NeuroTrader - The Construct™Docs
Open App

Core Concepts

Reading the Metrics

The platform separates four different kinds of readings: market & model conditions, platform health, execution readiness, and your account risk. This page defines every number and where it comes from.

The single most important distinction

Condition metrics (Signal Quality, Avg Confidence, Trading Conditions, Market Field, Coordination Health) describe the market and the models' conviction — they rise and fall with the market. Platform Status is the only tile that describes the service itself. "Weak" conditions with Platform Status "Online" means a healthy platform in a quiet market — the AI is deliberately waiting.

The four categories

CategoryQuestion it answersWhere it shows
Platform / service healthAre NeuroTrader services online and processing?Platform Status tile, System Events feed
Model / market confidenceHow much conviction do the engines have right now?Signal Quality, Avg Confidence, Model Health, Coordination Health, Market Field
Execution readinessHow favourable are conditions for trading?Trading Conditions score, Pre-Flight Review
User / account riskWhat limits protect my account?Settings → Risk Controls, Decisions page reasons

System page tiles

Signal QualityThe average signal strength across all AI engines, live. Good ≥ 85%, Fair ≥ 65%, below that Weak. Weak means the engines see little conviction in the market — they are online and analyzing the whole time.
Models ActiveHow many engines are currently online and reporting telemetry, and on how many assets.
Avg ConfidenceThe mean prediction confidence across the engines. Low values in choppy markets are normal and protective: trades only fire above your configured threshold.
Platform StatusWhether the AI service is online and processing cycles, with the running cycle count. This is the actual service-health reading — independent of market conditions.

The Trading Conditions score

A composite 0–100 score of how favourable the environment is for trading:

ComponentWeightSource
Model consensus35%How strongly the voting engines agree on direction
Signal quality35%Average engine signal strength
Your win rate30%Your account’s closed-trade history

Why fresh accounts start low

A new account has no closed trades, so the win-rate component contributes zero — the score is mathematically capped at 70 even in a perfect market. It builds as your history accumulates. A low score is the system being honest, not broken.

Beneath the score, three sub-readings:

ConsensusStrong / Med / Weak — engine agreement on direction.
StabilityHigh / Moderate / Low — derived from signal quality.
CautionLow / Raised / High — how careful the AI is being, driven by consensus. High caution means the AI is holding back; it is not a statement about your account being in danger.

Coordination Health

AlignedNo conflicts and strong consensus — the engines agree.
Low ConsensusNo open conflict, but conviction is weak. The AI stands down until agreement improves. The most common state in quiet markets.
ResolvingEngines actively disagree on an asset; the conflict is being evaluated.
AwaitingNo telemetry yet (e.g. right after a service start).

Market Field

The Coherence Model's view of the market as a whole: Global Coherence (how aligned price action is across assets), Field Strength, and a Condition label — Aligned, Fragmented, or Chaotic. "Chaotic" describes the market itself; the AI deliberately trades less in chaotic fields.

Per-engine Model Health

Each engine's own signal strength, confidence, and win rate. An engine marked Fair or Weak is online and analyzing — it simply has low conviction in current conditions. An engine that is actually offline is a platform issue and shows in Platform Status instead.

Portfolio metrics

Portfolio ValueCash plus the market value of open positions, live.
Total P&LRealized P&L across all closed trades plus unrealized P&L on open positions — all-time for the account.
Win RateWinning closed trades ÷ all decided closed trades.
30D Change / ReturnChange in end-of-day equity over the last 30 days, from hourly snapshots.
Max DrawdownThe largest peak-to-trough decline within the charted window.