Specification sheet Ref OE‑001 Chapter: Optimism, practiced
optimism
engineering.
A good outcome is a structure, not a feeling about a structure.
Premise
Scope of worksOptimism is usually filed under temperament. Some people have it, some do not, and the ones who do are held to be either lucky or naive.
Filed that way, it cannot be taught, cannot be checked, and cannot be relied on. It arrives when the weather is fine and leaves when the news is bad. Nothing you would put weight on.
The alternative is to treat the good outcome the way you would treat a bridge. It has a span. It carries load. It has a set of conditions under which it stands and a set under which it comes down, and both sets can be written out in advance by anyone willing to sit with the drawing long enough.
That reframing costs you the comfort of the mood and hands you something better: a list of things to do on Monday.
Load path
Fig. 1 / not to scaleRemove any one pier and the deck still stands, for a while, which is why the failure is hard to see coming. Remove two and the span becomes a mood again: something you feel about the outcome rather than something holding it up.
Failure modes
Observed in service| Mode | Presents as | Root cause | Remedy |
|---|---|---|---|
| Brittle | Total collapse on first contradicting fact | Claim was a wish, never load tested | State the falsifier before you need it |
| Diffuse | Everyone agrees it matters, nobody moves | No named owner on the pier | One name per load path |
| Blind | Confidence holds while conditions rot | Feedback loop longer than the decision cycle | Shorten the loop, not the report |
| Overtight | Small shock cascades to full stop | Zero slack; plan assumed the median case | Buy margin before you need margin |
| Ornamental | Vision documents accumulate; nothing ships | Optimism spent on the brief, not the footings | Convert one claim to one dated commitment |
| Ductile | Bends visibly, keeps carrying, gets repaired | Designed for the bad week, not the good one | Nominal condition. Continue. |
Method
Repeat until built-
Name the outcome in units
Not better, not sustainable, not thriving. A number, a date, a place, a person who would notice. If it cannot be stated in units, it cannot be built, only admired.
-
Write the failure first
Before the plan, the postmortem. Assume it went badly and describe exactly how. The list you produce in ten minutes is the specification for everything you have to defend against.
-
Find the binding constraint
One thing is holding the whole span. It is rarely the thing that is loudest. Effort spent anywhere else is decoration, however satisfying it feels while you are spending it.
-
Reduce the span
A decade is not buildable. A quarter is. Cut the distance between now and evidence until the gap is short enough that being wrong is cheap and being right is visible.
-
Instrument it
Decide in advance what would change your mind, and where that signal will show up. Undecided in advance means never, because by then you will have reasons.
-
Load it and watch
Put real weight on it early, while the failure is still small and survivable. A structure that has never carried anything is not proven safe. It is only untested.
Factor of safety
Design allowancesCertainty required
Low. You do not need to know it will work. You need to know what you will do when it does not.
Evidence required
High. Cheerfulness is free and proves nothing. The bar is the same bar you would hold anyone else to.
Margin held back
Substantial. Optimism without slack is a plan that only works on the days you were already going to be fine.
The pessimist and the optimist are both making a forecast. Only one of them is obliged to pick up a tool.
Sheet OE‑001 · issued for construction · supersedes all moods