhibits two simultaneous systematic biases: (a) severe undershoot in the price_tier:300_plus slice (median +158% residual, coverage 22.7%) where the model both under-predicts AND has bands far too narrow — high-confidence evidence of pairing/cross-printing contamination in the source data, mirrored by the Mystic Remora alternating-realised pattern in recent_samples; (b) consistent overshoot across modern_2003_2014, recent_2015_2020, and sub-A$100 price tiers (-27% to -55% median residuals) suggesting expensive comparables are leaking into cheap-printing predictions. The prob_profit calibration line is non-monotonic (decile [0.2-0.3) collapses to 4.6% actual vs 26.8% predicted) and unsuitable for position sizing. Vintage_pre_2003 (n=10,032) is near-balanced (-8.3%) and the dominant slice masking issues in aggregate. Recommend INVESTIGATE_AND_DERATE: investigate pairing/printing-resolution in the realised-outcome join (cross-printing leak is the leading hypothesis per CLAUDE.md priors), and operator-side derate Tier 3 picks in price_tier:300_plus and the modern_2003_2014 era until investigation completes. No retraining or prior changes per pre-reg shadow integrity.
#cardsetfolv1 sellv1 profitv1 roiv1 velv2 p25v2 p50v2 p75v2 widthv2 P(profit)v2 E[U]T1Δ p50−v1Review
4Riding the Dilu Horseptk·$856$2751%weak$11$29$75$633.2%-936·$-827—
3Riding the Dilu Horseptk·$856$2751%weak$48$98$199$1519.5%-838·$-758—
2Juggernautlea·$695$2841%weak$13$34$86$737.4%-591·$-661—
1Juggernautlea·$695$2841%weak$60$122$248$18824.4%-466·$-573—
6Library of Lenglea·$362$1251%weak$13$33$86$7213.1%-358·$-328—
26Control Magiclea·$346$420%weak$14$35$89$768.9%-527·$-312—
5Library of Lenglea·$362$1251%weak$36$73$149$11322.9%-300·$-289—
44Ravages of Warptk·$284$360%weak$11$28$72$618.5%-441·$-256—
25Control Magiclea·$346$420%weak$54$109$222$16924.2%-421·$-237—
48Chrome Mox2xm·$239$210%medium$8$20$51$436.1%-406·$-219—
43Ravages of Warptk·$284$360%weak$33$67$136$10316.0%-385·$-217—
16Warrior's Oathptk·$235$721%weak$11$27$69$5915.2%-250·$-207—
162Black Knightleb·$218$140%weak$12$31$78$6611.7%-364·$-187—
15Warrior's Oathptk·$235$721%weak$25$50$103$7822.1%-215·$-184—
18Cursed Landlea·$210$661%weak$10$27$68$5717.2%-215·$-183—
47Chrome Mox2xm·$239$210%medium$33$67$136$10318.3%-340·$-172—
58Purelaceleb·$193$320%weak$11$27$69$5814.0%-269·$-166—
17Cursed Landlea·$210$661%weak$22$45$92$7022.8%-186·$-165—
10Chrome Moxmrd·$207$390%strong$19$48$123$10424.3%-258·$-159—
9Chrome Moxmrd·$207$390%strong$24$49$100$7618.4%-250·$-158—
84Zhang Fei, Fierce Warriorptk·$179$250%weak$10$25$64$5413.3%-264·$-154—
22Polluted Deltaons·$200$240%strong$21$53$136$11524.7%-278·$-147—
21Polluted Deltaons·$200$240%strong$27$55$111$8419.0%-270·$-146—
57Purelaceleb·$193$320%weak$25$51$104$7920.7%-232·$-142—
12Lady Sunptk·$167$822%weak$10$25$65$5532.2%-89·$-141—
161Black Knightleb·$218$140%weak$38$78$158$12024.6%-295·$-140—
7City of Brass8ed·$145$913%weak$3$6$12$94.9%-78·$-139—
83Zhang Fei, Fierce Warriorptk·$179$250%weak$24$49$100$7620.4%-228·$-130—
8City of Brass8ed·$145$913%weak$8$21$54$4536.2%-60·$-124—
34Thicket Basilisklea·$144$390%weak$11$27$69$5920.1%-184·$-116—
11Lady Sunptk·$167$822%weak$26$53$108$8253.6%-48·$-114—
54Dingus Eggleb·$143$330%weak$11$29$75$6320.4%-195·$-113—
40Jade Monolithleb·$139$370%weak$10$27$68$5820.2%-180·$-113—
62Karmalea·$133$320%weak$11$29$73$6221.6%-178·$-104—
117Giant Growthlea·$133$170%weak$12$30$78$6619.6%-213·$-102—
133Uthden Trolllea·$126$160%weak$10$27$68$5818.0%-206·$-99—
66Zhang He, Wei Generalptk·$121$300%weak$9$24$60$5119.8%-164·$-98—
39Jade Monolithleb·$139$370%weak$21$42$86$6525.1%-156·$-97—
50Camouflagelea·$123$341%weak$10$27$68$5723.0%-153·$-97—
32Wanderlustlea·$120$401%weak$10$26$67$5725.7%-130·$-94—
126Zhao Zilong, Tiger Generalptk·$116$170%weak$9$24$61$5217.9%-186·$-92—
33Thicket Basilisklea·$144$390%weak$26$52$106$8031.2%-147·$-92—
90Lady Zhurong, Warrior Queenptk·$111$240%weak$9$24$60$5120.5%-156·$-88—
336Disenchantlea·$119$60%weak$14$35$91$7722.7%-208·$-84—
118Giant Growthlea·$133$170%weak$24$49$100$7624.8%-184·$-84—
65Zhang He, Wei Generalptk·$121$300%weak$19$38$78$5925.4%-140·$-83—
53Dingus Eggleb·$143$330%weak$30$62$126$9535.1%-147·$-81—
136Throne of Bonelea·$106$160%weak$10$27$68$5822.1%-161·$-79—
259Steal Artifactlea·$107$80%weak$11$29$74$6221.4%-182·$-78—
49Camouflagelea·$123$341%weak$22$45$92$7032.0%-124·$-78—
Δ p50−v1: large negative = v2 thinks v1 over-estimates the realised AU price. Highlight thresholds: red >30% (or >A$5); ochre >15% (or >A$2); green within band. v2 width column (p75 − p25): wider = model less confident; red >A$20 (or 60% of v1). Review column: latest verdict from any specialist agent in .claude/agents/; hover for agent name + confidence + revised band. Operator-evidence (engine free to read post-shadow).

agent	slice	recommendation	findings	n_pred	n_paired	conf	when
comparable-sales-enumerator`prediction 1aaccd4d`PROCEED_DEEP_DIVE121200.8210h ago Summary: 6 high-similarity (>=0.85) AU sold comparables for Fifth Dawn #114 nonfoil over 180d, ranging A$24.18 to A$47.12. Top-5 by similarity median ~A$45.53 (range A$24.18-A$47.12). Two LP-condition exact matches around A$24-30 (881d50a9, bda00fe0), three LP/MP exact matches A$45-47. Operator landed cost A$31.63 sits between the soft-floor LP cluster and the harder MP/no-#-but-named-set cluster. Cross-printing references cluster A$36-52 and support a A$40+ ceiling. EU CM EUR 16.57 (~A$27.34 floor), TCG USD 30.67 (~A$47.54 ceiling). 17 cross-printing rows and 1 SLD foil / 1 signed / 1 proxy excluded - verdict: deep dive justified.
anomaly-watcher`magic tier=0 1d`INVESTIGATE_AND_DERATE(high)3000.9010h ago Summary: RED. Three HIGH-severity findings. (1) eBay AU listings 20% of 7d baseline (58 vs 294) - likely scraper outage or missing EBAY_APP_ID/CERT on VPS. (2)(3) Pokemon foil cm_eur 100% NULL for pre-2014 AND 2014-2020 eras (n=243 combined predictions) - matches documented foil-column inversion bug, landed costs understated 3-5x. Operator next: SSH to VPS to verify EBAY_APP_ID/CERT and re-run eBay scrape; concurrently derate or suppress Pokemon foil picks in pre-2020 eras until column-mapping fix lands.
calibration-auditor`magic tier=3 90d`INVESTIGATE_AND_DERATE(high)74413225740.6510h ago Summary: Model exhibits two simultaneous systematic biases: (a) severe undershoot in the price_tier:300_plus slice (median +158% residual, coverage 22.7%) where the model both under-predicts AND has bands far too narrow — high-confidence evidence of pairing/cross-printing contamination in the source data, mirrored by the Mystic Remora alternating-realised pattern in recent_samples; (b) consistent overshoot across modern_2003_2014, recent_2015_2020, and sub-A$100 price tiers (-27% to -55% median residuals) suggesting expensive comparables are leaking into cheap-printing predictions. The prob_profit calibration line is non-monotonic (decile [0.2-0.3) collapses to 4.6% actual vs 26.8% predicted) and unsuitable for position sizing. Vintage_pre_2003 (n=10,032) is near-balanced (-8.3%) and the dominant slice masking issues in aggregate. Recommend INVESTIGATE_AND_DERATE: investigate pairing/printing-resolution in the realised-outcome join (cross-printing leak is the leading hypothesis per CLAUDE.md priors), and operator-side derate Tier 3 picks in price_tier:300_plus and the modern_2003_2014 era until investigation completes. No retraining or prior changes per pre-reg shadow integrity.

#	card	set	fol	v1 sell	v1 profit	v1 roi	v1 vel	v2 p25	v2 p50	v2 p75	v2 width	v2 P(profit)	v2 E[U]	T1	Δ p50−v1	Review
4	Riding the Dilu Horse	ptk	·	$856	$275	1%	weak	$11	$29	$75	$63	3.2%	-936	·	$-827	—
3	Riding the Dilu Horse	ptk	·	$856	$275	1%	weak	$48	$98	$199	$151	9.5%	-838	·	$-758	—
2	Juggernaut	lea	·	$695	$284	1%	weak	$13	$34	$86	$73	7.4%	-591	·	$-661	—
1	Juggernaut	lea	·	$695	$284	1%	weak	$60	$122	$248	$188	24.4%	-466	·	$-573	—
6	Library of Leng	lea	·	$362	$125	1%	weak	$13	$33	$86	$72	13.1%	-358	·	$-328	—
26	Control Magic	lea	·	$346	$42	0%	weak	$14	$35	$89	$76	8.9%	-527	·	$-312	—
5	Library of Leng	lea	·	$362	$125	1%	weak	$36	$73	$149	$113	22.9%	-300	·	$-289	—
44	Ravages of War	ptk	·	$284	$36	0%	weak	$11	$28	$72	$61	8.5%	-441	·	$-256	—
25	Control Magic	lea	·	$346	$42	0%	weak	$54	$109	$222	$169	24.2%	-421	·	$-237	—
48	Chrome Mox	2xm	·	$239	$21	0%	medium	$8	$20	$51	$43	6.1%	-406	·	$-219	—
43	Ravages of War	ptk	·	$284	$36	0%	weak	$33	$67	$136	$103	16.0%	-385	·	$-217	—
16	Warrior's Oath	ptk	·	$235	$72	1%	weak	$11	$27	$69	$59	15.2%	-250	·	$-207	—
162	Black Knight	leb	·	$218	$14	0%	weak	$12	$31	$78	$66	11.7%	-364	·	$-187	—
15	Warrior's Oath	ptk	·	$235	$72	1%	weak	$25	$50	$103	$78	22.1%	-215	·	$-184	—
18	Cursed Land	lea	·	$210	$66	1%	weak	$10	$27	$68	$57	17.2%	-215	·	$-183	—
47	Chrome Mox	2xm	·	$239	$21	0%	medium	$33	$67	$136	$103	18.3%	-340	·	$-172	—
58	Purelace	leb	·	$193	$32	0%	weak	$11	$27	$69	$58	14.0%	-269	·	$-166	—
17	Cursed Land	lea	·	$210	$66	1%	weak	$22	$45	$92	$70	22.8%	-186	·	$-165	—
10	Chrome Mox	mrd	·	$207	$39	0%	strong	$19	$48	$123	$104	24.3%	-258	·	$-159	—
9	Chrome Mox	mrd	·	$207	$39	0%	strong	$24	$49	$100	$76	18.4%	-250	·	$-158	—
84	Zhang Fei, Fierce Warrior	ptk	·	$179	$25	0%	weak	$10	$25	$64	$54	13.3%	-264	·	$-154	—
22	Polluted Delta	ons	·	$200	$24	0%	strong	$21	$53	$136	$115	24.7%	-278	·	$-147	—
21	Polluted Delta	ons	·	$200	$24	0%	strong	$27	$55	$111	$84	19.0%	-270	·	$-146	—
57	Purelace	leb	·	$193	$32	0%	weak	$25	$51	$104	$79	20.7%	-232	·	$-142	—
12	Lady Sun	ptk	·	$167	$82	2%	weak	$10	$25	$65	$55	32.2%	-89	·	$-141	—
161	Black Knight	leb	·	$218	$14	0%	weak	$38	$78	$158	$120	24.6%	-295	·	$-140	—
7	City of Brass	8ed	·	$145	$91	3%	weak	$3	$6	$12	$9	4.9%	-78	·	$-139	—
83	Zhang Fei, Fierce Warrior	ptk	·	$179	$25	0%	weak	$24	$49	$100	$76	20.4%	-228	·	$-130	—
8	City of Brass	8ed	·	$145	$91	3%	weak	$8	$21	$54	$45	36.2%	-60	·	$-124	—
34	Thicket Basilisk	lea	·	$144	$39	0%	weak	$11	$27	$69	$59	20.1%	-184	·	$-116	—
11	Lady Sun	ptk	·	$167	$82	2%	weak	$26	$53	$108	$82	53.6%	-48	·	$-114	—
54	Dingus Egg	leb	·	$143	$33	0%	weak	$11	$29	$75	$63	20.4%	-195	·	$-113	—
40	Jade Monolith	leb	·	$139	$37	0%	weak	$10	$27	$68	$58	20.2%	-180	·	$-113	—
62	Karma	lea	·	$133	$32	0%	weak	$11	$29	$73	$62	21.6%	-178	·	$-104	—
117	Giant Growth	lea	·	$133	$17	0%	weak	$12	$30	$78	$66	19.6%	-213	·	$-102	—
133	Uthden Troll	lea	·	$126	$16	0%	weak	$10	$27	$68	$58	18.0%	-206	·	$-99	—
66	Zhang He, Wei General	ptk	·	$121	$30	0%	weak	$9	$24	$60	$51	19.8%	-164	·	$-98	—
39	Jade Monolith	leb	·	$139	$37	0%	weak	$21	$42	$86	$65	25.1%	-156	·	$-97	—
50	Camouflage	lea	·	$123	$34	1%	weak	$10	$27	$68	$57	23.0%	-153	·	$-97	—
32	Wanderlust	lea	·	$120	$40	1%	weak	$10	$26	$67	$57	25.7%	-130	·	$-94	—
126	Zhao Zilong, Tiger General	ptk	·	$116	$17	0%	weak	$9	$24	$61	$52	17.9%	-186	·	$-92	—
33	Thicket Basilisk	lea	·	$144	$39	0%	weak	$26	$52	$106	$80	31.2%	-147	·	$-92	—
90	Lady Zhurong, Warrior Queen	ptk	·	$111	$24	0%	weak	$9	$24	$60	$51	20.5%	-156	·	$-88	—
336	Disenchant	lea	·	$119	$6	0%	weak	$14	$35	$91	$77	22.7%	-208	·	$-84	—
118	Giant Growth	lea	·	$133	$17	0%	weak	$24	$49	$100	$76	24.8%	-184	·	$-84	—
65	Zhang He, Wei General	ptk	·	$121	$30	0%	weak	$19	$38	$78	$59	25.4%	-140	·	$-83	—
53	Dingus Egg	leb	·	$143	$33	0%	weak	$30	$62	$126	$95	35.1%	-147	·	$-81	—
136	Throne of Bone	lea	·	$106	$16	0%	weak	$10	$27	$68	$58	22.1%	-161	·	$-79	—
259	Steal Artifact	lea	·	$107	$8	0%	weak	$11	$29	$74	$62	21.4%	-182	·	$-78	—
49	Camouflage	lea	·	$123	$34	1%	weak	$22	$45	$92	$70	32.0%	-124	·	$-78	—