MacroPulze

[CLOSED BETA]
Methods · Tutorials · Open Access_

Markets reward those who understand how to think under uncertainty. MacroPulze teaches the quantitative methods behind institutional analysis: hypothesise, formalise, confront with data, revise. Scientific method as investable skill.

Read more on the blog →

§ Goal

Build a self-improving model of the financial system itself

MacroPulze is not a collection of standalone tools. The explicit long-term goal is to construct an underlying generative model of the financial system — one that explains, predicts, and revises its own structure. The theoretical engine is the Free Energy Principle: every model, every indicator, every simulation on this platform is a component of a system that minimises surprise by continuously updating its beliefs against incoming evidence.

F = Eq[ln q(θ) − ln p(y, θ)] → min

The system maintains a probabilistic world model of macroeconomic structure, market regimes, and inter-agent dynamics. When new evidence arrives — a policy shift, a liquidity event, a structural break — the model updates. The tools you learn here are the components of that system.

i. Perceive

Evidence Extraction

Ingest filings, macro data, network topology. Score observations against the current model's predictions.

ii. Update

Belief Revision

Minimise free energy by adjusting priors — tighten distributions where the model is right, widen them where it's surprised.

iii. Act

Active Inference

Generate scenarios and counterfactuals that reduce expected uncertainty — the system seeks information, not confirmation.

Four learning paths. Each teaches a method — not an outcome.
$ ./learn.sh_
macro_simulation
Simulation & Scenario Design
Learn to build multi-agent simulations and stress tests that expose structural dependencies, generate scenarios, and surface parameter sensitivity across macroeconomic hypotheses.
evidence_extraction
Evidence & Calibration Methods
Master domain-adapted techniques for reading filings, policy documents, and research — scoring evidence against explicit assumptions with out-of-sample validation.
decision_frameworks
Uncertainty Quantification
Study the methods behind probability distributions, counterfactual reasoning, and audit trails. Understand how institutional analysts structure decisions under incomplete information.
ai_integration
AI-Augmented Workflows
Build knowledge bases, connect data pipelines via MCP Server, and design self-improving analytical workflows. Learn to use AI as a methodological partner — not a black box.
Structured learning paths from foundations to applied analysis. All tutorials use historical data — you build the skill, you make the decisions.
T-01
Foundations4 parts · ~3h

Bayesian Inference for Analysts

Prior beliefs, likelihood functions, posterior updating — the epistemic engine behind every tool on this platform.

T-02
Foundations3 parts · ~2h

Monte Carlo Simulation from Scratch

Random sampling, convergence, variance reduction. Build a simulation engine step by step and understand why point estimates lie.

T-03
Methods5 parts · ~4h

Probabilistic DCF — Valuation Under Uncertainty

From deterministic spreadsheet to full probability distribution. Model assumptions as distributions, propagate uncertainty, interpret confidence bands.

T-04
Methods4 parts · ~3h

Context-Aware RSI — Beyond the Oscillator

Why RSI alone is incomplete. Integrating global M2 liquidity and news sentiment into a context-dependent signal.

T-05
Methods6 parts · ~5h

Geometric Economics — Ricci Curvature on Financial Networks

Differential geometry meets systemic risk. Construct correlation networks, compute discrete Ricci curvature, and read structural fragility before it hits prices.

T-06
Applied3 parts · ~2h

Case Study: Banking Network Fragility 2006–2008

Apply geometric economics to pre-crisis banking data. Reconstruct the signals visible in network topology months before the collapse.

T-07
Applied4 parts · ~3h

Building an AI-Augmented Research Workflow

Connect MCP tools, structure a knowledge base, design retrieval pipelines, and build self-improving analytical loops.

λ

Method over opinion

Every tool and tutorial is grounded in formalised, testable methodology. We teach the reasoning — not the conclusion.

Collaborative

Members contribute analyses, discuss methods, and refine models together. Peer review, not guru culture.

Inclusive

Access to analytical methods should not be a privilege. No one is excluded by income.

Interactive learning tools — you provide the inputs, you interpret the outputs.
Valuation

Probabilistic DCF Engine

Monte Carlo-simulated valuations with full uncertainty quantification. Input your own assumptions, see the distribution they imply.

Indicators

Context-Aware RSI

RSI extended with global liquidity data (M2) and news sentiment — an indicator that understands context. Learn to read it, not follow it.

Network Analysis

Geometric Economics

Ricci curvature on financial networks: detect structural fragility in correlation topology before it shows up in prices.

Collaboration

Peer-Reviewed Analyses

Community-contributed method discussions, shared models, and a growing knowledge base under open license.

MacroPulze connects directly to Claude via MCP Server — turning your AI assistant into an analytical partner.
MCP

MacroPulze MCP Server

Model Context Protocol · Claude AI
Proactive

Change Notifications

Get notified when indicators shift, when network curvature crosses thresholds, or when new tutorials match your path.

On Demand

Method Queries

Ask Claude to explain any MacroPulze method, run a historical comparison, or walk through a tutorial step.

Workflow

Knowledge Base

Claude reads your notes, saved analyses, and bookmarked research. Build a personal analytical knowledge base.

Pipeline

Data & Agent Integration

Connect external data sources, trigger simulations, and chain analytical steps through natural conversation.

$ macropulze --connect claude.ai
# MCP server registered. 4 tools available.
$ macropulze --notify rsi_context_shift geometric_curvature new_tutorials
# Proactive notifications enabled.
$ macropulze --kb sync
# Knowledge base connected. 23 documents indexed.
"Science is a way of trying not to fool yourself — and you are the easiest person to fool."
— Richard Feynman
MacroPulze is an educational platform for quantitative methods. All content is provided for informational and educational purposes only and does not constitute investment advice, personal recommendations, or solicitation to buy or sell financial instruments. Tutorials use historical data to illustrate methodology. Users are solely responsible for their own investment decisions.