Direct Answer
The Sentiment Composite Explorer shows how a market sentiment composite is built from six evidence families, and lets you test how much the result depends on the weights, on any one family, and on missing or stale inputs. It shows each family's score, weight, freshness and contribution, reports completeness honestly instead of rescaling it away, and labels any custom weighting as a custom educational view. It runs on values you enter or on a labelled fictional sample, and it describes evidence, not a forecast.
Sentiment Composite Explorer: Weights, Sources and Sensitivity
Key Takeaways
- The published composite has six evidence families with fixed weights of 20, 20, 20, 15, 15, and 10 percent. The weights are set by methodology, not fitted to past returns.
- Correlated inputs live inside one family, so a new volatility measure cannot become an independent family under a new name.
- A missing family lowers completeness by its weight. The raw composite is never silently rescaled to look complete.
- Leave-one-family-out shows whether one family dominates the result. A wide spread means it does.
- Custom weights are a custom educational view, not the canonical Swoopr composite, and are labelled that way everywhere.
- Live feeds are not connected. The default numbers are a fictional illustrative sample.
Composite Explorer
Illustrative sample data. The values loaded below are fictional, with a fictional date of 2026-01-15, and describe no real market. Live feeds are not connected. For real readings, open the primary source for each family and enter your own normalized scores. Read the sentiment composite framework for how the families are chosen.
All calculation runs in your browser. Nothing you enter is sent to Swoopr Investment's servers, stored, or logged.
Methodology Drawer
Formulas, component weights, normalization, the stale rule, version history and the change log come straight from the published methodology the tool computes with, so this section cannot drift from the numbers above.
Open methodology, formulas and change log
How Is the Composite Calculated?
Each component is first normalized to a score from minus one (defensive evidence) to plus one (risk-seeking evidence) against its own history. Higher-is-defensive series, such as credit spreads, are oriented so a wide spread reads defensive. Rates and the dollar have no universal sign, so their scores are entered as supplied and the methodology never invents one.
A family score is the weighted sum of its component scores, with weights inside the family adding to 100 percent. The composite is the weighted sum of the available family scores. A family can never contribute more than its weight, because every component score is capped at one.
Freshness is judged against each series' own cadence. A weekly positioning report or a monthly margin statistic is not penalised for being older than a daily close. Only a missed release counts as stale, and the stale rule decides what happens next: exclude the family, reduce confidence, or keep it with an explicit degraded flag. Version 1.0.0 reduces confidence. See data latency and vintages for why reporting delays matter.
How to Read the Robustness Panel
- Leave-one-family-out range: the lowest and highest composite when each family is removed in turn. A narrow range means no single family decides the result.
- Largest family contribution: which family moves the raw composite most, as a share of the absolute contributions.
- Families pointing against the composite: families whose own score has the opposite sign to the composite by a material margin. Contradiction is information about disagreement, not an error.
- Stale or missing inputs: which families are absent, degraded, or only partly supplied.
- Custom weights versus published weights: in sensitivity mode, how far your weights move the answer on the same data.
Related reading on how families relate: cross-asset risk appetite, the Cross-Asset Risk Appetite Dashboard, and the source ladder that ranks the sources each input can come from.
What This Tool Does Not Do
- It does not fetch live data and does not show current market readings.
- It does not reproduce historical values or point-in-time vintages: there is no stored history on this page.
- It does not forecast, rank assets, or recommend any action. Zones describe evidence and imply no statistical precision.
- It cannot change which family a component belongs to, or the weights inside a family. Those are fixed by the published methodology version.
Frequently Asked Questions
What is a market sentiment composite?
A market sentiment composite combines several indicators into one number after putting them on a common scale. This explorer groups them into six evidence families (equity participation, credit, volatility, rates and dollar context, positioning, leverage), blends the components inside each family with fixed weights, then blends the families with published weights of 20, 20, 20, 15, 15, and 10 percent. Every step is visible, which is the point: a single gauge with no visible inputs cannot be checked.
Why are the family weights fixed in the published view?
The published weights are chosen by methodology, not fitted to past returns, and a published version never changes: a change becomes a new version with a change-log entry. Letting the view silently drift would make two readers see different numbers under one name. Sensitivity mode lets you change the weights to see how much the answer depends on them, and it labels the result as a custom educational view, not the canonical Swoopr composite.
What does completeness mean here?
Completeness is the share of family weight that is actually available. If the leverage family, worth 10 percent, is missing, completeness reads 90 percent. The raw composite is not rescaled to look complete. A second figure divides by completeness so the available families can be read on a minus one to plus one scale, and it is always shown next to the completeness it rests on.
Does the explorer use live market data?
No. Live feeds are not connected. The page computes from the values you enter and from a bundled sample that is fictional and labelled illustrative, with fictional dates. To check a real reading, open the primary source for each family and enter your own normalized scores.
Is the composite a forecast or a recommendation?
No. The zones describe where a continuous number falls: strong defensive evidence, moderately defensive, mixed or neutral, moderately risk-seeking, strong risk-seeking evidence. They describe the state of the evidence and carry no statistical precision, no forecast and no recommendation. Positioning and leverage in particular describe crowding and fragility, not the direction of the next move.
References
The family structure, weights and component list are Swoopr Investment's own published methodology, version 1.0.0, and the sample values are fictional. The one external source below describes how the positioning family's underlying futures data is reported and what its limits are.
- CFTC: Commitments of Traders Explanatory Notes: how the weekly futures positioning report is defined and what its categories mean.
Jurisdiction: United States. Last reviewed by the Swoopr Editorial Team in October 2026. This page is educational and is not personalized investment advice.