Q
Quantara
14ms

Learning Center & Agent Leaderboard

Calibrated Reliability • Expectancy Attribution • 9-Month Failure Taxonomy Analysis

Specialist Agent Leaderboard

Ranked by calibrated accuracy and net expectancy contribution (avoiding naive win rate ranking)

29 Evaluated Agents
AgentCalibrated AccExpectancy ContribWarning ValueOptimal RegimeDecisions
1Critic Agent
88.4%+₹84.20 / trd94.2%RANGE / CHOP142
2Theta Risk Agent
86.1%+₹62.10 / trd91.8%ALL OPTIONS98
3Price Action Agent
79.5%+₹54.40 / trd78.5%TREND_BULLISH180
4Options Liquidity Agent
84.0%+₹41.00 / trd89.0%HIGH_VOL98
5Technical Agent
72.3%+₹38.50 / trd71.2%TREND_CLEAN195
6Volume / RVOL Agent
75.8%+₹33.20 / trd80.1%BREAKOUT164
7Bear Agent
78.2%+₹29.80 / trd85.4%DISTRIBUTION120
8Bull Agent
71.0%+₹21.40 / trd64.0%EXPANSION155

Failure Taxonomy Recurrence

Most frequent root causes across 31 historical losses:

F13: Theta decay (Options)12 occurrences (38%)
F02: Early entry into chop7 occurrences (22%)
F09: Excessive spread/slippage5 occurrences (16%)
F04: False breakout reversal4 occurrences (13%)
F22: Ignored Critic conflict3 occurrences (11%)

Active Learning Hypotheses

Bounded parameter iterations proposed from autopsy data:

HYP-01: Theta Risk Weight ElevationACTIVE IN CHALLENGER

Increase Theta Agent weight to 0.20 to mitigate F13 option decay in late sessions.

HYP-02: Spread Tolerance TighteningSHADOW TESTING

Cap marketable option spread at 0.5% (down from 0.8%) to cut F09 slippage drag.

HYP-03: ORB Volume FilterOOS VALIDATION

Require 2.0x RVOL rather than 1.8x on morning equity breakouts to avoid F04 fakeouts.