One forensic-first book of Indian small, micro and select mid caps. The accounts are read for trouble before anything else, every name carries a written thesis published before the trade, and positions are held for months to years — until the thesis breaks.
Risk of permanent capital loss. Small and micro cap equities are less liquid and more volatile than large caps. A thesis can be wrong, a position can become hard to exit at any sensible price, and you can lose part or all of the money you commit. Nothing on this page is advice or a recommendation to buy any security.
The process
Most research starts with the story and checks the accounts later. Core reverses it. A name is read for trouble before it is read for opportunity — and most names never make it past that first read.
Cash conversion against reported profit, promoter pledging, related-party flow, auditor changes, receivable and inventory build, surveillance flags. A single serious governance or liquidity finding vetoes the name outright, however good the business looks.
Business quality, management and governance, balance sheet and cash, valuation, catalyst and liquidity. Every pillar is scored against cited, fetched evidence — never recall. A number that cannot be sourced is written down as unavailable rather than guessed.
The written thesis — what has to be true, what size, and the pre-committed level that says the thesis has failed — is published before the order is placed. Not after the fill. Not after the price moves.
The thesis breaks, or the pre-committed invalidation level is hit. Not because the position is boring, not because something else looks exciting. Filings on a held name force an immediate thesis re-check; conviction is re-scored at least once a quarter.
Why the ordering matters
Publishing the thesis first removes the option to edit the reasoning after the price moves.
Every investor believes they had a reason. Almost nobody can prove what that reason was before the outcome was known. Hindsight quietly rewrites it — the winner becomes conviction, the loser becomes bad luck.
A timestamped thesis published ahead of the trade closes that door. If a name works, the published reason is the one that worked. If it fails, the published reason is the one that failed, and the pre-committed invalidation level is there in writing alongside it. That ordering is the product's central integrity claim, and it is the one thing that cannot be retrofitted.
The book, as it stands
What follows is what the book is actually made of, and how far it sits from its own model. It is not a return record and nothing here is a projection. Where the book is off-model, the gap is shown at full size.
Loading the live model…
Chart 1 of 3
Loading…
Chart 2 of 3
All holdings
| Ticker | Bucket | Weight | Conviction |
|---|---|---|---|
| Loading… | |||
Loading…
Chart 3 of 3
Conviction scores are not yet published for this model version — not scored, not hidden. This chart fills in automatically once they are.
The honest state of the book
Right now the model holds — cash against a — target. That is never a market call and never a designed allocation — it is what happens when the research funnel has not yet produced enough names that clear their entry gate. See the live book for the day-by-day figure and the reasoning behind today's number.
The mandate is explicit about this: cash above target has to be explained by a named binding constraint at every session — no qualifying idea cleared the gates, the book is already carrying its maximum open risk, liquidity fails, market conditions are hostile. Cash is never allowed to be residue.
The same rule covers any bucket sitting under target — see Chart 1 above for today's gaps. A bucket under target because good names are hard to source in that part of the market is a research-supply constraint, not a view on the sector, and closing it by simply buying to target would break the process the product is built on.
Idle capital is a cost too, and the system is biased toward acting on evidence, not toward sitting still. But the bar for acting is evidence, not activity. When nothing clears, nothing is bought, and the reason is written down.
What high cash means for you
The minimum
The model publishes weights. Your account has to express those weights in whole shares. The minimum investment is therefore derived, not chosen: the highest share price ÷ target weight across the model — the smallest corpus at which every line buys at least one whole share. It is recomputed with live prices on every publish. Above it you can track; for tight tracking, ₹5,00,000 is recommended — below that, whole-share rounding still distorts the smallest lines.
The arithmetic
The smallest position a ₹2,00,000 account can hold in whole shares carries a 1.36% target weight. At that account size, a single share of that company is roughly 0.68% of the account — so the position lands at about 0.7% against a 1.36% target. That is a 50% tracking error on that line, and you cannot fix it: you cannot buy half a share. Three more positions miss by 0.4 to 0.5 percentage points at that size.
By contrast, a ₹10,00,000 account tracks the model within 0.2 percentage points on every line. ₹5,00,000 is where the error stops being structural. Figures from the model portfolio, 17 Jul 2026.
The minimum is a technical floor on how faithfully an account can track the model. It is not a measure of who is welcome, and a larger account does not buy better research — every account tracks the identical published weights.
What can go wrong
These are not boilerplate. They are the failure modes that actually apply to a concentrated book of Indian small and micro caps, stated plainly.
Not a dip you wait out — capital that does not come back. A small company can lose its market, its promoter, or its accounting credibility, and the value does not return. Forensic screening lowers the odds of walking into this. It does not remove them. You can lose part or all of what you commit.
Small and micro cap liquidity is not constant — it evaporates in precisely the conditions that make you want to sell. A position sized comfortably against normal volume can take days to exit in a falling market, at prices well below the screen. Position size is gated on traded volume for this reason, and a name that fails the liquidity gate is passed on however good it looks.
A holding period measured in years means living through falls that would end a shorter strategy. The process is built to hold through price weakness while the thesis holds — which means the book is designed to sit in drawdown rather than sell out of it. If you would need this money back during one, this is the wrong place for it.
Eleven positions, with the largest at 8.18% of the book — 0.18pp above the 8% single-name cap, on price appreciation rather than on a decision to add. Drift above the cap is what forces the model weight to be cut; that is precisely what happened to ECLERX on 17 Jul. A concentrated book means individual mistakes are felt, not diluted. That is deliberate — dilution is also the enemy of conviction — but it makes single-name risk real. A single-name cap of 8% is the limit, not the aspiration.
Publishing reasoning first makes it accountable. It does not make it correct. Cited evidence can be incomplete, a management can change, an industry can turn. Each position carries a pre-committed level at which the thesis is treated as failed — that is the discipline, not a guarantee against being wrong.
Nearly half the book is in cash today. If markets rise while the funnel stays dry, that cash earns nothing and the opportunity is gone. Waiting for evidence has a cost, and it is being paid right now.
How access works
Nothing about this product involves handing over custody of your money or discretion over your account.
An invite code
Access is invite-only. You receive a single-use code, valid for 30 days, and redeem it in the member portal.
A signed agreement
A separate legal agreement is signed via DocuSign before anything starts. Terms, scope and commercials live there — not in marketing copy.
You connect your own broker
You keep your own broker account in your own name. Your money is never pooled with anyone else's, including ours.
You choose how orders are placed
When the model moves, the resulting orders are sized to your account. What happens next is the execution mode you pick: we notify you and you place them yourself, or you approve each set before it goes out, or you switch on auto-execution and they are placed for you. Auto is off unless you turn it on, it takes its own separate consent, and you can withdraw it at any time.
Orders are placed in your own account
Whichever mode you chose, the orders go through your own connected broker's API. You can see every one of them in your own broker statement, because they are your orders.
The one thing to understand
Money never pools.
Funds sit in your own broker account, in your name, throughout. At settlement, money moves out of your account to pay for the shares you bought — exactly as it does on any trade you place yourself. What never happens is your money sitting in a common pot with other members' money, or with ours.
What you can always do
The rest of the range
Derivatives.
US-listed equity, via LRS.
Systematic momentum.
Getting in
If you have a single-use code, redeem it in the member portal. If you do not, write to us. We will send you the full process documentation and the terms, and you can decide for yourself whether Core is a fit.
Minimum investment ₹98,590 (computed live — every model line ≥1 whole share); ₹5,00,000 recommended for tight tracking. A separate agreement is signed via DocuSign before anything starts. Nothing here is advice or an offer.