⚡ AI WMFetching…
Six books
Six paper portfolios on the same nightly agent output, ₹10 lakh notional each, identical fills and costs, judged against the NIFTY 500 and the equal-weight universe at 30, 60 and 90 decision dates. Expectation written down before the first trade: Book 4 first, Book 3 second, Book 5 ahead of Book 1; Books 2 and 6 open. Kill: 10 points behind the null on a rolling 60 dates, twice. Promote: the best at 90 dates if ahead of both benchmarks. Overlay: target exposure 1.0 above / 0.5 below the equal-weight large-cap proxy's 200-day average, confirmed over 5 sessions.
Loading snapshot…
Book 1Engine as it isthe control
The production engine exactly as it runs: four agents per name, trader proposal, risk and tax checks, the entry gate. Prompts frozen. The control everything is measured against.
Book 2AI-first, rebuiltopen question
A properly briefed AI picks well: price-state facts with their historical meaning, a timing analyst instead of a technical rating, scorecard feedback, five guardrails, sizing by pillar agreement.
Book 3Fundamental grades → rulesexpected second
The one signal with measured forward information, gated by the one number the grade does not encode: eligible when the fundamental view says growth strong and risk not high AND the newest results filing shows revenue growth not decelerating versus the prior period; entries at least five sessions after the filing, equal weight, at most 30 names; exits only on a grade deterioration or a bearish sentiment event.
Book 4Fundamental ∩ momentumexpected first
Our two independent signals together: Book 3's gate (strong growth, risk not high, revenue growth not decelerating), above the 200-day average, ranked by 12-1 momentum, 30 names rebalanced monthly, half exposure while the NIFTY 500 is below its 200-day average.
Book 5Technical pullbackexpected ahead of Book 1
Technicals the right way round: leaders (above the 200-day average, momentum top quintile) bought after a down week or a close below the 20-day average; held at least 10 sessions, then sold at the 20-day average or after 20 sessions, with a 10% stop. The inverse of what the technical analyst did. The replay says this is the weakest book after real costs.
Book 6Macro + sector rotationopen question (2 quarters)
Top-down adds to bottom-up: weekly sector tilts from the macro strategist (bounded, turnover-capped), Book 3's names within sectors, with a Brinson split of allocation versus selection and a scorecard for the sector calls themselves.