Philosophy of the Canon Scoring System
First published 23 Aug 2026 · Last verified 29 Aug 2026
Lesson 63.1 — Why Ad Hoc Stock Picking Fails on NEPSE
At a health post in rural Nepal, a new patient arrives with fever, stomach pain, and fatigue. An experienced doctor could simply look at the patient, take a guess based on the last ten similar cases she remembers, and prescribe something. Most of the time, she might even be right. But hospitals do not run on "most of the time." They run on checklists — a fixed sequence of questions and tests: check temperature, check blood pressure, ask about travel history, test for malaria, test for typhoid, review medication history. The checklist does not replace the doctor's judgment. It disciplines it. It forces her to look at the same set of things, in the same order, every single time, so that a bad day, a distracting patient in the next bed, or a false first impression does not cause her to skip something that matters.
Investing in companies listed on the Nepal Stock Exchange (NEPSE — the country's only stock exchange, based in Kathmandu) has, for most retail investors, worked more like the doctor guessing from memory than the doctor using a checklist. An investor hears that a friend made money in a hydropower initial public offering (IPO — the first sale of a company's shares to the public). Another investor sees a bank's share price rising for three straight weeks and assumes it must be a "good company." A third buys a finance company's shares because the managing director is a relative of someone he trusts. None of this is analysis. It is pattern-matching on fragments of information, filtered through whatever the investor happened to notice that week.
This chapter opens Part XIII of this book, which introduces the Canon Score — a structured, repeatable, weighted framework this book will use, starting in Chapter 64, to evaluate the quality of individual companies listed on NEPSE. Before this book hands you that framework, it owes you an explanation of why such a framework is necessary at all. That is the entire purpose of this chapter: not to give you the scorecard, but to explain why a scorecard beats guessing, why ratios alone are not enough, and why a framework built for Wall Street or Dalal Street cannot simply be imported wholesale into Kathmandu.
Start with the plain problem of ad hoc analysis — "ad hoc" is a Latin phrase that simply means "for this particular case," done without a general system, made up on the spot. Ad hoc stock picking means: no fixed process, no consistent questions, no comparability from one company to the next. You analyse Company A by looking at its dividend history because that's what caught your eye. You analyse Company B by looking at its chairman's reputation because that's what a friend mentioned. You analyse Company C by looking at the share price chart because that's the only data you bothered to open. Three completely different lenses, three completely different "analyses," and no way to honestly compare A, B, and C against each other, because you never asked them the same questions.
Consider a concrete Nepali example. Suppose in early 2023 an investor was choosing between three hydropower companies newly listed on NEPSE. One had just completed construction and started generating revenue. One was still under construction with a delayed commissioning date. One had a long operating history and a strong monsoon-season output record. An ad hoc investor might buy the first one simply because "hydropower is the future of Nepal" — a true statement, but not a reason to prefer one hydropower company over another. He has made a sector-level judgment and mistaken it for a company-level judgment. This is one of the most common and costly errors on NEPSE: confusing "I like this sector" with "I have evaluated this company." Banking, hydropower, and life insurance are three of NEPSE's largest and most crowded sectors. Liking the sector tells you almost nothing about which of the twenty-plus companies inside it deserves your capital.
Ad hoc analysis has three specific failure modes worth naming, because each one recurs constantly on NEPSE.
The first is recency bias — the tendency to overweight whatever information arrived most recently, simply because it is freshest in memory. A company's share price has doubled in two months due to speculative buying ahead of a bonus share announcement (a bonus share is additional stock given free to existing shareholders, funded from the company's reserves, which increases the number of shares outstanding without adding new capital). The ad hoc investor sees the price chart, concludes the company must be strong, and buys at the top — never having looked at whether the underlying business actually improved.
The second is anecdote substitution — replacing systematic evidence with a single vivid story. "My uncle worked there for ten years and says it's well run" feels like real information. It may be true. It may also be completely irrelevant to whether the company's balance sheet is overleveraged or whether its promoter (the founding shareholder group that typically also runs the company — a term used constantly in Nepali company law and stock market commentary) has been quietly pledging shares against personal loans.
The third is halo transfer — letting one positive trait bleed into an unjustified positive impression of everything else. A bank with a famous, charismatic chief executive is assumed to also have strong risk controls, though the two have no necessary connection. Nepal's banking sector merger wave through the 2010s and 2020s, driven by Nepal Rastra Bank's (NRB — the central bank of Nepal, which regulates and supervises all banks and financial institutions) capital requirements, produced several mergers where a well-regarded acquiring bank absorbed a weaker one — investors who bought purely on the acquirer's reputation, without checking the combined entity's post-merger asset quality, were sometimes surprised by the non-performing loan (NPL — a loan on which the borrower has stopped making scheduled payments) burden they inherited.
None of this means intuition is worthless. A seasoned investor's gut feeling is often built on years of pattern recognition and is not to be dismissed. The problem is not intuition itself — it is intuition used as the entire process, with no structure to catch its blind spots. A structured score does not throw away intuition. It gives intuition a job: help interpret the score, flag things the score might miss, decide what to do at the margin. But the score itself is built the same way, every time, for every company, which is precisely what raw intuition cannot promise.
Lesson 63.2 — The Problem With Ratios in Isolation
Earlier parts of this book taught you how to read a balance sheet, an income statement, and a cash flow statement. You learned to compute ratios: return on equity (ROE — net profit divided by shareholders' equity, showing how efficiently a company turns owners' capital into profit), the price-to-earnings ratio (P/E — share price divided by earnings per share, showing how expensive a stock is relative to its profit), the debt-to-equity ratio (total debt divided by shareholders' equity, showing how leveraged a company is), and many others specific to banks, hydropower companies, and insurers.
Ratios are enormous progress over ad hoc analysis. A ratio is at least a number, computed the same way every time, comparable across companies. But ratios in isolation — meaning looked at one at a time, disconnected from each other and from everything qualitative — create a different and subtler problem: they can each look fine individually while the whole picture is not fine at all.
Think of this the way a doctor thinks about vital signs. Blood pressure alone does not diagnose a patient. Neither does temperature alone, or pulse alone. A patient can have a normal temperature and a dangerously low blood pressure at the same time — and a doctor who checks only temperature would send that patient home. The vital signs have to be read together, and even together, they have to be interpreted against the patient's history and symptoms, not treated as a mechanical pass/fail test.
The same is true of company ratios on NEPSE. Consider a hypothetical finance company with an eye-catching ROE of 22 percent — well above the sector average. Looked at alone, this ratio says: strong. But suppose that ROE is being generated by a debt-to-equity ratio of 9:1, meaning the company is earning that return almost entirely through heavy borrowing, not through operating skill. A small deterioration in loan quality could wipe out equity fast. The ROE ratio, taken alone, hid the real story. Only by reading ROE together with leverage, together with loan-loss provisioning, together with capital adequacy (a ratio, regulated by NRB, measuring how much capital a bank or finance company holds relative to its risk-weighted assets, to absorb losses) does the real picture emerge.
Isolated ratio-reading has three specific weaknesses.
First, ratios cannot see each other's blind spots unless someone deliberately puts them side by side. A ratio-only investor who checks P/E this week and debt-to-equity next month, in two separate mental exercises, may never notice that the company's low P/E (which looks cheap) exists precisely because the market has already priced in the leverage risk that the debt-to-equity ratio would have shown him, had he looked at both together.
Second, ratios are entirely backward-looking. They are computed from historical financial statements, typically reported quarterly or annually. A ratio tells you what happened in a completed reporting period. It says nothing directly about governance quality, management competency, promoter intentions, or how a company will behave in the next monsoon season, the next interest rate cycle set by NRB, or the next regulatory change from the Securities Board of Nepal (SEBON — the regulator overseeing the securities market, listed companies' disclosure obligations, and merchant banking activity in Nepal). Two companies can show identical five-year ROE histories and have completely different futures, because one has a disciplined, professional management team and the other has a promoter family known for related-party dealing.
Third, and most dangerous for NEPSE specifically, ratios computed from financial statements are only as reliable as the financial statements themselves. As earlier chapters on financial statement analysis discussed, disclosure quality varies significantly across NEPSE-listed companies. A ratio calculated from understated provisioning or optimistic asset valuation is not a conservative ratio — it is a wrong ratio wearing the costume of precision. A single number with two decimal places feels more trustworthy than it sometimes deserves to be.
This is not an argument against ratios — this book has spent many earlier chapters teaching you to compute and interpret them correctly, and you will keep using them. It is an argument against treating ratio-reading as if it were, by itself, a complete evaluation system. A ratio is an input. It is not a verdict. What NEPSE investors have generally lacked is not data — Nepal's listed companies publish quarterly reports, and CDSC (the Central Depository System and Clearing Limited, which holds dematerialized — electronic, paperless — share records for Nepali investors) and the exchange itself make trading and shareholding data available. What has been missing is a consistent method for turning scattered ratios and scattered qualitative facts into one coherent judgment about company quality. That gap is exactly what a scoring system exists to close.
Lesson 63.3 — The Philosophy of Decomposing Company Quality Into Dimensions
"Decompose" simply means to break something large and complicated into smaller, separately understandable parts. A mechanic does not diagnose a car by asking "is this a good car?" as a single yes/no question. She checks the engine, the brakes, the transmission, the electrical system, the tires — separately, one system at a time — and only then forms an overall judgment about the car's condition. Each system can be checked on its own terms, using tools suited to that system, and a fault in one system does not automatically hide a fault in another, because she is not trying to judge everything at once through a single foggy impression.
"Company quality" is exactly this kind of large, complicated thing. It is not one fact. It is a bundle of many different, only loosely related facts: how profitable the company is, how it is financed, how honestly it is governed, how easily you can buy and sell its shares, how exposed it is to Nepal-specific risks, how sensibly it has allocated capital over time, and what you are being asked to pay for all of that today. Trying to judge all of this at once, in your head, in a single impression, is exactly the ad hoc failure mode described in Lesson 63.1. The fix is decomposition: split company quality into distinct dimensions, score each dimension on its own terms using the evidence suited to it, and only then combine the dimension scores into one overall number.
This is not a novel idea invented for this book — it is standard practice among the world's most rigorous evaluators of companies and debt issuers. Global credit rating agencies, when they assign a credit rating to a company's debt, do not ask "is this company creditworthy?" as one question. They explicitly separate a company's business risk profile — the quality and stability of the underlying business, its competitive position, and the industry it operates in — from its financial risk profile — its leverage, cash flow adequacy, and capital structure — score each separately using different evidence, and only then combine the two into a final rating. A company can have an excellent business and a weak financial structure, or a mediocre business propped up by a very conservative balance sheet. Collapsing those two very different situations into one un-decomposed impression would destroy the very information that made the two-part analysis useful in the first place.
The same decomposition logic underlies "quality factor" investing, a well-studied approach used by large global asset managers, which does not ask "is this a quality company?" as one vague question either. It typically separates quality into sub-factors such as profitability, earnings stability, low leverage ("safety"), and consistency of shareholder payouts — measuring each with different data, then combining them. And it underlies ESG scoring (Environmental, Social, and Governance scoring — a framework used by index providers and asset managers to rate companies on non-financial risk), which explicitly scores the E, the S, and the G as separate pillars, because a company can be strong on one and weak on another, and an investor needs to know which is which, not just a single blended letter grade that hides the difference.
Three philosophical commitments follow from choosing decomposition as the design principle for the Canon Score, and it is worth stating them plainly now, even though the specific dimensions and point values are Chapter 64's job.
The first commitment is that dimensions must be genuinely distinct — each one should capture something the others do not, so that scoring them separately adds real information rather than just re-measuring the same underlying fact five different ways. A company's profitability and its governance quality are genuinely different things; a company that manipulates its own reported profitability blurs this line, which is itself a governance problem worth scoring on its own terms.
The second commitment is that each dimension should be scored using evidence appropriate to that dimension, not forced through a single generic lens. Profitability is scored primarily from financial statement data. Governance is scored partly from disclosure behaviour, related-party transaction history, and board structure — different evidence entirely. Trying to judge governance using only profit ratios, or judging profitability using only governance impressions, defeats the purpose of separating them in the first place.
The third commitment is that the dimensions must be combined, not left scattered. A decomposed score that never gets reassembled into one overall number is not more useful than no score at all — it just produces a pile of separate facts with no way to compare Company A's overall quality against Company B's. The combination step — how much weight each dimension carries in the final number — is where judgment about what matters most for a Nepali investor gets encoded. That weighting is exactly what Chapter 64 will lay out in full: seven dimensions, each scored on its own terms, combined into a single 0–100 Canon Score.
Table: Ad Hoc Analysis, Ratio-Only Screening, and Structured Scoring Compared
| Attribute | Ad Hoc "Gut Feel" Picking | Ratios Screened in Isolation | Structured, Decomposed Scoring |
|---|---|---|---|
| Repeatable across companies | No — different information used for each company | Partially — same ratios, but no fixed combination method | Yes — same dimensions, same weights, every time |
| Comparable across companies | No — impressions are not comparable numbers | Limited — ratios are comparable one at a time, not overall | Yes — a single combined score is directly comparable |
| Handles conflicting signals | Poorly — one loud fact dominates the impression | Poorly — a strong ratio can mask a weak one nearby | Explicitly — each dimension scored and weighted separately |
| Captures governance and qualitative risk | Inconsistently, if at all | Rarely — most ratios are purely financial | Yes — governance is scored as its own dimension |
| Vulnerable to recency and halo bias | Highly vulnerable | Less vulnerable, but ratio selection itself can be biased | Structurally resistant — the checklist does not change with mood |
| Produces an auditable record | No — reasoning is not written down | Partially — ratios can be recorded, but not the overall verdict | Yes — every score and its basis can be recorded and reviewed later |
| Adapted to Nepal-specific risk factors | Depends entirely on the individual investor's knowledge | No — generic ratios do not encode Nepal-specific risk | Yes, if deliberately designed to — this is Lesson 63.4's subject |
Lesson 63.4 — Why NEPSE Needs Its Own Scoring Framework, Not an Imported One
It would be far less work to simply take a company-scoring framework built by a large American or Indian financial institution and apply it directly to NEPSE-listed companies. The temptation is real: such frameworks already exist, are well tested, and carry institutional credibility. This book deliberately does not do that, and the reasons are not cosmetic. They come directly from structural features of the Nepali market that earlier parts of this book have already documented in detail.
The first reason is liquidity. Liquidity — how easily an asset can be bought or sold without materially moving its price — behaves very differently on NEPSE than on the New York Stock Exchange or the Bombay Stock Exchange. Many NEPSE-listed companies trade only a few thousand shares on an average day, and some trade far less. A scoring framework built for the US market can reasonably treat liquidity as background noise, because almost every listed company there trades enough volume that liquidity risk is a minor factor for a typical retail position. On NEPSE, liquidity is not background noise — it is frequently the difference between a paper gain and a gain you can actually realise. A company might score well on every financial metric and still be nearly impossible to exit in size without crashing its own price, especially given NEPSE's daily circuit filters (rules that halt trading in a stock once its price moves up or down by a set percentage in a session, which can prevent an investor from selling at all on a day of bad news). A Nepal-specific score has to weight liquidity explicitly, at a level of seriousness a framework built for deep, liquid markets never bothered to include.
The second reason is governance. This book's earlier chapters on corporate governance and promoter behaviour in Nepal — the sections of this book that examined ownership concentration, related-party transactions, and board independence — documented governance weaknesses that are structurally more common on NEPSE than in markets with longer histories of institutional shareholder activism, larger analyst coverage, and stronger minority-shareholder legal protections. Reports examining corporate governance in Nepal have repeatedly flagged weak board independence, concentrated promoter control, and inconsistent enforcement of disclosure norms as recurring, market-wide issues rather than isolated cases. A scoring framework imported from a market where institutional investors routinely hold managements accountable, and where minority shareholder litigation is a real deterrent, will systematically underweight governance risk when applied to NEPSE, because it was calibrated for a different accountability environment.
The third reason is concentration — both sector concentration and ownership concentration. NEPSE's market capitalisation is heavily concentrated in banking, hydropower, and life and non-life insurance. This book's chapters on sector analysis walked through why each of these sectors carries distinct structural risks: banks are exposed to NRB's monetary policy stance and periodic liquidity crunches tied to remittance inflow cycles (remittances — money sent home by Nepali migrant workers abroad, a major source of the foreign currency and bank deposits that fund NEPSE-listed banks' lending); hydropower companies are exposed to monsoon variability, transmission-line bottlenecks, and power purchase agreement terms with the Nepal Electricity Authority; insurers are exposed to actuarial and investment-portfolio risk specific to a shallow domestic capital market. A generic scoring framework built for a diversified market like the US, where technology, healthcare, industrials, and consumer companies each represent a meaningful share of the index, has no reason to build in sector-specific adjustments this heavy. On NEPSE, ignoring sector concentration means ignoring a defining feature of the market itself.
The fourth reason is structural: settlement, disclosure timeliness, free float (the proportion of a company's shares actually available for public trading, as opposed to shares locked up with promoters), and the depth of analyst coverage all differ meaningfully between NEPSE and larger markets. CDSC's dematerialization of shareholding has been a genuine modernisation of Nepal's market infrastructure, but the depth of independent equity research covering individual NEPSE-listed companies remains far thinner than in markets served by dozens of competing brokerage research desks. A framework assuming rich third-party analyst coverage as a supplementary check on management's claims is assuming something that, for most NEPSE companies, does not exist. The Canon Score has to be buildable primarily from what a diligent retail investor can actually obtain: company filings, SEBON and CDSC data, NEPSE trading data, and public disclosures — not from proprietary data feeds or analyst consensus estimates that simply do not exist for most Nepali companies.
None of this is a claim that NEPSE companies are inherently worse than companies listed on larger exchanges, or that Nepali corporate governance is beyond repair — both claims would be unfair and untrue, and this book has profiled well-governed, well-run Nepali companies throughout its earlier chapters. The claim is narrower and more precise: the risks that matter most for judging company quality differ in kind and in weight between NEPSE and larger, more liquid, more heavily regulated markets. A scoring system that does not build those differences in from the start will produce scores that are precise-looking and quietly wrong — the same failure this chapter warned about with isolated ratios in Lesson 63.2, now at the level of an entire framework rather than a single number.
Lesson 63.5 — What a Score Can and Cannot Tell You: Honest Limits
Every measurement tool has limits, and pretending otherwise is how tools get misused. A thermometer tells you a patient's temperature. It does not tell you what caused the fever, and it does not guarantee the patient will recover. A loan officer's credit scorecard tells a Nepali cooperative or bank whether a borrower's documented history fits the pattern of past reliable borrowers. It does not guarantee the borrower will repay — a job loss, a family emergency, or a bad harvest can still intervene after a high score was assigned. The scorecard narrows uncertainty. It does not eliminate it.
The Canon Score, once Chapter 64 lays it out in full, will work the same way. It is worth being completely honest about its limits now, before you ever see the first number attached to a real company, because a tool oversold is a tool that will eventually be blamed for failures that were never really its to prevent.
The first limit: a score is not a prediction. A high Canon Score describes the observable quality of a company today, based on the dimensions the framework measures. It does not promise that the share price will rise, and it does not promise the company will still deserve a high score in three years. Businesses change. A well-run hydropower company today can be poorly run after a change in senior management five years from now. A bank with strong asset quality today can deteriorate if it underwrites aggressively during the next credit cycle. Past scores describe the past and, at best, the present. They are not a contract about the future.
The second limit: a score is only as good as its inputs. This is the single most important honest caveat in this entire chapter, and it echoes directly back to Lesson 63.2's warning about ratios computed from unreliable financial statements. If a company's disclosed financial statements are inaccurate — whether through simple error, aggressive accounting choices, or deliberate misstatement — then any score computed from those statements will be inaccurate too, no matter how carefully the scoring framework itself was designed. A scoring system does not manufacture truth out of bad data. It organises whatever data it is given, faithfully, including any lies buried inside that data. This is precisely why later chapters, when they build out the Canon Score's governance dimension, will spend real effort on disclosure-quality checks and red flags — because the framework's own integrity depends on taking the reliability of its inputs seriously, not assuming it away.
The third limit: a single combined number can hide important detail if you stop looking at it too early. Two companies can arrive at the same overall Canon Score through very different paths — one strong on financial strength but weak on liquidity, the other the reverse. Chapter 64 will show you how to read the dimension-level breakdown behind the headline number precisely so this does not happen to you; the overall score is a starting point for further reading, not a replacement for it. An investor who only ever looks at the final number, and never opens up the dimensions behind it, has partly recreated the very problem decomposition was meant to solve — collapsing distinct information back into one undifferentiated impression, just with more decimal places attached.
The fourth limit: judgment still has the final word. A scoring system structures your analysis. It does not remove the need for you to think. If the Canon Score's governance dimension has not yet caught a very recent, still-unfolding scandal because the framework's inputs have not updated yet, a diligent investor who reads the news should not wait for the score to catch up before adjusting her own view. The score is a floor for discipline, not a ceiling on thinking. This book will say this again, more than once, in the chapters that follow, because it is the single most common way any scoring tool — in finance, in medicine, in credit — gets misused: treated as a final verdict instead of a structured input into a verdict a human being still has to make.
The fifth limit, specific to Nepal: data quality and update frequency constrain what any score can capture in real time. NEPSE companies report quarterly, not continuously. A score built on the most recent quarterly filing can be, at worst, close to three months stale by the time the next filing arrives. Fast-moving developments — a sudden change in NRB's monetary policy stance, a hydropower company's transmission line failure, a bank's sudden liquidity stress — can outrun the score's own update cycle. A structured score reduces the damage of ad hoc bias considerably. It does not, and cannot, turn investing into a mechanical, riskless exercise.
Lesson 63.6 — How the Canon Score Will Be Used Going Forward
With the philosophy now in place, it is worth previewing — briefly, without yet revealing the full rubric — how the Canon Score will actually function across the rest of this book, so you know what to expect as you move into Chapter 64 and beyond.
Chapter 64, immediately following this one, will lay out the Canon Score's full architecture: seven distinct dimensions of company quality, each scored on its own defined criteria, combined through explicit weights into a single overall score out of 100 points. You will learn exactly what evidence feeds each dimension, how points are assigned within each one, and how the seven dimension scores are combined into the headline number. Nothing about the specific weights or point allocations is revealed in this chapter deliberately — this chapter's job was philosophy, not mechanics — but you can expect the seven dimensions to track directly from everything this chapter has argued: a dimension addressing financial strength and profitability, a dimension addressing governance and promoter behaviour, a dimension addressing liquidity and tradability, a dimension addressing valuation reasonableness relative to what you are being asked to pay, a dimension addressing sector and business model durability in the Nepali context, a dimension addressing the growth trajectory, and a dimension addressing dividend and capital return discipline.
Table: Preview of the Seven Canon Score Dimensions (Full Rubric in Chapter 64)
| Dimension | Points | What It Broadly Captures |
|---|---|---|
| Financial Strength & Profitability | 20 | How efficiently and sustainably the company generates profit from the capital it employs |
| Governance & Promoter Behaviour | 15 | Board independence, related-party dealing, disclosure honesty, and promoter behaviour |
| Liquidity & Tradability | 10 | How easily an investor can actually buy and sell the stock in meaningful size |
| Valuation Reasonableness | 15 | Whether the current share price is reasonable relative to the company's demonstrated quality |
| Sector & Business Model Durability | 15 | Exposure to Nepal-specific structural risks — monetary policy, monsoon, remittances, regulation — and whether the business model endures |
| Growth Trajectory | 15 | Whether revenue and earnings growth is durable and well-financed, not just fast |
| Dividend & Capital Return Discipline | 10 | How consistently management actually returns profit to shareholders, and from genuine earnings |
| Total | 100 |
Chapter 65 will take this rubric and apply it, in worked detail, to real categories of NEPSE-listed companies — showing how the same seven-dimension framework produces different scores, and different insights, when applied to a bank, a hydropower company, and an insurer, each of which has its own dominant risks. Chapter 66 will address how to use the Canon Score operationally within a portfolio — how it complements, rather than replaces, the portfolio construction and risk management principles you learned in Part XII, and how to combine a company's Canon Score with position-sizing, diversification, and valuation judgment to make an actual buy, hold, or sell decision.
It is worth being explicit about what the Canon Score is not, one final time, before this chapter closes. It is not a market-timing signal — a high score does not tell you today is the day to buy, only that the underlying business, as measured, is of high quality. It is not a substitute for position sizing and diversification — even the highest-scoring company on NEPSE should not become your entire portfolio, for exactly the reasons Part XII spent four chapters explaining. And it is not a static, one-time judgment — company quality changes, and so a Canon Score calculated today needs to be revisited, not filed away as a permanent verdict. The Canon Score is, in the end, exactly what this chapter has argued a good scoring system should be: a disciplined, decomposed, Nepal-specific way of asking the right questions about a company, in the same order, every time — leaving the final decision, as it always must, to you.
Chapter recap
This chapter opened Part XIII of this book by making the case for why NEPSE investors need a structured, quantitative company-scoring system rather than relying on ad hoc, feeling-based stock picking. Using the analogy of a doctor's diagnostic checklist, the chapter showed that ad hoc analysis — buying a company because of a friend's tip, a rising share price chart, or a liked promoter's reputation — suffers from recency bias, anecdote substitution, and halo transfer, none of which can be checked, compared across companies, or corrected when wrong. The remedy is not to abandon judgment but to discipline it with a fixed, repeatable process, exactly as a checklist disciplines a diagnosis without replacing the diagnostician.
The chapter then showed that ratios, while a major improvement over pure guesswork, are not sufficient on their own when read in isolation. A ratio like ROE can look strong while hiding leverage risk that only shows up when read alongside a debt-to-equity ratio; ratios are backward-looking and only as reliable as the financial statements they are computed from, which is a genuine concern given variable disclosure quality across NEPSE-listed companies. The lesson was not to distrust ratios — this book will keep using them throughout — but to stop treating any single ratio, or any unconnected pile of ratios, as a complete verdict on a company.
From there the chapter introduced decomposition as the philosophical core of the Canon Score: breaking "is this a good company?" into several genuinely distinct, independently scorable dimensions — profitability, growth quality, governance, liquidity, sector exposure, capital allocation, and valuation — each judged on evidence suited to it, and then combined into one overall number. This mirrors established practice among credit rating agencies, which separate business risk from financial risk before combining them, and quality-factor investing frameworks and ESG scoring systems, which separate profitability, safety, and payout quality, or environmental, social, and governance pillars, rather than judging a company through one undifferentiated impression.
The chapter argued, at length, that this decomposed framework has to be built specifically for NEPSE rather than imported from a US or Indian model, because Nepal's market differs structurally in ways that matter: thinner liquidity and circuit filters that can trap an investor in a position, governance enforcement still catching up through instruments like the NEPSE Governance Code and SEBON's disclosure directives, heavy sector concentration in banking, hydropower, and insurance with each sector's own Nepal-specific risk drivers, and thinner independent research coverage than deeper markets enjoy. A framework calibrated to a different market's assumptions will systematically misjudge the risks that matter most on NEPSE.
The chapter was equally insistent on the limits of any scoring system, including the Canon Score once it is fully built. A score is not a prediction of future price performance; it is only as reliable as the inputs — mainly company disclosures — that feed it; a single headline number can hide important detail unless the dimensions behind it are examined; and judgment still has the final word when new information outruns a score's update cycle. The Canon Score structures analysis. It does not replace the investor's own thinking, any more than a checklist replaces the doctor holding it.
The next chapter, Chapter 64, "The Seven-Dimension Company Quality Score (0–100)," delivers the framework this chapter has only previewed. It will define each of the seven dimensions in full — Financial Strength & Profitability, Growth Quality & Sustainability, Corporate Governance & Promoter Integrity, Liquidity & Tradability, Sector & Macro Positioning, Capital Allocation & Management Track Record, and Valuation Discipline — specify exactly how each is scored from available NEPSE, SEBON, and CDSC data, and show precisely how the seven dimension scores combine into a single Canon Score out of 100 points. Chapters 65 and 66 will then apply that rubric to real categories of Nepali companies and show how the score integrates into the portfolio construction discipline built in Part XII.