Skip to Content
Last modified by Robert Schaub on 2026/02/08 08:31
Show last authors
1 {{info}}
2 **Current Implementation (v2.6.33)** - Simplified model without versioning. Uses 7-point symmetric verdict scale.
3 {{/info}}
4
5 = Evidence and Verdict Data Model =
6
7 {{mermaid}}
8
9 erDiagram
10 CLAIM ||--|| CLAIM_VERDICT : has
11 CLAIM_VERDICT }o--o{ FACT : supported_by
12 FACT }o--|| SOURCE : from
13
14 CLAIM {
15 string id_PK
16 string text
17 string type
18 string claimRole
19 boolean isCentral
20 string_array dependsOn
21 }
22
23 CLAIM_VERDICT {
24 string id_PK
25 string claimId_FK
26 string verdict
27 int truthPercentage
28 int confidence
29 string explanation
30 string_array supportingFactIds
31 string_array opposingFactIds
32 string contestationStatus
33 float harmPotential
34 }
35
36 FACT {
37 string id_PK
38 string sourceId_FK
39 string text
40 string quote
41 string relevance
42 boolean supports
43 string extractedContext
44 }
45
46 SOURCE {
47 string id_PK
48 string name
49 string domain
50 string url
51 float reliabilityScore
52 string bias
53 string factualReporting
54 }
55
56 {{/mermaid}}
57
58 = Verdict Generation Flow =
59
60 {{mermaid}}
61
62 flowchart TB
63 subgraph Research[Research Phase]
64 FACTS[Collected Facts]
65 SOURCES[Source Metadata]
66 end
67
68 subgraph Analysis[Analysis]
69 WEIGHT[Weight Evidence by source reliability]
70 CONTEST[Check Contestation doubted vs contested]
71 HARM[Assess Harm Potential]
72 end
73
74 subgraph Verdict[Verdict Generation]
75 CALC[Calculate Truth Percentage]
76 MAP[Map to 7-point Scale]
77 CONF[Assign Confidence]
78 end
79
80 subgraph Output[Result]
81 CLAIM_V[Claim Verdict]
82 ARTICLE_V[Article Verdict]
83 end
84
85 FACTS --> WEIGHT
86 SOURCES --> WEIGHT
87 WEIGHT --> CONTEST
88 CONTEST --> HARM
89 HARM --> CALC
90 CALC --> MAP
91 MAP --> CONF
92 CONF --> CLAIM_V
93 CLAIM_V --> ARTICLE_V
94
95 {{/mermaid}}
96
97 == 7-Point Verdict Scale ==
98
99 |= Verdict |= Truth % Range |= Description
100 | **TRUE** | 86-100% | Claim is well-supported by evidence
101 | **MOSTLY-TRUE** | 72-85% | Largely accurate with minor caveats
102 | **LEANING-TRUE** | 58-71% | More evidence supports than contradicts
103 | **MIXED** | 43-57% (high conf) | Roughly equal evidence both ways
104 | **UNVERIFIED** | 43-57% (low conf) | Insufficient evidence to determine
105 | **LEANING-FALSE** | 29-42% | More evidence contradicts than supports
106 | **MOSTLY-FALSE** | 15-28% | Largely inaccurate
107 | **FALSE** | 0-14% | Claim is refuted by evidence
108
109 == Contestation Status ==
110
111 * **Doubted**: Evidence is weak, uncertain, or ambiguous
112 * **Contested**: Strong evidence exists on both sides
113
114 == Source Reliability ==
115
116 Source reliability scores come from external **MBFC (Media Bias/Fact Check) bundle**:
117 * Pre-loaded reliability scores for known sources
118 * Configurable via ##source-reliability.ts##