The weekly letter · always free
One stock, valued the Graham way — in your inbox every week.
One North American stock a week through Benjamin Graham's checks, in plain English: is it cheap, which Canadian account to hold it in, and the honest verdict. Free. No spam.
What lands in your inbox
The classic Graham checks — P/E, price-to-book, margin of safety — with a clear cheap / fair / expensive call. Shown work, not a black box.
The Canadian part the big US newsletters skip: whether a stock belongs in your TFSA, RRSP, or non-registered account — and the tax reason why.
Including what's wrong with it. You get the full picture — flaws, flags, and all — not a highlight reel or a hype pitch.
Written by Curtis Travis — B.Comm, AACI (retired).
I trained and worked as a professional appraiser early in my career — putting a defensible value on an asset was the job. Now I bring that same discipline to stocks. Every issue I have written is below, unedited.
The record
Every issue so far, and what happened next
Read the back issues before you subscribe. Each row keeps the letter exactly as it was emailed and — from Issue #4 on — the workfile behind it: the full appraisal, every tool run, every warning sign printed rather than the flattering ones.
What this is, and what it isn't
- There is no back-test here. This app is new. Every row was published, dated, before anyone knew the outcome — and nothing has been added retroactively.
- Not every issue is a buy. Some said wait. Those are marked, because scoring a “this is too expensive” issue as though it were a recommendation would be dishonest.
- These are not returns. Nobody managed money here. It is a record of what was written and what the price did afterwards — a few weeks of which proves nothing about anything. Judge the reasoning, not the arithmetic.
Read this slowly. It is supposed to be slow.
Value investing works on a timescale that makes a table like this almost useless in the short run. Graham's framework assumes years, not weeks. A few percent either way over a month tells you nothing about whether the reasoning was sound — it tells you the market changed its mood, which it does daily and for no reason you can act on.
It follows that a falling price is not automatically bad news. If the business is still sound and you intend to keep buying, a lower price is the whole point. Buffett put it as a question in his 1997 letter to Berkshire shareholders: if you plan to eat hamburgers for years and don't own a cattle ranch, do you want beef to get more expensive or cheaper? Most people answer sensibly about lunch and get it exactly backwards about shares. His shorter version: “Be fearful when others are greedy and greedy when others are fearful.”
So the honest instruction is: don't score this page. Read what each issue argued, decide whether the argument holds, and check back in a few years. If a name here is down and the reasoning still stands, that is the situation Graham spent his career trying to find.
| Issue | Company | The call | Price then | Price now | Change |
|---|---|---|---|---|---|
#9 Oct 11, 2026 | Commercial Metals Company CMC A steel company with 247 straight quarterly dividends, priced at about 12 to 14 times earnings against 20 for Nucor, after borrowing $2 billion to buy its way into precast concrete. New debt, a tax credit that flatters earnings and a steel cycle are the catch; its October 15 report is the first test. | Looked cheap | $63.89 | $63.89 Oct 9, 2026 close | +0.0% |
#8 Oct 4, 2026 | EQT Corporation EQT The largest US natural gas producer, priced at 11 times earnings while the copper and uranium stocks chasing the same AI-power-demand story trade at 19 to 148 times. The CEO sold $15.0M in stock this year — the SEC filing shows the sales ran on a trading plan set up five months earlier. | Looked cheap | $50.17 | $52.83 Oct 9, 2026 close | +5.3% |
#7 Sep 27, 2026 | Grupo Aeroportuario del Centro Norte OMAB A regulated airport monopoly that fell 32% in a week when the Mexican government cut its fees in 2023 — and has earned 46%+ on equity every year since. The protection is monopoly cash flow: nobody can build a competing airport next to Monterrey's. Disclosure: Curtis held this position when the issue went out. | Contrarian bet | $103.03 | $90.28 Oct 9, 2026 close | -12.4% |
#6 Sep 20, 2026 | Fairfax Financial FFH.TO Cheap on the measures that fit an insurer — 1.25x book on a 16.9% return on equity, 7.4x earnings — with softening insurance prices as the main risk. A call on price, not on where rates go next. | Looked cheap | CA$2,244.95 | CA$2,250.89 Oct 9, 2026 close | +0.3% |
#5 Sep 13, 2026 | Lennar LEN Below book at 0.89x, with Berkshire buying more while underwater — the same quarter it bought a rival builder outright for $6.8B. A contrarian bet on housing, not a Graham bargain. Disclosure: Curtis held this position when the issue went out. | Contrarian bet | $79.60 | $76.69 Oct 9, 2026 close | -3.7% |
#4 Sep 6, 2026 | Dick's Sporting Goods DKS A contrarian turnaround after a 40%+ fall, backed by a cluster of insider buying — explicitly not a low-risk Graham bargain. Disclosure: Curtis held this position when the issue went out. | Contrarian bet | $139.15 | $136.65 Oct 9, 2026 close | -1.8% |
#3 Aug 30, 2026 | Stifel Financial SF Cheap once you read a bank correctly — the distress and net-net screens misfire on financials, and the app says so on the page. | Looked cheap | $81.29 | $70.34 Oct 9, 2026 close | -13.5% |
#2 Aug 23, 2026 | Ingredion INGR Genuinely cheap and high quality — Piotroski 8/9, ~21% under the Graham Number — with lukewarm insider selling as the catch. | Looked cheap | $106.84 | $95.09 Oct 9, 2026 close | -11.0% |
#1 Aug 19, 2026 | Canadian Tire CTC-A.TO Above fair value on the Graham Number, with no margin of safety. Not a buy at this price. | Said wait | CA$202.00 | CA$192.44 Oct 9, 2026 close | -4.7% |
9 issues published so far. Price now is the most recent closing price, recorded by hand with its date beneath it and refreshed with each Sunday issue. Each is shown in the stock's own currency and is not converted, and the change is price only, with dividends excluded. Curtis Travis is a retired AACI appraiser, not a licensed investment advisor — this is a record of published writing, not advice, and not a solicitation to buy anything.
The plan
What happens to all of this
The reason for writing an appraisal down and dating it is that you can be held to it afterwards. So the plan is to come back to every one of these in public, on a clock that was set before anybody knew the outcome.
Now
2026 →
One issue a week. Each keeps its letter, and from Issue #4 on, a workfile — the full appraisal behind the pick, every tool run, every warning sign printed rather than the flattering ones.
8 months
spring 2027
The first look back. Long enough that the argument has been tested, short enough that I still remember precisely why I made it — the only window in which anyone is honest about their own reasoning.
8 years
2034
Long enough for a business to have been through a full cycle, a recession and at least one management change. The first horizon on which a value argument can fairly be marked right or wrong.
18 years
2044
A generation. Some of the companies here will have been taken over, broken up, or gone under. Those are the entries worth the most — a record that keeps only the survivors teaches nothing.
28 years
2054
By then the arithmetic has settled every argument on this page, and luck has had long enough to cancel itself out. The letters will still say exactly what they said in 2026, because nothing on this site gets edited after the fact.
Between those, there is a shorter loop. Every workfile names, in writing and in advance, the one thing that would prove it wrong — and where that test was itself wrong on the first pass, the file says so rather than swapping in a better one. When the named thing happens, the workfile gets a dated addendum underneath the original text and the original stays put. A report that quietly updates itself is a brochure. A marked-up one is evidence.
And the point of it is the method, not the verdict. Every issue runs the same tools in the same order on a different company and shows the working, so that after a year of reading them you can do this without me. That is the whole idea, and it is why the letters stay up and the workfiles are free to anyone who wants the long version.
The intervals are borrowed from a position-sizing rule I have used for years — 8, 18, 28. As dates they forecast nothing. What they are is long, and fixed in advance, which is the only property that matters.
Next issue goes out Sunday. It is free, and it is the same work you have just read.
Prefer to poke around first? Try the free valuation tool →