A rules-based public-equity strategy — audited across twenty-nine market years on file, then deployed with personal capital. This page is the story of how it was built and what the first live year did.
I am Abdulmajid Al Qurashi. By day, Deputy CEO of a private business with four hundred stores in sixteen countries. The rest of the time, I built this.
Before any of this, the question was the one every retail investor asks. The advice from Bogle to Buffett's own letter to his executors converges on the same answer: just hold the index. Beating the market consistently is hard, the data say you will not, and the cheapest passive fund will quietly outperform most people who try. So the honest first question was — should I bother?
What I noticed when I looked closely was that the index is an average of two populations. Some stocks compounded steadily for years. Others bled for a decade. The S&P 500's return is what survives the cancellation between them. The real question, then, was not can I beat the market. It was — can I filter the index, keep the names whose arithmetic is on the rising side, drop the ones bleeding, and let the average shift?
That is what sent me to the books. ↓
Fifty books. Forty thousand pages. The famous authors did not agree on a single number.
Graham, Buffett, Klarman, and the dozens of practitioners I read alongside them used the same vocabulary — margin of safety, intrinsic value, quality — and meant materially different numbers by it. Even Buffett had quietly reversed his own framework half a career in. I am trained as a civil engineer. A bridge that holds in one drawing and falls in the next is not engineering. It is opinion.
The plan was not to pick an author. An author is only as good as the decade he wrote in. The plan was to specify every rule the literature argued about, then test each one mechanically against every market year on file, and keep only the rule that survived.
Is the company growing faster than the market has priced in? Not faster than its peers. Faster than its own share price already assumes. The rule reads operating numbers — revenue, earnings, cash from operations — and asks whether the trajectory is steeper than the implied curve in the current quote.
Is it trading below a conservative estimate of what it is actually worth? Margin of safety in the engineering sense — built into the design, not hoped for after the fact. Graham borrowed the phrase from structural engineering for a reason; the rule borrows the discipline.
Can the balance sheet survive a bad year? Cash on hand, interest coverage, debt that does not break under stress. The company has to still be there in twelve months — because the rule does not trade out of a position when it is wrong, it holds.
Most years, across thousands of US-listed candidates, only a handful of businesses pass all three filters at once. The exact formulas and numeric thresholds remain proprietary — the philosophy is public, the mechanism stays inside.
The cadence is fixed. Five to seven names each April, equal weight, twelve-month hold. No trading, no override, no adjustment in between. The rule does not respond to news. It does not respond to drawdowns. It does not respond to the operator changing his mind in October. The whole point is that nothing in the discretionary tradition is allowed to enter the loop.
The rule runs across every US-listed name. Five to seven pass. They are bought, equal weight, on day one of April.
The seven names are published the day they are bought. No private picks. The record is open from minute one.
The portfolio is marked against the S&P 500 every market day on the live ledger. The result forms in public, in real time.
Twelve months later the positions are closed. The result enters the historical record. Then the rule runs again.
There is nothing to sign up for and nothing for sale. The site you are reading is the record — published in public so it cannot be quietly rewritten if a year goes badly.
The Arabic word for the merchant's balance — an instrument as old as commerce itself, by which traders in every old culture weighed gold and goods. It is also, plainly, what a rule-based system is: a set of criteria that weigh a candidate and produce a decision. The operator does not overrule the scales. The scales are set; one reads what they say.
Six positive, one contained at −24%. No trading, no override, no adjustment.
Price return over the identical twelve-month window — the default benchmark for US equity performance.
Selected by the rule on 1 April 2025, equal weight, held to 1 April 2026 with no trading. See each line on the live ledger →
Year One landed at +68.92% — above the rule's historical median, well below the +83.99% mean (which the +394.5% recovery vintage of 2020 pulls upward). Inside the distribution, not above it. The full backtest, every year, is on the Results page.
One year is a single data point. It is not proof the rule works. It is evidence that the first live vintage behaved consistent with what the twenty-nine-year backtest predicts for a median-to-good year — nothing more, nothing less. Six of the seven names finished positive. One closed at −24% and the rule held it through the full twelve months without override. The full year-by-year backtest, including drawdowns, hit rates, and worst-case stress tests, is on the Results page.
A short detour from the live ledger, into the historical record. These are three illustrative picks the rule surfaced across twenty-nine audited years — chosen here to show the same arithmetic working in very different markets. None of these are Year One holdings.
The iPhone did not yet exist. Streaming was a footnote in an annual report. The rule does not read stories — it read three numbers, and the three numbers lined up.
Most investors were watching the megacaps at the opening of the COVID dislocation. The rule found a smaller business growing through the shock, trading below the math.
Each time the operating numbers had moved ahead of where the market had priced the stock. When the math re-opened, the rule re-entered. When it closed, the rule left.
A name that looked correct on every filter in April 2025 deteriorated through the summer and never recovered. The rule does not trade out of a position when it is wrong. It does not respond to drawdowns. It does not respond to news. The position was held the full twelve months and closed in April 2026 at the loss it had earned. This is the discipline the method requires — and the discipline the operator must keep, in the months when the market is telling him he is wrong, for the rule to mean anything at all.
A careful reader should be asking what could break this. I have asked it too, and this is the honest answer.
One live year is a single observation. Five live years will start to be a signal. Ten will be a record. Until then, the most defensible thing I can claim about the rule is that its first live vintage behaved consistent with what the twenty-nine-year backtest predicted — nothing stronger.
The twenty-nine-year backtest, however carefully constructed, is still a reconstruction of the past. The past carries selection bias I cannot fully eliminate: companies that went bankrupt between 1995 and now are not in every database with the same completeness as the survivors. I have corrected for survivorship where I could, declared the correction in the methodology where I applied it, and assumed the result is slightly overstated where I could not.
The rule itself may have over-fit some subtle feature of the audited window. The 2020s do not look like the 2000s, which did not look like the late 1990s. No amount of stress-testing reveals over-fit until the rule is run, unchanged, into a regime nobody has yet seen. The mitigation is not in the audit. It is in the public live record — each closed year an observation against an unrevised rule.
The year-end numbers also hide the path. The backtest’s worst full year was −0.57% — but that is an artefact of measuring once a year. The daily reconstruction shows the concentrated rule fell roughly 53% intra-year at the depth of the 2020 shock before recovering and finishing the year far ahead. A real holder lived through that decline. Anyone who could not have held through a 53% drawdown could not have earned the return.
And the most consequential caveat of all is the human one. This method does not work if I lose the discipline to hold a position when the market is telling me, for nine months out of twelve, that I am wrong. The rule is mechanical; the operator is not. Every year that the rule produces a contained loss is a year the operator has to sit through it without acting. That is the test the method actually faces, and it is not a test the backtest can run.
These four limits are why the work is published the day it runs. Selections are time-stamped before they perform; positions are marked daily; the closing return enters the record at year-end whether it flatters the rule or embarrasses it. The author cannot quietly revise what is already public.
— A.A.Q., April 2026
Seven new names, selected on 1 April 2026 by the same rule. Equal weight. Held for twelve months without adjustment. Marked against the S&P 500 every market day.
This year's seven names are committed at entry and disclosed a year later, once the year has closed — what you see live is the aggregate return, never a live tip. The full named record for the most recent closed year is on the live ledger. Aggregate as of —; equal-weighted at inception and held without rebalancing for the full year.
There is no subscription, no mailing list, no paid access. This site is a record — of a question I spent five years trying to answer honestly, of the method I arrived at, and of what happens each year when the rule meets the market.
If you want to interrogate the research, the full method and the twenty-nine-year backtest are one click away. If you want to watch the experiment as it runs, the live ledger updates every market day. If you would like to write, I am reachable directly.
Mīzān is a personal investment-research record published by Abdulmajid Al Qurashi for informational purposes only. It is not investment advice, not a solicitation, and not a recommendation to buy, sell, or hold any security. The author is not a licensed financial adviser, and this site is not a regulated financial product. There is nothing for sale and nothing to subscribe to.
Past performance — whether backtested or realised — is not indicative of future results, and all investing carries risk, including the loss of principal. Anyone making investment decisions should consult a qualified, licensed professional in their own jurisdiction.
The live record reflects only the author's own capital and his own decisions.