Using the Canon Score in Portfolio Decisions

First published 23 Aug 2026 · Last verified 29 Aug 2026

Lesson 66.1 — Using the Canon Score as a Screening Filter

Every year, tens of thousands of Nepali students sit for entrance examinations for medicine, engineering, and the civil service. Almost none of them get in on the strength of the exam score alone. The score does something narrower and more useful: it decides who even gets called for the next round — the interview, the document check, the merit list. A student who scores 40 out of 100 does not get invited to argue that she is secretly brilliant. A student who scores 85 is not admitted on the spot either; she still has to clear the interview, produce her citizenship papers, and pass the medical test. The score is a filter, not a verdict.

This is exactly the job the Canon Score should do in your portfolio. Chapter 64 gave you a rigorous, seven-dimension, 0-to-100 rubric — Financial Strength & Profitability, Governance & Promoter Behaviour, Liquidity & Tradability, Valuation Reasonableness, Sector & Business Model Durability, Growth Trajectory, and Dividend & Capital Return Discipline. Chapter 65 showed you how to adjust that rubric for the quirks of banks, hydropower companies, microfinance institutions, and insurers. By the time you reach this chapter, you should already know how to produce a number for any NEPSE-listed company you care to study. The question this chapter answers is simple: now that you have the number, what do you actually do with it?

The first and most basic use is screening — plain English: sorting a large pile of candidates into "worth studying further" and "not worth my time," before you spend real hours reading annual reports. A screening filter is a first pass, not a final judgment. Think of the loan officer at a village-level microfinance branch. Before she visits a borrower's home to check the goat shed and the vegetable plot, she checks whether the applicant's basic paperwork clears a minimum bar — citizenship, land ownership proof, no default history with another lender. Candidates who fail that first pass are set aside before anyone spends a day walking to a village. Candidates who pass move on to the real work: the home visit, the character reference, the judgment call.

For your equity sleeve — the actively selected portion of your portfolio you built in Chapter 60, as distinct from the index-tracking or passive core — a Canon Score screen works the same way. A practical rule many disciplined investors use, echoing a real institutional practice: global "quality factor" index providers such as MSCI construct indexes by ranking companies on a composite quality score (built from return on equity, earnings stability, and low balance-sheet leverage) and then including only names above a chosen percentile. They do not claim the score picks winners. They use it to exclude the bottom of the pile before any other analysis happens. You can borrow that discipline at home, on a spreadsheet, for a portfolio of Nepali equities.

Set two thresholds, not one. A "watch list minimum" and a "buy list minimum." A reasonable starting point, calibrated to the 0-100 rubric from Chapter 64:

• Below 45: Exclude. Do not track it as a candidate. If you already own it, treat the low score as a loud signal to open the Chapter 61 review process, covered further in Lesson 66.3.
• 45 to 59: Watch list only. You track the company, you read its quarterly disclosures, but you commit no fresh capital. A rebound above 60 upgrades it; a further slide confirms your decision to stay out.
• 60 to 64: Borderline eligible. Worth a full research file, but position sizes should stay small and conditional, as set out in Lesson 66.2.
• 65 and above: Eligible for standard buy-list treatment, sized according to the table in the next lesson.

Why 60 and 65, and not 50 and 55? Because the seven dimensions of the Canon Score are deliberately weighted toward things that protect capital first — financial strength, governance, and liquidity carry heavy weight before growth and momentum-type factors ever enter the picture. A company scoring in the high 50s on this rubric is not merely "average." It usually means at least one of governance, financial strength, or liquidity is genuinely weak, and weakness in those particular dimensions is precisely the kind of weakness that turns into a permanent loss of capital rather than a temporary paper loss. Setting your buy-list bar at 65 rather than 50 is how you make sure the screen is actually doing its job of cutting away danger, not just cutting away small numbers.

Two caveats belong here, both grounded in what you learned in Chapter 65. First, apply the screen within the sector, not only across the whole market. Hydropower companies, by the nature of NEPSE's own thin secondary market for many run-of-river issuers, will structurally score lower on the Liquidity & Tradability dimension than a large commercial bank, even when the underlying project economics are sound. A hydropower stock scoring 62 overall, with the drag coming entirely from thin trading volume rather than governance or financial weakness, deserves a different read from a bank scoring 62 because its capital adequacy ratio is thin and a director resigned without explanation. The total number is the same; the story behind it is not. Always open the dimension breakdown before you trust the headline figure — a habit Lesson 66.5 will insist on again.

Second, remember that a screening filter answers only "should I look closer," never "should I buy now." A company can clear 70 on the Canon Score and still be priced at a level that makes it a poor purchase this month — the Valuation Reasonableness dimension inside the score captures relative valuation discipline, but it does not replace the entry-timing and margin-of-safety thinking from earlier parts of this book. The filter narrows your universe. It does not write the order ticket.

One more practical point about screening in the Nepali context: your candidate universe is smaller than in a market like India or the United States. NEPSE lists a few hundred companies, heavily concentrated in banks, hydropower, and financial institutions. A strict 65-point cutoff applied blindly across every sector could leave your equity sleeve dominated by only two or three industries, defeating the diversification work you did in Chapter 60. The fix is not to lower the bar. The fix is to build sector-specific watch lists using the Chapter 65 adjustments, so that "the best available manufacturing company" and "the best available bank" are each judged against a fair, sector-appropriate version of the rubric, even while the absolute screening thresholds stay the same for entry into the buy list.

Lesson 66.2 — Tying Position Size to Score — A Worked Sizing Table

A bank credit officer sizing a business loan does not ask only "does this borrower qualify." She asks "how much can this borrower responsibly carry." A borrower with a strong repayment history and stable cash flow might be approved for a much larger loan than a first-time borrower with a thinner file — but even the strongest borrower runs into Nepal Rastra Bank's single-obligor exposure limits, which cap how much of a bank's capital can be lent to any one borrower or group, no matter how creditworthy. Quality raises the ceiling within the loan officer's discretion. It never removes the ceiling itself.

Your Canon Score should work on your portfolio the same way. Chapters 60 and 61 set hard ceilings on how large any single position, and any single sector, is allowed to grow inside your equity sleeve — ceilings that exist to protect you from concentration risk regardless of how good a story a company tells. Nothing in this chapter overrides those ceilings. What the Canon Score adds is a disciplined way to decide where, within that ceiling, a given holding should sit. A company scoring 90 can responsibly be sized close to your maximum single-stock ceiling. A company scoring 66 should sit well below it, even if you are confident in the story, simply because the score itself is telling you the margin of safety is thinner.

Here is a worked sizing table you can adapt directly, assuming — consistent with the concentration framework from Chapter 60 — a single-stock ceiling of roughly 8 percent of the equity sleeve and a single-sector ceiling of roughly 25 to 30 percent:

Two things about this table deserve emphasis, because a table this clean invites lazy, mechanical use — the very trap Lesson 66.5 warns against.

First, the size ranges are ceilings within a band, not entitlements. A company scoring 92 does not automatically deserve the full 8 percent weight on day one. You still build the position in tranches, still respect your entry-price discipline from earlier chapters, and still watch how the position behaves relative to the rest of your sleeve as you add to it. The Canon Score tells you how large the position is allowed to become over time. It does not tell you to buy it all at once.

Second, sector caps sit above single-stock sizing and must be checked every time you add. Suppose your banking-sector exposure is already at 24 percent of the equity sleeve against a 25 percent sector ceiling, and a fifth bank clears the screen with a score of 88. The individual-stock sizing table says you could add up to 8 percent. The sector ceiling says you have only 1 percent of room left. The sector ceiling wins. This is not a flaw in the scoring system — it is exactly the kind of guardrail concentration limits exist to provide, and a high Canon Score was never designed to override it.

There is also a case for sizing down deliberately even within a band, based on which dimensions are driving the score. A company at 78 with an even, balanced profile across all seven dimensions is a different animal from a company at 78 that scores near-perfect on Growth Trajectory and Valuation but merely adequate on Governance & Promoter Behaviour. Both clear the same screening bar. A cautious investor will size the second company toward the bottom of its band's range, not the top, because concentrated strength in growth and cheapness cannot fully substitute for a thinner governance cushion — governance weakness is exactly the kind of risk that shows up suddenly, not gradually, and Chapter 64 weighted it heavily for that reason.

Lesson 66.3 — Score Deterioration as a Sell and Review Trigger

A patient with a chronic condition does not wait for a doctor's advice only when she feels sick. She gets a blood test every few months, even when she feels fine, because the numbers can move before the symptoms do. If her cholesterol or blood sugar jumps sharply between two routine tests, a good doctor does not necessarily prescribe a drastic change immediately — but she does insist on a proper follow-up: repeat the test, ask what changed in diet or medication, rule out a lab error, and only then decide on a course of action. The jump itself is not the diagnosis. It is the trigger that forces a proper look.

This is the model for using score deterioration in your portfolio. A deterioration trigger is a preset drop in a held company's Canon Score, measured between two re-scoring dates, that forces you to open a formal review — not a rule that forces you to sell automatically. The distinction matters enormously. Chapter 61 built a review discipline precisely so that difficult decisions get made with a clear head, on a schedule, rather than in a panic or, just as dangerously, through inertia and denial. Score deterioration is one of the cleanest, most objective ways to trip that review discipline into action.

A workable trigger rule, calibrated to the same rubric from Chapter 64, has two legs:

• A total-score trigger: any drop of 10 points or more in a single re-scoring cycle (typically one quarter), or any drop that moves the holding down a full band in the sizing table from Lesson 66.2 — for example, from Standard Buy List into Conditional/Satellite, or from Conditional into Watch List.
• A single-dimension trigger: a sharp, isolated collapse in the Governance & Promoter Behaviour dimension or the Financial Strength & Profitability dimension, even if the total score has not yet fallen by 10 points, because these two dimensions carry the heaviest weight for a reason — they are the dimensions most associated with permanent capital loss rather than ordinary cyclical wobble.

This two-leg approach borrows directly from how credit analysts at rating agencies operate, and it is worth understanding the parallel because it is genuinely instructive. When a bond issuer's fundamentals weaken, a rating agency frequently does not jump straight to a downgrade. It first places the issuer "on watch" or assigns a "negative outlook" — a formal signal that a review is underway and a downgrade is being actively considered, without committing to one yet. Only after that review concludes does an actual rating change occur. Separately, many global bond funds are mandated to hold only "investment-grade" debt; the moment an issuer's rating is cut below the investment-grade cutoff (from BBB- to BB+, for instance — a bond that crosses this line is informally called a "fallen angel"), those funds are contractually forced to sell, regardless of their own view of the company's prospects, simply because the mandate says so. Your Canon Score deterioration trigger should behave like the "negative outlook" step, not the fallen-angel step: it forces a disciplined, unhurried review. It does not, by itself, force a sale. You are not running a fund bound by someone else's mandate. You are the analyst doing the review — the whole point of building your own scoring system in Chapter 64 was to keep that judgment in your own hands.

Once the trigger fires, the review itself should ask one central question: is this deterioration temporary and cyclical, or is it structural? A hydropower company's Financial Strength score can dip sharply after one poor monsoon season and lower-than-expected generation — a real event, but one that reverses the following year if the underlying asset and its power purchase agreement are sound. That is cyclical. A bank whose Governance score drops because an independent director resigned abruptly without explanation, followed by a qualified audit opinion on loan loss provisioning, is showing something structural — a warning about the quality of information you are being given, not a one-off bad quarter. The first case argues for patience, possibly even for holding through the dip if the position size was already sensible. The second argues for trimming toward the exit, following the sell discipline built in Chapter 61.

Two guardrails prevent this trigger from being misused, in either direction.

The first guardrail is against denial: do not "explain away" every deterioration trigger just because you like the company or have held it a long time. The whole reason to build a preset numerical trigger, rather than relying on gut feel about when to review a holding, is that gut feel tends to rationalise staying in a familiar, comfortable position exactly when the discipline should be at its sharpest. If your own review keeps concluding "it's fine, nothing to see here" quarter after quarter while the score keeps drifting lower, that pattern is itself useful information about your objectivity, not just about the company.

The second guardrail is against overreaction: do not treat every trigger as an automatic sell. A single disappointing quarter that trips the 10-point threshold, followed by a review that turns up a plausible and temporary explanation, is exactly the situation this two-step process — trigger, then review — was designed to handle sensibly. Selling reflexively on every score wobble converts a long-term investment discipline into short-term, cost-heavy trading, and NEPSE's brokerage costs and settlement mechanics make that an expensive habit.

Lesson 66.4 — Building a Re-Scoring Calendar and Habit

A score you calculated once, at the time of purchase, and never revisited is not a risk management tool. It is a museum piece — accurate the day it was made, and steadily less true every day after. Chapter 62 built a systematic investing calendar so that contributions, rebalancing, and reviews happen on a schedule rather than whenever mood or memory permits. Re-scoring needs the same calendar discipline, because the entire value of the deterioration trigger in Lesson 66.3 depends on comparing scores measured at regular, predictable intervals — a comparison that is meaningless if some holdings get re-scored every month and others get re-scored only when something goes wrong.

Anchor the re-scoring calendar to events that already happen on a fixed rhythm in the Nepali market, so the habit rides on infrastructure that already exists rather than competing with it for your attention.

Most NEPSE-listed companies publish unaudited quarterly financial statements within roughly a month of each quarter-end, in line with SEBON's disclosure timeline. That gives you four natural re-scoring windows a year, one after each quarter's unaudited results land. A full re-score of every holding and every active watch-list company, using the fresh quarterly numbers, should happen in each of these four windows. This is also the natural point to update the spreadsheet described in Lesson 66.2 and to check every trigger from Lesson 66.3.

Layer a deeper annual re-score on top of the quarterly cycle, timed to each company's audited annual report and AGM — typically clustered in the Poush-to-Falgun window for many Nepali companies, when audited financials, the directors' report, dividend and bonus share declarations, and any changes to the board are all disclosed together. This is the richest data drop of the year for the Governance & Promoter Behaviour and Dividend & Capital Return Discipline dimensions in particular, since both depend heavily on information — related-party transactions, auditor's notes, AGM resolutions — that only appears in the full annual filing, not the abbreviated quarterly one.

Add a third, irregular layer: ad hoc re-scoring triggered by material news, whenever it occurs, rather than waiting for the next scheduled window. A rights issue announcement, an auditor change, a promoter share pledge or its release, an NRB directive affecting a bank or development bank's capital or provisioning requirements, a credit rating action by a domestic agency such as ICRA Nepal or CARE Ratings Nepal, or a hydropower company's PPA renegotiation with the Nepal Electricity Authority — any of these can shift a Canon Score meaningfully before the next quarterly window arrives, and none of them should wait three months for a look.

One habit compounds the value of re-scoring more than any other: keep a score history, not just a current score. A single Canon Score snapshot tells you where a company stands today. A score history — the same company's score at each of the last six or eight quarterly windows, laid out in a simple line — tells you the direction of travel, and direction is often more informative than level. A company holding steady at 68 for two years is a very different holding from a company that fell from 84 to 68 over the same period, even though both show a score of 68 today. The first looks like a stable, moderate-quality business. The second looks like a business in the middle of a genuine decline that simply has not yet crossed your deterioration trigger threshold. Chapter 61's review discipline works far better when it can see the trend line, not just the last data point.

Finally, treat the Dashain-Tihar period the way Chapter 62 already treats it for contribution scheduling — as a natural pause point, not a working period. Most Nepali households and many company secretariats slow down markedly during this stretch. Rather than fighting that rhythm, build your annual deep-review date either just before Dashain, using the most recent quarterly data as a stocktake before the festival season, or just after Tihar, once markets and company activity have resumed their normal pace. Fighting the calendar you actually live in is a losing habit; working with it is how a systematic discipline survives for years rather than fizzling out after two quarters.

Lesson 66.5 — Honest Limitations — What the Score Cannot Do

Chapter 63 opened Part XIII with a warning worth repeating here at the close, because a well-built tool is exactly the kind of thing investors are tempted to trust too much. The Canon Score is a structure for judgment. It is not a substitute for judgment, and it was never designed to be one. A student who tops the entrance exam still has to survive medical school, still has to actually learn to treat patients — the exam score predicted readiness, it did not deliver competence. The Canon Score works the same way: it predicts which companies deserve your attention and roughly how much capital they can responsibly carry. It cannot do the remaining work of being a careful, skeptical investor.

Six honest limitations deserve to be stated plainly, precisely because a numerical score invites false confidence.

First, the score is only as good as the disclosure it is built from, and disclosure quality in the Nepali market is genuinely uneven. A large commercial bank with decades of listed history, professional investor-relations practice, and heavy analyst coverage gives you rich, comparable data. A recently listed hydropower company, three years into commercial operation, gives you a much thinner run of audited numbers, and a smaller manufacturing or hospitality company may disclose the bare regulatory minimum and nothing more. The same numeric score built on five years of clean data and the same score built on eighteen months of thin data are not equally trustworthy, even though the spreadsheet shows the same figure.

Second, the score cannot fully price crisis-level liquidity risk. The Liquidity & Tradability dimension measures ordinary trading conditions — average daily turnover, bid-ask spread, free float. In a genuine market panic, NEPSE's circuit breakers and thin order books mean that even stocks with historically decent liquidity scores can gap down with no buyers at any price for several sessions in a row. A liquidity score built on calm-market data quietly assumes calm markets continue. They do not always.

Third, the score cannot anticipate sudden regulatory or political shifts. An NRB monetary tightening cycle, a change to loan loss provisioning norms, a hydropower PPA renegotiation, or a shift in remittance-linked banking regulation can all move a company's prospects overnight, in ways no backward-looking rubric could have flagged the prior quarter. The score updates only as fast as your re-scoring calendar runs; regulation and politics do not wait for your calendar.

Fourth, a rising score built on one or two exceptionally strong quarters can be mistaken for durable improvement when it is really a temporary spike — a bumper hydropower generation year from unusually heavy monsoon rainfall, or a one-off gain from a land or asset sale flattering a bank's profitability ratios for a single period. The Growth Trajectory and Financial Strength dimensions can both look temporarily excellent on numbers that will not repeat. Read the notes to the financial statements, not just the ratios, before trusting a sudden jump.

Fifth, remember that the qualitative dimensions — governance chief among them — still involve human judgment, and human judgment carries bias. Two careful investors scoring the same board composition and the same related-party disclosure can reasonably land a few points apart. The score creates useful discipline and a shared vocabulary; it does not create false precision. Treat a Canon Score of 71 versus 74 as "roughly similar quality," not as a meaningfully different verdict.

Sixth, and most important: the score cannot make the final call for you, and it should not be allowed to. It structures your attention, ranks your candidates, sizes your positions, and flags deterioration for review. The decision to buy, to hold through a difficult quarter, or to exit — that remains yours, informed by everything the score cannot capture: a phone call with someone who knows the company, a site visit to a hydropower project, a careful read of an auditor's qualified opinion, or simply a mature judgment about whether Nepal-specific risk in a given sector has quietly changed in a way no rubric yet reflects.

Lesson 66.6 — Closing Synthesis of Part XIII and a Worked End-to-End Example

Walk through a complete, worked example that ties Lessons 66.1 through 66.5 together, following one investor's equity sleeve through a full year.

Sabina, a Kathmandu-based investor, runs an equity sleeve of roughly NPR 15 lakh, built under the Chapter 60 framework with an 8 percent single-stock ceiling and a 28 percent single-sector ceiling. She maintains a Canon Score spreadsheet, re-scored every quarter under the Chapter 66.4 calendar, covering four current holdings and three watch-list candidates.

At the start of the fiscal year, her holdings and their most recent scores look like this:

Two quarters in, Sabina's scheduled re-score turns up a material change. Everest Trust Bank's score falls from 71 to 56 — an 18-point drop, crossing two bands at once, tripping the total-score deterioration trigger from Lesson 66.3. On review, she finds the cause: a sharp rise in non-performing loans in one regional branch cluster, combined with a mid-cycle change of statutory auditor announced outside the normal AGM process. Applying the temporary-versus-structural test from Lesson 66.3, both signals point toward structural concern rather than a single bad quarter — asset quality deterioration concentrated in one lending pocket, paired with an unexplained auditor change, is the kind of combination that historically does not reverse quickly. Following Chapter 61's review discipline, she trims the position from 5 percent down to 2 percent over the following month, in line with the Watch List sizing guidance from Lesson 66.2, rather than exiting all at once or holding the full position hoping for a recovery.

At the same re-scoring window, Sagarmatha Laghubitta — previously a Watch List candidate at 58 — improves to 64 after a strong quarterly disclosure showing tightened loan-loss provisioning and a clean AGM with no governance red flags. It now clears the Conditional/Satellite screening bar from Lesson 66.1. Sabina opens a small 1.5 percent starter position, sized toward the bottom of its band per the sizing table, consistent with a single quarter of improvement rather than a long track record.

By year-end, her equity sleeve reflects a portfolio actively shaped by the screen, the sizing table, and the deterioration trigger working together — not a static buy-and-forget list, and not a portfolio churned by every quarterly wobble, but one adjusted with the same unhurried discipline a credit analyst would apply to a loan book under periodic review.

This worked example is really the closing argument for all of Part XIII. Chapter 63 established why a scoring system belongs in a Nepali retail investor's toolkit at all — because judgment applied inconsistently, company by company, mood by mood, is judgment that eventually fails you, and a structured rubric is how you make your own analysis repeatable and honest with yourself. Chapter 64 built the actual instrument: seven dimensions, weighted deliberately toward capital protection before growth, producing one comparable number across very different businesses. Chapter 65 admitted that no single generic rubric fits every sector equally well, and adjusted the weights and thresholds for the structural realities of banks, hydropower, microfinance, and insurance — recognising, for instance, that a hydropower company's cash flows run on hydrology and PPA terms, not the same drivers as a bank's net interest margin. This chapter closes the loop by putting the finished score to work exactly where it belongs: screening candidates before they earn your research time, sizing positions within the hard concentration ceilings from Chapter 60, triggering — never dictating — the review discipline from Chapter 61, and living on a calendar borrowed from Chapter 62's systematic rhythm, rather than sitting idle until a crisis forces a look.