AI Market Intelligence

Model forecasts, verified evidence, and runtime context translated into structured market briefs.

⌁

Market Pulse

Primary Focus

BTCUSDT

Bias

LONG

Model Agreement

100%

High

Expected Move (H=1)

+0.88%

Evidence Quality

Medium

Reason: direction accuracy 50.4%

Runtime Status

Monitoring

No open exposure

Market at a Glance

2 market tickers synced

₿

BTCUSDT — AI Market Brief

Timeframe: H=1 (1 Day)Last Update: Sep 19, 2026

A. Current Model View

Direction / BiasLONG

Expected Move+0.88%

Forecast Close$81,593.57

Reference Close$80,883.86

IssuedSep 18, 2026

TargetSep 19, 2026

Model Agreement100%

Active ModelNova-II (in Fusion-3N) · Astra-II (in Fusion-4) · Vega-II (in Fusion-3A) · Nova-II (in Fusion-3C) · Orion-II (in Fusion-3B)

Forecast vs Reference

B. Historical Reliability (BTCUSDT H=1)

Verified Forecast Rows3,240

Verified Days180

Models in Pool18

Avg Error (MAPE)1.66%

Median Error1.25%

Direction Accuracy50.40%

Touch Rate (Hit)69.85%

Recent MAPE2.32%

Days with Avg Error <1%50.0%

Forecast Rows <1% Error54.2%

Summary: Evidence 180d. Charts: overview recent rows (144 rows). Daily sub-1% uses average day error; row sub-1% uses individual forecast rows.

Direction Accuracy Trend

Hover points to see direction accuracy %.

Error Distribution

54.2%

<1%

16.7%

1–2%

3.5%

2–3%

14.6%

3–5%

11.1%

>5%

C. Scenario Map

Base Scenario (60%)

Models keep a LONG bias while price behavior remains consistent with the reference area near $80,883.86. Target context: $81,593.57.

Upside Extension (25%)

A stronger upside extension is supported only if agreement stays firm and recent error does not expand materially.

Downside Invalidation Risk (15%)

The scenario weakens if price rejects the forecast path, volatility expands sharply, or evidence quality deteriorates against the downside path.

D. Risk Context

⚡Volatility RiskNormal

▣News / Event RiskMedium

⇄Spread RiskLow

⌁Model AgreementHigh

↯Historical Acc. (Dir.)Moderate

〽Recent Model TrendDeteriorating

Overall Risk: MEDIUM

E. Analyst Note

BTCUSDT shows a LONG model bias for the next daily target, with expected move +0.88% and model agreement is 100%. Historical evidence quality is medium, so the useful value is in combining current direction with verified error history, volatility context, and runtime filters.

This is research context, not a trade signal.

What Changed Since Last Update

↑

Bias Change

SHORT → LONG

↑

Expected Move

-0.55% → +0.88%

↑

Model Agreement

current 100%

Recent MAPE

6.04% → 2.32%

Runtime Status

Monitoring → Monitoring

Model Reliability Snapshot (BTCUSDT H+1)

Model	Direction Acc.	Avg Error	Recent MAPE	Touch Rate
Fusion-4	55.56%	1.41%	0.96%	80.00%
Nova-II · in Fusion-3C	55.56%	1.44%	1.02%	78.89%
Orion-II · standalone	47.78%	1.40%	1.01%	93.33%
Fusion-3C	54.44%	1.45%	1.01%	76.67%
Orion-II · in Fusion-4	51.11%	1.45%	1.04%	82.22%

View full evidence →

Weekly AI Research Brief

Computed from public Evidence summary and 90d model leaderboard. It does not mean an asset is a trading signal.

ⓘ

Lowest Avg Error

BTCUSDT · 1.66% MAPE

Highest Avg Error

ETHUSDT · 2.16% MAPE

◎

Best Direction Accuracy

ETHUSDT · 51.14% direction

⇅

Weakest Direction Accuracy

BTCUSDT · 50.40% direction

≋

Most Stable Model

Orion-II · BTCUSDT · 1.40% MAPE

⌁

Where Runtime Filtered Out

ETHUSDT

⌬

Where Models Disagreed

No major disagreement in public forecast pool

◈

Strongest Agreement

BTCUSDT · 100%

▣

Evidence Note

Lowest MAPE is not the same as strongest trading signal; direction accuracy, touch rate, runtime state, and risk must be read separately.

View full weekly report →

What ProsQuant is not

Research output, not an investment instruction.

Not financial adviceNot a black-box signal sellerNot a profit guaranteeNot a broker or execution venue

Research only. Not financial advice. Not investment advice. Past performance is not indicative of future results.