HScreen

rendered 2026-08-23 04:19 UTC

The screen every researched name · ranked · decides nothing

researched names on the screen36
fund / absolute barALPHA · 10.0%
A near-miss is not hidden and a broken verdict is not a filter — both are shown with their numbers so the judgement stays with you. gap is expected return less the bar. beta is the measured common-factor loading: negative means the name diversifies the AI-capex factor.
nameER%gapbetaverdictstatussectorwhy
2330.TW22.3+12.3+0.035brokenQUEUEDTechnologyadversary verdict=broken
GOOGL20.7+10.7-0.006brokenQUEUEDCommunication Servicesadversary verdict=broken
QCOM17.8+7.8+0.098weakenedHELDTechnology6.40% of alpha
VST17.7+7.7+0.108weakenedHELDUtilities6.40% of alpha
AMAT17.2+7.2+0.190brokenQUEUEDTechnologyadversary verdict=broken
ORCL16.5+6.5+0.060weakenedHELDTechnology3.00% of alpha
ETN15.0+5.0+0.091weakenedHELDIndustrials6.40% of alpha
AMD14.9+4.9+0.197weakenedQUEUEDTechnologyben_veto (2026-08-18): option_gate FALSE — the only hard negative in t
NVDA14.8+4.8+0.145weakenedHELDTechnology10.67% of alpha
8035.T14.4+4.4+0.056weakenedHELDTechnology6.40% of alpha
META13.8+3.8+0.000weakenedHELDCommunication Services3.00% of alpha
AVGO13.4+3.4+0.148brokenQUEUEDTechnologyadversary verdict=broken
6146.T12.9+2.9+0.043weakenedQUEUEDTechnologygross exposure cap: only 0.00% of capital unallocated, below position_
MSFT12.8+2.8-0.012weakenedHELDTechnology3.00% of alpha
ASML12.8+2.8+0.136brokenQUEUEDTechnologyadversary verdict=broken
005930.KS12.7+2.7+0.069weakenedHELDTechnology10.67% of alpha
CEG12.7+2.7+0.084weakenedHELDUtilities4.80% of alpha
ANET12.3+2.3+0.145weakenedQUEUEDTechnologygross exposure cap: only 0.00% of capital unallocated, below position_
MU12.1+2.1+0.239weakenedHELDTechnology3.00% of alpha
AMZN11.4+1.4-0.000brokenQUEUEDConsumer Cyclicaladversary verdict=broken
ARM10.8+0.8+0.169weakenedQUEUEDTechnologydropped by the D-023 rescale: 3.00% x 0.5333 = 1.60%, below position_m
LRCX10.8+0.8+0.206weakenedHELDTechnology3.00% of alpha
KLAC10.7+0.7+0.187weakenedHELDTechnology3.00% of alpha
MRVL10.5+0.5+0.248weakenedQUEUEDTechnologytheme 'foundry-logic' weight cap 35% binding
LIN9.1-0.9-0.032brokenSCREENED-OUTBasic Materialsadversary verdict=broken; expected_return 9.1% (probability_weighted,
GEV8.8-1.2+0.126weakenedSCREENED-OUTUtilitiesexpected_return 8.7% (probability_weighted, annualised from 8.8% over
ON8.6-1.4+0.164weakenedSCREENED-OUTTechnologyexpected_return 8.6% (probability_weighted, annualised from 8.6% over
VRT8.3-1.7+0.214weakenedSCREENED-OUTIndustrialsexpected_return 8.3% (probability_weighted, annualised from 8.3% over
RTX7.2-2.8-0.032weakenedSCREENED-OUTIndustrialsexpected_return 7.2% (probability_weighted, annualised from 7.2% over
PWR7.1-2.9+0.098brokenSCREENED-OUTIndustrialsadversary verdict=broken; expected_return 7.1% (probability_weighted,
DIS7.1-2.9-0.036weakenedSCREENED-OUTCommunication Servicesexpected_return 7.1% (probability_weighted, annualised from 7.1% over
000660.KS7.0-3.0+0.104brokenSCREENED-OUTTechnologyadversary verdict=broken; expected_return 7.0% (probability_weighted,
INTC6.9-3.1+0.139weakenedSCREENED-OUTTechnologyexpected_return 6.9% (probability_weighted, annualised from 6.9% over
DLR6.2-3.8+0.006weakenedSCREENED-OUTReal Estateexpected_return 6.2% (probability_weighted, annualised from 6.2% over
EQIX3.6-6.4-0.012weakenedSCREENED-OUTReal Estateexpected_return 3.6% (probability_weighted, annualised from 3.6% over
2327.TW-2.2-12.2+0.042weakenedSCREENED-OUTTechnologycoverage_tier=T3-DATA-ONLY not in ['T1-FULL', 'T2-PARTIAL', 'T3-PRIMAR

ALPHA — weights and spread 15 names · gross 75.74% · cash 24.26%

sectorweight%share of grossnames
Technology49.1464.9%005930.KS, 8035.T, KLAC, LRCX, MSFT, MU, NVDA, ORCL, QCOM
Utilities11.2014.8%CEG, VST
Industrials6.408.4%ETN
Real Estate3.004.0%VICI
Energy3.004.0%EPD
Communication Services3.004.0%META
measured pairs105
average pair co-movement0.362
factor beta (PC1)0.0891
Not two bets. These pairs co-move at 0.75 or above on the canonical FX-corrected matrix — at that level they are one position held twice:
  KLAC · LRCX +0.898
  LRCX · MU +0.755
Different sectors, one trade. A sector table alone would call these diversified — the measurement disagrees:
  ETN (Industrials) · KLAC (Technology) +0.611
  ETN (Industrials) · LRCX (Technology) +0.609

INCOME — weights and spread 20 names · gross 73.0% · cash 27.0%

sectorweight%share of grossnames
Consumer Defensive21.8029.9%GIS, KO, MO, PEP, PG, PM
Energy19.6026.8%0857.HK, CVX, EPD, XOM
Communication Services8.8012.1%0941.HK, VZ
Real Estate8.8012.1%VICI
Technology8.0011.0%CSCO, IBM, MSFT, QCOM
Healthcare4.005.5%AMGN, JNJ
Utilities2.002.7%0002.HK
measured pairs190
average pair co-movement0.2403
factor beta (PC1)-0.0373
Not two bets. These pairs co-move at 0.75 or above on the canonical FX-corrected matrix — at that level they are one position held twice:
  CVX · XOM +0.843

Household both sleeves, summed dollars

sector% of invested
Technology38.42
Energy15.19
Consumer Defensive14.66
Utilities8.87
Real Estate7.93
Communication Services7.93
Industrials4.30
Healthcare2.69

Evidence attribution is the depot evidence actually about this company?

universe names measured184
below the 40.0% on-company floor41
A pack assembled from other companies' documents produces a confident, fully-cited dossier about the wrong business. Measured per name as the share of retrieved depot chunks that mention the subject company. The analyst now refuses to run below the floor.
nameon-companychunksnote
2327.TW18%2/11a thesis already exists on this evidence
Rendered from canonical artifacts only (T1) — no terminal database exists. Every figure is read from the system's own files at render time; where an artifact is missing the panel names the dependency rather than showing an empty box. No execution path exists anywhere in this interface (T5), not even a link.