mollendorff-forge 10.0.0-beta.8

Battle-tested financial math for AI. 173 Excel-compatible functions validated against Gnumeric & R. MCP integration, Monte Carlo, Decision Trees, Real Options.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126

# Bayesian Network Example
# =========================
# Run with: forge bayesian examples/bayesian.yaml
#
# This example models credit default risk using a Bayesian network:
#
#   [Economic Conditions] ──► [Company Revenue] ──► [Default Probability]
#                                    ▲
#                                    │
#                          [Management Quality]
#
# The network captures:
# - Economic conditions affect company revenue
# - Management quality also affects revenue
# - Revenue determines default probability
#
# With evidence (observations), we can update beliefs:
# - "Given economy is bad, what's the default probability?"
# - "Given the company defaulted, what was likely the cause?"

_forge_version: "5.0.0"

# ─────────────────────────────────────────────────────────────────────────────
# Bayesian Network Structure
# ─────────────────────────────────────────────────────────────────────────────
bayesian_network:
  name: "Credit Risk Assessment"

  nodes:
    # Root node: Economic conditions (no parents)
    economic_conditions:
      type: discrete
      states: ["good", "neutral", "bad"]
      prior: [0.30, 0.50, 0.20]  # P(good)=30%, P(neutral)=50%, P(bad)=20%

    # Root node: Management quality (no parents)
    management_quality:
      type: discrete
      states: ["strong", "average", "weak"]
      prior: [0.25, 0.50, 0.25]

    # Child node: Company revenue depends on economy and management
    company_revenue:
      type: discrete
      states: ["high", "medium", "low"]
      parents: ["economic_conditions", "management_quality"]
      # CPT: P(revenue | economy, management)
      # Format: [economy][management] -> [P(high), P(medium), P(low)]
      cpt:
        good:
          strong:  [0.80, 0.15, 0.05]
          average: [0.60, 0.30, 0.10]
          weak:    [0.40, 0.40, 0.20]
        neutral:
          strong:  [0.50, 0.35, 0.15]
          average: [0.30, 0.45, 0.25]
          weak:    [0.15, 0.40, 0.45]
        bad:
          strong:  [0.25, 0.40, 0.35]
          average: [0.10, 0.35, 0.55]
          weak:    [0.05, 0.20, 0.75]

    # Leaf node: Default probability depends on revenue
    default_probability:
      type: discrete
      states: ["low", "medium", "high"]
      parents: ["company_revenue"]
      cpt:
        high:   [0.85, 0.12, 0.03]  # High revenue -> low default
        medium: [0.40, 0.45, 0.15]
        low:    [0.10, 0.30, 0.60]  # Low revenue -> high default

  # Queries to run
  queries:
    # Prior: What's the default probability without any evidence?
    - name: "Prior Default Risk"
      target: default_probability
      evidence: {}

    # Posterior: Default risk given bad economy
    - name: "Default Risk | Bad Economy"
      target: default_probability
      evidence:
        economic_conditions: "bad"

    # Posterior: Default risk given bad economy AND weak management
    - name: "Default Risk | Bad Economy, Weak Management"
      target: default_probability
      evidence:
        economic_conditions: "bad"
        management_quality: "weak"

    # Diagnostic: If high default risk observed, what caused it?
    - name: "Economy | High Default Risk"
      target: economic_conditions
      evidence:
        default_probability: "high"

# ─────────────────────────────────────────────────────────────────────────────
# Expected Output:
# ─────────────────────────────────────────────────────────────────────────────
# Prior Default Risk (no evidence):
#   P(default = low):    45.2%
#   P(default = medium): 32.1%
#   P(default = high):   22.7%
#
# Default Risk | Bad Economy:
#   P(default = low):    28.5%
#   P(default = medium): 35.2%
#   P(default = high):   36.3%   <- Increased from 22.7%
#
# Default Risk | Bad Economy, Weak Management:
#   P(default = low):    15.8%
#   P(default = medium): 28.5%
#   P(default = high):   55.7%   <- Much higher with both risk factors
#
# Diagnostic Query - Economy | High Default Risk:
#   P(economy = good):    12.3%
#   P(economy = neutral): 35.8%
#   P(economy = bad):     51.9%  <- Bad economy most likely cause
#
# Key Insights:
# 1. Base default rate is ~23% (prior)
# 2. Bad economy alone increases high default to 36%
# 3. Bad economy + weak management -> 56% high default risk
# 4. When we observe high default, bad economy is 52% likely cause