Strategy reference

F&O Intelligence — the owner's governing charter (RECORDED VERBATIM)

Origin: 🧑 RAMANA — wholly his. Every rule, gate and phase on this page is quoted from two

messages he wrote on 2026-08-13; the house has contributed no method here, only the transcription.

The single derived paragraph (§12) is marked as mine inline. See the origins notes.

<!-- ONE LINE, DELIBERATELY: the extractor Pat reads this with is line-anchored, so a wrapped

definition ships a mid-sentence fragment to a reader. Do not re-wrap this paragraph. -->

One-line definition: A planned evening research system that would narrow the futures-and-options universe to a short, explained watchlist for the next trading day, saying what the evidence agrees on, what contradicts it, and what would prove it wrong — and treating "nothing worth acting on tomorrow" as a proper answer; none of it is built yet, so this page records the specification only.

**Status: PARTLY BUILT (phases 1–7 of his twelve; see §12). This page remains a RECORD, not a design

— every quote below is his and is never edited to match what got built.**

**Provenance: the desk, 2026-08-13, session *"Design"* (design-audit-ecosystem-88266d), two messages

at 16:37 and 17:11 — 17,643 characters.** Recovered 2026-08-13 by transcript mining pass 6

(the internal reference §4).

⚠️ Why this page exists. He wrote a complete strategy charter in his own words and **nothing in

the internal reference §2 an attribution without a quote is a claim, not a record, so the charter

is reproduced here substantially verbatim. **Where any later summary disagrees with the quotes on

this page, the quotes win.**

Related but NOT this: the dvpt notes (delivery), the F&O data estate (per-stock OI,

participant OI, SLB — ingested nightly, NSE-primary), and the **Phase-0 verdict already on record:

PCR selects weakly; maxpain / basis / OI-change fail.** That verdict is evidence *for* this charter's

central complaint, not against it.

---

§1 The objective, in his words

*"Build an uncertainty-aware F&O research, decision and learning system that can reduce the

eligible Indian F&O universe each evening into a small, explainable shortlist for the next

trading day.*

*The shortlist is not an automatic trade instruction. I will still inspect charts and apply

judgment. The system must show **why a stock deserves attention, what evidence overlaps with

existing Patearn strategies, what contradicts the thesis, what would invalidate it**, and whether

the evidence is strong enough to act on. **"No sufficiently strong trade tomorrow" must be a valid

and respected result.**"*

The eleven questions the system must answer (verbatim):

*"Which stocks deserve inspection tomorrow? · Why are they more interesting than the remaining F&O

universe? · Which independent evidence lanes agree? · **Are those lanes genuinely independent, or

are they duplicates of the same information?** · What does the derivatives market appear to be

pricing? · Where may existing positions experience economic pressure? · What evidence contradicts

the proposed direction? · What chart or price behavior must confirm the setup? · What would

invalidate it? · How did comparable historical situations perform after realistic costs? · **How

well calibrated was the system's previous confidence?**"*

§2 The framing principles (from the 16:37 message)

*"Do not compare every stock using the same PCR/OI threshold. Each stock must be compared

against **its own historical behavior, sector context, volatility, liquidity, lot size, and expiry

behavior**.*

*Daily long buildup/short covering is not enough. I want **cumulative interpretation across

expiries, rollovers, and historical positioning**.*

*The dashboard must be rich, precise, compact, and story-driven.*

*Use only data we actually have or can legally source from primary/official sources. **Do not

invent unavailable data. Mark data gaps clearly.**"*

*(The primary-source clause is Guardrail #8 restated by him, unprompted.)*

§3 🔴 THE INDEPENDENT INTELLIGENCE CONTRACT — the part most likely to be ignored

*"**My ideas and examples are starting hypotheses, not the boundary of the work and not instructions

to implement blindly.***

*For every idea I provide: 1. Identify the deeper question behind it. 2. **Challenge its

validity and hidden assumptions.** 3. Generate materially different ways to answer the question.

4. Identify data and relationships I did not mention. 5. Find interactions with existing Patearn

evidence. 6. State the strongest competing explanation. 7. Recommend whether to test, improve,

combine, replace, defer or reject it. 8. Define what evidence would falsify it.*

*After handling my ideas, run a separate Innovation Pass… **Do not wait for me to enumerate

every feature.** Proactively identify missing questions, better formulations, unusual but defensible

combinations… Separate established methods from your own new hypotheses.*

*Novelty alone is not valuable. An innovation may advance only if it has a plausible mechanism,

primary-source data feasibility, a falsifiable prediction, a validation method and an acceptable

risk of misleading the user."*

⚠️ **This is the same instruction as the OWNER DOCTRINE §5d (one symptom is a class) and § F4

(*"why should we confine to the component we have to live with?"*), stated as a contract.** It also

supersedes any reading of the 16:37 message as a feature list — he says so explicitly.

§4 🔴 THE EPISTEMIC CONTRACT — four labels, and the OI limitation stated by him

*"Every material conclusion must be classified as: OBSERVED — directly present in primary-source

data · DERIVED — reproducibly calculated from observed data · INFERRED — a probabilistic

interpretation dependent on stated assumptions · UNKNOWABLE — not responsibly recoverable from

available data.*

***Never present option buying, option writing, institutional ownership, support, resistance, dealer

positioning, paired strategies, accumulation or distribution as established fact merely from

aggregate OI.***

*Every open contract has both a long and a short. Price-plus-OI labels such as long buildup and

short covering are heuristics, not identification of the initiating trader. The system must

expose this limitation and attach **confidence, evidence quality, competing interpretations and

uncertainty intervals**."*

§5 POSITION ECONOMICS — his central requirement

*"Equal OI must not be treated as equal strength, conviction, risk, pressure or pain."*

Distinguish, in his list: *raw OI · rupee-notional OI · delta-equivalent OI · gamma-, vega- and

theta-weighted exposure · estimated entry-premium-adjusted OI · expiry breakeven distance · distance

from spot to breakeven · distance normalized by expected move or volatility · estimated

mark-to-market pressure · margin- and capital-weighted pressure · liquidity-adjusted exposure ·

scenario and stress losses.*

*"For a short call, estimated expiry breakeven is strike plus received premium. For a short put, it

is strike minus received premium. **But breakeven is an expiry concept: current pain must also

account for mark-to-market premium changes, IV, time, Greeks, margin changes, hedges and

liquidity.***

*Investigate OI-vintage reconstruction. When OI increases, create estimated entry cohorts

using contemporaneous executable or best-available prices. When OI decreases, test **FIFO, LIFO,

proportional and price-sensitive cohort retirement. Report a range when these assumptions

disagree.***

*Investigate probable straddle, strangle, spread and roll detection… **Call these structures

probable, not detected facts**, unless the data truly supports identification. Provide pairing

confidence and alternative explanations.*

*Build position-pressure and pain surfaces across spot, IV and time scenarios. **Do not reduce

this multidimensional information to a simplistic "OI wall."**"*

His own worked example, verbatim:

*"If a stock is at 500, lot size is 1000, and 500 PE premium is 5, then premium received is 5000 and

assignment exposure is 500000. I want the system to judge whether that premium is worth the risk

based on the stock's own history."*

§6 CONFLUENCE — successive gates, never an additive score

*"Do not create a conviction score by casually adding correlated indicators. Use successive

gates:*

*1. Existing Patearn strategies and charts generate candidates. 2. Futures and options provide

confirmation or contradiction. 3. Position economics estimates pressure, crowding and

vulnerability. 4. Volatility determines whether risk is cheaply or richly priced. 5. **Liquidity and

execution can veto a setup.** 6. Events, MWPL, settlement and margins can veto or downgrade it.

7. Historical calibration evaluates comparable point-in-time cases. 8. **Conflict analysis prevents

false confidence. 9. The final system produces Review First, Watch/Incomplete and Rejected**

groups.*

*Measure overlap between signals. Five variations of OI are not five confirmations. Estimate

redundancy, correlation and incremental information before calling evidence "confluence."*"

Every shortlisted stock must carry (verbatim): *proposed direction or neutral opportunity · setup

and originating Patearn strategy · independent confirming evidence · contradictory evidence ·

evidence quality and data freshness · calibrated confidence, not rhetorical confidence · relevant

price/volatility scenarios · chart confirmation required · invalidation condition · liquidity and

execution assessment · event and settlement risks · comparable historical sample size and outcomes ·

why it ranked above alternatives · what would change the conclusion · **whether the proper decision is

no trade**.*

§7 VALIDATION AND LEARNING

*"Every predictive hypothesis must be evaluated point-in-time with: survivorship and look-ahead

controls · corporate-action correctness · realistic costs, spreads, slippage and liquidity ·

simple baselines and doing nothing · walk-forward or genuinely held-out testing · performance by

stock, regime, expiry distance and liquidity bucket · false-positive rate and coverage · maximum

favorable and adverse excursion · confidence calibration · sensitivity to modeling assumptions ·

independent reproduction where practical.*

*Do not optimize only for returns. Evaluate drawdown, tail loss, stability, turnover, capacity

and decision usefulness. **Preserve negative findings in the strategy ledger. Descriptive context

may remain useful even when it has no trading edge, but it must not silently contribute predictive

conviction.***

*Paper-forward-test the daily shortlist before treating it as actionable. Store exactly what the

system knew at decision time, subsequent outcomes, invalidations and **whether abstaining was

correct**."*

⚠️ *"Must not silently contribute predictive conviction"* is the descriptive-only fence in his own

words — the rule that governs Credibility, Accumulation, Wolfe and seasonal.

§8 PRODUCT ARCHITECTURE — and his own restraint on adding altitude

*"F&O is substantial enough to require a dedicated workspace, **but do not immediately add another

global top-level altitude merely because the feature set is large.***

*Preferred information architecture: Markets → F&O Intelligence, with internal views:

**Tomorrow · Futures and rollover · Options positioning · Volatility · Position economics and

probable structures · Participants and market structure · Research and validation.***

*Each stock dossier should retain a stock-specific F&O section… This creates both

market-to-stock discovery and stock-to-evidence investigation.*

*Avoid one huge page and avoid a separate page for every metric. Follow Patearn's existing

design language, registry, Surface Playbook, Pat coverage, education, provenance, CSV and

honest-empty requirements. **Preserve existing routes and sacred surfaces. Extend additively through

isolated modules.**"*

*(He restates Guardrail #9 and the additive rule himself.)*

§9 THE WORKING METHOD — twelve phases, his sequence

*"Do not attempt the complete system in one uncontrolled implementation pass."*

1. Re-grounding and stale-assumption report · 2. Complete data census and feasibility matrix ·

3. Epistemic and inference contract · 4. Position-economics research prototypes · 5. Probable-structure

and rollover research · 6. Volatility and distribution research · 7. F&O workspace information

architecture · 8. One end-to-end stock-level vertical slice · 9. Point-in-time falsification and

backtests · 10. Confluence engine using only approved evidence · 11. Daily paper-forward-test and

calibration · 12. Broader rollout only after gates pass.

At each phase (verbatim): *state what I requested · restate the deeper objective · **identify what

I failed to ask** · contribute independent ideas · list assumptions challenged · define experiments

and gates · record rejected ideas and reasons · identify the most valuable next investigation ·

*"Use other available models or agents as an adversarial research council… explorer, quantifier,

skeptic, data auditor, backtester, replication agent and verifier. **Do not use model agreement as

proof.**"*

§10 THE GOVERNING PRINCIPLE — his closing sentence, verbatim

***"You are responsible not only for implementing what I have imagined, but also for discovering

what I have not yet imagined, while proving that every proposed innovation deserves to exist."***

---

§11 What he also asked for in the 16:37 message — the deliverable list

Seven numbered outputs, *"in tables wherever possible"*: 1. an F&O signal taxonomy (14 named

signals, each with required data · formula/logic · historical comparison method · confidence score ·

failure cases · visualization idea) · 2. stock-relative metrics (PCR/OI/IV percentile,

premium richness, strike distance, lot-size capital exposure, expiry behavior, historical assignment

risk, rollover persistence) · 3. an option-writing risk framework ending in an

*"avoid / watch / acceptable / attractive"* label · 4. dashboard sections, each with *what it

shows · why it matters · what action the user takes next · what charts/tables are needed* ·

5. ten differentiated ideas phased 1/2/later · 6. a data audit classifying every field as

*likely already available · needs verification · missing but sourceable · not allowed / avoid* ·

7. top-5 highest-impact features, top-5 risks/misinterpretations, and open questions.

*"Keep the answer concise but complete. Use structured tables. **Do not give generic textbook

explanations.**"*

---

§12 Status, stated honestly

*Updated 2026-08-14 (S-FNO-P7). The rows moved; his words above did not.*

🔴 The Phase-0 failure is the charter's own argument. Those tests used **one threshold across all

stocks and daily** OI change — precisely the two things §2 rules out. *That is my reading, not his

words, and it is a hypothesis to test rather than a reason to re-run anything.*

Nothing on this page may be treated as a design decision. It is what he asked for, in his words,

so the next session does not have to guess — and so no future summary can quietly replace it.

🔴 **And that rule now has teeth it did not have on 2026-08-13: part of this charter is built, so the

temptation from here is to edit his words toward what shipped.** The §12 table is the only place that

moves. If a future reader finds a quote above that matches the product suspiciously well, check it

against the 2026-08-13 transcript before trusting it — a charter that drifts to match its

implementation stops being able to say the implementation is wrong.

September 13, 2026 — separate F&O Lab implementation

Implementation note, not a change to the verbatim charter above: /markets/fno/lab adds a separate, owner-preview workspace. It reads exact stock/index dates and expiries; offers 17 structural option examples and a six-leg user-directed calculator; compares explicit portfolio shocks; exports assumptions with results; and freezes private assessments. The existing F&O views remain available, with a visible link to the lab. Expiry arithmetic is distinct from execution, settlement feasibility and historical performance. Source coverage is never converted into a win rate. The full conditional-history, failure-atlas and forward-learning program remains open in the internal reference §12.10.