> ## Documentation Index
> Fetch the complete documentation index at: https://docs.beliefsystems.xyz/llms.txt
> Use this file to discover all available pages before exploring further.
# What is Belief Index?
> A family of rules-based indices that measure market-implied probabilities of real-world events – benchmarks of event risk, computed from prediction market data.
Belief Index is a family of **rules-based indices** published by Belief Systems. Each index – called a **series** – tracks a themed basket of prediction market contracts and condenses them into a single published number: a benchmark of how markets are pricing a coherent category of real-world event risk, such as the trajectory of U.S. monetary policy or the outcome of a national election.
If you have used the S\&P 500 to read the equity market, or the VIX to read volatility, the concept is familiar: rather than inspecting hundreds of individual instruments, you read one rules-based number whose construction is published, systematic, and independently verifiable.
This documentation covers the full production methodology: market selection, weighting, NAV
computation, index levels, reconstitution, and known limitations. Belief Systems publishes index
data and analytics to the general public; separately, a small number of invited participants hold
interests linked to certain series through a private, invitation-only **Alpha Program**
(documented in its own section). Nothing in this documentation is an offer to sell or a
solicitation of an offer to buy any security. See [Disclosures](/risk/disclosures).
## The Measurement Problem
Prediction markets are emerging as one of the most efficient mechanisms for aggregating beliefs about future events – elections, economic data releases, policy decisions, and more. But as a *data source*, they present the same problem equity markets did before the first stock indices: hundreds of individual instruments and no summary statistic.
A single venue hosts hundreds of individual markets, each with its own order book, liquidity profile, and resolution terms. The aggregate signal – "how do markets price rate-cut risk?" – is scattered across dozens of related contracts with no principled way to read them together.
There is no single number that captures a thematic question – such as "how is the market pricing
the 2026 easing cycle?" or "what are aggregate Republican electoral prospects for 2028?" – across
multiple related events at once. Every market must be found, read, and mentally weighted by hand.
Individual prediction markets settle at exactly \$1 or \$0. Read in isolation, a single market is
a noisy, discontinuous series that can jump to zero overnight on one surprise. A diversified
basket turns a set of binary readings into a continuous, interpretable signal.
Each market has its own resolution date, settlement rules, and oracle. Tracking a theme over time means constantly replacing resolved contracts with live ones – and doing so without introducing artificial jumps into the series. This is precisely the problem index reconstitution and chain-linking solve.
## What a Belief Index Does
Belief Index applies the core disciplines of index construction – published selection criteria, systematic weighting, transparent valuation, and rules-based maintenance – to prediction markets.
Each series bundles 5-20 related prediction markets around a coherent theme, selected according
to [published criteria](/indices/selection-and-weighting) including liquidity, resolution
clarity, and thematic relevance.
Constituent weights follow a published formula with concentration caps. Where judgment enters –
theme definition, [judgment screens](/indices/selection-and-weighting#judgment-screens) – it is
exercised once, at composition time, under published criteria; it is never applied continuously
or invisibly.
The basket is valued at periodic NAV windows – currently every 30 minutes, subject to change –
from observable order book midprices, using a [published methodology](/indices/nav-methodology).
Every formula is disclosed; every computation is independently verifiable. No black boxes.
Each series publishes an **index level**, rebased to 100 at inception, that moves proportionally
with the underlying basket. Resolved markets fold their settlement values into the level;
reconstitutions are chain-linked so composition changes never create artificial jumps.
## Series Examples
Each series tracks a coherent theme, so the index level answers a category-level question rather than an event-level one:
| Series | Theme | Typical Markets |
| ----------------------------------------------------------- | ------------------------ | ------------------------------------------------------------ |
| **U.S. Presidential Election Republican Expectations 2028** | Election outcomes | Candidate victories, party-level results, primary outcomes |
| **U.S. Monetary Policy Dovish Stance Expectations** | Interest rate trajectory | Fed rate decisions, inflation targets, employment thresholds |
| **Global Conflict Risk Escalation Expectations** | Geopolitical risk | Escalation events, military actions, ceasefire outcomes |
An election-focused series answers "what are aggregate Republican electoral prospects for 2028, right now?" – one number, computed by published rules, comparable across time, for a question no listed instrument prices at all. Themes like elections, policy decisions, and geopolitical events have no futures curve or swap market; a Belief Index series is often the only continuous, citable measurement of them. See [Where the Signal Is Irreplaceable](/why-belief-index#where-the-signal-is-irreplaceable).
## Reading an Index vs. Reading Individual Markets
| Feature | Individual Prediction Markets | Belief Index |
| --------------------------------- | ----------------------------------------- | ---------------------------------------------------------------------------- |
| **Signal** | One binary event at a time | Themed aggregate of 5-20 markets |
| **Jump noise** | Discontinuous (\$0 or \$1) | Smoothed across multiple events |
| **Monitoring burden** | Track each order book and resolution date | One published series per theme |
| **Valuation** | Check each order book yourself | Published level at regular windows |
| **Verification** | Trust the venue | Published methodology, on-chain constituent holdings, replicable computation |
| **Continuity across resolutions** | Series ends when the market settles | Reconstitution with chain-linked continuity |
## The Order of Operations
In equities, the index came decades before the index fund: Dow Jones began publishing averages in 1896; the first index fund arrived in the 1970s. The benchmark is what makes everything downstream possible – research, journalism, risk measurement, and eventually investable products built by issuers who license the index.
The sequence matters for a second reason: a benchmark's authority accrues with its track record, and a track record cannot be created retroactively. Prediction market order books are not archived the way equity prices are – every valuation window published today is a point-in-time data point that no one, including us, could reconstruct later. See [A History That Cannot Be Rebuilt](/why-belief-index#a-history-that-cannot-be-rebuilt).
Belief Systems is built in that order. The indices are published, informational, and free to cite. Issuers who want to reference a series as the underlying for a product can do so under license through [Index Services](https://beliefsystems.xyz/institutional/license), and a small invitation-only Alpha Program operates privately under the conditions described in [Disclosures](/risk/disclosures).
Belief Index levels are informational benchmarks computed from observable market data –
theoretical values, not execution prices. See [Risk Factors](/risk/risk-factors) for methodology
limitations.
## Key Principles
* **Transparent** – The valuation methodology is [fully published](/indices/nav-methodology). Anyone can replicate any published level using public market data and a spreadsheet.
* **Systematic** – Index composition and weighting follow published, rules-based methodology. Where human judgment enters – theme definition and the documented [judgment screens](/indices/selection-and-weighting#judgment-screens) – it is exercised once, in public, under published criteria, not continuously and invisibly. That is what makes the resulting number citable.
* **Theme-driven** – Each series tracks a coherent risk domain, so the index level reads as an answer to a specific macro or event-driven question.
* **Point-in-time** – Levels are recorded as of each valuation window and are never silently restated; methodology revisions are dated and prospective-only. The archive is built live, window by window – see [Benchmark Governance](/trust/benchmark-governance).
* **Investable-grade discipline** – Valuations are computed with forward-pricing windows, double-entry fund accounting, and continuous reconciliation against on-chain holdings: the operational standard an investable product would require, applied to a published index from day one.
## Getting Started
How an index is produced, from market data to published level.
The case for rules-based benchmarks of event risk.
How themed baskets are constructed, weighted, and maintained.
The transparent, verifiable methodology behind every published level.