AI Council — Agent 5 of 8
Setup Surgeon
Setup edge & decay
Aggregate PnL can mask dying setups. The Setup Surgeon ranks every Playbook setup, flags edge decay, and names which strategies to scale, optimize, or kill.
Operates at the setup level — where edge actually lives or dies. Reads pre-computed per-setup metrics: win rate, profit factor, expectancy, Sharpe/Sortino, parameter drift (recent vs. older), compliance_summary from setup_validity, quant_regime_dependency, volatility regime splits, portfolio allocation, dead/star setup summaries, and What-If optimal SL/TP when available.
What the Setup Surgeon actually outputs
Not a chat summary — a structured forensic report from your canonical trade tape, with grades, benchmarks, and prioritized recommendations.
Sample Setup Surgeon report
Illustrative output from a populated audit — same structure as the in-app agent renderer.
Setup Surgeon's Assessment
Five setups analyzed: two stars, one dead, two maintenance candidates. Pullback Long shows regime split — strong in trend, negative in chop with 71% non-compliance. Headline PnL still positive because ORB Long carries 58% of dollar edge.
Benchmark analysis
What you're doing well
Star setup: PF 2.15, 45 trades, stable parameter drift — scale candidate.
Positive expectancy across trend regimes; 82% setup_validity compliance.
Areas needing attention
71% of chop-regime trades non-compliant — taking invalid instances destroys edge.
Dead setup: 14 trades, negative expectancy — kill candidate with ~$340/mo savings.
Recommended actions
Kill Mean Reversion Short — negative expectancy over 14 trades; reallocate trade count to ORB Long.
Expected impact: Stops ~$340/mo bleed; concentrates edge on top performer.
Trade Pullback Long only in trend regimes — skip chop until compliance recovers above 80%.
Expected impact: Recovers chop-regime −0.4R drag without abandoning the setup entirely.
What it reads from your tape
The Setup Surgeon receives pre-aggregated metrics from your completed trades — not a generic chat summary. Every input comes from your canonical trade tape.
- →Per-setup win rate, profit factor, expectancy, and total P&L
- →Recent vs. older performance splits and parameter drift classification
- →Compliance rates and invalid-setup R from setup_validity tags
- →Quant regime dependency (trend/chop/vol) and volatility regime splits
- →Portfolio allocation — trade-count % and P&L contribution % per setup
What it detects
- Parameter drift — declining expectancy or win rate over time
- Dead setups (10+ trades, negative expectancy or PF < 1)
- Star setups (10+ trades, PF ≥ 1.5, positive expectancy)
- Regime-sensitive setups — edge that only works in specific conditions
- Compliance degradation inflating loss rate on otherwise valid patterns
What it delivers
- Setup assessment with profitable/breakeven/dead counts
- Ranked setup table with SCALE_UP / MAINTAIN / REDUCE / KILL recommendations
- Dead setup list with reason_to_kill and potential monthly savings
- Parameter drift alerts, regime analysis, and portfolio allocation advice
Find which setups carry your edge — and which ones are bleeding you dry
Run the Setup Surgeon on your completed trades. Inspect the full AI Council free on a populated demo account — read-only, no credit card.
Example finding
“Pullback Long: +0.6R in trend, −0.4R in chop — 71% of chop trades non-compliant; parameter drift declining over last 18 trades.”
How it fits the Council
The Setup Surgeon runs in parallel with six other specialists on the same audit period. Each agent sees metrics tuned to its domain. When specialists disagree, they debate transparently in The Situation Room before the Chief Coaching Officer synthesizes everything into a dollar-ranked Kill List.
- 1Seven specialists analyze pre-computed metrics in parallel
- 2Disagreements surface in The Situation Room debate transcript
- 3Chief Coaching Officer ranks preventable leakage into your Kill List ($/month)
Meet the other agents
Performance Analyst
Profitability & statistical edge
Learn more Agent 2Behavioral Psychologist
Discipline & emotional leakage
Learn more Agent 3Execution Tactician
Timing & fill quality
Learn more Agent 4Risk Assassin
Capital preservation
Learn more Agent 6Regime Cartographer
Market environment fit
Learn more Agent 7Entry & Exit Judge
Opportunity capture forensics
Learn more Agent 8Chief Coaching Officer
Synthesis & prioritization
Learn moreFrequently asked questions
Frequently asked questions
What is setup edge decay?
Edge decay occurs when a setup's expectancy or win rate declines over time — often because market structure changed, parameters drifted, or you started taking lower-quality instances. The Setup Surgeon detects this via parameter_drift (recent vs. older splits) before aggregate PnL turns negative.
How does Setup DNA relate to this agent?
Setup DNA is the in-app tab that visualizes per-setup metrics from your trade tape. The Setup Surgeon is the AI agent that interprets those pre-computed metrics — drift classification, dead/star lists, compliance rates, and regime splits — and outputs ranked recommendations.
What compliance data does it use?
When you tag trades with setup_validity (valid vs. invalid instance), the agent computes compliance_summary per setup: valid %, invalid count, and invalid avg R. Quant regime blocks also show invalid_setup_pct per trend/chop/vol tag — so you see whether losses come from bad instances or bad environments.
How is this different from the Regime Cartographer?
The Setup Surgeon ranks and kills setups — expectancy, drift, compliance, and allocation at the strategy level. The Regime Cartographer maps market environment (ATR percentiles, session behavior, volatility clusters) to explain when and why setups stop working across the portfolio. Setup Surgeon cites quant_regime_dependency per setup; Regime Cartographer goes deeper on environment attribution.
How many trades do I need per setup?
Basic metrics appear from the first trade. Dead-setup classification requires 10+ completed trades with negative expectancy or PF < 1. Star setup classification needs 10+ trades with PF ≥ 1.5 and positive expectancy. Parameter drift needs 5+ trades in each recent/older half.
See Setup Surgeon on your trade history
Inspect the full AI Council on a populated account — every agent visible, read-only, no credit card.