Lesson 8 of 8
Three rails for one company
The finished set. Three conditions on one named filer, each level derived from a reason rather than read off the history, each with its review date attached.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- The output of the whole bridge is three or four written conditions on one company, not an opinion about credit.
- Two of the three here are automated and one is checked by hand, which is the honest answer at this filer.
- Every level is derived from a reason, and the reason is written on the same line so a reader in two years can see why.
Five courses end here. This is the finished document for one company, written out in full, with the reasoning visible beside each level.
One filer's position, from its latest annual report
- Operating cash flow to debt0.95x
- Debt to equity0.34x
- Net debt to EBITDA0.48x
The reading this follows from
At this filer, as of the fiscal 2025 annual report: nothing falls due in the first two rungs of the five-year maturity schedule, total debt is 18,396 million dollars against cash of 9,642 million, and operating cash flow for the year was 17,525 million. Fixed asset turnover is 0.80 times and capital expenditure was 42.42 percent of revenue.
So the load-bearing question here is not a dated obligation. It is whether cash generation holds up well enough to fund a capital programme of that size and service the borrowings, at a company whose profit is a geared version of its sales.
Rail one: cash generation against debt
Figure. Operating cash flow to debt. Direction. At or above. Level. 0.55.
Why that level. The same ratio was 0.46 in the fiscal year ended 29 August 2024 and 0.95 in the year ended 28 August 2025. The reason for holding is that the business generates enough cash through a cycle to service and retire its borrowings without needing the credit market. A level set between the two filed readings tests that reason: it sits above the weaker year, so it would have fired then, and below the stronger year, so it is not fired by ordinary variation at the current level.
Review. Each quarterly and annual filing. Automated. Yes.
Rail two: the capital programme against the cash it needs
Figure. Capital expenditure to operating cash flow. Direction. At or below. Level. 150 percent.
Why that level. The filed readings are 79 percent in fiscal 2022, 492 percent in fiscal 2023, 99 percent in fiscal 2024 and 90 percent in fiscal 2025. Three of the four sit under 100 percent and one, a trough year, was five times that. A level of 150 percent is above ordinary variation and well below the trough reading, so it separates a company funding its programme from a company not funding it, which is the distinction the reason cares about.
What a breach would establish. That the year's capital programme was not funded from the year's cash. The third lesson of the capital intensity course explains why that happens structurally at this kind of filer, so a breach here is expected in a trough and is not by itself evidence that the reason failed.
Review. Each annual filing, because capital spending is lumpy and a quarter is noise. Automated. Yes.
Rail three: the maturity schedule
Figure. The five-year maturity schedule. Direction. None. Trigger. The first rung ceasing to be zero without a source of funds identified.
Why no level. The fifth lesson of this course: a condition written on a level cannot see a date, and here the relevant fact is that the first two rungs are currently zero. Total debt could rise without that changing, and the first rung could become material without total debt moving.
Review. The annual filing, by hand. Automated. No, and the thesis says so, because a rail with no backward test and no automation is one a reader in two years must not mistake for a monitored one.
What this document is, and is not
It is a record of why a position is held and what would refute it, in figures a stranger can check against a named filing. It contains no view about credit conditions, no date, and no statement about any price.
It is not a valuation and it is not a decision. The capstones do the first and nobody here does the second.
Check yourself
Why is a level set between two filed readings better than one set below both?
- It fires more often, which gives more information
- Because it would have fired in the weaker year and not in the stronger one, which is the minimum evidence that it tests anything
- Because it matches the peer median
Show the answer
Answer: B.
A level below both would never have fired across the available history, which the retro-test lesson identifies as a reason to change it before relying on it.Why does the capital expenditure rail carry a note about expected breaches?
- Because it is not automated
- Because a capital-intensive filer predictably spends more than it generates in a trough, so a breach can be read against that expectation rather than against nothing
- Because the level was chosen from the peer group
Show the answer
Answer: B.
A condition whose expected behaviour is written down in advance is more useful than one whose behaviour is a surprise.Why does the thesis record that the third rail is not automated?
- To explain why it has no level
- So that a reader in two years does not mistake an unmonitored condition for a monitored one and read its silence as evidence
- Because manual rails are reviewed less often
Show the answer
Answer: B.
A holding rail and an unchecked rail produce the same silence, which is the most expensive of the four wrong readings of a quiet inbox.Sources
- Micron Technology, Inc., annual reports on Form 10-K for fiscal years 2022 through 2025, for debt, cash, equity, operating cash flow, capital expenditure and the five-year maturity schedules: fiscal 2022 accession 0000723125-22-000048; fiscal 2023 accession 0000723125-23-000054; fiscal 2024 accession 0000723125-24-000027; fiscal 2025 accession 0000723125-25-000028.
What this leaves out
- This is a heterodox tradition. The mainstream account of the same phenomenon differs, and nothing here is a forecast or a view on any security.
- The three conditions are written to show how a level is derived from a reason. They are not conditions anybody should adopt, the reason attached to them is a stated example rather than anyone's actual reason, and a different reason produces different levels from the same filings.
- Two filed years is the minimum evidence a derived level can rest on and it is thin. A longer window is better wherever the filed history supports one.
- Nothing here is a view on Micron or on any security.