AI QUORUM TERMINAL

Agent Leaderboard & Brier Score Calibration

Specialist research models generate structured signals evaluated by statistical accuracy (Brier Scores: 0.0 = perfect, 1.0 = worst).

Onchain Flow Agent

0.128

35% Ensemble Weight

Quant Regime Agent

0.151

25% Ensemble Weight

News Impact Agent

0.164

20% Ensemble Weight

Macro Tape Agent

0.186

20% Ensemble Weight

Agent Ensemble Weight Sandbox

Adjust individual model weights to simulate how agent ensemble consensus recalculates under custom research priorities.

Onchain Flow Model (72%)35% Weight
Quant Regime Model (63%)25% Weight
News Impact Model (66%)20% Weight
Macro Tape Model (53%)20% Weight
Recalculated Consensus

64.8%

Ensemble Brier Score

0.153

Onchain flow
BULLISH
Forecast Probability

72%

Brier Score

0.128

Validator inflows and DEX routing depth improved for three consecutive sessions.

Verified Evidence Stream

  • Stake-weighted liquidity +11%
  • Priority fee pressure easing
  • Jupiter route depth stable
Confidence: 84%Strength: +82pt
Quant regime
BULLISH
Forecast Probability

63%

Brier Score

0.151

Momentum remains constructive, but realized vol is still outside comfort zone.

Verified Evidence Stream

  • Trend z-score 1.8
  • Vol percentile 68
  • Correlation cluster elevated
Confidence: 81%Strength: +71pt
News impact
BULLISH
Forecast Probability

66%

Brier Score

0.164

Headline velocity favors expansion, with ETF flows carrying most of the lift.

Verified Evidence Stream

  • ETF inflow streak
  • Protocol outage risk low
  • Policy attention moderate
Confidence: 78%Strength: +76pt
Macro tape
NEUTRAL
Forecast Probability

53%

Brier Score

0.186

Rates and dollar pressure are not hostile, but not supportive enough to chase aggressively.

Verified Evidence Stream

  • DXY flat
  • Risk assets mixed
  • Liquidity impulse neutral
Confidence: 69%Strength: +49pt