How to use this page
Read the extract first, then the application and limits sections, and only then decide whether the thesis is strong enough for action or only for context.
Extractable overview
What EdgeVisor does: it turns raw market pricing into a structured thesis: crowd price, model estimate, evidence stack, caveats, and action framing.
What changes by category: politics, crypto, and macro can attach fresh monitored news when the mapping and timing are valid. Sports usually relies more on market structure than on external citations.
What the user should infer: an EdgeVisor pick is not a command. It is a decision surface that shows where the disagreement is, what supports it, and where it can fail.
Pipeline stages
The public output starts long before the UI. The ingestion loop refreshes active markets and enrichments, then the model stack produces category-aware signals from market structure, baseline context, timing, and outside evidence when that evidence is available.
Those signals feed a research summary that separates supporting evidence, counter-evidence, citations, source links, and confidence notes. The decision layer then decides whether the setup is tradable, informational, or too thin to surface.
| Layer | What it adds | Why the user should care |
|---|---|---|
| Market ingestion | Live price, liquidity, category, timing, partner comparisons | Defines the starting consensus and whether the market is even worth evaluating |
| Analyst stack | Baseline context, timing, liquidity context, and behavioral market checks | Shows whether the current price looks crowded, stale, thin, or structurally wrong |
| Research summary | Supporting evidence, counter-evidence, citations, confidence notes | Turns a number into an auditable thesis instead of a black-box estimate |
| Decision layer | Tradable vs informational framing, readiness, caveats | Tells you whether the setup looks actionable or only useful for context |
What the user sees
The user-facing explanation includes price, estimate, edge label, recommendation text, evidence bullets, confidence notes, and sometimes citations. Those fields are there so a user can inspect why a pick exists and where the thesis is fragile.
A positive edge does not automatically mean "buy now." It means the model stack sees a meaningful disagreement with the crowd price. The next question is whether that disagreement is supported by aligned evidence, acceptable caveats, and a category where the product actually has signal depth.
- Read the gap first: start with market price vs estimate, because that defines the thesis you are even evaluating.
- Audit the support: check whether the thesis is backed by evidence, external citations, market-structure signals, or only a thin estimate gap.
- Read the caveat before action: if confidence notes or category limits weaken the setup, treat it as informational even if the headline gap looks large.
In politics especially, EdgeVisor can surface an informational thesis even when the final tradable edge is weaker. That is why some picks should be treated more like research context than a direct trading prompt.
Limits
EdgeVisor is strongest when the market has enough liquidity, the category logic is well mapped, and the system has either strong structure signals or fresh external evidence. It is weaker when a category has thin context, sparse external support, or highly efficient crowd pricing.
- Current market sources: live Polymarket pricing and market structure data, plus partner comparisons such as Preddy when available.
- Current evidence sources: monitored RSS buckets for selected categories, such as White House, Federal Reserve, and BLS, only when the market mapping and freshness window support them.
- Honesty rule: if the live product does not produce a given evidence layer, the docs should not imply that it does.
Frequently asked questions
No. External news citations depend on category mapping and fresh RSS items. Some categories are mostly market-structure driven.
Not exactly. The explanation layer surfaces context and evidence for humans, while the decision engine applies gating, tiers, and risk constraints.