HOW HALO WORKS

See the market through its best traders.

Halo turns scattered public prediction-market activity into a clear intelligence layer: who has an edge, where conviction is building, and how to follow it within precise risk limits.

PUBLIC ACTIVITYTRADER EVIDENCEHALO INTELLIGENCE
THE SIMPLE VERSION

Markets reveal probability.
Behavior reveals conviction.

A market price tells you what the crowd believes now. Halo studies the behavior behind that price—the traders, entries, timing, sizing, and agreement patterns that can make a move worth understanding.

Observe

Read public market and trader activity.

Evaluate

Measure evidence, consistency, and risk.

Detect

Find independent high-quality agreement.

Control

Apply explicit portfolio guardrails.

01 / DATA ENGINE

One market.
A richer picture.

Halo normalizes official Polymarket market metadata, public account activity, and order-book pricing into a common event model. That makes evidence comparable across markets without pretending every market behaves the same way.

01Sources

Gamma metadata
Data API activity
CLOB pricing

02Normalize

Markets, outcomes,
positions, fills,
timestamps

03Halo layer

Scores, signals,
context, and
risk decisions

Freshness is part of the data. Halo presents source and update context alongside live observations. If a value is unavailable, the product does not invent a replacement.

02 / HALO SCORE

A compact view
of observable edge.

94

The Halo Score is a 0–100 evidence-weighted ranking designed for comparison—not a promise of future returns. It rewards repeatable performance, useful history, and risk-aware behavior while reducing confidence when the evidence is thin.

ComponentWeightWhat it asks
Profitability24%Realized outcomes and return profile
Consistency19%Repeatability across markets and time
Drawdown control18%Loss depth, recovery, and downside behavior
Experience13%Observable activity and resolved-market history
Closing performance11%Entry quality relative to later market pricing
Account + breadth15%History, diversification, and evidence quality

Weights describe the current scoring framework and can evolve as the methodology is validated. Market context and evidence depth still matter.

03 / CONVERGENCE SIGNALS

When strong,
independent views align.

A Convergence Signal forms when multiple qualifying traders independently build exposure to the same outcome. Halo de-duplicates wallets, compares timing, aggregates visible exposure, and surfaces the evidence as a readable signal.

HIGH CONVICTION

Agreement is useful only when you can inspect it.

Every signal is paired with its direction, participating evidence, observed timing, market price, and confidence context.

LIVE EVIDENCE UPDATED 18s AGO

5 qualifying traders

aligned on the same outcome

  • DirectionYES
  • Market price64¢
  • ConfidenceHIGH
Independent

Repeated or linked observations should not inflate a signal.

Time-aware

Fresh, clustered entries carry different meaning from stale positions.

Explainable

The underlying observations remain visible beside the conclusion.

04 / COPY ENGINE

From source trade
to bounded decision.

Halo's portfolio experience models what strategy-following could look like under a user's own limits. It is a policy engine, not a blind multiplier.

  1. 01
    Read the source event

    Direction, entry, size, timing, and current market state.

  2. 02
    Run policy checks

    Eligibility, score threshold, liquidity, concentration, and loss limits.

  3. 03
    Calculate target size

    Scale exposure against portfolio capital and copy percentage.

  4. 04
    Record the decision

    Show copied, reduced, skipped, or paused—with a reason.

05 / RISK CONTROLS

The limit is part
of the strategy.

01Capital allocation
02Max position size
03Copy percentage
04Daily loss limit
05Minimum Halo Score
06Drawdown pause
07Exit mirroring
08Portfolio kill switch

More activity is not the goal.

Better-filtered activity is.

06 / TRUST LAYER

Inspectable
by default.

Halo separates observed data from calculated intelligence. The interface is designed to preserve provenance, timestamps, calculation context, and the difference between live observations and paper-portfolio modeling.

signal.json
{
  "source": "public_market_activity",
  "observed_at": "2026-08-16T…Z",
  "participants": 5,
  "direction": "YES",
  "confidence": "high",
  "mode": "observed"
}
OPEN ON GITHUB

Read the public engineering overview.

Architecture, methodology, data boundaries, and risk principles.

FAQ

Good questions
are part of the edge.

Does Halo use live prediction-market data?

Yes. Halo's live surfaces use official public Polymarket endpoints for market metadata, public trader activity, and pricing. Source availability and update timing can vary.

Is the portfolio real-money execution?

The current portfolio experience is a paper simulation for exploring sizing and risk policies. It is labeled as paper mode and does not place trades or hold funds.

Does a high Halo Score guarantee performance?

No. It summarizes observable historical evidence. Prediction markets remain uncertain, and past performance cannot guarantee future results.

Why use multiple independent traders?

Agreement can be more informative when it comes from genuinely separate decision-makers. Halo uses independence, timing, and evidence quality to reduce noisy consensus.

THE INTELLIGENCE LAYER

Find the signal.
Understand the evidence.

Explore Halo GitHub X